integration of plasmids into the bacteriophage t4 genome · integration of plasmids into the...

Copyright 0 1994 by the Genetics Society of America

Integration of Plasmids Into the Bacteriophage T4 Genome

Helen W. E. Kreuzer' and Kenneth N. Kreuzer

Department of Microbiology, Duke University Medical Center, and the Duke University Program in Genetics, Durham, North Carolina 27710

Manuscript received July 10, 1992 Accepted for publication August 29, 1994

ABSTRACT We have analyzed the integration of plasmids into the bacteriophage T4 genome via homologous

recombination. Asjudged by genetic selection for a plasmid-borne marker, a mutation in phage gene U V S X 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 T 4 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 UvsXYdependent 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 strandexchange proteins (UvsX and Red, respectively).

G ENETIC recombination plays a prominent role in the life cycle of bacteriophage T4 (for review, see

MOSIG 1987,1994). Independent segregation occurs between genetic markers separated by as little as 10 ki- lobase pairs (kb) , demonstrating that every 169-kb progeny phage genome participated in multiple recombination events in the course of a single infection. Mutations that reduce recombination also impair phage growth, blocking DNA synthesis at late times of infection when the majority of phage replication normally occurs. In addition, recombination events between phage DNA molecules are thought to be required to produce the concatemeric DNA necessary for packaging. Finally, the repair of damaged T4 DNA by post-replication recombination mechanisms is also dependent upon phage recombination proteins.

Studies with multiply mutant phage have implied that T4 uses two or three recombination pathways that utilize different subsets of T4 recombination proteins ( H A " LEI" and BERGER 1975; CUNNINGHAM and BERGER 1977; also see MOSIG 1994 for recent review). Table 1 summa- rizes gene products that have been implicated in T4 recombination. Mutations that inactivate each protein have previously been shown to influence recombination between coinfecting phage DNAs, and biochemical activities have been uncovered for each. UvsX is a strand-exchange protein with many similarities to the RecA protein of Escherichia coli (FUJISAWA et al. 1985; GRIFFITH and FORMOSA 1985; YONESAKI and MINAGAWA 1985; YONESAKI et al. 1985; FORMOSA and ALBERTS 1986;

' Current address: Department of Biology, Salem College, WinstonSalem, North Carolina 27108

Genetics 138: 983-992 (December, 1994)

HINTON and NOSSAL 1986). UvsY is an accessory protein which greatly stimulates binding of UvsX to a single-stranded DNA-(ssDNA-) gp32 complex (HARRIS and GRIFFITH 1989; YONESAKI and MINAGAWA 1989; KODADEK et al. 1989). Mutations in uvsX and/or UUSY reduce recombination between rZ1 markers on coinfecting phage chromosomes by only two- to fourfold (HAMLETT and BERGER 1975; CUNNINGHAM and BERGER 1977). The product of phage gene 32is the archetypal ssDNA binding protein (ALBERTS and FREY 1970). Gp32 is important in UvsX-promoted strand invasion and, by itself, can facilitate renaturation of homologous single-stranded DNA molecules (ALBERTS and FREY 1970). Amber mutations in gene 32 are lethal to the phage and block recombination as judged by electron microscopic and density-shift analyses of intracellular phage DNA (BRESCHKIN and MOSIG 1977; DAN- NENBERG and MOSIG 1981). Under semi-permissive conditions, a gene 32 amber mutation reduces recombination between rZZ markers by about threefold (BERGER et al. 1969). The products of genes 46 and 47 reportedly form a membrane-associated exonuclease (MICKELSON and WIBERG 1981; also see ALBRIGHT and GEIDUSCHEK 1983), and mutations that inactivate gp46 and/or gp47 reduce recombination between phage rZZ markers by approximately six- to eightfold (SHAH and BERGER 1971; HOSODA et al. 1971; CUNNINGHAM and BERGER 1977; WAKEM and EBISUZAKI 1981). The product of gene 59 has recently been shown to assemble the T4 helicase/primase complex onto ssDNA in vitro (J. BARRY and B. M. ALBERTS, personal communication) and is thereby thought to play an important role in recombination-dependent T4 DNAreplication (see

