[Frontiers in Bioscience 1, d48-58, March 1, 1996]
Reprints
PubMed
CAVEAT LECTOR



DnaA- and PriA-DEPENDENT PRIMOSOMES: TWO DISTINCT REPLICATION COMPLEXES FOR REPLICATION OF Escherichia coli CHROMOSOME

Hisao Masai and Ken-ichi Arai.

Department of Molecular and Developmental Biology, Institute of Medical Science, University of Tokyo, 4-6-1, Shirokanedai, Minato-ku, Tokyo, 108, Japan

Received 01/17/96; Accepted 02/17/96; On-line 03/01/96

1. Summary
2. Introduction
3. phiX174-type (PriA-dependent) primosome
3.1 Primosomal proteins for assembly of the phiX174-type primosome
3.2 n'-pas isolated from various plasmids
4. DnaA-dependent primosome
4.1 oriC primosome
4.2 ABC primosome
4.3 ABC primosome can support progression of replication forks
5. DnaA/oriC-independent replication of the E. coli chromosome
6. Genetic study of primosomal proteins required for PriA-dependent primosome
6.1 dnaT (protein i)
6.2 PriA (protein n')
7. Two primosomes for replication of the E. coli chromosome and various plasmid replicons
8. PriA-dependent primosome for replication under hard life
9. Future perspectives
10. Acknowledgments
11. References

1. SUMMARY

Enzymatic analyses of primosome assembly at chromosomal and plasmid origins as well as that at single-stranded replication origins revealed the presence of two distinct primosomes in Escherichia coli for primer RNA synthesis and duplex unwinding. A DnaA-dependent primosome is assembled at oriC, the chromosomal origin of Escherichia coli, as well as at the A site, a single-stranded DNA hairpin containing a dnaA box sequence within its stem. In contrast, PriA protein recognizes a hairpin, called n'-pas (primosome assembly site), and initiates assembly of the phiX174-type PriA-dependent primosome in conjunction with other prepriming proteins. Genetic analyses of the prepriming proteins required specifically for the latter primosome strongly suggested that it is responsible for RecA-dependent, DnaA/oriC-independent replication of the Escherichia coli chromosome. Furthermore, primosome assembly in replication of various plasmids may also be classified into either DnaA-dependent or PriA-dependent type. We propose that Escherichia coli possesses two distinct, mutually exclusive primosomes which are differentially utilized by the chromosome as well as by the plasmids. PriA protein appears to be conserved in a wide range of prokaryotic species, and we will also discuss possible biological function of the PriA-dependent primosome in the process of responses to DNA damages.

2. INTRODUCTION

The process of DNA replication requires assembly of a replication complex, primosome, at the origin. The primosome is believed to be responsible for primer RNA synthesis of lagging strand and for duplex DNA unwinding, which are mediated by primase and DNA helicase, respectively, present in the protein complex. In conjunction with twin DNA polymerases and a swiverase which relieves torsional tension accumulating in advance of a replication fork, a primosome may contribute to efficient concurrent replication of leading and lagging strands at a replication fork. The molecular architecture of the primosomes have been studied in most detail in E. coli, leading to discovery of two structurally distinct primosomes. Molecular genetic analyses of protein components of the primosomes revealed their biological function of these protein complexes in replication of the E. coli chromosome as well as in that of plasmid replicons.

3. phiX174-type (PriA-dependent) primosome

3.1 Primosomal proteins for assembly of the phiX174-type primosome (Table 1)

In E. coli, a primosome was first discovered from the study of SS to RF replication of single-stranded DNA phage phiX174 (1, 2). Assembly of this primosome is initiated by recognition of a specific hairpin structure (called n' site or n'-pas [primosome assembly site]) present on the phiX174 genome by protein n' (later redesignated as PriA) (3, 4, 5), which is followed by assembly of a preprepriming complex through association of proteins n (PriB), n" (PriC), and i (DnaT). A heterohexmeric DnaB-DnaC protein joins the complex to form a preprimosome, which moves along a DNA strand while periodically associating with primase to generate primer RNAs (1, 2, 6, 7). The preprimosome can be physically isolated by gel filtration, which can support efficient primer RNA synthesis and DNA chain elongation upon addition of primase and DNA polymerase III holoenzyme (6, 7).

Table1: Components of the two primosomes for E. coli chromosomal replication
Site of primosome assemblyoriC, A-siten'-pas, D-loop, R-loop
RecognitionDnaA proteinPriA protein
Auxilliary proteins for loading helicase(HU for oriC)PriB
PriC
DnaT
HelicaseDnaB/DnaC
PrimingPrimase

With both primosomes, DNA chains are elongated by DNA polymerase III holoenzyme.

3.2 n'-pas isolated from varoius plasmids

n'-pas on the phiX174 DNA can adopt a stable secondary structure which is resistant to exonuclease VII digestion and can activate ATPase activity of PriA protein (3, 8). On ColE1 plasmid, sequences which could support efficient SS to RF conversion of an origin-defective M13 phage were discovered (9). The sequences were discovered on a related plasmid, pBR322, which could stimulate ATPase activity of PriA protein when converted to single-strand (10).

