Recombination-dependent DNA replication stimulated by double-strand breaks in bacteriophage T4.

We analyzed the mechanism of recombination-dependent DNA replication in bacteriophage T4-infected Escherichia coli using plasmids that have sequence homology to the infecting phage chromosome. Consistent with prior studies, a pBR322 plasmid, initially resident in the infected host cell, does not replicate following infection by T4. However, the resident plasmid can be induced to replicate when an integrated copy of pBR322 vector is present in the phage chromosome. As expected for recombination-dependent DNA replication, the induced replication of pBR322 required the phage-encoded UvsY protein. Therefore, recombination-dependent plasmid replication requires homology between the plasmid and phage genomes but does not depend on the presence of any particular T4 DNA sequence on the test plasmid. We next asked whether T4 recombination-dependent DNA replication can be triggered by a double-strand break (dsb). For these experiments, we generated a novel phage strain that cleaves its own genome within the nonessential frd gene by means of the I-TevI endonuclease (encoded within the intron of the wild-type td gene). The dsb within the phage chromosome substantially increased the replication of plasmids that carry T4 inserts homologous to the region of the dsb (the plasmids are not themselves cleaved by the endonuclease). The dsb stimulated replication when the plasmid was homologous to either or both sides of the break but did not stimulate the replication of plasmids with homology to distant regions of the phage chromosome. As expected for recombination-dependent replication, plasmid replication triggered by dsbs was dependent on T4-encoded recombination proteins. These results confirm two important predictions of the model for T4-encoded recombination-dependent DNA replication proposed by Gisela Mosig (p. 120-130, in C. K. Mathews, E. M. Kutter, G. Mosig, and P. B. Berget (ed.), Bacteriophage T4, 1983). In addition, replication stimulated by dsbs provides a site-specific version of the process, which should be very useful for mechanistic studies.     

Characterization of a defective phage system for the analysis of bacteriophage T4 DNA replication origins

We have developed a defective phage system for the isolation and analysis of phage T4 replication origins based on the T4-mediated transduction of plasmid pBR322. During the initial infection of a plasmid-containing cell, recombinant plasmids with T4 DNA inserts are converted into fully modified linear DNA concatamers that are packaged into T4 phage particles, to create defective phage (transducing particles). In order to select T4 replication origins from genomic libraries of T4 sequences cloned into the plasmid pBR322, we searched for recombinant plasmids that transduce with an unusually high efficiency, reasoning that this should select for T4 sequences that function as origins on plasmid DNA after phage infection. We also selected for defective phage that can propagate efficiently with the aid of a coinfecting helper phage during subsequent rounds of phage infection. which should select for T4 sequences that can function as origins on the linear DNA present in the defective phage. Several T4 inserts were isolated repeatedly in one or both of these selective procedures, and these were mapped to particular locations on the T4 genome. When plasmids were selected in this way from genomic libraries constructed using different restriction nucleases, they contained overlapping segments of the T4 genome, indicating that the same T4 sequences were selected. The inserts in two of the selected plasmids permit a very high frequency of transduction from circular plasmids: these have been shown to contain a special type of T4 replication origin.     

Tertiary initiation of replication in bacteriophage T4. Deletion of the overlapping uvsY promoter/replication origin from the phage genome