984 H. W. E. Kreuzer and K. N. Kreuzer

TABLE 1

Tkncoded proteins involved in recombination

Protein Reported function‘

uvsx UVSY

DNA strand-exchange protein UvsX accessory protein

gP32 gp46/47

Single-stranded DNA binding protein

gp59 Putative exonuclease Assembly of helicase/primase

gp39/52/60 T4 DNA topoisomerase gp49 Endonuclease VII; cleaves branched structures ‘See text for references.

below). Mutations in gene 59 reduce recombination between rIImarkers by approximately two- to fourfold (SHAH 1976; CUNNINGHAM and BERGER 1977). Phage genes 39,52 and 60 encode the three subunits of the T4 type I1 DNA topoisomerase, which is an ATP- dependent DNA relaxing enzyme (KREUZER and HUANC 1983). Mutations that inactivate the T4 topoisomerase stimulate recombination between rZ1 markers by approximately two- to fivefold, perhaps because of changes in T4 DNA topology (MUFTI and BERNSTEIN 1974; LEUNC et al. 1975). Finally, T4 gene 49 encodes endonuclease VII, an enzyme that preferentially cleaves branched and mismatched DNA molecules in vitro (MIZUUCHI et al. 1982;JENSCH and KEMPER 1986; SOLAR0 et aL 1993). Gp49 is nearly essential for phage infection, presumably because the protein resolves recombination intermediates and other branched molecules as a prerequi- site to phage DNA packaging (KEMPER and BROWN 1976).

Recent studies on T4 DNA replication reveal two important connections between recombination and replication (for review, see KREUZER and MORRICAL 1994). First, phage replication origins generate recombination hotspots in the phage chromosome (YAP and KREUZER

1991), suggesting that a replication intermediate can stimulate recombination. Second, the majority of phage DNA replication requires recombination proteins and is proposed to involve the conversion of recombination intermediates into replication forks (LUDER and MOSIG 1982; MOSIC 1983). In a normal infection, this recombinationdependent replication is triggered when the single-stranded 3’ end of a newly replicated genome invades homologous DNA in the same or a different DNA molecule. Plasmids bearing homology to the infecting phage chromosome, but no T4 origin sequence, replicate during T4 infection (MATTSON et al. 1983a; KREUZER et al. 1988b). This “non-origin” plasmid replication is dependent upon phage recombination proteins and provides a convenient model system for studying the process of recombination-dependent replication (KREUZER et al. 1988b). One goal of the experiments described in this communication is to explore the role of particular phage gene products in recombination- dependent replication by comparing their involvement in the replication and in the recombination of non-origin plasmids.

We decided to approach T4 recombination mechanisms by studying the integration of a plasmid into the T4 chromosome, a recombination event that can be easily measured when the plasmid carries a selectable marker appropriate for phage T4 (SELICK et al. 1988). Plasmid integration can occur, at least in principle, by a simple reciprocal recombination event (see MATTSON

et al. 1983b; SELICK et al. 1988). It is possible that plasmid integration can also occur by a more complex series of steps. For example, MATTSON et al. (1983b) demon- strated that plasmid-phage recombination often gener- ates non-viable phage that carry multiple copies of inserted plasmid, and an intramolecular excision event could presumably convert a fraction of these non-viable phage into viable integrants with only a single plasmid insert.

In this communication, we investigate the requirements for plasmid integration into the T4 genome. The results suggest that integration can occur via two distinct pathways, one dependent upon the phage UvsX and UvsY gene products, and the other independent of UvsXY but dependent on the presence of a T4 replication origin on the plasmid.

MATERIALS AND METHODS

Materials: Restriction enzymes and DNA ligase were pur- chased from commercial sources. L broth contained NaCl (10 g/liter), Bacto-Tryptone (10 g/liter), and yeast extract (5 g/liter); it was supplemented with MgSO, (0.1% w/v) and L-tryptophan (20 mg/liter) for T4 adsorption and/or with am- picillin (25 mg/liter in liquid and 40 mg/liter in solid media) for selection of pBR322derived plasmids.