Figure 1: Nucleotide sequences and possible stem-loop structures of n'-pas isolated from various plasmid DNAs. Eight newly isolated n'-pas in addition to previously reported ones were classified into five groups on the basis of their primary structures. n'-pas of phiX174 phage, R100 plasmid and F•f2b belong to group I, n'-pas on the lagging strand of ColE1-type plasmids (ssiA) and n' sites of F•f5, F•f7 and ColE2 plasmid to group II, n'-pas of the F•f2a to group III, n'-pas on the leading strand of ColE1-type plasmids (ssiB) to group IV, and n'-pas from Rts1 plasmid to group V. Each group is represented by the sequences of the first listing described above. The sequences indicated by boxes show some sequence similarities, although none of them are conserved among all the groups. The group I and group II were described before (10).

Later, it was demonstrated that these sequences could support assembly of the phiX174-type primosome in vitro (11, 12). We have screened the genomes of various plasmids for the presence of n'-pas by utilizing a derivative of single-stranded M13 phage DNA which lacks the complementary strand origin (13). This defective phage generates only small, turbid plaques by itself. However, upon cloning of a sequence capable of directing efficient primer RNA synthesis, it can form big, clear plaques (14). Enzymatic analyses of primer RNA synthesis revealed that many of these so called ssi (single-strand initiation sequences) were functionally equivalent to n'-pas (15). Comparison of phage- and plasmid-derived n'-pas lead to classification of them into five groups on the basis of their primary structures (15) (Figure 1). n'-pas sequences in each group are capable of forming a stable secondary structure. The significance of these secondary structures was suggested by the conservation of the base paring in the stems in spite of divergence of primary sequences within the members of each group.

4. DnaA-dependent primosome

4.1 oriC primosome (Table 1)

The 4700 kb E. coli chromosome is replicated from oriC, located at 83 min on the chromosome, and initiation at oriC strictly depends on DnaA protein (16). ATP-bound DnaA protein binds to the oriC, and changes the conformation of the origin sequences (17, 18). Localized melting of the three repeats of the 13mer sequences permits loading of DnaB/DnaC helicase. The oriC preprimosome, generated at the oriC sequence with DnaA, DnaB and DnaC proteins, can be isolated by sucrose gradient and addition of primase and DNA polymerase holoenzyme III together with SSB and gyrase to the isolated preprimosome can sustain DNA synthesis of the entire plasmid (19).

4.2 ABC primosome (table 1)

A ssi cloned from R6K plasmid did not support assembly of the phiX174-type primosome in vitro, nor did it support primer RNA synthesis with any known priming enzymes. We discovered that SS to RF replication of a single-stranded phage containing this ssi was dependent on DnaA protein. We were able to reconstitute the replication with purified proteins including DnaA, DnaB, DnaC, primase, SSB and DNA polymerase III holoenzyme (20). The ssi is capable of forming a secondary structure, and its stem contains a dnaA box sequence. DnaA protein specifically recognizes the stem and forms a complex which is isolatable by gel filtration. Hence, we have named this ssi A site and the complex ABC primosome.

In contrast to oriC plasmid replication in vitro, which requires ATP-form of DnaA protein, the ABC primosome can be assembled on A site with ADP-form of DnaA protein and the DnaA-A site complex can be isolated in the absence of ATP. However, formation of an isolatable preprimosome complex requires the presence of ATP or ATPgS, which stabilizes heterohexamers formed with DnaB and DnaC proteins. DnaB protein is delivered to the DnaA-A site complex by virtue of its association with DnaC, which may interact with DnaA protein. The preprimosome can translocate on SSB-coated single-stranded DNA with energy supplied by hydrolysis of ATP, dATP or dCTP (21). The hydrolysis of the nucleotide is required also to facilitate the release of DnaC protein from DnaB-DnaC complex (22).

DnaB protein hydrolyzes ATP, CTP and GTP but not deoexynucleotides (23, 24). In contrast, the isolated ABC preprimosome hydrolyzes dATP, dCTP and to some extent dGTP, in consistent with its ability to utilize these nucleotides for helicase activity (21). Helicase activity of the ABC preprimosome is maximally activated by 100 to 200 µM ATP, whereas more than 1 mM of ATP is required for maximum activation of the helicase activity by DnaB protein alone (24). The preprimosome contains DnaA and DnaB proteins but most likely not DnaC protein. Among the proteins required for the primosome assembly, only DnaB protein is capable of hydrolyzing nucleotides. These results indicate that assembly of the ABC preprimosome somehow leads to functional and/or structural alteration of DnaB helicase, which enables it to utilize low concentration of ATP as well as deoxyribonucleotides. dATP can not only support helicase activity but also support priming and replication by the ABC primosome, since the A site-dependent replication, which has been suppressed by the presence of ATPgS, can be reactivated by dATP in the absence of any ribonucleotides (21).

4.3 ABC primosome can support progression of replication forks

pBR322 plasmid contains an n'-pas on the lagging strand template, and efficient lagging strand synthesis depends on the assembly of the phiX174-type primosome at this n'-pas, in the absence of which replication intermediates containing only the nascent leading strand are accumulated in vitro (25, 26). Furthermore, lagging strand synthesis in vitro is completely suppressed in the presence of anti-DnaT protein antibody, which inhibits assembly of the phiX174-type primosome. Replacement of the n'-pas with A site on pBR322 restored the activity to synthesize lagging strand in the absence of DnaT protein (20). The same plasmid is capable of replication in the priA1::kan strain, in which the wild-type pBR322 cannot be replicated due to lack of the phiX174-type primosome assembly (27). These results demonstrate that ABC primosome can replace the phiX174-type primosome in replication of pBR322 for lagging strand synthesis and duplex unwinding at the replication fork.

5. DnaA/oriC-independent replication of the E. coli chromosome

The E. coli chromosome can be replicated in a manner independent of DnaA and oriC under certain conditions (Table 2). These altered modes of the chromosomal replication are called stable DNA replication (SDR), since they can continue stably in the absence of protein synthesis (28, 29). In the presence of DNA damaging agents or any treatment which temporarily halts the progression of replication forks, inducible SDR (iSDR) is observed. iSDR depends on RecA as well as on RecB and RecC. It is proposed that iSDR is initiated from a double-strand break (DSB) which is introduced in or near oriM (origin of SDR) (29). The linear end is converted to single-strand by exonuclease activity of RecBCD complex, which is assimilated into a D-loop by action of RecA protein (30). On the other hand, in an E. coli mutant which contains reduced activity of RNaseH encoded by rnhA, constitutive SDR (cSDR) is observed (31). cSDR is proposed to be initiated from R-loops, which are efficiently removed by rnhA-encoded RNaseH activity in the wild-type strain but persist in rnhA mutants. E. coli cells can survive lack of DnaA and/or oriC, when rnhA is mutated, indicating that growth of E. coli can be supported solely by cSDR. Then, what is the nature of the protein complex responsible for replication initiated from D-loops or R-loops?

Table 2: DnaA- and PriA-type primosomes in replication of E. coli chromosome and plasmids
DnaA-typePriA-type
ChromosomeoriCoriMs (iSDR) <- DNA damages
oriKs (cSDR) <- rnhA1-
PlasmidsF
pSC101
R6K
RK2
Rts1 etc.
ColE1
pBR322
RSF1030
ColE2

oriMs and oriKs, which are present at multiplie locations on the E. coli chromosome, are replication origins for iSDR and cSDR respectively.

6. Genetic study of primosomal proteins required for PriA-dependent primosome

6.1 dnaT (protein i)

In order to examine the physiological function of the phiX174-type primosome in replication of the E. coli chromosome, we undertook molecular genetic analyses of primosomal proteins for the phiX174-type primosome. First, we isolated the gene for protein i, which was mapped next to dnaC (32, 33). Further genetic analyses indicated that protein i is encoded by dnaT. dnaT and dnaC constitute an operon and are cotranscribed together with the p18 gene downstream of dnaC from a promoter present upstream of dnaT (33). dnaT was originally isolated as being defective in induction of iSDR (34) . This result suggests for the first time that the phiX174-type primosome may function in iSDR.

6.2 PriA (protein n')

n' protein recognizes the n'-pas and triggers assembly of the phiX174-type primosome. We isolated the gene for this protein and designated it priA (4, 35). priA is a previously unknown gene located at 88.7 min on the E. coli chromosome. PriA protein belongs to the DEXH-type RNA/DNA helicase family (Figure 2A) and it does have intrinsic ATPase and DNA helicase activities which are specifically stimulated by n'-pas.

Figure 2 B
Escherichia coli
Haemophilus Influenzae
Rhodospirillum rubrum
1

.......... .......... .MpvahVALP VPLPRtFDYl lPeGMt.vka
.......... .......... .MkIVrVALa VPLPRlFDYf VPddvs.lqi
maglpalppg srelfpedah AepVVaVlLP lPLagayDYk VPaGMarpav
50
Escherichia coli
Haemophilus Influenzae
Rhodospirillum rubrum
51

GcRVRVPFGk QqERIGIVvs vsdASElPln eLKAvvevLD sePvFThsvW
GmRVlVPFGt Qk.RvaIVad fptkSdvPed kLKAilqpLD laPlFTpiyW
GtlVRVPlGr reE.IGvVwg .agAgEtPpe rLKplig.fp ecPplpaplr
100
Escherichia coli
Haemophilus Influenzae
Rhodospirillum rubrum
101

rlLlWAAdYY hhPiGDVLFh ALPilLRqGr pAa.NApmWy wfateQgqAv
dwLhWAAnYY qaglGDVLFq ALPVkLRnGe sAvkNdrtfw RitdagknAl
afidWvAaYt vqPpGaVLrm ALsV..paal eApppAlgWr RpsagQraAg
150
Escherichia coli
Haemophilus Influenzae
Rhodospirillum rubrum
151

d........l nsLKRSPKQQ qALAaLrqgk iwr....Dqv RtleFndAal
k........q GeLKRSkKQa eALqyLsetd lek....gnn ...dFssAiw
qraegqgplp GgarlSPgrQ rvLAvLddhp glpfagaDla ReaavgpAvv
200
Escherichia coli
Haemophilus Influenzae
Rhodospirillum rubrum
201

qALrKKGLcd ...lasetpe fsdWrtNyav ..sgeRLrLN teQAtAvGai
sALkaKGfiE eiTiqtnpls wqqrlgNnPi vnaenRLtLN kQQAlAfsql
aAmaKaGLlE avT......r sneWspqaPd adrpgpLlsa dQQAaAdGlr
250
Escherichia coli
Haemophilus Influenzae
Rhodospirillum rubrum
251

hsaadtFSaW LLaGVTGSGK TEVYLsvlEn vLaqGKQALV mVPEIGLTPQ
lfhsg.FnvW LLdGVTGSGK TEiYLqyIEE iLksGKQvLV LVPEIGLTPQ
taldqgFSgl LLeGVTGSGK TEVYfeaIaE tLrrGrQALV LlPEIaLaaQ
300
Escherichia coli
Haemophilus Influenzae
Rhodospirillum rubrum
301

TiaRFreRFN ApveVLHSgL nDseRLsAWl kAknGeAAIV IGTRSALFTP
TvqRFkvRFN veidVLHSnL tDtqRLyvWd rArsGqsAIV IGTRSALFTq
wprRFadRFg AapvqwHSqm gaaaRrrAWr avalGrApvV vGaRSALFlP
350
Escherichia coli
Haemophilus Influenzae
Rhodospirillum rubrum
351

FkNLGvIviD EEHDSSYKQQ EGWRYHARDL AVyRAhseqI PiiLGSATPa
FsNLGaIIlD EEHDSSYKQQ dsWRYHARDL AiVlAqklnI svlmGSATPS
ypdLGlIIvD EEHDSafkQe EGvpYnARDm AVVRArlggf PavLaSATPS
400
Escherichia coli
Haemophilus Influenzae
Rhodospirillum rubrum
401

LETlcNVQQk KYRlLrLtRR AGNArpaiqh VLDLKgQkvQ ..........
LEsInNVQnG KYqHLVLskR AGNstalrhf ViDLKnQniQ ..........
LETIeNarQG rYRHLVLpRR hGgAempeit lLDLrrappQ kwlptdfagp
450
Escherichia coli
Haemophilus Influenzae
Rhodospirillum rubrum
451

.......... .......... .......... .......... .....aGLaP
.......... .......... .......... .......... .....nGLSk
ggseglaapg gandeaeeqk apppsptasp sptaspspma rrarlgwLSP
500
Escherichia coli
Haemophilus Influenzae
Rhodospirillum rubrum
501

aLITRMrqHL qAdNQViLFL NRRGFAPaLL CHdCGWIAeC PRCdhyYTlH
PLleRMkaHL ekGNQVLLFL NRRGFAPvLL CHeCGWIAqC PhCekpYTyH
PLITaveetL aAGeQVLLFL NRRGyAPltL CrsCGhrlkC PRCtawlveH
550
Escherichia coli
Haemophilus Influenzae
Rhodospirillum rubrum
551

QaQhhLRCHH CdsQrPvPRQ CPsCGSTH.L VPvGLGTEQL EqTLApLFPg
QhQnvLRCHH CGaQktIPRQ CgdCGSTH.L VttGLGTEQL EETLktLFPh
rrdgrLRCHH CGyQqPIPet CPaCGvadsL aPcGpGvErL aEeaAhrFPk
600
Escherichia coli
Haemophilus Influenzae
Rhodospirillum rubrum
601

vpisRIDRDT TSRKGaLEqq LaevhrGgar ILIGTQMLAK GHHFPdVTLV
ysvaRIDRDs TSRKGkLEgy LedIqqGksq ILIGTQMLAK GHHFPnVTLV
armdvaasDT vtgpkeaaal atrIanhdid liIGTQimAK GHHFPliTLV
650
Escherichia coli
Haemophilus Influenzae
Rhodospirillum rubrum
651

ALlDVDgALF SaDFRsaERf AQLYtQVAGR AGRAgKQGEV VLQTHhPeHP
ALVnVDsALF SlDFRAeERl AQLYiQVAGR AGRAdKQGEV VLQTHyPdHP
gvVDgDlgLt ggDlRAsERt hQLlhQVAGR AGRAerpGrV liQTvdPgHP
700
Escherichia coli
Haemophilus Influenzae
Rhodospirillum rubrum
701

LLqTLLykGY dAFAEqrLAE RrmMqLPPwT shVivrAedh nnqhAplfLQ
LLtTLLanGY qAFAketLql RhsMgLPPFT fqalikAqar hSdlAencLs
vmeaLasgdp alFlEveaAE RqalamPPFg rlValvisge dSar....vQ
750
Escherichia coli
Haemophilus Influenzae
Rhodospirillum rubrum
751

QlrnlilSsp ladekLwVLG PVPAlapKrg GRwRWQiLLQ HPSRVrLQhi
QiAdffqSkq itG..LqmLG PmPAPfsKka GqyRWQLLLQ HPSRmtLQka
avAaalgraa pmGpgLdVLG PVPAPlamlr GRhRhrLLLk aargVkvQpv
800
Escherichia coli
Haemophilus Influenzae
Rhodospirillum rubrum
801

ingtLaLint iPdsrkVKwv LDVDPIegz.
lReyqq..ae leknsqVrli LDVDPqdlsz
vRhwLsLvsi pPg...VKvq vDVDPIsfl.
830
Blocks in blue are conserved motifs for helicases I, Ia, II, III, IV, and VI consequetively as indicated in A
Blocks in green are the conserved cysteine-rich region, V

Figure 2: Schematic representation of the structure of E. coli PriA protein and comparison of its primary structure with possible homologues form other bacterial species. A Six conserved sequence motifs for DNA/RNA helicases and the clusters of cysteine residues for a possible Zinc-finger structure are shown by gray and black boxes, respectively. The numbers indicate the amino acid number from the N terminus. B:The amino acid sequences of the E. coli PriA and those of two candidate homologues from Haemophilus influenzae and Rhodospirillum rubrum are aligned for maximum identity. The amino acid residues identical in at least two genes are shown with capital letters. Sequences from Haemophilus (Fleischmann et al., Science 269, 496 [1995]) and Rhodospirillum (Falk et al., Biochem. J. 228, 391 [1985]) are under the accession number HIU32718 and X02499, respectively, in Genbank. Homology with the Rhodospirillum rubrum gene was previously reported (4).

An additional feature of the structure of PriA protein is the four repeats of CXXC (C, cysteine) which are inserted between the helicase conserved motifs. This region is likely to constitute a Zinc finger-like structure. Mutagenesis study indicated the cysteine residues are essential for priA function (36). Search of data base lead to identification of likely homologues of priA from two species, one from Haemophilus influenzae and the other from a photosynthetic bacteria, Rhodospirillum rubrum (Figure 2B). Identity (I) and similarity (S) are as follows:

E. coli (733 aa) vs. Haemophilus (710 aa) : I, 52%; S, 71 %

E. coli (733 aa) vs. Rhodospirillum (812 aa): I, 39%; S, 58%

The conserved motifs for helicases and CXXC sequences are precisely conserved among all the three proteins, indicating the importance of those sequences for priA function. Inactivation of the chromosomal copy of the priA gene by insertion of a kanamycine resistance gene (priA1::kan) can be tolerated for viability of the E. coli cells (37, 38), but priA1::kan cells grow extremely poorly and do not grow on rich medium (27). They are highly sensitive to DNA damaging agents such as UV and mytomycin C and exhibit filamentous morphology (39). Requirement of protein i for iSDR prompted us to examine involvement of PriA protein for iSDR. We measured iSDR in the priA1::kan strain, and discovered that iSDR was completely lost in the mutant strain (27). It was recovered by introduction of a plasmid carrying the wild-type priA gene. We constructed a double-mutant of rnhA224 and priA1::kan and measured cSDR. cSDR was also completely wiped out in this strain (27). Furthermore, cSDR-dependent growth of rnhA224 and dnaA(ts) double mutant at 42°C was also lost by priA1::kan mutation.These results demonstrate that PriA protein is essential for both iSDR and cSDR. They, in conjunction with requirement of dnaT for iSDR, strongly suggest that the phiX174-type primosome is responsible for replication forks during the course of SDR. A possible additional role of PriA protein in chromosomal replication could be "reloading" of replication complexes which may fall off the forks during the course of their propagation along the DNA template more than 2000 kb in length, although this possibility has not been experimentally tested.

7. Two primosomes for replication of the E. coli chromosome and various plasmid replicons

Replication of the E. coli chromosome under normal growth condition appears to be conducted solely by the DnaA-dependent primosome and dispensability of PriA protein for viability indicates that the phiX174-type primosome is not involved in DnaA/oriC-dependent replication (4, 35). On the other hand, PriA-dependent SDR does not require dnaA function. Thus, DnaA-dependent and PriA-dependent primosomes function independently of each other. The E. coli cells can be replicated with either protein complex, depending on the environment or genetic background. Replication of various plasmids generally depends on host replication proteins except for plasmid-encoded initiators that specifically recognize cognate replication origins. Many replication origins contain one or more DnaA box sequences within the minimum replication origin sequences and their replication does require DnaA protein (40). Replication of these DnaA-dependent replicons, such as F, pSC101, R6K, RK2 or Rts1, does not require priA function, whereas replication of pBR322, RSF1030, and ColE2, which is independent of dnaA function, strictly requires PriA protein (27). These indicate that replication of plasmid replicons also can be classified into either DnaA-type or PriA-type (Table 2). In the former class, DnaA protein may help assembly of a replication complex similar to oriC-type or ABC primosome by binding to the DnaA box present within the replication origin, while, in the latter class, the phiX174-type primosome may be assembled at an n'-pas on the lagging strand template. We propose that E. coli cells possess two functionally similar, independent primosomes, which are differentially utilized by the chromosome as well as by various plasmids (Table 2).

8. PriA-dependent primosome for replication under hard life

PriA-dependent replication of the E. coli chromosome (iSDR and cSDR) occurs only under specialized conditions. Especially, iSDR is specifically induced by DNA damages, and requires RecA, RecB and RecC proteins (41, 42, 43). It is hyperactivated in a recD mutant (43). These genetic requirements are extremely similar to those of adaptive (or induced) mutatgenesis which also requires the same set of recombination proteins (44, 45), which lead us to propose that iSDR is the mode of chromosomal replication, which is induced under hard environmental condition to help E. coli cells better adapt with the environment by increased mutation rate. It was previously reported that iSDR is error-prone (41). It will be important to determine whether priA is required for induced mutagenesis.

priA1::kan cells are extremely sensitive to DNA damaging agents such as UV and mytomycin C, strongly indicating that PriA protein is involved in repair of DNA damages, especially in recombination-dependent repair of double-strand breaks (DSB). Furthermore, the priA mutant exhibits decreased Hfr conjugation and decreased P1 transduction efficiency, indicating that priA is required for general homologous recombination processes (39). We have concluded that the phiX174-type primosome generally functions in DNA replication which is associated with recombination.

