1. ABSTRACT

Matrix vesicles (MVs) induce calcification during endochondral bone formation. Experimental methods for structural, compositional, and functional analysis of MVs are reviewed. MV proteins, enzymes, receptors, transporters, regulators, lipids and electrolytes are detailed. MV formation is considered from both structural and biochemical perspectives. Confocal imaging of Ca²⁺ and H⁺ were used to depict how living chondrocytes form MVs. Biochemical studies revealed that coordinated mitochondrial Ca²⁺ and Pi metabolism produce MVs containing a nucleational complex (NC) of amorphous calcium phosphate, phosphatidylserine and annexin A5 - all critical to the mechanism of mineral nucleation. Reconstitution of the NC and modeling with unilamellar vesicles reveal how the NC transforms into octacalcium phosphate, regulated by Mg²⁺, Zn²⁺ and annexin A5. Extravasation of intravesicular mineral is mediated by phospholipases and tissue-nonspecific alkaline phosphatase (TNAP). In the extravesicular matrix, hydroxyapatite crystal propagation is enhanced by cartilage collagens and TNAP, which destroys inhibitory PPi, and by metalloproteases that degrade proteoglycans. Other proteins also modulate mineral formation. Recent findings from single and multiple gene knockouts of TNAP, NPP1, ANK, PHOSPHO1, and Annexin A5 are reviewed.

2. INTRODUCTION

Skeletal formation is a highly complex process that involves cellular proliferation and differentiation - like development of all tissues - but superimposed on this is the need for mineralization to enable bone to physically support bodily functions. There are two major types of bone formation: endochondral and intramembranous. Endochondral bone formation begins as a cartilaginous anlage and is observed during formation of long bones such as those of the limbs and ribs. Intramembranous bone formation is characteristic of "flat" bones such as the skull and vertebra. Within each type of bone formation there are variations depending on the species and the nature of the bone being considered. Skeletal development involves growth, expansion and remodeling of the bony structure. Specialized types of cells are involved at each stage.

While there are many commonalities in the mechanism of mineral deposition in various species and genera within the vertebrate order, there are also significant variations depending on size and rate of skeletal formation. Recently, for assessment of the roles of the various proteins and enzymes involved, skeletal research has focused increasingly on the use of gene deletions of putative key targets in highly-inbred laboratory mice. Of necessity, many heterozygous characteristics of native wildtype animals have had to be eliminated to produce the genetic homogeneity required for knockout studies. To ensure reproducible results, these laboratory animals must be reared in a highly protective environment with optimal nutrition, temperature and cleanliness, and with minimal stress. For direct biochemical analyses of growth plate mineralization, fetal bovine tissues were initially used; however, much of the subsequent research utilized developing bones of hybrid, broiler-strain chickens because of the exceptional amounts of material readily available at very low cost. This has enabled direct biochemical analysis of key components critical for elucidation of this highly complex system. Others, in order to overcome the problem of insufficient tissue mass, have resorted to the use of rachitic rats. However, none of these experimental approaches is necessarily ideal. In comparison to the mouse, the human skeleton is massive, over one-thousand times larger, and it lasts over one-hundred times longer. Thus, it is doubtful that mineral formation in a diminutive vertebrate like the mouse is entirely analogous to that in humans. On the other hand, the extreme demands for rapid growth and mineralization of bone in the broiler-strain chicken may well require specialized features that are not required for skeletal development of the mouse - or the human - except perhaps during the pubertal growth spurt. Finally, the use of rachitic rat growth plate tissue, while expedient, is problematic because of the fact that the the tissue, by definition, is abnormal.

It is also important to understand that in endochondral calcification, induction of calcium deposition is not directly analogous to that in true bone formation. Endochondral calcification is a primary, but provisional process; its end product, calcified cartilage, is more heavily calcified, but is mechanically weaker than true bone. It is destined to be largely replaced by cancellous bone. Endochondral calcification is mediated by growth plate chondrocytes; it is a rapid de novo process that begins with Ca²⁺ and Pi ions in solution. It involves overcoming of a sizeable nucleation barrier that requires the direct mediation of cellular metabolic activity. It utilizes extracellular vesicles and ends with formation of large amounts of microcrystalline, Ca²⁺-deficient, acid-phosphate-rich apatite deposits embedded in a proteoglycan and cartilage-specific collage-rich matrix. In contrast, true bone formation is mediated by osteoblasts; it is a secondary process that typically requires the presence of pre-existing mineral. During initial cancellous bone formation, osteoblasts deposit mineral upon a calcified cartilage scaffold; they guide formation of biological apatite into a predominantly type I collagen matrix. This leads to formation of the mechanically robust composite structure characteristic of bone. During embryonic intramembranous and subsequent periosteal bone formation, osteoblasts - not chondrocytes - facilitate and guide mineral formation.

The discovery that discrete extracellular nanostructures are the site of initial mineral deposition in growth plate cartilage of long bones was made independently over 40 years ago in the late 1960's by Anderson (1, 2) and Bonucci (3, 4). While there was some initial controversy, it soon became apparent that these structures are typically invested by a trilaminar membrane; thus the name, "matrix vesicles - MVs", became their accepted title. Since their initial discovery in growth plate cartilage, MVs have been found to be the locus of the first mineral deposits in intramembranous bone formation (5), in fracture callus (6-8), in developing dentin (5, 9), in developing deer antler (10), in pathological calcifications in aortic valves (11) and in osteosarcoma (12, 13). In addition, in other studies (14, 15) the initial site of mineral deposition in calcifying turkey leg tendon was also shown to begin within MVs and spread to the adjacent type I collagen fibrils.

The primary focus of this review is on de novo mineral formation in the epiphyseal growth plate of vertebrates. This is a cell-mediated process, discovered in the rat to be initiated by MVs (1, 3). However, this discovery was antedated by almost a decade by Irving et al. who found intensely staining Sudanophilic material in the extracellular matrix at the calcification front of the tibial growth plate of the rat and chicken, respectively (Figure 1A-B). This unusual lipid staining pattern was subsequently observed at the calcification front in all vertebrate species investigated (16-18). As will be described later, the Sudanophilic material is mineral-associated lipid, shown to be primarily acidic phospholipids - particularly phosphatidylserine (PS) (19-23). Although initially regarded as an artifact caused by the affinity of Ca²⁺ for PS (24-27), subsequent in vivo studies by Eisenberg et al. (28) established that this association was of biological origin. The ensuing finding that MVs have high levels of phospholipids enriched in PS (29-31), corroborated the finding of Sudanophilia at the calcification front, linking it to the presence of MVs.

It is important to point out that MVs are not typically associated with mineral deposition beyond the early inductive phase of calcification. Thus, MVs are not associated with the massive mineralization that occurs later in bone, dentin or cementum, nor are they association with mineralization events occurring during enamel formation. Thus their biological role appears confined to that of induction of solid mineral formation in tissues where no mineral phase previously existed. Despite a general acceptance of this role for MVs, they may not be responsible for the induction of mineral deposition in all vertebrate calcifying tissues. It is quite likely that some types of pathological calcification are brought about by other means.

3. MORPHOLOGY OF MATRIX VESICLES

Since their discovery over four decades ago, a great deal of work has been focused on MVs in an attempt to characterize their basic structural and chemical features. Much attention also has been focused on the mechanism of MV formation and their mode of action. First, to be discussed will be two sections on morphological and biochemical characterization. Later sections will deal with MV formation and the mechanism and regulation of MV-induced calcification.

3.1. Conventional transmission electron microscopy

Studies that led to the discovery of MVs were morphological in nature and the first observations made of MVs were of conventionally fixed and stained transmission electron microscopy (TEM) preparations (1-4). In general, these methods revealed MVs to be somewhat heterogeneous structures ranging from about 100 to 300 nm in diameter, typically enclosed by a trilaminar membrane, containing varying amounts of relatively amorphous osmiophilic material. Some, but by no means all of these vesicles were found to contain electron-dense crystalline structures. These mineral crystals were frequently associated with the inner surface of the lipid bilayer. Some studies have utilized La³⁺ as a probe for Ca²⁺ binding sites associated with MV mineral deposition (32, 33). La³⁺ is a rare-earth cation with an ionic radius similar to Ca²⁺; it binds with high affinity to Ca²⁺-binding sites such as acidic phospholipids of the MV membrane as well as focal proteoglycan aggregates. However, its charge-density and coordination chemistry are quite different from those of Ca²⁺ (34, 35). La³⁺ has no biological ion porter and thus does not penetrate intact membranes or enter living cells; in fact, submicromolar levels of La³⁺ block Ca²⁺-release-activated channels (36). While La³⁺ binds tenaciously to Ca²⁺-binding sites in the extracellular matrix and displaces Ca²⁺ (33), it apparently does not displace Ca²⁺ in microcrystalline apatitic deposits. In these studies it was noted that there were two types of MVs; some were dense and stained heavily with La³⁺; others were "light" and unstained. Thyberg and Friberg (37, 38) also made note of the heterogeneity of MVs and concluded that there were two types of vesicles, one of which derived from lysosomes. While many have noted heterogeneity in MV structures, there is now a much clearer understanding of their origin.

3.2. Cryofixation, freeze-substitution electron microscopy

Conventional tissue processing for TEM observation typically involves fixation with glutaraldehyde, post-fixing with osmium tetroxide, dehydration steps in alcohols, embedding in plastic, ultra-thin sectioning, flotation of the sections on aqueous medium, and frequently post-staining with uranyl acetate or lead acetate to increase electron density. It is now evident that these steps can cause major alteration in the native structure of the vesicles and their enclosed mineral, as well as in the surrounding tissue. To avoid these problems, several alternative approaches have been employed to reduce artifacts. Hunziker and others (39-41) reported the use of cryofixation and freeze-substitution for the preservation of details of MV structure. In their methods, rapid (40 msec) high-pressure freezing at liquid N temperature was used to prevent ice crystal formation. This was followed by dehydration at -90^oC with anhydrous methanol containing uranyl acetate (0.05%) and glutaraldehyde (2%), with gradual warming steps to -60^oC and -30^oC, replacement of methanol with acetone and infiltration with resin at +4�C. Such techniques prevented collapse of the proteoglycan network in the extracellular areas and the shrinkage of chondrocytes seen in conventional methods. MVs appeared to be attached to the surrounding matrix proteins (both proteoglycans and collagen fibrils); the trilaminar membrane was clearly evident. Similar cryofixation methods have been used by Arsenault et al. (42, 43), but with the use of "slam-freezing" in which freshly excised tissue was rapidly brought into contact with a copper block prechilled to <25�K with liquid helium. In addition, high resolution elemental mapping of the resultant preparations was performed using the electron spectroscopic imaging (ESI) technique of Ottensmeyer and Andrew (44). MVs again showed intimate association with matrix proteoglycans and collagen, and frequently contained what appeared to be mineral deposits (Figure 2A). Elemental mapping showed the presence of both Ca and P in electron-dense material in the vesicle lumen; however the intensity and localization of the two elements did not precisely coincide (Figure 2B) (40, 42).

Another problem attendant with tissue preservation using cryofixation and freeze substitution techniques concerns the handling of tissue sections. Typically these are floated on aqueous medium before being picked up by the coated grids. While macromolecules trapped in the plastic resin are well preserved, this step permits the loss of soluble electrolytes and small molecules from tissue specimens. To avoid this aqueous step, both ethylene glycol and glycerol have been used as flotation media. In a detailed study of these different anhydrous methods, Morris et al. (45) showed that with either freeze substitution, cryo-ultramicrotomy, or anhydrous ethylene glycol tissue processing, MVs were the locus of the first mineral crystallites seen in calcifying growth plate cartilage. The vesicles were shown to contain both Ca and P by energy-dispersive X-ray analysis. It should be mentioned here that one of the problems with the use of ethylene glycol processing after cryo-ultramicrotomy is that of poor membrane preservation. Subsequently, papers by Arsenault and Hunziker (40, 41), and Akisaka et al. (46) exploited the rapid cryofixation, freeze-substitution method, avoiding the use of post-staining and aqueous flotation methods. With these methods it became clear that much of the apparent heterogeneity seen in MVs with the various staining methods arises from artifacts directly attributable to the stain itself. Although the contrast was weak, the appearance of the unfixed, unstained MVs handled anhydrously after cryofixation and freeze-substitution was relatively uniform (46). Many contained uniform amorphous electron-dense material, with some vesicles containing small crystallites. These were best seen using selected-area dark field imaging (40). Perhaps of greatest importance, staining was shown by EDAX X-ray microanalysis to totally displace both Ca and P from the vesicles (40). In light of these findings, interpretation of conventionally prepared transmission electron micrographs showing MVs in tissue sections of mineralizing tissues must be done with great caution.

3.3. Freeze-fracture studies

Another technique that has yielded important information about MV structure is freeze-fracture. The first studies on MVs using this technique were reported in 1978 by Cecil and Anderson (47) and Borg et al. (48) at the Second Conference on Matrix Vesicles. These, and subsequent studies (49-51) have consistently revealed the presence of abundant MVs in the extracellular matrix of growth plate cartilage. The observed MVs fractured in a manner consistent with a lipid bilayer-enclosed structure. They possess well-defined intramembranous particles distributed more abundantly on the outer convex protoplasmic (PF) face than on the concave inner exoplasmic face (EF) of the vesicle membrane (Figure 3). The distribution of intramembranous particles in MVs was similar to that seen on the plasma membrane of the microvilli from which the vesicles appear to derive. The importance of this observation is that it reveals that the MV membrane retains the normal inside-outside orientation during the vesiculation process. This finding has major bearing on the role of Ca²⁺-dependent ATPase in Ca²⁺ loading by MVs.

Freeze-fracture studies also suggest that MVs derive not only by budding from cellular processes, and from cell disintegration (as indicated by some electron microscopic methods), but also from protrusions at smooth surfaces (46). Further, freeze-fracture studies revealed a close association between MVs and collagen fibers (47, 48). Remnants of MVs can be detected within islets of developing mineral deposits (50). Finally, freeze-fracture studies by Akisaka et al. (46) suggest that MVs can be categorized with respect to intramembranous particle distribution. MVs in the reserve and proliferative regions tended to be more uniform and had a random distribution of intramembranous particles on the EF and PF faces. In the hypertrophic region, MVs tended to be of two types: those with no intramembranous particles and those in which they were aggregated. MVs with aggregated intramembranous particles were associated with mineralization; those free of intramembranous particles were not. These data suggest that the intramembranous particles may be involved in mineralization and raise the important question as to which proteins are present in these particles. This question has not been addressed.

4. ISOLATION OF MATRIX VESICLES

One of the early obstacles to biochemical characterization of MVs was the lack of reliable methods for isolation of pure, chemically intact, biologically functional structures. This problem stemmed from the fact that MVs, as their name implies, are embedded among extracellular matrix connective tissue macromolecules such as proteoglycans and collagens. Further, MVs comprise only a minor fraction of the total mass of the tissue. Initially, the general approach used to obtain MVs from tissues involved use of crude collagenase, plus or minus other enzymes to digest the tissue slices, followed by differential centrifugation (52-54), with or without further subfractionation. To avoid the use of proteases, others used tissue homogenization followed by differential centrifugation and further subfractionation using either sucrose gradient methods (55, 56) or Percoll gradients (57) to produce MV-enriched microsomes (MVEM). Another method involved simple centrifugation of the spent medium from primary cultures of growth plate chondrocytes to produce "media vesicles" (MeV) (58). A major problem with MVs isolated with crude collagenase, as originally described by Ali et al. (52) was that they showed evidence of substantial damage to most of the resident proteins, even though their lipid and electrolyte content appeared to be less seriously affected. Significantly, this protease digestion method did not significantly damage the activity of alkaline phosphatase, which apparently was protected by its large 18-22 kDa glycan "umbrella" (59). In order to initiate mineral deposition, these MVs required the addition of high levels of organic phosphate substrates, typically 10 mM beta-glycerophosphate. On the other hand, the non-enzymatically released MVEMs, while very active in their ability to induce mineral formation without use of organic phosphate substrates, suffered from the presence of significant contamination of subcellular membrane fragments (55). Attempts to purify MVEM by a variety of methods met with limited success. Some success was achieved through use of Percoll gradients that do not cause the osmotic effects of sucrose, but there was difficulty with contaminating Percoll in the MVEM fraction (57). Subsequent work using much milder tissue digestion methods have enabled the isolation of MV that retain their complement of normal proteins (60-62), as well as the ability to rapidly initiate mineral formation. They also appear to have minimal contamination with intracellular membrane fragments.

4.1. Crude collagenase digestion methods

While there has been considerable controversy as to which is the best method of MV isolation, there probably is no 'best' method; each can serve some useful purpose. For example, the original crude collagenase digestion method of Ali et al. (52), while yielding vesicles that appeared quite homogeneous by conventional electron microscopy, caused major damage to external vesicle proteins. Nevertheless, the method was of value for analysis of MV lipids (29-31) and electrolytes (63, 64), and for certain MV enzymes (e.g. tissue-nonspecific alkaline phosphatase, TNAP) that are resistant to the nonspecific proteases present in crude collagenase (54). While collagenase-released MVs (CRMVs) have continued to be used by several groups for metabolic studies on the induction of mineralization, the damage caused to the external proteins and ion porters by crude collagenase undoubtedly alter their ability to induce mineralization and must be interpreted with caution.

4.2. Non-collagenase-dependent methods

To avoid problems arising from damage caused by the nonspecific proteases present in crude collagenase, MVEM were isolated by tissue homogenization and differential centrifugation. Such preparations very actively accumulated Ca²⁺ and inorganic phosphate (Pi), and rapidly induced mineral formation from synthetic lymph without the addition of organic phosphate substrates (55, 56). However, they contained intracellular membrane fragments, and could have entrapped extravesicular mineral during the homogenization process. Such MVEM were studied extensively. Preparations from the proliferating zone of growth plate cartilage neither accumulated significant Ca²⁺, nor Pi, nor induced calcification (55); MVEM from the hypertrophic zone readily induced mineralization. This showed that the ability to induce mineralization was not a general property of microsomes. Further, Ca²⁺ uptake by active MVEM could be inactivated simply by exposure to proteases (65). This revealed that induction of mineralization by MVEM did not result from contaminating mineral; otherwise the proteases would have exposed it and enabled even more rapid mineralization. However, these preparations were not homogeneous. Attempts to purify them using sucrose gradients, while eliminating many subcellular fragments, destroyed their ability to mineralize. The osmotic effects of the high sucrose concentrations caused major loss of Ca²⁺ and Pi from the vesicles (65-67). This revealed that the Ca²⁺ and Pi present were not stable; otherwise they would have survived this gradient procedure. Thus, although sucrose step-gradients enabled the preparation of fractions essentially identical by TEM to CRMV and had similar TNAP activity, they lacked the original ability of MVEM to induce mineral deposition (55, 56). In contrast, isosmotic Percoll gradients enabled the fractionation of MVEM into two major subpopulations, one of which readily induced mineralization, the other not (57). While this fractionation procedure improved the purity of the isolated MV fractions, Percoll contaminated the samples.

4.3. Cell culture methods

The use of growth plate chondrocyte cultures that release vesicles into the culture medium was also used to gain some insight into the protein and lipid composition of cell-released vesicles and in the mechanism of MV formation (68-72). However, a significant problem was that vesicles released into the culture medium (MeV) are not identical to those trapped in the matrix (71, 73-75). One significant difference is that these released vesicles differ in their lipid composition (73), as well as their ability to accumulate mineral ions. Another is that they had no detectable annexins, lipid-dependent Ca²⁺-binding proteins characteristic of MV isolated from native tissue, or the matrix of cultured chondrocytes (76).

4.4. Modified collagenase digestion methods

Progressive refinements in the original collagenase digestion method have been made that have significantly reduced the deleterious effects of the nonspecific proteases present in crude bacterial collagenases (61, 62). These entail reduction in the concentration of collagenase and the length of exposure, as well as certain post-digestion manipulations that facilitate the release of MV entrapped in the partially digested matrix. Such preparations rapidly mineralize when incubated in synthetic cartilage lymph in the absence of organic phosphate substrates. Protein patterns of these MVs are reproducible, showing numerous characteristic major bands. Mineralization of these MVs is highly sensitive to protease treatment; low levels of chymotrypsin or trypsin rapidly ablate their ability to mineralize (60, 61). And like MVEM, these vesicles are highly sensitive to pH; exposure to pH 6 buffers for only a few minutes almost totally destroys their ability to induce mineralization, even when subsequently supplied with organic phosphate substrates (77). Despite these improvements, there is some protease damage to these MV preparations, albeit not to such an extent as to destroy their ability to induce mineral deposition. This modified procedure for MV isolation has been used successfully by Balcerzak et al. (78) for proteomic analysis of MV proteins.

4.5. Other isolation methods

Various other approaches have been reported in the past for the purification of CRMVs. Anderson's group has utilized rachitic rat epiphyseal growth plate cartilage as a tissue source for isolation of CRMVs (79-82). Shapiro et al. have employed isoelectric focusing of CRMV (83). Vaananen and Korhonen used Percoll gradients to purify CRMV from lysosomal contaminants (84). Slavkin et al. (85) used Bio-Gel A-15m size-exclusion chromatography to purify CRMV released from predentine; and Deutsch and Sela et al. (86) used a combination of collagenase digestion, differential centrifugation and sucrose step-gradient fractionation to isolate MVs from alveolar bone. Very recently, Xiao et al. isolated MV from mineralizing osteoblasts for proteomic analysis (87). Thus, many different techniques have been used to isolate MV from various mineralizing tissues. Keeping this in mind, the proteins, enzymes, lipids, and electrolytes found in various MV preparations will now be explored. For most biochemical characterizations reported in this review, the MVs were isolated from normal, 6-8 week old, rapidly growing broiler-strain chickens.

5. MATRIX VESICLE PROTEINS

A major impediment to the analysis of MV proteins was the difficulty of isolating pure vesicles from the tissue without concomitant damage to the constitutive proteins. Through modifications in the isolation procedures for MV isolation - as was just discussed - significant progress in the characterization of MV proteins has been made. Very recently proteomic analysis of MVs has greatly expanded our perception of the complexity of the patterns of proteins expressed. This methodology has revealed the large number of proteins present, which now can be categorized into subclasses with implications as to their importance to MV formation and function. While proteomic analysis of MV proteins has become of great value in assessing the panoply of proteins present, certain important proteins are nevertheless missed. Therefore, in this review, biochemical, enzymatic, and proteomic analytical data will be used to present an integrated view of the nature and importance of the proteins detected.

5.1. Early SDS-PAGE Studies

Several early studies explored the sodium dodecyl sulfate - polyacrylamide gel electrophoresis (SDS-PAGE) patterns of proteins present in MV using crude collagenase digestion (88-90), as well as MVEM (55-57, 65, 68) and culture media vesicles(58, 69, 70, 91). While there was considerable variability in the patterns of MV proteins from the various tissues, animal species, and MV preparations, certain protein bands nevertheless were consistently observed; these include two or three intensely staining bands in the 30-38 kDa range, two strongly staining bands at 40-48 kDa, and bands frequently appearing in the 55 kDa, 90-100 kDa and 130-150 kDa range. Initially little effort was made to identify these proteins, although there were some estimates based on the apparent size of the proteins. Work by Muhlrad et al. (92) provided evidence for the presence of actin in MV isolated from rat alveolar bone, but beyond assessment of enzyme activity, limited attempt was made to assign a function to the proteins identified in MVs.

5.2. Isolation and Identification of Major MV Proteins

It should be mentioned here that early studies with MVEM showed that trypsin treatment reduced uptake of Ca²⁺ much more than uptake of Pi (57, 65)). When later studies again showed that protease treatment of actively calcifying MV fractions markedly impaired their ability to mineralize and that the effect of protease treatment was not due to damage to alkaline phosphatase activity (66), effort was begun in earnest to isolate, identify and characterize proteins critical to MV ion accumulation (76). The approach used was based on improved methods of isolation (See 4.4. above) that resulted in mineralization-competent MV; it compared several types of biologically active MV preparations (58, 66, 71, 76, 91). Proteins isolated from intact, but non-homogeneous MVEM prepared by nonprotease-dependent methods, were compared with those from pure, but partially damaged CRMV. Highly purified proteins from both types of MV preparations were used to raise monospecific polyclonal antibodies (59-61, 93). These antibodies, and well-characterized antibodies directed toward known proteins, were then used to identify cross-reactive proteins in each type of MV preparation. In this manner, intact samples of MV proteins were isolated from MVEM preparations; incomplete but pure MV proteins were isolated from the partially damaged collagenase-released MV (61). This technique was used successfully to identify several major groups of MV proteins.

5.3. Sequential extraction, separation, and characterization of major MV proteins

Using sequential extractions with buffered salt, non-ionic detergent and ethylene glycol tetra-acetate (EGTA) solutions, Genge et al. and Wu et al. (61, 94) developed methods that enabled the proteins of the isolated MV to be selectively extracted in a step-wise fashion. For example, treatment of isolated CRMV with hypertonic salt solutions selectively released bands migrating on SDS-PAGE at an apparent MW of 130-150 kDa (95). These released proteins were identified using both monoclonal and polyclonal antibodies to be type II collagen and its fragments. Subsequent extraction of the vesicle pellet with low ionic strength solutions released two major MV proteins migrating on SDS-PAGE at apparent MW of 40-48 kDa (96). These water-extractable proteins were identified to be proteoglycan link and hyaluronic acid binding region protein using both monoclonal antibodies and N-terminal amino acid sequencing of the proteins. Further extraction of the residual MV pellet with Triton X-100 detergent-containing buffers released TNAP and a variety of other quantitatively minor uncharacterized proteins. Next, extraction of the residual pellet with EGTA under specific pH and ionic strength conditions selectively released two classes of phospholipid-dependent Ca²⁺-binding proteins (61, 94, 97) (Figures 4A-C). Variations on this extraction scheme also enabled the selective extraction of TNAP largely free of other MV proteins (59).

5.4. Proteomic characterization of matrix vesicle proteins

The chemical dissection just delineated led to identification and characterization of many major MV proteins. The development of a proteomic approach proved to be of great value in more completely identifying the array of proteins present in authentic MVs. However, it is critically important in any proteomic study to utilize carefully purified, homogenous, functionally active MVs capable of inducing rapid mineralization. Two very recent studies on the MV proteome come to mind: the Xiao et al. studies (87, 98, 99) using vesicles isolated from cultured mouse cranial osteoblasts (both collagenase-released and culture media MV), and the 2008 Balcerzak et al. study (78) using MVs isolated from femurs of chicken embryos. While a huge number of proteins (>1000) were implicated by the Xiao et al. studies, the purity and functional activity of the vesicles analyzed were not clearly established. Neither was there any clear analysis of the probability that these proteins were from bona fide MVs. Also, there was no indication of whether these vesicles were capable of inducing mineralization - a critical feature of their known biological activity. Regrettably, there is ample reason to believe that these vesicle fractions were seriously contaminated with cell and matrix debris. In contrast, the purity of the MVs isolated by Balcerzak et al. from chick embryo femurs was much better documented; the vesicles were clearly shown to induce mineral formation, and the SDS-PAGE profiles were very similar to those of previously published reports on protein patterns of functional MVs (60, 61). The number of proteins identified by Balcerzak et al. was still large (126 gene products); but, these were each presented with an indication of their abundance, the reliability of their identification, and their likely function.

Nevertheless, it is important to note that despite the plethora of proteins reported in these studies, several key enzymes and proteins known be present in MVs from their biological activities were not detected. For example, two known Pi transporters, an L-type Ca²⁺ channel, and a variety of phospholipases and growth factors known to be present in growth plate chondrocytes and MVs were not detected by these proteomic analyses. Their absence therefore raises a significant note of caution. Possible reasons for failing to detect these proteins are: 1) that ion transporters are relatively large multi-transmembrane integral proteins that may not have been properly dissolved in the detergents used; and 2) that very small proteins like the phospholipases and growth factors may have been lost during preparation of the vesicles for analysis. With these caveats, the array of proteins identified in the chick embryo femur MV by Balcerzak et al. will be presented, along with their Mascot Scores (100). The Mascot Score (MS) = -10*Log(P), where P is the probability that the observed match is a random event (101) A big score means that the probability for a given protein to be in MVs is high. Thus, the MS is shown in brackets next to each protein identified by the proteomic analysis.

6. MATRIX VESICLE-ASSOCIATED EXTRACELLULAR MATRIX PROTEINS

Surprisingly, extracellular matrix proteins represent some of the major MV proteins identified by proteomic analysis. They appear to be proteins attached to MVs via various linkages. However, their presence in isolated MV fractions is in close agreement with previous studies that documented the strong binding between MVs and certain matrix proteins present in growth plate cartilage (95, 96, 102, 103).

6.1. Type VI collagen

The proteomic discovery of the abundance of Type VI collagen in the MV fraction is the first indication of its close relationship with MVs (78). This analysis revealed the presence of each of the three collagen chains: alpha-1 (Mascot Score, 1621), alpha-2 (1556), and alpha-3 (1023) in the MV fraction, indicative that a substantial amount of Type VI collagen is associated with MVs. Type VI collagen is a major component of microfibrils present in the capsule that defines the pericellular matrix surrounding individual chondrons (104). Mutant forms of type VI collagen appear to be causally involved in various classes of muscular dystrophy (105-108). Of special interest to growth plate mineralization, Type VI collagen receptors on the chondrocyte surface (NG2-proteoglycan) (109, 110) act as transducers to mitochondria within the cells, affecting their permeation transition pore for Ca²⁺ (111). As will be discussed later, this appears to have an important influence on mitochondrial Ca²⁺ release (112).

6.2. Type X collagen

Type X collagen has long been associated with the hypertrophic region of growth plate chondrocytes, and was assumed to be involved either in the onset or regulation of mineral formation (113-115). Recent studies by Genge et al. using synthetic MV nucleational complexes found that Type X collagen had a small inhibitory effect on the onset and rate of mineral formation; however it significantly extended the length of the "rapid formation" period, thereby altering the kinetics of the overall amount of mineral formation (116). Only one representative of type X collagen: alpha-1 was observed in the MV proteome; it's Mascot score (142) was weak and indicated that only minimal amounts of Type X collagen were present in the MV fraction (78). In light of the documented tight binding of Type X collagen to MV (95, 102, 103) it is likely that the use of bacterial collagenases during isolation of MV caused the Type X collagen to become largely degraded. More will be said about type X collagen later.

6.3. Proteoglycan link protein and aggrecan core protein

Proteoglycan link protein and aggrecan core protein were first shown to be tightly associated with MVs by Wu et al. in 1991 (96). Both proteins were found to be major components of isolated functionally active MVs and could be dissociated by exposure to low ionic strength buffer. These early findings were confirmed by the recent proteomic analyses of avian MVs (78). Their Mascot Scores, (1040) for proteoglycan link protein, and (966) for aggrecan core protein, indicate a strong presence in the isolated MV fraction. Proteoglycan link protein is known to stabilize aggregates of proteoglycan monomers with hyaluronic acid in the extracellular cartilage matrix. Similarly, aggrecan core protein serves as the attachment site of chondroitin sulfate and keratin sulfate chains in the large proteoglycan macromolecule. MVs are known to be embedded in the extracellular matrix around hypertrophic chondrocytes; freeze-clamp ultrastructural studies have confirmed the filamentous attachments between the MV and the extracellular matrix (43). Thus, it was not surprising that proteoglycan-related proteins would be observed in the MV proteome. Small amounts of additional proteoglycan-related proteins were also identified by proteomic analysis (78).

6.4. Fibrillin-1 and Fibrillin-2

The fibrillins have also been observed by proteome analysis of isolated MVs; the Mascot Scores of Fibrillin-1 (549) and Fibrillin-2, (672) indicate a significant presence in MVs. Fibrillins are present in the extracellular matrix of growth plate cartilage (115, 117-119). Mutations in the fibrillin gene have been linked to the Marfan syndrome (120). Flash-frozen ultrastructural examination of growth plate cartilage has revealed that MVs are attached to microfibrils in the extracellular matrix (41, 44). Although not established morphologically, it is probable that the fibrillins are present in combination with type VI collagen in the pericellular domain and are attached to MVs in the extracellular matrix.

7. THE ANNEXINS - ACIDIC-PHOSPHOLIPID-DEPENDENT Ca²⁺-BINDING PROTEINS

The dominant, most abundant proteins detected by the proteomic assay were the annexins; they compromise a large family of acidic phospholipid-dependent Ca²⁺-binding proteins. The Mascot Score (2877) of annexin A5 (AnxA5) was the highest score for any MV protein; however, annexin A6 (AnxA6) (1798) and annexin A2 (AnxA2) (956) were also high-ranking (78). These findings are in close agreement with earlier analyses of proteins present in nucleationally-active MVs made by Genge et al. in 1989 showing that the annexins were major proteins and that they bound tightly to PS in a Ca²⁺-dependent manner (Figure 4A-C, Figure 5) (60). However, proteomic analyses have also identified three additional annexins not evident from the earlier studies: annexin A1 (AnxA1) (1197), annexin A11 (AnxA11) (846), and annexin A4 (AnxA4) (327).

7.1. Annexin A5

AnxA5 is a ~33 kDa protein that has been extensively studied during the past 20 years. Its 3-D crystalline structure was established already 20 years ago and revealed a predominant cluster of alpha-helices that define a hydrophilic pore through the center of the protein (121, 122). This pore was thought to serve as a Ca²⁺ channel in the MV membrane, even though it exhibited no transmembrane domains (123-125). AnxA5 is known to form triskelion arrayed trimers that bind to the surface of acidic phospholipid-containing bilayers (126-130). There is evidence that these trimers may be involved in the nucleational core of MVs and contribute to the ability of MVs to induce mineral formation (131). AnxA5 becomes exposed to the external surface of cells undergoing apoptosis (132-135), and as such has been used to monitor programmed cell death. More than any other protein so far studied, AnxA5 greatly accelerates the nucleational activity of the acidic phospholipid-Ca²⁺-Pi complexes present in the nucleational core that triggers de novo calcium phosphate mineral formation in MVs (116, 131, 136). AnxA5 specifically overcomes the inhibitory effects of Mg²⁺ that otherwise greatly slow the induction of mineral formation (137). The key roles of AnxA5 in growth plate mineralization will be discussed later.

7.2. Annexin A6

AnxA6 is a genetically-fused double version of an ancient annexin with eight instead of four domains; it has a MW of ~68 kDa (138, 139). The crystal structure of AnxA6 has been determined (140) and reveals six Ca²⁺-binding sites in AB loops analogous to the three present in AnxA5 (141). AnxA6 has been implicated in mediating endosome aggregation and vesicle fusion during exocytosis (142). Expression of AnxA6 in cells reduces cell proliferation (143) and thus may play that role in growth plate cartilage in the zones of maturation and hypertrophy where cell division ceases. Recent studies on calcification in vascular smooth muscle cells have shown that their MVs are rich in AnxA6 and AnxA2; they stimulate mineralization when associated with type I, but not type II collagen (144). Curiously, these vascular smooth muscle MVs had no AnxA5 present - in stark contrast to cartilage MVs. This recent study indicates that AnxA6 (and possibly AnxA2) may be important for adherence to type I collagen in smooth muscle calcification.

7.3. Annexin A2

Like other annexins, AnxA2 is a ~36 kDa Ca²⁺-dependent phospholipid-binding protein abundantly present in MVs (94, 145). AnxA2 has been implicated in diverse cellular processes involving linkage of membrane-associated protein complexes to the actin cytoskeleton, e.g. ion channel formation, and cell-matrix interactions (146). Another of its functions appears to be facilitating exocytosis of intracellular proteins to the extracellular domain (147). Studies by Hale et al. (69) reveal that dissociation of actin microfibrils from the cell membrane stimulates MV formation; AnxA2 has been implicated in the detachment of membrane plaques from actin filaments, leading to the formation of cell-surface blebs (148). The available data strongly suggest that AnxA2 is directly involved in the formation of MVs. Studies to date do not indicate that AnxA2 enhances cartilage MV mineralization (136, 145), although it is possible that it may enhance mineralization of MVs formed by vascular smooth muscle (144).

7.4. Annexin A1

AnxA1, also known as Lipocortin I, was found to be abundant in MVs by proteomic analysis (78). It has a molecular weight of approximately 40 kDa and preferentially localizes to the cytosolic face of the plasma membrane upon elevation of intracellular Ca²⁺ (149, 150). AnxA1 has inhibitory activity toward phospholipase A₂ (151, 152). Since phospholipase A₂ is required for the biosynthesis of the eicosanoid mediators of inflammation, prostaglandins and leukotrienes, AnxA1 has potential anti-inflammatory activity (151, 153). Glucocorticoids are known to stimulate production of lipocortin (153). Prostaglandins are known to be synthesized by growth plate chondrocytes (154). It is apparent that the presence of lipocortin (AnxA1) in MVs, and the cells that produce them, serves as a regulatory mechanism controlling the inflammatory process during growth plate development (155) - perhaps also being involved in regulation of angiogenesis and vascular invasion (156) after mineralization of the cartilage matrix.

7.5. Annexin A11 and Annexin A4

AnxA11 and AnxA4 are other members of the annexin family found by proteomic analysis to be a significant presence in MV (78). AnxA11 is a 56-kD protein that has been shown to interact with PDCD6 (Programmed Cell Death Protein 6), a Ca²⁺-binding protein belonging to the penta-EF-hand protein family that participates in glucocorticoid-induced programmed cell death (157). Thus, AnxA11 appears to work in synergism with AnxA5 during the process of programmed cell death that appears to accompany MV formation. AnxA4 was also detected at low levels in the MV proteome. Its role in MV function is uncertain. The involvement of specific annexins in MV-mediated mineral formation will be discussed in more detail later in this review.

8. MATRIX VESICLE ENZYMES

A wide variety of enzyme activities have been found in MV; these fall into several general classes: phosphatases, nucleotide triphosphate pyrophosphohydrolase, phospholipases, esterases, proteases and peptidases, and a wide variety of others. Forty years ago Ali et al. (52) upon successful isolation of MVs made a study of their constituent enzymes (53). This work indicated that MVs were highly enriched in TNAP, ATPase and pyrophosphatase activities. (However, it is important to note here that any enzyme that can hydrolyze ATP can be considered to be an "ATPase"; similarly any enzyme capable of hydrolyzing PPi is by definition a "pyrophosphatase". As will discussed later, these activities appear to result from only one enzyme - TNAP). MVs also exhibited several other phosphatase activities, but not acid phosphatase. Subsequent broader screening of MV enzyme activities in two independent studies using both collagenase-released (158) and non-collagenase-released, sucrose gradient-purified MV (56) observed that MVs did not exhibit significant activities of intracellular organelle-related enzymes such as NADPH-cytochrome c reductase (endoplasmic reticulum), succinate dehydrogenase, NADH-cytochrome c reductase and cytochrome oxidase (mitochondria), cathepsin D and β-glucuronidase (lysosomes). Additional studies described the activities of two ubiquitous enzymes, lactate dehydrogenase (159) and carbonic anhydrase (160) in isolated MV. On the other hand, very recent proteomic studies on MVs have detected the physical presence of a large array of enzymes, including small amounts of many cytosolic enzymes, as well as significant amounts of peptidases and amino transferases (78).

8.1. MV alkaline phosphatase

Special interest in alkaline phosphatase stems from the fact that as early as 1923 its activity was found to be associated with the mineralization process (161). Much later when MVs were established as sites of de novo mineralization, they were found to possess intense TNAP activity (52, 53). Recent proteomic analysis has shown that TNAP has by far the most dominant physical presence of any enzyme in MV (MS = 2549) (78). TNAP is one of four AP isozymes typically found in mammals. Unequivocal evidence that TNAP is directly involved in mineralization during skeletal development in vivo has come from recent gene knockout studies in mice (162, 163). Some have concluded that the primary function of TNAP is due to its pyrophosphohydrolase activity (164, 165); however, there is evidence that other aspects of its activity may contribute significantly to mineral formation.

8.1.1. Molecular structure

Like other vertebrate alkaline phosphatases, TNAP is a homodimer with a core protein molecular mass of ~57 kDa. It is now well established from genetic data that the human form contains 534 amino acids (166); and like other AP isozymes, TNAP is covalently attached to the outer membrane of cells and MVs by a glycosyl-phosphatidylinositol tether near the C-terminal end of the enzyme (167). TNAP is post-transcriptionally glycosylated at several sites with substantial levels of N-linked oligosaccharides to yield a glycoprotein with total molecular mass of ~80 kDa. The avian MV enzyme occurs in two different membrane-bound glycosylated forms of 77 and 81.5 kDa; these represent the 59 kDa protein core to which are attached either four or five 4.5 kDa N-linked oligosaccharides, respectively (59). These oligosaccharides form a shield that largely protects TNAP from proteolytic inactivation during MV isolation. These findings are consistent with earlier studies by Wu et al. (59) and Hsu et al. (168, 169).

8.1.2. Amino acid sequence

MV-TNAP has not been directly sequenced; but the sequences of TNAP isozymes from various vertebrate species have been deduced from cDNAs. There is significant homology between E. coli, rat, human, and chicken TNAP, especially in the catalytic, metal-binding regions. There are differences in other regions of the enzyme. Between mammalian species, the sequences of TNAP have shown a high degree of identity (>90%); however differences between chicken cartilage and mammalian TNAP are such that antibodies to the mammalian TNAP only weakly cross-reacted with chicken cartilage isozyme; similarly, mammalian cDNA probes do not hybridize well with the avian cartilage form (59).

8.1.3. Three-Dimensional (3-D) structure

To date, this has not been determined; however, structures have been solved for the E. coli enzyme (170, 171), for human placental AP (172, 173) and some other APs. These studies reveal a compact, dimeric enzyme with four metal-binding sites in or near the catalytic region of each subunit. There are two Zn²⁺ ions and one Mg²⁺ at the catalytic site of each subunit, and a stabilizing Ca²⁺ ion located 10 angstroms from the active site (172). This Ca²⁺-binding property of TNAP was foreshadowed by the much earlier observations of Stagni (174) and de Bernard et al. (175) who implicated this in the mineralization mechanism. Because of the heavy glycosylation of the vertebrate APs, no diffraction-quality crystals of the intact enzyme have been obtained. However, the intimacy between TNAP and the developing mineral phase during MV mineralization (76) underlines the importance of establishing the 3-D structure of the enzyme.

8.1.4. Disposition in MV membrane

Disposition of TNAP in membranes is key to understanding its function. Studies by Low et al. (167) revealed that AP is an ectoenzyme, covalently anchored to the plasma membrane by a glycosylphosphatidylinositol (GPI) linkage to the external lipid bilayer. TNAP can be selectively removed from the MV membrane by the action of a highly specific bacterial phospholipase C which attacks only this GPI link (66). Despite its external localization, AP is remarkably insensitive to protease attack because of the glycan "umbrella" previously described. Further, attachment of TNAP via its GPI-based anchor appears to facilitate mineral formation to an apatitic form indistinguishable from that produced by intact MVs (176). These findings are consistent with earlier work showing that activity of TNAP dies off in direct proportion to the extent of MV mineralization (76). This finding indicates an intimate, self-destructive physical association occurs between membrane-bound MV-TNAP and the mineral that is being formed.

8.1.5. Catalytic properties

The broad spectrum of phosphatase activity in MV, including pyrophosphohydrolase, is due primarily to that of TNAP (54, 177). These findings are in accord with earlier work by Moss et al. (178, 179) who found that TNAP possessed both pyrophosphatase as well as general phosphoesterase activity. MV TNAP has significant catalytic activity toward PPi and other phosphorylated substrates under extracellular conditions, as has been often demonstrated by studies that show enhancement of MV calcification when supplied with organic phosphate substrates (82, 180). Moreover, the finding that knockout of TNAP in mice gives rise to a distinct form of osteomalacia (181-183) directly confirms its physiological significance. However it is important to note that despite the lack of TNAP, normal mineral formation nevertheless occurred within the MVs - but it did not spread to the surrounding matrix (163). Coupled with the observation of the self-destructive physical association between TNAP and MV mineral formation (76), it is evident that TNAP is required for the progression of mineral from the MV lumen to the extracellular matrix.

8.1.6. Collagen-binding properties

MV-TNAP has been shown to bind with high-affinity to types II and X collagen (69, 95, 184). This is consistent with evidence that TNAP is localized on the external face of the MV membrane (66). Licia Wu found that when either type II or X collagen was incubated with intact MVs, each co-sedimented with the vesicles (75). Further, TNAP in detergent extracts of collagenase-released MV was found to bind tightly to columns of immobilized native type II or X collagen (95). A collagen-binding loop evident in 3-D TNAP structure appears to be responsible (185). Several other MV proteins also bind to type II and X collagens, but not to type I collagen - findings that help explain how MVs are attached to the cartilage extracellular matrix and facilitate the spread of mineral formation into the extravesicular matrix. Electron microscopic studies using freeze-quenching methods show that MVs are attached to the matrix by filamentous protein strands (40-42, 46).

8.2. Nucleotide pyrophosphatase/phosphodiesterase (NPP1, PC1)

Many studies have revealed the presence of NPP1 activity in MVs (165, 186-193). In fact, recent data indicate that the ATPase activity found in MVs can be largely attributed to TNAP and NPP1 (190,193). Recent proteomic analyses of MVs isolated from chicken embryo femurs have confirmed the presence of NPP1 (Mascot Score, 684), however, at only about one-fourth the level of TNAP (78). This ecto-enzyme hydrolyzes nucleotide triphosphates to produce nucleotide monophosphates + PPi extracellularly. There the released PPi would act as a HAP crystal growth inhibitor (194); however, being a substrate for TNAP it would be rapidly hydrolyzed (189). NPP1 accounts for nearly all the ecto-pyrophosphatase activity of chondrocytes (195) and the release of PPi is thought to provide a means for regulating apatitic mineral formation (194). More will be made of this in the discussion of the role of PPi in regulation of mineral formation. NPP1 is a type 2 membrane glycoprotein, a member of a small multigene family of nucleotide pyrophosphatase/phosphodiesterases (196).  It is a 240 kDa homodimer comprised of 2 disulfide-bonded 120 kDa subunits (197-198).  NPP1 is expressed strongly in bone and cartilage, but most notably in chondrocytes (165).  It has been shown to possess both alkaline phosphodiesterase I (EC 3.1.4.1) and nucleotide pyrophosphatase (EC 3.6.1.9) activities (199).  NPP1 also has been found to exhibit auto-phosphorylation activity, which inactivates its other two enzymatic activities (200, 201).  It has been suggested that auto-phosphorylation, which occurs at low level ATP concentrations, is a regulatory mechanism preventing depletion of nucleotides when they are scarce.

8.3. PHOSPHO-1 (phosphoethanolamine/phosphocholine phosphatase)

PHOSPHO-1 is a recently identified phosphatase that exhibits enzymatic activity in MVs (202) and represents a significant amount of the protein present. Deduced by proteomic analysis, PHOSPHO-1 has a Mascot Score of (420), about one-sixth the level of TNAP (78). PHOSPHO-1 is a 267 amino acid, 29.7 kDa member of the haloacid dehalogenase (HAD) super family of Mg²⁺-dependent hydrolases (203). The active site of PHOSPHO-1 has been shown by crystallographic and mutation analysis to consist of three aspartate residues at positions 32, 34 and 203 coordinated with Mg²⁺ (203). Because gene expression of PHOSPHO-1 is upregulated as much as 120-fold in mineralizing cells as compared to soft tissue cells, the enzyme is now thought to generate Pi within MV (204). Very recent (2010) studies suggest that inhibitors of PHOSPHO-1 can block mineralization of embryonic long bones of chicks, as assessed by alcian blue/alizarin red staining (205). This finding, coupled with the recent study indicating that inhibition of PHOSPHO-1 can block MV mineralization (203), suggest that PHOSPHO-1 is involved in long bone mineralization in vivo. It is therefore of interest to understand how this enzyme contributes, given the predominant activity of TNAP. The main substrates for PHOSPHO-1, phosphoethanolamine (PEA) and phosphocholine (PCho) are products of the action of phospholipase C on membrane phospholipids, phosphatidylcholine (PC) and phosphatidylethanolamine (PE) (206). Recent studies by Wu et al. have shown that the PE and PC content of MVs decreases rapidly during MV-mediated mineral formation, accompanied by an accumulation of 1,2-diacylglycerol and 1-monoacylglycerol (207). These findings present clear evidence that phospholipase C-dependent breakdown in MV phospholipids accompanies MV mineralization. Taken together, these findings reveal that combined actions of phospholipase C and PHOSPHO-1 release Pi (as well as ethanolamine and choline) from membrane phospholipids, making them available for use in biomineralization. Further, recent studies show that Pi uptake by mineralizing MVs is not strictly Na⁺-dependent; Pi uptake is enhanced when choline⁺ is substituted for Na⁺ (208). This will be covered in more detail in the section dealing with Pi uptake during MV formation.

8.4. Acid phosphatase

Several other MV phosphatases also have received significant attention in the past. In a 1995 review, Anderson cited studies that indicate several types of phosphatases: pyrophosphatase (PPiase), 5'-AMPase, nucleotide triphosphate phosphatase (NTPPase), as well as Ca²⁺-ATPase (209) are present in MVs. While it is probable that PPiase and 5'-AMPase simply represent manifestations of TNAP, the activity of acid phosphatase is clearly separate. Early studies by Thyberg and Friberg (37) revealed the presence of acid phosphatase histochemically in MV. Subsequent biochemical studies by Ali (158),Watkins et al. (56) and Deutsch et al. (86) have detected modest levels in MVs. On the other hand, using Percoll fractionation, Vaananen separated acid phosphatase-containing from other MV (84). Thus some MVs contain acid phosphatase; it is perhaps those derived from lysosomal extrusion (46) as suggested by Thyberg and Friberg (37). The fact that proteomic analysis failed to detect acid phosphatase in mineralizing MVs from chicken cartilage (78) makes it doubtful that MVs containing acid phosphatase activity function in the induction of mineral formation.

8.5. Proteases and peptidases

The presence of peptidase (210) and protease (211) activities associated with MV have been confirmed by proteomic studies on MVs isolated from actively mineralizing chicken bones, showing a significant (MS = 1476) total level (78). This supports the concept that enzymatic degradation of the proteoglycan matrix of growth plate cartilage is important in facilitating mineral formation, an idea raised over 40 years ago (212, 213) prior to their discovery in MVs (214). These aminopeptidases, which exhibit activity toward a variety of synthetic peptides and amino acyl esters (210), are blocked by N-carbobenzoxylation. Such enzymes are commonly detected in plasma membrane fractions. In studies by Katsura et al. a neutral metalloprotease isolated from growth plate MVs readily degraded proteoglycans (211). It had a molecular mass of 33 kDa, a pH optimum of 7.2, and was latent within MVs until activated by detergents. The fact that the activity is latent has important physiological implications. It is probable that it becomes activated when the vesicle membrane ruptures after onset of apatite crystal formation. Thus, initiation of MV-induced mineralization becomes coupled with degradation of the inhibitory proteoglycan matrix. This concept will be developed further when the regulatory role of proteoglycans is discussed.