S t r a i n s E . coli strain CR63 (supD) was originally from the collection of B. M. ALBERTS (University of California, San Fran- cisco). E. coli strains AB1 (nonsuppressing) and MCSl (supD) are isogenic except that MCSl contains the supD marker (which was transduced from CR63) and a mutation of unknown origin causing proline auxotrophy (KREUZER et al. 1988a). JC5029 (Hfr thr-300 relAl rpsE2300 spoT1 ilu-318 thi-1 X ) and JC5088 (recA56; otherwise isogenicwith JC5029) were from the E. coli Genetic Stock Center (Yale University) (CLARK 1967; WILLETTS et al. 1969).

T4 strain K10 contains the following mutations: amB262 (gene 38), amS29 (gene 51 ), nd28 (denA), and rZZPT8 (an approximately 3-kb deletion that inactivates den& rZZB, rZZA and ndd; SELICK et al. 1988). T4 K10-uusYA1 is a K10 derivative which contains a 0.12-kb deletion of the gene uusY promoter and initiation codon (also removes replication origin orz(uvsY); for a complete description, see KREUZER et al. 1988a). The construction of T4 K10 derivatives with the following additional mutations has also been described (KREUZER et al. 1988b): a m 1 1 ( uusX) , amBl4 (gene 46), amHL628 (gene 59) and amNl16 (gene 39); theT4KlO-amE727 (gene 49) mutant was generated by the same procedure.

Generation of T4 HE11 and T4 HEll-uvsYA1: T4 HE11 [ 2461 nd28 (denA) rZZPT8 (denB-rZZ deletion)] was generated by crossing T4 HE10 [24A1 nd28 (denA) rZIPT8 (den& rZZ deletion) amB262 (gene 38) amS29 (gene 51) ; ENCMAN and KREUZER 19931 to a 38’ 51’ derivative of T4 K10; this cross eliminated the amber mutations in genes 38 and 51. The presence of the 24A1 mutation was monitored by the failure to grow in the absence of a gp24providing plasmid (pKK032 or pKKl208) and confirmed by Southern hybridization, the denB-rZZ deletion mutation (rZZPT8) by testing for exclusion

Plasmid Integration into T4 Genome 985

on a A lysogen (BENZER 1955), and the denA mutation ( nd28) by hydroxyurea sensitivity (WARNER et al . 1970). T4 strain HEll-uusYA1 [24A1 nd28 (denA) rIIPT8 (denB-rlldeletion) uvsYAl] was generated by crossing T4 HEll to T4 K10- uusYA1 at a ratio of 1 T4 K10-uvsYA1 to IO T4 HEll to favor the T4 HE1 1 genetic background. Among the segregants, the 24A1 mutation was monitored as described above and the uusYAl mutation was tentatively identified by the minute- plaque phenotype and the failure to generate large plaques in a cross-streak with a known uusY mutant. The presence of the uusYA1 mutation in T4 HE1 1-uusYA1 was confirmed by South- ern hybridization analysis, as described (see KREUZER et al. 1988a). Stocks of all T4 strains carrying the gene 24 deletion were propagated on E. coli MCSl containing the gp24 providing plasmid pKK032 (KREUZER and ALBERTS 1986; ENG MAN and KREUZER 1993; also see below).

Generation of T4 H E 1 1 derivatives with additional mutations: The T4 HEll derivatives used in this study were generated from genetic crosses between T4 HE11 and T4 K10 derivatives carrying the indicated mutation. For example, T4 HE1 1 was crossed with T4 K10-amBl4 (gene 4 6 ) at a ratio of ten T4 HE1 1 to one T4 KlO-amBI4 to favor the T4 HE11 genetic background. Among the segregants, the gene 4 6 , 3 8 , and 51 mutations were tested by cross-streaking with the corresponding single mutants to check for recombination. The gene 24 deletion was monitored through the requirement for a gp24providing plasmid, the denA mutation by sensitivity to hydroxyurea (WARNER et al . 1970), and the denB-rIZ deletion by exclusion on a A lysogen (BENZER 1955).