Our results indicate that the E. coli chromosome which is normally replicated by the DnaA-oriC pathway switches its replication mode to the PriA-dependent one in response to the environment (Figure 3). This temporal switch of the replication mode may permit the E. coli cells to adjust to "hard life" condition in two important ways. In individual cells within the population, it will enable them to repair lesions on DNA through a recombination-dependent manner, which requires PriA-dependent DNA synthesis. It will also increase the chance of survival as a species by enhanced mutation rate caused by PriA-dependent replication. In life-cycle of bacterial cells, this sort of switch of replication modes may occur more frequently than anticipated in response to changes of the environment and similar response may also occur in eukaryotic cells (see below).

Figure 3: Switch of replication modes in response to environmental conditions in E. coli. In E. coli, the chromosome is normally replicated by the high fidelity DnaA-dependent machinery. However, upon encountering "harsh" environment such as the presence of DNA damage-inducing agents or possibly the absence of sufficient nutrition, it is temporally switched to PriA-dependent one, which may have lower fidelity and is associated with recombination-dependent repair of DNA lesions. See text for details.

9. Future perspectives

How does PriA protein initiate DNA replication from D-loops or R-loops? The cloned oriM sequences do not have n'-pas (46, 47), and there may not be any functional n'-pas on the E. coli chromosome, since extensive screening using an origin-defective M13 phage vector failed to detect any defined sequences with n'-pas activity (N. Nomura et al., unpublished results). Therefore, it is unlikely that n'-pas on D-loops or on R-loops are recognized by PriA protein. We consider a possibility that PriA protein recognizes a part of the structural features of D-loops and R-loops. Absolute requirement of PriA protein for replication of pBR322 lacking n'-pas in the wild-type and in rnhA mutants supports this idea, since replication of this plasmid presents a model for D-loop or R-loop dependent replication.

Interestingly, we noticed that the nucleotide binding domain of PriA protein possesses a low but significant homology to that of RecG protein, which is an RNA helicase specific for R-loops (data not shown; 48). Our preliminary data suggests that PriA protein interacts with an R-loop structure, although ATPase is not activated by this interaction (H. Masai et al., unpublished result). The presence of PriA homologues in two other distantly related bacteria suggests that PriA-dependent replication from D-loops and R-loops may be conserved throughout the eubacteria species. It will be essential to develop an in vitro replication system in which replication is specifically initiated from a D-loop or from an R-loop.

Multiple modes of primosome assembly, which we have demonstrated in E. coli, may operate in replication of the eukaryotic chromosomes as well. Chromosomal replication in eukaryotic cells is strictly regulated so that it is coordinated with various cell cycle events. At the same time, choice of replication origins and sequence specificity in origin recognition can vary depending on developmental stages (49) or state of transcription in the vicinity of the origin sequences (50, 51). In early cleaving embryo, the S phases are completed in a matter of several minutes. This extraordinary rate of DNA replication is achieved by increasing the numbers of replication origins fired during the short S phase. The replication origin in the intergenic space of b-globin region is fired early in S phase in blood cells where the gene is actively transcribed, whereas it is replicated late in the cells which do not express globins (52, 53). Flexibility in origin selection was also indicated by genetic studies of yeast Saccharomyces cerevisiae, in which it was shown that deletion of an actively firing origin from a chromosome can be generally torelated, and that in some cases normally inactive replication origins can be activated to compensate for the loss of active origins (54). It is not known whether these variations in origin usage reflect the switch of replication modes or changes in the chromatin structures in the origin region. We also know very little about the origin selection and mode of initiation during premeiotic DNA synthesis, which could be different from those of mitotic DNA replication. It is of interest whether a recombination-dependent replication pathway similar to the one discovered in E. coli exists in eukaryotic cells. If PriA protein is an "initiator" for DNA replication from D-loops and R-loops, which are commonly found in eukaryotic cells as well, it may well be conserved in eukaryotes.

10. ACKNOWLEDGMENTS

We thank Dr. Tokio Kogoma of University of New Mexico School of Medicine for collaboration on genetic study of E. coli priA gene as well as for sharing information of unpublished results. We also thank Tsuyoshi Miyake of our department and Albert Zlotnik of DNAX Research Institue for help in data bank search.

11. REFERENCES

1. Arai, K., Low, R. & Kornberg, A. Movement and site selection for priming by the primosome in phage phiX174 DNA replication. Proc. Natl. Acad. Sci. USA 78, 707-711 (1981b).

2. Arai, K. & Kornberg, A. Unique primed start of phage phiX174 DNA replication and mobility of the primosome in a direction opposite chain synthesis Proc. Natl. Acad. Sci. USA 78, 69-73 (1981b).

3. Shlomai, J. & Kornberg, A. An Escherichia coli replication protein that recognizes a unique sequence within a hairpin region in phiX174 DNA. Proc. Natl. Acad. Sci. USA 77, 799-803 (1980).

4. Lee, E.H., Masai, H., Allen, J.G.C. & Kornberg, A. The priA gene encoding the primosomal replicative n' protein of Escherichia coli. Proc. Natl. Acad. Sci. USA 87, 4620-4624(1990).