8.6. Lactate dehydrogenase and other glycolytic enzymes

Nearly 20 years ago studies by Hosokawa et al. (159, 215) documented the presence of lactate dehydrogenase (LDH) in isolated MV. While no other glycolytic enzymes were detected at the time, very recent proteomic studies have not only detected significant levels of LDH (Mascot Score = 855), but seven of the eleven glycolytic enzymes normally present in cells were also observed (78). In addition, several enzymes were found that pertain to the synthesis of the proteoglycans that are abundant in cartilage. These enzymes appear to be adventitious carryovers that occur during MV formation; thus, the abundant presence of LDH speaks of the relatively hypoxic state of growth plate chondrocytes prior to the time of MV formation (216).

8.7. Carbonic anhydrase

Carbonic anhydrase (CA) was first implicated in endochondral calcification by Cuervo in Howell's group nearly 40 years ago (217) in conjunction with studies on the pH and pCO₂ of the extracellular fluid. CA has since been observed in growth plate cartilage in two isozymic forms (218-220). CA isozyme II was notably present in the extracellular matrix of hypertrophic and calcifying cartilage, the locus of MVs. Sauer et al. (160) were the first to document CA activity in isolated MVs, observing levels comparable to those reported earlier for avian growth plate cartilage. Treatment of the isolated vesicles with acetazolamide, a CA inhibitor, delayed but did not prevent MV mineralization. Thus it is probable that CA modulates the activity of MVs, but may not to be absolutely essential for MV function.

8.8. Phospholipases

Observations on the lipid composition of growth plate cartilage (20) and MVs (30, 221) have provided evidence for the presence of significant phospholipase activity, even though none of these enzymes has been isolated physically. In MVs obtained from chicken growth plate cartilage, 10-15% of the total phospholipids are lysophospholipids, indicative of the presence of significant phospholipase A activity (29-31, 222). Clues regarding the nature of phospholipases in MVs come from studies on changes in lipid composition of growth plate chondrocytes during MV formation - as well as during induction of mineral formation by MVs. For example, depletion in PC and increase in lysophospholipids characteristic of MVs are already evident in microvilli from which MVs derive (73). Further, induction of mineral formation by MVs causes a progressive decline in phospholipid levels (207). This marked reduction is accompanied by the buildup of free fatty acids (FFA), 1,2-diacylglycerols (1,2-DAG) and 1-monoacylglycerols (1-MAG). The two phospholipids undergoing the most extensive hydrolysis during mineral formation were phosphatidylinositol (PI) and sphingomyelin (SM). PE and PC also were extensively degraded, but not as rapidly. These findings parallel those made with frozen-thawed slices of calcifying epiphyseal cartilage (223).

These studies lead to several conclusions. 1) The high levels of lysophospholipids present in growth plate cartilage and MVs provide direct evidence for the presence of phospholipase A (PLA) activity. 2) The progressive decline in PC and PE, coupled with the buildup of 1,2-DAG, provides evidence for the presence of phospholipase C (PLC) activity in MV (207). 3) The lack of accumulation of lysophosphatidylcholine (LPC) and lysophosphatidylethanolamine (LPE) during induction of mineral formation by MVs, coupled with the increase in 1-MAG, is evidence for the presence of lysophospholipase C (LPLC) activity. 4) The very rapid decline in SM reveals the presence of sphingomyelinase activity. 5) The extremely rapid decrease in the level of PI, without buildup of lysophosphatidylinositol (LPI), suggests degradation by PI-specific PLC activity. 6) The unique pattern of PS and lysophosphatidylserine (LPS) metabolism in which buildup occurs during MV mineralization points to active base-exchange PS synthase enzymes that transfer serine to LPE or PE to form LPS or PS. 7) Levels of phosphatidic acid (PA) in MVs are generally low, indicating low phospholipase D (PLD) activity (207); however, Balcerzak et al. recently describe phosphorylation-dependent PLD activity in MVs (224).

8.8.1. Phospholipase A (PLA)

PLAs represents a large class of typically small (13-14 kDa) Ca²⁺-dependent enzymes that hydrolyze the sn-1 or sn-2 ester of glycerophospholipids to produce a fatty acid and a lysophospholipid (monoacyl phosphoglyceride) (225-226). PLA₁ causes the release of a free fatty acid from the sn-1 position of glycerol to produce 2-monoacyl phosphoglycerides; PLA₂ causes release of fatty acid from the sn-2 position of glycerol to produce 1-monoacyl phosphoglycerides. (Typically the fatty acid present in the sn-2 position is polyunsaturated and serves as a substrate for eicosanoid biosynthesis). Attempts to physically isolate the PLA enzyme present in growth plate cartilage have so far failed. However, characterization of its activity has met with limited success. MV-PLA activity is Ca²⁺-dependent and is highly selective for intramembranous, as opposed to externally added phospholipid substrates - implying that it was an integral membrane protein (227). The enzyme had optimal activity at pH 8, displayed A₂ specificity, and hydrolyzed PC in preference to PE or other membrane phospholipids. Its properties resemble activities reported for PLA₂ activity isolated from human rheumatoid synovial fluid (229-230). Schwartz and Boyan (231-232) explored the effect of two vitamin D metabolites, 1,25-dihydroxyvitamin D₃ (1,25-(OH)D₃) and 24,25-dihydroxyvitamin D₃ (24,25-(OH)D₃) on the activity of PLA₂ in cultured chondrocytes. They found a marked enrichment in PLA₂ activity in MV, compared to the plasma membrane. Further, 1,25-(OH)D₃ at levels of 10^-8 and 10^-9 M significantly increased PLA₂ activity in MV, but not in the plasma membrane. Thus, vitamin D metabolites influence the activity of PLA₂ in growth plate chondrocytes. It is of further interest that the annexins act as PLA₂ inhibitors (233), protecting PS from degradation during MV formation by interfering with PLA₂ binding to PS (60, 151).

8.8.2. Phospholipase C (PLC)

PLC acts on glycerophospholipids to release 1,2-DAG and the phosphorylated polar head group of the parent phospholipid. There are two broad classes of PLC. One type rapidly hydrolyzes phosphatidylinositol 4,5 bis-phosphate (PIP₂) to produce 1,2-DAG and 1,4,5-inositol trisphosphate (IP₃) - products that are involved in transient receptor-mediated PI-cycle cell signaling (234-239). A slower acting PC-specific PLC has been shown to be involved in regulation of the cell cycle, cell division and apoptosis (240-245). PC-PLC acts on PC to produce 1,2-DAG and PCho. PC-PLCs from both rat (246) and human (247) fibroblast-like cell lines have been characterized. The human homolog is a 66 kDa enzyme that is directly involved in regulating the cell cycle in fibroblasts. It is upregulated by mitogens such as platelet-derived growth factor (PDGF). Although the action of PIP₂-specific PLC has been studied in chondrocytes (239, 248), there have been no studies on PLC activity toward PC or PE, two far more abundant lipids in growth plate cells and MV. Breakdown of MV lipids during in vitro calcification causes progressive buildup of both 1,2-DAG and 1-MAG, findings indicative of the presence of PLC and lysophospholipase C (LPLC) activity, respectively (207). While neither enzyme has been characterized in growth plate chondrocytes, PC-PLC and PE-PLC have been extensively studied in other tissues where they appear to be involved in cellular proliferation (249-250). In view of the rapid cell division that occurs in the proliferative zone of growth plate cartilage, it is not surprising to see evidence of PLC activity in MVs. The recent finding that PHOSPHO1 targets hydrolysis of PEA and PCho, two specific products of PLC action on PE and PC, respectively, reveals a direct connection that appears to contribute to the function of MV in inducing mineral formation.

8.8.3. Lysophospholipase C (LPLC)

One of the enigmas observed during incubation of growth plate cartilage slices, or of MVs during induction of mineral formation, was the finding that despite the significant loss of PC and PE, there was little concomitant accumulation of LPC or LPE (207). While this might be explained by the presence of the alternative PC-PLC or PE-PLC pathway, the extent of accumulation of 1,2-DAG was not sufficient to explain the loss of these lipids. On the other hand, the obvious accumulation of 1-MAG during MV mineralization provides a more plausible alternative explanation - namely the presence of LPLC activity. Recently, a study of the ubiquitous NPP family of ectoenzymes, of which the previously described PC-1 (NPP1) is a member, reported the presence of additional family members, NPP2, NPP6 and NPP7 (251). These family members, rather than attacking nucleotide phosphodiesters, attacked lysophospholipids. NPP6 had activity toward LPC with polyunsaturated acyl chains. It also hydrolyzed sphingosylphosphocholine (SPC), the lyso form of sphingomyelin (SM) (251). The products of PC-LPLC (NPP6) action on LPC are PCho + 1-MAG; those on SPC are PCho + sphingosine; action on GPC yielded PCho + glycerol. Thus, NPP6 specifically cleaves the P-O bond opposite the choline residue. The significance of NPP6 to MV lipid metabolism is that it would provide a mechanism for further degradation of PCho resulting from breakdown of both PC and SM present in MVs. And as noted before, PCho is a preferred substrate for PHOSPHO1 and would release Pi needed for supporting mineral formation.

8.8.4. Sphingomyelinase (SMase)

SM is specifically enriched in MVs compared to the plasma membrane of growth plate chondrocytes from which they derive (73). However, SM is also one of the lipids most rapidly degraded when MVs are incubated in synthetic cartilage lymph and allowed to mineralize (207). This indicates that a highly active SMase must be operative during MV mineralization. SMases hydrolyze SM to ceramide (N-acyl sphingosine) and PCho. Since the pH of the MV incubation was 7.5, it can be assumed that the activity observed is from one of the neutral SMases (nSMase). In fact three distinct nSMase isoforms have been identified: nSMase-1, -2, and -3; they differ in subcellular localization, tissue specificity, and enzymatic properties (252-253). A likely candidate for the MV enzyme is nSMase-2, a Mg²⁺-dependent isozyme, localized to the plasma membrane (254-257). Its activity is enhanced by phosphatidylserine (PS) (258), a lipid specifically enriched in the MV inner leaflet (259). nSMase-2 is also known to be localized to the inner leaflet of the plasma membrane (260-261). Upregulation of nSMase-2 has been shown to cause arrest of the cell cycle in the G0/G1 phase (262); upregulation of nSMase-2 is also associated with programmed cell death (263). Based on gene sequencing, nSMase-2 is now known to be identical with sphingomyelin phosphodiesterase-3 (SMPD3). Two independent studies have shown that knockout of the SMPD3 gene in mice causes skeletal defects: chondrodysplasia and dwarfism (264) or osteogenesis and dentinogenesis imperfecta (265). These findings may relate to the role of nSMase-2 in MVs. Features most directly relevant to MV mineralization are: 1) its product, PCho, is a key substrate for PHOSPHO1, a MV enzyme that releases Pi into the vesicle lumen; and 2) that breakdown of SM would help destabilize the MV membrane and facilitate outgrowth of mineral from the vesicle lumen into the extracellular matrix.

8.8.5. Other MV phospholipases

As noted above, PI is very rapidly degraded during the onset of MV mineralization (207); after only 4 h incubation 80% of the PI is lost. This loss is not accompanied by any increase in LPI, suggesting that PI may be degraded by a PI-specific PLC activity.

8.9. Phosphatidylserine synthases

Based on changes that occur in the levels of phosphatidylserine (PS) during MV mineral formation, there must be unique pathways for metabolizing this important lipid. PS binds to Ca²⁺ and Pi to form lipid-Ca²⁺-Pi complexes (266) with the ability to rapidly nucleate hydroxyapatite (HA) mineral formation when incubated in synthetic cartilage lymph (SCL) (137). During MV mineralization, after initial rapid and almost total loss of non-complexed PS, the levels of PS, and its LPS form, progressively increased during the remainder of the incubation period and were largely complexed with the newly forming mineral (207). These are remarkable findings. The only plausible explanation for this increase is the presence of the non-energy-dependent, Ca²⁺-dependent, base-exchange mechanism of PS-synthase previously noted in growth plate cartilage (267) and MVs (268). The free amino acid, serine, undergoes Ca²⁺-dependent exchange with either the ethanolamine group of PE, or the choline group of PC, to form PS (269). ATP is not required for this reaction. Normally, PS is synthesized in mitochondria-associated membranes by action of either of two PS synthases that cause exchange of serine for choline (PS synthase-1) or for ethanolamine (PS synthase-2) (270). It is probable these PS synthases become incorporated into the plasma membrane upon transfer of Ca²⁺ (and Pi) during MV formation (8, 271-275). PS synthases are ~56 kDa, multi-pass integral membrane proteins (276-279) that could be transferred directly to the MVs. That this actually occurs is evident from the fact that (³H)serine is incorporated into PS of isolated MVs; it occurs by a mechanism that requires Ca²⁺, but not nucleotide triphosphates (268). This provides a plausible explanation for the ~5-fold enrichment of PS in MVs compared to the cells from which they derive (29, 30). Intraorganelle transport of PS, known to occur in all cells (280), also helps explain how PS synthesized in mitochondria-associated membranes becomes transferred to the plasma membrane and ultimately to MVs.

8.10. Phospholipid flippases and scramblases

Flippases are P-type ATPases that mediate the unidirectional "flipping" of PS from the outer leaflet of the plasma membrane to maintain the asymmetric transbilayer distribution of PS (281). On the other hand, scramblases are enzymes responsible for the bidirectional translocation of phospholipids between the two leaflets of the lipid bilayer of cell membranes (282-284). Scramblases are activated by Ca²⁺, do not require ATP for activity, and tend to randomize the distribution of PS in the bilayer (285). In humans, phospholipid scramblases constitute a family of five homologous proteins. In normal cells, the inner-leaflet facing the inside of the cell contains negatively-charged PS and neutral PE; the outer-leaflet, facing the extracellular environment, contains PC and SM. Scramblases transport negatively-charged phospholipids from the inner-leaflet to the outer-leaflet, and vice versa; thus, they have the capacity to randomize the sidedness of phospholipids in membrane bilayers (286). A recent report describes phospholipid scramblase activity in growth plate chondrocytes indicating its association with TNAP-containing lipid rafts during MV formation (287). Since PS is an integral component of the nucleational complex that triggers mineral formation by MVs (288), exposure to the extravesicular fluid could be a key event that enables the newly formed crystalline mineral to propagate into the extracellular matrix. Failure to do this is an apparent feature in the TNAP knockout mouse; the paradoxical finding is that while mineral forms normally within the MVs, it does not become externalized (289). Taken with the finding that scramblase is associated with TNAP in the cell membrane, externalization of newly forming mineral in MVs must require the presence of TNAP - as well as scramblase activity. The only study in which the sidedness of the two key amine-containing phospholipids, PS and PE, has been analyzed in the MV membrane revealed that ca. 80% of the PS remains localized within the MVs, whereas 50% of the PE was external (259). The fact that PS in MVs remains largely internal is significant because one of the signal features of cellular apoptosis is the externalization of PS to the outer leaflet (132, 290). Hence, MVs are not typical apoptotic bodies. The PS-Ca²⁺-Pi complexes formed during MV construction resist scramblase activity because the complexed calcium phosphate stabilizes the membrane.

9. MATRIX VESICLE MEMBRANE SURFACE RECEPTORS AND TRANSPORTERS

9.1. Cell Surface receptors

Proteomic analysis of isolated MVs also revealed the presence of numerous cell surface receptors, but none would be considered to be quantitatively major proteins (78). The most abundant receptor was CSG-HT7 (Mascot Score = 521), a cell surface glycoprotein known to be associated with the ingrowth of vascular endothelial cells (291-292). It is probable that the presence of this MV protein serves as part of an angiogenesis signal, once mineralization of the cartilage matrix has occurred. The next most abundant cell surface protein (Mascot Score = 433) was CDON, a cell surface receptor of the immunoglobulin (Ig)/fibronectin type III repeat family; however, its specific function has not been elucidated (293-294). Numerically, the most well represented receptors in avian MVs were the four integrins: alpha-5, beta-1, alpha-3, and beta-5, which were detected at (Mascot Scores) of 369, 337, 169, and 71, respectively. These integrins undoubtedly facilitate anchoring of the MVs to the extracellular matrix collagens (295-300). Another receptor, CD36, (Mascot Score 131), is an integral membrane protein found on the surface of many vertebrate cell types. It is a member of the class B scavenger receptor family of cell surface proteins that bind to many ligands including collagen. It has been shown very recently to be a pattern recognition receptor in hypertrophic chondrocytes (301). CD36 may function in opsinization of MV remnants during chondroclastic resorption of calcified cartilage at the time of replacement by cancellous bone. All of these receptors appear to be vestiges of the plasma membrane of the hypertrophic chondrocytes; they appear to function in growth plate development rather than in MV mineralization per se.

9.2. ATP-driven ion pumps and transporters

Many different transporters were detected in the proteome of MVs. They ranged in abundance from Mascot Scores (MS) of 619 down to 68. All appear to represent porters present in the plasma membrane of the chondrocytes from which the MVs derived. Of the 23 transporters detected, the most abundant, quantitatively, were five Na⁺/K⁺ ATPases (alpha-1, (MS - 619), alpha-2, (MS - 397), beta-3, (MS - 193), beta-1, (MS - 127), and beta-2, (MS - 96)), and three H⁺-ATPases (subunit B, (MS - 248), subunit A, (MS - 271), and subunit S1, (MS - 94)). To be active in MVs, these ion pump ATPases would require the presence of ATP, which could arise only from synthesis by the parent chondrocytes. However, the levels of ATP in MVs are very low (302) and are insufficient to drive Na⁺ extrusion and K⁺ uptake. This is borne out by direct analysis of the electrolytes of MVs, which were found to be much richer in Na⁺ than K⁺ - the opposite of what occurs in the cells from which they derive (63). Thus, although abundant, it is apparent that these ATP-driven ion pumps do not support ion transport activity in MVs. Also notably abundant was a choline transporter (MS - 409), which may have important activity in conjunction with the product of the PHOSPHO-1 enzyme that hydrolyzes P-choline and P-ethanolamine to Pi and their respective bases (193, 202). It is noteworthy that choline has been found to be a better co-substrate for Pi uptake by MVs than was Na⁺ (208, 303).

9.3. Channel proteins and solute-carrier transporters

Three significantly abundant ion channels were detected by proteomic analysis of MVs. Of special interest was TRPV4 (MS - 350), a transient potential cation channel that appears to be important in regulation of osmotic pressure (304-305). In the maturing growth plate chondrocytes, the presence of TRPV4 very likely has an important function in the major osmotic swelling that occurs upon hypertrophy when cell volume expands over 9-fold going from the proliferating to the hypertrophic state (306). Also of interest are CLIC4 (MS - 359) - the Channel 4 intracellular Cl^- transporter (307-309); this channel protein transports Cl^- in exchange for HCO₃^- (310-311), and appears to be important in buffering the internal pH of MVs. The third is VDAC-2, (MS - 381) - the Channel 2 voltage-dependent anion-selective channel present in the mitochondrial outer membrane (312-313). This channel may play a part in the mitochondrial permeability transition pore that becomes activated prior to formation of MVs by hypertrophic chondrocytes (314). Although the Mascot Score of VDAC-2 is not large, it may represent a mitochondrial outer membrane protein incorporated into MVs. The most numerous of all the transporters detected by proteomic analysis of isolated avian growth plate MVs were the solute-carrier family (SLC-family). In total nine different SLC carriers were detected. The most abundant was SLC 16A1 (MS - 283), a transporter known to transport lactate, pyruvate and other monocarboxylates in the plasma membrane of many cell types (315-316). Other members of the SLC family are transporters for neutral amino acids (SLC38A3, (MS - 176)), alanine/serine/cysteine/threonine (SLC1A4, (MS - 90)), organic cations, e.g. carnitine (SLC22A16, (MS - 177), glycine (SLC6A9, (MS - 137)), Na⁺-dependent citrate (SLC13A5, (MS - 96)), Na⁺-independent cationic amino acids (SLC7A3, (MS -74)), and equilibration of nucleosides (SLC29A1, (MS - 71)). Each of these carriers are of undoubted importance during the proliferation of the growth plate chondrocytes, but probably have only minor effects on the activity of released MVs. They appear to be carryovers from the plasma membrane of the chondrocytes. However, within the SLC families of carriers, of special interest to MV function are two Zn²⁺ transporters - SLC39A14, (MS - 80) and SLC39A8, (MS - 79). Zn²⁺ is remarkably abundant in MVs (317) and it is possible that these carriers are involved in its uptake, most likely into the chondrocytes before MV formation.

9.4. Ca²⁺ channel proteins

Several types of Ca²⁺ channels are differentially expressed in growth plate chondrocytes. In early stages of growth plate development, Ca²⁺ channels of the L-type are evident; however as the calcification front is approached, the putative AnxA5 type of channel becomes dominant. While there is no evidence for L-type Ca²⁺ channels in MVs, as noted earlier, AnxA5, AnxA2 and AnxA6 - some of the most dominant proteins in cartilage MVs - may serve that function, although recent data render this unlikely. This is because AnxA5 failed to support Ca²⁺ uptake in liposomes modeled after MVs, and which readily acquired Ca²⁺ when treated with several Ca²⁺ ionophores (318).

9.4.1. L-Type Ca²⁺ channels in growth plate chondrocytes

Studies in the early 1990s indicated that L-type voltage-dependent Ca²⁺ channels are operative in growth plate chondrocytes (319-320); this was initially based on their response to various benzothiazepines that selectively block L-type Ca²⁺ channels. Their presence was later confirmed using immunostaining with antibodies to the alpha subunit, revealing that they are expressed at all stages of the growth plate (321). These channels appear to participate in chondrocyte proliferation and early differentiation. The alpha-1 subunit of L-type Ca²⁺ channels mediates influx of Ca²⁺ into cells upon membrane polarization (322-323). This subunit contains 24 transmembrane segments that form a highly selective pore through which Ca²⁺ passes into the cell (324-325). The structure of L-type Ca²⁺ channels is highly complex and consists of alpha-1, alpha-2/delta, beta, and gamma subunits in a 1:1:1:1 ratio (326). In resting zone chondrocytes, combined activities of L-type Ca²⁺ channels (327) and Ca²⁺-ATPase (328) appear to maintain mean intracellular Ca²⁺ activity at around 50 nM; however, as the cells proliferate and approach the zone of hypertrophy, average basal Ca²⁺ levels rise to >100 nM (329). Confocal in situ imaging of proliferating chondrocytes in slices of PBS-bathed tibial growth plate tissue reveal dynamic fluctuations in Ca²⁺ activities in different regions of the cell (306). However, L-type Ca²⁺ channels would not be operative in MVs; they lack the voltage differential needed to drive Ca²⁺ uptake.

.

9.4.2. Annexins as Ca²⁺ channels

As alluded to earlier, SDS-PAGE analyses and initial characterizatoin by Genge et al. revealed that the dominant proteins in MVs isolated from growth plate cartilage were AnxA5, AnxA2 and AnxA6 (60, 61). AnxA5 was shown to be progressively and powerfully expressed in growth plate cartilage (Figure 6) (94); cells in the proliferative zone scarcely expressed the protein. Maximal expression was seen in the hypertrophic zone where MV formation is most active. Studies by Arispe et al. were the first to provide apparent evidence that AnxA5 exhibits Ca²⁺ channel activity when inserted into acidic phospholipid bilayers (125) - behavior predicted by Huber, based on the hydrophilic pore in its crystal structure (123) - but the experimental conditions required were not physiological. Eight years later, studies by Kirsch et al. confirmed elevated expression of AnxA5 in hypertrophic chondrocytes (330); they subsequently suggested that AnxA5-mediated Ca²⁺ influx may regulate growth plate hypertrophy and apoptosis (331). However, in studies using large unilamellar liposomal models of MVs designed to demonstrate under physiological conditions that AnxA5 exhibits Ca²⁺ channel activity, AnxA5 failed to support significant Ca²⁺ uptake (318). So far, only under artificial in vitro conditions does AnxA5 appear to exhibit Ca²⁺ channel activity. Its primary function appears to be in the nucleational core that enables MVs to induce mineral formation (288); incorporation of AnxA5 markedly accelerates the onset of mineral formation (131, 136, 332).

9.5. Inorganic phosphate (Pi) transporters

Uptake of Pi by all living cells is essential for survival and growth, inasmuch as phosphate is essential for synthesis of DNA, RNA, as well as all nucleotides involved in cellular metabolism and replication. In the "housekeeping" functions of vertebrate cells, Pi transporters use an inwardly directed electrochemical Na⁺ current to support Pi influx. This Na⁺ gradient, produced by the action of a Na⁺/K⁺ ATPase in the plasma membrane, carries Pi into living cells. However in MVs, Na⁺ levels are as high as that of the extravesicular fluid (63); thus there is no Na⁺ gradient to drive Pi uptake.

9.5.1. Na⁺-dependent Pi transporters

In vertebrates, two unrelated families of Na⁺-dependent transporters carry out general Pi transport. Remarkably, they transport different Pi species: divalent HPO₄^2- is carried by type II Na⁺/Pi cotransporters; monovalent H₂PO₄^- is carried by type III (PiT-1) Na⁺/Pi cotransporters (333). Montessuit et al. were the first to study Pi transport in MVs (334). In MVs isolated from cartilage of mildly rachitic chickens, Pi uptake appeared to be a saturable process that could be driven by establishing an artificial Na⁺ gradient (334). It appeared that Pi uptake in both MVs and chondrocytes was Na⁺-dependent (334, 335). Subsequent studies by Guo et al. have confirmed the presence of a type III Na⁺/Pi transporter using PCR analysis of RNA isolated from normal chicken tibial growth plate cartilage (336). This Pi transporter is a close homolog of human PiT-1 (Glvr-1, type III Na⁺/Pi porter) showing 77% sequence identity. It appears to be a 73.5 kDa MW integral membrane protein with 12-transmembrane domains. However, in situ hybridization revealed that this Na⁺-dependent PiT-1 Pi porter was only expressed strongly in the resting and proliferating zones; it was barely present in the hypertrophic zone (Figure 7B) (336). Subsequent studies on MVs isolated from normal chick growth plate cartilage showed that Pi transport was not strictly Na⁺-dependent. Substitution of Li⁺ or K⁺ for Na⁺ had minimal effect on Pi uptake; N-methyl D-glucamine (NMG⁺) was totally inhibitory, whereas choline⁺ was clearly stimulatory (208). Several known inhibitors of Pi transport were also used in this study. As controls, several putative intermediates in the MV mineralization pathway - ACP, PS-Ca²⁺-Pi complex (PS-CPLX) and HAP - were studied. When incubated in synthetic cartilage lymph (SCL) matched in electrolyte composition and pH to that of the native fluid, PS-CPLX and HAP caused almost immediate uptake of Pi; pure ACP required ~1 h; but MVs displayed rapid Pi uptake only after ~3 h incubation. Phosphonoformate (PFA) delayed the onset and rate of Pi uptake by the MVs with an IC₅₀ of 3-6 microM; however, it also inhibited Pi uptake by the mineralization intermediates with only slightly higher IC₅₀'s of 8.5-14 microM. Thus, PFA proved not to be a specific Pi transport inhibitor. While Na⁺ substitutions had minimal effect on HAP or CPLX-seeded mineral formation, with ACP, NMG⁺ totally blocked and choline⁺ stimulated mineralization, just as with MVs. Thus, ACP appeared to mimic MV Pi uptake.

9.5.2. Na⁺-independent Pi transporters

Studies on confluent, primary cultures of normal growth plate chondrocytes also detected a second type of Pi transport activity, expressed shortly before the onset of mineralization (303, 336). While the Na⁺-dependency of Pi uptake early in cultures of the chondrocytes was confirmed, a different Na⁺-independent Pi transporter was expressed as the cells approached hypertrophy. In contrast to Na⁺-dependent Pi transporters whose activity is upregulated by Pi insufficiency (337, 338), this porter was upregulated by elevated Pi levels typical of the extracellular fluid of growth plate cartilage. The activity of this Na⁺-independent Pi transporter was Cl^-dependent; but Cl^- was not co-transported (336). It had a slightly alkaline pH optimum (7.6-8.0) appropriate for the pH of growth plate extracellular fluid (63, 339). Choline⁺ was nearly as effective as Na⁺ as a co-substrate (336). The apparent K_m for Pi was 0.59�0.01 mM; its V_max was 0.99�0.01 nanomoles/mg protein/min. (For comparison with non-calcifiable cells, Pi uptake by Chinese hamster ovary cells tested under identical conditions had an apparent K_m for Pi of 0.046 mM and a V_max of 0.34 nanomoles/mg protein/min.) The activity of this Pi porter appears to represent the Na⁺-independent action of the mitochondrial PiC transporter that is active when mature growth plate chondrocytes load large amounts of Ca²⁺ and Pi prior to the onset of growth plate mineralization. Support for this concept comes from the finding that the mitochondrial PiC Pi transporter is strongly up-regulated in hypertrophic chondrocytes, the opposite of the Na⁺-dependent PiT-1 plasma membrane Pi transporter (Figure 7A) (336). This will be discussed further when MV formation is considered.

10. MATRIX VESICLE REGULATORY PROTEINS

About a dozen proteins involved in regulation of chondrocytes were found by proteomic analysis to be present in significant levels in MVs (78). While it can be assumed that these proteins are carry-overs from the cells that formed the MVs, they do not appear to exert a major effect on mineral formation. However, some appear to be involved in the formation of MVs.

10.1. Syntenin

The most abundant regulatory protein identified in MVs was syntenin (MS - 1158), a protein that links syndecan to the cytoskeletal protein, actin (340-341). Studies have shown that syntenin, syndecan, and actin co-localize to filamentous cell process in endothelial cells (342-343); these short cellular "villi" are closely analogous to those seen in freeze-fracture images of growth plate chondrocytes (50). However, there have been no studies to date on the role of syntenin in growth plate development. Early studies showed that drugs that disrupt actin filament formation clearly enhance release of MV in growth plate chondrocyte cultures (69, 73). On the other hand, syndecan-3, a cell-surface heparan sulfate proteoglycan that binds to syntenin, appears to play an important role in chondrocyte proliferation (344-345). Syntenin is a 32 kDa cellular PDZ-containing adaptor protein (346) that binds to the cytoplasmic domains of the vertebrate syndecans; its binding to syndecan requires the cooperation of both PDZ domains of syntenin (346). Syndecans act as co-receptors involved in docking of heparin-binding growth factors and adhesion molecules (348). Of special interest to MV formation, the development of cell membrane protrusions requires dephosphorylation of tyrosine residues in syndecan-1 (349). Apparently, at the tip of these processes, syntenin binds to syndecan, which in turn binds to type VI collagen in the extracellular matrix (348), leaving MVs attached at the points of contact. It is likely that additional matrix proteins (e.g. fibrillin, versican or proteoglycan link proteins), as well as the integrins, are also involved in MV formation.

10.2. G_i Protein, alpha-2

The alpha-2 subunit of G_i protein is a heterotrimeric subunit that inhibits the production of cAMP from ATP. Gi-alpha-2, with a proteomic score (MS - 660) appears to occur in significant abundance in MVs (78). In the cellular environment of growth plate cells, the function of G_i-alpha-2 is mainly to suppress cAMP dependent pathways by inhibiting adenylate cyclase activity, decreasing the production of cAMP from ATP. In turn, this would result in decreased activity of cAMP-dependent protein kinases (350). Thus, the ultimate effect of G_i is the opposite of cAMP-dependent protein kinases. In the cell, G_i protein also appears to play a role in activation of the phospholipase C pathway (351, 352). Considering its relative abundance in MVs, Gi-alpha-2 may well be important in activating the phospholipase C activity responsible for the breakdown of the MV phospholipids (PC and PE) that occurs during MV mineralization (207). (See discussion of MV phospholipases, Section 7).

10.3. Tyr/Trp mono-oxygenase activation proteins, etc.

The theta and beta subunits of the mono-oxygenase activating protein are relatively abundant in MVs, (MS - 725) and (MS -307), respectively. They are protein kinases that enhance the activity of Tyr/Trp hydrolases that would lead to the formation of serotonin and Dopa in GP chondrocytes. These gene products belong to the 14-3-3 (MS - 539) family of proteins that mediate signal transduction by binding to phosphoserine-containing proteins. This highly conserved protein family is found in plants and mammals. These enzymes appear to be carryovers from the plasma membranes of the chondrocytes that form MVs.

11. CYTOSKELETAL PROTEINS

Cytoskeletal proteins have been repeatedly implicated in the formation of membrane-derived vesicle-like structures in a variety of systems, including red blood cells, intestinal brush border cells, liver cells, and a number of other cell types (353-355). Actin was first identified as a constituent of MVs in 1982 by Muhlrad (92). Actin was subsequently implicated in MV formation by studies on primary cultures of growth plate chondrocytes showing that agents affecting the stability of actin microfilaments significantly altered MV formation (69). Cytochalasin D, a fungal metabolite that causes depolymerization of actin microfilaments greatly enhanced MV formation. In contrast, phalloidin, an agent that stabilizes actin microfilaments, significantly decreased MV formation. Another drug, colchicine, which destabilizes microtubules, also stimulated the release of MVs. In an earlier study, Hale showed that the localization of cytoskeletal proteins in cultured growth plate chondrocytes indicated that both microtubules and microfilaments were involved in the formation of cell processes from which MV were shown to derive (68). Subsequent studies, in which the enzyme, protein, and lipid composition of chondrocyte microvilli and MVs were compared, indicated that vesicles must derive from these structures. Twenty five years later, proteomic analysis of MVs confirmed that numerous cytoskeletal proteins are present in collagenase-released MVs from embryonic chick femoral cartilage (78). Of these, by far the most abundant was beta-actin (MS = 1139). Among other cytoskeletal proteins detected were: filamin B, beta (MS - 398), actinins alpha-1 and -4 (MS - 698), tubulins beta-6 and alpha-8 (MS - 840), radixin (MS - 439), gelsolin/vilin related proteins (MS - 433), as well as other minor proteins. The author acknowledges the online use of Wikipedia as a valuable resource for information on several of these cytoskeletal proteins.

11.1. Actin

Actin is a globular, ~42-kDa protein found in nearly all eukaryotic cells; it is typically present in high concentrations (over 100 microM). Actins are one of the most highly-conserved proteins, differing by no more than 20% in species as diverse as algae and humans. Actin represents the monomeric subunit of two types of filaments in cells: microfilaments, and thin filaments; it is a critical part of the contractile apparatus in muscle cells. Actins participate in important cellular processes including muscle contraction, cell motility, cell division and cytokinesis, vesicle and organelle movement, cell signaling, and the establishment and maintenance of cell junctions and cell shape. Many of these actions are mediated by extensive and intimate interactions between actin and cellular membranes. In vertebrates, three main groups of actin isoforms, alpha, beta, and gamma have been identified. The alpha-actins are found in muscle tissues and are a major constituent of its contractile apparatus. The beta- and gamma-actins co-exist in most cell types as components of the cytoskeleton and are mediators of internal cell motility. Beta-actin is one of six different actin isoforms that have been identified in humans and is one of the nonmuscle cytoskeletal actins. As noted above, beta-actin is a dominant MV cytoskeletal protein; and as noted earlier in cultured growth plate chondrocytes, drugs that block formation of actin microfilaments significantly stimulate the release of MVs (68, 69, 73).

11.2. Filamin B, beta

This protein, also known as FLNB, is a cytoplasmic protein that binds to actin. It regulates intracellular communication and signaling by cross-linking with actin to allow direct communication between the cell membrane and cytoskeletal network. Filamin B is thought to be important for proper skeletal development (356). Mutations in the FLNB gene are involved in several lethal bone dysplasias, including boomerang dysplasia and atelosteogenesis type I (357). The finding of filamin B in MVs (MS - 398) appears to be linked to the abundant presence of beta-actin.

11.3. Actinin

Actinin belongs to the spectrin gene superfamily that represents a diverse group of cytoskeletal proteins, including the alpha- and beta-spectrins and dystrophins. Actinins, alpha-1 and -4 are actin-binding proteins with various roles in different cell types. In nonmuscle cells, these cytoskeletal isoforms are found in microfilament bundles and adherens-type junctions, where they are involved in binding actin to the membrane. In contrast, the skeletal-, cardiac-, and smooth-muscle isoforms of actinin are localized to the Z-disc and analogous dense bodies, where they help anchor the myofibrillar actin filaments. The presence of both actinin alpha-1 (MS - 440) and alpha-4 (MS - 258) in MVs again appears to relate to their high affinity for actin.

11.4. Tubulin

Tubulin is one of several members of a small family of globular proteins. The most common members of the tubulin family are alpha- and beta-tubulin, proteins that make up microtubules. Each has a molecular weight of approximately 55 kDa. Microtubules are assembled from dimers of alpha- and beta-tubulin. These subunits are slightly acidic with an isoelectric point between 5.2 and 5.8 (358). To form microtubules, the dimers of alpha- and beta-tubulin bind to GTP (359) and assemble onto the (+) ends of microtubules while in the GTP-bound state (360). After the dimer is incorporated into the microtubule, the molecule of GTP bound to the beta-tubulin subunit can hydrolyze into GDP through inter-dimer contacts along the microtubule protofilament (361). Whether the beta-tubulin member of the tubulin dimer is bound to GTP or GDP influences the stability of the dimer in the microtubule. Dimers bound to GTP tend to assemble into microtubules, while dimers bound to GDP tend to fall apart; thus, the GTP cycle is essential for the dynamic instability of the microtubule network. Work by Hale et al. (69) reveals that tubulin depolymerization stimulates MV formation. Since GP chondrocytes that form MVs are typically in an energy-depleted state, the guanine nucleotides would tend to be in the GDP form and would thus favor disassembly of microtubules, and hence stimulate MV formation. In the MV proteome, three tubulin forms were detected: general tubulin (MS - 296), beta-6 (MS - 456), and alpha-8 (MS - 89). Since tubulin is essential for cell division, it would clearly be essential during chondrocyte proliferation. The presence of tubulin in MVs may result from being in disassembled form in the cytosol that was encapsulated during MV formation.

11.5. Radixin

Radixin is a cytoskeletal protein that acts as an adapter and signaling protein (362). It functions as a plasma membrane-cytoskeletal cross-linker in actin-rich cell surface structures. It appears to be required for cytoskeletal functions such as transmembrane signaling, growth regulation, differentiation, cell motility, adhesion and proliferation. Radixin is a member of the ERM (ezrin-radixin-moesin) protein family that share highly similar amino acid sequences. Radixin is a 583 amino acid (~80 kDa) protein with an extremely long alpha-helical domain (residues 310-477, ~240 Angstroms in length). The N-terminal FERM domain (residues 1-309) is 85-90% identical to those of ezrin and moesin and acts as a conditional membrane-binding site. The shorter C-terminal domain (residue 478-583) binds to f-actin; thus radixin forms a bridge between phosphatidylinositol 4,5-bisphosphate (PIP₂) in the plasma membrane and actin. Thus, it is not surprising to find that radixin has a significant proteome score (MS - 439); it is apparent that it plays a role in the binding of actin to the plasma membrane.

11.6. Gelsolin and proteins with gelsolin/villin domains

There were significant levels of proteins with gelsolin/villin domains in the MV proteome (MS - 356), with lesser levels of gelsolin itself (MS - 77). Gelsolin is an actin-binding protein that regulates actin filament assembly and disassembly. Gelsolin is one of the most potent members of the actin-severing gelsolin/villin superfamily; it severs f-actin with nearly 100% efficiency (363). It is an 82-kD protein with six homologous subdomains, referred to as S1-S6; each subdomain is composed of a five-stranded beta-sheet, flanked by two alpha-helices, one positioned perpendicular with respect to the strands and one positioned parallel. The N-terminus (S1-S3) forms an extended beta-sheet, as does the C-terminal (S4-S6) (364). As an important actin regulator, gelsolin plays a role in podosome formation (365-366). Podosomes are protrusions from the cell membrane that extend into the extracellular matrix (363). Apparently growth plate chondrocytes in the process of MV formation first produce actin-rich structures akin to these structures. Gelsolin inhibits apoptosis by blocking mitochondrial membrane potential loss and cytochrome C release (367). Prior to cell death (such as occurs upon chondrocyte hypertrophy), mitochondria lose membrane potential and become more permeable. Gelsolin can impede the release of cytochrome C, obstructing the signal amplification that would lead to apoptosis. Actin can be cross-linked into a gel by actin-cross-linking proteins. Gelsolin can turn this gel into a sol, hence the name gelsolin. Thus, although present in modest levels, gelsolin likely plays an active role in regulating MV formation.

11.7. Plastins

Plastins comprise a family of actin-binding proteins that are conserved throughout eukaryote evolution and are expressed in most tissues of higher eukaryotes. In humans, two ubiquitous plastin isoforms (L and T) have been identified. The L isoform of plastin is expressed only in hemopoietic cell lineages, while the T isoform has been found in all other normal cells of solid tissues that have replicative potential (fibroblasts, growth plate chondrocytes, endothelial cells, etc.). The C-terminal 570 amino acids of the T-plastin and L-plastin proteins are 83% identical and contain a potential Ca²⁺-binding site near the N terminus (368). Of special interest to MV formation, T-plastin 3 appears to contribute to the assembly and maintenance of the actin cytoskeleton of plasma membrane protrusions that are characteristic of growth plate chondrocytes in vivo and from which MVs appear to bleb. The proteomic score of plastin 3 is not high (MS - 151), but the protein would appear to be involved with MV formation.

12. MATRIX VESICLE LIPIDS

The early discovery of extracellular lipids at sites of mineralization by Irving & Wuthier (16-19) found its explanation in the discovery of membrane invested MVs (1-4). The development of facile methods for quantitation of the complex mixture of lipids present in MVs was critical to the success in this endeavor (222, 369-370). Early investigations of the lipid composition of MVs (29-31) established several unusual features. MVs were enriched in nonpolar lipids and depleted in phospholipids compared to the plasma membrane, or the total cell. The ratio of free cholesterol to phospholipids in MVs was double that of the plasma membrane, and nearly 5-fold higher than in the cells. Thus some MV lipids must not reside in the vesicle membrane, but rather in some other form (proteolipids?) within the vesicle lumen (74, 93). This may explain: 1) the marked osmiophilicity of the luminal contents of many MV (3), 2) the presence of a "low-density nucleation factor" in MVEM (66), and 3) the abundant presence of milk-fat globule protein in the proteome of MVs (78).

12.1. Phospholipids

MV phospholipid composition was significantly different from that of the cell, or 'membrane' fractions (29-31). Among the notable differences were an enrichment in phosphatidylserine (PS) (3.2-4.5 fold), sphingomyelin (SM) (2.2-2.5 fold), and total lysophospholipids (LPL) (1.8-2.2 fold), as well as depletion in phosphatidylcholine (PC) (0.6-0.7 fold) - when compared to the cells (Table 1). A similar enrichment in SM and LPL was evident in hypertrophic vs. proliferating cells of the growth plate, suggesting that phospholipases are active in the process of hypertrophy, a phenomenon that precedes MV formation. This finding agrees with earlier studies showing elevated phospholipase activity in the hypertrophic zone of growth plate cartilage (223).

12.1.1 Fatty acid composition of MV phospholipids

Fatty acid analyses of the LPL of MVs showed enrichment of saturated fatty acids (stearic) and reduction in unsaturated fatty acids (oleic and palmitoleic) compared to the diacyl forms (227). This indicates that the LPL must arise from action of phospholipase A₂ type activity. The abundant presence of zwitterionic type LPLs, and the depletion of PC in MVs, indicates that phospholipase A₂ must be active during MV formation. More will be made of this in the discussion of the mechanism of MV formation.

12.1.2. Physical form of MV phospholipids

Studies also have been directed toward elucidation of the physical form of phospholipids in MV (64). This was done to ascertain whether the acidic phospholipids present in MV were complexed with the newly forming calcium phosphate mineral. Earlier studies on the phospholipids of growth plate cartilage had shown that a large portion of the acidic phospholipids were complexed with the mineral and were extractable only after demineralization of the tissue (20, 371). These mineral-complexed phospholipids had a turnover pattern different from that of the non-complexed lipids (28). In fact, nearly half of the PS and PI present in MVs were complexed with the newly forming mineral phase (64). About 10% of the total Ca²⁺ and Pi in the vesicles was extractable into organic solvents as acidic phospholipid-Ca²⁺-Pi complexes (CPLX). This phenomenon is based on the fact that Pi markedly enhances the affinity of Ca²⁺ for acidic phospholipids like PS, giving rise to PS-CPLXs that are soluble in organic solvents (266). CPLXs have been shown to have potent mineral nucleating properties (131, 136, 221, 288, 372). In forming these complexes, it is critical that Pi be present prior to exposure of PS to Ca²⁺ (373). If Pi is not present first, PS forms a ternary (PS)₂-Ca²⁺ complex that excludes Pi binding and has no nucleational activity. This is an important clue to the sequence of events that leads to formation of PS-CPLX and formation of MVs.

12.1.3. Topological distribution of MV phospholipids

Studies by Majeska et al. also have shown that the bulk of MV PS resides in the inner membrane leaflet; surprisingly, PE a lipid typically found in the inner leaflet of the plasma membrane is equally abundant on the external membrane (259). However, this redistribution appears to result from the activity of scramblases recently shown to be present in MVs (287). The internal localization of PS is consistent with findings in many other tissues (374-376) and has important physiological ramifications. The formation of the PS-CPLXs in MVs would occur inside the vesicle lumen and would not be accessible to the extracellular fluid until the membrane was breached. While phospholipase and scramblase activities appear to be involved in MV formation, since much of the PS is complexed with Ca²⁺ and Pi, it is resistant to the action of these enzymes, as has been shown by Boskey et al. (377).

12.1.4. Ca²⁺-binding properties of MV phospholipids

While it was known that PS binds with moderate affinity to Ca²⁺ (24-27), Holwerda et al. (373) determined that the K_f of Ca²⁺ for PS was 3 x 10⁴ using biphasic solvent partition assays. Through use of ¹³C and ³¹P nuclear magnetic resonance (NMR) studies of Ca²⁺-binding to PS and PE, it was established that both the carboxylate and phosphate groups were essential for Ca²⁺ binding; and that the amino group of PS exerts a significant repulsive effect (373). Later, it was discovered that the annexins bind to Ca²⁺ in an acidic phospholipid-dependent manner (378-379). The affinity of binding between the annexins and the acidic lipids is such that physiological changes in cytosolic Ca²⁺ ion activity can markedly alter the association between these proteins and PS on the inner surface of the plasma membrane. Acidic phospholipid-dependent Ca²⁺-binding annexins have been found to constitute a major fraction of the MV proteins (60, 61) and have PS-dependent Ca²⁺-binding constants in the 250-500 nM range (Figure 5). One of the earliest biochemical clues implicating acidic phospholipids in the mechanism of growth plate calcification was the discovery that substantial amounts of acidic phospholipids, and PS in particular, are complexed with Ca²⁺ in the mineralizing tissue (19, 20). Growth plate cartilage was dissected from successively more calcified zones: resting, proliferating, hypertrophic, and calcified cartilage, and compared with cancellous and compact bone. With increasing mineralization, progressively more and more of the acidic phospholipids could not be extracted until the tissues were demineralized with EDTA (371). Most of the Ca²⁺-bound phospholipids are present in the form of PS-CPLXs.

12.2. Nonpolar lipids

Compositional analysis of the nonpolar lipid fraction from MV (30) revealed significant differences from that of the plasma membrane or the total cell. Triacylglycerols (TG) were the major nonpolar lipid in cells, accounting for 44.0�0.7% of the total; this was followed by free cholesterol (19.3�1.9%), free fatty acids (10.2�2.0%), cholesterol esters (8.4�0.8%) and 1,2-diacylglycerols (6.4�0.4%). Only trace amounts of 1,3-diacylglycerol (0.3�0.2%) were present, although there was ~10% of partially characterized hydroxysterols (223). In contrast, in MV, free cholesterol was dominant, accounting for 31.7�4.8% of the nonpolar lipid. Levels of TG in MV were only half that seen in the cells (22.4�3.6%); levels of free fatty acids, cholesterol esters and 1,2-diacylglycerols were not significantly changed, whereas the level of 1,3-diacylglycerols increased significantly to 2.5�0.7% of the total nonpolar lipid. Thus, MVs were significantly richer in free cholesterol and depleted in triacylglycerols compared to the plasma membrane and total cells. The finding of marked enrichment in free cholesterol, the high levels of SM and polyhexosyl ceramides (see next), as well GPI-anchored TNAP in MV (207) suggests that formation of lipid platforms is involved in MV formation (287) and may contribute to the nucleation of mineral formation by MVs.

12.3. Glycolipids

The non-phosphorus-containing polar lipid composition of MVs, plasma membranes and total cells of avian growth plate cartilage has been only minimally explored (30). While these polar lipids account for a significant amount of the total lipid, they have not been well characterized. (There is no evidence that these glycolipids are complexed with the mineral phase of MVs.) MVs and the plasma membrane fraction of growth plate chondrocytes were found to be enriched in the polyhexosyl ceramides (24.6% in MVs vs. 7.0% in cells) and acidic glycosyl ceramides (24.1% in MVs vs. 8.4% in cells). Consistent with this early finding of acidic glycosyl ceramides in MVs, over 30 years later studies on detergent-resistant membrane fractions of growth plate chondrocytes report the presence of ganglioside-1 (GM1) as the principal neuraminic acid-containing glycosyl ceramide (287). These membrane fractions also contain GPI-anchored TNAP that becomes incorporated into MVs. These chemical features, as well as morphological studies, all indicate that MVs derive from the plasma membrane.

13. MATRIX VESICLE ELECTROLYTES

Endochondral calcification is a complex process in which Ca²⁺ and Pi (HPO₄^2- and H₂PO₄^-) in solution are processed and brought together to induce precipitation of a mineral phase that ultimately forms hydroxyapatite (Ca²⁺₁₀∙PO₄^3-₆∙OH^-₂, HAP), the principal mineral of bone. To properly understand the role of MVs in the induction of growth plate mineralization, the levels of MV electrolytes 1) need to be compared with: 2) those present in the intracellular fluid of chondrocytes from which the MVs form, 3) the extracellular fluid (ECF) where mineralization occurs, and 4) the blood plasma - the central source of electrolytes that is controlled by various endocrine systems to maintain relatively constant levels. Presented in Table 2 are chemical analyses of levels of major ion species in these four compartments. Because the precipitation process involves interactions between the ionic activity of Ca²⁺ and HPO₄^2-, and also other ionic species, the levels of these and other electrolytes are expressed as: a) total, b) ultrafilterable and c) ion activities. Mean values for Na⁺, K⁺, Mg²⁺, Ca²⁺, Pi, Cl^- and HCO₃^- from growth plates of rat (339) and chicken cartilage (63) are presented. Supporting data obtained from ultrastructural electron probe analyses (380) are also considered. The physical properties of the mineral present in MVs are also discussed in detail.