Tertiary initiation of bacteriophage T4 DNA replication is resistant to the RNA polymerase inhibitor rifampicin and apparently involved in the activity of recombination hot spots in the T4 genome (Kreuzer, K. N., and Alberts, B. M. (1985) Proc. Natl. Acad. Sci. U. S. A. 82, 3345-3349). One of the origins that function by the tertiary mechanism maps at the promoter for gene uvs Y. A deletion and a linker-insertion mutation in the uvsY promoter/origin region were generated by in vitro manipulations and then placed into the T4 genome using the insertion/substitution system (Selick, H. E., Kreuzer, K. N., and Alberts, B. M. (1988) J. Biol. Chem. 263, 11336-11347). Both resulting phage strains are uvsY- mutants, but they differ in that one has a deletion of the minimal tertiary origin and the other does not. The effects of the uvsY mutations on tertiary origin activity were assayed by infecting tertiary origin plasmid-bearing Escherichia coli with the two phage mutants. The tertiary origin plasmids replicated extensively after infection by either uvsY- phage mutant, demonstrating that the uvsY protein is not required for tertiary initiation. The extent of plasmid replication was increased dramatically as a result of either mutation, indicating that the uvsY protein plays some negative role in either the initiation or subsequent processing of plasmid replicative intermediates. The phage strain with an origin deletion induced the replication of a tertiary origin plasmid with which it shared no homology. Therefore, plasmid-phage recombination is not required for the replication of tertiary origin plasmids. The replication of a tertiary origin plasmid is also shown to be independent of the phage genes uvsX, 59, and 46, but markedly reduced by mutations in the T4-induced topoisomerase.     

Plasmid models for bacteriophage T4 DNA replication: requirements for fork proteins.

Bacteriophage T4 DNA replication initiates from origins at early times of infection and from recombinational intermediates as the infection progresses. Plasmids containing cloned T4 origins replicate during T4 infection, providing a model system for studying origin-dependent replication. In addition, recombination-dependent replication can be analyzed by using cloned nonorigin fragments of T4 DNA, which direct plasmid replication that requires phage-encoded recombination proteins. We have tested in vivo requirements for both plasmid replication model systems by infecting plasmid-containing cells with mutant phage. Replication of origin and nonorigin plasmids strictly required components of the T4 DNA polymerase holoenzyme complex. Recombination-dependent plasmid replication also strictly required the T4 single-stranded DNA-binding protein (gene product 32 [gp32]), and replication of origin-containing plasmids was greatly reduced by 32 amber mutations. gp32 is therefore important in both modes of replication. An amber mutation in gene 41, which encodes the replicative helicase of T4, reduced but did not eliminate both recombination- and origin-dependent plasmid replication. Therefore, gp41 may normally be utilized for replication of both plasmids but is apparently not required for either. An amber mutation in gene 61, which encodes the T4 RNA primase, did not eliminate either recombination- or origin-dependent plasmid replication. However, plasmid replication was severely delayed by the 61 amber mutation, suggesting that the protein may normally play an important, though nonessential, role in replication. We deleted gene 61 from the T4 genome to test whether the observed replication was due to residual gp61 in the amber mutant infection. The replication phenotype of the deletion mutant was identical to that of the amber mutant. Therefore, gp61 is not required for in vivo T4 replication. Furthermore, the deletion mutant is viable, demonstrating that the gp61 primase is not an essential T4 protein.     

The generation of concatemeric plasmid DNA in Bacillus subtilis as a consequence of bacteriophage SPP1 infection.

Bacteriophage SPP1 infection of Bacillus subtilis cells bearing plasmids induces the synthesis of multigenome-length plasmid molecules. Two independent pathways can account for this synthesis. In one of those, homology to the phage genome is required, whereas in the other such homology is not a prerequisite. In wild type cells both modes overlap. In dnaB(Ts), at non permissive temperature, or in recE polA strains the main concatemeric plasmid replication mode is the homology-dependent plasmid (hdp) mode. The rate of recombination-dependent concatemeric plasmid DNA synthesis is a consequence of a phage-plasmid interaction which leads to chimeric phage::plasmid DNA. The second mode, which is an homology-independent plasmid (hip) mode seems to be triggered upon the synthesis of a phage encoded product(s) (e.g. inactivation of the exonuclease V enzyme).     

Initiation of DNA replication at cloned origins of bacteriophage T7

Bacteriophage T7 DNA replication is initiated at a site 15% of the distance from the genetic left end of the chromosome. This primary origin contains two tandem T7 RNA polymerase promoters (phi 1.1A and phi 1.1B) followed by an A + T-rich region. When the primary origin region is deleted replication initiates at secondary origins. We have analyzed the ability of plasmids containing cloned fragments of T7 to replicate after infection of Escherichia coli with bacteriophage T7. All cloned T7 fragments that support plasmid replication contain a T7 promoter but a T7 promoter alone is not sufficient for replication. Replication of plasmids containing the primary origin is dependent on T7 DNA polymerase and gene 4 protein (helicase/primase) and a portion of the A + T-rich region. The other T7 fragments that support plasmid replication after T7 infection are promoter regions phi OR, phi 13 and phi 6.5 (secondary origins). When both the primary and secondary origins are present simultaneously on compatible plasmids, replication of each is temporally regulated. Such regulation may play a role during T7 DNA replication.     

Fate of cloned bacteriophage T4 DNA after phage T4 infection of clone-bearing cells

Plasmid pBR322 replication is inhibited after bacteriophage T4 infection. If no T4 DNA had been cloned into this plasmid vector, the kinetics of inhibition are similar to those observed for the inhibition of Escherichia coli chromosomal DNA. However, if T4 DNA has been cloned into pBR322, plasmid DNA synthesis is initially inhibited but then resumes approximately at the time that phage DNA replication begins. The T4 insert-dependent synthesis of pBR322 DNA is not observed if the infecting phage are deleted for the T4 DNA cloned in the plasmid. Thus, this T4 homology-dependent synthesis of plasmid DNA probably reflects recombination between plasmids and infecting phage genomes. However, this recombination-dependent synthesis of pBR322 DNA does not require the T4 gene 46 product, which is essential for T4 generalized recombination. The effect of T4 infection on the degradation of plasmid DNA is also examined. Plasmid DNA degradation, like E. coli chromosomal DNA degradation, occurs in wild-type and denB mutant infections. However, neither plasmid or chromosomal degradation can be detected in denA mutant infections by the method of DNA--DNA hybridization on nitrocellulose filters.     

Integration of Plasmids into the Bacteriophage T4 Genome

We have analyzed the integration of plasmids into the bacteriophage T4 genome via homologous recombination. As judged by genetic selection for a plasmid-borne marker, a mutation in phage gene uvsX or uvsY essentially blocked the integration of a plasmid with homology to the T4 genome but no phage replication origin (non-origin plasmid). The strict requirement for these two proteins suggests that plasmid integration can proceed via a strand-invasion reaction similar to that catalyzed in vitro by the T4-encoded strand-exchange protein (UvsX) in concert with UvsY and gp32. In contrast to the results with the non-origin plasmid, a mutation in uvsX or uvsY reduced the integration of a T4 replication origin-containing plasmid by only 3-10-fold. These results suggest that the origin-containing plasmid integrates by both the UvsXY-dependent pathway used by the non-origin plasmid and by a UvsXY-independent pathway. The origin-containing plasmid integrated into the phage genome during a uvsX-or uvsY-mutant infection of a recA-mutant host, and therefore origin-dependent integration can occur in the absence of both phage- and host-encoded strand-exchange proteins (UvsX and RecA, respectively).     

DNA helicase requirements for DNA replication during bacteriophage T4 infection.

The lytic bacteriophage T4 uses multiple mechanisms to initiate the replication of its DNA. Initiation occurs predominantly at replication origins at early times of infection, but there is a switch to genetic recombination-dependent initiation at late times of infection. The T4 insertion-substitution system was used to create a deletion in the T4 dda gene, which encodes a 5'-3' DNA helicase that stimulates both DNA replication and recombination reactions in vitro. The deletion caused a delay in T4 DNA synthesis at early times of infection, suggesting that the Dda protein is involved in the initiation of origin-dependent DNA synthesis. However, DNA synthesis eventually reached nearly wild-type levels, and the final number of phages produced per bacterium was similar to that of the wild type. When the dda mutant phage also contained a mutation in T4 gene 59 (a gene normally required only for recombination-dependent DNA replication), essentially no DNA was synthesized. Recent in vitro studies have shown that the gene 59 protein loads a component of the primosome, the T4 gene 41 DNA helicase, onto DNA. A molecular model for replication initiation is presented that is based on our genetic data.     

Rolling-circle replication of bacterial plasmids.

Many bacterial plasmids replicate by a rolling-circle (RC) mechanism. Their replication properties have many similarities to as well as significant differences from those of single-stranded DNA (ssDNA) coliphages, which also replicate by an RC mechanism. Studies on a large number of RC plasmids have revealed that they fall into several families based on homology in their initiator proteins and leading-strand origins. The leading-strand origins contain distinct sequences that are required for binding and nicking by the Rep proteins. Leading-strand origins also contain domains that are required for the initiation and termination of replication. RC plasmids generate ssDNA intermediates during replication, since their lagging-strand synthesis does not usually initiate until the leading strand has been almost fully synthesized. The leading- and lagging-strand origins are distinct, and the displaced leading-strand DNA is converted to the double-stranded form by using solely the host proteins. The Rep proteins encoded by RC plasmids contain specific domains that are involved in their origin binding and nicking activities. The replication and copy number of RC plasmids, in general, are regulated at the level of synthesis of their Rep proteins, which are usually rate limiting for replication. Some RC Rep proteins are known to be inactivated after supporting one round of replication. A number of in vitro replication systems have been developed for RC plasmids and have provided insight into the mechanism of plasmid RC replication.     

Rolling circle replication of single-stranded DNA plasmid pC194.

A group of small Staphylococcus aureus/Bacillus subtilis plasmids was recently found to replicate via a circular single-stranded DNA intermediate (te Riele et al., 1986a). We show here that a 55 bp region of one such plasmid, pC194, has origin activity when complemented in trans by the plasmid replication protein. This region contains two palindromes, 5 and 14 bp long, and a site nicked by the replication protein. DNA synthesis presumably initiated at the nick in the replication origin can be terminated at an 18 bp sequence homologous to the site of initiation, deriving from another plasmid, pUB110, or synthesized in vitro. This result suggests that, similar to the Escherichia coli single-stranded DNA phages, pC194 replicates as a rolling circle. Interestingly, there is homology between replication origins and replication proteins of pC194 and the phage phi mX174.     

Replication origin of a single-stranded DNA plasmid pC194.

The replication of the single-stranded (ss) DNA plasmid pC194 by the rolling circle mechanism was investigated using chimeric plasmids that possess two pC194 replication origins. One of the origins was intact, whereas the other was either intact or mutated. The origins were activated by inducing synthesis of the pC194 replication protein, under the control of lambda phage pL promoter. Initiation of pC194 replication at one origin and termination at the other generated circular ssDNA molecules smaller than the parental chimeric plasmid. From the nature and the amount of ssDNA circles, the activity of an origin could be assessed. Our results show that (i) the signal for initiation of pC194 replication is more stringent than that for termination; (ii) the sequence and structure of the origin are important for its activity and (iii) successful termination of one replication cycle is not followed by reinitiation of another. This last observation differentiates a ssDNA plasmid (pC194) from a ssDNA phage (phi X174).     

DNA helicase mutants of bacteriophage T4 that are defective in DNA recombination

We previously isolated a plasmid-borne, recombination-deficient mutant derivative of the bacteriophage T4 DNA helicase gene 41. We have now transferred this 41rrh1 mutation into the phage genome in order to characterize its mutational effects further. The mutation impairs a recombination pathway that is distinct from the pathway involving uvsX, which is essential for strand transfer, and it also eliminates most homologous recombination between a plasmid and the T4 genome. Although 41rrh1 does not affect T4 DNA replication from some origins, it does inactivate plasmid replication that is dependent on ori(uvsY) and ori(34), as well as recombination-dependent DNA replication. Combination of 41rrh1 with some uvsX alleles is lethal. Based on these results, we propose that gene 41 contributes to DNA recombination through its role in DNA replication. Received: 3 February 1999 / Accepted: 20 July 1999     

Genetic complementation by cloned bacteriophage T4 late genes.