5. Zavitz, K.H. & Marians, K.J. Dissecting the functional role of PriA protein-catalyzed primosome assembly in Escherichia coli DNA replication. Mol. Microbiol. 5, 2869-2873 (1991).

6. Arai, K., Low, R., Kobori, J., Shlomai, J. & Kornberg, A. Mechanism of dnaB protein action. V. Association of dnaB protein, protein n', and other prepriming proteins in the primosome of DNA replication. J. Biol. Chem. 256, 5273-5280 (1981a).

7. Allen, G.C. & Kornberg, A. Assembly of the primosome of DNA replication in Escherichia coli. J. Biol. Chem. 268, 19204-19209 (1993).

8. Shlomai, J. & Kornberg, A. A prepriming DNA replication enzyme of Escherichia coli. II. Action of protein n': J. Biol. Chem. 225, 6794-6798 (1980).

9. Nomura, N. & Ray, D.S. Expression of a DNA strand initiation sequence of ColE1 plasmid in a single-stranded DNA phage. Proc. Natl. Acad. Sci. USA 77, 6566-6570 (1980).

10. Zipursky, S.L. & Marians, K.J. Identification of two Escherichia coli factor Y effector sites near the origins of replication of the plasmids ColE1 and pBR322. Proc. Natl. Acad. Sci. USA 77, 6521-6525 (1980).

11. Masai, H., Nomura, N. & Arai, K. The ABC primosome: a novel priming system employing dnaA, dnaB, dnaC, and primase on a hairpin containing a dnaA box sequence. J. Biol. Chem. 265, 15134-15144 (1990).

12. Zipursky, S.L. & Marians, K.J. Escherichia coli factor Y sites of plasmid pBR322 can function as origins of DNA replication. Proc. Natl. Acad. Sci. USA 78, 6111-6115 (1981).

13. Nomura, N., Low, R.L. & Ray, D.S. Identification of ColE1 DNA sequences that direct single strand-to-double strand conversion by a phiX174 type mechanism. Proc. Natl. Acad. Sci. USA 79, 3153-3157 (1982).

14. Nomura, N., Masai, H., Inuzuka, M., Miyazaki, C., Ohtsubo, E., Itoh, T., Sasamoto, S., Matsui, M., Ishizaki, R. & Arai, K. Identification of eleven single-stranded initiation sequences (ssi) for priming of DNA replication in the F, R6K, R100 and ColE2 plasmids. Gene 108, 15-22 (1991).

15. Nomura, N., Low, R.L. & Ray, D. Selective cloning of ColE1 DNA initiation sequences using the cloning vector M13ÆE101. Gene 239-246 (1982).

16. Masai, H., Nomura, N., Kubota, Y. & Arai, K. Roles of phiX174-type primosome- and G4 type primase- dependent primings in initiation of lagging and leading strand syntheses of DNA replication. J. Biol. Chem. 265, 15124-15133 (1990).

17. Fuller, R.S., Kaguni, J.M. & Kornberg, A. Enzymatic replication of the origin of the Escherichia coli chromosome. Proc. Natl. Acad. Sci. USA 78, 7370-7374 (1981).

18. Sekimizu, K., Bramhill, D. & Kornberg, A. ATP activates dnaA protein in initiating replication of plasmids bearing the origin of the E. coli chromosome. Cell 50, 259-265 (1987).

19. Bramhill, D. & Kornberg, A. Duplex opening by dnaA protein at novel sequences in initiation of replication at the origin of E. coli. Cell 52, 743-755(1988a).

20. Sekimizu, K., Bramhill, D. & Kornberg, A. Sequential early stages in the in vitro initiation of replication at the origin of the Escherichia coli chromosome. J. Biol. Chem. 263, 7124-7130 (1988).

21. Masai, H. & Arai, K. DnaA-dependent assembly of the ABC primosome at the A site, a single-stranded DNA hairpin containing a dnaA box. Eur. J. Biochem. 230, 384-395 (1995).

22. Wahle, E., Lasken, R.S. & Kornberg, A. The dnaB-dnaC replication protein complex of Escherichia coli. II. Role of the complex in mobilizing dnaB functions. J. Biol. Chem. 264, 2469-2475 (1989).

23. Arai, K. & Kornberg, A. Mechanism of dnaB protein action. II. ATP hydrolysis by dnaB protein dependent on single- or double-stranded DNA. J. Biol. Chem. 256, 5253-5259 (1981a).

24. LeBowitz, J.H. & McMacken, R. The Escherichia coli dnaB replication protein is a DNA helicase. J. Biol. Chem. 261, 4738-4748. (1986)

25. Masai, H. & Arai, K. Initiation of lagging-strand synthesis for pBR322 plasmid DNA replication in vitro is dependent on primosomal protein i encoded by dnaT. J. Biol. Chem. 263, 15016-15023 (1988).

26. Minden, J.S. & Marians, K.J. Replication of pBR322 with purified proteins: requirement for topoisomerase I in the maintenance of template specificity. J. Biol. Chem. 260, 9316-9325 (1985).

27. Masai, H., Asai, T., Kubota, Y., Arai, K. & Kogoma, T. Escherichia coli PriA protein is essential for inducible and constitutive stable DNA replication. EMBO J. 13, 5338-5345. (1994)

28. Kogoma, T. & Lark, K.G. Characterization of the replication of Escherichia coli DNA in the absence of protein synthesis: stable DNA replication. J. Mol. Biol. 94, 243-256 (1975).

29. Asai, T. & Kogoma, T. D-loops and R-loops: alternative mechanisms for the initiation of chromosome replication in Escherichia coli. J. Bacteriol. 176, 1807-1812(1994).

30. Asai, T., Bates, D.B. & Kogoma, T. Double-strand breaks induce homologous recombination function-dependent DNA replication in Escherichia coli. Cell 78 1051-1061 (1994a).

31. Kogoma, T. RNaseH-defective mutants of Escherichia coli. J. Bacteriol. 166, 361-363. (1986)

32. Masai, H., Bond, M.W. & Arai, K. Cloning of the Escherichia coli gene for primosomal protein i: the relationship to dnaT, essential for chromosomal DNA replication. Proc. Natl. Acad. Sci. USA 83, 1256-1260 (1986).

33. Masai, H. & Arai, K. Operon structure of dnaT and dnaC essential for normal and stable DNA replication of Escherichia coli chromosome. J. Biol. Chem. 263, 15083-15093 (1988).

34. Lark, K.G. & Lark, C.A. recA-dependent DNA replication in the absence of protein synthesis: characterics of a dominant lethal replication mutation, dnaT, and requirement for recA+ function. Cold Spring Harb. Symp. Quant. Biol. 43, 537-547 (1978).

35. Nurse, P., DiGate, R.J., Zavitz, K.H. & Marians, K.J. Molecular cloning and DNA sequence analysis of Escherichia coli priA, the gene encoding the primosomal protein replication factor Y. Proc. Natl. Acad. Sci. USA 87, 4615-4619 (1990).

36. Zavitz, K.H. & Marians, K.J. Helicase-deficient cysteine to glycine substitution mutants of Escherichia coli replication protein priA retain single-stranded DNA-dependent ATPase activity J. Biol. Chem. 268, 4337-4336 (1993).

37. Lee, E.H. & Kornberg, A. Replication deficiencies in priA mutants of Escherichia coli lacking primosomal replication n' protein. Proc. Natl. Acad. Sci. USA 88, 3029-3032 (1991).

38. Nurse, P., Zavitz, K.H. & Marians, K.J. Inactivation of Escherichia coli PriA DNA replication protein induces the SOS response. J. Bacteriol. 173, 6686-6693 (1991).

39. Bramhill, D. & Kornberg, A. A model for initiation at origins of DNA replication. Cell 54, 915-918 (1988b).

40. Kogoma, T., Cadwell, G., Bernard, K. G., & Asai, T. Requirement of the DNA replication protein, PriA, for homologous recombination and double-strand break repair. J. Bacteriol. 178, in press (1996).

41. Kogoma, T., Torrey, T.A. & Connaughton, M.J. Induction of UV-resistant DNA replication in Escherichia coli: induced stable DNA replication as an SOS function. Mol. Gen. Genet. 176, 1-9 (1979).

42. Magee, T.R., Asai, T., Malka, D. & Kogoma, T. DNA damage-inducible origins of DNA replication in Escherichia coli EMBO J. 11, 4219-4225 (1992).

43. Magee, T.R. & Kogoma, T. Requirement of recBC enzyme and an elevated level of activated recA for induced stable DNA replication in Escherichia coli. J. Bacteriol. 172, 1834-1839 (1990).

44. Cairns, J., Overbaugh, J. & Miller, S. The origin of mutants Nature 335, 142-145 (1988).

45. Harris, R.S., Longerich, S. & Rosenberg, S.M. Recombination in adaptive mutation. Science 264, 258-260 (1994).

46. Asai, T., Imai, M. & Kogoma, T. DNA damage-inducible replication of the Escherichia coli chromosome is initiated at separable sites within the minimal oriC. J. Mol. Biol. 235, 1459-1469 (1994b).

47. Asai, T., Sommer, S., Bailone, A. & Kogoma, T. Homologous recombination-dependent initiation of DNA replication from DNA damage-inducible origins in Escherichia coli. EMBO J. 12, 3287-3295 (1993).

48. Hong, X., Cadwell, G.W. & Kogoma, T. Escherichia coli RecG and RecA proteins in R-loop formation. EMBO J. 14, 2385-2392 (1995).

49. Hyrien, O., Maric, C. & Mechali, M. Transition in specification of embryonic metazoan DNA replication origins. Science 270, 994-997 (1995).

50. Goldman, M.A. The chromatin domain as a unit of gene regulation. Bioassays 9, 50-55 (1988).

51. DePamphilis, M.L. Eukaryotic DNA replication: anatomy of an origin. Annu. Rev. Biochem. 62, 29-63 (1993).

52. Epner, E., Forrester, W.C. & Groudine, M. Proc. Natl. Acad. Sci. USA 85, 8081-8085 (1988).

53. Kitsberg, D., Selig, S., Keshet, I. & Ceder, H. Replication structure of the human b-globin gene domain. Science 366, 588-590 (1993).

54. Fangman, W.L. & Brewer, B.J. Activation of replication origins within yeast chromosomes. Annu. Rev. Cell Biol. 7, 375-402 (1991).