13.1. Chemical analyses of intracellular electrolytes in growth plate chondrocytes

Electrolyte levels in living cells constitute an analytical challenge because of their dynamic character. This is particularly true of cells entrapped in a dense proteoglycan- and type II collagen-rich matrix such as occurs in growth plate cartilage. Different approaches have been taken to solve this problem. First, by careful release using controlled enzymatic digestion of the matrix proteins, chondrocytes can be obtained in essentially intact form, particularly if they are quickly restored to an ionic environment (e.g. synthetic cartilage lymph, SCL) that matches the electrolyte composition and osmolarity of the native extracellular fluid. However, when feasible, ion activities are even better assessed in situ using fluorescent probes specific for the ion in question. Using the first approach, we performed zonal dissection of the tibial growth plates of young (6-8 week old) rapidly growing chickens to obtain tissue from the proliferative and hypertrophic zones. The cells were released and equilibrated in SCL containing glucose to restore native electrolyte values (63). ³H-labeled H₂O and ¹⁴C-labeled inulin (a high molecular weight inert polysaccharide that does not penetrate into living cells) were then used to enable determination of the intracellular volume for accurate electrolyte concentrations in these living cells (63).

Table 2 presents data from that study. Shown in the first two columns are the cellular concentrations of the major electrolytes. Immediately obvious is the fact that the proliferating cells, like all normal vertebrate cells, were rich in K⁺ (~120 mM) and relative low in Na⁺ (~40 mM). K⁺ levels in the hypertrophic cells were slightly lower (~100 mM) and Na⁺ levels higher, evidently because ATP levels were reduced as these cells approach their apoptotic demise. The levels of total Mg²⁺ were relatively high (~9 mM) typical of living cells; total Ca²⁺ levels was much lower (1.5-2.0 mM). However, both of these divalent cations are largely protein bound - as is evident in the marked reduction in the ultrafiltrates (Mg²⁺ = ~1.5 mM, Ca²⁺ = 0.10-0.15 mM). Concerning the anions, note that intracellular Cl^- (~30 mM) is markedly lower than is seen in the extracellular fluid (~110 mM) - as is typical of all living vertebrate cells. The only remarkable difference from typical soft tissue cells is the presence of high levels of Pi (22-23 mM), a value consistently observed in growth plate chondrocytes. This appears to be a modification designed to facilitate subsequent mineral formation. Note further that overall average cellular Ca²⁺ ion activities (0.4-0.5 micromolar), determined in situ using a Ca²⁺-specific fluorescent probe in thin slices of live growth plate tissues incubated in phosphate-buffered saline (306), are only 1/250 - 1/300 of the ultrafilterable levels. However, average Ca²⁺ activity levels are poor indicators of the dynamics of chondrocyte Ca²⁺ metabolism. Major differences in local intracellular Ca²⁺ levels in different zones of the growth plate have important bearing on MV formation.

13.2. Chemical analyses of electrolytes in isolated matrix vesicles

In comparing electrolyte levels in MVs with those in the cells from which the vesicles arise, note that the Ca²⁺ and Pi levels represent those present in nascent vesicles. Much additional acquisition of Ca²⁺ and Pi occurs rapidly upon exposure of the MVs to the extracellular fluid. Note first that the levels of total Pi (35-40 mM) in the MVs are not all that much higher than those present in the cells (20-25 mM). In contrast, note the markedly higher levels of total Ca²⁺ (60-70 mM) in MVs compared to the average levels of total Ca²⁺ (1-2 mM) in the cells. (How this remarkable enrichment occurs will be delineated in a subsequent section). The levels of total Ca²⁺ and Pi in MVs are also much higher than those in the ECF. This seems paradoxical from a thermodynamic perspective - until one examines MV Ca²⁺ ion activity, which is only 1.2-1.5% of that in the extracellular fluid. (MV Ca²⁺ ion activity levels were analyzed in situ as just described for the chondrocytes.) Another remarkable finding is that levels of Na⁺ (~165 mM) and K⁺ (~20 mM) in the MVs are just the reverse of what is seen in the chondrocytes from which they arise. MV Na⁺ and K⁺ levels are very similar to those in the ECF and blood plasma. These levels of MV Na⁺ and K⁺ show that despite the presence of Na⁺/K⁺-ATPase protein in MVs noted previously (78), the pump is obviously not operational. This is due to lack of ATP needed to drive the ion pump. Other MV electrolyte features justify comment. The pH of the nascent MVs is only about 6.8, similar to that in the perimeter of the hypertrophic cells from which the vesicles arise (307). However, after exposure to the ECF, the intravesicular pH rises significantly to ~7.6 (77). This appears to be in response to carbonic anhydrase activity that is part of a two-component system for regulating intravesicular pH (160). In addition to carbonic anhydrase that enables CO₂ to combine with H₂O to form carbonic acid (H₂CO₃), there is a HCO₃^-/Cl^- exchange channel (311) that enables entrance of HCO₃^- to buffer H⁺ released when Ca²⁺ reacts with HPO₄^2-, to form ACP (Ca²⁺₃∙PO₄^3-₂) ion clusters in route to OCP and HAP formation. This specific intravesicular buffer system is vital to mineral induction.

13.3. Chemical analyses of electrolytes in the extracellular fluid and blood plasma

As is true of all vertebrate species, Na⁺ and Cl^- comprise the major ion-pair in blood plasma (Table 2), with levels of ~150 mM and ~120 mM, respectively. In cartilage ECF Na⁺ levels are significantly lower (115-117 mM) because of extensive binding to proteoglycans that are present in high levels (25-35 mg/mL). The levels of Na⁺ in ultrafiltrates of blood plasma are significantly reduced - compared to the total - because of binding to plasma proteins (e.g. albumin and globulins). Cl^- levels in ultrafiltrates are much less affected, presumably because proteins generally do not tightly bind Cl^-. Levels of ultrafilterable K⁺ in cartilage ECF are higher than those in blood plasma; they appear to be caused by Ca²⁺-activation of a K⁺ channel in growth plate chondrocytes (381). The levels of total Ca²⁺ in blood plasma (3.2 mM) and cartilage ECF (3.1-3.3 mM) are very similar; however, major reduction is seen in ultrafilterable Ca²⁺ in cartilage ECF (1.2-1.3 mM) compared to plasma (2.0 mM). In cartilage ECF only about 35-45% of the Ca²⁺ is not protein bound (i.e. ultrafilterable), whereas in blood plasma about 67% is ultrafilterable. In cartilage fluid this reduces the activity of Ca²⁺, but the proteoglycan-bound Ca²⁺ can act as a reservoir - available for subsequent mineral growth. On the other hand, the levels of Pi in the ECF ultrafiltrate are 10-15% higher than the levels in plasma, presumably from intracellular leakage. This would provide a significant boost in the ability of the ECF to support apatitic mineral growth. Note that substantial levels of HCO₃^- appear in both blood plasma and cartilage ECF; this anion provides part of the important buffer systems in both tissue fluids. Some data presented in Table 2 are derived from chemical analyses of electrolytes in extracellular fluid of growth plates made by Howell and Pita using their elegant micropuncture techniques (340). The bulk of the data are from expressed extracellular fluid from premineralizing and mineralizing zones of rapidly growing broiler-strain chickens (63, 64). Direct chemical analyses of electrolytes were also made on isolated cells and MVs (63, 64, 318) as well as MV-enriched microsomes isolated by non-enzymatic methods (55, 56).

13.4. Ultrastructural analyses of Ca²⁺ and P in chondrocytes and MVs

Electron probe analyses of in situ Ca and P in the cells and matrix of chicken growth plate cartilage have been reported (381). It is important to note that use of this method to determine concentrations of Ca²⁺ and P is problematic because of the difficulty in measuring the volume of tissue analyzed. Nevertheless, the general patterns obtained by chemical and microprobe analyses agree quite well, qualitatively. Chemical analyses of MVs isolated from the proliferative and hypertrophic zones yielded levels of total Ca²⁺ (43 � 3, and 98 � 14 mM, respectively) (63); electron probe in situ estimates of MV levels from similar zones reported values of Ca²⁺ 277 � 46, and 882 � 101 mM, respectively - values 5-9-fold higher. Similarly, chemical analyses of total inorganic P (Pi) in MV isolated from the two zones were 30 � 4, and 50 � 10 mM, respectively; electron probe estimates of total P (organic and inorganic) yielded values of 173 � 32, and 603 � 69 mM, respectively - values 6-12-fold higher. Thus, while the quantitative values obtained do not agree, qualitatively they show the same pattern. In either case, it is obvious that the levels of Ca²⁺ and Pi present far exceed what could occur free in solution. Using a special potassium pyroantimonate Ca²⁺ stain, Brighton and Hunt carefully studied the levels of Ca²⁺ in chondrocytes at different zones of epiphyseal cartilage of rat costal cartilage (Figure 8A-C) (382-384). Their findings showed that growth plate chondrocytes accumulated Ca²⁺ in their mitochondria reaching maximal levels in the hypertrophic cells just prior to the onset of mineral formation. Their studies indicated that this stored Ca²⁺ was transferred via cytoplasmic processes to the extracellular matrix in the form of Ca²⁺-loaded MVs. Although controversial at the time of first discovery subsequent biochemical and confocal studies using ion probes for Ca²⁺ have essentially confirmed these early observations.

13.5. Chemical analyses of microelements in MVs

Sauer, in particular, focused attention on levels of various microelements in MVs (317). These studies were prompted by the discovery that chelators of transition metal ions (e.g. o-phenanthroline) markedly stimulated the rate of Ca²⁺ loading by mineralizing MVs (76). These analytical studies revealed remarkable levels of total Zn²⁺ (1.0-1.5 mM) in MVs. Levels of other trace elements were at background values, although low levels of total Cu²⁺ (50-100 microM) were detected. Further studies revealed that addition of 5 microM Zn²⁺ to the mineralizing medium markedly inhibited MV Ca²⁺ loading (385). This will be detailed later when the regulatory role of Zn²⁺ in the induction of mineralization by MV is considered.

13.6. Physico-chemical properties of MV mineral forms

Knowledge of the chemical form of the mineral ions at different stages of MV development is essential for proper understanding of the mechanism of MV mineral formation. For example, crystalline HAP is not the first mineral formed by MV; to the contrary, there is strong evidence that ACP and complexes of it with certain proteins and lipids are the first solid- phase mineral form present in MVs. This is evident from the following data.

13.6.1. X-ray Diffraction Analyses

A variety of physico-chemical methods have been applied to ascertain the physical form of Ca²⁺ and Pi present in MV. Early X-ray powder diffraction studies on freeze-dried collagenase-released MV isolated from avian growth plate cartilage revealed that no crystalline HAP mineral is present initially (63) (Figure 9). Nevertheless, chemical analyses reveal that nascent vesicles contained large amounts of Ca²⁺ and Pi - much more than could exist in solution phase. It is now apparent that this is largely amorphous (ACP), but because of its diffuse X-ray diffraction pattern (386), it is obscured in the diffuse diffraction halo of the MV proteins and lipids. X-ray diffraction studies on MV-enriched microsomal fractions purified by isosmotic Percoll density gradient fractionation, also revealed the absence of crystalline mineral (57). However, upon incubation in SCL for 16 h these MVs accumulated large amounts of Ca²⁺ and Pi, which X-ray diffraction analysis then clearly showed was HAP-like crystalline mineral.

13.6.2. FT-IR and FT-raman analyses

Using solid-state spectroscopic (FT-IR and FT-Raman) methods, Sauer et al. also examined the initial mineral phases present in nascent MVs, as well as that subsequently formed after incubation in SCL (387-388). MVs prepared by mild collagenase digestion accumulate large amounts of Ca²⁺ and Pi from SCL - without the addition of organic phosphate substrates (76). These vesicles were incubated in SCL and harvested at different stages of ion accumulation, lyophilized, incorporated into KBr pellets, and examined by FT-IR. The baseline spectrum of the nascent MV was computer-subtracted from that of MV at progressive stages of mineralization, revealing the formation of mineral by the vesicles. The first detectable crystalline phase was found to be an HPO₄^2--containing octacalcium phosphate (OCP)-like mineral (Figure 10); this finding was later confirmed using FT-Raman spectroscopy (388). In these intact MVs, no obvious spectra of ACP were evident, presumably again because the patterns of lipids and proteins obscured it. However, when the nucleational core of MVs was isolated and characterized (See Section 16), FT-IR analyses provided clear evidence for the presence of ACP. However, the mechanism of mineral formation by MVs is not a straight-forward process; it involves a number of staged events that progressively lead to the formation of biological apatite.

13.7. Solubility properties of nascent MV mineral

A variety of studies have explored the solubility properties of the initial MV mineral. First, analyses of CRMVs using biphasic solvent partition of electrolytes revealed that the great majority of the Ca²⁺ (>90%) was in an insoluble form, including about 8-10% of the Ca²⁺ being present as acidic phospholipid-Ca²⁺-Pi complexes (64). Less than 5% of the MV Ca²⁺ appeared to be free in solution. In contrast, much of the MV Pi and Mg²⁺ (about 50%, each) and Na⁺ and K⁺ (70-80%, each) were free in solution. These findings suggested that at least half of the Pi and Ca²⁺ present in nascent MVs are not initially combined together; about 40% of the total Ca²⁺ appeared to be bound to proteins and lipids. Subsequent studies by McLean et al. (67) supported these observations. Sucrose gradient analysis of MV-enriched microsomes revealed that in the low-density fractions, much of the Ca²⁺ and Pi were not combined; Ca²⁺ appeared to be largely bound to proteins, whereas most Pi appeared to be unbound in the vesicle fluid. Second, in the McLean study, brief exposure (10 min) of the MVs to isosmotic citrate (pH 6.0) buffer removed 85-90% of the Pi and 70-80% of the Ca²⁺ (67). This destroyed their ability to subsequently accumulate mineral ions. Such brief, mild treatments should not dissolve crystalline HAP protected within the MV lumen, but would readily dissolve ACP, which is very acid labile. Subsequent studies on the properties of nascent MV mineral have led to the discovery of a nucleational core that is directly responsible for induction of crystalline mineral. How it was discovered and how it functions will be described later. However next, attention will focus on how MVs are formed by growth plate chondrocytes.

14. FORMATION OF MATRIX VESICLES

Formation of MVs is a complex physiological process that involves a series of progressive changes in growth plate chondrocytes that ultimately lead to a modified form of apoptosis. The cells shed calcifiable MVs into the extracellular matrix - capable of inducing mineral formation. MV formation stems from the fact that in the proliferative zone of the growth plate, chondrocytes become embedded in a highly impervious matrix of their own making. This proteoglycan-rich gel impedes diffusion of O₂, albumin-bound fatty acids (389), and other nutrients and electrolytes to the proliferating cells. Thus, during their rapid growth, the epiphyseal chondrocytes become separated from their vascular supply and become hypoxic (390-391). They express high levels of glycolytic enzymes (392-394) and become dependent on anaerobic glycolysis for energy (216). The cells adapt to their cloistered, hypoxic environment; they maintain a normal energy charge (216). However, upon penetration of blood vessels to near the maturing chondrocytes, delivery of nutrients, Ca²⁺, and Pi - and especially of O₂ - leads to a sudden increase in oxidative metabolism. Delivery of O₂ to these hypoxic cells causes oxidative stress; their mitochondria begin to store Ca²⁺; becoming loaded with Ca²⁺, they can no longer efficiently produce ATP. The cells undergo a physiological "ischemia-reperfusion" crisis. The cells become depleted of ATP; they swell and become hypertrophic. They generate reactive oxygen species that along with elevated Pi from ATP hydrolysis trigger the opening of the mitochondrial permeation transition pore (395). This leads to rapid unloading of the stored Ca²⁺ and Pi, and this in turn leads to a type of cellular apoptosis (396-397). During this event Ca²⁺-loaded vesicles are released from the cells into the matrix, forming MVs. This introduction is a snap-shot from a physiological perspective. Presented next, is a morphological view of MV formation. Following that, are various biochemical studies that describe how the cells form the MVs and enable them to induce rapid mineral formation - once they are released into the extracellular matrix.

14.1. Electron microscopic studies

The first studies that directly addressed the question of how MVs are formed were those of Bonucci in 1970 (4). Using serial TEM sections of growth plate cartilage, data were obtained that indicated MVs form from cellular projections, either by fragmentation, or by budding from the tips. Studies by Thyberg and Friberg (37, 38) a short time later presented evidence that there may be different types of MV. These early TEM studies indicated that MVs largely derive by exfoliation of vesicles from the outer membrane of growth plate chondrocytes. As noted previously, freeze-fracture analyses of growth plate cartilage contributed definitive evidence that the distribution of membrane-associated particles on cellular processes was identical to that on MVs (Figure 3) (47, 48, 50, 55). This revealed that MVs arise from the cell by mechanisms that do not alter the inside/outside orientation of the membrane. Subsequent TEM studies of MV formation by Akisaka et al. (46) presented evidence that some MVs form by extrusion of preformed osmiophilic cytoplasmic material. Images presented by Arsenault (40, 42, 43) using slam-freezing techniques revealed the presence of high levels of Ca²⁺ and Pi in MVs prior to crystal formation (Figure 2). Ca²⁺ and Pi were not always co-localized, suggesting that they were not present only as calcium phosphate mineral per se - a finding in accord with solubility studies noted previously (64, 67). Studies in which tissue sections were collected by flotation on nonaqueous media and left unstained also suggested that Ca²⁺ and Pi in MV were not in a crystalline form (46).

14.2. Biochemical studies

Much of the evidence providing data on the chemical processes by which MV formation occurs has come from metabolic studies. These have been both in vivo and in vitro experiments demonstrating that incorporation of precursors into MV lipids, electrolytes and proteins is a very active, rapid process.

14.2.1. Phospholipid metabolic studies

The phospholipid composition of MV shows many similarities to that of the plasma membrane, with some notable exceptions (73). While supporting morphological studies that MVs derive from the plasma membrane, metabolic studies have provided more explicit information. In vivo studies on the incorporation of ³²Pi into phospholipids of chondrocytes and MV have revealed that MV formation is a dynamic process in which labeling of MV PC and PE - which tend to be depleted - was nearly as rapid as that of the cellular lipids. In contrast, labeling of PS and SM - which are enriched in MV - was significantly slower (398). This paradox suggests that MV formation involves processes in which the plasma membrane becomes modified by action of phospholipases that selectively degrade PC and PE, while retaining PS and SM. The very limited ability of isolated MVs to incorporate radiolabeled fatty acid lipid precursors demonstrated that the rapid labeling of MVs seen in vivo resulted from prelabeling of the MV membrane lipids while they were part of the cell (268). However, isolated MVs can incorporate ¹⁴C-serine into both PS and its lyso-form, LPS, via Ca²⁺-dependent non-ATP-dependent base-exchange (PS synthase). This reveals that MVs retain the ability to synthesize PS, a key Ca²⁺-binding lipid, enabling its further enrichment in the MV membrane.

14.2.2. Cell culture studies

Further insight into the mechanism of MV formation has come from studies on primary cultures of growth plate chondrocytes. However, it is important to distinguish between vesicles released into the culture media (media vesicles, MeVs) from those that become entrapped in the extracellular matrix (matrix vesicles, MVs). MeVs do not appear capable of inducing mineral formation, whereas MVs clearly do (74, 399). It is noteworthy that both MeVs (68, 91), and MVs are rich in TNAP. The kinetics of MeV formation can be modulated by a variety of drugs that affect the cytoskeletal system and the acidic phospholipid-dependent Ca²⁺-binding proteins (annexins). Cytochalasin D, a drug that blocks actin microfilament formation, markedly enhanced the rate of MeV formation, whereas phalloidin, a drug that prevents actin filament disassembly, clearly inhibited it (69). Coupled with the observation that isolated chondrocyte microvilli have a phospholipid composition almost identical to that of collagenase-released MVs (73), this provides biochemical confirmation of freeze-fracture studies that show MVs being formed from cellular microvilli (47, 50). While MVs contain some actin (78), chondrocyte microvilli contain abundant filamentous actin (69). This reveals that MVs form by mechanisms that involve depolymerization of actin filaments. On the other hand, studies using tricyclic amine drugs that selectively bind to the annexins also stimulate release of MeVs from cultured chondrocytes (70). Taken together, these findings indicate that processes that uncouple the annexins from the cytoskeletal network must be involved in MV formation. Further, the discovery that ascorbate, a vitamin required for normal collagen synthesis and fibril formation, strongly stimulates MV formation and matrix mineralization by cultured chondrocytes (75), suggests that MV formation is linked with synthesis of extracellular matrix collagens (400-401). The demonstration that both TNAP and at least two of the MV annexins (AnxA5 and AnxA6) bind tightly to both type II and X collagen (95, 102), suggests that MV formation involves the attachment of proteins present in microvilli-like extensions of the chondrocyte plasma membrane to matrix collagens (95, 402-403). Further, the finding of various integrins by proteomic analyses of MVs (78) indicates that they also contribute MV formation. Thus, binding of extracellular proteins to receptors on the outer membrane of the chondrocytes facilitates MV formation.

14.3. Role of cellular Ca²⁺ and Pi metabolism in MV formation

There is clear evidence that cellular metabolism of both Pi and Ca²⁺ is involved in MV formation and subsequent mineralization (404). Here is a brief preview of what will be discussed on this topic. When mitochondrial Ca²⁺ is released via the mitochondrial transition pore into the Pi rich cytosol, the released Ca²⁺ interacts simultaneously with both Pi and PS forming PS-Ca²⁺-Pi complexes, as well as with PS and AnxA5 to form PS-Ca²⁺-AnxA5 complexes. Both types of complexes are attached to the PS-rich inner leaflet of the plasma membrane of the chondrocytes becoming incorporated into and facilitating MV formation. Much of the mitochondrial Ca²⁺ released into the cytosol appears to become tied up in these complexes. Confocal imaging of Ca²⁺ ion activity in the hypertrophic cells show that central cytosolic Ca²⁺ levels drop to <10^-7 M, but peripheral Ca²⁺ levels rise markedly to ~7.5 x 10^-6 M (307). Because of this Ca²⁺ depletion, cytoskeletal structural proteins like f-actin depolymerize and allow blebbing of MVs from the plasma membrane (69, 73). Upon binding of TNAP to types II and X collagens (95) and proteoglycan core and link protein (95), the tips of the membrane-enveloped cellular microvilli detach and form MVs. These vesicles envelop various cytoplasmic proteins, the PS-Ca²⁺-Pi complexes, as well as Pi, Mg²⁺, and other electrolytes, and attach tightly to the extracellular matrix proteins. Subsequently, phospholipase activity in the MV membrane (207) apparently enable externalization of PS-Ca²⁺-annexin complexes that is facilitated by binding to extracellular proteins like types II and X collagen (405). Based on recent proteomic data (78) it is evident that type VI collagen is also involved in MV attachment.

14.3.1. Histology of cellular Ca²⁺ metabolism during MV formation

The early discovery by Martin and Matthews (271-272) that growth plate chondrocytes accumulate significant amounts of Ca²⁺ in the form of mitochondrial granules prior to the onset of extracellular mineralization in the growth plate provided the first indication of how cells load Ca²⁺ into MVs. The fact that all mitochondria possess the ability to accumulate massive amounts of Ca²⁺ had been known for some time (406-407). Further, studies on Ca²⁺ accumulation by isolated growth plate mitochondria by Shapiro and Lee (408-409) showed that even under hypoxic conditions these mitochondria could accumulate Ca²⁺ - significantly more than those from noncalcifying cells. And as was noted earlier, ultrastructural studies by Brighton and Hunt (382-384) using pyroantimonate staining (Figure 8A-C), as well as the electron probe studies by Hargest et al. (380), revealed that the accumulated mitochondrial Ca²⁺ becomes subsequently transferred to MV during growth plate development. These ultrastructural findings were confirmed by Whitson et al. (410) using a completely different Ca²⁺ stain which showed that Ca²⁺ accumulated on the inner aspect of the plasma membrane; it became concentrated at the termini of cellular processes from which MV appeared to form. Nearly a dozen different investigations using a variety of chemical and histological methodologies document that mitochondria play a key role in growth plate calcification. These were critically reviewed by Matthews nearly a quarter of a century ago at the Fourth International Conference on Matrix Vesicles (411). And the evidence continues to mount.

14.3.2. Confocal imaging of Ca²⁺ in living growth plate chondrocytes

Imaging studies published first in 1993 (404), and reconfirmed in 1997 (306), document that growth plate chondrocytes exhibit highly active metabolism of Ca²⁺. Laser confocal imaging of sections of living avian growth plate cartilage using a sensitive Ca²⁺ probe (Indo 1) at low magnification revealed marked differences in Ca²⁺ ion activities within individual chondrocytes in different regions of the growth plate (Figure 11A-B). More detailed study revealed that in the proliferative zone, while there were generally lower overall levels of Ca²⁺ (Figure 12A), there were also marked differences in Ca²⁺ levels within sharply defined regions, presumably corresponding to endoplasmic reticulum, mitochondria and nuclei. Although marked variations were seen between different regions within the cells and between individual cells, few cells had overall high levels and the nuclei obviously had very low Ca²⁺ levels. In the maturation zone there was a marked increase in the differences in Ca²⁺ levels of individual chondrocytes (Figure 12B); some had very high overall levels; others had quite low levels; some had intermediate levels similar to those in the proliferative zone. Ca²⁺ levels near the plasma membrane were clearly the highest. In individual cells, overall Ca²⁺ levels rose and fell within a matter of minutes; cells with high levels of Ca²⁺ tended to loose it, while those with low levels tended to gain it (Figure 13A). In the greatly enlarged cells of the hypertrophic zone, the previous Ca²⁺ oscillations appeared to have transferred much of the Ca²⁺ to the cell perimeter (plasma membrane), where it was then shed as highly-Ca²⁺-enriched microstructures (MVs) into the extracellular matrix (Figure 12C). The volume of the cells in this region of the growth plate was almost ten-fold larger than those of the proliferative zone (306). In the calcifying zone, as the cells shed Ca²⁺-loaded MVs, the central cytosol became markedly lower in Ca²⁺ than in the preceding zone (Figure 12D). The overall impression was that after a crescendo of rising and falling Ca²⁺ levels, the cells exfoliated the accumulated Ca²⁺ via a burst of Ca²⁺-loaded vesicles that were shed from their perimeter (Figure 13B).

14.3.3. Biochemistry of mitochondrial Ca²⁺ metabolism

TEM studies of the growth plate showed clear evidence of Ca²⁺ loading in mitochondria, with subsequent transfer to MVs (383, 384). On the other hand early biochemical studies by Lee and Shapiro indicate that functional mitochondria can be isolated from epiphyseal chondrocytes that are able to carry out oxidative phosphorylation (412-413). These isolated chondrocyte mitochondria accumulated maximal Ca²⁺ when ATP and a respiratory substrate (e.g. citrate or alpha-ketoglutarate) were available (409). Uncouplers of oxidative phosphorylation, or inhibitors of respiration, impeded Ca²⁺ uptake - as is seen in mitochondria isolated from non-calcifying cells. Further, there was heterogeneity in the extent of Ca²⁺ loading (i.e. density) of the growth plate mitochondria. Ca²⁺ accumulation increased from the proliferative zone to the early hypertrophic zone of the growth plate (414). Then in the hypertrophic zone where MVs most rapidly form, mitochondrial levels abruptly fell - a finding later confirmed by Hargest (380). Accumulation of excessive Ca²⁺ is known to inhibit mitochondrial respiration - as is true in all mitochondria (415). However, growth plate mitochondria appear to be more tolerant of Ca²⁺ accumulation than most (409). The general mechanisms by which mitochondria transport Ca²⁺ have been thoroughly reviewed by Gunter and Pfeiffer (416-417). Mitochondria primarily utilize a uniporter system for transporting Ca²⁺ into the inner matrix, driven by the electro-chemical gradient produced by mitochondrial respiration. It can be best described as a "very fast gated pore" (416). This property was studied in growth plate mitochondria by Lee and Shapiro (418) who found that Ca²⁺ uptake was similar to that from noncalcifying cells. They observed that inhibitors of mitochondrial respiration caused rapid loss of greater than 80% of the Ca²⁺. However, in mitochondria isolated from the "most calcified" zone there was increased "nonspecific Ca²⁺ binding"; and, a high percentage of the accumulated Ca²⁺ was "non-extrudable", apparently in the form of mitochondrial granules - in agreement with the ultrastructural observations noted previously. There are two primary mechanisms by which mitochondria are able to release stored Ca²⁺; these are Na⁺-dependent and Na⁺-independent pathways - evident in all cells studied (416). For most cells, including growth plate chondrocytes up to the zone of maturation, mitochondria primarily regulate cytosolic Ca²⁺ ion activities by these two Ca²⁺ carriers. However, another more substantial pathway of mitochondrial Ca²⁺ release is of special importance to growth plate mineralization. This is the so called "mitochondrial permeability transition" (MPT), a mechanism involving a marked change in the permeability of the inner mitochondrial membrane. It involves a dramatic loss of membrane potential, as well as rapid leakage of K⁺, Mg²⁺, Ca²⁺, Pi and other small ions and molecules from the mitochondrial matrix (419-420). This massive release of mitochondrial electrolytes is closely associated with the onset of apoptosis in many cell types (421-424); however, it is of key importance in the formation of MVs by growth plate chondrocytes. The MPT is brought about by first preloading of mitochondria with Ca²⁺; and then by inducing Ca²⁺ release via an agent that activates the MPT pore. Among the various cofactors known to function as Ca²⁺-releasing agents in the MPT, the most potent and best studied is inorganic phosphate (Pi) (416).

14.4. Growth plate cellular Pi metabolism

Thus, cellular Pi metabolism is deeply entwined with Ca²⁺ metabolism. Cells require Pi for a plethora of basic cellular functions ranging from synthesis of DNA and RNA to metabolic intermediates directly involved in energy metabolism. Over and beyond this in the growth plate, Pi is needed for formation of the calcium phosphates that initiate solid phase formation during bone growth. Growth plate cartilage has been shown to possess unusually high levels of both cytosolic (20-25 mM) and extracellular (2.2-2.5 mM) Pi (63). In fact, studies by Mansfield et al. (425-426) have documented that exposure of terminally differentiated, cultured tibial growth plate chondrocytes to even modestly elevated levels (3 mM) of extracellular Pi induces 'apoptotic' cell death within 48 h. Apoptosis of these cells is mediated via the MPT mechanism (314). It is noteworthy that the release of mitochondrial Ca²⁺ via the PMT also activates uptake of Ca²⁺ by the endoplasmic reticulum, which in turn releases Ca²⁺ into the cytosol (427). These findings explain how Ca²⁺, first stored in mitochondria of growth plate chondrocytes, becomes released. It causes elevation in cytosolic Ca²⁺ levels - ultimately at the cell periphery, as confocal imaging of Ca²⁺ metabolism has indicated (306). Of direct interest to growth plate mineralization, Type VI collagen receptors on the cell surface (109, 110) act as transducers to mitochondria, affecting opening of the MPT pore for Ca²⁺ (111). This appears to trigger mitochondrial Ca²⁺ unloading (112).

14.4.1. Chondrocyte mitochondrial PiC Pi-transporter

To better understand the processes by which mitochondria take up Pi, the PiC transporter will now be discussed. Mitochondrial acquisition of Pi is essential for oxidative phosphorylation - an energy-capture process vital to all aerobic eukaryotic cells. Mitochondrial uptake of Pi is largely accomplished by PiC, an H⁺/Pi symporter (or Pi/OH^- antiporter) (428-429). PiC is an integral membrane protein, a homodimer that resides in the inner mitochondrial membrane that translocates or exchanges Pi from the cytosol into the mitochondrial matrix (430). PiC is Na⁺-independent and is responsible for the rapid import of Pi used for ATP synthesis. The mitochondrial PiC transporter has been cloned and sequenced from chicken growth plate cartilage (336). The molecular mass of the mature protein is 34,985. Hydropathy plots indicate that PiC is a highly compact protein with 6 transmembrane domains; both the N-terminal and C-terminal ends reside in the mitochondrial matrix lumen. In situ hybridization analysis reveals that mitochondria PiC is most strongly expressed in the maturation/prehypertrophic zone, although it is expressed at low levels throughout the growth plate (Figure 7A) (336). However, note that at the terminus where the cells become apoptotic and vascular penetration occurs, many cells are devoid of PiC message. Of pertinence here, analysis of PiC gene expression in primary mineralizing cultures of chicken growth plate chondrocytes reveal that the gene is only weakly expressed early in culture; it becomes progressively more highly expressed as the cultures mature and calcify. These features match the Na⁺-independent Pi porter described by Wu et al. Thus it is apparent that this high-capacity mitochondrial PiC porter is the one whose expression is enhanced by elevated Pi levels in the culture medium (303).

14.4.2. Chondrocyte plasma membrane Pi-transporters

At this point, we need to return to Pi uptake and transport in the plasma membrane of growth plate chondrocytes. Recall that two Na⁺-dependent (type III Glvr-1 and type II) plasma membrane Pi transporters (334-335, 426, 431) are present in growth plate chondrocytes and MVs. Subsequent studies have confirmed the presence of the PiT type III (Glvr-1) Pi porter in growth plate chondrocytes, but in situ hybridization studies reveal that it is expressed strongly only in the resting and proliferative zones; there is almost no expression in the maturation/hypertrophic zone (Figure 7B) (336). Again, note that it is also expressed in the early growth plate chondrocyte cultures, but declines as the cells mature and begin to induce calcification (303). As indicated above, these studies also uncovered evidence for a Na⁺-independent Pi transporter whose expression was upregulated as the cells approached hypertrophy. In contrast to the Na⁺-dependent Pi transporters that are upregulated by falling Pi levels, expression of this porter is induced by culturing the chondrocytes in media containing slightly elevated levels of Pi (2-3 mM) - levels typical of the extracellular fluid of growth plate cartilage. This Na⁺-independent Pi transporter had a pH optimum of 7.6-8.0, appropriate for the pH of growth plate extracellular fluid (63, 338).

14.4.3. Growth Plate Chondrocyte Mitochondrial Permeability Transition (MPT) Pore

The key Pi-sensitive MPT pore appears to be directly involved in the apoptotic demise of hypertrophic chondrocytes (314). Under conditions of mitochondrial Ca²⁺ overload, especially when accompanied by oxidative stress, elevated Pi levels and ATP deletion, this non-specific MPT pore opens in the inner mitochondrial membrane (432-436). This allows rapid release of mitochondrial electrolytes, including both Pi and Ca²⁺. Immediately, ATP synthesis effectively stops; and as noted before, there is clear evidence that hypertrophic growth plate chondrocytes rapidly become energy depleted. Several independent studies reveal that the levels of ATP in hypertrophic cells are greatly reduced compared to resting or proliferating cells (302, 437-439). Hydrolysis of ATP, which is typically present in 5-10 mM levels in healthy, active cells, would release large amounts of Pi. And direct analyses of the Pi levels of isolated growth plate chondrocytes reveal the presence of high cytosolic levels in the range of 20-25 mM (63). This has been confirmed by ³¹P-NMR analysis of growth plate chondrocytes (404, 438). However later in this process, breakdown of cellular phospholipids through the action of phospholipases (223) and the action of PHOSPHO-1 (202-203), may also contribute to the elevated Pi levels in the cytosol of growth plate chondrocytes. Taken together, these findings reveal an intricate scenario in which Ca²⁺-loaded mitochondria, subjected to oxidative stress, and exposed to elevated levels of cytosolic Pi, release their electrolytes via the MPT into the cytosol. In this process, they induce a specialized form of apoptosis in which the released Ca²⁺ and Pi combine with PS and AnxA5 to form complexes that contribute to MV formation, and ultimately to mineral formation. Upon Ca²⁺ release, microvilli-like cellular membrane extensions are formed, a process that depends on the activity of phospholipases and scramblases that modify the membrane lipid composition, as well as the inside/outside orientation of the lipids in the membrane. During this complex event, membrane-associated proteins like TNAP become externalized allowing them to interact with matrix proteins. Upon binding to the extracellular collagens (Types II, VI, X) these cellular extensions bleb off to form of vesicles - MVs - primed with sufficient levels of both Ca²⁺ and Pi in the form of a nucleational core of ACP-PS-AnxA5 complexes that serve as the nidus for triggering extracellular mineral formation.

14.5. Growth plate cellular H⁺ (pH) metabolism

The pH of both the extracellular and intracellular fluids also has a profound effect on the process of MV formation. Because growth plate chondrocytes are embedded in a highly viscous tissue comprised of large amounts of high molecular weight proteoglycans, their pH is largely governed by their anaerobic metabolic activity. However, in weight-bearing joints, the on-off pressure exerted by body movements enables exchange of fluid despite the limited vasculature of the tissue. This exchange of fluid ameliorates the tendency toward acidification. There are also zonal differences in cellular metabolism ranging from the quiescent "resting" zone through the zone of rapid cellular proliferation to the zone of hypertrophy where the cells undergo apoptotic formation of MVs. These changes in metabolic activity influence the cellular pH, which has been documented by confocal imaging using the fluorescent probe, BCECF.

14.5.1. Confocal imaging of H⁺ (pH) in living growth plate chondrocytes

The pH of growth plate chondrocytes in situ has been measured using confocal imaging with BCECF as a fluorescent probe (306). Low magnification confocal imaging shows that in different regions of the growth plate there are sharp differences in pH between individual cells (Figure 14A-B). At higher magnification, it became evident that this occurred in various intracellular compartments where pH levels ranged from pH 6.2-8.0. In the proliferating zone, average cytosolic pH (pH 6.95�0.05) was generally lower than in the other zones, although there were many focal regions of lower and higher pH (Figure 15A). For example, the pH of the nuclei of the cells tended to be low (pH 6.5-6.7), a finding not as apparent in the maturing and hypertrophic zones. In the zone of maturation, the average pH (7.39�0.04) was significantly higher and was characterized by the presence of many focal regions of distinct alkalinity (pH 7.8-8.0) (Figure 15B). In the hypertrophic zone, the overall pH (7.19�0.04) was slightly lower than in the maturation zone. While there were many regions of alkalinity, notice that in the periphery of the cells, the pH was significantly more acidic (Figure 15C). In the zone where extracellular calcification was already occurring, the intracellular pH was lowered even further (6.76�0.03); there were few focal areas of alkalinity (Figure 15D). These findings are interpreted as follows:

In the proliferating zone, the key cellular activity is replication, whose rate is governed by both genes and growth factors (440). Cell replication requires a constant source of raw materials for DNA and RNA synthesis, as well as metabolic energy. Because of their cloistered avascular environment, there is low O₂ tension in these cells (216, 390, 441). Most of their energy is derived from glycolytic production of ATP (216). The generally lower cellular pH in the proliferative zone appears to result from glycolytic production of lactic acid. In the zone of maturation, different forces are operative. Cell division has largely ceased; and the cells, because of their closer proximity to the invading blood supply, are affected by the increasingly available supply of O₂, nutrients, Ca²⁺ and Pi (395). There is an awakening of mitochondrial oxidative metabolism, as well as the active acquisition and storage of Ca²⁺ within the mitochondria (382, 409). With the increased metabolic activity, there is synthesis of a variety of new proteins that become involved in the mineral-forming process. At this stage, the cells begin to undergo a measurable increase in volume (306). Apparently due to the sharply increased oxidative phosphorylation, these maturing cells show marked alkalinization in areas that appear to be organelles like the mitochondria and the endoplasmic reticulum (306). At this stage there is maximal Ca²⁺ storage by the mitochondria (409), which adversely affects their ability to produce ATP (408). When the demand for ATP exceeds the ability of the impaired Ca²⁺-loaded mitochondria to supply, levels of Pi build up in the cytosol (63, 438, 442) setting up eventual apoptosis. In the cell's struggle to control cytosolic Ca²⁺, there is a cyclical rise and fall of Ca²⁺ levels (306) (Figure 13A-B).

Approaching the zone of hypertrophy a crisis point is reached that triggers the opening of the mitochondrial transition pore described earlier (314). ATP synthesis ceases. A flood of stored Ca²⁺, electrolytes, nucleotides, and small proteins leaves the damaged mitochondria. The rising levels of cytosolic Ca²⁺ activate intracellular proteases and phospholipases that in turn begin to release large amounts of free amino acids - as well as damage the intracellular membranes. The level of ATP plummets as metabolic demand exceeds the available supplies (395, 442). Intracellular levels of Pi increase dramatically (63, 438). With the rise in amount of low molecular weight solutes, there is influx of water that causes marked osmotic swelling; the cells become hypertrophic. At this time, there is movement of both Ca²⁺ and H⁺ to the cell periphery where pH levels fall (306). Facilitated by the presence of the annexins, Ca²⁺ and Pi bind to PS on the inner leaflet of the plasma membrane to form PS-Ca²⁺-Pi complexes (PS-CPLXs) (60, 61). Dissociation of cytoskeletal actin fibrils facilitates the blebbing of MVs from cell processes (69, 73). The newly forming vesicles carry with them the annexins (60, 61), the PS-CPLXs (64), the encompassed electrolytes (Pi, Mg²⁺, Ca²⁺, K⁺, Na⁺) (63), as well as various cytosolic enzymes (e.g. the family of glycolytic enzymes) and proteins (78). Because the pH near the plasma membrane of these cells is slightly acidic (pH 6.5-6.9), even though the levels of Ca²⁺ and Pi are sufficiently high, ACP cannot form (77). However at this pH, Mg²⁺-containing PS-CPLXs can form readily (266) and become a key component of the nucleational core of MVs (288, 443). Upon release of the MVs into the extracellular space where the pH is more alkaline (pH 7.5-7.6) (63, 339), the high levels of enclosed Ca²⁺ and Pi readily combine to form ACP. But because of abundant Mg²⁺, ACP is stabilized and does not immediately dismutate to OCP or HAP (444). In the zone of calcification when the hypertrophic chondrocytes are "spent" and the extracellular matrix around them begins to calcify around the MVs, the levels of intracellular Ca²⁺ markedly decline (Figure 12D, Figure 13B) and the aveage pH of the cytosol falls to 6.76�0.03 (Figure 15D). There are few focal areas of high pH such as was seen in the maturing and hypertrophic cells. These cells generally are eliminated by macrophages during the invasion of the capillaries from the marrow cavity. In avian species, however, there is evidence that some growth plate cells survive, become reprogrammed (445-446), and express osteoblast-like character (447), producing type I instead of type II collagen, etc. Attention must now be directed toward the mechanism by which newly formed MVs induce de novo crystalline mineral - once they are released into the extracellular matrix.

15. THE KINETICS OF MINERAL FORMATION BY MATRIX VESICLES

15.1. Differing views of MV mineral formation

Many types of in vitro studies have been conducted to explore the ability of isolated MV to accumulate Ca²⁺ and Pi and form crystalline mineral deposits. It should come as no surprise that numerous differences have been reported with regard to: 1) the requirement for ATP (82, 187, 448) or other organic phosphates (66, 449), 2) the regulatory effects of Zn²⁺ (317, 385, 450), 3) the kinetics of mineral ion accumulation (332, 451-453), and 4) the nature of the first mineral phase produced (57, 65, 77, 387-388, 454). Most of these differences can be traced to the differing methods of MV preparation, the species from which they were obtained, the extent of damage imposed on the structures during isolation from the tissue, and the method used for assessing the timing, rate, and amount of mineral formation. Rather than dwell on these differences, attention will be focused on MVs that have been released from normal avian cartilage using carefully controlled levels and timing of protease exposure. These MVs can be obtained in a good yield, in a high state of purity, and induce mineral formation within a few hours when incubated in a synthetic cartilage lymph (SCL) matched to that of the extracellular fluid in growth plate, with no added organic P substrates required (60-62, 76). Their activity will be compared with MV-enriched microsomes that have not been exposed to exogenous proteases.

15.2. Radio-isotope analyses of the kinetics of Ca²⁺ and Pi accumulation

MVs isolated by modified collagenase digestion of growth plate cartilage from broiler-strain chickens, or MVEM isolated by the previously-described nonenzymatic tissue homogenization methods, acquire mineral ions from synthetic lymph and rapidly induce the formation of crystalline solid phase mineral in a sequence of predictable stages (Figure 16A) (62, 65-67, 317, 332) Studies measuring simultaneous uptake of ⁴⁵Ca²⁺ and ³²Pi during MV mineralization indicated that an OCP-like, low (1.33) Ca/P-ratio mineral forms during the early phase of rapid mineral ion accumulation (Figure 16B) (65, 455). Later during the slow uptake period, the Ca/P ratio rose to ~1.60, indicating progressive transformation into HAP-like mineral (Figure 16B). These findings are in close accord to studies using FT-IR, which reached the same conclusion (Figure 10) (387). Both approaches reveal that mineral formation by MVs is not a simple process; it involves several solid phases prior to the formation of the thermodynamically stable HAP, the major mineral of mature bone. Based on extensive studies, the kinetics of MV calcification generally involve the following sequence of events: 1) exchange of ions at the vesicle membrane surface, 2) selective acquisition of Ca²⁺ in preference to Pi during a lag period in which very limited ion uptake occurs, 3) a transient induction period in which the Ca²⁺/Pi uptake ratio rapidly declines, indicative of selective Pi uptake, followed by 4) a rapid accumulation of both Ca²⁺ and Pi at a Ca²⁺/Pi stoichiometry of close to 1.33, and 5) finally, an extended period of slower acquisition of both mineral ions at a gradually rising Ca²⁺/Pi ratio approaching that of Ca²⁺-deficient apatite (~1.60) by the end of 24 h. Interpretation of these five stages of MV mineral formation now follows.

15.2.1. Stage 1 - initial exchange period

When isolated MVEM are exposed to ⁴⁵Ca²⁺- or ³²Pi-labeled SCL, there was initially a rapid, but very limited uptake of both ions that corresponds to an exchange with ions bound to proteins and/or acidic phospholipids at the vesicle surface (Figure 16A) (65, 449, 456). During this brief period, ⁴⁵Ca²⁺/³²Pi uptake ratios rapidly increased from very low values (<0.7) to values typically above 1.5, indicating a preferential binding of Ca²⁺ (Figure 16B) (57, 455).

15.2.2. Stage 2 - the lag period

Stage 1 is followed by a lag period of variable duration during which little further accumulation of Ca²⁺ and Pi occurs. The lag period can be extended in several ways that do not preclude eventual induction of mineral formation. For example, by adding low (i.e. 5 micromolar) levels of Zn²⁺ to the SCL in which the MVs are incubated, Ca²⁺ accumulation is essentially blocked (Figure 17A). The inhibitory effect of Zn²⁺ can be rapidly eliminated by treatment of MVs with o-phenanthroline, a group 2B metal ion chelator that binds tightly to Zn²⁺ (Figure 17B) (76, 317, 385). Thus, one explanation of the lag period is that it represents the time needed for displacement of Zn²⁺ by Ca²⁺ at key sites that enables rapid ion acquisition. Zn²⁺ binds tightly to AnxA5 in the presence of PS (Figure 4C) (60). While Zn²⁺ has been shown to block the apparent Ca²⁺-channel activity of AnxA5 in planar PS bilayers (125), the high levels of Ca²⁺ used and the excessive amount of Zn²⁺ required, cast doubt that this is the physiological action of AnxA5. The lag period also can be greatly extended by exposing active MV to proteases such as trypsin or chymotrypsin (76). In this case, when ion uptake finally begins, it does so at a much lower rate than in the untreated control. However, as noted previously, if these protease-inactivated MVs are treated with detergents that permeabilize the vesicle membrane, mineral formation then rapidly ensues (288, 457). This indicates that the lag period is a rate-limiting step that represents the inability of mineral ions to pass through the vesicle membrane and reach the nucleational core. In modeling MV mineralization using liposomes encapsulating Pi-buffered intracellular electrolytes (See Section 18), Blandford et al. explored several methods for enabling Ca²⁺ to enter the liposomes (318). The most rapid means was found to be by addition of phospholipase A₂ to the incubation mixture; it triggered almost immediate rapid uptake of ⁴⁵Ca²⁺. Since latent phospholipases are evident in MVs, activation of these enzymes would be natural way to breach the membrane block. In fact, in studies of changes that occur in MVs during mineral formation, Wu et al. found that phospholipids levels began to decline upon introduction of the MVs into the SCL, accompanied by increased levels of FFA and 1,2-DAG (207). This is evidence of phospholipase activity, which would increase membrane leakiness and allow entrance of Ca²⁺ and Pi. Thus, the event that signals the end of the lag period appears to be partial degradation of the MV membrane brought about by activation of endogenous phospholipases. This allows access of Ca²⁺ and Pi to the nucleation site, which then can initiate rapid onset of mineralization.

15.2.3. Stage 3 - the induction phase

However, approaching the end of the lag period there is a transient sharp decline in the Ca²⁺/Pi uptake ratio from ~1.5 to 1.25 (Figure 16B) (57, 455). This is indicative of conversion of ACP to crystalline OCP. In fact, for the induction of rapid mineral formation, two conditions need to be met. First, there must be a means by which the extravesicular Ca²⁺ and Pi can gain access to the nucleation core. Second, there needs to be present within the MV lumen crystalline nuclei that can engender crystal growth. FT-IR analyses reveal subtle conformational changes in MV structure during this period - apparently caused by induction of the first crystalline nuclei (387). Two factors can prevent the conversion of the ACP-containing nucleational core to OCP crystalline nuclei: 1) elevating the pH of SCL to 8.0 (77), which delays induction indefinitely; or 2) incorporation of Mg²⁺ (131, 137), which delays onset of rapid mineral formation for at least 10-12 h. In typical MV mineralization induction of rapid mineral formation occurs within 1-3 h after exposure of the vesicles to the extracellular fluid.

15.2.4. Stage 4 -the rapid acquisition phase

After induction has occurred, rapid acquisition of both Ca²⁺ and Pi ensue. This phase corresponds to the proliferation of the first crystalline mineral phase - an OCP-like, HPO₄^2--containing calcium phosphate mineral (387). The Ca²⁺/Pi stoichiometry of ion acquisition during this rapid mineral formation period is close to 1.33, characteristic of OCP (450, 457). Both FTIR and FT-Raman spectroscopic analyses of this early mineral phase are consistent with the formation of a poorly-crystalline OCP-like mineral (Figure 10) (387-388). This stage generally lasts for 2-3 h (Figure 16A) and coincides with rapid OCP formation. The acquired ⁴⁵Ca²⁺ by the MVs is not exchangeable with ion-exchange resins until the beginning of the plateau phase (67), indicating that it must occur within the vesicle lumen. This is supported by the fact that the ionic environment within the vesicle lumen would favor growth of OCP in preference to HAP. The fluid is rich in Mg²⁺ and Pi (63, 64), and contains low levels of Zn²⁺ (317) - a combination of factors known to suppress HAP and favor OCP-crystal formation (385). It is also evident that the acidic phospholipid-rich inner membrane, in combination with the MV annexins, facilitates growth of OCP-like crystals along the membrane fluid interface. Crystal growth within MVs is typically seen in close association with the vesicle membrane (2). Such crystal growth, in conjunction with Ca²⁺-activated phospholipases, would lead to further destabilization and rupture of the vesicle membrane. Crystal growth is also closely associated with TNAP, whose catalytic activity declines progressively as mineralization ensues (76).

15.2.5. Stage 5 - the plateau phase

At the end of the rapid uptake phase there is an extended period of slow accumulation of both Ca²⁺ and Pi, with gradually rising Ca²⁺/Pi uptake stoichiometry (Figure 16A) (57, 65, 449, 455). The plateau phase appears to coincide with the complete breaching of the vesicle membrane and direct exposure of the nascent crystals to the extravesicular environment. This interpretation is based on the finding that only after this point does the acquired Ca²⁺ become readily exchangeable with ion-exchange resins (67). The nature of the mineral phase formed during the plateau phase is a poorly crystalline HAP; FTIR patterns confirm that it is HAP formed from an OCP precursor, rather than from ACP (Figure 10). The obvious decrease in the rate of mineral ion uptake relates to several factors: 1) the progressive depletion of mineral ions available from the SCL, 2) the apparently slow, gradual dismutation of the OCP to HAP, and 3) the adsorption of MV regulatory proteins to the growing HAP crystals. Typically, during 24 h incubation, at most 30-40% of the original Ca²⁺ and Pi in the SCL become incorporated into the newly forming mineral (57, 65, 385, 449, 456). In contrast, if pure HAP crystals are seeded into protein-free SCL, acquisition of about 70% of the mineral ions occurs during a similar incubation period. This marked difference suggests that various factor(s) associated with MVs limit the extent of crystal growth.

16. THE NUCLEATIONAL CORE: THE DRIVING FORCE IN MV Ca²⁺ AND Pi ACCUMULATION

A key question with regard to MV mineralization is what drives the accumulation of Ca²⁺ and Pi by MV? At the outset, let it be stated that it is highly doubtful that ATP-driven active transport processes are involved. This conclusion is based on several facts. First, as has been already argued, the orientation of the Ca²⁺/Mg²⁺ ATPase pump in the vesicle membrane would be facing outward - favoring loss, not gain, of Ca²⁺. (There is no evidence that the plasma membrane becomes everted during MV formation.) Second, while there is strong evidence for the abundant presence of Na⁺/K⁺-ATPase pump proteins in MVs - based on proteomic studies (78), it is obvious that they are not operative. Third, there is insufficient ATP in the extravesicular fluid to power such an ion pump (302). Supporting this contention is the finding that MVs have high levels of Na⁺ and low levels of K⁺ (63), the opposite of what would occur if the Na⁺/K⁺ pump were operative. Thus, the most logical force for driving Ca²⁺ uptake by MVs is the nucleational core, which provides "sink conditions" - i.e. by reducing the ion activity of Ca²⁺ within the vesicle lumen to levels significantly lower than those in the extravesicular fluid. MVs create sink conditions by forming an insoluble, metastable Ca²⁺-Pi phase within the vesicle lumen. Strange as it may seem, while total Ca²⁺ (65 � 8 mM) levels in newly formed MV are much higher than those in the extravesicular fluid (3.2�0.5 mM) (63), the ionic activity of Ca²⁺ in the MV lumen is much lower. Based on in situ confocal imaging of MVs, the Ca²⁺ activities are 7.4 � 0.3 micromolar (306). These levels are far lower than the 1,000 � 200 micromolar (i.e. 1.0 � 0.2 mM) Ca²⁺ activity present in the extracellular fluid (339). Thus, the vast majority (~99%) of Ca²⁺ in MVs must be tied up in insoluble forms (64); these are embodied in the nucleational core.

16.1. Discovery of the nucleational core

A long series of observations ultimately led to the discovery that it was the presence of a nucleational core that drives MV Ca²⁺ and Pi accumulation. The roots of the discovery of the nucleational core go back to 1977 when it was first documented that MVs contain sizeable amounts of Ca²⁺, Pi, and other mineral ions (63); this mineral was amorphous to x-ray diffraction (Figure 9), and occurred in various solubility states (64). A significant breakthrough occurred in 1986 when the marked acid lability of the mineral in MVs was first documented (66). The mineral was shown to be inside the vesicle lumen, and when lost, the MVs would not mineralize (67). Ca²⁺ uptake by MV was shown to be highly sensitive to pH, being limited to a range very similar to that observed with ACP (77). The next step was a major improvement in MV isolation, yielding purer, highly active, mineralization-competent vesicles (76). Importantly, the ability of these MVs to induce mineral formation could be readily ablated by chymotrypsin. Amazingly, if these damaged vesicles were then treated with detergents, their nucleational competence was restored (288). This series of observations led to the conclusion that a "nucleational core" was responsible for accumulation of Ca²⁺ and Pi by MVs.

The discovery process evolved in this way. First, one of the most effective ways to prevent Ca²⁺ ion uptake by isolated MVs was to briefly (ca. 10 min) expose them to isosmotic, pH 6 citrate buffer (66). After this simple treatment, their ability to accumulate Ca²⁺ (and Pi) was almost totally abrogated. Such treatment was shown to remove 80-85% of the total Ca²⁺ and Pi from the vesicles (67). These findings revealed that the main contributor to the driving force for Ca²⁺ uptake must be an insoluble, but highly acid-labile Ca²⁺-Pi-rich phase. Second, it was found that accumulation of mineral ions by MVs from SCL only occurred within a very narrow pH range - between 7.4-7.8 (77) (Figure 18A). Either below or above this range, the rate of Ca²⁺ uptake was dramatically reduced; for example, at pH 8.0 there was simply no uptake of ⁴⁵Ca²⁺ by MVs. (In contrast, upon seeding of HAP at pH 8.0, there is a slow, steady accumulation of ⁴⁵Ca²⁺ for at least 24 h (77)). On the other hand, it became evident that below pH 7.4, the solubility of the mineral phase in MVs was too high to reduce Ca²⁺ activity in the intravesicular fluid to levels that enabled Ca²⁺ acquisition. Above pH 7.8, it was evident that the stability of this phase was so great that it could not convert to a crystalline state that could acquire additional Ca²⁺. Between pH 7.6 and 7.8 this phase exhibited optimal low solubility and sufficient metastability to support rapid Ca²⁺ and Pi acquisition from the extravesicular fluid. It is physiologically significant that the pH of the extracellular fluid in the hypertrophic region of the growth plate is poised at 7.6 (63, 339); this is the center of the active pH range for MV mineralization. Third, the only known calcium phosphate phase with the properties described above is ACP (amorphous calcium phosphate) (444). However, in intact MVs, ACP is in a hidden form not readily discernible with FT-IR or FT-Raman (387-388) or X-ray diffraction (63). It was only by careful biochemical dissection of the nucleational core from isolated MVs that the true nature of the nucleational complex was revealed.

16.2. Isolation of the nucleational core

The following chemical dissection procedure describes how the nucleational core was isolated (Figure 19) (288, 443, 457). It entailed sequential extraction of active MVs: 1) with hyaluronidase to remove hyaluronate, 2) with water to dissociate proteoglycan link and core proteins, 3) with pure collagenase to remove attached collagens, and 4) with 1 M NaCl to dissociate other vesicle proteins. (Critically, the 1 M NaCl had to be made in Ca²⁺-containing SCL to prevent dissolution of the ephemeral mineral.) Finally, 5) a detergent, e.g. CHAPS, was used to remove MV membrane phospholipids. The properties of the nucleational core isolated by this procedure - taken with data from the preceding studies - revealed that this was the material responsible for induction of mineral formation by MV.

16.3. Chemical characterization of the nucleational core

What are the morphological, chemical and physical properties of the nucleational core? While the just cited studies revealed much of what is currently known about this ephemeral material, much more is now understood.

16.3.1. Electrophoretic Analysis of Proteins of the nucleational core

Proteins remaining in the nucleational core appear to contribute significantly to its ability to lower the intravesicular Ca²⁺ ion activity. The most obvious of these are the MV annexins. And indeed, SDS-PAGE electrophoretic analysis of the nucleational core reveals that AnxA5 is the main protein component (Figure 20). AnxA5 has been shown to bind to Ca²⁺ in the presence of acidic phospholipids with a dissociation constant in the range of 5 x 10^-7 M (Figure 5) (60). Taken with its abundance in MVs (78), it is evident that it could trap significant amounts of Ca²⁺. Importantly, Pi was shown not to interfere with the formation of these ternary PS-Ca²⁺-annexin complexes (Figure 4C) (61). Before development of the procedure for isolating the nucleational core an earlier type of biochemical dissection of MVs revealed that the MV annexins can only be removed by using powerful Ca²⁺-chelating agents such as EGTA (Figure 4A) (60, 61). Further, this family of proteins is uniquely able to be partitioned into organic phases under highly acidic conditions, revealing their amphipathic nature (93). In fact, the annexins cannot be removed from MVs without destroying the ability of the MVs to mineralize. Thus, it is not surprising that AnxA5 was the major protein in the nucleational core.

16.3.2. Chromatographic analysis of lipids in the nucleational core

On the other hand, the detergent, CHAPS used to expose the nucleational core, extracted nearly all of the resident phospholipids. The only lipids that remained were free cholesterol and an unidentified lipid. After 1M-NaCl treatment, CHAPS displaced most of the phospholipid. (Another detergent, Digitonin, while removing most of the cholesterol, left the nucleational core phospholipids intact.) However, the fact that the nucleational core was still functional after CHAPS treatment indicates that CHAPS must have substituted for the phospholipids in the core structure, somehow preserving its activity. However, in the native nucleational core, significant amounts of PS-CPLX are present. About 10% of the total Ca²⁺ and Pi in MV is in the form of organic solvent-soluble PS-ACP complexes (64); the great majority of the Ca²⁺ (85-90%) and about half of the Pi and Mg²⁺ in MV are insoluble in organic solvents - i.e. present as non-lipid-soluble complexes. It is now evident that the PS-ACP complex (PS-CPLX) is a key component of the nucleational core of native MVs.

16.4. Physical characterization of the nucleational core

Several types of physical measurements were made on the CHAPS-isolated nucleational core. TEM and FTIR, as well as ³¹P-NMR and its solubility properties were used to analyze the structure of the nucleational core (443, 457).

16.4.1. Transmission electron microscopy (TEM)

TEM reveal that the nucleational core is composed of aggregates of particles ~110 nm in diameter, with ~1.0 nm diameter subunits apparently composed of clusters of Ca²⁺ and Pi ions (457). The ~1.0 nm diameter of the subunit is similar in size to the core structure of ACP measured by radial distribution function analysis (458). This finding supports the conclusion that ACP is a major component of the nucleational core. There was no evidence for the presence of crystalline mineral in the nucleational core.

16.4.2. FT-IR analysis

FT-IR spectral analysis of the nucleational core revealed the presence of a number of characteristic features: strong, sharp C-H stretch bands (2922 and 2851 cm^-1) characteristic of acyl chains (lipids), strong amide bands (~1654 and ~1545 cm^-1) characteristic of proteins, a strong P-O stretch band (1054 cm^-1) characteristic of all calcium phosphate minerals, and a broad, unsplit P-O absorption band (~558 cm^-1) characteristic of ACP (Figure 21). These data reveal chemical features that distinguish it from those of the MVs from which it was isolated. The most obvious difference is the marked increase in the intensity of the unsplit P-O absorption band at ~558 cm^-1 and the strong P-O stretch band at ~1054 cm^-1; these absorbencies finally reveal the obvious presence of ACP in the nucleational core - in contrast to its being masked in intact MVs. Further, the strong absorbencies at ~1654 cm^-1 and ~1545 cm^-1 represent the amide-I C=O stretch bands and the amide-II N-H stretch bands, respectively, clearly show that significant amounts of protein are also present in the nucleational core. The C-H stretch bands at 2922 and 2851 cm^-1 are not as intense as those in native MVs, in agreement with the reduction in lipids evident in TLC chromatograms of the nucleational core. Further, the sharp strong bands at 3507 and 3441 cm^-1 evident in the nucleational core, but not native MVs, appear to be due to the presence of the three hydroxyl moieties in the sterol nucleus of the CHAPS detergent. This supports the contention that the detergent displaced phospholipids in the nucleational core. The intense broad band at 3332 cm^-1 represents the presence of tenacious water, despite vacuum drying.

16.4.3. ³¹P-NMR analysis

The mineral initially present in MVs was also studied by solid-state, magic-angle ³¹P-NMR using proton cross-polarization (Figure 22) (457). A distinct central peak at 3.2 ppm characteristic of all calcium phosphates, including ACP, OCP or HAP, was evident, but x-ray diffraction studies of either isolated MVs (Figure 9C) (63) or the nucleational core (457) showed no evidence of crystallinity. The cross-polarization spinning sideband pattern was highly distinctive and matched that of the PS-CPLX; it did not resemble that of any of the known crystalline calcium phosphates (459-460). By measurement of the relative size of the central peak and the spinning sidebands and comparing them with known mixtures of ACP and PS-CPLX, it was estimated that only about 8-10% of the total MV-P was as the PS-CPLX; the bulk, about 90%, was present as ACP. The proportion of MV-P present as CPLX closely agrees with the ~5-10% value obtained by solvent partition of MV by Wuthier and Gore (64) noted earlier.

16.4.4. pH Sensitivity of the nucleational core

One of the most effective ways to destroy the nucleational core is to expose it to isosmotic, pH 6 citrate buffer (288). After this simple treatment, its ability to induce mineral formation was totally destroyed. This reveals that it is highly sensitive to even mildly acidic conditions. On the other hand, at pH 8, the nucleational core is highly stable and insoluble. As was discussed earlier, the nucleational activity of the MVs and the nucleational core is operative only within a very narrow pH range - between 7.4-7.8 (77) (Figure 18B). Either below or above this range, its ability to nucleate mineral formation is dramatically reduced. Below pH 7.4, the solubility of the nucleational core is too high to accrete Ca²⁺ and Pi from the intravesicular fluid to enable crystal growth. Above pH 7.8, the stability of the nucleational core is so great that it can not convert to a crystalline form that could acquire additional Ca²⁺. Only between pH 7.6 and 7.8 does the nucleational core exhibit sufficiently low solubility and adequate metastability to induce rapid mineral formation. As noted previously, the extracellular fluid at the site of calcification in the growth plate in fact has a pH of 7.6 (63, 339); this enables a maximal rate of mineral formation. Thus, in accord with all the data just presented, the solubility properties of the nucleational core also match those of ACP (444); but the story gets more complicated.

16.5. Summary of nucleational core structure and composition

To summarize, SDS-PAGE electrophoresis, X-ray diffraction, TEM, FTIR and ³¹P-NMR findings - as well as the solubility properties of the nucleational core - all clearly indicate that ACP is a major component of the nucleational core. Localized within the vesicle lumen, this quasi-stable highly hydrated solid phase of amorphous mineral forms the bulk of the nucleational core. However, ³¹P-NMR and the early solvent partition assays indicate that only about 10% of the Pi is present in the form of PS-CPLX. Off hand, this small percentage suggests that PS-CPLX would have little to do with nucleation of mineral formation. However as we shall soon discover, this small critical amount of PS-complex is the trigger for mineral nucleation. The insolubility of the nucleational core is the initial driving force for Ca²⁺ and Pi accumulation. Its metastability is what enables the nucleational core to convert to crystalline OCP and drive rapid mineral formation.

17. RECONSTITUTION OF THE NUCLEATIONAL CORE: ROLE OF PS-Ca²⁺-Pi COMPLEX (PS-CPLX) IN MV CRYSTAL NUCLEATION

To fully understand how the native MVs work, it has been necessary to take them apart and put them back together again. Being able to chemically dissect and then reconstitute MVs and their key components provides assurance that a proper understanding of the modus operandi of these intriguing biological nanostructures has been achieved. After it was possible to consistently isolate functionally operational MVs that induced mineral formation in a predictable manner in an ionic environment closely matched to that occurring in the native tissue, it became possible to disassemble the native vesicles by biochemical dissection. As noted previously, this led to the important discovery that a nucleational core exists in MVs that enables them to induce mineral formation - even in the absence of the protective environment of the vesicle lumen. The next challenge was to reconstruct from pure components a functional nucleational core that was capable of inducing mineral formation of a type and character that matched that of the native vesicles. The final challenge to demonstrating that the mechanism of MVs was truly understood was to reconstitute functional synthetic vesicles (liposomes) from pure precursors that were capable of inducing the same mineral type produced by native MVs. This has been largely accomplished, although there will always be room for continued refinements to make the constructs match ever more closely the native structure.

17.1. Reconstitution of the nucleational core

The discovery of the nucleational core (NC) prompted attempts to reconstitute it from known pure materials. This required extensive characterization of its components (288, 457). As already noted, these studies revealed that key components of the NC were: ACP, AnxA5, and PS-CPLX. Also evident were regulatory factors that alter the rate at which the NC could induce mineral formation. A key regulatory component was found to be Mg²⁺, the principal intracellular divalent cation. Another important factor was the lipid composition of the NC; for example, PS was the most efficacious lipid. Dilution of PS with other phospholipids normally present in MVs, or nonpolar lipids such as free cholesterol, had a progressively inhibitory effect on the activity of the synthetic NC. The first effort to reconstitute the NC was made in 1996 by Wu et al. (443), who centered on the PS-Ca²⁺-Pi complex. Various pure lipids, electrolytes and proteins were combined to form a synthetic nucleationally active complex, analyzing its rate and extent of mineral formation. While inclusions of PE or SM with PS were strongly inhibitory, co-incorporation of AnxA5 increased nucleation activity, but much more was needed to refine the reconstruction of the physiological complex.

17.2. Mathematical analysis of the kinetics of mineral formation

It was not until a decade later that a systematic attempt was made to reconstitute the NC using a biomimetic approach. To accurately ascertain the effects of each permutation, a more precise method for measuring the onset time, the kinetic rate and the amount of mineral formed needed to be devised. This led to the development of a high throughput method for studying the kinetics of mineral formation using in vitro modeling of MV mineralization (332). The method was based on light scattering by nascent crystallites at 340 nm (461) that enabled monitoring of mineral formation at regular intervals without disturbing the system using an automated plate reader. It yielded precise replicate values that typically agreed within <5%. Using a Mg²⁺-containing biomimetic PS-CPLX as the nucleating material, it was found that mineral formation followed a sigmoidal pattern - as it does with authentic MVs. There was a quiescent induction period; then rapid formation ensued for a limited time; this was followed by a distinct decline in rate that continued to slow, ultimately reaching a maximal asymptotic value. Key to quantization of mineral formation was the use of first-derivative analysis that defined the induction time, the rate, the length, and the amount of initial mineral formation (Figure 23). Through use of a five-parameter logistic curve-fitting algorithm, the maximal amount of mineral formation could be predicted accurately.

17.3. The Physical and nucleational properties of pure PS-CPLX

Studies aimed at reconstituting the nucleational core revealed that the entity most responsible for nucleation of crystalline mineral was PS-CPLX. TEM analysis of the pure synthetic PS-CPLX revealed extended membrane-like structures composed of PS-CPLX units 7.6-8.0 Angstroms in diameter that form quasi-crystalline arrays; these appear to be nucleating mineral (arrows, Figure 24). Evidence of this short-range two-dimensional order comes from FT-IR analyses which consistently reveal splitting of the 560 - 600 cm^-1 P-O bond resonance. Pure Mg²⁺-free PS-CPLX can nucleate calcium phosphate microcrystalline mineral almost immediately upon exposure to SCL, very much like that induced by HAP (Figure 25) (444). Thus, while the PS-CPLX comprises only a quantitatively minor component of the NC (289, 458), it is vital to its nucleation properties.

17.4. Requirements for PS-CPLX formation

During MV formation, PS-CPLXs can readily form at pH 6.7 (266), conditions like that present in the perimeter of growth plate chondrocytes (Figure 15C). In contrast, the ACP component of the nucleational core cannot form at this pH and can occur only after the vesicles are released into the extracellular fluid where the pH (7.6) favors ACP formation. Other requirements for PS-CPLX formation were studied recently by Wu et al. (131) using the just described in vitro model system (332). A key requirement in the formation of PS-CPLX is that excess Pi must be present with the lipid before introduction of Ca²⁺ (266, 443, 462). This feature occurs in cells during formation of MVs. There, levels of cytoplasmic Pi are high (63) and levels of Ca²⁺ are low (306) - prior to Ca²⁺ being released during the mitochondrial permeation transition that engenders MV formation (306, 314, 383). If Ca²⁺ interacts with PS in the absence of Pi, it leads to formation of 2:1 PS-Ca²⁺ complexes that have no nucleational activity (266, 373). This reveals that only under circumstances in which high cytosolic Pi is present, is it possible to form PS-CPLX in vivo. On the other hand, addition of high levels of Ca²⁺ to neutral solutions rich in Pi leads to formation of amorphous calcium phosphate (ACP) an ephemeral phase that spontaneously converts to HA (386, 444) unless stabilized with various agents such as Mg²⁺ (463-464), certain proteins (465-467), and acidic lipids such as PS (266). Interaction with PS during ACP formation spontaneously leads to production of PS-CPLX (221). However, PS-CPLXs do not form instantly when Ca²⁺ comes in contact with PS and Pi.

17.5. Molecular rearrangements during PS-CPLX formation

In recent studies by Wu et al. (131), micelles of PS in a K⁺- and Pi-rich intracellular buffer were used to model in vivo PS-CPLX formation. Addition of 4 Ca²⁺ per PS was minimally required to form nucleationally competent complexes. During the initial 10-min "ripening" period, light-scattering data revealed significant contraction in the size of the nascent PS-CPLX particles. FT-IR analyses showed that this was accompanied - not only by major deprotonation of HPO₄^2- to PO₄^3- - but with rearrangement of the phosphate groups, producing features that mimic those seen in crystalline calcium phosphate mineral (i.e. splitting of the P-O band in the 560-600 cm^-1 region (Figure 26) (131). Molecular dynamic simulations indicated that the initial interactions between Ca²⁺ and Pi, and PS (i.e. its carboxyl, amino, and phosphodiester groups) produced interatomic distances in the incipient PS-CPLX that agreed closely with previous RDF-EXAFS (radial distribution function x-ray-absorption fine structure) measurements on the mature complex (468) - except for an elongated Ca-N distance. During this critical period, atomic rearrangements involving deprotonation of the PS amino group, led to contraction and stabilization of the PS-ACP complex. Molecular simulations revealed the following: interaction between Ca²⁺ and the (PS-NH₃⁺∙∙∙HPO₄^2-)^- ion-pair leads to formation of a (Ca²⁺∙∙∙PS-NH₃⁺∙∙∙HPO₄^2-)⁺ adduct. Because of its net positive charge, deprotonation of the PS-NH₃⁺ (pK_a = 9.8) rather than the HPO₄^2- (pK_a = 12.3) group would lead to (Ca²⁺∙∙∙PS-NH₂∙∙∙HPO₄^2-)⁰ as a product. The electrically neutral R-NH₂ group, having a pair of unshared electrons, would interact with Ca²⁺, forming a shortened Ca-N bond. Such Ca-N bonding is an important feature of strong chelating agents like EDTA, but in PS-CPLX the affinity for Ca²⁺ is weaker and the reaction occurs more slowly. Contraction in the size of the polar head group would enable construction of planar hexagonal assemblies that closely match the lattice parameters of OCP and HAP, forming an almost perfect nucleational agent - as was evident from the rapidity in which mineral formation occurred upon addition of the PS-CPLX to SCL (Figure 25). These events obviously occur intracellularly during the critical period after release of mitochondrial Ca²⁺ that sets in motion the formation of PS-CPLX on the inner surface of the plasma membrane and becomes incorporated into the MVs when they are released from the cell. In the more alkaline (pH 7.5) extracellular environment, the excess Ca²⁺ and Pi in the nascent MVs rapidly form ACP, which becomes the quantitatively major component of the NC. However, the nucleational properties ascribed to the NC for crystal formation largely reside in the PS-CPLX.

17.6. Effect of Mg²⁺ incorporation on the nucleational activity of PS-CPLX

A complicating factor in understanding crystal nucleation is the fact that in vivo ACP and PS-CPLX both contain high levels of Mg²⁺ (64); this is true simply because Mg²⁺ is far more abundant than Ca²⁺ in the cytoplasm. Mg²⁺ stabilizes both ACP and the PS-CPLX in the NC, inhibiting their conversion to crystalline calcium phosphates. As noted above, Mg²⁺-free PS-CPLXs are almost as potent as HAP in nucleating crystal formation, initiating mineral formation in a matter of seconds (Figure 25). Mg²⁺-containing PS-Ca²⁺-Pi complexes are much less active, requiring much more time; incorporation of a little as 0.5 mM Mg²⁺ is sufficient to extend the induction time by 3 h; increasing this to 0.75 mM or higher delay induction for many additional hours (Figure 27A) (137). How Mg²⁺ does this is readily apparent from FT-IR analysis, which reveal that it prevents the formation of the highly ordered P-O bonds evident as splitting of the 560-600 cm^-1 resonances observed in all crystalline calcium phosphates (Figure 27C).

17.7. Role of AnxA5 in MV crystal nucleation

Here is where AnxA5 has been shown to play a critical role; its presence counteracts the inhibitory effects of Mg²⁺. If AnxA5 is incorporated at the time of PS-CPLX formation, the induction time to mineral formation can be reduced from 16 h to only 3-4 h (Figure 28) (136). This acceleration in nucleation is of key importance in rapidly growing species like broiler-strain chickens where longitudinal growth can be up to 50-60 micrometers per hour. Since hypertrophic chondrocytes are ~20 micrometers in diameter, this means that mineralization of 2-3 cells must be completed every hour to maintain structural integrity. How can AnxA5 so greatly enhance the rate of mineral formation? There are several important clues. First, high resolution scanning electron micrographs of AnxA5 associated with PS-rich bilayers reveal an orderly hexagonal pattern of triskelion shaped trimers on the membrane surface (126, 128-130). Second, the dimensions of this hexagonal pattern closely coordinate with the dimensions of hexagonal crystal structures of OCP and HAP. Third, at the interfacial regions where AnxA5 interacts with Ca²⁺ and the PS polar head groups, domains are created that apparently exclude Mg²⁺. This is because in the presence of PS, AnxA5 has no affinity for Mg²⁺ - in stark contrast to its high affinity for Ca²⁺ (60, 61). For example, if AnxA5 and PS micelles are co-incubated with low levels of Ca²⁺, they rapidly form complexes that co-sediment with the PS liposomes. In contrast, if Mg²⁺ is substituted for Ca²⁺ even at much higher levels, AnxA5 does not bind Mg²⁺ or co-sediment with the PS liposomes (Figures 4B-C). This is because the coordination chemistry of Mg²⁺ is very different from that of Ca²⁺ (469-471). Whereas Ca²⁺ can form flexible bond lengths and angles with as many as 7 different ligands, Mg²⁺ coordinates with only 6 ligands in a highly rigid, precise geometry (472-473). Thus, Mg²⁺, because of its rigid coordination geometry, blocks the nucleational activity of PS-CPLX, and prevents ACP dismutation to HAP or OCP.

17.8. Mechanism of MV nucleation of OCP

A key event occurs during the induction phase that immediately precedes rapid mineral formation (332). At this time the PS-CPLXs, in conjunction with AnxA5, come together to produce a template from which OCP-like crystals rapidly proliferate. As previously noted, the Ca²⁺/Pi stochiometry of Ca²⁺ and Pi accumulation during the rapid mineral formation period is 1.33 (Figure 10; Figure 16B), much lower than that of stoichiometric HAP (1.67) (57, 65). In fact the stoichiometry of Ca²⁺/Pi during the rapid formation period is 1.33 - very close to the Ca²⁺/Pi ratio of OCP (1.33) (317, 385, 449). However, it is evident that the initial mineral formed is not simple stoichiometric OCP, which when mature forms large platy crystals (474-476). Solid-state ³¹P-NMR analyses of the initial mineral forms by Glimcher's group (477) show that the chemical shifts of the HPO₄^2- groups match those of both OCP (isotropic) and dicalcium phosphate dihydrate (anisotropic). The data reveal a HPO₄^2--rich Ca²⁺ phosphate that is distinct from these well established crystalline phases. This is consistent with FT-IR analyses of initial MV mineral formation, which also indicates a HPO₄^2--rich mineral phase (387). Thus, the initial crystalline phase is unique and transient, but over a matter of a few hours progressively converts to a non-stoichiometric, carbonate-containing HAP with a Ca²⁺/Pi ratio of ~1.60 (160).

18. MODELING OF MATRIX VESICLES USING SYNTHETIC UNILAMELLAR LIPOSOMES

The ultimate challenge in understanding how MVs work is to reconstruct functional biomimetic models from pure components in proportions like those present in native vesicles. Our initial efforts involved incorporation of pure lipids, electrolytes and protein constituents into an emulsion made from a mild detergent, n-octyl beta-D-glucopyranoside (318). Using dialysis in an intracellular-like high-Pi buffer to remove the detergent, large (~300 nm diameter), stable, unilamellar vesicles (LUV) were formed; they encapsulated the Pi-rich buffer, and can include proteins like AnxA5. The LUV membrane contained PC, PS and free cholesterol (7:2:2, molar ratio); they enclosed 25-100 mM Pi, and had a ~300 nm diameter - each simulating values found in native MVs. When these LUVs were tested for their ability to induce mineral formation, incubation in SCL for periods as long a 48 h caused no ⁴⁵Ca²⁺ uptake. This documented that the LUVs were stable and not leaky.

18.1. Effects of Ca²⁺ ionophores on ⁴⁵Ca²⁺ uptake

To enable entrance of Ca²⁺, various ionophores were initially used. Mineral formation was assessed by measuring the amount of ⁴⁵Ca²⁺ uptake, as well as FT-IR analysis of the LUVs, following incubation in SCL. If the Ca²⁺ ionophore, A23187 was added to SCL, rates of ⁴⁵Ca²⁺ uptake occurred that were comparable to those of native MVs (Figure 29A). After incubation of these A23187-activated LUVs for 24 h, sufficient Ca²⁺ uptake occurred to produce substantial amounts of mineral (318). Significantly, the mineral formed had an acid phosphate (HPO₄^2-) rich, OCP-like FTIR spectrum essentially identical to that of native MVs (Figure 29B). Thus, this simple LUV model mimicked two key features of MV mineral formation: 1) the soluble Pi-rich luminal fluid was sufficient to drive Ca²⁺ uptake, and 2) entrance of Ca²⁺ into this Pi-rich environment formed an OCP-like mineral just like that seen in native MVs.

18.2. Minimal effects of annexin A5 on ⁴⁵Ca²⁺ Uptake

Based on the assumption that AnxA5 was functioning as a Ca²⁺ channel in the MV membrane, attempt was made to model MV Ca²⁺ uptake using AnxA5 added to the SCL. Surprisingly, it only weakly supported ⁴⁵Ca²⁺ uptake by the LUVs (319), which was in stark contrast to ionophore A23187 which enabled rapid accumulation of ⁴⁵Ca²⁺. Thinking that the protein perhaps needed to be present within the lumen, we encapsulated the AnxA5 during formation of the LUVs and tested them for their ability to take up ⁴⁵Ca²⁺. Surprisingly, even here AnxA5 also failed to produce significant ⁴⁵Ca²⁺ uptake. At first glance, these findings appear to be at variance with those of Kirsch who reported that AnxA5 stimulated Ca²⁺ flux across liposomal membranes (478). However, in those studies only tiny amounts of Ca²⁺ were accumulated (detectable only by a Ca²⁺-sensitive fluorescent probe, FURA). There was no encapsulated Pi in those liposomes, and no attempt was made to determine if AnxA5 enabled sufficient Ca²⁺ uptake to induce mineral formation.

18.3. Effects of detergents and phospholipase A₂

There were two other important findings from the Blandford study (318). First, treatment of the Pi-loaded liposomes with the detergent, CHAPS, greatly stimulated ⁴⁵Ca²⁺ uptake. This occurred after a 24 h control incubation of the LUVs in the absence of detergents in which no ⁴⁵Ca²⁺ uptake occurred. Thus, treatment with CHAPS led to rapid and extensive uptake of ⁴⁵Ca²⁺. If the effects of CHAPS were studied sequentially with those of ionophore A23187 in the same LUV preparation, the effects of the two agents were additive. However, from a physiological perspective, the most significant finding of this study was that treatment with phospholipase A₂ caused rapid and extensive ⁴⁵Ca²⁺ accumulation. Again, this occurred after a control 24 h incubation of the liposomes in ⁴⁵Ca²⁺-labled SCL in the absence of either ionophores or strong detergents in which no ⁴⁵Ca²⁺ uptake occurred (Figure 29A). These findings clearly show that disruption of the vesicle membrane is sufficient to trigger rapid Ca²⁺ uptake. At the time these experiments were done, AnxA5 was thought to be an important Ca²⁺ channel in MVs (125, 479). While we were greatly puzzled by its inability to support ⁴⁵Ca²⁺ uptake by the LUVs, we had not yet discovered the extensive breakdown of phospholipids that occurs in native MVs when incubated in SCL (207). It is now apparent that breakdown of membrane phospholipids via action of phospholipase activity is an important part of the natural mechanism for allowing Ca²⁺ access to the lumen of the vesicles. However, as noted above (Figure 28), reconstitution of a biomimetic nucleational core clearly reveals that AnxA5 exerts a potent stimulatory effect on its nucleational activity. Its principal effect is to markedly accelerate the onset of mineral formation by enhancing the nucleation of OCP formation within the MV lumen.

18.4. Proteoliposomal MV Models by other groups

As noted above, Kirsch et al. (478) produced large unilamellar liposomes for studying the ability of AnxA5 to mediate Ca²⁺ uptake using FURA as an indicator. Very recently, Simao et al. reported on the production of proteoliposomes as models of MVs for study of initiation of skeletal mineralization (190, 479). However, in none of these model studies was there any measurement of the kinetics of mineral formation or of the nature of the mineral formed by their MV models. In the Kirsch et al. study there was no indication that the level of Ca²⁺ uptake caused by the incorporated AnxA5 was sufficient to induce mineral formation. In the Simao et al studies, although they were able to incorporate specific enzymes (TNAP, NPP1) into the liposomes, and show that these enzymes exhibited appropriate activity when incorporated (190, 479), there was no indication that these enzymes enabled mineral formation by the liposomes. In another recent study, MVs were isolated from cultured osteoblasts derived from wild-type and TNAP-, NPP1- and PHOSPHO1-gene-deleted mice and studied for their ability to hydrolyze putative phosphosubstrates and induce mineral formation (193). However, rather than study the ability of intact MVs to induce mineral formation, these authors first lysed the MVs with deionized water and then added this lysate to a calcification medium that contained at least 3 mM (and most likely 10 mM) AMP, as well as 2.2 mM Ca²⁺ and 1.6 mM Pi. Given these artificial conditions, TNAP would have generated sufficient Pi from the added AMP (plus that from the calcifying medium) to raise the Ca²⁺ x Pi ion product to a level where spontaneous mineral precipitation would occur. Freeze-clamp zonal analysis of growth plate cartilage by Shapiro's group (480) have shown that levels of AMP are only 0.25-0.65 mM, values nearly 10-fold lower than those used in this study.

19. REGULATION OF MV CALCIFICATION

Many factors are involved in the regulation of the nucleation of mineral formation by MVs - and the proliferation and growth of crystals in the extravesicular matrix. Those primarily influencing MV nucleation include: a) the inhibitory effects of Mg²⁺, Zn²⁺, and PPi, b) the aforementioned synergistic effects of the MV annexins and PS, and c) the inhibitory effects of PE and other lipid components of the MV membrane. Those specifically influencing subsequent crystal growth and proliferation are: d) PPi - a powerful inhibitor of apatitic crystal growth, e) the proteoglycans that encompass and limit spread of the newly forming mineral in cartilage, and f) the collagens that facilitate and guide crystal growth to form the strong mineral-collagen composite typical of bone. In addition, there are a series of mineral-regulating proteins that bind to apatite and generally inhibit crystal growth. The effects of nearly all of these regulators are complex and to one degree or another affect both nucleation and crystal growth.

19.1. Factors that primarily effect nucleation

The factors that primarily influence the ability of MV to nucleate mineral formation will be discussed first. Experimental studies in which these effects were most clearly delineated were those aimed at reconstituting the native nucleational core present in MVs (116). As a control, ACP was prepared from a Mg²⁺-containing phosphate buffer modeled after the composition of the cytosol of growth plate chondrocytes (63); it was used as the nucleator to which the various lipid-containing ACP complexes were compared. Parameters studied were: induction time, rate of rapid mineral formation, nucleational potential and amount of final mineral formed. In this Mg²⁺-containing system, surprisingly, incorporation of PS during formation of the ACP to form the PS-Ca²⁺-Pi complex significantly impeded mineral formation, decreasing the rate of rapid mineral growth and the final amount of mineral formed (116). However, incorporating AnxA5 along with PS, led to synergistic acceleration of the onset of mineral formation and enhancement of the rate of mineral formation - when compared with the ACP combined with AnxA5 alone (Figure 28). Thus the marked stimulatory effect of incorporating PS in this Mg²⁺-containing native ACP was dependent on the co-presence of AnxA5.

19.2. Effects of phospholipid composition

As noted earlier, by varying the composition of the lipid mixture used in forming the annexin-containing ACP-based nucleational complex it was documented that pure PS was most effective in stimulating nucleation (443). Dilution with PE, and the other lipids present in MVs progressively reduced the effectiveness of these nucleators. Thus, it is not fortuitous that PS is consistently enriched in all MV preparations, and that in vivo mechanisms (i.e. PS synthase) are in place to facilitate PS formation - even in the absence of ATP (207, 268). Of physiological importance, the inhibitory effect of incorporating PE into the complex was essentially overcome by the incorporation of AnxA5 (443).

19.3. Stimulatory effect of different annexins on MV nucleation

As was evident above (Figure 28), AnxA5 is important for enhancing the nucleational activity of biomimetic Mg²⁺-containing nucleational complexes. Although the presence of Mg²⁺ is ignored in many studies, it is the major divalent cation present in all living cells - including growth plate chondrocytes. When the protein composition of nucleationally active MVs was first characterized, annexins A5, A2 and A6 were identified as being quantitatively major proteins (60, 61, 94); subsequent proteomic studies have confirmed this and also revealed the presence of additional annexins - A1, A11 and A4 (78, 87). One of the key features of all of the annexins is their acidic phospholipid-dependent Ca²⁺-binding affinity. For example, half-maximal binding affinity of Ca²⁺ for AnxA5 in the presence of PS is ~5 x 10^-7 M (Figure 5) (60). Inorganic P does not interfere with this lipid-dependent Ca²⁺ binding property (Figure 4C). It was thus of major interest to determine the ability of the different MV annexins to stimulate mineral formation when incorporated into synthetic nucleational complexes modeled after those present in native growth plate MVs. In a recent study, the effects of native chicken cartilage annexins A5, A2 and A6 were compared with those of native human placental AnxA5 and avian liver AnxA5 (136). It was found that all of the AnxA5s potently activated the nucleational activity of Mg²⁺-containing PS-Ca²⁺-Pi complexes. From a physiological perspective, however, native avian cartilage AnxA5 was the most effective form of the annexins. It is significant that avian cartilage annexin A6 was less active than AnxA5, and that annexin A2 had little or no stimulatory activity.

19.4. Effects of non-apatitic electrolytes

Polyvalent electrolytes that occur as minor constituents of biological apatites can exert a profound inhibitory influence on formation and growth of calcium phosphate crystals. The three electrolytes that exert the most direct influence on this process are Mg²⁺, Zn²⁺, and PPi - although CO₃^2- substitution appears to be an almost universal biological "contaminant" that regulates apatitic crystal size (481-484). From a biological perspective, the key issue is the relationship between the levels needed to significantly affect the various mineralization parameters and the levels known to be present in biological fluids. Here it is important to distinguish between the levels present in blood serum (or more specifically, the tissue lymph) and those present within the cells where MVs are formed (e.g. growth plate chondrocytes). More relevant still are the levels present in ultrafiltrates of these fluids, eliminating the effects of ions bound to macromolecules like proteins and proteoglycans or anionic phospholipids. Even more relevant are the actual chemical ion activities that are modulated by low molecular weight solutes like cellular metabolites. Chemical analysis of MVs, their constituent PS-Ca²⁺-Pi complexes, and the intracellular and extracellular fluid of growth plate tissue all reveal the presence of substantial levels of Mg²⁺, Zn²⁺ and PPi. The most relevant comparisons are between the inhibitory IC₅₀ values of these ions for different mineralization parameters and the levels present in the bathing fluid.

19.4.1. Effect of Mg²⁺

MVs are formed by growth plate cells whose cytosolic levels of Mg²⁺ are much higher than those of Ca²⁺ (Table 2) (63). Thus, all PS-Ca²⁺-Pi complexes formed in vivo incorporate both Mg²⁺ and Zn²⁺; in fact analyses of the complexes present in avian MVs have shown that Mg²⁺ is almost as abundant as Ca²⁺ (64). Our recent studies have revealed that this level of Mg²⁺ is profoundly inhibitory to the nucleational activity of PS-Ca²⁺-Pi complexes, greatly reducing their rate (Figure 27A) (137). This inhibitory effect is physiologically important because as noted above rapidly growing species like broiler-strain chickens have such fast longitudinal growth that speedy nucleation is essential for mineral formation to keep pace. Thus, delays in the induction time of as little as 1-2 h, or 20% reduction in the rate of mineral formation, if consistent, would lead to the development of rickets in these birds. Mg²⁺ is the predominant intracellular divalent cation in all living cells. Even during the differentiation, maturation and hypertrophy of growth plate chondrocytes the cytosolic levels of ultrafilterable Mg²⁺ are over 10-fold higher than those of Ca²⁺. This is true at the time of heavy loading and release of mitochondrial Ca²⁺ that triggers the blebbing of Ca²⁺, Pi, and PS-CPLX-primed MVs. In fact, in MVs that form from these cells, the ratio of soluble Mg²⁺/Ca²⁺ was ~9/1 (64); this level of Mg²⁺ is sufficient to essentially block the conversion of ACP to HAP. While the molar ratio of Mg²⁺/Ca²⁺ in the insoluble fraction was much lower (0.15�0.03), this level was still sufficient to delay the conversion of ACP to HAP for a matter of days (444, 463). And as just illustrated, recent studies on synthetic PS-Ca²⁺-Pi complexes showed that incorporation of low, subphysiological levels of Mg²⁺ led to profound inhibition of their nucleational activity (137). FT-IR analyses showed that Mg²⁺ altered their structure, greatly disordering the 560-600 cm^-1 region (nu₄ antisymmetric P-O bending modes) characteristic of crystalline phosphate groups (Figure 27C). Mg²⁺ prevented the splitting of this region, indicating destruction of the short-range crystal-like P-O structure normally present in Mg²⁺-free PS-Ca²⁺-Pi complexes. FT-IR analyses also showed an increase in the 1062 cm^-1 disordered nu₃ HPO₄^2- groups and a decrease in the 1032 cm^-1 ordered nu₃ PO₄^3- groups of the complex. As noted earlier, packing of clusters of PS-Ca²⁺-Pi complexes in the MV membrane produces assemblies that form sites for OCP mineral nucleation. Mg²⁺ obviously interferes with the ordering of P-O resonances at these sites; however, as we have just seen, if AnxA5 is also incorporated into the PS-Ca²⁺-Pi complexes, the effects of Mg²⁺ are largely abrogated (Figure 28) (116, 136). This is because AnxA5 has no affinity for PS in the presence of Mg²⁺. It is also important to note that when Mg²⁺ was present in the extravesicular lymph its inhibitory effects were far less; supraphysiological levels as high as 2.5 mM cause minimal delay in the onset of mineral formation (137). These findings document that AnxA5 is a key factor that enables Mg²⁺-rich MVs to form HAP.

19.4.2. Effect of Zn²⁺

Studies of Sauer et al. have shown that levels of extravesicular Zn²⁺ as low as 5 microM are sufficient to totally block MV mineral formation (Figure 17A); this Zn²⁺ blockage can be rapidly overcome by subsequent treatment with the Zn²⁺ chelator, o-phenanthroline (Figure 17B) (317). Paradoxically, as noted previously, analysis of MVs revealed the presence of surprisingly high levels (1.57�0.17 mM) of total Zn²⁺ within MVs. While most of the Zn²⁺ (~60%) was associated with the insoluble phase, a second phase (~40%) could be lost upon exposure of the MVs to Ca²⁺-free media. Evidentally, this portion of Zn²⁺ binding within the MVs was Ca²⁺-dependent (385). However, no Zn²⁺ was detected in expressed growth plate cartilage lymph using highly sensitive atomic emission spectrometry (317). Nonetheless, this intravesicular Zn²⁺ had minimal inhibitory effect on MV mineralization; these vesicles readily induced mineral formation (317). Our recent studies confirm this (137); FT-IR analyses reveal that Zn²⁺ incorporated into PS-CPLX has no disruptive effect on its 560-600 cm^-1 crystal-like structure (Figure 27C), and had minimal effect on mineral formation (Figure 27B). Normal levels of Zn²⁺ incorporated into the nucleational core of MVs did not interfere with mineral formation. On the other hand, as noted above, the presence of low micromolar levels of Zn²⁺ in extravesicular lymph caused profound inhibition of MV mineralization. The key question is how? Blockage of the Ca²⁺ channel activity of AnxA5 at first seemed logical - based on electrophysiological studies by Arispe et al. showing Zn²⁺ blockage of AnxA5 Ca²⁺ conductance (125) - presumably because of tight Zn²⁺ binding to Ca²⁺ sites in the ion channel. But there is a more plausible explanation. As noted earlier, in the presence of Ca²⁺, AnxA5 trimers form planar hexagonal arrays on the surface of PS-rich lipid bilayers (Figures 30A, 31A-B) (129, 130). These planar arrays of AnxA5 trimers appear to be key mineral nucleation sites on the PS-rich MV membrane. Importantly, in studies of the crystal structure of Zn²⁺-AnxA5 adducts (485), while the 3-D structure of the AnxA5 monomers is not altered, Zn²⁺ does bind to a coordination site formed at the interface of those monomers - three His-286 residues (Figure 30C) - distorting the trimer structure (Figures 30B, 31C-D). This seriously disrupted the normally planar hexagonal arrays (Figure 31A-B), which would readily explain why addition of o-phenanthroline can so rapidly restore mineral formation by Zn²⁺-inhibited MVs (Figure 17B) (317). Chelation of Zn²⁺ would allow rapid reformation of the planar AnxA5 trimers and the hexagonal nucleation array, because Zn²⁺ does not alter the AnxA5 monomer structure.

19.4.3. Effect of pyrophosphate (PPi)

From a metabolic perspective, it is important to realize that PPi is the byproduct of several key activating enzymes in all cells - including growth plate chondrocytes - involved in glycogen, proteoglycan, DNA and RNA synthesis, as well as activation of fatty acid catabolism. Most of these reactions require phosphatase enzymes to cleave PPi, which pull the reactions forward. PPi is also formed via NPP1 during chondrocyte and osteoblast activity (165, 193). PPi adsorbs tightly to HAP (486-487) and is capable of blocking crystal growth (488-490). In fact, the concept that PPi is involved in the regulation of skeletal mineral deposition goes back almost 50 years to the hypotheses by Fleisch (194, 491-496). Zonal analysis of PPi and Pi in fetal bovine tissue revealed that during the onset of mineralization in growth plate cartilage the amount of PPi and Pi per unit dry weight both increased, but the ratio of PPi to Pi declined precipitously (497) (Figure 32). While TNAP was already known to hydrolyze PPi (178, 498), it was not established that PPi was actually destroyed in vivo by TNAP during growth plate mineralization. In fact, the levels of PPi in growth plate tissue actually increase - but far less rapidly than levels of Pi (497). Further, in hypophosphatasia, a disease of TNAP deficiency, it was known that there was not only defective bone mineralization, but that levels of PPi were elevated in both urine (499) and blood plasma (500). Studies by Ali et al. showing that TNAP is highly enriched in MVs (53, 158) were followed by those of Register et al. comparing the inhibitory effects of PPi and two non-hydrolyzable bisphosphonates (EHDP and Cl₂MDP) on ⁴⁵Ca²⁺ and ³²Pi uptake by TNAP-rich MV-enriched microsomes (501). These studies showed that while inhibition with EHDP and Cl₂MDP persisted for many hours, inhibition by PPi did not; thus revealing that hydrolysis of PPi by MV enzymes was able to overcome its inhibition of apatitic mineral deposition.

However, later it was discovered that MVs isolated from osteoarthritic cartilage or chondrocalcinosis, were capable of depositing Ca²⁺-PPi dihydrate (CPPD) (502). Thus the conclusion that PPi was a key regulator of cartilage mineralization became problematic. It became even more so when MVs isolated from human osteoarthritic cartilage were found to produce both CPPD and HAP in vitro (503). While the formation of CPPD crystals was closely correlated with the presence of nucleoside triphosphate pyrophosphohydrolase (NPP1 - PC1), it became evident that formation of CPPD required the presence of ATP as a substrate (503). The story became even more complex with the discovery that in progressive ankylosis (ANKH) (504-509) the gene expressed was a defective mutant PPi transporter in the chondrocyte plasma membrane. This defect prevented PPi from escaping from the cytosol to the extracellular fluid - where it would have presumably blocked unwanted soft tissue calcification. From gene knockout studies conducted since then it has become progressively clear that TNAP has an important function in hydrolysis of PPi at sites of beginning mineral formation (162, 163). Further studies have shown that co-knockout of PC1 and/or the ANKH gene and TNAP can partially correct the mineralization defects caused by the absence of each alone. While these studies are tantalizing, the findings are not all clear cut. More will be made of this later when several types of gene knockout studies are considered.

19.5. Inhibitory effects of proteoglycans on mineral formation

Early on it was evident that proteoglycans present in the extracellular matrix of growth plate cartilage exerted a powerful inhibitory effect on mineral formation (510-512). Evidence for this role came from the following observations: a) the extracellular fluid collected from the growth plate completely inhibited mineral growth when seeded with mineral isolated from calcifying sites of growth plate mineral formation; b) this inhibitory effect could be abolished by prior treatment with hyaluronidase; and c) the inhibitor reacted positively to antibodies to proteoglycan (511). This work was preceded by a systematic study of the specific interaction between cartilage protein-polysaccharides (proteoglycans) and freshly precipitating calcium phosphate mineral (513). It became evident that certain components of the proteoglycan assembly in cartilage could prevent sedimentation of the precipitating mineral. These studies indicated that proteoglycans confine and limit the spread of mineral formation. They do not bind to specific sites on the developing mineral or prevent crystal growth (e.g. like PPi or proteins like osteonectin); they encompass the mineral and present a powerful diffusion barrier. Calcified cartilage exemplifies the effect of proteoglycans; it actually has a higher percentage mineral content than true bone (514-515) - yet half of the proteoglycan is still present in its matrix. Clearly, significant removal of proteoglycans does occur during cartilage calcification. It is largely accomplished by the presence of matrix metalloproteases (516-520) and makes room for apatitic mineral growth. However, extensive degradation of the proteoglycans is not required (521). Calcified cartilage stains strongly with alcian blue, indicating that much of the chondroitin sulfate is still present after calcification is completed.

19.6. Stimulatory effects of collagens on MV mineralization

Early transmission electron microscopy studies showed that the mineral crystallites of bone were in close registration with the periodicity of type I collagen (522), which has been repeatedly confirmed (523). Since it was evident that these two components were intimately related in bone, this supported the earlier hypothesis that type I collagen was the nucleator of HAP during bone formation (524). This compelling hypothesis held sway for at least a decade until 1967 when the first reports were made of MVs being the sites of de novo apatite mineralization during growth plate development (1, 3). Subsequent studies revealing that MVs bind tightly to type II and X collagens present in growth plate cartilage (95) enhanced earlier evidence that the C-propeptide of type II collagen was somehow involved in mineral nucleation (525). These studies were in accord with the findings that type II collagen binds to MV proteins like AnxA5 (95, 102, 402-403). Now nearly two decades later, this topic was reinvestigated using AnxA5-, ACP- and the PS-containing complexes as cofactors in the mineralization process (116). From these studies it became evident that neither type II nor type X collagen accelerated the onset of mineral formation. In fact, type X collagen tended to slow down the induction of mineral formation. It became apparent that while the cartilage collagens had no ability to accelerate the induction of mineral formation, they clearly facilitated mineral growth once it had been induced by the PS-Ca²⁺-Pi complexes (116). The effects were dosage-dependent; higher levels of type II collagen enhanced mineral formation in a progressive manner. And, although type X collagen delayed the onset of mineral formation, once it had been initiated, the duration of mineral formation was extended so that in the end it clearly enhanced the amount of total mineral formed. It was further observed that the presence of the non-helical telopeptides were indeed important to the stimulatory effects of type II collagen; pepsin treatment to removal these telopeptides significantly reduced the ability of type II collagen to stimulate mineral formation (116). These observations confirm the long held view of Poole et al. (525-528) that the C-propeptide of type II collagen is important to cartilage calcification. Since type I collagen is not significantly present in growth plate cartilage, in the above studies it was not investigated. However, in mineral deposition at other sites where type I collagen is a dominant structural protein, it recently has been found that binding of MV annexins to type I collagen enhanced mineral deposition (144). In this study on pathological calcification in vascular smooth muscle cells, it was evident that calcification was mediated by annexins A2 and A6 and was dependent on type I, not type II collagen (144). Thus, it is now apparent that in different circumstances interaction of different annexins with specific collagen types appears to be required. In this light, it would be of great interest to determine whether annexin A6 is involved in MV-mediated mineralization of turkey leg tendons (529), as well as in osteoblastic mineral deposition where, again, the predominant collagen is type I.

19.7. Effects of non-collagenous bone-related proteins

Many other non-collagenous proteins have been isolated from bone and have been considered to be of potential importance in regulation of HAP formation during bone formation and/or preventing unwanted soft tissue calcifications. While most of these bone-related proteins have not been directly tested for their effects on MV mineralization, most have been studied with regard to regulation of the general mineralization process (530). Since the focus of this review is specifically on MV mineralization, these proteins will not be discussed here; however, because they have obvious importance to overall control of mineralization, they will be mentioned. These proteins include osteonectin (531), osteopontin (532), bone sialoprotein (533-534), chondrocalcin (528, 535-536), and two vitamin K-dependent gamma carboxyglutamate (Gla)-containing proteins - bone Gla protein (BGP) (537-538) and matrix Gla protein (MGP) (539). Most of these proteins appear to act primarily as inhibitors of HAP crystal growth, although bone sialoprotein appears to have the ability to stimulate mineral formation (540). The most recent data indicate that phosphorylation of Ser-136 of BSP is critical for its nucleational properties (541).

20. FURTHER EVALUATION OF THE ROLES OF KEY MATRIX VESICLE PROTEINS

Several MV proteins and enzymes that appear to have important roles in MV mineralization have been subjected to gene knockout in the mouse. Although each was described earlier, their importance to the process of MV mineralization warrants further discussion in the light of these recent gene deletion studies. The proteins in mind are TNAP, AnxA5, PS-synthase, and PHOSPHO-1. Further, because of its newly appreciated role in conjunction with PHOSPHO-1 and phospholipase C; the remarkable stimulation by phospholipase A₂ in liposomal models of MV mineralization also merits further consideration.

20.1. Tissue nonspecific alkaline phosphatase (TNAP)

TNAP has received a great deal of attention over the past ~40 years, particularly since the discovery of its abundance in MVs (53). It has been the subject of numerous reviews, the most recent being a comprehensive and thorough review published in 2010 by Michael Whyte (542). Although the importance of TNAP to skeletal mineralization has been confirmed by gene knockout studies in the mouse, several unexpected issues have surfaced that warrant further discussion.

20.1.1. Role of TNAP in PPi hydrolysis

There is now almost universal acceptance that a major role of TNAP in bone is in the hydrolysis of PPi at sites of mineral formation. This ability of TNAP is important because PPi is such a potent inhibitor of apatitic crystal growth. In pure TNAP knockout in mice there is typically severe osteomalacia; simple reduction of systemic PPi by transgenic over-expression of TNAP using ApoE-Tnap completely prevents development of skeletal abnormalities (543). In fact, the concept that TNAP causes local hydrolysis of PPi is supported by studies going back at least 35 years. Even from our group it was shown that isolated pure TNAP had the ability to hydrolyze PPi at physiological pH (54, 177, 544) and that in the growth plate where TNAP is highly expressed, the ratio of PPi/Pi declines precipitously at the site of mineralization (Figure 32) (497). Now, however, with the ability to co-delete genes for TNAP and related PPi-generating or -transporting proteins in mice (165, 543), the key role of TNAP in skeletal mineralization is well established. Yet, although hypophosphatasia often displays increased excretion of PPi accompanying skeletal osteomalacia (542), recent findings by Yadav et al. (545) reveal that in PHOSPHO1 deficiency, reduction of circulating PPi does not overcome the mineralization deficit. Similarly, while co-deletion of NPP1 or ANK with TNAP largely prevented the vertebral osteomalacia induced by knockout of TNAP (182, 183, 546), mineralization of the appendicular skeleton was not fully normalized. Sustained osteomalacia of the long bones persisted, despite major improvement in the axial skeleton (162). These studies reveal that TNAP is involved in more than simple removal of PPi from mineralization sites in appendicular long bones.

20.1.2. Additional roles of TNAP

There is another paradox in hypophosphatasia. In TNAP knockout, mineralization of the developing skeleton is typically normal at birth (542). Somehow the maternal support system within the intrauterine environment enables normal mineral formation despite the fetal genetic defect. In these TNAP gene-ablated mice, despite the progressive osteomalacia, ultrastructural analysis reveals that in the growth plate, within MVs there are nevertheless normal amounts of mineral formed (289). However, the mineral fails to propagate from the protective enclave within the MVs. These findings indicate that skeletal mineralization must be a two-step process: 1) the de novo induction of mineral formation within the protective enclave of the MV lumen, and 2) the propagation of induced mineral into the extravesicular matrix - a concept expressed by Anderson in his recent review of the role of MV in osteogenesis (547). Thus, TNAP is not required for the induction step, but is essential for propagation of the apatitic mineral. Even in the absence of TNAP, normal mineral formation appears to occur within MVs. While it has been assumed that the lack of TNAP allowed persistent ambient PPi to block mineral crystal growth away from the MV nucleation site, the recent studies of Yadav et al. (545) show that this is not an adequate explanation. (From a thermodynamic perspective, once HAP crystals are formed, by a process of secondary nucleation, ordered nanoscale ion clusters are shed from the surface of these crystals and enable rapid spread of mineral formation - provided that crystal growth is not blocked by an inhibitor such as PPi.)

However, here we need to consider an even earlier enigmatic study on the die-off of TNAP during MV mineralization. Over 20 years ago we reported that during MV-induced mineralization there is a concomitant die-off of TNAP activity (76). We have repeatedly observed this phenomenon, but no other group has either confirmed or refuted it; nor has a satisfactory explanation been provided. However, now taken together with the above-stated failure of mineral to spread from the MV lumen, it is evident that TNAP must have yet another role - the extravasation of mineral from within the MV lumen. But in so doing, TNAP loses its enzymatic activity. The most obvious explanation of this phenomenon is that crystal growth from the MV lumen proceeds through and damages the structure of TNAP, leading to its denaturation. When we made this discovery, we attempted revive the activity of TNAP after mineralization had occurred by first removing the mineral using EDTA treatment, followed by restoration of the Zn²⁺ and Mg²⁺ that are known to be crucial to its enzymatic activity (76). We could only partially recover TNAP activity, indicating that mineral formation caused irreversible damage to the TNAP. Exactly how this occurs is currently unclear, but probably involves the Ca²⁺-binding (174, 175) and Pi- binding (active) sites in the enzyme. This fact points to the need to solve the 3-D structure of skeletal TNAP. With such data in hand, quantum mechanical based molecular modeling should enable the elucidation of how TNAP brings about the extravasation of the mineral crystals from the vesicle lumen.

20.1.3. Improperly assigned roles of TNAP

There is yet one further issue that needs to be discussed here - one where previous and current reports are frequently in error - namely that TNAP is needed to hydrolyze organic phosphate substrates to provide Pi for induction of mineral formation. This error stems from an ignorance of the levels of phosphorylated substrates available in the extracellular fluid where mineralization occurs. Several studies that have addressed this issue and reveal that the levels of organic phosphate esters in the extracellular fluid are in the micromolar (not millimolar) range (302, 437, 439, 442, 480). Despite this, many investigators continue to use 3-10 mM (grossly excessive amounts) of beta-glycerophosphate (a nonphysiological substrate) to drive mineralization, oblivious of the fact that they are using an abnormal substrate in unwarranted levels when compared to those of extracellular biological substrates. Under these artificial conditions, what is interpreted as induction of mineralization is simply a manifestation of TNAP activity, which releases sufficient levels of Pi that anomalous mineral deposition is induced. These levels of organic phosphate substrates do not occur normally in vivo in the extracellular fluid. Thus, what is thought to be a demonstration of normal induction of mineral formation is in reality a misleading experimental artifact.

20.2. Annexin A5 (AnxA5)

The acidic phospholipid-dependent Ca²⁺-binding proteins - the annexins - are predominant proteins in MVs. The most physically abundant protein, AnxA5 (60, 94), is present in larger amounts than TNAP or any other MV protein (78). In fact, AnxA5 is the principal protein present in the nucleational core of native MVs (457). As has been demonstrated in this review, it has the remarkable ability to accelerate the onset and rate of mineral formation induced by the nucleational core. Thus, one would logically conclude that AnxA5 must play a major role in MV mineralization. However, there is now major confusion with regard to its physiological role; some regard AnxA5 as being irrelevant - having no biological function at all.

20.2.1. The questionable role of AnxA5 in Ca²⁺ Entrance into MVs

It has been widely assumed that AnxA5 serves as the Ca²⁺ channel in growth plate chondrocytes and MVs (94, 103, 145, 478). This was based on the presence of a hydrophilic channel in the 3-D structure of AnxA5 (and most other annexins) that is presumed to function as a Ca²⁺ channel (121). In fact, AnxA5 has been reported to enable Ca²⁺ acquisition by isolated MVs (330, 478, 548). However, questions are now arising as to whether the principal role of AnxA5 is in fact as a Ca²⁺ channel. First, AnxA5 is the principal protein in the nucleational core of native MVs; this core has the ability to induce mineral formation independent of the MV membrane or any Ca²⁺ channel activity (288, 443, 457). Second, incorporation of AnxA5 into synthetic nucleational complexes (modeled after those present in native MV) markedly accelerates the rate of mineral formation (Fig. 28) (116, 131, 136). Thus, while AnxA5 clearly enhances mineral formation by MVs, it is doubtful that it does so by acting as a Ca²⁺ ion channel. Third, although the studies by Arispe et al. showed multiconductance Ca²⁺ channel activity by pure AnxA5 in supported PS-containing phospholipid bilayers (125), there is a problem relating these findings to in vivo conditions. This is because the reported Ca²⁺ conductance studies involved use of nonphysiological conditions such as 100 mM CsCl and 70 mM CaCl₂ to demonstrate ion channel activity. Further, to cause inhibition of this channel by Zn²⁺ required 2 mM Zn²⁺- 400-fold higher than what is needed to inhibit MV Ca²⁺ uptake. Fourth, more recent studies using large unilamellar liposomes as models of MV mineralization, revealed that encapsulation of AnxA5 within the liposomes, or addition of AnxA5 to the mineralizing fluid, failed to support significant Ca²⁺ acquisition (318). In contrast, ionophores like A12387 enabled rapid Ca²⁺ acquisition by the liposomes and induced definite mineral formation (Fig. 29). In addition, unlike L-type Ca²⁺ channels that have subunits containing multiple transmembrane domains that enable selective transport of Ca²⁺ (321, 324), there are no membrane-spanning domains in the 3-D structure of the annexins (123, 124). Finally, fifth, there is quantitatively far more annexin present in MVs than should be required to function as Ca²⁺ channels - if they were active carriers (60, 61, 78, 485). Proteomic analysis shows that the annexins are in far greater abundance than any of the numerous other transporters detected (78). This suggests that if they function as Ca²⁺ transporters, the annexins must be highly inefficient in this role.

20.2.2. Activation of the nucleational core by AnxA5

In contrast, as was described earlier, the most remarkable effect of AnxA5 is its acceleration and enhancement of mineral formation by the reconstituted nucleational core of MVs (Fig. 28) (116, 136, 332). Its ability to overcome the inhibitory effects of the ubiquitous presence of Mg²⁺ and the non-Ca²⁺-binding MV lipids is nothing less than remarkable (131, 137). Since MVs are able to trigger mineralization in the hostile environment of the proteoglycan-rich extracellular matrix, and contain such abundant amounts of AnxA5, it not surprising that AnxA5 accelerates and enhances rapid mineral formation. In fact, all studies so far performed with respect to the nucleational core of MVs strongly indicate that it has a key role for stimulating mineral formation. As noted previously in the analysis of MV formation, there is also a straight-forward explanation of why AnxA5 becomes incorporated during MV formation, given its Ca²⁺-dependent PS-binding ability.

20.2.3. Does AnxA5 in fact have a physiological function in vivo?

Because of its obvious stimulatory effects on in vitro mineral formation - as well as its apparently key role in apoptotic processes - it was widely assumed that deletion of the AnxA5 gene would be lethal. Hence, it was a shock to read the dogmatic assertion that AnxA5 "is not essential for skeletal development" by Brachvogel et al. in 2003 (549). Even seven years after that original report, and the recent report that double knockout of AnxA5 and AnxA6 had no discernible effect on the mice (550), these findings still seem implausible. If one is to believe the implications of these reports, both AnxA5 and AnxA6 are irrelevant not only to skeletal mineralization, but to any other physiological function. This is implausible not only because of the their abundant expression in growth plate chondrocytes, but because of their abundance in cells of all phyla so far examined. There have been hundreds of reports indicating biological functions for these and other annexins; many indicate involvement in skeletal formation. Are all these findings invalid? They document activities of proteins whose genes have been expressed and conserved over millions of years of evolution. To date, neither of these annexin gene knockout findings has been confirmed by any other group. However, assuming that they are valid, what rational explanations can be postulated to explain the reported lack of function?

20.2.4. Possible alternative routes to growth plate and skeletal calcification

Knowing of the absolute essentiality of skeletal mineralization for survival, it is plausible to speculate that backup mechanisms for induction of mineral formation exist and can be called into play if the normal pathway fails. Taking this view, it is plausible that the annexin knockout mice induced mineralization by a MV-independent mechanism. Both annexin knockout reports lacked ultrastructural detail in the histology of the mineralization site; this was particularly true of the initial report (549). While the second report (550) was more comprehensive, it still provided inadequate ultrastructural analysis of how the KO mice induced mineralization. In both studies the authors assumed the presence of of MVs, but provided no clear evidence that they were present at sites of initial mineral formation. Thus, an alternative backup mechanism may explain why apparently normal mineralization occurred in these mice.

20.2.5. Effects of Knockout of anxa5 and anxa6 on expression of other genes

In their most recent report, the Brachvogel group reported that despite having no discernible physiological effect on the mice, co-deletion of both AnxA5 and AnxA6 caused marked up-regulation of three poorly characterized genes: Depdc6 (+26-fold), Ptdsr (+18-fold) and Pold3 (+14-fold), as well as marked down-regulation of three other little understood genes: BC045135 (-33-fold), NAP115274-1 (-30-fold) and Unc13a (-11-fold) (550). What is the significance of these findings? The authors provide no clue. The best studied of the upregulated genes is Ptdsr, a putative receptor for phosphatidylserine (PS) (551-556). Did the PS receptor (Ptdsr) substitute for AnxA5? Recall that PS is the Ca²⁺-binding acidic phospholipid previously implicated in binding with AnxA5 in the nucleational core of MVs (64, 457). While it would have been natural to assume that Ptdsr protein would bind PS and thereby provide a substitute for AnxA5, this turns out not to be the case (557-558). While the function of Ptdsr in apoptosis has been studied, the reports have been inconsistent (557, 559-560). It has now been concluded that Ptdsr is a misnomer and that the gene does not produce a receptor for PS. Thus, the putative upregulation of the Ptdsr gene in AnxA5 and AnxA5 knockout mice seems irrelevant. There are currently no clues as to why the expression of the other genes was altered.

20.2.6. What would have been the effects of expression of defective annexin mutants?

Another possibility is that pathology may become evident only if a defective mutant form of AnxA5 is expressed. The Belluoccio report, although comprehensive, did not rigorously exclude the possibility that some other member of this large gene family may have filled in for the two missing members. This story may be akin to what is observed in type X collagen gene deletion (561-565). Deletion of type X collagen caused no apparent growth plate or skeletal pathology (561); however, expression of various defective forms of type X collagen presented with significant skeletal pathology (566-570). Their conclusion was that only in the presence of an abnormal gene product was pathology observed. This may well be the case for the AnxA5 and AnxA6 gene knockout mice. This phenomenon can be envisioned using an analogy of workers in a factory assembly line. If a worker reports that he cannot work, management can bring in a replacement and maintain production. However, if the worker arrives ill, takes his place and is unable do the work, his job will not be performed and production will be blocked. Therefore, would it not be more instructive to over-express defective forms of AnxA5 - e.g. those with mutated PS-dependent Ca²⁺-binding sites or perhaps altered "Ca²⁺ channel" structure - to determine their effects?

20.2.7. Is the need for AnxA5 stress-dependent?

Still another reason for the lack of effect of AnxA5 knockout may be because these highly inbred "wildtype" mice are maintained in a highly protective environment with all of their nutritional and other physical requirements met. If they were exposed to stresses like those of the ancestral wild mouse that affect skeletal development (e.g. Ca²⁺ or Pi deficiency, or some developmental stress) then differences between the wildtype and AnxA5-deleted mice may well become evident. A plausible explanation is that skeletal growth in a small rodent does not require the stimulus of AnxA5. The stress of needing very rapid bone growth in much larger, fast growing species like the broiler-strain chicken or long legged mammals like humans may make the requirement for AnxA5 critical. In fact, the principal effect of AnxA5 on mineral nucleation is to markedly accelerate the onset of mineral formation (Figure 28), (136, 332). Thus in a small animal like the mouse, where demands for maximal mineralization are unnecessary, AnxA5 catalytic stimulus may be superfluous. However, this still does not explain why all other bodily processes were unaffected. It seems implausible that the annexins should have been conserved over millions of years of evolution in many genera if they served no useful purpose.

20.3. Phosphatidylserine (PS) synthases

Perhaps the most unusual enzymes involved in MV mineralization are the PS synthases. These enzymee are required for the formation of PS, the key Ca²⁺-binding acidic phospholipid intimately involved in formation of PS-CPLX.

20.3.1. PS synthases do not require high-energy nucleotides, but requires Ca²⁺

In contrast to all other biosynthetic enzymes in lipid synthesis, these enzymes do not require high-energy nucleotides to drive PS formation; however the reactions are Ca²⁺-dependent (571). This fact becomes important in the growth plate at the time of chondrocyte hypertrophy because levels of ATP and other nucleoside triphosphates are very low (442). PS synthases are enzymes that catalyze the Ca²⁺-dependent exchange of the amino acid, serine, for the base present in the polar head group of either of two neutral membrane phospholipids, phosphatidylethanolamine (PE) or phosphatidylcholine (PC). As noted previously, there are two types of PS synthases (PSS) - PSS1 which exchanges serine for choline in PC, and PSS2 which exchanges serine for ethanolamine in PE (269). PSS2 activity is evident in growth plate chondrocytes (267) and MVs (268); it must be involved in the increase in PS that occurs during MV mineralization (207). PS is required for synthesis of the PS-CPLXs involved in MV mineral nucleation. Early studies showed that synthesis of PS can occur in isolated MVs, and can be prevented by treatment with EDTA. These findings document the presence of PSS activity in native MVs (268).

20.3.2. PSS1 and PSS2 gene knockouts

Of further interest, recent studies have shown that PSS1-deficient or PSS2-deficient mice are viable, fertile, and have a normal life span; however, crosses of Pss1^-/- and Pss2^-/- mice produced no live offspring; this revealed that double knockout mice are not viable (572). However in Pss1^-/-/Pss2^+/- and Pss1^+/-/Pss2^-/- mice, where the ability to incorporate serine into PS was reduced by 65-91% with concomitant reduction in both PS and PE levels, the mice were still viable. These findings have bearing on the lack of effects of double knockout of AnxA5 and AnxA6, indicating that the activities of these annexins are not co-dependent. While any effects of PSS1/PSS2 double knockout on skeletal formation were not considered, these studies nevertheless inject a note of caution in the interpretation of negative results in gene knockout studies in mice.

20.4. Phospholipase A and C activities

The remarkable stimulatory effect on induction of mineral formation by simple addition of phospholipase A₂ to reconstituted liposomal models of MVs (Figure 29) (318), as well as the rapid degradation of MV phospholipids during induction of mineral formation (207) point to a heretofore little appreciated consequence of the lipid degradative activity evident in growth plate cartilage (223), in MV formation (73) and in MV mineralization (207).

20.4.1. Membrane lysis and Ca²⁺ access to the nucleational core

Because it was logical to assume that entrance of Ca²⁺ into MVs was mediated by the presence of AnxA5, the lack of ability of this protein to enable Ca²⁺ entrance into synthetic unilamellar liposomal models of MVs was unexpected (318). However in retrospect, it was unlikely that such an abundant protein in MVs as AnxA5 should be required for Ca²⁺ entry. On the other hand, the extraordinary stimulatory effect that simple addition of phospholipase A₂ to reconstituted liposomal models of MVs has on induction of mineral formation (Figure 29) (318) should not have been so surprising. It is apparent that synergistic activation of latent intravesicular phospholipase activity by influx of Ca²⁺, such as was evident in the rapid breakdown of various phospholipids during MV mineralization, (207), would provide an efficient means for accessing Ca²⁺ and Pi to the nucleational core - once thermodynamically stable OCP crystals were established within the MV lumen.

20.4.2. Role in the egress of mineral from the vesicle lumen

It is highly probable that phospholipase activity is also an important factor in the egress of mineral from within the vesicle lumen; such activity would cause lysis of the membrane barrier. However, nothing is simple with MVs, because as was noted earlier, TNAP associated with lipid rafts in the MV membrane also appears to be involved (76); and it is also evident that Ca²⁺-activated phospholipid scramblase activity (287) also may get into the act.

20.4.3. Salvage of phospholipid P for mineral formation

But coupled with the breakdown of membrane phospholipids is another little appreciated phenomenon - the catabolism of phospholipids also enables salvage of Pi for MV mineral formation. Phosphate, stored in the linkage of the polar head group of phospholipids of all membranes, including MVs, is made available for enhancing mineral formation by the release of phosphocholine and phosphoethanolamine through the action of phospholipase C on PC and PE, respectively (207). The mechanism by which this occurs will now be discussed.

20.5. PHOSPHO-1

PHOSPHO-1 is another important enzyme present in MVs (202-204, 573). It has the remarkable ability to coordinate breakdown of the MV membrane phospholipids with formation of mineral.

20.5.1. Salvage of Pi and choline⁺ from phospholipid breakdown

PHOSPHO1 has the ability to hydrolyze phosphoethanolamine and phosphocholine, byproducts of phospholipase C action on PC and PE, to release Pi - as well as the amine bases, choline⁺ and ethanolamine. What makes the action of PHOSPHO1 especially significant is the finding that Pi transport into hypertrophic chondrocytes and MVs is not strictly Na⁺-dependent; in fact Pi uptake was more active with choline⁺ than with Na⁺ (208, 303). Thus, the hydrolytic products of phosphocholine by PHOSPHO1 not only release Pi needed form mineral formation, but choline⁺ that can enhance Pi uptake by MVs. These are physiologically important functions.

20.5.2. Inhibition of PHOSPHO1 activity

Very recently it has been reported that inhibition of PHOSPHO1 activity resulted in impaired skeletal mineralization during limb development of the chick (205). This plausible conclusion was based on the finding that the inhibitor, Lansoprazole, markedly reduced mineralization of the long bones of the legs and wings. Unfortunately, Lansoprazole is by no means specific for PHOSPHO1 and is in fact a powerful inhibitor of H⁺/K⁺-ATPase (574-576). Thus, it is equally probable that the drug blocked H⁺-transport in the growth plate chondrocytes and interfered with MV Ca²⁺ loading (306). As was discussed previously, H⁺ metabolism is notable different in successive zones of the growth plate and appears to be vital for proper MV formation. Thus, although the findings that Lansoprazole inhibits mineralization are tantalizing, they are open to question.

20.5.3. PHOSPHO1 gene knockout

However, very recent findings reveal that knockout of PHOSPHO1 gene causes significant skeletal pathology (545). Deletion of the PHOSPHO1 gene produced numerous physiological effects: generalized mild osteomalacia accompanied by bowing of the long bones with evidence of frequent fractures, as well as growth retardation, and reduced food and water consumption. The specific effects on bone mineralization were not enormous, but led to obvious weakening of the skeletal structures - as evidence by the progressive scoliosis of the spine and the numerous fracture callouses observed. While there were minimal effects on most biochemical parameters in the blood serum, there was a transient increase in Pi levels - perhaps due to reduced mineral formation. One curious side effect was a slight reduction in blood plasma TNAP levels, as well as small increases in blood levels of NPP1 and PPi.

20.5.4. Co-modulation of TNAP and PHOSPHO1 gene expression

While over-expression of TNAP superimposed on the PHOSPHO1 knockout corrected the PPi level in blood, it had no beneficial effect on the osteomalacia. This reveals that the effects of PPi on mineral formation are not simple. From this study it is clear that TNAP and PHOSPHO1, although both acting as phosphoesterases, operate independently. However, superimposing haploinsufficiency of TNAP (Akp2^+/-) with PHOSPHO1 (Phospho1^-/-) knockout caused a progressive worsening of the skeletal defects. A further finding was that there were differences in the skeletal defects observed with individual Akp2^-/-and Phospho1^-/- knockout mice: there was a more severe reduction in vertebral mineralization in Phospho1^-/- than in Akp2^-/-knockouts.

20.5.5. Effect of PHOSPHO1 on MV mineralization

With regard to its effect on MV mineralization, PHOSPHO1 knockout did not block it; in fact ultrastructural studies seem to indicate that, if anything, the mineral formed within the MV was more intense. However, like TNAP knockout, the mineral did not spread from the MVs. In a further study where both PHOSPHO1 and TNAP were concomitantly ablated, there was an almost total block in skeletal mineralization; none of the 272 conceptions continued to live birth. Analysis of the embryonic Phospho1^-/-/ Akp2^-/- pups showed an almost total lack of skeletal mineralization; there was no evidence of outgrowth of mineral from within the MVs. This effect cannot be due to an excess of PPi. The data clearly indicate that the activities of both TNAP and PHOSPHO1 are concomitantly required for skeletal mineralization in the mouse. However, more detailed biochemical study will be required to definitively establish how the combined activities of these two key enzymes contributes to function of MVs in this complex process.

21. PERSPECTIVES

The processes involved in growth plate mineralization depicted in this review have proven to be remarkably complex and the mechanisms involved have generated frequent controversy. Nevertheless, it is now generally accepted that the entities most responsible for the de novo induction of crystalline mineral in most vertebrate hard tissues are MVs. While it is universally recognized that the resident cells are involved, there has been a longstanding question whether they are directly involved in the processing of Ca²⁺ and Pi prior to the production of MVs. Data presented in this review should put that question to rest - at least in avian skeletal development. They document that during growth plate development the chondrocytes initially require Pi for synthesis of DNA, RNA and nucleotides essential for proliferation. Upon termination of cell division, the nucleotides begin to be degraded; ³¹P-NMR and chemical analyses reveal that Pi builds up in the cytoplasm. Simultaneously, in situ confocal imaging reveals that the chondrocytes acquire and process Ca²⁺ at an accelerated pace until in the early hypertrophic zone this climaxes with the cells exfoliating from their plasma membrane the accumulated Ca²⁺ and Pi as ion-loaded vesicles. This mineral phase, present as a nucleational core within the vesicle lumen, is primarily noncrystalline calcium phosphates complexed with both PS and the annexins, which facilitate the induction of crystalline mineral formation. When MVs are formed by the cells they contain not only Ca²⁺ and Pi, but also other electrolytes that significantly affect the kinetics of mineral formation; these include Mg²⁺, Zn²⁺, and PPi, as well as HCO₃^-. How these ions influence mineral formation are discussed, as is the nature of the successive calcium phosphate minerals that form during the mineral induction process. Also discussed are the roles of the plethora of enzymes now known to reside in MVs; however the major focus is on about a half a dozen that have particular importance in mineral induction. What emerges is a remarkably complex and beautifully orchestrated series of cellular events that enable the rapid elongation of bones while maintaining their physical integrity. Production of mineralization-competent MVs enables mineral formation to keep pace with cellular proliferation, preventing development of rickets or osteomalacia. Thus, endochondral bone formatioin involves the provisional calcification of the cartilaginous matrix by chondrocytes - i.e. the formation of calcified cartilage - a more heavily calcified, but mechanically weaker temporary scaffolding that maintains mechanical stability until the truly robust type I collagen-hydroxyapatite composite, bone, is generated by osteoblasts. One of the noteworthy features of this cell-mediated process is that it involves features that are observed pathologically in the ischemia-reperfusion of stroke and cardiac vascular blockage, but are utilized here to induce normal mineral formation in a highly regulated manner during rapid bone growth. This review covers a long period of discovery and involves nearly 50 years of personal participation in elucidation of this remarkable process. It has been a special privilege to have been involved in many of the experimental discoveries that now provide the landscape of understanding. It is important however for workers currently in the bone field to master the fundamentals of the physical chemistry of the mineral forming process, as well of the cellular physiology, biology and biochemistry that are all essential for proper understanding of this highly complex process. Biological mineralization not only involves features characteristic of all living soft tissues cells, but superimposes on these the need to transform ions in solution into the robust inorganic mineral-collagen composite material (bone) that provides the framework in which our bodies function.

22. ACKNOWLEDGEMENTS

I want to acknowledge first and foremost my incalculable debt to my Lord and Savior, Jesus Christ, for his mercy and grace, and to the Holy Spirit, who has guided and enabled me to have faith in God, our Creator. I thank Him for granting me the privilege of studying and helping to elucidate this remarkable area of vertebrate anatomy, namely the growth plate, and the mechanism by which the cells in this tissue bring about, concomitantly, longitudinal growth and deposition of mineral to maintain mechanical strength as these structures are being formed. It was He who guided me in 1955 to graduate school at the University of Wisconsin, where under the tutelage of my major professor, Paul H. Phillips, Ph.D., I was introduced to hard tissues. Prof. Phillips directed me to study the long-term effects of low dosages of fluoride on both rats and dogs. He also encouraged me to take classes in Histology and Physiology in the School of Medicine in Madison, to be completed with courses in Physical Chemistry and Organic Chemistry laboratory, as part of my basic training. The broad perspective so obtained has been an invaluable asset in approaching the analysis of the complex physico-chemical processes that occur during bone development and mineralization. Next, I want to express my gratitude for the patience, kindness and wisdom of Prof. James T. Irving, M.D., my postdoctoral mentor at Forsyth Dental Center, who in 1960 revealed to me his remarkable finding of intense Sudanophilia at sites of mineral deposition in the extracellular matrix of hypertrophic chondrocytes of the growth plate in rats. He asked me to determine what this material was and what it was doing there. This was the beginning of a quest that has taken 50 years of study - through the efforts of a small army of students and colleagues to solve - at least to the level of understanding depicted in this review.

That these studies have involved a host of participants is evident from the 583 citations in the reference section. It was through the efforts of this outstanding group of investigators in many institutions and disciplines, who attacked this problem from many different angles that our understanding is what it is today. However, I want to acknowledge the specific contributions of members of my own group - spanning a period from 1960 to the present. Among the earliest was John M. Cotmore, Ph.D., who in 1968 discovered the potentiating effect of inorganic phosphate (Pi) on the binding of Ca²⁺ to phosphatidylserine (PS), a key lipid we had earlier discovered was bound with newly forming bone mineral. The PS-Ca²⁺-Pi complexes so formed can now be assigned as key components of the nucleational core of MVs. Next was the critical 1970 finding by Eugene Eisenberg, M.D. - from in vivo studies on the incorporation of ³²Pi into bone lipids of rapidly growing chickens - that the labeling pattern of Ca²⁺-complexed lipids was distinctly different from that of non-complexed lipids, proving that the association between PS and Ca²⁺ was biological, not an adventitious artifact. Next, the outstanding contributions of two of my esteemed technical assistants, Paul H. King, in Boston, MA, and Michael S. Giancola, in Burlington, VT must be recognized - Paul for his invaluable assistance in the early lipid chromatography, and Mike for his skill in performing the double-isotope labeling studies that enabled accurate determination of the electrolyte levels within the cells and MVs, as well as those in the extracellular fluid (ECF) from different zones of the growth plate. Added to this list are the many graduate students whose specific contributions cannot be mentioned individually here, but are cited in the list of references. Next, I must recognize four of my long-term post-doctoral colleagues: Yoshinori Ishikawa (1978-1998), Licia N.Y. Wu (1987-present), Glenn R. Sauer (1987-2000) and Brian R. Genge (1988-present), who each have been major contributors to this story. Yoshi analyzed the levels of free amino acids in the ECF, documenting their stimulatory effect on MV formation, and with Licia, developed primary cultures of growth plate chondrocytes capable of inducing mineralization under physiological conditions. Glenn used FT-IR characterization of the mineral phase and subsequent study of the role of Zn²⁺ in MVs to elucidate key steps in MV function. Licia and Brian, working as a team together, developed definitive techniques for isolation, dissection, characterization, and reconstitution of MV; their subsequent studies on the kinetics of mineral formation, the role of the annexins, the nucleational core, and regulatory effects of Mg²⁺ defined critical aspects of mineral formation by MVs. In particular, their skill, talents, and meticulous attention to details - and their long, hard hours of work - are what made much of this report possible. Finally, I give special credit and deep appreciation to my dear wife, Shirley Jo Westerfield Wuthier, who for 54 years has been my most loving companion and most insightful critic - one who has enabled me to devote the years of labor described in this fascinating, but complex report. She, more than any other person has encouraged me to invest the effort required to pull together this epic report and had the patience to allow me to do it.

Last, I gratefully acknowledge the 30 years of continuous support of National Institutes of Health, NIH Grant AR18983, for research on the Role of Matrix Vesicles in skeletal mineralization.

23. REFERENCES

1. Anderson, H. C.: Electron microscopic sudies of induced cartilage development and calcification. J Cell Biol, 35, 81-101 (1967)
doi:10.1083/jcb.35.1.81
PMid:6061727    PMCid:2107116

2. Anderson, H. C.: Vesicles associated with calcification in the matrix of epiphyseal cartilage. J Cell Biol, 41, 59-72 (1969)
doi:10.1083/jcb.41.1.59
PMid:5775794    PMCid:2107736

3. Bonucci, E.: Fine structure of early cartilage calcification. J Ultrastruct Res, 20, 33-50 (1967)
doi:10.1016/S0022-5320(67)80034-0

4. Bonucci, E.: Fine structure and histochemistry of "calcifying globules" in epiphyseal cartilage. Z. Zellforsch. Mikrosk. Anat., 103, 192-217 (1970)
doi:10.1007/BF00337312

5. Bernard, G. W. & D. C. Pease: An electron microscopic study of initial intramembranous osteogenesis. Am J Anat, 125, 274-290 (1969)
doi:10.1002/aja.1001250303
PMid:5790994

6. Ketenjian, A. Y., A. M. Jafri & C. Arsenis: Studies on the mechanism of callus cartilage differentiation and calcification during fracture healing. Orthop Clin North Am, 9, 43-65 (1978)
PMid:643267

7. Chai, B. F. & X. M. Tang: Ultrastructural investigation on experimental fracture healing. VI. Electron microscopic observation on matrix vesicles. Chin Med J (Engl), 97, 805-12 (1984)

8. Brighton, C. T. & R. M. Hunt: Histochemical localization of calcium in the fracture callus with potassium pyroantimonate. Possible role of chondrocyte mitochondrial calcium in callus calcification. J Bone Joint Surg Am, 68, 703-15 (1986)
PMid:2424916

9. Sisca, R. F. & D. V. Provenza: Initial dentin formation in human deciduous teeth. An electron microscope study. Calcif Tissue Res, 9, 1-16 (1972)
doi:10.1007/BF02061941
PMid:4336177

10. Sayegh, F. S., G. C. Solomon & R. W. Davis: Ultrastructure of intracellular mineralization in the deer's antler. Clin Orthop Relat Res, 99, 267-84 (1974)
doi:10.1097/00003086-197403000-00030
PMid:4133065

11. Kim, K. M.: Calcification of matrix vesicles in human aortic valve and aortic media. Fed Proc, 35, 156-62 (1976)
PMid:1248649

12. Sela, J. & I. A. Bab: The relationship between extracellular matrix vesicles and calcospherities in primary mineralization of neoplastic bone tissue. TEM and SEM studies on osteosarcoma. Virchows Arch A Pathol Anat Histol, 382, 1-9 (1979)
doi:10.1007/BF01102736

13. Fedde, K. N.: Human osteosarcoma cells spontaneously release matrix-vesicle-like structures with the capacity to mineralize. Bone Miner, 17, 145-51 (1992)
doi:10.1016/0169-6009(92)90726-T

14. Landis, W. J. & M. J. Song: Early mineral deposition in calcifying tendon characterized by high voltage electron microscopy and three-dimensional graphic imaging. J Struct Biol, 107, 116-27 (1991)
doi:10.1016/1047-8477(91)90015-O

15. Arsenault, A. L., B. W. Frankland & F. P. Ottensmeyer: Vectorial sequence of mineralization in the turkey leg tendon determined by electron microscopic imaging. Calcif Tissue Int, 48, 46-55 (1991)
doi:10.1007/BF02555795
PMid:2007226

16. Irving, J. T.: A histological staining method for sites of calcification in teeth and bone. Arch Oral Biol, 1, 89-96 (1959)
doi:10.1016/0003-9969(59)90001-9

17. Irving, J. T. & R. E. Wuthier: Further observations on the Sudan black stain for calcification. Arch Oral Biol, 5, 323-324 (1961)
doi:10.1016/0003-9969(61)90067-X

18. Irving, J. T.: The sudanophilic material at sites of calcification. Arch Oral Biol, 8, 735-743 (1963)
doi:10.1016/0003-9969(63)90005-0

19. Wuthier, R. E. & J. T. Irving: Lipids in developing calf bone. J. Dental Res, 43, 814-815, abstract (1964)

20. Wuthier, R. E.: Lipids of mineralizing epiphyseal tissues in the bovine fetus. J. Lipid Res, 9, 68-78 (1968)
PMid:5637432

21. Shapiro, I. M. & R. E. Wuthier: A study of the phospholipids of bovine dental tissues. II. Developing bovine foetal dental pulp. Arch Oral Biol, 11, 513-9 (1966)
doi:10.1016/0003-9969(66)90157-9

22. Shapiro, I. M., R. E. Wuthier & J. T. Irving: A study of the phospholipids of bovine dental tissues. I. Enamel matrix and dentine. Arch Oral Biol, 11, 501-12 (1966)
doi:10.1016/0003-9969(66)90156-7

23. Irving, J. T. & R. E. Wuthier: Histochemistry and biochemistry of calcification with special reference to the role of lipids. Clin Orthop Relat Res, 56, 237-260 (1968)
doi:10.1097/00003086-196801000-00026
PMid:4172398

24. Abramson, M. B., R. Katzman, C. E. Wilson & H. P. Gregor: Ionic properties of aqueous dispersions of phosphatidic acid. J Biol Chem, 239, 4066-72 (1964)
PMid:14247649

25. Abramson, M. B., R. Katzman & H. P. Gregor: Aqueous dispersions of phosphatidylserine. Ionic properties. J Biol Chem, 239, 70-6 (1964)
PMid:14114874

26. Nash, H. A. & J. M. Tobias: Phospholipids Membrane Model: Importance of phosphatidylserine and its cation exchanger nature. Proc Natl Acad Sci U S A, 51, 476-80 (1964)
doi:10.1073/pnas.51.3.476

27. Hendrickson, H. S. & J. G. Fullington: Stabilities of metal complexes of phospholipids: Ca (II), Mg (II), and Ni (II) complexes of phosphatidylserine and triphosphoinositide. Biochemistry, 4, 1599-605 (1965)
doi:10.1021/bi00884a021
PMid:4286195

28. Eisenberg, E., R. E. Wuthier, R. B. Frank, J. T. Irving: Time study of in vivo incorporation of 32P-orthophosphate into phospholipids of chicken epiphyseal tissues. Calcif Tissue Res, 6, 32-48 (1970)
doi:10.1007/BF02196182
PMid:4993508

29. Peress, N. S., H. C. Anderson & S. W. Sajdera: The lipids of matrix vesicles from bovine fetal epiphyseal cartilage. Calcif Tissue Res, 14, 275-81 (1974)
doi:10.1007/BF02060301
PMid:4367314

30. Wuthier, R. E.: Lipid composition of isolated epiphyseal cartilage cells, membranes and matrix vesicles. Biochim Biophys Acta, 409, 128-43 (1975)
PMid:1182191

31. Wuthier, R. E.: Lipids of matrix vesicles. Fed Proc, 35, 117-21 (1976)
PMid:765160

32. Morris, D. C. & J. Appleton: The effects of lanthanum on the ultrastructure of hypertrophic chondrocytes and the localization of lanthanum precipitates in condylar cartilages of rats fed on normal and rachitogenic diets. J Histochem Cytochem, 32, 239-47 (1984)
PMid:6693756

33. Gomez, S., J. M. Lopez-Cepero, G. Silvestrini, P. Mocetti & E. Bonucci: Matrix vesicles and focal proteoglycan aggregates are the nucleation sites revealed by the lanthanum incubation method: a correlated study on the hypertrophic zone of the rat epiphyseal cartilage. Calcif Tissue Int, 58, 273-82 (1996)
PMid:8661960

34. Petersheim, M. & J. Sun: On the coordination of La3+ by phosphatidylserine. Biophys J, 55, 631-6 (1989)
doi:10.1016/S0006-3495(89)82860-7

35. Bentz, J., D. Alford, J. Cohen & N. Duzgunes: La3+-induced fusion of phosphatidylserine liposomes. Close approach, intermembrane intermediates, and the electrostatic surface potential. Biophys J, 53, 593-607 (1988)
doi:10.1016/S0006-3495(88)83138-2

36. Aussel, C., R. Marhaba, C. Pelassy & J. P. Breittmayer: Submicromolar La3+ concentrations block the calcium release-activated channel, and impair CD69 and CD25 expression in CD3- or thapsigargin-activated Jurkat cells. Biochem J, 313 (Pt 3), 909-13 (1996)
PMid:8611174    PMCid:1216997

37. Thyberg, J. & U. Friberg: Ultrastructure and acid phosphatase activity of matrix vesicles and cytoplasmic dense bodies in the epiphyseal plate. J Ultrastruct Res, 33, 554-73 (1970)
doi:10.1016/S0022-5320(70)90181-4

38. Thyberg, J. & U. Friberg: Electron microscopic enzyme histochemical studies on the cellular genesis of matrix vesicles in the epiphyseal plate. J Ultrastruct Res, 41, 43-59 (1972)
doi:10.1016/S0022-5320(72)90037-8

39. Hunziker, E. B., W. Herrmann, R. K. Schenk, M. Mueller & H. Moor: Cartilage ultrastructure after high pressure freezing, freeze substitution, and low temperature embedding. I. Chondrocyte ultrastructure - implications for the theories of mineralization and vascular invasion. J Cell Biol, 98, 267-76 (1984)
doi:10.1083/jcb.98.1.267
PMid:6707090

40. Arsenault, A. L. & E. B. Hunziker: Electron microscopic analysis of mineral deposits in the calcifying epiphyseal growth plate. Calcif Tissue Int, 42, 119-26 (1988)
doi:10.1007/BF02556344
PMid:2450627

41. Hunziker, E. B., W. Herrmann, L. M. Cruz-Orive & A. L. Arsenault: Image analysis of electron micrographs relating to mineralization in calcifying cartilage: theoretical considerations. J Electron Microsc Tech, 11, 9-15 (1989)
doi:10.1002/jemt.1060110103
PMid:2915265

42. Arsenault, A. L., F. P. Ottensmeyer & I. B. Heath: An electron microscopic and spectroscopic study of murine epiphyseal cartilage: analysis of fine structure and matrix vesicles preserved by slam freezing and freeze substitution. J Ultrastruct Mol Struct Res, 98, 32-47 (1988)
doi:10.1016/S0889-1605(88)80932-7

43. Arsenault, A. L. & D. M. Kohler: Image analysis of the extracellular matrix. Microsc Res Tech, 28, 409-21 (1994)
doi:10.1002/jemt.1070280507
PMid:7919528

44. Ottensmeyer, F. P. & J. W. Andrew: High-resolution microanalysis of biological specimens by electron energy loss spectroscopy and by electron spectroscopic imaging. J Ultrastruct Res, 72, 336-48 (1980)
doi:10.1016/S0022-5320(80)90069-6

45. Morris, D. C., H. K. Vaananen & H. C. Anderson: Matrix vesicle calcification in rat epiphyseal growth plate cartilage prepared anhydrously for electron microscopy. Metab Bone Dis Relat Res, 5, 131-7 (1983)
doi:10.1016/0221-8747(83)90014-0

46. Akisaka, T., H. Kawaguchi, G. P. Subita, Y. Shigenaga & C. V. Gay: Ultrastructure of matrix vesicles in chick growth plate as revealed by quick freezing and freeze substitution. Calcif Tissue Int, 42, 383-93 (1988)
doi:10.1007/BF02556357
PMid:3135108

47. Cecil, R. N. A. & H. C. Anderson: Freeze-fracture studies of matrix vesicles in epiphyseal growth plate. Metab Bone Dis Relat Res, 1, 89-95 (1978)
doi:10.1016/0221-8747(78)90043-7

48. Borg, T. K., R. B. Runyan & R. E. Wuthier: Correlation of freeze-fracture and scanning electron microscopy of epiphyseal chondrocytes. Calcif Tissue Res, 26, 237-41 (1978)
doi:10.1007/BF02013264
PMid:750066

49. Takagi, M., Y. Kasahara, H. Takagi & Y. Toda: Freeze-fracture images of matrix vesicles of epiphyseal cartilage and non-calcified tracheal cartilage. J Electron Microsc (Tokyo), 28, 165-75 (1979)

50. Borg, T. K., R. B. Runyan & R. E. Wuthier: A freeze-fracture study of avian epiphyseal cartilage differentiation. Anat Rec, 199, 449-57 (1981)
doi:10.1002/ar.1091990402
PMid:7270908

51. Katchburian, E. & N. J. Severs: Matrix constituents of early developing bone examined by freeze fracture. Cell Biol Int Rep, 7, 1063-70 (1983)
doi:10.1016/0309-1651(83)90012-7

52. Ali, S. Y., S. W. Sajdera & H. C. Anderson: Isolation and characterization of calcifying matrix vesicles from epiphyseal cartilage. Proc Natl Acad Sci U S A, 67, 1513-20 (1970)
doi:10.1073/pnas.67.3.1513

53. Ali, S. Y., H. C. Anderson & S. W. Sajdera: Enzymic and electron-microscopic analysis of extracellular matrix vesicles associated with calcification in cartilage. Biochem J, 122, 56P (1971)
PMid:4256901    PMCid:1176871

54. Majeska, R. J. & R. E. Wuthier: Studies on matrix vesicles isolated from chick epiphyseal cartilage. Association of pyrophosphatase and ATPase activities with alkaline phosphatase. Biochim Biophys Acta, 391, 51-60 (1975)
PMid:237558

55. Wuthier, R. E., R. E. Linder, G. P. Warner, S. T. Gore & T. K. Borg: Nonenzymatic method for isolation of matrix vesicles: Characterization and initial studies on Ca and P orthophosphate metabolism. Metab Bone Dis Relat Res, 1, 125-136 (1978)
doi:10.1016/0221-8747(78)90049-8

56. Watkins, E. L., J. V. Stillo & R. E. Wuthier: Subcellular fractionation of epiphyseal cartilage: Isolation of matrix vesicles and profiles of enzymes, phospholipids, calcium and phosphate. Biochim Biophys Acta, 631, 289-304 (1980)
PMid:7407248

57. Warner, G. P., H. L. Hubbard, G. C. Lloyd & R. E. Wuthier: Pi and Ca metabolism by matrix vesicles prepared from chicken epiphyseal cartilage microsomes by isosmotic Percoll density-gradient fractionation. Calcif Tissue Int, 35, 327-338 (1983)
doi:10.1007/BF02405054
PMid:6871763

58. Wuthier, R. E., J. E. Chin, J. E. Hale, T. C. Register, L. V. Hale & Y. Ishikawa: Isolation and characterization of calcium-accumulating matrix vesicles from chondrocytes of chicken epiphyseal growth plate cartilage in primary culture. J Biol Chem, 260, 15972-9 (1985)
PMid:3905800

59. Wu, L. N., W. B. Valhmu, G. C. Lloyd, B. R. Genge & R. E. Wuthier: Isolation of two glycosylated forms of membrane-bound alkaline phosphatase from avian growth plate cartilage matrix vesicle-enriched microsomes. Bone Miner, 7, 113-25 (1989)
doi:10.1016/0169-6009(89)90069-X

60. Genge, B. R., L. N. Wu & R. E. Wuthier: Identification of phospholipid-dependent calcium-binding proteins as constituents of matrix vesicles. J Biol Chem, 264, 10917-21 (1989)
PMid:2499584

61. Genge, B. R., L. N. Wu & R. E. Wuthier: Differential fractionation of matrix vesicle proteins. Further characterization of the acidic phospholipid-dependent Ca2+-binding proteins. J Biol Chem, 265, 4703-10 (1990)
PMid:2155235

62. Balcerzak, M., J. Radisson, G. Azzar, D. Farlay, G. Boivin, S. Pikula & R. Buchet: A comparative analysis of strategies for isolation of matrix vesicles. Anal Biochem, 361, 176-82 (2007)
doi:10.1016/j.ab.2006.10.001
PMid:17194438

63. Wuthier, R. E.: Electrolytes of isolated epiphyseal chondrocytes, matrix vesicles and extracellular fluid. Calcif Tissue Res, 23, 125-133 (1977)
doi:10.1007/BF02012777
PMid:890549

64. Wuthier, R. E. & S. T. Gore: Partition of inorganic ions and phospholipids in isolated cell, membrane and matrix vesicle fractions: Evidence for Ca-Pi-acidic phospholipid complexes. Calcif Tissue Res, 24, 163-171 (1977)
doi:10.1007/BF02223311
PMid:597754

65. Warner, G. P. & R. E. Wuthier: Metabolism of 32Pi and 45Ca by matrix vesicles isolated from epiphyseal cartilage by nondigestive methods. In: Third International Conference on Matrix Vesicles, Eds: Ascenzi, A., E. Bonucci & B. deBernard, Wichtig Editore, Milano, Italy, 47-51 (1981).