To generate the T4 HEll-uusYA1 derivatives carrying additional mutations, T4 HE1 1-uusYA1 was crossed with the appropriate T4 HE1 1 derivative at a ratio of 1:l. The desired amber mutations, the gene 24 deletion, and the denA and denB mutations were tested as described above. The presence of the uusY deletion was monitored by testing for the failure to generate large plaques in cross-streaks with known uvsY mutants, and/or by testing for complementation of the uvsY minute-plaque phenotype by a UvsY-providing plasmid.

Plasmids: All plasmids used in this study are derivatives of pBR322. Plasmid pKK032 is pBR322 with the T4 EcoRI fragment 106.984-109.432 kb inserted at the vector EcoRI site (KREUZER and ALBERTS 1986). Plasmid pKK1208 is identical to pKK032 except that pKK1208 also contains a 247-bp (EcoRI- HzndIII) fragment (T4 genome coordinates 152.881-153.127 kb) with the T4 replication origin ori(34) inserted between the EcoRI and HindIII sites of the vector (see MENKENS and KREu- ZER 1988). Plasmid pHE1105 was constructed by linearizing pKK032 with EcoRI and then ligating the linear DNA to a 730-bp EcoRI fragment containing part of the T4 frd gene; the frd DNA fragment was purified from an EcoRI digest of pBSEOf+ (SELICK et al. 1988), and originally consisted of an EcoRI-Hind111 fragment of the T4 genome (map coordinates 145.321-146.020 kb). Plasmid pHEl11 was constructed by linearizing pHEl105 with EcoRI and ligating to an approximately 250-bp EcoRI fragment containing ori(34) (fragment identical to origin fragment in pKK1208 except that an EcoRI linker is present at the HindIII site; see MENKENS and KREUZER 1988).

Plasmid integration experiments: Cells containing the indicated plasmid were grown at 37" in L broth containing am- picillin to a density of 4 X lo8 cells/ml and then infected with the indicated T4 strain at a multiplicity of 3 plaque-forming units (pfu) per cell. After a Smin incubation without shaking (for phage adsorption), the infected cultures were incubated for 1 hr at 37" with vigorous shaking. The cultures were then chilled on ice and lysis was completed by adding chloroform. Total pfu were measured by plating on E. coli MCSl containing plasmid pKKl208; this plasmid contains gene 24 and

A. gene 24

T4 24A1 T4 24A1 requires gene 24 plasmid in host cells

lntegrant grows on plasmid-free host

B. gene 24 gene 24

gene 24 gene 24 -

ori(34)

FIGURE 1.-Gene 24 integration assay. (A) During infection by T4 HE11 and derivatives, homologous recombination between the plasmid-borne frd fragment and the corresponding chromosomal gene results in the integration of the gene 24- containing plasmid into the phage chromosome. (B) The plasmids used in the gene 24 integration assay are described in MATERIALS AND METHODS.

ori(34) and allows for optimal growth of gene 24 deletion phage (ENGMAN and KREUZER 1993). The total pfu/ml values shown have been corrected for the dilution caused by addition of phage lysate. Integrants were measured by plating on E. coli MCSl or CR63. In several parallel experiments, the number of phage that fail to attach to cells during the adsorption period was found to be between about 2 X lo6 and 4 X 10' per ml. This small number of unattached phage was not subtracted from the pfu/ml values shown in the tables.

RESULTS

The gene 24 integration assay: The T4 insertion/ substitution (I/S) system was designed as a means of introducing in vitro-generated mutations into the T4 chromosome, but can also be used to conveniently measure the integration of plasmids into the phage genome (SELICK et al. 1988). Homologous recombination between the I/S phage chromosome and a plasmid bearing a T4 DNA insert results in integrant phage that can be selected by virtue of a plasmid-borne suppressor tRNA gene ( supF) . