[Frontiers in Bioscience 3, d376-398, March 26, 98] |
TRANSCRIPTION BY RNA POLYMERASE I
Katherine M. Hannan, Ross D. Hannan and Lawrence I. Rothblum
Henry Hood Research Program, Department of Molecular and Cellular Physiology, Penn State College of Medicine, Weis Centre for Research, 100 N. Academy Avenue, Danville, PA 17822-2618
Received 3/18/98 Accepted 3/23/98
5. REGULATION OF rDNA TRANSCRIPTION
Potential sites for regulation of ribosome synthesis include transcription of the ribosomal precursor genes (45S and 5S), pre-ribosomal splicing, and assembly of the ribosomal subunits, and transport from the nucleus to the cytoplasm. However, in the majority of cases ribosome synthesis has been shown to be regulated largely at the level of transcription of the ribosomal genes (rDNA). Theoretically, regulation of rDNA transcription can involve: i) changes in chromatin structure; ii) alterations in the amount, localization, or activity of RNA polymerase I; and/or iii) similar alterations in the associated transcription factors. Moreover, recent studies suggest that the rDNA transcription apparatus can assemble (or colocalize) on the rDNA, but not be actively transcribing, suggesting that there may also be mechanisms for inhibiting transcription (23,168,169,170).
Chromatin may regulate gene activity by limiting the access of transcription factors to their DNA binding sites on the promoter. However, this may not be the case for all transcription factors since UBF can associate with either naked or nucleosome associated rDNA. Like other genes, the ribosomal genes upon activation require a chromatin modification. For example, inactive Xenopus rRNA genes have nucleosomes occupying the complete 40S transcribed region and most of the intergenic spacer, including the repetitive enhancers. Transcribed genes do not appear to contain nucleosomes. Both UBF and TTF-1 may play roles in altering the structure of the chromatin. The association of UBF with chromatin in vitro results in the displacement of the linker histone H1, without affecting the core histones. The spacer promoter may also play a role in opening the enhancer chromatin to activating factors and thus be involved in an early stages of gene activation (2). To date, limited experimental techniques are available to fully examine and understand chromatin’s role in the regulation of rDNA transcription.
5.2. RNA Polymerase I
Efficent rDNA transcription requires active RNA polymerase I. Only a fraction of the total amount of RNA polymerase I purified from cells is capable of participating in transcription in vitro. As stated above Acanthamoeba castellanii encyst when starved, and their rate of rDNA transcription decreases concomitant with an increase in the content of a modified form of RNA polymerase I (PolA) and a decrease in PolB. PolA does not support specific transcription in vitro (66,72), while the other form, PolB, can initiate both specific and non-specific transcription (7,66,72,171). The difference between these two forms is unclear and has been be ascribed to a modification of one of the subunits of the enzyme (A. castellanii), and/or a change in its association with other RNA polymerase I associated factors.
Interestingly, several yeast RNA polymerase I subunits, including A190, A43, A34.5, ABC23 and AC19 and possibly C53 (62), as well as the A194 subunit of mammalian RNA polymerase I (67) are modified by phosphorylation. However, the role of phosphorylation in RNA polymerase I activity has yet to be established. Alternatively, as mentioned above, the difference between PolA and PolB in mammals may be due to changes in either their association with RNA polymerase I-associated factor(s) or to a change in that factor. These activities have been referred to as TFIC, Factor C*, TIF-IA, TIF-IC or PAF53 (15,78,80). Specific examples of such modifications are discussed below.
5.3. RNA Polymerase I Associated Factors
5.3.1. TFIC, TIF-IA and Factor C*
In a number of studies the rate of rDNA transcription was shown to correlate with the activity of TFIC, TIF-IA or Factor C*. For example, reduced rDNA transcription observed with cyclohexamide or glucocorticoid treatment of mouse P1798 lymphosarcoma cells was attributed to a decrease in either the amount and/or activity of TFIC (15,80). Interestingly, serum starvation of the same cell line did not alter TFIC activity even though rDNA transcription was reduced (74). Thus, regulation of RNA polymerase I activity by TFIC, in these cells, appears to be stimulus-dependent. rDNA transcription is also reduced in response to treatment with cyclohexamide or in post confluent suspension cultures of L 1210 and Ehrlich ascites cells. In these cells, this correlated with reduced Factor C* activity (72). Similarly, post confluent suspension cultures of Ehrlich ascites cells demonstrated reduced rates of rDNA transcription and reduced levels of TIF-IA activity (73,74,171). Paradoxically, if cells are arrested in mitosis by nocododazole they exhibit a high level of TIF-IA activity but a reduced rate of rDNA transcription (78).