66. Register, T. C., F. M. McLean, M. G. Low & R. E. Wuthier: Roles of alkaline phosphatase and labile internal mineral in matrix vesicle-mediated calcification. Effect of selective release of membrane-bound alkaline phosphatase and treatment with isosmotic pH 6 buffer. J Biol Chem, 261, 9354-60 (1986)
PMid:3722200

67. McLean, F. M., P. J. Keller, B. R. Genge, S. A. Walters & R. E. Wuthier: Disposition of preformed mineral in matrix vesicles. Internal localization and association with alkaline phosphatase. J Biol Chem, 262, 10481-8 (1987)
PMid:3611080

68. Hale, J. E., J. E. Chin, Y. Ishikawa, P. R. Paradiso & R. E. Wuthier: Correlation between distribution of cytoskeletal proteins and release of alkaline phosphatase-rich vesicles by epiphyseal chondrocytes in primary culture. Cell Motil, 3, 501-12 (1983)
doi:10.1002/cm.970030517
PMid:6420067

69. Hale, J. E. & R. E. Wuthier: Role of cytoskeletal depolymerization in matrix vesicle formation by primary cultures of epiphyseal growth plate chondrocytes. In: Cell Mediated Calcification and Matrix Vesicles, Ed: Ali, S. Y., Elsevier Science Publ., Amsterdam, The Netherlands, 237-240 (1986).

70. Schalk, E. M. & R. E. Wuthier: Effect of trifluoperazine and other drugs on matrix vesicle formation by chicken growth plate chondrocytes in primary cell culture. Biochem Pharmacol, 35, 2373-9 (1986)
doi:10.1016/0006-2952(86)90464-8

71. Ishikawa, Y., J. E. Chin, E. M. Schalk & R. E. Wuthier: Effect of amino acid levels on matrix vesicle formation by epiphyseal growth plate chondrocytes in primary culture. J Cell Physiol, 126, 399-406 (1986)
doi:10.1002/jcp.1041260310
PMid:3949889

72. Garimella, R., X. Bi, N. Camacho, J. B. Sipe & H. C. Anderson: Primary culture of rat growth plate chondrocytes: an in vitro model of growth plate histotype, matrix vesicle biogenesis and mineralization. Bone, 34, 961-70 (2004)
PMid:3543016

73. Hale, J. E. & R. E. Wuthier: The mechanism of matrix vesicle formation. Studies on the composition of chondrocyte microvilli and on the effects of microfilament-perturbing agents on cellular vesiculation. J Biol Chem, 262, 1916-25 (1987)
PMid:2592380

74. Boyan, B. D., Z. Schwartz, L. D. Swain, D. L. Carnes, Jr. & T. Zislis: Differential expression of phenotype by resting zone and growth region costochondral chondrocytes in vitro. Bone, 9, 185-94 (1988)
PMid:3192545

75. Wu, L. N., G. R. Sauer, B. R. Genge & R. E. Wuthier: Induction of mineral deposition by primary cultures of chicken growth plate chondrocytes in ascorbate-containing media. Evidence of an association between matrix vesicles and collagen. J Biol Chem, 264, 21346-55 (1989)

76. Genge, B. R., G. R. Sauer, L. N. Wu, F. M. McLean & R. E. Wuthier: Correlation between loss of alkaline phosphatase activity and accumulation of calcium during matrix vesicle-mediated mineralization. J Biol Chem, 263, 18513-9 (1988)
PMid:18095356

77. Valhmu, W. B., L. N. Wu & R. E. Wuthier: Effects of Ca/Pi ratio, Ca2+ x Pi ion product, and pH of incubation fluid on accumulation of 45Ca2+ by matrix vesicles in vitro. Bone Miner, 8, 195-209 (1990)
doi:10.1016/0169-6009(90)90105-O
PMid:1146961    PMCid:1912651

78. Balcerzak, M., A. Malinowska, C. Thouverey, A. Sekrecka, M. Dadlez, R. Buchet & S. Pikula: Proteome analysis of matrix vesicles isolated from femurs of chicken embryo. Proteomics, 8, 192-205 (2008)
doi:10.1002/pmic.200700612
PMid:200279

79. Anderson, H. C., R. Cecil & S. W. Sajdera: Calcification of rachitic rat cartilage in vitro by extracellular matrix vesicles. Am J Pathol, 79, 237-54 (1975)

80. Hsu, H. H. & H. C. Anderson: A simple and defined method to study calcification by isolated matrix vesicles. Effect of ATP and vesicle phosphatase. Biochim Biophys Acta, 500, 162-72 (1977)

81. Hsu, H. H. & H. C. Anderson: Calcification of isolated matrix vesicles and reconstituted vesicles from fetal bovine cartilage. Proc Natl Acad Sci U S A, 75, 3805-8 (1978)
doi:10.1073/pnas.75.8.3805

82. Hsu, H. H., N. P. Camacho & H. C. Anderson: Further characterization of ATP-initiated calcification by matrix vesicles isolated from rachitic rat cartilage. Membrane perturbation by detergents and deposition of calcium pyrophosphate by rachitic matrix vesicles. Biochim Biophys Acta, 1416, 320-32 (1999)
doi:10.1016/S0005-2736(98)00235-1
PMid:174957

83. Kakuta, S., D. Malamud, E. E. Golub & I. M. Shapiro: Isolation of matrix vesicles by isoelectric focusing in Pevikon-Sephadex. Bone, 6, 187-91 (1985)

84. Vaananen, H. K. & L. K. Korhonen: Matrix vesicles in chicken epiphyseal cartilage. Separation from lysosomes and the distribution of inorganic pyrophosphatase activity. Calcif Tissue Int, 28, 65-72 (1979)
doi:10.1007/BF02441219
PMid:17096383

85. Slavkin, H. C., R. D. Croissant, P. Bringas, P. Matosian, P. Wilson, W. Mino & H. Guenther: Matrix vesicle heterogeneity: possible morphogenetic functions for matrix vesicles. Fed Proc, 35, 127-34 (1976)
PMid:159774

86. Deutsch, D., I. Bab, A. Muhlrad & J. Sela: Purification and further characterization of isolated matrix vesicles from rat alveolar bone. Metab Bone Dis Relat Res, 3, 209-14 (1981)
doi:10.1016/0221-8747(81)90010-2
PMid:6111382

87. Xiao, Z., C. E. Camalier, K. Nagashima, K. C. Chan, D. A. Lucas, M. J. de la Cruz, M. Gignac, S. Lockett, H. J. Issaq, T. D. Veenstra, T. P. Conrads & G. R. Beck, Jr.: Analysis of the extracellular matrix vesicle proteome in mineralizing osteoblasts. J Cell Physiol, 210, 325-35 (2007)
doi:10.1002/jcp.20826

88. Bab, I. A., A. Muhlrad & J. Sela: Ultrastructural and biochemical study of extracellular matrix vesicles in normal alveolar bone of rats. Cell Tissue Res, 202, 1-7 (1979)
doi:10.1007/BF00239215

89. Sela, J., I. A. Bab & A. Muhlrad: A comparative study on the occurrence and activity of extracellular matrix vesicles in young and adult rat maxillary bone. Calcif Tissue Int, 33, 129-34 (1981)
doi:10.1007/BF02409425
PMid:6814727

90. Muhlrad, A., A. Setton, J. Sela, I. Bab & D. Deutsch: Biochemical characterization of matrix vesicles from bone and cartilage. Metab Bone Dis Relat Res, 5, 93-9 (1983)
doi:10.1016/0221-8747(83)90008-5
PMid:2037607

91. Chin, J. E., J. E. Hale, Y. Ishikawa, T. C. Register & R. E. Wuthier: Origin of matrix vesicles: Characterization of mineralization-inducing vesicles and their production by primary cultures of chicken growth plate chondrocytes. In: Endocrine Control of Bone and Calcium Metabolism, Eds: D. V. Cohn, Potts, J.T., Jr., Fujita, T., Elsevier Science Publ., Amsterdam, The Netherlands, 454-458 (1984).
PMid:1386488

92. Muhlrad, A., I. A. Bab, D. Deutsch & J. Sela: Occurrence of actin-like protein in extracellular matrix vesicles. Calcif Tissue Int, 34, 376-81 (1982)
doi:10.1007/BF02411271
PMid:1845989

93. Genge, B. R., L. N. Wu, H. D. Adkisson, IV & R. E. Wuthier: Matrix vesicle annexins exhibit proteolipid-like properties. Selective partitioning into lipophilic solvents under acidic conditions. J Biol Chem, 266, 10678-85 (1991)
PMid:1985942

94. Genge, B. R., X. Cao, L. N. Wu, W. R. Buzzi, R. W. Showman, A. L. Arsenault, Y. Ishikawa & R. E. Wuthier: Establishment of the primary structure of the major lipid-dependent Ca2+ binding proteins of chicken growth plate cartilage matrix vesicles: identity with anchorin CII (annexin V) and annexin II. J Bone Miner Res, 7, 807-19 (1992)
doi:10.1002/jbmr.5650070710
PMid:8267590

95. Wu, L. N., B. R. Genge, G. C. Lloyd & R. E. Wuthier: Collagen-binding proteins in collagenase-released matrix vesicles from cartilage. Interaction between matrix vesicle proteins and different types of collagen. J Biol Chem, 266, 1195-203 (1991)

96. Wu, L. N., B. R. Genge & R. E. Wuthier: Association between proteoglycans and matrix vesicles in the extracellular matrix of growth plate cartilage. J Biol Chem, 266, 1187-94 (1991)
PMid:19068243

97. Cao, X., B. R. Genge, L. N. Wu, W. R. Buzzi, R. M. Showman & R. E. Wuthier: Characterization, cloning and expression of the 67-kDA annexin from chicken growth plate cartilage matrix vesicles. Biochem Biophys Res Commun, 197, 556-61 (1993)
doi:10.1006/bbrc.1993.2515
PMid:16118637

98. Xiao, Z., T. P. Conrads, G. R. Beck & T. D. Veenstra: Analysis of the extracellular matrix and secreted vesicle proteomes by mass spectrometry. Methods Mol Biol, 428, 231-44 (2008)
doi:10.1007/978-1-59745-117-8_13
PMid:15193542

99. Xiao, Z., J. Blonder, M. Zhou & T. D. Veenstra: Proteomic analysis of extracellular matrix and vesicles. J Proteomics, 72, 34-45 (2009)
doi:10.1016/j.jprot.2008.11.011
PMid:3166834

100. Elias, J. E., W. Haas, B. K. Faherty & S. P. Gygi: Comparative evaluation of mass spectrometry platforms used in large-scale proteomics investigations. Nat Methods, 2, 667-75 (2005)
doi:10.1038/nmeth785
PMid:4027096

101. Perkins, D. N., D. J. Pappin, D. M. Creasy & J. S. Cottrell: Probability-based protein identification by searching sequence databases using mass spectrometry data. Electrophoresis, 20, 3551-67 (1999)
doi:10.1002/(SICI)1522-2683(19991201)20:18<3551::AID-ELPS3551>3.0.CO;2-2

102. Kirsch, T. & M. Pfaffle: Selective binding of anchorin CII (annexin V) to type II and X collagen and to chondrocalcin (C-propeptide of type II collagen). Implications for anchoring function between matrix vesicles and matrix proteins. FEBS Lett, 310, 143-7 (1992)
doi:10.1016/0014-5793(92)81316-E

103. Kirsch, T. & R. E. Wuthier: Stimulation of calcification of growth plate cartilage matrix vesicles by binding to type II and X collagens. J Biol Chem, 269, 11462-9 (1994)
PMid:8157677

104. Poole, C. A., S. Ayad & R. T. Gilbert: Chondrons from articular cartilage. V. Immunohistochemical evaluation of type VI collagen organisation in isolated chondrons by light, confocal and electron microscopy. J Cell Sci, 103 (Pt 4), 1101-10 (1992)
PMid:1487492

105. Jobsis, G. J., H. Keizers, J. P. Vreijling, M. de Visser, M. C. Speer, R. A. Wolterman, F. Baas & P. A. Bolhuis: Type VI collagen mutations in Bethlem myopathy, an autosomal dominant myopathy with contractures. Nat Genet, 14, 113-5 (1996)
doi:10.1038/ng0996-113
PMid:8782832

106. Lamande, S. R., J. F. Bateman, W. Hutchison, R. J. McKinlay Gardner, S. P. Bower, E. Byrne & H. H. Dahl: Reduced collagen VI causes Bethlem myopathy: a heterozygous COL6A1 nonsense mutation results in mRNA decay and functional haploinsufficiency. Hum Mol Genet, 7, 981-9 (1998)
doi:10.1093/hmg/7.6.981
PMid:9580662

107. Camacho Vanegas, O., E. Bertini, R. Z. Zhang, S. Petrini, C. Minosse, P. Sabatelli, B. Giusti, M. L. Chu & G. Pepe: Ullrich scleroatonic muscular dystrophy is caused by recessive mutations in collagen type VI. Proc Natl Acad Sci U S A, 98, 7516-21 (2001)
doi:10.1073/pnas.121027598
PMid:11381124    PMCid:34700

108. Nishino, I. & E. Ozawa: Muscular dystrophies. Curr Opin Neurol, 15, 539-44 (2002)
doi:10.1097/00019052-200210000-00004
PMid:12351997

109. Nishiyama, A. & W. B. Stallcup: Expression of NG2 proteoglycan causes retention of type VI collagen on the cell surface. Mol Biol Cell, 4, 1097-108 (1993)
PMid:8305732    PMCid:275746

110. Burg, M. A., E. Tillet, R. Timpl & W. B. Stallcup: Binding of the NG2 proteoglycan to type VI collagen and other extracellular matrix molecules. J Biol Chem, 271, 26110-6 (1996)
doi:10.1074/jbc.271.42.26110
PMid:8824254

111. Bernardi, P. & P. Bonaldo: Dysfunction of mitochondria and sarcoplasmic reticulum in the pathogenesis of collagen VI muscular dystrophies. Ann N Y Acad Sci, 1147, 303-11 (2008)
doi:10.1196/annals.1427.009
PMid:19076452

112. Bernardi, P. & A. Rasola: Calcium and cell death: the mitochondrial connection. Subcell Biochem, 45, 481-506 (2007)
doi:10.1007/978-1-4020-6191-2_18

113. Habuchi, H., H. E. Conrad & J. H. Glaser: Coordinate regulation of collagen and alkaline phosphatase levels in chick embryo chondrocytes. J Biol Chem, 260, 13029-34 (1985)
PMid:4055731

114. Reginato, A. M., I. M. Shapiro, J. W. Lash & S. A. Jimenez: Type X collagen alterations in rachitic chick epiphyseal growth cartilage. J Biol Chem, 263, 9938-45 (1988)
PMid:3133372

115. Salo, L. A., J. Hoyland, S. Ayad, C. M. Kielty, A. Freemont, P. Pirttiniemi, T. Kantomaa, M. E. Grant & J. T. Thomas: The expression of types X and VI collagen and fibrillin in rat mandibular condylar cartilage. Response to mastication forces. Acta Odontol Scand, 54, 295-302 (1996)
doi:10.3109/00016359609003541
PMid:8923923

116. Genge, B. R., L. N. Wu & R. E. Wuthier: Mineralization of annexin-5-containing lipid-calcium-phosphate complexes: modulation by varying lipid composition and incubation with cartilage collagens. J Biol Chem, 283, 9737-48 (2008)
doi:10.1074/jbc.M706523200
PMid:18250169    PMCid:2442302

117. Ruggiero, F., B. Petit, M. C. Ronziere, J. Farjanel, D. J. Hartmann & D. Herbage: Composition and organization of the collagen network produced by fetal bovine chondrocytes cultured at high density. J Histochem Cytochem, 41, 867-75 (1993)
PMid:8315278

118. Keene, D. R., C. D. Jordan, D. P. Reinhardt, C. C. Ridgway, R. N. Ono, G. M. Corson, M. Fairhurst, M. D. Sussman, V. A. Memoli & L. Y. Sakai: Fibrillin-1 in human cartilage: developmental expression and formation of special banded fibers. J Histochem Cytochem, 45, 1069-82 (1997)
PMid:9267468

119. Trask, B. C., T. M. Trask, T. Broekelmann & R. P. Mecham: The microfibrillar proteins MAGP-1 and fibrillin-1 form a ternary complex with the chondroitin sulfate proteoglycan decorin. Mol Biol Cell, 11, 1499-507 (2000)
PMid:10793130    PMCid:14862

120. Francke, U. & H. Furthmayr: Genes and gene products involved in Marfan syndrome. Semin Thorac Cardiovasc Surg, 5, 3-10 (1993)
PMid:8425000

121. Huber, R., M. Schneider, I. Mayr, J. Romisch & E. P. Paques: The calcium binding sites in human annexin V by crystal structure analysis at 2.0 A resolution. Implications for membrane binding and calcium channel activity. FEBS Lett, 275, 15-21 (1990)
doi:10.1016/0014-5793(90)81428-Q

122. Huber, R., J. Romisch & E. P. Paques: The crystal and molecular structure of human annexin V, an anticoagulant protein that binds to calcium and membranes. Embo J, 9, 3867-74 (1990)
PMid:2147412    PMCid:552154

123. Huber, R., R. Berendes, A. Burger, H. Luecke & A. Karshikov: Annexin V-crystal structure and its implications on function. Behring Inst Mitt, 91, 107-25 (1992)
PMid:1388018

124. Voges, D., R. Berendes, A. Burger, P. Demange, W. Baumeister & R. Huber: Three-dimensional structure of membrane-bound annexin V. A correlative electron microscopy-X-ray crystallography study. J Mol Biol, 238, 199-213 (1994)
doi:10.1006/jmbi.1994.1281
PMid:8158649

125. Arispe, N., E. Rojas, B. R. Genge, L. N. Wu & R. E. Wuthier: Similarity in calcium channel activity of annexin V and matrix vesicles in planar lipid bilayers. Biophys J, 71, 1764-75 (1996)
doi:10.1016/S0006-3495(96)79377-3

126. Reviakine, I. I., W. Bergsma-Schutter & A. Brisson: Growth of protein 2-D crystals on supported planar lipid bilayers imaged in situ by AFM. J Struct Biol, 121, 356-61 (1998)
doi:10.1006/jsbi.1998.4003
PMid:9705879

127. Sopkova, J., M. Vincent, M. Takahashi, A. Lewit-Bentley & J. Gallay: Conformational flexibility of domain III of annexin V at membrane/water interfaces. Biochemistry, 38, 5447-58 (1999)
doi:10.1021/bi982760g
PMid:10220332

128. Reviakine, I., W. Bergsma-Schutter, C. Mazeres-Dubut, N. Govorukhina & A. Brisson: Surface topography of the p3 and p6 annexin V crystal forms determined by atomic force microscopy. J Struct Biol, 131, 234-9 (2000)
doi:10.1006/jsbi.2000.4286
PMid:11052896

129. Oling, F., J. S. Santos, N. Govorukhina, C. Mazeres-Dubut, W. Bergsma-Schutter, G. Oostergetel, W. Keegstra, O. Lambert, A. Lewit-Bentley & A. Brisson: Structure of membrane-bound annexin A5 trimers: a hybrid cryo-EM - X-ray crystallography study. J Mol Biol, 304, 561-73 (2000)
doi:10.1006/jmbi.2000.4183
PMid:11099380

130. Oling, F., W. Bergsma-Schutter & A. Brisson: Trimers, dimers of trimers, and trimers of trimers are common building blocks of annexin a5 two-dimensional crystals. J Struct Biol, 133, 55-63 (2001)
doi:10.1006/jsbi.2000.4337
PMid:11356064

131. Wu, L. N., B. R. Genge & R. E. Wuthier: Analysis and molecular modeling of the formation, structure, and activity of the phosphatidylserine-calcium-phosphate complex associated with biomineralization. J Biol Chem, 283, 3827-38 (2008)
doi:10.1074/jbc.M707653200
PMid:18077457

132. Martin, S. J., C. P. Reutelingsperger, A. J. McGahon, J. A. Rader, R. C. van Schie, D. M. LaFace & D. R. Green: Early redistribution of plasma membrane phosphatidylserine is a general feature of apoptosis regardless of the initiating stimulus: inhibition by overexpression of Bcl-2 and Abl. J Exp Med, 182, 1545-56 (1995)
doi:10.1084/jem.182.5.1545
PMid:7595224

133. Fabisiak, J. P., V. E. Kagan, V. B. Ritov, D. E. Johnson & J. S. Lazo: Bcl-2 inhibits selective oxidation and externalization of phosphatidylserine during paraquat-induced apoptosis. Am J Physiol, 272, C675-84 (1997)

134. Zhang, G., V. Gurtu, S. R. Kain & G. Yan: Early detection of apoptosis using a fluorescent conjugate of annexin V. Biotechniques, 23, 525-31 (1997)
PMid:9298227

135. Johnson, K., K. Pritzker, J. Goding & R. Terkeltaub: The nucleoside triphosphate pyrophosphohydrolase isozyme PC-1 directly promotes cartilage calcification through chondrocyte apoptosis and increased calcium precipitation by mineralizing vesicles. J Rheumatol, 28, 2681-91 (2001)
PMid:11764218

136. Genge, B. R., L. N. Wu & R. E. Wuthier: In vitro modeling of matrix vesicle nucleation: synergistic stimulation of mineral formation by annexin A5 and phosphatidylserine. J Biol Chem, 282, 26035-45 (2007)
doi:10.1074/jbc.M701057200
PMid:17613532

137. Wu, L. N., B. R. Genge & R. E. Wuthier: Differential effects of zinc and magnesium ions on mineralization activity of phosphatidylserine calcium phosphate complexes. J Inorg Biochem, 103, 948-62 (2009)
doi:10.1016/j.jinorgbio.2009.04.004
PMid:19477528

138. Smith, P. D. & S. E. Moss: Structural evolution of the annexin supergene family. Trends Genet, 10, 241-6 (1994)
doi:10.1016/0168-9525(94)90171-6

139. Smith, P. D., A. Davies, M. J. Crumpton & S. E. Moss: Structure of the human annexin VI gene. Proc Natl Acad Sci U S A, 91, 2713-7 (1994)
doi:10.1073/pnas.91.7.2713

140. Avila-Sakar, A. J., C. E. Creutz & R. H. Kretsinger: Crystal structure of bovine annexin VI in a calcium-bound state. Biochim Biophys Acta, 1387, 103-16 (1998)
doi:10.1016/S0167-4838(98)00111-3

141. Crompton, M. R., R. J. Owens, N. F. Totty, S. E. Moss, M. D. Waterfield & M. J. Crumpton: Primary structure of the human, membrane-associated Ca2+-binding protein p68 a novel member of a protein family. Embo J, 7, 21-7 (1988)
PMid:3258820    PMCid:454210

142. Bandorowicz-Pikula, J. & S. Pikula: Modulation of annexin VI-driven aggregation of phosphatidylserine liposomes by ATP. Biochimie, 80, 613-20 (1998)
doi:10.1016/S0300-9084(98)80014-X

143. Theobald, J., P. D. Smith, S. M. Jacob & S. E. Moss: Expression of annexin VI in A431 carcinoma cells suppresses proliferation: a possible role for annexin VI in cell growth regulation. Biochim Biophys Acta, 1223, 383-90 (1994)
doi:10.1016/0167-4889(94)90099-X

144. Chen, N. X., K. D. O'Neill, X. Chen & S. M. Moe: Annexin-mediated matrix vesicle calcification in vascular smooth muscle cells. J Bone Miner Res, 23, 1798-805 (2008)
doi:10.1359/jbmr.080604
PMid:18597635    PMCid:2685487

145. Kirsch, T., G. Harrison, E. E. Golub & H. D. Nah: The roles of annexins and types II and X collagen in matrix vesicle-mediated mineralization of growth plate cartilage. J Biol Chem, 275, 35577-83 (2000)
doi:10.1074/jbc.M005648200
PMid:10956650

146. Merrifield, C. J., U. Rescher, W. Almers, J. Proust, V. Gerke, A. S. Sechi & S. E. Moss: Annexin 2 has an essential role in actin-based macropinocytic rocketing. Curr Biol, 11, 1136-41 (2001)
doi:10.1016/S0960-9822(01)00321-9

147. Senda, T.: Mechanisms of secretory granule transport and exocytosis in anterior pituitary cells. Ital J Anat Embryol, 100 Suppl 1, 219-29 (1995)
PMid:11322296

148. Chasserot-Golaz, S., N. Vitale, E. Umbrecht-Jenck, D. Knight, V. Gerke & M. F. Bader: Annexin 2 promotes the formation of lipid microdomains required for calcium-regulated exocytosis of dense-core vesicles. Mol Biol Cell, 16, 1108-19 (2005)
doi:10.1091/mbc.E04-07-0627
PMid:15635098    PMCid:551477

149. Christian, H. C., A. D. Taylor, R. J. Flower, J. F. Morris & J. C. Buckingham: Characterization and localization of lipocortin 1-binding sites on rat anterior pituitary cells by fluorescence-activated cell analysis/sorting and electron microscopy. Endocrinology, 138, 5341-51 (1997)
doi:10.1210/en.138.12.5341
PMid:9389519

150. Serres, M., C. Comera & D. Schmitt: Annexin 1 regulation in human epidermal cells. Cell Mol Biol (Noisy-le-grand), 40, 701-6 (1994)

151. Kim, K. M., D. K. Kim, Y. M. Park, C. K. Kim & D. S. Na: Annexin-I inhibits phospholipase A2 by specific interaction, not by substrate depletion. FEBS Lett, 343, 251-5 (1994)
doi:10.1016/0014-5793(94)80566-0

152. Bastian, B. C., C. Sellert, A. Seekamp, J. Romisch, E. P. Paques & E. B. Brocker: Inhibition of human skin phospholipase A2 by "lipocortins" is an indirect effect of substrate/lipocortin interaction. J Invest Dermatol, 101, 359-63 (1993)
doi:10.1111/1523-1747.ep12365541
PMid:8370974

153. Solito, E., R. De Caterina, D. Giannessi, P. L. Paggiaro, R. Sicari & L. Parente: Studies on the induction of lipocortin-1 by glucocorticoids. Ann Ist Super Sanita, 29, 391-4 (1993)
PMid:8172456

154. Northington, F. K., T. D. Oglesby, Y. Ishikawa & R. E. Wuthier: Localization of prostaglandin synthetase in chicken epiphyseal cartilage. Calcif Tissue Res, 26, 227-36 (1978)
doi:10.1007/BF02013263
PMid:108004

155. Teixeira, M. M., A. M. Das, J. M. Miotla, M. Perretti & P. G. Hellewell: The role of lipocortin-1 in the inhibitory action of dexamethasone on eosinophil trafficking in cutaneous inflammatory reactions in the mouse. Br J Pharmacol, 123, 538-44 (1998)
doi:10.1038/sj.bjp.0701625
PMid:9504395    PMCid:1565186

156. Yi, M. & J. E. Schnitzer: Impaired tumor growth, metastasis, angiogenesis and wound healing in annexin A1-null mice. Proc Natl Acad Sci U S A, 106, 17886-91 (2009)
doi:10.1073/pnas.0901324106
PMid:19805119    PMCid:2764877

157. Ruan, G. R., H. S. Zhao, Y. Chang, J. L. Li, Y. Z. Qin, Y. R. Liu, S. S. Chen & X. J. Huang: Adenovirus-mediated PDCD5 gene transfer sensitizes K562 cells to apoptosis induced by idarubicin in vitro and in vivo. Apoptosis, 13, 641-8 (2008)
doi:10.1007/s10495-008-0206-9
PMid:18401719

158. Ali, S. Y.: Analysis of matrix vesicles and their role in the calcification of epiphyseal cartilage. Fed Proc, 35, 135-42 (1976)
PMid:1248646

159. Hosokawa, R., Y. Uchida, S. Fujiwara & T. Noguchi: Lactate dehydrogenase isoenzymes are present in matrix vesicles. J Biol Chem, 263, 10045-7 (1988)
PMid:3392004

160. Sauer, G. R., B. R. Genge, L. N. Wu & J. E. Donachy: A facilitative role for carbonic anhydrase activity in matrix vesicle mineralization. Bone Miner, 26, 69-79 (1994)
doi:10.1016/S0169-6009(08)80163-8

161. Robison, R.: The possible significance of hexosephosphoric esters in ossification. Biochem J, 17, 286-93 (1923)
PMid:16743183    PMCid:1259346

162. Anderson, H. C., D. Harmey, N. P. Camacho, R. Garimella, J. B. Sipe, S. Tague, X. Bi, K. Johnson, R. Terkeltaub & J. L. Millan: Sustained osteomalacia of long bones despite major improvement in other hypophosphatasia-related mineral deficits in tissue nonspecific alkaline phosphatase/nucleotide pyrophosphatase phosphodiesterase 1 double-deficient mice. Am J Pathol, 166, 1711-20 (2005)
PMid:15920156    PMCid:1602415

163. Anderson, H. C., J. B. Sipe, L. Hessle, R. Dhanyamraju, E. Atti, N. P. Camacho & J. L. Millan: Impaired calcification around matrix vesicles of growth plate and bone in alkaline phosphatase-deficient mice. Am J Pathol, 164, 841-7 (2004)
PMid:14982838    PMCid:1613274

164. Hessle, L., K. A. Johnson, H. C. Anderson, S. Narisawa, A. Sali, J. W. Goding, R. Terkeltaub & J. L. Millan: Tissue-nonspecific alkaline phosphatase and plasma cell membrane glycoprotein-1 are central antagonistic regulators of bone mineralization. Proc Natl Acad Sci U S A, 99, 9445-9 (2002)
doi:10.1073/pnas.142063399
PMid:12082181    PMCid:123160

165. Harmey, D., L. Hessle, S. Narisawa, K. A. Johnson, R. Terkeltaub & J. L. Millan: Concerted regulation of inorganic pyrophosphate and osteopontin by akp2, enpp1, and ank: an integrated model of the pathogenesis of mineralization disorders. Am J Pathol, 164, 1199-209 (2004)
PMid:15039209    PMCid:1615351

166. Weiss, M. J., P. S. Henthorn, M. A. Lafferty, C. Slaughter, M. Raducha & H. Harris: Isolation and characterization of a cDNA encoding a human liver/bone/kidney-type alkaline phosphatase. Proc Natl Acad Sci U S A, 83, 7182-6 (1986)
doi:10.1073/pnas.83.19.7182

167. Low, M. G. & D. B. Zilversmit: Role of phosphatidylinositol in attachment of alkaline phosphatase to membranes. Biochemistry, 19, 3913-8 (1980)
doi:10.1021/bi00558a004
PMid:7407077

168. Hsu, D. S. & S. S. Chen: Heterogeneity of alkaline phosphatase observed by high-performance liquid chromatography. J Chromatogr, 328, 409-12 (1985)
doi:10.1016/S0021-9673(01)87418-5

169. Hsu, H. H., P. A. Munoz, J. Barr, I. Oppliger, D. C. Morris, H. K. Vaananen, N. Tarkenton & H. C. Anderson: Purification and partial characterization of alkaline phosphatase of matrix vesicles from fetal bovine epiphyseal cartilage. Purification by monoclonal antibody affinity chromatography. J Biol Chem, 260, 1826-31 (1985)
PMid:3968087

170. Sowadski, J. M., M. D. Handschumacher, H. M. Murthy, B. A. Foster & H. W. Wyckoff: Refined structure of alkaline phosphatase from Escherichia coli at 2.8 A resolution. J Mol Biol, 186, 417-33 (1985)
doi:10.1016/0022-2836(85)90115-9

171. Kim, E. E. & H. W. Wyckoff: Reaction mechanism of alkaline phosphatase based on crystal structures. Two-metal ion catalysis. J Mol Biol, 218, 449-64 (1991)
doi:10.1016/0022-2836(91)90724-K

172. Llinas, P., E. A. Stura, A. Menez, Z. Kiss, T. Stigbrand, J. L. Millan & M. H. Le Du: Structural studies of human placental alkaline phosphatase in complex with functional ligands. J Mol Biol, 350, 441-51 (2005)
doi:10.1016/j.jmb.2005.04.068
PMid:15946677

173. Llinas, P., M. Masella, T. Stigbrand, A. Menez, E. A. Stura & M. H. Le Du: Structural studies of human alkaline phosphatase in complex with strontium: implication for its secondary effect in bones. Protein Sci, 15, 1691-700 (2006)
doi:10.1110/ps.062123806
PMid:16815919    PMCid:2242561

174. Stagni, N., G. Furlan, F. Vittur, M. Zanetti & B. de Bernard: Enzymatic properties of the Ca2+-binding glycoprotein isolated from preosseous cartilage. Calcif Tissue Int, 29, 27-32 (1979)
doi:10.1007/BF02408052
PMid:116741

175. de Bernard, B., M. Gherardini, G. C. Lunazzi, C. Modricky, L. Moro, E. Panfili, P. Pollesello, N. Stagni & F. Vittur: Role of the Ca++-binding alkaline phosphatase in the mechanism of cartilage calcification. In: Cell Mediated Calcification and Matrix Vesicles. Ed: Y. Ali, Elsevier Science Publishers B.V., Amsterdam, The Netherlands, 95-100 (1986).

176. Harrison, G., I. M. Shapiro & E. E. Golub: The phosphatidylinositol-glycolipid anchor on alkaline phosphatase facilitates mineralization initiation in vitro. J Bone Miner Res, 10, 568-73 (1995)
doi:10.1002/jbmr.5650100409
PMid:7610927

177. Cyboron, G. W. & R. E. Wuthier: Purification and initial characterization of intrinsic membrane-bound alkaline phosphatase from chicken epiphyseal cartilage. J Biol Chem, 256, 7262-7268 (1981)
PMid:7251597

178. Eaton, R. H. & D. W. Moss: Organic pyrophosphates as substrates for human alkaline phosphatases. Biochem J, 105, 1307-1312 (1967)
PMid:16742561    PMCid:1198456

179. Moss, D. W., R. H. Eaton, J. K. Smith & L. G. Whitby: Association of inorganic-pyrophosphatase activity with human alkaline-phosphatase preparations. Biochem J, 102, 53-7 (1967)
PMid:6030299    PMCid:1270209

180. Anderson, H. C.: Matrix vesicles and calcification. Curr Rheumatol Rep, 5, 222-6 (2003)
doi:10.1007/s11926-003-0071-z
PMid:12744815

181. Whyte, M. P.: Hypophosphatasia and the role of alkaline phosphatase in skeletal mineralization. Endocr Rev, 15, 439-61 (1994)
doi:10.1210/er.15.4.439
PMid:7988481

doi:10.1210/edrv-15-4-439
PMid:7988481

182. Narisawa, S., N. Frohlander & J. L. Millan: Inactivation of two mouse alkaline phosphatase genes and establishment of a model of infantile hypophosphatasia. Dev Dyn, 208, 432-46 (1997)
doi:10.1002/(SICI)1097-0177(199703)208:3<432::AID-AJA13>3.0.CO;2-1

183. Fedde, K. N., L. Blair, J. Silverstein, S. P. Coburn, L. M. Ryan, R. S. Weinstein, K. Waymire, S. Narisawa, J. L. Millan, G. R. MacGregor & M. P. Whyte: Alkaline phosphatase knock-out mice recapitulate the metabolic and skeletal defects of infantile hypophosphatasia. J Bone Miner Res, 14, 2015-26 (1999)
doi:10.1359/jbmr.1999.14.12.2015
PMid:10620060

184. Vittur, F., N. Stagni, L. Moro & B. de Bernard: Alkaline phosphatase binds to collagen; a hypothesis on the mechanism of extravesicular mineralization in epiphyseal cartilage. Experientia, 40, 836-7 (1984)
doi:10.1007/BF01951980

185. Crawford, K., H. Weissig, F. Binette, J. L. Millan & P. F. Goetinck: Tissue-nonspecific alkaline phosphatase participates in the establishment and growth of feather germs in embryonic chick skin cultures. Dev Dyn, 204, 48-56 (1995)
PMid:8563025

186. Murphree, S., H. H. Hsu & H. C. Anderson: In vitro formation of crystalline apatite by matrix vesicles isolated from rachitic rat epiphyseal cartilage. Calcif Tissue Int, 34 Suppl 2, S62-8 (1982)

187. Siegel, S. A., C. F. Hummel & R. P. Carty: The role of nucleoside triphosphate pyrophosphohydrolase in in vitro nucleoside triphosphate-dependent matrix vesicle calcification. J Biol Chem, 258, 8601-7 (1983)
PMid:6134731

188. Kanabe, S., H. H. Hsu, R. N. Cecil & H. C. Anderson: Electron microscopic localization of adenosine triphosphate (ATP)-hydrolyzing activity in isolated matrix vesicles and reconstituted vesicles from calf cartilage. J Histochem Cytochem, 31, 462-70 (1983)
PMid:6219157

189. Caswell, A. M. & R. G. Russell: Identification of ecto-nucleoside triphosphate pyrophosphatase in human articular chondrocytes in monolayer culture. Biochim Biophys Acta, 847, 40-7 (1985)
doi:10.1016/0167-4889(85)90150-8

190. Simao, A. M., M. C. Yadav, S. Narisawa, M. Bolean, J. M. Pizauro, M. F. Hoylaerts, P. Ciancaglini & J. L. Millan: Proteoliposomes harboring alkaline phosphatase and nucleotide pyrophosphatase as matrix vesicle biomimetics. J Biol Chem, 285, 7598-609 (2010)
doi:10.1074/jbc.M109.079830
PMid:20048161    PMCid:2844207

191. Vaingankar, S. M., T. A. Fitzpatrick, K. Johnson, J. W. Goding, M. Maurice & R. Terkeltaub: Subcellular targeting and function of osteoblast nucleotide pyrophosphatase phosphodiesterase 1. Am J Physiol Cell Physiol, 286, C1177-87 (2004)
doi:10.1152/ajpcell.00320.2003
PMid:15075217

192. Garimella, R., X. Bi, H. C. Anderson & N. P. Camacho: Nature of phosphate substrate as a major determinant of mineral type formed in matrix vesicle-mediated in vitro mineralization: An FTIR imaging study. Bone, 38, 811-7 (2006)

193. Ciancaglini, P., M. C. Yadav, A. M. Simao, S. Narisawa, J. M. Pizauro, C. Farquharson, M. F. Hoylaerts & J. L. Millan: Kinetic analysis of substrate utilization by native and TNAP-, NPP1-, or PHOSPHO1-deficient matrix vesicles. J Bone Miner Res, 25, 1-9 (2009)
PMid:13893487

194. Fleisch, H. & S. Bisaz: Mechanism of calcification: inhibitory role of pyrophosphate. Nature, 195, 911 (1962)
doi:10.1038/195911a0

195. Terkeltaub, R., M. Rosenbach, F. Fong & J. Goding: Causal link between nucleotide pyrophosphohydrolase overactivity and increased intracellular inorganic pyrophosphate generation demonstrated by transfection of cultured fibroblasts and osteoblasts with plasma cell membrane glycoprotein-1. Relevance to calcium pyrophosphate dihydrate deposition disease. Arthritis Rheum, 37, 934-41 (1994)
doi:10.1002/art.1780370624

196. Rehemtulla, A. & R. J. Kaufman: Protein processing within the secretory pathway. Curr Opin Biotechnol, 3, 560-5 (1992)
doi:10.1016/0958-1669(92)90086-X
PMid:6292294

197. Goding, J. W. & F. W. Shen: Structure of the murine plasma cell alloantigen PC-1: comparison with the receptor for transferrin. J Immunol, 129, 2636-40 (1982)
PMid:3104326

198. van Driel, I. R. & J. W. Goding: Plasma cell membrane glycoprotein PC-1. Primary structure deduced from cDNA clones. J Biol Chem, 262, 4882-7 (1987)
PMid:11027689

199. Gijsbers, R., H. Ceulemans, W. Stalmans & M. Bollen: Structural and catalytic similarities between nucleotide pyrophosphatases/phosphodiesterases and alkaline phosphatases. J Biol Chem, 276, 1361-8 (2001)
doi:10.1074/jbc.M007552200

200. Fridrich, D., M. Kern, J. Fritz, G. Pahlke, N. Kohler, P. Winterhalter & D. Marko: The epidermal growth factor receptor and human topoisomerases represent potential cellular targets of oligomeric procyanidins. Mol Nutr Food Res, 51, 192-200 (2007)
doi:10.1002/mnfr.200600186
PMid:16461032

201. Muller, G. & W. Frick: Signalling via caveolin: involvement in the cross-talk between phosphoinositolglycans and insulin. Cell Mol Life Sci, 56, 945-70 (1999)
doi:10.1007/s000180050485

202. Roberts, S., S. Narisawa, D. Harmey, J. L. Millan & C. Farquharson: Functional involvement of PHOSPHO1 in matrix vesicle-mediated skeletal mineralization. J Bone Miner Res, 22, 617-27 (2007)
doi:10.1359/jbmr.070108
PMid:17227223

203. Stewart, A. J., R. Schmid, C. A. Blindauer, S. J. Paisey & C. Farquharson: Comparative modelling of human PHOSPHO1 reveals a new group of phosphatases within the haloacid dehalogenase superfamily. Protein Eng, 16, 889-95 (2003)
doi:10.1093/protein/gzg126
PMid:14983068

204. Houston, B., E. Seawright, D. Jefferies, E. Hoogland, D. Lester, C. Whitehead & C. Farquharson: Identification and cloning of a novel phosphatase expressed at high levels in differentiating growth plate chondrocytes. Biochim Biophys Acta, 1448, 500-6 (1999)
doi:10.1016/S0167-4889(98)00153-0

205. Macrae, V. E., M. G. Davey, L. McTeir, S. Narisawa, M. C. Yadav, J. L. Millan & C. Farquharson: Inhibition of PHOSPHO1 activity results in impaired skeletal mineralization during limb development of the chick. Bone, 46, 1146-55 (2010)
PMid:16054448

206. Roberts, S. J., A. J. Stewart, R. Schmid, C. A. Blindauer, S. R. Bond, P. J. Sadler & C. Farquharson: Probing the substrate specificities of human PHOSPHO1 and PHOSPHO2. Biochim Biophys Acta, 1752, 73-82 (2005)
PMid:11714705

207. Wu, L. N., B. R. Genge, M. W. Kang, A. L. Arsenault & R. E. Wuthier: Changes in phospholipid extractability and composition accompany mineralization of chicken growth plate cartilage matrix vesicles. J Biol Chem, 277, 5126-33 (2002)
doi:10.1074/jbc.M107899200

208. Wu, L. N., G. R. Sauer, B. R. Genge, W. B. Valhmu & R. E. Wuthier: Effects of analogues of inorganic phosphate and sodium ion on mineralization of matrix vesicles isolated from growth plate cartilage of normal rapidly growing chickens. J Inorg Biochem, 94, 221-35 (2003)
doi:10.1016/S0162-0134(03)00003-5
PMid:7634645

209. Anderson, H. C.: Molecular biology of matrix vesicles. Clin Orthop Relat Res266-80 (1995)
PMid:6652555

210. Hirschman, A., D. Deutsch, M. Hirschman, I. A. Bab, J. Sela & A. Muhlrad: Neutral peptidase activities in matrix vesicles from bovine fetal alveolar bone and dog osteosarcoma. Calcif Tissue Int, 35, 791-7 (1983)
doi:10.1007/BF02405125
PMid:5331119

211. Katsura, N. & K. Yamada: Isolation and characterization of a metalloprotease associated with chicken epiphyseal cartilage matrix vesicles. Bone, 7, 137-43 (1986)
PMid:4226655

212. Hirschman, A. & D. D. Dziewiatkowski: Protein-polysaccharide loss during endochondral ossification: immunochemical evidence. Science, 154, 393-5 (1966)
doi:10.1126/science.154.3747.393
PMid:5567494

213. Jibril, A. O.: Proteolytic degradation of ossifying cartilage matrix and the removal of acid mucopolysaccharides prior to bone formation. Biochim Biophys Acta, 136, 162-5 (1967)

214. Hirschman, A. & M. Hirschman: Aminopeptidase profile and protease activity in rat cartilage at physiological pH. Isr J Med Sci, 7, 403-5 (1971)

215. Hosokawa, R., H. Ohashi-Takeuchi, N. Yamada, Y. Uchida, S. Fujiwara & T. Noguchi: Lactate dehydrogenase isoenzymes in matrix vesicles (review). Bone Miner, 17, 177-81 (1992)
doi:10.1016/0169-6009(92)90732-S
PMid:4998681

216. Rajpurohit, R., C. J. Koch, Z. Tao, C. M. Teixeira & I. M. Shapiro: Adaptation of chondrocytes to low oxygen tension: relationship between hypoxia and cellular metabolism. J Cell Physiol, 168, 424-32 (1996)
doi:10.1002/(SICI)1097-4652(199608)168:2<424::AID-JCP21>3.0.CO;2-1

217. Cuervo, L. A., J. C. Pita & D. S. Howell: Ultramicroanalysis of pH, pCO2 and carbonic anhydrase activity at calcifying sites in cartilage. Calcif Tissue Res, 7, 220-31 (1971)
doi:10.1007/BF02062609
PMid:6816093

218. Vaananen, H. K.: Immunohistochemical localization of carbonic anhydrase isoenzymes I and II in human bone, cartilage and giant cell tumor. Histochemistry, 81, 485-7 (1984)
doi:10.1007/BF00489754
PMid:6801112

219. Anderson, R. E., H. Schraer & C. V. Gay: Ultrastructural immunocytochemical localization of carbonic anhydrase in normal and calcitonin-treated chick osteoclasts. Anat Rec, 204, 9-20 (1982)
doi:10.1002/ar.1092040103
PMid:1212633

220. Gay, C. V., R. E. Anderson, H. Schraer & D. S. Howell: Identification of carbonic anhydrase in chick growth-plate cartilage. J Histochem Cytochem, 30, 391-4 (1982)

221. Wuthier, R. E.: Effect of phospholipids on the transformation of amorphous calcium phosphate to hydroxapatite in vitro. Calcif Tissue Res, 19, 197-210 (1975)
doi:10.1007/BF02564004
PMid:4689119

222. Genge, B. R., L. N. Wu & R. E. Wuthier: Separation and quantification of chicken and bovine growth plate cartilage matrix vesicle lipids by high-performance liquid chromatography using evaporative light scattering detection. Anal Biochem, 322, 104-15 (2003)
doi:10.1016/S0003-2697(03)00462-7
PMid:16997299

223. Wuthier, R. E.: The role of phospholipids in biological calcification: distribution of phospholipase activity in calcifying epiphyseal cartilage. Clin Orthop Relat Res, 90, 191-200 (1973)
PMid:2925633

224. Balcerzak, M., S. Pikula & R. Buchet: Phosphorylation-dependent phospholipase D activity of matrix vesicles. FEBS Lett, 580, 5676-80 (2006)
doi:10.1016/j.febslet.2006.09.018
PMid:2606907

225. Kramer, R. M., C. Hession, B. Johansen, G. Hayes, P. McGray, E. P. Chow, R. Tizard & R. B. Pepinsky: Structure and properties of a human non-pancreatic phospholipase A2. J Biol Chem, 264, 5768-75 (1989)

226. Komada, M., I. Kudo, H. Mizushima, N. Kitamura & K. Inoue: Structure of cDNA coding for rat platelet phospholipase A2. J Biochem, 106, 545-7 (1989)
PMid:8474071

227. Stillo, J. V.: Characterization and metabolism of phospholipids in chicken epiphyseal cartilage: Studies on phospholipase A activity. Department of Chemistry and Biochemistry, University of South Carolina, Columbia (1980).
PMid:1428121

228. Nevalainen, T. J., F. Marki, P. T. Kortesuo, M. G. Grutter, S. Di Marco & A. Schmitz: Synovial type (group II) phospholipase A2 in cartilage. J Rheumatol, 20, 325-30 (1993)

229. Pruzanski, W., K. Scott, G. Smith, I. Rajkovic, E. Stefanski & P. Vadas: Enzymatic activity and immunoreactivity of extracellular phospholipase A2 in inflammatory synovial fluids. Inflammation, 16, 451-7 (1992)
doi:10.1007/BF00918971
PMid:3258818

230. Vignon, E., J. C. Balblanc, P. Mathieu, P. Louisot & M. Richard: Metalloprotease activity, phospholipase A2 activity and cytokine concentration in osteoarthritis synovial fluids. Osteoarthritis Cartilage, 1, 115-20 (1993)
doi:10.1016/S1063-4584(05)80026-3
PMid:3264240

231. Schwartz, Z. & B. Boyan: The effects of vitamin D metabolites on phospholipase A2 activity of growth zone and resting zone cartilage cells in vitro. Endocrinology, 122, 2191-8 (1988)
doi:10.1210/endo-122-5-2191
PMid:2945590

232. Schwartz, Z., D. L. Schlader, L. D. Swain & B. D. Boyan: Direct effects of 1,25-dihydroxyvitamin D3 and 24,25-dihydroxyvitamin D3 on growth zone and resting zone chondrocyte membrane alkaline phosphatase and phospholipase A2 specific activities. Endocrinology, 123, 2878-84 (1988)
doi:10.1210/endo-123-6-2878
PMid:6095092

233. Kristensen, T., C. J. Saris, T. Hunter, L. J. Hicks, D. J. Noonan, J. R. Glenney, Jr. & B. F. Tack: Primary structure of bovine calpactin I heavy chain (p36), a major cellular substrate for retroviral protein-tyrosine kinases: homology with the human phospholipase A2 inhibitor lipocortin. Biochemistry, 25, 4497-503 (1986)
doi:10.1021/bi00364a007
PMid:1411571

234. Berridge, M. J.: The Albert Lasker Medical Awards. Inositol trisphosphate, calcium, lithium, and cell signaling. Jama, 262, 1834-41 (1989)
PMid:16527821

235. Berridge, M. J. & R. F. Irvine: Inositol trisphosphate, a novel second messenger in cellular signal transduction. Nature, 312, 315-21 (1984)
doi:10.1038/312315a0
PMid:2981853

236. Nishizuka, Y.: Intracellular signaling by hydrolysis of phospholipids and activation of protein kinase C. Science, 258, 607-14 (1992)
doi:10.1126/science.1411571
PMid:16005036

237. Frey, R. S., X. Gao, K. Javaid, S. S. Siddiqui, A. Rahman & A. B. Malik: Phosphatidylinositol 3-kinase gamma signaling through protein kinase Czeta induces NADPH oxidase-mediated oxidant generation and NF-kappaB activation in endothelial cells. J Biol Chem, 281, 16128-38 (2006)
doi:10.1074/jbc.M508810200
PMid:10748000

238. O'Rourke, F. A., S. P. Halenda, G. B. Zavoico & M. B. Feinstein: Inositol 1,4,5-trisphosphate releases Ca2+ from a Ca2+-transporting membrane vesicle fraction derived from human platelets. J Biol Chem, 260, 956-62 (1985)
PMid:13129817

239. Ekstein, J., E. Nasatzky, B. D. Boyan, A. Ornoy & Z. Schwartz: Growth-plate chondrocytes respond to 17beta-estradiol with sex-specific increases in IP3 and intracellular calcium ion signalling via a capacitative entry mechanism. Steroids, 70, 775-86 (2005)
doi:10.1016/j.steroids.2005.04.007
PMid:16840176

240. Sjoholm, A., Q. Zhang, N. Welsh, A. Hansson, O. Larsson, M. Tally & P. O. Berggren: Rapid Ca2+ influx and diacylglycerol synthesis in growth hormone-mediated islet beta-cell mitogenesis. J Biol Chem, 275, 21033-40 (2000)
doi:10.1074/jbc.M001212200

241. Du, C., Q. Zhao, S. Araki, S. Zhang & J. Miao: Apoptosis mediated by phosphatidylcholine-specific phospholipase C is associated with cAMP, p53 level, and cell-cycle distribution in vascular endothelial cells. Endothelium, 10, 141-7 (2003)
doi:10.1080/713715227
PMid:19347873

242. Liu, X., Q. Zhao, S. Araki, S. Zhang & J. Miao: Contrasting effects of phosphatidylinositol- and phosphatidylcholine-specific phospholipase C on apoptosis in cultured endothelial cells. Endothelium, 13, 205-11 (2006)
doi:10.1080/10623320600760423
PMid:19626662

243. Szumilo, M. & I. Rahden-Staron: (Biological role of phosphatidylcholine-specific phospholipase C in mammalian cells). Postepy Hig Med Dosw (Online), 62, 593-8 (2008)
PMid:9110992

244. Fu, D., Y. Ma, W. Wu, X. Zhu, C. Jia, Q. Zhao, C. Zhang & X. Z. Wu: Cell-cycle-dependent PC-PLC regulation by APC/C (Cdc20)-mediated ubiquitin-proteasome pathway. J Cell Biochem, 107, 686-96 (2009)
doi:10.1002/jcb.22163
PMid:15350536

245. Sun, C., N. Wang, J. Huang, J. Xin, F. Peng, Y. Ren, S. Zhang & J. Miao: Inhibition of phosphatidylcholine-specific phospholipase C prevents bone marrow stromal cell senescence in vitro. J Cell Biochem, 108, 519-28 (2009)
doi:10.1002/jcb.22282
PMid:16325216

246. van Dijk, M. C., F. J. Muriana, J. de Widt, H. Hilkmann & W. J. van Blitterswijk: Involvement of phosphatidylcholine-specific phospholipase C in platelet-derived growth factor-induced activation of the mitogen-activated protein kinase pathway in Rat-1 fibroblasts. J Biol Chem, 272, 11011-6 (1997)
doi:10.1074/jbc.272.17.11011

247. Ramoni, C., F. Spadaro, B. Barletta, M. L. Dupuis & F. Podo: Phosphatidylcholine-specific phospholipase C in mitogen-stimulated fibroblasts. Exp Cell Res, 299, 370-82 (2004)
doi:10.1016/j.yexcr.2004.05.037

248. Boyan, B. D., L. Wang, K. L. Wong, H. Jo & Z. Schwartz: Plasma membrane requirements for 1alpha,25 (OH)2D3 dependent PKC signaling in chondrocytes and osteoblasts. Steroids, 71, 286-90 (2006)
doi:10.1016/j.steroids.2005.09.018
PMid:20053388

249. Murphy, E. J., K. M. Brindle, C. J. Rorison, R. M. Dixon, B. Rajagopalan & G. K. Radda: Changes in phosphatidylethanolamine metabolism in regenerating rat liver as measured by 31P-NMR. Biochim Biophys Acta, 1135, 27-34 (1992)
doi:10.1016/0167-4889(92)90162-5
PMid:3521686

250. Hafez, M. M. & M. E. Costlow: Phosphatidylethanolamine turnover is an early event in the response of NB2 lymphoma cells to prolactin. Exp Cell Res, 184, 37-43 (1989)
doi:10.1016/0014-4827(89)90361-3
PMid:2674489

251. Sakagami, H., J. Aoki, Y. Natori, K. Nishikawa, Y. Kakehi, Y. Natori & H. Arai: Biochemical and molecular characterization of a novel choline-specific glycerophosphodiester phosphodiesterase belonging to the nucleotide pyrophosphatase/phosphodiesterase family. J Biol Chem, 280, 23084-93 (2005)
doi:10.1074/jbc.M413438200
PMid:15788404

252. Liu, B. & Y. A. Hannun: Inhibition of the neutral magnesium-dependent sphingomyelinase by glutathione. J Biol Chem, 272, 16281-7 (1997)
doi:10.1074/jbc.272.26.16281
PMid:9195931

253. Zhang, P., B. Liu, G. M. Jenkins, Y. A. Hannun & L. M. Obeid: Expression of neutral sphingomyelinase identifies a distinct pool of sphingomyelin involved in apoptosis. J Biol Chem, 272, 9609-12 (1997)
doi:10.1074/jbc.272.15.9609
PMid:9092485

254. Clarke, C. J. & Y. A. Hannun: Neutral sphingomyelinases and nSMase2: bridging the gaps. Biochim Biophys Acta, 1758, 1893-901 (2006)
doi:10.1016/j.bbamem.2006.06.025
PMid:16938269

255. Clarke, C. J., C. F. Snook, M. Tani, N. Matmati, N. Marchesini & Y. A. Hannun: The extended family of neutral sphingomyelinases. Biochemistry, 45, 11247-56 (2006)
doi:10.1021/bi061307z
PMid:16981685

256. Clarke, C. J., B. X. Wu & Y. A. Hannun: The neutral sphingomyelinase family: Identifying biochemical connections. Adv Enzyme Regul (In press)
PMCid:2877797

257. Wu, B. X., C. J. Clarke & Y. A. Hannun: Mammalian Neutral Sphingomyelinases: Regulation and roles in cell signaling responses. Neuromolecular Med, 12, 320-330
doi:10.1007/s12017-010-8120-z
PMid:20552297

258. Marchesini, N., C. Luberto & Y. A. Hannun: Biochemical properties of mammalian neutral sphingomyelinase 2 and its role in sphingolipid metabolism. J Biol Chem, 278, 13775-83 (2003)
doi:10.1074/jbc.M212262200
PMid:12566438

259. Majeska, R. J., D. L. Holwerda & R. E. Wuthier: Localization of phosphatidylserine in isolated chick epiphyseal cartilage matrix vesicles with trinitrobenzenesulfonate. Calcif Tissue Int, 27, 41-6 (1979)
doi:10.1007/BF02441159
PMid:111786

260. Tani, M. & Y. A. Hannun: Analysis of membrane topology of neutral sphingomyelinase 2. FEBS Lett, 581, 1323-8 (2007)
doi:10.1016/j.febslet.2007.02.046
PMid:17349629    PMCid:1868537

261. Tani, M. & Y. A. Hannun: Neutral sphingomyelinase 2 is palmitoylated on multiple cysteine residues. Role of palmitoylation in subcellular localization. J Biol Chem, 282, 10047-56 (2007)
doi:10.1074/jbc.M611249200
PMid:17272284

262. Marchesini, N., W. Osta, J. Bielawski, C. Luberto, L. M. Obeid & Y. A. Hannun: Role for mammalian neutral sphingomyelinase 2 in confluence-induced growth arrest of MCF7 cells. J Biol Chem, 279, 25101-11 (2004)
doi:10.1074/jbc.M313662200
PMid:15051724

263. Mikati, M. A., M. Zeinieh, R. A. Habib, J. El Hokayem, A. Rahmeh, M. El Sabban, J. Usta & G. Dbaibo: Changes in sphingomyelinases, ceramide, Bax, Bcl (2), and caspase-3 during and after experimental status epilepticus. Epilepsy Res, 81, 161-6 (2008)
doi:10.1016/j.eplepsyres.2008.05.009
PMid:18603412

264. Stoffel, W., B. Jenke, B. Holz, E. Binczek, R. H. Gunter, J. Knifka, J. Koebke & A. Niehoff: Neutral sphingomyelinase (SMPD3) deficiency causes a novel form of chondrodysplasia and dwarfism that is rescued by Col2A1-driven smpd3 transgene expression. Am J Pathol, 171, 153-61 (2007)
doi:10.2353/ajpath.2007.061285
PMid:17591962    PMCid:1941606

265. Aubin, I., C. P. Adams, S. Opsahl, D. Septier, C. E. Bishop, N. Auge, R. Salvayre, A. Negre-Salvayre, M. Goldberg, J. L. Guenet & C. Poirier: A deletion in the gene encoding sphingomyelin phosphodiesterase 3 (Smpd3) results in osteogenesis and dentinogenesis imperfecta in the mouse. Nat Genet, 37, 803-5 (2005)
doi:10.1038/ng1603
PMid:16025116

266. Cotmore, J. M., G. Nichols, Jr. & R. E. Wuthier: Phospholipid-calcium phosphate complex: enhanced calcium migration in the presence of phosphate. Science, 172, 1339-41 (1971)
doi:10.1126/science.172.3990.1339
PMid:5580215

267. Wuthier, R. E. & J. W. Cummins: In vitro incorporation of (3H)serine into phospholipids of proliferating and calcifying epiphyseal cartilage and liver. Biochim Biophys Acta, 337, 50-9 (1974)
PMid:4373052

268. Wuthier, R. E., F. H. Wians, Jr., M. S. Giancola & S. S. Dragic: In vitro biosynthesis of phospholipids by chondrocytes and matrix vesicles of epiphyseal cartilage. Biochemistry, 17, 1431-6 (1978)
doi:10.1021/bi00601a011
PMid:417732

269. Vance, J. E.: Phosphatidylserine and phosphatidylethanolamine in mammalian cells: two metabolically related aminophospholipids. J Lipid Res, 49, 1377-87 (2008)
doi:10.1194/jlr.R700020-JLR200
PMid:18204094

270. Vance, J. E.: Molecular and cell biology of phosphatidylserine and phosphatidylethanolamine metabolism. Prog Nucleic Acid Res Mol Biol, 75, 69-111 (2003)
doi:10.1016/S0079-6603(03)75003-X

271. Matthews, J. L., J. H. Martin, H. W. Sampson, A. S. Kunin & J. H. Roan: Mitochondrial granules in the normal and rachitic rat epiphysis. Calcif Tissue Res, 5, 91-9 (1970)
doi:10.1007/BF02017539
PMid:5310405

272. Martin, J. H. & J. L. Matthews: Mitochondrial granules in chondrocytes. Calcif Tissue Res, 3, 184-93 (1969)
doi:10.1007/BF02058661
PMid:5769904

273. Matthews, J. L., J. H. Martin & E. J. Collins: Metabolism of radioactive calcium by cartilage. Clin Orthop Relat Res, 58, 213-23 (1968)
doi:10.1097/00003086-196805000-00023

274. Matthews, J. L. & J. H. Martin: Intracellular transport of calcium and its relationship to homeostasis and mineralization. An electron microscope study. Am J Med, 50, 589-97 (1971)
doi:10.1016/0002-9343(71)90114-8

275. Matthews, J. L., J. H. Martin, J. A. Lynn & E. J. Collins: Calcium incorporation in the developing cartilaginous epiphysis. Calcif Tissue Res, 1, 330-6 (1968)
doi:10.1007/BF02008105

276. Stone, S. J., Z. Cui & J. E. Vance: Cloning and expression of mouse liver phosphatidylserine synthase-1 cDNA. Overexpression in rat hepatoma cells inhibits the CDP-ethanolamine pathway for phosphatidylethanolamine biosynthesis. J Biol Chem, 273, 7293-302 (1998)
doi:10.1074/jbc.273.13.7293
PMid:9516423

277. Stone, S. J. & J. E. Vance: Cloning and expression of murine liver phosphatidylserine synthase (PSS)-2: differential regulation of phospholipid metabolism by PSS1 and PSS2. Biochem J, 342 (Pt 1), 57-64 (1999)
doi:10.1042/0264-6021:3420057
PMid:10432300    PMCid:1220436

278. Kuge, O., K. Hasegawa, T. Ohsawa, K. Saito & M. Nishijima: Purification and characterization of Chinese hamster phosphatidylserine synthase 2. J Biol Chem, 278, 42692-8 (2003)
doi:10.1074/jbc.M307270200
PMid:12912985

279. Saito, K., M. Nishijima & O. Kuge: Genetic evidence that phosphatidylserine synthase II catalyzes the conversion of phosphatidylethanolamine to phosphatidylserine in Chinese hamster ovary cells. J Biol Chem, 273, 17199-205 (1998)
doi:10.1074/jbc.273.27.17199
PMid:9642289

280. Voelker, D. R.: Interorganelle transport of aminoglycerophospholipids. Biochim Biophys Acta, 1486, 97-107 (2000)
PMid:10856716

281. Daleke, D. L.: Phospholipid flippases. J Biol Chem, 282, 821-5 (2007)
doi:10.1074/jbc.R600035200
PMid:17130120

282. Sahu, S. K., S. N. Gummadi, N. Manoj & G. K. Aradhyam: Phospholipid scramblases: an overview. Arch Biochem Biophys, 462, 103-14 (2007)
doi:10.1016/j.abb.2007.04.002
PMid:17481571

283. Wolfs, J. L., P. Comfurius, J. T. Rasmussen, J. F. Keuren, T. Lindhout, R. F. Zwaal & E. M. Bevers: Activated scramblase and inhibited aminophospholipid translocase cause phosphatidylserine exposure in a distinct platelet fraction. Cell Mol Life Sci, 62, 1514-25 (2005)
doi:10.1007/s00018-005-5099-y
PMid:15971000

284. Zwaal, R. F., P. Comfurius & E. M. Bevers: Surface exposure of phosphatidylserine in pathological cells. Cell Mol Life Sci, 62, 971-88 (2005)
doi:10.1007/s00018-005-4527-3
PMid:15761668

285. Zhou, Q., J. Zhao, J. G. Stout, R. A. Luhm, T. Wiedmer & P. J. Sims: Molecular cloning of human plasma membrane phospholipid scramblase. A protein mediating transbilayer movement of plasma membrane phospholipids. J Biol Chem, 272, 18240-4 (1997)
doi:10.1074/jbc.272.29.18240
PMid:9218461

286. Contreras, F. X., L. Sanchez-Magraner, A. Alonso & F. M. Goni: Transbilayer (flip-flop) lipid motion and lipid scrambling in membranes. FEBS Lett (2009)

287. Damek-Poprawa, M., E. Golub, L. Otis, G. Harrison, C. Phillips & K. Boesze-Battaglia: Chondrocytes utilize a cholesterol-dependent lipid translocator to externalize phosphatidylserine. Biochemistry, 45, 3325-36 (2006)
doi:10.1021/bi0515927
PMid:16519527

288. Wu, L. N., T. Yoshimori, B. R. Genge, G. R. Sauer, T. Kirsch, Y. Ishikawa & R. E. Wuthier: Characterization of the nucleational core complex responsible for mineral induction by growth plate cartilage matrix vesicles. J Biol Chem, 268, 25084-94 (1993)
PMid:8227072

289. Anderson, H. C., H. H. Hsu, D. C. Morris, K. N. Fedde & M. P. Whyte: Matrix vesicles in osteomalacic hypophosphatasia bone contain apatite-like mineral crystals. Am J Pathol, 151, 1555-61 (1997)
PMid:9403706    PMCid:1858375

290. Cotter, G., S. Verhaegen, M. Clynes & K. Kavanagh: Membrane changes associated with the early stages of apoptosis in HEp-2 cells decrease susceptibility to adherence by Candida albicans. J Med Vet Mycol, 35, 219-24 (1997)
doi:10.1080/02681219780001171

291. Jefferies, D., B. Houston, D. Lester, C. C. Whitehead, B. H. Thorp, M. Botman & C. Farquharson: Expression patterns of chondrocyte genes cloned by differential display in tibial dyschondroplasia. Biochim Biophys Acta, 1501, 180-8 (2000)
PMid:10838191

292. Seulberger, H., F. Lottspeich & W. Risau: The inducible blood-brain barrier specific molecule HT7 is a novel immunoglobulin-like cell surface glycoprotein. Embo J, 9, 2151-8 (1990)
PMid:2357963    PMCid:551936

293. Kang, J. S., M. Gao, J. L. Feinleib, P. D. Cotter, S. N. Guadagno & R. S. Krauss: CDO: an oncogene-, serum-, and anchorage-regulated member of the Ig/fibronectin type III repeat family. J Cell Biol, 138, 203-13 (1997)
doi:10.1083/jcb.138.1.203
PMid:9214393    PMCid:2139939

294. Wegorzewska, M., R. S. Krauss & J. S. Kang: Overexpression of the immunoglobulin superfamily members CDO and BOC enhances differentiation of the human rhabdomyosarcoma cell line RD. Mol Carcinog, 37, 1-4 (2003)
doi:10.1002/mc.10121
PMid:12720294

295. Durr, J., S. Goodman, A. Potocnik, H. von der Mark & K. von der Mark: Localization of beta 1-integrins in human cartilage and their role in chondrocyte adhesion to collagen and fibronectin. Exp Cell Res, 207, 235-44 (1993)
doi:10.1006/excr.1993.1189
PMid:7688313

296. Enomoto, M., P. S. Leboy, A. S. Menko & D. Boettiger: Beta 1 integrins mediate chondrocyte interaction with type I collagen, type II collagen, and fibronectin. Exp Cell Res, 205, 276-85 (1993)
doi:10.1006/excr.1993.1087
PMid:8387015

297. Loeser, R. F.: Chondrocyte integrin expression and function. Biorheology, 37, 109-16 (2000)
PMid:10912183

298. Shakibaei, M.: Integrin expression on epiphyseal mouse chondrocytes in monolayer culture. Histol Histopathol, 10, 339-49 (1995)
PMid:7599432

299. Shakibaei, M., B. Zimmermann & H. J. Merker: Changes in integrin expression during chondrogenesis in vitro: an immunomorphological study. J Histochem Cytochem, 43, 1061-9 (1995)
PMid:7560884

300. Wang, Y., F. Middleton, J. A. Horton, L. Reichel, C. E. Farnum & T. A. Damron: Microarray analysis of proliferative and hypertrophic growth plate zones identifies differentiation markers and signal pathways. Bone, 35, 1273-93 (2004)
PMid:15589209

301. Cecil, D. L., C. T. Appleton, M. D. Polewski, J. S. Mort, A. M. Schmidt, A. Bendele, F. Beier & R. Terkeltaub: The pattern recognition receptor CD36 is a chondrocyte hypertrophy marker associated with suppression of catabolic responses and promotion of repair responses to inflammatory stimuli. J Immunol, 182, 5024-31 (2009)
doi:10.4049/jimmunol.0803603
PMid:19342682    PMCid:2698125

302. Hubbard, H. L.: Characterization of the extracellular fluid of chicken epiphyseal growth plate cartilage: Studies on metabolites and electrolytes. Department of Chemistry & Biochemistry, University of South Carolina, 1-125 (1984).

303. Wu, L. N., Y. Guo, B. R. Genge, Y. Ishikawa & R. E. Wuthier: Transport of inorganic phosphate in primary cultures of chondrocytes isolated from the tibial growth plate of normal adolescent chickens. J Cell Biochem, 86, 475-89 (2002)
doi:10.1002/jcb.10240
PMid:12210754

304. Cameron, T. L., D. Belluoccio, P. G. Farlie, B. Brachvogel & J. F. Bateman: Global comparative transcriptome analysis of cartilage formation in vivo. BMC Dev Biol, 9, 20 (2009)
doi:10.1186/1471-213X-9-20
PMid:19272164    PMCid:2662817

305. Phan, M. N., H. A. Leddy, B. J. Votta, S. Kumar, D. S. Levy, D. B. Lipshutz, S. H. Lee, W. Liedtke & F. Guilak: Functional characterization of TRPV4 as an osmotically sensitive ion channel in porcine articular chondrocytes. Arthritis Rheum, 60, 3028-37 (2009)
doi:10.1002/art.24799

306. Wu, L. N., M. G. Wuthier, B. R. Genge & R. E. Wuthier: In situ levels of intracellular Ca2+ and pH in avian growth plate cartilage. Clin Orthop Relat Res310-24 (1997)
PMid:9020233

307. Chalothorn, D., H. Zhang, J. E. Smith, J. C. Edwards & J. E. Faber: Chloride intracellular channel-4 is a determinant of native collateral formation in skeletal muscle and brain. Circ Res, 105, 89-98 (2009)
doi:10.1161/CIRCRESAHA.109.197145
PMid:19478202    PMCid:2729269

308. Tung, J. J., O. Hobert, M. Berryman & J. Kitajewski: Chloride intracellular channel 4 is involved in endothelial proliferation and morphogenesis in vitro. Angiogenesis, 12, 209-20 (2009)
doi:10.1007/s10456-009-9139-3
PMid:19247789

309. Suh, K. S., M. Mutoh, K. Nagashima, E. Fernandez-Salas, L. E. Edwards, D. D. Hayes, J. M. Crutchley, K. G. Marin, R. A. Dumont, J. M. Levy, C. Cheng, S. Garfield & S. H. Yuspa: The organellular chloride channel protein CLIC4/mtCLIC translocates to the nucleus in response to cellular stress and accelerates apoptosis. J Biol Chem, 279, 4632-41 (2004)
doi:10.1074/jbc.M311632200
PMid:14610078

310. Kopito, R. R. & H. F. Lodish: Primary structure and transmembrane orientation of the murine anion exchange protein. Nature, 316, 234-8 (1985)
doi:10.1038/316234a0
PMid:2410791

311. Kopito, R. R. & H. F. Lodish: Structure of the murine anion exchange protein. J Cell Biochem, 29, 1-17 (1985)
doi:10.1002/jcb.240290102
PMid:3840489

312. Chandra, D., G. Choy, P. T. Daniel & D. G. Tang: Bax-dependent regulation of Bak by voltage-dependent anion channel 2. J Biol Chem, 280, 19051-61 (2005)
doi:10.1074/jbc.M501391200
PMid:15757910

313. Premkumar, A. & R. Simantov: Mitochondrial voltage-dependent anion channel is involved in dopamine-induced apoptosis. J Neurochem, 82, 345-52 (2002)
doi:10.1046/j.1471-4159.2002.00966.x
PMid:12124435

314. Rajpurohit, R., K. Mansfield, K. Ohyama, D. Ewert & I. M. Shapiro: Chondrocyte death is linked to development of a mitochondrial membrane permeability transition in the growth plate. J Cell Physiol, 179, 287-96 (1999)
doi:10.1002/(SICI)1097-4652(199906)179:3<287::AID-JCP6>3.0.CO;2-T

315. Jansen, S., M. Pantaleon & P. L. Kaye: Characterization and regulation of monocarboxylate cotransporters Slc16a7 and Slc16a3 in preimplantation mouse embryos. Biol Reprod, 79, 84-92 (2008)
doi:10.1095/biolreprod.107.066811
PMid:18385447

316. Graham, C., I. Gatherar, I. Haslam, M. Glanville & N. L. Simmons: Expression and localization of monocarboxylate transporters and sodium/proton exchangers in bovine rumen epithelium. Am J Physiol Regul Integr Comp Physiol, 292, R997-1007 (2007)
doi:10.1152/ajpregu.00343.2006
PMid:17008462

317. Sauer, G. R., H. D. Adkisson, B. R. Genge & R. E. Wuthier: Regulatory effect of endogenous zinc and inhibitory action of toxic metal ions on calcium accumulation by matrix vesicles in vitro. Bone Miner, 7, 233-44 (1989)
doi:10.1016/0169-6009(89)90080-9

318. Blandford, N. R., G. R. Sauer, B. R. Genge, L. N. Wu & R. E. Wuthier: Modeling of matrix vesicle biomineralization using large unilamellar vesicles. J Inorg Biochem, 94, 14-27 (2003)
doi:10.1016/S0162-0134(02)00629-3

319. Messler, H. H., W. Koch & K. J. Munzenberg: Analogous effects of organic calcium antagonists and magnesium on the epiphyseal growth plate. Clin Orthop Relat Res135-41 (1990)
PMid:2394040

320. Duriez, J., B. Flautre, M. C. Blary & P. Hardouin: Effects of the calcium channel blocker nifedipine on epiphyseal growth plate and bone turnover: a study in rabbit. Calcif Tissue Int, 52, 120-4 (1993)
doi:10.1007/BF00308320
PMid:8443687

321. Mancilla, E. E., M. Galindo, B. Fertilio, M. Herrera, K. Salas, H. Gatica & A. Goecke: L-type calcium channels in growth plate chondrocytes participate in endochondral ossification. J Cell Biochem, 101, 389-98 (2007)
doi:10.1002/jcb.21183
PMid:17243114

322. Adams, B. A., T. Tanabe & K. G. Beam: Ca2+ current activation rate correlates with alpha 1 subunit density. Biophys J, 71, 156-62 (1996)
doi:10.1016/S0006-3495(96)79212-3

323. Yamada, Y., K. Masuda, Q. Li, Y. Ihara, A. Kubota, T. Miura, K. Nakamura, Y. Fujii, S. Seino & Y. Seino: The structures of the human calcium channel alpha 1 subunit (CACNL1A2) and beta subunit (CACNLB3) genes. Genomics, 27, 312-9 (1995)
doi:10.1006/geno.1995.1048
PMid:7557998

324. Fallon, J. L., M. R. Baker, L. Xiong, R. E. Loy, G. Yang, R. T. Dirksen, S. L. Hamilton & F. A. Quiocho: Crystal structure of dimeric cardiac L-type calcium channel regulatory domains bridged by Ca2+* calmodulins. Proc Natl Acad Sci U S A, 106, 5135-40 (2009)
doi:10.1073/pnas.0807487106
PMid:19279214    PMCid:2654391

325. Kwok, T. C., K. Hui, W. Kostelecki, N. Ricker, G. Selman, Z. P. Feng & P. J. Roy: A genetic screen for dihydropyridine (DHP)-resistant worms reveals new residues required for DHP-blockage of mammalian calcium channels. PLoS Genet, 4, e1000067 (2008)
doi:10.1371/journal.pgen.1000067
PMid:18464914    PMCid:2362100

326. Dolphin, A. C.: A short history of voltage-gated calcium channels. Br J Pharmacol, 147 Suppl 1, S56-62 (2006)
doi:10.1038/sj.bjp.0706442
PMid:16402121    PMCid:1760727

327. Zanetti, M., R. Camerotto, D. Romeo & B. De Bernard: Active extrusion of Ca2+ from epiphysial chondrocytes of normal and rachitic chickens. Biochem J, 202, 303-7 (1982)
PMid:6807290    PMCid:1158112

328. Akisaka, T. & C. V. Gay: Ultrastructural localization of calcium-activated adenosine triphosphatase (Ca2+-ATPase) in growth-plate cartilage. J Histochem Cytochem, 33, 925-32 (1985)
PMid:3160764

329. Gunter, T. E., M. J. Zuscik, J. E. Puzas, K. K. Gunter & R. N. Rosier: Cytosolic free calcium concentrations in avian growth plate chondrocytes. Cell Calcium, 11, 445-57 (1990)
doi:10.1016/0143-4160(90)90077-8

330. Kirsch, T., B. Swoboda & H. Nah: Activation of annexin II and V expression, terminal differentiation, mineralization and apoptosis in human osteoarthritic cartilage. Osteoarthritis Cartilage, 8, 294-302 (2000)
doi:10.1053/joca.1999.0304
PMid:10903884

331. Wang, W., J. Xu & T. Kirsch: Annexin-mediated Ca2+ influx regulates growth plate chondrocyte maturation and apoptosis. J Biol Chem, 278, 3762-9 (2003)
doi:10.1074/jbc.M208868200
PMid:12446691

332. Genge, B. R., L. N. Wu & R. E. Wuthier: Kinetic analysis of mineral formation during in vitro modeling of matrix vesicle mineralization: effect of annexin A5, phosphatidylserine, and type II collagen. Anal Biochem, 367, 159-66 (2007)
doi:10.1016/j.ab.2007.04.029
PMid:17585866

333. Virkki, L. V., J. Biber, H. Murer & I. C. Forster: Phosphate transporters: a tale of two solute carrier families. Am J Physiol Renal Physiol, 293, F643-54 (2007)
doi:10.1152/ajprenal.00228.2007
PMid:17581921

334. Montessuit, C., J. Caverzasio & J. P. Bonjour: Characterization of a Pi transport system in cartilage matrix vesicles. Potential role in the calcification process. J Biol Chem, 266, 17791-7 (1991)
PMid:1833387

335. Montessuit, C., J. P. Bonjour & J. Caverzasio: Pi transport regulation by chicken growth plate chondrocytes. Am J Physiol, 267, E24-31 (1994)

336. Guo, Y.: Characterization of phosphate transport, and isolation of cDNAs encoding the mitochondrial Pi transporter and pyrophosphate transporter from chicken growth plate chondrocytes. Department of Chemistry and Biochemistry, University of South Carolina, 1-188 (2002).

337. Kavanaugh, M. P., D. G. Miller, W. Zhang, W. Law, S. L. Kozak, D. Kabat & A. D. Miller: Cell-surface receptors for gibbon ape leukemia virus and amphotropic murine retrovirus are inducible sodium-dependent phosphate symporters. Proc Natl Acad Sci U S A, 91, 7071-5 (1994)
doi:10.1073/pnas.91.15.7071

338. Giral, H., Y. Caldas, E. Sutherland, P. Wilson, S. Breusegem, N. Barry, J. Blaine, T. Jiang, X. X. Wang & M. Levi: Regulation of rat intestinal Na-dependent phosphate transporters by dietary phosphate. Am J Physiol Renal Physiol, 297, F1466-75 (2009)
doi:10.1152/ajprenal.00279.2009
PMid:19675183    PMCid:2781338

339. Howell, D. S., J. C. Pita, J. F. Marquez & J. E. Madruga: Partition of calcium, phosphate, and protein in the fluid phase aspirated at calcifying sites in epiphyseal cartilage. J Clin Invest, 47, 1121-32 (1968)
PMid:5645857    PMCid:297264

340. Grootjans, J. J., P. Zimmermann, G. Reekmans, A. Smets, G. Degeest, J. Durr & G. David: Syntenin, a PDZ protein that binds syndecan cytoplasmic domains. Proc Natl Acad Sci U S A, 94, 13683-8 (1997)
doi:10.1073/pnas.94.25.13683

341. Zimmermann, P., D. Tomatis, M. Rosas, J. Grootjans, I. Leenaerts, G. Degeest, G. Reekmans, C. Coomans & G. David: Characterization of syntenin, a syndecan-binding PDZ protein, as a component of cell adhesion sites and microfilaments. Mol Biol Cell, 12, 339-50 (2001)
PMid:11179419    PMCid:30947

342. Stier, S., G. Totzke, E. Grunewald, T. Neuhaus, S. Fronhoffs, A. Sachinidis, H. Vetter, K. Schulze-Osthoff & Y. Ko: Identification of syntenin and other TNF-inducible genes in human umbilical arterial endothelial cells by suppression subtractive hybridization. FEBS Lett, 467, 299-304 (2000)
doi:10.1016/S0014-5793(00)01177-7

343. Yuan, K., T. M. Hong, J. J. Chen, W. H. Tsai & M. T. Lin: Syndecan-1 up-regulated by ephrinB2/EphB4 plays dual roles in inflammatory angiogenesis. Blood, 104, 1025-33 (2004)
doi:10.1182/blood-2003-09-3334
PMid:15126321

344. Kirsch, T., E. Koyama, M. Liu, E. E. Golub & M. Pacifici: Syndecan-3 is a selective regulator of chondrocyte proliferation. J Biol Chem, 277, 42171-7 (2002)
doi:10.1074/jbc.M207209200
PMid:12194984

345. Pacifici, M., T. Shimo, C. Gentili, T. Kirsch, T. A. Freeman, M. Enomoto-Iwamoto, M. Iwamoto & E. Koyama: Syndecan-3: a cell-surface heparan sulfate proteoglycan important for chondrocyte proliferation and function during limb skeletogenesis. J Bone Miner Metab, 23, 191-9 (2005)
doi:10.1007/s00774-004-0584-1
PMid:15838620

346. Beekman, J. M. & P. J. Coffer: The ins and outs of syntenin, a multifunctional intracellular adaptor protein. J Cell Sci, 121, 1349-55 (2008)
doi:10.1242/jcs.026401
PMid:18434645

347. Grootjans, J. J., G. Reekmans, H. Ceulemans & G. David: Syntenin-syndecan binding requires syndecan-synteny and the co-operation of both PDZ domains of syntenin. J Biol Chem, 275, 19933-41 (2000)
doi:10.1074/jbc.M002459200
PMid:10770943

348. San Antonio, J. D., M. J. Karnovsky, S. Gay, R. D. Sanderson & A. D. Lander: Interactions of syndecan-1 and heparin with human collagens. Glycobiology, 4, 327-32 (1994)
doi:10.1093/glycob/4.3.327

349. Sulka, B., H. Lortat-Jacob, R. Terreux, F. Letourneur & P. Rousselle: Tyrosine dephosphorylation of the syndecan-1 PDZ binding domain regulates syntenin-1 recruitment. J Biol Chem, 284, 10659-71 (2009)
doi:10.1074/jbc.M807643200
PMid:19228696    PMCid:2667753

350. Hagiwara, H., A. Inoue, S. Nakajo, K. Nakaya, S. Kojima & S. Hirose: Inhibition of proliferation of chondrocytes by specific receptors in response to retinoids. Biochem Biophys Res Commun, 222, 220-4 (1996)
doi:10.1006/bbrc.1996.0725
PMid:8670186

351. Pouyssegur, J., J. C. Chambard, G. L'Allemain, I. Magnaldo & K. Seuwen: Transmembrane signalling pathways initiating cell growth in fibroblasts. Philos Trans R Soc Lond B Biol Sci, 320, 427-36 (1988)
doi:10.1098/rstb.1988.0086

352. Roche, S., J. McGlade, M. Jones, G. D. Gish, T. Pawson & S. A. Courtneidge: Requirement of phospholipase C gamma, the tyrosine phosphatase Syp and the adaptor proteins Shc and Nck for PDGF-induced DNA synthesis: evidence for the existence of Ras-dependent and Ras-independent pathways. Embo J, 15, 4940-8 (1996)
PMid:8890167    PMCid:452231

353. Dunn, M. J.: Red blood cell calcium and magnesium: effects upon sodium and potassium transport and cellular morphology. Biochim Biophys Acta, 352, 97-116 (1974)
doi:10.1016/0005-2736(74)90182-5

354. Prentki, M., C. Chaponnier, B. Jeanrenaud & G. Gabbiani: Actin microfilaments, cell shape, and secretory processes in isolated rat hepatocytes. Effect of phalloidin and cytochalasin D. J Cell Biol, 81, 592-607 (1979)
doi:10.1083/jcb.81.3.592
PMid:572368

355. Burgess, D. R.: Reactivation of intestinal epithelial cell brush border motility: ATP-dependent contraction via a terminal web contractile ring. J Cell Biol, 95, 853-63 (1982)
doi:10.1083/jcb.95.3.853
PMid:7153249

356. Lu, J., G. Lian, R. Lenkinski, A. De Grand, R. R. Vaid, T. Bryce, M. Stasenko, A. Boskey, C. Walsh & V. Sheen: Filamin B mutations cause chondrocyte defects in skeletal development. Hum Mol Genet, 16, 1661-75 (2007)
doi:10.1093/hmg/ddm114
PMid:17510210

357. Bicknell, L. S., T. Morgan, L. Bonafe, M. W. Wessels, M. G. Bialer, P. J. Willems, D. H. Cohn, D. Krakow & S. P. Robertson: Mutations in FLNB cause boomerang dysplasia. J Med Genet, 42, e43 (2005)
doi:10.1136/jmg.2004.029967
PMid:15994868    PMCid:1736093

358. Williams, R. C., Jr., C. Shah & D. Sackett: Separation of tubulin isoforms by isoelectric focusing in immobilized pH gradient gels. Anal Biochem, 275, 265-7 (1999)
doi:10.1006/abio.1999.4326
PMid:10552916

359. Nogales, E., K. H. Downing, L. A. Amos & J. Lowe: Tubulin and FtsZ form a distinct family of GTPases. Nat Struct Biol, 5, 451-8 (1998)
doi:10.1038/nsb0698-451
PMid:9628483

360. Heald, R. & E. Nogales: Microtubule dynamics. J Cell Sci, 115, 3-4 (2002)
PMid:11801717

361. Howard, J. & A. A. Hyman: Dynamics and mechanics of the microtubule plus end. Nature, 422, 753-8 (2003)
doi:10.1038/nature01600
PMid:12700769

362. Hoeflich, K. P. & M. Ikura: Radixin: cytoskeletal adopter and signaling protein. Int J Biochem Cell Biol, 36, 2131-6 (2004)
doi:10.1016/j.biocel.2003.11.018

363. Sun, H. Q., M. Yamamoto, M. Mejillano & H. L. Yin: Gelsolin, a multifunctional actin regulatory protein. J Biol Chem, 274, 33179-82 (1999)
doi:10.1074/jbc.274.47.33179
PMid:10559185

364. Kiselar, J. G., P. A. Janmey, S. C. Almo & M. R. Chance: Visualizing the Ca2+-dependent activation of gelsolin by using synchrotron footprinting. Proc Natl Acad Sci U S A, 100, 3942-7 (2003)
doi:10.1073/pnas.0736004100
PMid:12655044    PMCid:153027

365. Hai, C. M. & Z. Gu: Caldesmon phosphorylation in actin cytoskeletal remodeling. Eur J Cell Biol, 85, 305-9 (2006)
doi:10.1016/j.ejcb.2005.08.008
PMid:16546574

366. Spinardi, L. & P. C. Marchisio: Podosomes as smart regulators of cellular adhesion. Eur J Cell Biol, 85, 191-4 (2006)
doi:10.1016/j.ejcb.2005.08.005
PMid:16546561

367. Koya, R. C., H. Fujita, S. Shimizu, M. Ohtsu, M. Takimoto, Y. Tsujimoto & N. Kuzumaki: Gelsolin inhibits apoptosis by blocking mitochondrial membrane potential loss and cytochrome c release. J Biol Chem, 275, 15343-9 (2000)
doi:10.1074/jbc.275.20.15343
PMid:10809769

368. Lin, C. S., T. Park, Z. P. Chen & J. Leavitt: Human plastin genes. Comparative gene structure, chromosome location, and differential expression in normal and neoplastic cells. J Biol Chem, 268, 2781-92 (1993)
PMid:8428952

369. Wuthier, R. E.: Purification of lipids from nonlipid contaminants on Sephadex bead columns. J Lipid Res, 7, 558-61 (1966)
PMid:5965298

370. Wuthier, R. E.: Two-dimensional chromatography on silica gel-loaded paper for the microanalysis of polar lipids. J Lipid Res, 7, 544-50 (1966)
PMid:5965296

371. Wuthier, R. E.: Zonal analysis of phospholipids in the epiphyseal cartilage and bone of normal and rachitic chickens and pigs. Calcif Tissue Res, 8, 36-53 (1971)
doi:10.1007/BF02010121
PMid:5135577

doi:10.1007/BF02010120
PMid:5135576

372. Boskey, A. L. & A. S. Posner: In vitro nucleation of hydroxyapatite by a bone calcium-phospholipid-phosphate complex. Calcif Tissue Res, 22 Suppl, 197-201 (1977)
doi:10.1007/BF02064064

373. Holwerda, D. L., P. D. Ellis & R. E. Wuthier: Carbon-13 and phosphorus-31 nuclear magnetic resonance studies on interaction of calcium with phosphatidylserine. Biochemistry, 20, 418-23 (1981)
doi:10.1021/bi00505a030
PMid:7470491

374. Gordesky, S. E. & G. V. Marinetti: The asymetric arrangement of phospholipids in the human erythrocyte membrane. Biochem Biophys Res Commun, 50, 1027-31 (1973)
doi:10.1016/0006-291X(73)91509-X

375. Rawyler, A., P. H. van der Schaft, B. Roelofsen & J. A. Op den Kamp: Phospholipid localization in the plasma membrane of Friend erythroleukemic cells and mouse erythrocytes. Biochemistry, 24, 1777-83 (1985)
doi:10.1021/bi00328a031
PMid:3859334

376. Hullin, F., M. J. Bossant & N. Salem, Jr.: Aminophospholipid molecular species asymmetry in the human erythrocyte plasma membrane. Biochim Biophys Acta, 1061, 15-25 (1991)
doi:10.1016/0005-2736(91)90263-8

377. Boskey, A. L., W. Ullrich, L. Spevak & H. Gilder: Persistence of complexed acidic phospholipids in rapidly mineralizing tissues is due to affinity for mineral and resistance to hydrolytic attack: in vitro data. Calcif Tissue Int, 58, 45-51 (1996)
doi:10.1007/BF02509545
PMid:8825238

378. Gerke, V. & K. Weber: Calcium-dependent conformational changes in the 36-kDa subunit of intestinal protein I related to the cellular 36-kDa target of Rous sarcoma virus tyrosine kinase. J Biol Chem, 260, 1688-95 (1985)
PMid:2981869

379. Glenney, J. R., Jr.: Phosphorylation of p36 in vitro with pp60src. Regulation by Ca2+ and phospholipid. FEBS Lett, 192, 79-82 (1985)
doi:10.1016/0014-5793(85)80047-8

380. Hargest, T. E., C. V. Gay, H. Schraer & A. J. Wasserman: Vertical distribution of elements in cells and matrix of epiphyseal growth plate cartilage determined by quantitative electron probe analysis. J Histochem Cytochem, 33, 275-86 (1985)
PMid:3980981

381. Long, K. J. & K. B. Walsh: A calcium-activated potassium channel in growth plate chondrocytes: regulation by protein kinase A. Biochem Biophys Res Commun, 201, 776-81 (1994)
doi:10.1006/bbrc.1994.1768
PMid:8003014

382. Brighton, C. T. & R. M. Hunt: Mitochondrial calcium and its role in calcification. Histochemical localization of calcium in electron micrographs of the epiphyseal growth plate with K-pyroantimonate. Clin Orthop Relat Res, 100, 406-16 (1974)
doi:10.1097/00003086-197405000-00057
PMid:4134194

383. Brighton, C. T. & R. M. Hunt: Histochemical localization of calcium in growth plate mitochondria and matrix vesicles. Fed Proc, 35, 143-7 (1976)
PMid:1248647

384. Brighton, C. T. & R. M. Hunt: The role of mitochondria in growth plate calcification as demonstrated in a rachitic model. J Bone Joint Surg Am, 60, 630-9 (1978)
PMid:681381

385. Sauer, G. R., L. N. Wu, M. Iijima & R. E. Wuthier: The influence of trace elements on calcium phosphate formation by matrix vesicles. J Inorg Biochem, 65, 57-65 (1997)
doi:10.1016/S0162-0134(96)00080-3

386. Eanes, E. D.: Thermochemical studies on amorphous calcium phosphate. Calcif Tissue Res, 5, 133-45 (1970)
doi:10.1007/BF02017543
PMid:5423622

387. Sauer, G. R. & R. E. Wuthier: Fourier transform infrared characterization of mineral phases formed during induction of mineralization by collagenase-released matrix vesicles in vitro. J Biol Chem, 263, 13718-24 (1988)
PMid:2843533

388. Sauer, G. R., W. B. Zunic, J. R. Durig & R. E. Wuthier: Fourier transform Raman spectroscopy of synthetic and biological calcium phosphates. Calcif Tissue Int, 54, 414-20 (1994)
doi:10.1007/BF00305529
PMid:8062160

389. Adkisson, H. D. t., F. S. Risener, Jr., P. P. Zarrinkar, M. D. Walla, W. W. Christie & R. E. Wuthier: Unique fatty acid composition of normal cartilage: discovery of high levels of n-9 eicosatrienoic acid and low levels of n-6 polyunsaturated fatty acids. Faseb J, 5, 344-53 (1991)
PMid:2001795

390. Brighton, C. T. & R. B. Heppenstall: Oxygen tension in zones of the epiphyseal plate, the metaphysis and diaphysis. An in vitro and in vivo study in rats and rabbits. J Bone Joint Surg Am, 53, 719-28 (1971)
PMid:5580029

391. Haselgrove, J. C., I. M. Shapiro & S. F. Silverton: Computer modeling of the oxygen supply and demand of cells of the avian growth cartilage. Am J Physiol, 265, C497-506 (1993)

392. Meyer, W. L. & A. S. Kunin: The inductive effect of rickets on glycolytic enzymes of rat epiphyseal cartilage and its reversal by vitamin D and phosphate. Arch Biochem Biophys, 129, 438-46 (1969)
doi:10.1016/0003-9861(69)90200-8

393. Sampson, H. W. & M. S. Cannon: Zonal analysis of metabolic profiles of articular-epiphyseal cartilage chondrocytes: a histochemical study. Histochem J, 18, 233-8 (1986)
doi:10.1007/BF01676232
PMid:3017899

394. Kim, J. K., J. C. Haselgrove & I. M. Shapiro: Measurement of metabolic events in the avian epiphyseal growth cartilage using a bioluminescence technique. J Histochem Cytochem, 41, 693-702 (1993)
PMid:8468450

395. Shapiro, I. M., E. E. Golub, B. Chance, C. Piddington, O. Oshima, O. C. Tuncay, P. Frasca & J. C. Haselgrove: Linkage between energy status of perivascular cells and mineralization of the chick growth cartilage. Dev Biol, 129, 372-9 (1988)
doi:10.1016/0012-1606(88)90384-3

396. Teixeira, C. C., K. Mansfield, C. Hertkorn, H. Ischiropoulos & I. M. Shapiro: Phosphate-induced chondrocyte apoptosis is linked to nitric oxide generation. Am J Physiol Cell Physiol, 281, C833-9 (2001)

397. Teixeira, C. C., H. Ischiropoulos, P. S. Leboy, S. L. Adams & I. M. Shapiro: Nitric oxide-nitric oxide synthase regulates key maturational events during chondrocyte terminal differentiation. Bone, 37, 37-45 (2005)
PMid:890550

398. Wuthier, R. E., R. J. Majeska & G. M. Collins: Biosynthesis of matrix vesicles in epiphyseal cartilage. I. In vivo incorporation of 32P orthophosphate into phospholipids of chondroxyte, membrane, and matrix vesicle fractions. Calcif Tissue Res, 23, 135-9 (1977)
doi:10.1007/BF02012778
PMid:7759559

399. Wu, L. N., Y. Ishikawa, G. R. Sauer, B. R. Genge, F. Mwale, H. Mishima & R. E. Wuthier: Morphological and biochemical characterization of mineralizing primary cultures of avian growth plate chondrocytes: evidence for cellular processing of Ca2+ and Pi prior to matrix mineralization. J Cell Biochem, 57, 218-37 (1995)
doi:10.1002/jcb.240570206
PMid:1962572

400. Shapiro, I. M., P. S. Leboy, T. Tokuoka, E. Forbes, K. DeBolt, S. L. Adams & M. Pacifici: Ascorbic acid regulates multiple metabolic activities of cartilage cells. Am J Clin Nutr, 54, 1209S-1213S (1991)
PMid:15869914

401. Boskey, A. L., D. Stiner, S. B. Doty & I. Binderman: Requirement of vitamin C for cartilage calcification in a differentiating chick limb-bud mesenchymal cell culture. Bone, 12, 277-82 (1991)
PMid:3868947

402. von der Mark, K., J. Mollenhauer, P. K. Muller & M. Pfaffle: Anchorin CII, a type II collagen-binding glycoprotein from chondrocyte membranes. Ann N Y Acad Sci, 460, 214-23 (1985)
doi:10.1111/j.1749-6632.1985.tb51169.x
PMid:6325473

403. Mollenhauer, J., J. A. Bee, M. A. Lizarbe & K. von der Mark: Role of anchorin CII, a 31,000-mol-wt membrane protein, in the interaction of chondrocytes with type II collagen. J Cell Biol, 98, 1572-9 (1984)
doi:10.1083/jcb.98.4.1572
PMid:8429379

404. Wuthier, R. E.: Involvement of cellular metabolism of calcium and phosphate in calcification of avian growth plate cartilage. J Nutr, 123, 301-9 (1993)
PMid:9230933

405. von der Mark, K. & J. Mollenhauer: Annexin V interactions with collagen. Cell Mol Life Sci, 53, 539-45 (1997)
doi:10.1007/s000180050069

406. Lehninger, A. L., C. S. Rossi & J. W. Greenawalt: Respiration-dependent accumulation of inorganic phosphate and Ca ions by rat liver mitochondria. Biochem Biophys Res Commun, 10, 444-8 (1963)
doi:10.1016/0006-291X(63)90377-2
PMid:14217896

407. Carafoli, E., C. S. Rossi & A. L. Lehninger: Cation and anion balance during active accumulation of Ca++ and Mg++ by isolated mitochondria. J Biol Chem, 239, 3055-61 (1964)

408. Shapiro, I. M. & N. H. Lee: Effects of Ca2+ on the respiratory activity of chondrocyte mitochondria. Arch Biochem Biophys, 170, 627-33 (1975)
doi:10.1016/0003-9861(75)90158-7
PMid:165026

409. Shapiro, I. M. & N. H. Lee: Calcium accumulation by chondrocyte mitochondria. Clin Orthop Relat Res323-9 (1975)
doi:10.1097/00003086-197501000-00044
PMid:6221186

410. Simmons, D. J., C. Arsenis, S. W. Whitson, S. E. Kahn, A. L. Boskey & N. Gollub: Mineralization of rat epiphyseal cartilage: a circadian rhythm. Miner Electrolyte Metab, 9, 28-37 (1983)

411. Matthews, J. L.: Role of Mitochondria in Calcification. In: Cell Mediated Calcification and Matrix Vesicles, Ed: S. Y. Ali, Elsevier Science Publishers B. V., Amsterdam, The Netherlands. 115-118 (1986).

412. Lee, N. H. & I. M. Shapiro: A method for isolating respiring mitochondria from epiphyseal cartilage. Anal Biochem, 58, 117-22 (1974)
doi:10.1016/0003-2697(74)90447-3
PMid:4155997

413. Lee, N. H. & I. M. Shapiro: Oxidative phosphorylation by chondrocyte mitochondria. Calcif Tissue Res, 16, 277-82 (1974)
doi:10.1007/BF02008235
PMid:999870

414. Shapiro, L. M., A. Burke & N. H. Lee: Heterogeneity of chondrocyte mitochondria. A study of the Ca2+ concentration and density banding characteristics of normal and rachitic cartilage. Biochim Biophys Acta, 451, 583-91 (1976)

415. Lehninger, A. L.: Role of phosphate and other proton-donating anions in respiration-coupled transport of Ca2+ by mitochondria. Proc Natl Acad Sci U S A, 71, 1520-4 (1974)
doi:10.1073/pnas.71.4.1520

416. Gunter, T. E. & D. R. Pfeiffer: Mechanisms by which mitochondria transport calcium. Am J Physiol, 258, C755-86 (1990)
PMid:11798034

417. Pfeiffer, D. R., T. E. Gunter, R. Eliseev, K. M. Broekemeier & K. K. Gunter: Release of Ca2+ from mitochondria via the saturable mechanisms and the permeability transition. IUBMB Life, 52, 205-12 (2001)
doi:10.1080/15216540152846019
PMid:357726

418. Lee, N. H. & I. M. Shapiro: Ca2+ transport by chondrocyte mitochondria of the epiphyseal growth plate. J Membr Biol, 41, 349-60 (1978)
doi:10.1007/BF01871999
PMid:7410387

419. Beatrice, M. C., J. W. Palmer & D. R. Pfeiffer: The relationship between mitochondrial membrane permeability, membrane potential, and the retention of Ca2+ by mitochondria. J Biol Chem, 255, 8663-71 (1980)
PMid:6177691

420. Beatrice, M. C., D. L. Stiers & D. R. Pfeiffer: Increased permeability of mitochondria during Ca2+ release induced by t-butyl hydroperoxide or oxalacetate. The effect of ruthenium red. J Biol Chem, 257, 7161-71 (1982)
PMid:10491114

421. Bernardi, P., L. Scorrano, R. Colonna, V. Petronilli & F. Di Lisa: Mitochondria and cell death. Mechanistic aspects and methodological issues. Eur J Biochem, 264, 687-701 (1999)
doi:10.1046/j.1432-1327.1999.00725.x
PMid:9171922

422. Hirsch, T., I. Marzo & G. Kroemer: Role of the mitochondrial permeability transition pore in apoptosis. Biosci Rep, 17, 67-76 (1997)
doi:10.1023/A:1027339418683

423. Lemasters, J. J., A. L. Nieminen, T. Qian, L. C. Trost, S. P. Elmore, Y. Nishimura, R. A. Crowe, W. E. Cascio, C. A. Bradham, D. A. Brenner & B. Herman: The mitochondrial permeability transition in cell death: A common mechanism in necrosis, apoptosis and autophagy. Biochim Biophys Acta, 1366, 177-96 (1998)
doi:10.1016/S0005-2728(98)00112-1
PMid:9239543

424. Zamzami, N., T. Hirsch, B. Dallaporta, P. X. Petit & G. Kroemer: Mitochondrial implication in accidental and programmed cell death: apoptosis and necrosis. J Bioenerg Biomembr, 29, 185-93 (1997)
doi:10.1023/A:1022694131572

425. Mansfield, K., R. Rajpurohit & I. M. Shapiro: Extracellular phosphate ions cause apoptosis of terminally differentiated epiphyseal chondrocytes. J Cell Physiol, 179, 276-86 (1999)
doi:10.1002/(SICI)1097-4652(199906)179:3<276::AID-JCP5>3.0.CO;2-#
PMid:12039849

426. Mansfield, K., C. C. Teixeira, C. S. Adams & I. M. Shapiro: Phosphate ions mediate chondrocyte apoptosis through a plasma membrane transporter mechanism. Bone, 28, 1-8 (2001)
PMid:2670944

427. Bowser, D. N., S. Petrou, R. G. Panchal, M. L. Smart & D. A. Williams: Release of mitochondrial Ca2+ via the permeability transition activates endoplasmic reticulum Ca2+ uptake. Faseb J, 16, 1105-7 (2002)
PMid:1985946

428. Ferreira, G. C., R. D. Pratt & P. L. Pedersen: Energy-linked anion transport. Cloning, sequencing, and characterization of a full length cDNA encoding the rat liver mitochondrial proton/phosphate symporter. J Biol Chem, 264, 15628-33 (1989)
PMid:8132488

429. Pratt, R. D., G. C. Ferreira & P. L. Pedersen: Mitochondrial phosphate transport. Import of the H+/Pi symporter and role of the presequence. J Biol Chem, 266, 1276-80 (1991)
PMid:2361480

430. Ferreira, G. C. & P. L. Pedersen: Phosphate transport in mitochondria: past accomplishments, present problems, and future challenges. J Bioenerg Biomembr, 25, 483-92 (1993)
doi:10.1007/BF01108405
PMid:9309682

431. Caverzasio, J., C. Montessuit & J. P. Bonjour: Stimulatory effect of insulin-like growth factor-1 on renal Pi transport and plasma 1,25-dihydroxyvitamin D3. Endocrinology, 127, 453-9 (1990)
doi:10.1210/endo-127-1-453

432. Halestrap, A. P., C. P. Connern, E. J. Griffiths & P. M. Kerr: Cyclosporin A binding to mitochondrial cyclophilin inhibits the permeability transition pore and protects hearts from ischaemia/reperfusion injury. Mol Cell Biochem, 174, 167-72 (1997)
doi:10.1023/A:1006879618176
PMid:9013575

433. Halestrap, A. P., P. M. Kerr, S. Javadov & K. Y. Woodfield: Elucidating the molecular mechanism of the permeability transition pore and its role in reperfusion injury of the heart. Biochim Biophys Acta, 1366, 79-94 (1998)
doi:10.1016/S0005-2728(98)00122-4
PMid:18667415

434. Halestrap, A. P., K. Y. Woodfield & C. P. Connern: Oxidative stress, thiol reagents, and membrane potential modulate the mitochondrial permeability transition by affecting nucleotide binding to the adenine nucleotide translocase. J Biol Chem, 272, 3346-54 (1997)
doi:10.1074/jbc.272.6.3346
PMid:19265700

435. Leung, A. W., P. Varanyuwatana & A. P. Halestrap: The mitochondrial phosphate carrier interacts with cyclophilin D and may play a key role in the permeability transition. J Biol Chem, 283, 26312-23 (2008)
doi:10.1074/jbc.M805235200
PMid:3213627

436. Halestrap, A. P.: What is the mitochondrial permeability transition pore? J Mol Cell Cardiol, 46, 821-31 (2009)
doi:10.1016/j.yjmcc.2009.02.021

437. Matsumoto, H., K. DeBolt & I. M. Shapiro: Adenine, guanine, and inosine nucleotides of chick growth cartilage: relationship between energy status and the mineralization process. J Bone Miner Res, 3, 347-52 (1988)
doi:10.1002/jbmr.5650030315
PMid:7593225

438. Pollesello, P., B. de Bernard, M. Grandolfo, S. Paoletti, F. Vittur & B. J. Kvam: Energy state of chondrocytes assessed by 31P-NMR studies of preosseous cartilage. Biochem Biophys Res Commun, 180, 216-22 (1991)
doi:10.1016/S0006-291X(05)81279-3
PMid:1400624

439. Hatori, M., C. C. Teixeira, K. Debolt, M. Pacifici & I. M. Shapiro: Adenine nucleotide metabolism by chondrocytes in vitro: Role of ATP in chondrocyte maturation and matrix mineralization. J Cell Physiol, 165, 468-74 (1995)
doi:10.1002/jcp.1041650304
PMid:3616560

440. Wu, L. N., B. R. Genge, Y. Ishikawa & R. E. Wuthier: Modulation of cultured chicken growth plate chondrocytes by transforming growth factor-beta 1 and basic fibroblast growth factor. J Cell Biochem, 49, 181-98 (1992)
doi:10.1002/jcb.240490211
PMid:6871381

441. Shapiro, I. M. & A. Boyde: Mineralization of normal and rachitic chick growth cartilage: vascular canals, cartilage calcification and osteogenesis. Scanning Microsc, 1, 599-606 (1987)
PMid:9084669

442. Shapiro, I. M., E. E. Golub, M. May & J. L. Rabinowitz: Studies of nucleotides of growth-plate cartilage: evidence linking changes in cellular metabolism with cartilage calcification. Biosci Rep, 3, 345-51 (1983)
doi:10.1007/BF01122899

443. Wu, L. N., B. R. Genge, G. R. Sauer & R. E. Wuthier: Characterization and reconstitution of the nucleational complex responsible for mineral formation by growth plate cartilage matrix vesicles. Connect Tissue Res, 35, 309-15 (1996)
doi:10.3109/03008209609029205

444. Termine, J. D., R. A. Peckauskas & A. S. Posner: Calcium phosphate formation in vitro. II. Effects of environment on amorphous-crystalline transformation. Arch Biochem Biophys, 140, 318-25 (1970)
doi:10.1016/0003-9861(70)90072-X

445. Wu, L. N., Y. Ishikawa, B. R. Genge, T. K. Sampath & R. E. Wuthier: Effect of osteogenic protein-1 on the development and mineralization of primary cultures of avian growth plate chondrocytes: Modulation by retinoic acid. J Cell Biochem, 67, 498-513 (1997)
doi:10.1002/(SICI)1097-4644(19971215)67:4<498::AID-JCB8>3.0.CO;2-N
PMid:16425255

446. Wu, L. N., Y. Ishikawa, D. Nie, B. R. Genge & R. E. Wuthier: Retinoic acid stimulates matrix calcification and initiates type I collagen synthesis in primary cultures of avian weight-bearing growth plate chondrocytes. J Cell Biochem, 65, 209-30 (1997)
doi:10.1002/(SICI)1097-4644(199705)65:2<209::AID-JCB7>3.0.CO;2-Y
PMid:8824294

447. Shapiro, I. M., C. S. Adams, T. Freeman & V. Srinivas: Fate of the hypertrophic chondrocyte: microenvironmental perspectives on apoptosis and survival in the epiphyseal growth plate. Birth Defects Res C Embryo Today, 75, 330-9 (2005)
doi:10.1002/bdrc.20057
PMid:6693404

448. Hsu, H. H. & H. C. Anderson: Evidence of the presence of a specific ATPase responsible for ATP-initiated calcification by matrix vesicles isolated from cartilage and bone. J Biol Chem, 271, 26383-8 (1996)
doi:10.1074/jbc.271.42.26383
PMid:1793679

449. Register, T. C., G. P. Warner & R. E. Wuthier: Effect of L- and D-tetramisole on 32Pi and 45Ca uptake and mineralization by matrix vesicle-enriched fractions from chicken epiphyseal cartilage. J Biol Chem, 259, 922-8 (1984)
PMid:11165936

450. Hsu, H. H. & H. C. Anderson: Effects of zinc and divalent cation chelators on ATP hydrolysis and Ca deposition by rachitic rat matrix vesicles. Bone, 17, 473-7 (1995)
PMid:8579959

451. Ali, S. Y., A. Wisby, L. Evans & J. Craig-Gray: The sequence of calcium and phosphorus accumulation by matrix vesicles. Calcif Tissue Res, 22 Suppl, 490-3 (1977)
doi:10.1007/BF02064139

452. Anderson, H. C.: Mineralization by matrix vesicles. Scan Electron Microsc953-64 (1984)
PMid:6385226

453. Balcerzak, M., E. Hamade, L. Zhang, S. Pikula, G. Azzar, J. Radisson, J. Bandorowicz-Pikula & R. Buchet: The roles of annexins and alkaline phosphatase in mineralization process. Acta Biochim Pol, 50, 1019-38 (2003)
PMid:14739992

454. Hsu, H. H. & H. C. Anderson: The deposition of calcium pyrophosphate and phosphate by matrix vesicles isolated from fetal bovine epiphyseal cartilage. Calcif Tissue Int, 36, 615-21 (1984)
doi:10.1007/BF02405376
PMid:6098356

455. Wuthier, R. E.: Mechanism of matrix vesicle-mediated mineralization in cartilage. ISI Atlas of Science: Biochemistry, 1, 231-241 (1988)

456. Register, T. C. & R. E. Wuthier: Effect of vanadate, a potent alkaline phosphatase inhibitor, on 45Ca and 32Pi uptake by matrix vesicle-enriched fractions from chicken epiphyseal cartilage. J Biol Chem, 259, 3511-8 (1984)
PMid:6561198

457. Wu, L. N., B. R. Genge, D. G. Dunkelberger, R. Z. LeGeros, B. Concannon & R. E. Wuthier: Physicochemical characterization of the nucleational core of matrix vesicles. J Biol Chem, 272, 4404-11 (1997)
doi:10.1074/jbc.272.7.4404
PMid:9020163

458. Betts, F., N. C. Blumenthal, A. S. Posner, G. L. Becker & A. L. Lehninger: Atomic structure of intracellular amorphous calcium phosphate deposits. Proc Natl Acad Sci U S A, 72, 2088-90 (1975)
doi:10.1073/pnas.72.6.2088

459. Aue, W. P., A. H. Roufosse, M. J. Glimcher & R. G. Griffin: Solid-state phosphorus-31 nuclear magnetic resonance studies of synthetic solid phases of calcium phosphate: potential models of bone mineral. Biochemistry, 23, 6110-4 (1984)
doi:10.1021/bi00320a032
PMid:6525349

460. Roufosse, A. H., W. P. Aue, J. E. Roberts, M. J. Glimcher & R. G. Griffin: Investigation of the mineral phases of bone by solid-state phosphorus-31 magic angle sample spinning nuclear magnetic resonance. Biochemistry, 23, 6115-20 (1984)
doi:10.1021/bi00320a033
PMid:6525350

461. Brecevic, L. & H. Furedi-Milhofer: Precipitation of calcium phosphates from electrolyte solutions. II. The formation and transformation of the precipitates. Calcif Tissue Res, 10, 82-90 (1972)
doi:10.1007/BF02012538

462. Boskey, A. L. & A. S. Posner: Optimal conditions for Ca-acidic phospholipid-PO4 formation. Calcif Tissue Int, 34 Suppl 2, S1-7 (1982)

463. Boskey, A. L., Posner, A.S.: Magnesium stabilization of amorphous calcium phosphate: A kinetic study. Materials Research Bulletin, 9, 907-916 (1974)
doi:10.1016/0025-5408(74)90169-X

464. Posner, A. S., F. Betts & N. C. Blumenthal: Role of ATP and Mg in the stabilization of biological and synthetic amorphous calcium phosphates. Calcif Tissue Res, 22 Suppl, 208-12 (1977)
doi:10.1007/BF02064066

465. Termine, J. D., R. E. Wuthier & A. S. Posner: Amorphous-crystalline mineral changes during endochondral and periosteal bone formation. Proc Soc Exp Biol Med, 125, 4-9 (1967)
PMid:6028913

466. Lyster, R. L., S. Mann, S. B. Parker & R. J. Williams: Nature of micellar calcium phosphate in cows' milk as studied by high-resolution electron microscopy. Biochim Biophys Acta, 801, 315-7 (1984)
PMid:6477967

467. McGann, T. C., R. D. Kearney, W. Buchheim, A. S. Posner, F. Betts & N. C. Blumenthal: Amorphous calcium phosphate in casein micelles of bovine milk. Calcif Tissue Int, 35, 821-3 (1983)
doi:10.1007/BF02405131
PMid:6652558

468. Taylor, M. G., K. Simkiss, J. Simmons, L. N. Wu & R. E. Wuthier: Structural studies of a phosphatidylserine-amorphous calcium phosphate complex. Cell Mol Life Sci, 54, 196-202 (1998)
doi:10.1007/s000180050143
PMid:9539964

469. Andersson, M., A. Malmendal, S. Linse, I. Ivarsson, S. Forsen & L. A. Svensson: Structural basis for the negative allostery between Ca (2+)- and Mg (2+)-binding in the intracellular Ca (2+)-receptor calbindin D9k. Protein Sci, 6, 1139-47 (1997)
doi:10.1002/pro.5560060602
PMid:9194174    PMCid:2143716

470. Babu, A., H. Su, Y. Ryu & J. Gulati: Determination of residue specificity in the EF-hand of troponin C for Ca2+ coordination, by genetic engineering. J Biol Chem, 267, 15469-74 (1992)
PMid:1639788

471. Fahie, K., R. Pitts, K. Elkins & D. J. Nelson: Molecular dynamics study of Ca (2+) binding loop variants of silver hake parvalbumin with aspartic acid at the "gateway" position. J Biomol Struct Dyn, 19, 821-37 (2002)
PMid:11922838

472. Buchbinder, J. L., J. Baraniak, P. A. Frey & G. H. Reed: Stereochemistry of metal ion coordination to the terminal thiophosphoryl group of adenosine 5'-O- (3-thiotriphosphate) at the active site of pyruvate kinase. Biochemistry, 32, 14111-6 (1993)
doi:10.1021/bi00214a006
PMid:8260493

473. Soultanas, P., M. S. Dillingham, S. S. Velankar & D. B. Wigley: DNA binding mediates conformational changes and metal ion coordination in the active site of PcrA helicase. J Mol Biol, 290, 137-48 (1999)
doi:10.1006/jmbi.1999.2873
PMid:10388562

474. Iijima, M., Y. Moriwaki & Y. Kuboki: Effect of some physico-chemical properties of matrix on lengthwise and oriented growth of octacalcium phosphate crystal. Connect Tissue Res, 38, 171-9; discussion 201-5 (1998)
doi:10.3109/03008209809017034
PMid:11063025

475. Nelson, D. G. & J. D. McLean: High-resolution electron microscopy of octacalcium phosphate and its hydrolysis products. Calcif Tissue Int, 36, 219-32 (1984)
doi:10.1007/BF02405321
PMid:6430505

476. Suzuki, O., H. Yagishita, T. Amano & T. Aoba: Reversible structural changes of octacalcium phosphate and labile acid phosphate. J Dent Res, 74, 1764-9 (1995)
doi:10.1177/00220345950740110801
PMid:8530738

477. Wu, Y., M. J. Glimcher, C. Rey & J. L. Ackerman: A unique protonated phosphate group in bone mineral not present in synthetic calcium phosphates. Identification by phosphorus-31 solid state NMR spectroscopy. J Mol Biol, 244, 423-35 (1994)
doi:10.1006/jmbi.1994.1740
PMid:7990131

478. Kirsch, T., H. D. Nah, D. R. Demuth, G. Harrison, E. E. Golub, S. L. Adams & M. Pacifici: Annexin V-mediated calcium flux across membranes is dependent on the lipid composition: implications for cartilage mineralization. Biochemistry, 36, 3359-67 (1997)
doi:10.1021/bi9626867
PMid:9116015

479. Simao, A. M., M. C. Yadav, P. Ciancaglini & J. L. Millan: Proteoliposomes as matrix vesicle biomimetics to study the initiation of skeletal mineralization. Braz J Med Biol Res, 43, 234-41
PMid:20401430

480. Matsumoto, H., Golub, E.E., Shapiro. I.M.: Superoxide required for matrix vesicle formation. In: Cell Mediated Calcification and Matrix Vesicles, Ed: S. Y. Ali, Elsevier Science Publishers D.V., Amsterdam, The Netherlands, 241-245 (1985).

481. McConnell, D.: The problem of the carbonate apatites. Science, 119, 913-914 (1954)
doi:10.1126/science.119.3104.913-a
PMid:17738591

482. McConnell, D. & A. S. Posner: Carbonate in apatites. Science, 134, 213-215 (1961)
doi:10.1126/science.134.3473.213
PMid:17818721

483. Bachra, B. N., O. R. Trautz, D. McConnell, W. J. Frajola & D. W. Deamer: Carbonic anhydrase and the precipitation of apatite. Science, 137, 337-8 (1962)
doi:10.1126/science.137.3527.337
PMid:13863741

484. McConnell, D., J. D. Termine & A. S. Posner: Infrared Absorption of Carbonate Apatite. Science, 155, 607-608 (1967)
doi:10.1126/science.155.3762.607
PMid:17737411

485. Ortlund, E., Chai, G., Genge, B., Wuthier, R.E., Lebioda, L.: Crystal structure of chicken annexin A5 in complex with functional modifiers Ca2+ and Zn2+ reveal Zn2+-induced formation of non-planar assemblies. Annexins, 1, 183-190 (2004)

486. Burton, F. G., M. W. Neuman & W. F. Neuman: On the possible role of crystals in the origins of life. I. The adsorption of nucleosides, nucleotides and pyrophosphate by apatite crystals. Curr Mod Biol, 3, 20-6 (1969)
PMid:4309914

487. Neuman, M. W., W. F. Neuman & K. Lane: On the possible role of crystals in the origins of life. 3. The phosphorylation of adenosine to AMP by apatite. Curr Mod Biol, 3, 253-9 (1970)
PMid:4318633

488. Jung, A., S. Bisaz & H. Fleisch: The binding of pyrophosphate and two diphosphonates by hydroxyapatite crystals. Calcif Tissue Res, 11, 269-80 (1973)
doi:10.1007/BF02547227
PMid:4350498

489. Bisaz, S., R. Felix, N. M. Hansen & H. Fleisch: Disaggregation of hydroxyapatite crystals. Biochim Biophys Acta, 451, 560-6 (1976)
PMid:187229

490. McGaughey, C. & E. C. Stowell: Effects of polyphosphates on the solubility and mineralization of HA: Relevance to a rationale for anticaries activity. J Dent Res, 56, 579-87 (1977)
doi:10.1177/00220345770560060501
PMid:197131

491. Fleisch, H. & S. Bisaz: Isolation from urine of pyrophosphate, a calcification inhibitor. Am J Physiol, 203, 671-5 (1962)
PMid:13945462

492. Fleisch, H., D. Schibler, J. Maerki & I. Frossard: Inhibition of aortic calcification by means of pyrophosphate and polyphosphates. Nature, 207, 1300-1 (1965)
doi:10.1038/2071300b0
PMid:4287147

493. Fleisch, H., J. Maerki & R. G. Russell: Effect of pyrophosphate on dissolution of hydroxyapatite and its possible importance in calcium homeostasis. Proc Soc Exp Biol Med, 122, 317-20 (1966)
PMid:4292457

494. Fleisch, H., R. G. Russell & F. Straumann: Effect of pyrophosphate on hydroxyapatite and its implications in calcium homeostasis. Nature, 212, 901-3 (1966)
doi:10.1038/212901a0

495. Fleisch, H., R. G. Russell, S. Bisaz, P. A. Casey & R. C. Muhlbauer: The influence of pyrophosphate analogues (diphosphonates) on the precipitation and dissolution. Calcif Tissue Res Suppl:10-10a (1968)
doi:10.1007/BF02065192

496. Francis, M. D., R. G. Russell & H. Fleisch: Diphosphonates inhibit formation of calcium phosphate crystals in vitro and pathological calcification in vivo. Science, 165, 1264-6 (1969)
doi:10.1126/science.165.3899.1264
PMid:4308521

497. Wuthier, R. E., S. Bisaz, R. G. Russell & H. Fleisch: Relationship between pyrophosphate, amorphous calcium phosphate and other factors in the sequence of calcification in vivo. Calcif Tissue Res, 10, 198-206 (1972)
doi:10.1007/BF02012549
PMid:4344764

498. Eaton, R. H. & D. W. Moss: Inhibition of the orthophosphatase and pyrophosphatase activities of human alkaline-phosphatase preparations. Biochem J, 102, 917-921 (1967)
PMid:16742510    PMCid:1270344

499. Russell, R. G.: Excretion of inorganic pyrophosphate in hypophosphatasia. Lancet, 2, 461-4 (1965)
doi:10.1016/S0140-6736(65)91422-4

500. Russell, R. G., S. Bisaz, A. Donath, D. B. Morgan & H. Fleisch: Inorganic pyrophosphate in plasma in normal persons and in patients with hypophosphatasia, osteogenesis imperfecta, and other disorders of bone. J Clin Invest, 50, 961-9 (1971)
doi:10.1172/JCI106589
PMid:4324072    PMCid:292015

501. Register, T. C. & R. E. Wuthier: Effect of pyrophosphate and two diphosphonates on 45Ca and 32Pi uptake and mineralization by matrix vesicle-enriched fractions and by hydroxyapatite. Bone, 6, 307-12 (1985)

502. Derfus, B. A., J. W. Rachow, N. S. Mandel, A. L. Boskey, M. Buday, V. M. Kushnaryov & L. M. Ryan: Articular cartilage vesicles generate calcium pyrophosphate dihydrate-like crystals in vitro. Arthritis Rheum, 35, 231-40 (1992)
doi:10.1002/art.1780350218
PMid:9701631

503. Derfus, B., S. Kranendonk, N. Camacho, N. Mandel, V. Kushnaryov, K. Lynch & L. Ryan: Human osteoarthritic cartilage matrix vesicles generate both calcium pyrophosphate dihydrate and apatite in vitro. Calcif Tissue Int, 63, 258-62 (1998)
doi:10.1007/s002239900523
PMid:10894769

504. Ho, A. M., M. D. Johnson & D. M. Kingsley: Role of the mouse ank gene in control of tissue calcification and arthritis. Science, 289, 265-70 (2000)
doi:10.1126/science.289.5477.265
PMid:11326272

505. Nurnberg, P., H. Thiele, D. Chandler, W. Hohne, M. L. Cunningham, H. Ritter, G. Leschik, K. Uhlmann, C. Mischung, K. Harrop, J. Goldblatt, Z. U. Borochowitz, D. Kotzot, F. Westermann, S. Mundlos, H. S. Braun, N. Laing & S. Tinschert: Heterozygous mutations in ANKH, the human ortholog of the mouse progressive ankylosis gene, result in craniometaphyseal dysplasia. Nat Genet, 28, 37-41 (2001)
doi:10.1038/88236
PMid:12297987

506. Pendleton, A., M. D. Johnson, A. Hughes, K. A. Gurley, A. M. Ho, M. Doherty, J. Dixey, P. Gillet, D. Loeuille, R. McGrath, A. Reginato, R. Shiang, G. Wright, P. Netter, C. Williams & D. M. Kingsley: Mutations in ANKH cause chondrocalcinosis. Am J Hum Genet, 71, 933-40 (2002)
doi:10.1086/343054
PMid:12707589

507. Williams, C. J.: Familial calcium pyrophosphate dihydrate deposition disease and the ANKH gene. Curr Opin Rheumatol, 15, 326-31 (2003)
doi:10.1097/00002281-200305000-00023
PMid:16462526

508. Zaka, R. & C. J. Williams: Role of the progressive ankylosis gene in cartilage mineralization. Curr Opin Rheumatol, 18, 181-6 (2006)
doi:10.1097/01.bor.0000209432.36355.6e
PMid:16869722

509. Gurley, K. A., H. Chen, C. Guenther, E. T. Nguyen, R. B. Rountree, M. Schoor & D. M. Kingsley: Mineral formation in joints caused by complete or joint-specific loss of ANK function. J Bone Miner Res, 21, 1238-47 (2006)
doi:10.1359/jbmr.060515
PMid:4886646    PMCid:322268

510. Howell, D. S., J. C. Pita, J. F. Marquez & R. A. Gatter: Demonstration of macromolecular inhibitors of calcification and nucleational factors in fluid from calcifying sites in cartilage. J Clin Invest, 48, 630-41 (1969)
doi:10.1172/JCI106021
PMid:4921198    PMCid:322719

511. Pita, J. C., L. A. Cuervo, J. E. Madruga, F. J. Muller & D. S. Howell: Evidence for a role of proteinpolysaccharides in regulation of mineral phase separation in calcifying cartilage. J Clin Invest, 49, 2188-97 (1970)
doi:10.1172/JCI106437
PMid:20078296

512. Howell, D. S. & J. C. Pita: Role of proteoglycans in calcification of cartilage. Ups J Med Sci, 82, 97-8 (1977)
doi:10.3109/03009737709179080
PMid:4225030

513. Di Salvo, J. & M. Schubert: Specific interaction of some cartilage proteinpolysaccharides with freshly precipitating calcium phosphate. J Biol Chem, 242, 705-10 (1967)
PMid:5344407

514. Wuthier, R. E.: A zonal analysis of inorganic and organic constituents of the epiphysis during endochondral calcification. Calcif Tissue Res, 4, 20-38 (1969)
doi:10.1007/BF02279103
PMid:5135576

515. Wuthier, R. E.: Zonal analysis of electrolytes in epiphyseal cartilage and bone of normal and rachitic chickens and pigs. Calcif Tissue Res, 8, 24-35 (1971)
doi:10.1007/BF02010120
PMid:5135577

doi:10.1007/BF02010121
PMid:6387076

516. Ehrlich, M. G., A. L. Armstrong & H. J. Mankin: Partial purification and characterization of a proteoglycan-degrading neutral protease from bovine epiphyseal cartilage. J Orthop Res, 2, 126-33 (1984)
doi:10.1002/jor.1100020203
PMid:1571846

517. Dean, D. D., Z. Schwartz, O. E. Muniz, R. Gomez, L. D. Swain, D. S. Howell & B. D. Boyan: Matrix vesicles are enriched in metalloproteinases that degrade proteoglycans. Calcif Tissue Int, 50, 342-9 (1992)
doi:10.1007/BF00301632

518. Dean, D. D., Z. V. Schwartz, O. E. Muniz, R. Gomez, L. D. Swain, D. S. Howell & B. D. Boyan: Matrix vesicles contain metalloproteinases that degrade proteoglycans. Bone Miner, 17, 172-6 (1992)
doi:10.1016/0169-6009(92)90731-R

519. Boskey, A. L., D. Stiner, I. Binderman & S. B. Doty: Effects of proteoglycan modification on mineral formation in a differentiating chick limb-bud mesenchymal cell culture system. J Cell Biochem, 64, 632-43 (1997)
doi:10.1002/(SICI)1097-4644(19970315)64:4<632::AID-JCB11>3.0.CO;2-E

520. D'Angelo, M., Z. Yan, M. Nooreyazdan, M. Pacifici, D. S. Sarment, P. C. Billings & P. S. Leboy: MMP-13 is induced during chondrocyte hypertrophy. J Cell Biochem, 77, 678-93 (2000)
doi:10.1002/(SICI)1097-4644(20000615)77:4<678::AID-JCB15>3.0.CO;2-P

doi:10.1002/(SICI)1097-4644(20000615)77:4<678::AID-JCB15>3.3.CO;2-G

521. Howell, D. S., J. C. Pita, J. F. Marquez, J. E. Madruga & F. J. Muller: Evidence for altered proteinpolysaccharide complexes (PPC) in fresh osteoarthritic cartilage in human hips. Calcif Tissue Res Suppl:127 (1970)
doi:10.1007/BF02152385
PMid:5704407

522. Glimcher, M. J.: A basic architectural principle in the organization of mineralized tissues. Clin Orthop Relat Res, 61, 16-36 (1968)
doi:10.1097/00003086-196811000-00004

523. Landis, W. J., K. J. Hodgens, J. Arena, M. J. Song & B. F. McEwen: Structural relations between collagen and mineral in bone as determined by high voltage electron microscopic tomography. Microsc Res Tech, 33, 192-202 (1996)
doi:10.1002/(SICI)1097-0029(19960201)33:2<192::AID-JEMT9>3.0.CO;2-V

doi:10.1002/(SICI)1097-0029(19960201)33:2<192::AID-JEMT9>3.3.CO;2-T

524. Glimcher, M. J., Hodge, A.J., Schmitt, F.O.: Molecular aggregation states in relation to mineralization: The collagen-hydroxyapatite system studied in vitro. Proc Natl Acad Sci U S A, 43, 860-867 (1957)
doi:10.1073/pnas.43.10.860
PMid:2672883

525. Poole, A. R., Y. Matsui, A. Hinek & E. R. Lee: Cartilage macromolecules and the calcification of cartilage matrix. Anat Rec, 224, 167-79 (1989)
doi:10.1002/ar.1092240207
PMid:3494735

526. Hinek, A., A. Reiner & A. R. Poole: The calcification of cartilage matrix in chondrocyte culture: studies of the C-propeptide of type II collagen (chondrocalcin). J Cell Biol, 104, 1435-41 (1987)
doi:10.1083/jcb.104.5.1435
PMid:2630170

527. Kujawa, M. J., M. Weitzhandler, A. R. Poole, L. Rosenberg & A. I. Caplan: Association of the C-propeptide of type II collagen with mineralization of embryonic chick long bone and sternal development. Connect Tissue Res, 23, 179-99 (1989)
doi:10.3109/03008208909002417
PMid:3800925    PMCid:1147077

528. Van der Rest, M., L. C. Rosenberg, B. R. Olsen & A. R. Poole: Chondrocalcin is identical with the C-propeptide of type II procollagen. Biochem J, 237, 923-5 (1986)

529. Landis, W. J., K. J. Hodgens, M. D. McKee, A. Nanci, M. J. Song, S. Kiyonaga, J. Arena & B. McEwen: Extracellular vesicles of calcifying turkey leg tendon characterized by immunocytochemistry and high voltage electron microscopic tomography and 3-D graphic image reconstruction. Bone Miner, 17, 237-41 (1992)
doi:10.1016/0169-6009(92)90743-W
PMid:8694787    PMCid:1217486

530. Hunter, G. K., P. V. Hauschka, A. R. Poole, L. C. Rosenberg & H. A. Goldberg: Nucleation and inhibition of hydroxyapatite formation by mineralized tissue proteins. Biochem J, 317 (Pt 1), 59-64 (1996)
PMid:3068009

531. Termine, J. D., H. K. Kleinman, S. W. Whitson, K. M. Conn, M. L. McGarvey & G. R. Martin: Osteonectin, a bone-specific protein linking mineral to collagen. Cell, 26, 99-105 (1981)
PMid:14139700

532. Termine, J. D.: Non-collagen proteins in bone. Ciba Found Symp, 136, 178-202 (1988)
PMid:5862222

533. Herring, G. M.: Comparison of Bovine Bone Sialoprotein and Serum Orosomucoid. Nature, 201, 709 (1964)
doi:10.1038/201709a0
PMid:6368573

534. Williams, P. A. & A. R. Peacocke: The physical properties of a glycoprotein from bovine cortical bone (bone sialoprotein). Biochim Biophys Acta, 101, 327-35 (1965)
PMid:6242444

535. Poole, A. R., I. Pidoux, A. Reiner, H. Choi & L. C. Rosenberg: Association of an extracellular protein (chondrocalcin) with the calcification of cartilage in endochondral bone formation. J Cell Biol, 98, 54-65 (1984)
doi:10.1083/jcb.98.1.54

536. Poole, A. R., I. Pidoux, A. Reiner, H. Choi & L. C. Rosenberg: The association of a newly discovered protein, called chondrocalcin, with cartilage calcification. Acta Biol Hung, 35, 143-9 (1984)

537. Price, P. A., A. A. Otsuka, J. W. Poser, J. Kristaponis & N. Raman: Characterization of a gamma-carboxyglutamic acid-containing protein from bone. Proc Natl Acad Sci U S A, 73, 1447-51 (1976)
doi:10.1073/pnas.73.5.1447

doi:10.1073/pnas.73.10.3374

538. Price, P. A., J. W. Poser & N. Raman: Primary structure of the gamma-carboxyglutamic acid-containing protein from bovine bone. Proc Natl Acad Sci U S A, 73, 3374-5 (1976)
doi:10.1073/pnas.73.10.3374

doi:10.1073/pnas.73.5.1447
PMid:9084679

539. Price, P. A., M. R. Urist & Y. Otawara: Matrix Gla protein, a new gamma-carboxyglutamic acid-containing protein which is associated with the organic matrix of bone. Biochem Biophys Res Commun, 117, 765-71 (1983)
doi:10.1016/0006-291X(83)91663-7
PMid:3006731

540. Goldberg, H. A., K. J. Warner, M. J. Stillman & G. K. Hunter: Determination of the hydroxyapatite-nucleating region of bone sialoprotein. Connect Tissue Res, 35, 385-92 (1996)
doi:10.3109/03008209609029216
PMid:7034958

541. Baht, G. S., J. O'Yong, A. Borovina, H. Chen, C. E. Tye, M. Karttunen, G. A. Lajoie, G. K. Hunter & H. A. Goldberg: Phosphorylation of Ser136 is critical for potent bone sialoprotein-mediated nucleation of hydroxyapatite crystals. Biochem J, 428, 385-395 (2010)
doi:10.1042/BJ20091864
PMid:20377527

542. Whyte, M. P.: Physiological role of alkaline phosphatase explored in hypophosphatasia. Ann N Y Acad Sci, 1192, 190-200 (2010)
doi:10.1111/j.1749-6632.2010.05387.x
PMid:20392236

543. Murshed, M., D. Harmey, J. L. Millan, M. D. McKee & G. Karsenty: Unique coexpression in osteoblasts of broadly expressed genes accounts for the spatial restriction of ECM mineralization to bone. Genes Dev, 19, 1093-104 (2005)
doi:10.1101/gad.1276205

544. Cyboron, G. W., M. S. Vejins & R. E. Wuthier: Activity of epiphyseal cartilage membrane alkaline phosphatase and the effects of its inhibitors at physiological pH. J Biol Chem, 257, 4141-4146 (1982)
PMid:6917850

545. Yadav, M. C., Simao, A.M.S., Narisawa, S., Huesa, C., McKee, M.D., Farquharson, C. and Millan, J.L.: Loss of skeletal mineralization by the simultaneous ablatioin of PHOSPHO1 and alkaline phosphatase function - A unified model of the mechanisms of initiation of skeletal calcification. J Bone Miner Res, (In Press), (2010)
doi:10.1002/jbmr.195
PMid:20684022

546. Waymire, K. G., J. D. Mahuren, J. M. Jaje, T. R. Guilarte, S. P. Coburn & G. R. MacGregor: Mice lacking tissue non-specific alkaline phosphatase die from seizures due to defective metabolism of vitamin B-6. Nat Genet, 11, 45-51 (1995)
doi:10.1038/ng0995-45
PMid:7550313

547. Anderson, H. C., R. Garimella & S. E. Tague: The role of matrix vesicles in growth plate development and biomineralization. Front Biosci, 10, 822-37 (2005)
doi:10.2741/1576
PMid:15569622

548. Kirsch, T.: Annexins - their role in cartilage mineralization. Front Biosci, 10, 576-81 (2005)
doi:10.2741/1553
PMid:15574394

549. Brachvogel, B., J. Dikschas, H. Moch, H. Welzel, K. von der Mark, C. Hofmann & E. Poschl: Annexin A5 is not essential for skeletal development. Mol Cell Biol, 23, 2907-13 (2003)
doi:10.1128/MCB.23.8.2907-2913.2003
PMid:12665588    PMCid:152556

550. Belluoccio, D., I. Grskovic, A. Niehoff, U. Schlotzer-Schrehardt, S. Rosenbaum, J. Etich, C. Frie, F. Pausch, S. E. Moss, E. Poschl, J. F. Bateman & B. Brachvogel: Deficiency of annexins A5 and A6 induces complex changes in the transcriptome of growth plate cartilage but does not inhibit the induction of mineralization. J Bone Miner Res, 25, 141-53 (2010)
doi:10.1359/jbmr.090710
PMid:19580468

551. De, S. R., M. A. Ajmone-Cat, A. Nicolini & L. Minghetti: Expression of phosphatidylserine receptor and down-regulation of pro-inflammatory molecule production by its natural ligand in rat microglial cultures. J Neuropathol Exp Neurol, 61, 237-44 (2002)
PMid:11895038

552. Fadok, V. A., D. L. Bratton, D. M. Rose, A. Pearson, R. A. Ezekewitz & P. M. Henson: A receptor for phosphatidylserine-specific clearance of apoptotic cells. Nature, 405, 85-90 (2000)
doi:10.1038/35011084
PMid:10811223

553. Hoffmann, P. R., A. M. deCathelineau, C. A. Ogden, Y. Leverrier, D. L. Bratton, D. L. Daleke, A. J. Ridley, V. A. Fadok & P. M. Henson: Phosphatidylserine (PS) induces PS receptor-mediated macropinocytosis and promotes clearance of apoptotic cells. J Cell Biol, 155, 649-59 (2001)
doi:10.1083/jcb.200108080
PMid:11706053    PMCid:2198875

554. Hoffmann, P. R., J. A. Kench, A. Vondracek, E. Kruk, D. L. Daleke, M. Jordan, P. Marrack, P. M. Henson & V. A. Fadok: Interaction between phosphatidylserine and the phosphatidylserine receptor inhibits immune responses in vivo. J Immunol, 174, 1393-404 (2005)
PMid:15661897

555. Li, M. O., M. R. Sarkisian, W. Z. Mehal, P. Rakic & R. A. Flavell: Phosphatidylserine receptor is required for clearance of apoptotic cells. Science, 302, 1560-3 (2003)
doi:10.1126/science.1087621
PMid:14645847

556. Otsuka, M., K. Goto, S. Tsuchiya & Y. Aramaki: Phosphatidylserine-specific receptor contributes to TGF-beta production in macrophages through a MAP kinase, ERK. Biol Pharm Bull, 28, 1707-10 (2005)
doi:10.1248/bpb.28.1707

557. Bose, J., A. D. Gruber, L. Helming, S. Schiebe, I. Wegener, M. Hafner, M. Beales, F. Kontgen & A. Lengeling: The phosphatidylserine receptor has essential functions during embryogenesis but not in apoptotic cell removal. J Biol, 3, 15 (2004)
doi:10.1186/jbiol10
PMid:15345036    PMCid:549712

558. Hahn, P., J. Bose, S. Edler & A. Lengeling: Genomic structure and expression of Jmjd6 and evolutionary analysis in the context of related JmjC domain containing proteins. BMC Genomics, 9, 293 (2008)
doi:10.1186/1471-2164-9-293
PMid:18564434    PMCid:2453528

559. Williamson, P. & R. A. Schlegel: Hide and seek: the secret identity of the phosphatidylserine receptor. J Biol, 3, 14 (2004)
doi:10.1186/jbiol14
PMid:15453906    PMCid:549716

560. Kunisaki, Y., S. Masuko, M. Noda, A. Inayoshi, T. Sanui, M. Harada, T. Sasazuki & Y. Fukui: Defective fetal liver erythropoiesis and T lymphopoiesis in mice lacking the phosphatidylserine receptor. Blood, 103, 3362-4 (2004)
doi:10.1182/blood-2003-09-3245
PMid:14715629

561. Rosati, R., G. S. Horan, G. J. Pinero, S. Garofalo, D. R. Keene, W. A. Horton, E. Vuorio, B. de Crombrugghe & R. R. Behringer: Normal long bone growth and development in type X collagen-null mice. Nat Genet, 8, 129-35 (1994)
doi:10.1038/ng1094-129
PMid:7842010

562. Warman, M. L., M. Abbott, S. S. Apte, T. Hefferon, I. McIntosh, D. H. Cohn, J. T. Hecht, B. R. Olsen & C. A. Francomano: A type X collagen mutation causes Schmid metaphyseal chondrodysplasia. Nat Genet, 5, 79-82 (1993)
doi:10.1038/ng0993-79
PMid:8220429

563. Chan, D., Y. M. Weng, A. M. Hocking, S. Golub, D. J. McQuillan & J. F. Bateman: Site-directed mutagenesis of human type X collagen. Expression of alpha1 (X) NC1, NC2, and helical mutations in vitro and in transfected cells. J Biol Chem, 271, 13566-72 (1996)
doi:10.1074/jbc.271.23.13566
PMid:8662807

564. McIntosh, I., M. H. Abbott, M. L. Warman, B. R. Olsen & C. A. Francomano: Additional mutations of type X collagen confirm COL10A1 as the Schmid metaphyseal chondrodysplasia locus. Hum Mol Genet, 3, 303-7 (1994)
doi:10.1093/hmg/3.2.303
PMid:8004099

565. Nielsen, V. H., C. Bendixen, J. Arnbjerg, C. M. Sorensen, H. E. Jensen, N. M. Shukri & B. Thomsen: Abnormal growth plate function in pigs carrying a dominant mutation in type X collagen. Mamm Genome, 11, 1087-92 (2000)
doi:10.1007/s003350010212
PMid:11130976

566. Jacenko, O., P. LuValle, K. Solum & B. R. Olsen: A dominant negative mutation in the alpha 1 (X) collagen gene produces spondylometaphyseal defects in mice. Prog Clin Biol Res, 383B, 427-36 (1993)
PMid:8115362

567. Jacenko, O., P. A. LuValle & B. R. Olsen: Spondylometaphyseal dysplasia in mice carrying a dominant negative mutation in a matrix protein specific for cartilage-to-bone transition. Nature, 365, 56-61 (1993)
doi:10.1038/365056a0
PMid:8361538

568. Jacenko, O. & B. R. Olsen: Transgenic mouse models in studies of skeletal disorders. J Rheumatol Suppl, 43, 39-41 (1995)
PMid:7752133

569. Olsen, B. R.: Mutations in collagen genes resulting in metaphyseal and epiphyseal dysplasias. Bone, 17, 45S-49S (1995)
PMid:11733375    PMCid:1850580

570. Jacenko, O., D. Chan, A. Franklin, S. Ito, C. B. Underhill, J. F. Bateman & M. R. Campbell: A dominant interference collagen X mutation disrupts hypertrophic chondrocyte pericellular matrix and glycosaminoglycan and proteoglycan distribution in transgenic mice. Am J Pathol, 159, 2257-69 (2001)
PMid:15052332

571. Vance, J. E. & D. E. Vance: Phospholipid biosynthesis in mammalian cells. Biochem Cell Biol, 82, 113-28 (2004)
doi:10.1139/o03-073
PMid:18343815

572. Arikketh, D., R. Nelson & J. E. Vance: Defining the importance of phosphatidylserine synthase-1 (PSS1): unexpected viability of PSS1-deficient mice. J Biol Chem, 283, 12888-97 (2008)
doi:10.1074/jbc.M800714200
PMid:1382017

573. Stewart, A. J., S. J. Roberts, E. Seawright, M. G. Davey, R. H. Fleming & C. Farquharson: The presence of PHOSPHO1 in matrix vesicles and its developmental expression prior to skeletal mineralization. Bone, 39, 1000-7 (2006)
PMid:2039199    PMCid:245037

574. Barradell, L. B., D. Faulds & D. McTavish: Lansoprazole. A review of its pharmacodynamic and pharmacokinetic properties and its therapeutic efficacy in acid-related disorders. Drugs, 44, 225-50 (1992)
doi:10.2165/00003495-199244020-00007
PMid:2537417

575. Iwahi, T., H. Satoh, M. Nakao, T. Iwasaki, T. Yamazaki, K. Kubo, T. Tamura & A. Imada: Lansoprazole, a novel benzimidazole proton pump inhibitor, and its related compounds have selective activity against Helicobacter pylori. Antimicrob Agents Chemother, 35, 490-6 (1991)

576. Nagaya, H., H. Satoh, K. Kubo & Y. Maki: Possible mechanism for the inhibition of gastric (H+ + K+)-adenosine triphosphatase by the proton pump inhibitor AG-1749. J Pharmacol Exp Ther, 248, 799-805 (1989)
PMid:3301497

577. Morse, A.: Formic acid-sodium citrate decalcifiecation and butyl alcohol dehydration of teeth and bones for sectioning in paraffin. J Dent Res, 24, 143 (1945)
doi:10.1177/00220345450240030501
PMid:3838314

578. Arsenault, A. L.: Fine structure and elemental maps of the calcifying epiphysis preserved by slam freezing and freeze substitution. In: The Chemistry and Biology of Mineralized Tissues, Ed: William T. Butler, Ebsco Media, Inc., Birmingham, AL, 364-367 (1984).

579. Carraway, K. L., 3rd, Y. Liu, D. Puett, K. L. Carraway & C. A. Carraway: Phenothiazine binding by a homolog of calpactin, the pp60src tyrosine kinase substrate. Faseb J, 1, 46-50 (1987)

580. Grynkiewicz, G., M. Poenie & R. Y. Tsien: A new generation of Ca2+ indicators with greatly improved fluorescence properties. J Biol Chem, 260, 3440-50 (1985)

581. Concannon, B.: Nuclear magnetic resonance investigations of 1. Early stages of biomineralization, and 2. Rat mammary healthy and tumorous cells and tissues. Department of Chemistry and Biochemistry, University of South Carolina, 1-79 (1996).

582. Pleshko, N., A. Boskey & R. Mendelsohn: Novel infrared spectroscopic method for the determination of crystallinity of hydroxyapatite minerals. Biophys J, 60, 786-93 (1991)
doi:10.1016/S0006-3495(91)82113-0
PMid:8579897

583. Bisaz, S., R. G. Russell & H. Fleisch: Isolation of inorganic pyrophosphate from bovine and human teeth. Arch Oral Biol, 13, 683-96 (1968)
doi:10.1016/0003-9969(68)90146-5
PMid:16837257

Key Words: Matrix Vesicles, Ca²⁺, cellular Ca²⁺, Ca²⁺ Channels, Inorganic Phosphate, Pi, Pi Transporters, Matrix Vesicle Formation, Mitochondria, Annexins, Annexin A5, Calcification, Mg²⁺, Zn²⁺, Pyrophosphate, PPi, Alkaline Phosphatase, Collagen, types VI, X, and II, Liposomal Models, Phospholipases A₂, Phospholipases C, phosphatidylserine, PS, PS-Ca²⁺-Pi Complexes, Nucleation, Nucleational Core, Molecular Simulation, Amorphous Calcium Phosphate, Octacalcium Phosphate, Hydroxyapatite