Bacteriophage T4 containing nonsense mutations in late genes was found to be genetically complemented by four conjugate T4 genes (7, 11, 23, or 24) located on plasmid or phage vectors. Complementation was at a very low level unless the infecting phage carried a denB mutation (which abolishes T4 DNA endonuclease IV activity). In most experiments, the infecting phage also had a denA mutation, which abolishes T4 DNA endonuclease II activity. Mutations in the alc/unf gene (which allow dCMP-containing T4 late genes to be expressed) further increased complementation efficiency. Most of the alc/unf mutant phage strains used for these experiments were constructed to incorporate a gene 56 mutation, which blocks dCTP breakdown and allows replication to generate dCMP-containing T4 DNA. Effects of the alc/unf:56 mutant combination on complementation efficiency varied among the different T4 late genes. Despite regions of homology, ranging from 2 to 14 kilobase pairs, between cloned T4 genes and infecting genomes, the rate of formation of recombinants after T4 den:alc phage infection was generally low (higher for two mutants in gene 23, lower for mutants in gene 7 and 11). More significantly, when gene 23 complementation had to be preceded by recombination, the complementation efficiency was drastically reduced. We conclude that high complementation efficiency of cloned T4 late genes need not depend on prior complete breakage-reunion events which transpose those genes from the resident plasmid to a late promoter on the infecting T4 genome. The presence of the intact gene 23 on plasmids reduced the yield of T4 phage. The magnitude of this negative complementation effect varied in different plasmids; in the extreme case (plasmid pLA3), an almost 10-fold reduction of yield was observed. The cells can thus be said to have been made partly nonpermissive for this lytic virus by incorporating a part of the viral genome.     

Evidence for a rolling-circle mechanism of phage DNA synthesis from both replicative and integrated forms of CTXphi

The genes encoding cholera toxin, the principal virulence factor of Vibrio cholerae, are part of the circular single-stranded DNA genome of CTXphi. In toxigenic V. cholerae strains, the CTXphi genome is typically found in integrated arrays of tandemly arranged CTX prophages. Infected cells that lack a chromosomal integration site harbour the CTXphi genome as a plasmid (pCTX). We studied the replication of pCTX and found several indications that this plasmid replicates via a rolling-circle (RC) mechanism. The initiation and termination sites for pCTX plus-strand DNA synthesis were mapped to a 22 bp sequence that contains inverted repeats and a nonanucleotide motif found in the plus-strand origins of several RC replicons. Furthermore, similar to other RC replicons, replication of plasmids containing duplicated pCTX origins resulted in the deletion of sequences between the two origins and the formation of a single chimeric origin. Our previous work revealed that CTX prophage arrays give rise to hybrid CTX virions that contain sequences derived from two adjacent prophages. We now report that the boundaries between the sequences contributed to virions by the upstream and the downstream prophages in an array correspond to the site at which synthesis of plus-strand pCTX DNA is initiated and terminated. These data support the model that plus-strand CTXphi DNA is generated from chromosomal prophages via a novel process analogous to RC replication.     

Initiation signals for complementary strand DNA synthesis on single-stranded plasmid DNA.