In general the amounts of the activities referred to as TFIC, TIF-IA, Factor C* and TIF-IC correlated with rDNA transcription and thus appear to be critical in regulating the ability of RNA polymerase I to initiate specific transcription. However, further knowledge of the exact functions of these factors and their relationship with RNA polymerase I or other transcription factors, will remain limited until they have been studied in great detail.
As mentioned above, PAF53 is one of a group of recently purified proteins that associate with RNA polymerase I. To date two lines of evidence support a role for PAF53 in the regulation of rDNA transcription. 1) There is a positive correlation between the accumulation of PAF53 in the nucleoli of 3T3 cells and the rate of rDNA transcription; and 2) PAF53 is isolated in a complex with PolB (the transcriptionally active form of RNA polymerase I), but not with PolA (68). It has been suggested that PAF53 mediates an interaction between RNA polymerase I and UBF (68). However, these results were obtained in vitro and await additional corroboration.
5.4. rDNA Trans-acting Factors:
There has been only one published report of a physiologically relevant alteration in the amount or activity of SL-1. A priori one might predict that SL-1 would be a primary target for regulation in that: i) it is the RNA polymerase I paralogue of TFIID; and ii) it is absolutely required for rDNA transcription. Zhai et al. (173) have reported that SV40 large T antigen can bind to SL-1 and activate rDNA transcription (173). In contrast, SL-1 activity has been examined in extracts from growing and non-growing Ehrlich ascites cells and found to be unchanged (114). However, since the SL-1 subunits has only recently been cloned, studies on its regulation are still in their infancy. It is likely, as with the core RNA polymerase II transcription factors, that it will prove to be regulated at either the post-translational level (e.g., phosphorylation, acetylation) or by changes in its specific interaction with other positive or negative regulators of rDNA transcription.
The hypothesis that regulating UBF activity in the cell might have an effect on rDNA transcription is controversial. For instance, there are believed to be 10 000-100, 000 copies of UBF in the cell (22). This number, is in vast excess when compared to both the number of active ribosomal genes and to the estimated number of SL-1 and RNA polymerase I complexes (1,5). This would suggest that UBF is not a rate-limiting component of the rDNA transcription apparatus. However, these estimates are based on the amount of UBF present in rapidly dividing, immortal cell lines. The cellular content of UBF in differentiated cells such as, adult hepatocytes, neonatal and adult cardiomyocytes is significantly lower than that observed in immortal cell lines (D. O’Mahony, R. Hannan, and L. Rothblum, unpublished observation). Moreover, the transfection and overexpression of UBF1 in neonatal cardiomyocytes is sufficient to stimulate transcription from a reporter construct for rDNA transcription in a dose dependent manner (90). Such observations have led groups to examine if UBF is a potential target for regulation during altered growth conditions.
In theory, the cellular activity of UBF can be regulated by either altering the amount of UBF available to transactivate the rDNA promoter or by changing the activity of an individual molecule by postranslational modifications such as phophorylation. In fact, both have of these mechanisms have been demonstrated to occur. Moreover, while these two mechanisms are not mutually exclusive, they appear to be dependent on both the cell type and stimulus being examined.
Regulation of UBF Content: Numerous studies have demonstrated a correlation between the cellular content of UBF and rDNA transcription. For example, the differentiation of L6 myoblasts into myotubes correlates with a decrease in UBF mRNA which precedes the decrease in UBF content and rDNA transcription (11). At the same time, myosin heavy chain protein accumulates, the mRNA level of myogenin increases, and transcription of the tubulin, r-protein L32, and 5S rRNA genes do not change (11). Thus, the observed decrease in UBF content during differentiation is not due to a general decrease in gene expression or translation (11).
Serum starvation of cells, such as 3T6 cells, reduces rDNA transcription due to a decrease in the availability of the mitogenic factors found in serum that these cells require for growth. This decrease correlated with a decrease in the cellular content of UBF. Refeeding serum-starved 3T6 cells with serum restored UBF content, which preceded the elevation of rDNA transcription to levels observed in control cells. Accumulation of UBF protein was found to result from regulation at the level of transcription of the UBF gene, in a manner similar to that of c-myc (12).
The regulation of rDNA transcription has been studied in LNCAP cells, an androgen dependent cell line. Nuclear run-on data demonstrated that DHT treatment of these cells increases rDNA transcription which correlated with an increase in UBF cellular content (174). In addition, extracts of prostate cells from orchiectomized rats showed a decrease in rDNA transcription and UBF protein. However, if the rats were treated with testosterone these levels did not decrease (174). Thus, androgens appear to, at least in part, stimulate rRNA synthesis by regulating the quantities UBF.
A correlation between UBF content and the regulation of rDNA transcription has been extensively studied in primary cultures of neonatal cardiomyocytes. When neonatal cardiomyocytes are treated with various growth promoting stimuli such as adrenergic agents, they undergo hypertrophy. This is associated with an elevated protein synthetic capacity due to increased ribosome biogenesis, which is achieved by increasing rDNA transcription (8,143). A good correlation is observed between the degree to which cells grow (hypertrophy) in response to a growth stimulus and the degree to which rDNA transcription is increased. Phenylephrine affected neither the content of RNA polymerase I nor UBF phosphorylation. However, there were significant increases in the cellular contents of UBF mRNA and protein which correlated, both temporally and quantitatively, with changes in rDNA transcription (8). This correlation was confirmed by the observation that overexpressing UBF1, in the absence of hypertrophic stimuli, increased the activity of a cotransfected reporter for rDNA transcription (90).
Regulation of UBF Phosphorylation: Stimulation of neonatal cardiomyocytes with two other hypertrophic agents, phorbol 12-myristate 13-acetate (PMA) or endothelin-1 (ET-1) does not change the cellular content of UBF. Instead a significant increase in UBF phosphorylation was observed. This increase in UBF phosphorylation correlated, both temporally and quantitatively, with elevated rDNA transcription. The affects were not seem until 6-12 h after the onset of PMA or ET-1 treatment (143), suggesting that they did not result from the activation of protein kinase C. These findings emphasize that even within one cell type the mechanisms utilized to regulate rDNA transcription and UBF activity are stimuli specific.
A correlation between the phosphorylation status of UBF and rDNA transcription has been observed in other cell culture systems. For example, the decreased rate of rDNA transcription that accompanies serum starvation of CHO cells correlates with a slow decrease in UBF phosphorylation, in the absence of changes in cellular content (13,88). The addition of serum restores both rDNA transcription and the degree of UBF phosphorylation. Pulse-chase experiments demonstrated that the decrease in UBF phosphorylation was due to a reduction in phosphorylation, and not the result of "active" dephosphorylation. Similarly, treatment of vascular smooth muscle cells (VSMC) with Angiotensin II (AII), a hypertrophic stimulus, rapidly (within 30 min) increases both rDNA transcription and UBF phosphorylation in the absence of changes in UBF content (175). The activity of CKII, an enzyme which phosphorylates UBF in vitro, is not altered in those cells, suggesting that either a serine kinase other than CKII is responsible for AII stimulation of UBF phosphorylation or the ability of CKII to specifically phosphorylate UBF was being regulated.
None of the above studies have established if there are qualitative changes in the specific serine residues phosphorylated. In addition, while in vitro experiments have demonstrated that phosphorylated UBF is more transcriptionally active than dephosphorylated UBF, it remains to be determined whether alterations in the phosphorylation state are necessary or sufficient to effect changes in rDNA transcription rates in vivo. Future studies will have to define the sites phosphorylated and identify the enzymes responsible for phosphorylating UBF in vivo.
Sequestration of UBF: Recent studies indicate that a direct measurement of the total cellular content of UBF or its degree of phosphorylation may not necessarily correlate with the amount of UBF available to transactivate the rDNA promoter or the specific activity of UBF. This conclusion stems from the observation that UBF can be sequestered into an inactive complex with Rb110. Rb110 functions as a tumor suppresser and is a negative regulator of growth (1,144,176), acting at the G1 checkpoint. It is the underphosphorylated (hypophosphorylated) form of Rb110 which is the most active. This form predominates in quiescent cells, while the hyperphosphorylated form is prevalent in actively growing cells (1,144,176). The hypophosphoryated form of Rb predominates in Go and G1 phase cells. Hyperphosphorylated Rb predominates in the G2, M and S phases of the cell cycle (176).
The initial observations of an interaction between UBF and Rb110 came from studies on the regulation of rDNA transcription in differentiating U937 cells (10). It was noted that when U937 cells differentiate, Rb110 accumulates in the nucleolus and rDNA transcription decreases in the absence of changes in UBF content. This study also demonstrated, using cell-free transcription assays, that the addition of Rb110 to an extract containing limiting amounts of Rb resulted in inhibition of UBF-dependent rDNA transcription in vitro (10). Coimmunopreciptiation experiments demonstrated an increase in the association between Rb110 and UBF with differentiation (10). Moreover, affinity chromatography experiments demonstrated that this interaction was specific and required the A/B pocket of Rb. This was deduced since Rb110209, a biologically inactive form of Rb110 containing a cysteine to phenylalanine mutation at amino acid 706 (in the A/B pocket), did not interact with UBF and did not inhibit rDNA transcription in vitro. In addition, the UBF-Rb interaction could be inhibited by a synthetic peptide that has been shown to interact with the A/B pocket and block the interaction of other proteins with the pocket (10,144).
A second study, by a different group, confirmed that that Rb could inhibit rDNA transcription and that this was due to an interaction between Rb and UBF (177). They also found that Rb inhibited UBF binding to the rDNA promoter, but did not affect the ability of UBF to interact with SL-1 or RNA polymerase I (177). Interestingly, in these experiments Rb209 was just as affective as Rb, and the authors concluded that the CO2H-terminal domain of Rb, and not the A/B pocket, was required for the interaction between Rb and UBF (177). This finding contradicts the results of the initial study (10), and the observation that Rb209 (as found in H209 cells) cannot be coimmunoprecipitated with UBF (K. Hannan, L. Jefferson, and L. Rothblum, manuscript in preparation). These discrepancies suggest that further experiments are required in order to determine the exact mechanism involved in Rb110 regulation of rDNA transcription.
5.5. Other Factors:
5.5.1. Ku/E1BF and CPBF
In some reports Ku/E1BF markedly inhibited rDNA transcription (146,178,179), while in others it stimulated transcription (145,148). The first finding corroborates in vivo studies which demonstrated that the expression of Ku/E1BF correlates negatively with the proliferation state of the cell (3,180). However, the question remains as to how Ku/E1BF acts as both a positive and negative regulator of rDNA transcription. It has been suggested that low concentrations of Ku/E1BF have a positive effect on rDNA transcription whereas higher concentrations repress transcription (178,179,181).
Anti-Ku antibodies can precipitate a repressor activity from HeLa cells, and stimulate rDNA transcription. The addition of UBF can also overcome Ku/E1BF repression (147), thus suggesting that one mechanism by which UBF may enhance rDNA transcription is by releasing Ku/E1BF repression rather than by directly stimulating transcription (168). Since Ku/E1BF interacts with the UBF and SL-1 binding sites on the rDNA promoter it may compete with them for their DNA-binding sites on the rDNA promoter (147). In this model, a high concentration of Ku/E1BF would titrate the UBF and/or SL-1 binding sites and reduce rDNA transcription. One other study, which examined Ku/E1BF in serum-starved rat NISI cells, suggested that there are probably two forms of Ku/E1BF; one which enhances (Ku/E1BFc*) and one which inhibits transcription (Ku/E1BFs: isolated from serum starved cells) (179). The difference between these two forms is unclear and may involve alternative splicing or a post translational modification of Ku/E1BF (179). It is likely that, in this case, the effects of Ku/E1BF on rDNA transcription reflect the ratio of active (Ku/E1BFc*) to repressive (Ku/E1BFs) forms in the cell. Ghoshal and Jacob have reported that heat shock (42oC, 3h) repressed rDNA transcription and leads to a reduction (90%) in E1BF, demonstrating a correlation between the regulation of E1BF and rDNA transcription (182).
Interestingly, when Ku/E1BF is complexed with DNA-dependent PK the complex represses rDNA transcription to a greater extent than Ku/E1BF alone (150,151). It is possible that the enhanced repression may be due to DNA-dependent PK phosphorylation of certain components of the RNA polymerase I transcription complex.
Cell extracts lacking CPBF are rDNA transcriptionally inactive and the subsequent addition of CPBF restores transcription in an dose dependent fashion (3). Interestingly, when the human homologue of CPBF, USF1 is overexpressed as a homodimer in CHO cells it represses rDNA transcription. However, when USF1 forms heterodimers with USF2, rDNA transcription is stimulated. It is possible that the form of dimer may affect the ability of USF to bind the E boxes of the rDNA promoter and thus alter transcription (3,153).
The effect of topoisomerase on rDNA transcription have been examined in a number of systems. Treatment of HeLa cells with topoisomerase I-specific inhibitors, such as camptothecin, rapidly inhibits 45S rRNA synthesis which is reversible with drug removal (183). Interestingly, topoisomerase I coprecipitates with TBP (184), a subunit of the transcription factor SL-1, and has been reported to copurify with RNA polymerase I (185,186). These results suggest that it may be a component of an RNA polymerase I holoenzyme involved in the formation of the pre-initiation complex and thus rDNA transcription. While that model would suggest that topoisomerase I plays a positive role in rDNA transcription, there is additional evidence that it may negatively regulate transcription. Topoisomerase II has been found to bind to the CPE of the rDNA promoter and inhibit transcription by preventing pre-initiation complex formation (158). This process is counteracted by UBF. UBF may be competing with topoisomerase for the same DNA binding sites, thus if UBF is bound to the promoter Topo II is unable to repress transcription (158). The inhibition of topoisomerase I was also shown to generate a graded decrease (5’ to 3’) in the number of RNA polymerase I molecules associated with the transcription unit. This has been interpreted as evidence that the inhibition of topoisomerase I results in the inhibit elongation (183). In addition, mutagenesis studies have demonstrated that both topoisomerase I and II are important for rRNA synthesis in S. cerivisiae (156).