The bacteriophage 0X174 origin for (+) strand DNA synthesis, when inserted in a plasmid, is in vivo a substrate for the initiator A protein, that is produced by infecting phages. The result of this interaction is the packaging of single-stranded plasmid DNA into preformed phage coats. These plasmid particles can transduce 0X-sensitive cells; however, the transduction efficiency depends strongly on the presence in the packaged DNA strand of an initiation signal for complementary strand DNA synthesis. A plasmid with the complementary (-) strand origin of 0X inserted in the same strand as the viral (+) origin transduces 50-100 times more efficient than the same plasmid without the (-) origin of 0X. The transduction efficiency of such a particle is comparable to the infection efficiency of the phage particle. It is shown that in this system the 0X (-) origin can be replaced by the complementary strand origins of the bacteriophages G4 and M13. We have used this system to isolate sequences, from E. coli plasmids (pACYC177, CloDF13, miniF and OriC) and from the E. coli chromosome that can function as initiation signals for the conversion of single-stranded plasmid DNA to double-stranded DNA. All isolated origins were found to be dependent for their activity on the dnaB, dnaC and dnaG proteins. We conclude that these signals were all primosome-dependent origins and that primosome priming is the major mechanism for initiation of the lagging strand DNA synthesis in E. coli. The assembly of the primosome depends on the sequence-specific interaction of the n' protein with single-stranded DNA. We have used the isolated sequences to deduce a consensus recognition sequence for the n' protein. The role of a possible secondary structure in this sequence is discussed.     

Characterization and comparative sequence analysis of replication origins from three large Bacillus thuringiensis plasmids.

The replication origins of three large Bacillus thuringiensis plasmids, derived from B. thuringiensis HD263 subsp. kurstaki, have been cloned in Escherichia coli and sequenced. The replication origins, designated ori 43, ori 44, and ori 60, were isolated from plasmids of 43, 44, and 60 MDa, respectively. Each cloned replication origin exhibits incompatibility with the resident B. thuringiensis plasmid from which it was derived. Recombinant plasmids containing the three replication origins varied in their ability to transform strains of B. thuringiensis, Bacillus megaterium, and Bacillus subtilis. Analysis of the derived nucleotide and amino acid sequences indicates that the replication origins are nonhomologous, implying independent derivations. No significant homology was found to published sequences of replication origins derived from the single-stranded DNA plasmids of gram-positive bacteria, and shuttle vectors containing the three replication origins do not appear to generate single-stranded DNA intermediates in B. thuringiensis. The replication origin regions of the large plasmids are each characterized by a single open reading frame whose product is essential for replication in B. thuringiensis. The putative replication protein of ori 60 exhibits partial homology to the RepA protein of the Bacillus stearothermophilus plasmid pTB19. The putative replication protein of ori 43 exhibits weak but extensive homology to the replication proteins of several streptococcal plasmids, including the open reading frame E replication protein of the conjugative plasmid pAM beta 1. The nucleotide sequence of ori 44 and the amino acid sequence of its putative replication protein appear to be nonhomologous to other published replication origin sequences.     

UvsW protein regulates bacteriophage T4 origin-dependent replication by unwinding R-loops

The UvsW protein of bacteriophage T4 is involved in many aspects of phage DNA metabolism, including repair, recombination, and recombination-dependent replication. UvsW has also been implicated in the repression of origin-dependent replication at late times of infection, when UvsW is normally synthesized. Two well-characterized T4 origins, ori(uvsY) and ori(34), are believed to initiate replication through an R-loop mechanism. Here we provide both in vivo and in vitro evidence that UvsW is an RNA-DNA helicase that catalyzes the dissociation of RNA from origin R-loops. Two-dimensional gel analyses show that the replicative intermediates formed at ori(uvsY) persist longer in a uvsW mutant infection than in a wild-type infection. In addition, the inappropriate early expression of UvsW protein results in the loss of these replicative intermediates. Using a synthetic origin R-loop, we also demonstrate that purified UvsW functions as a helicase that efficiently dissociates RNA from R-loops. These and previous results from a number of studies provide strong evidence that UvsW is a molecular switch that allows T4 replication to progress from a mode that initiates from R-loops at origins to a mode that initiates from D-loops formed by recombination proteins.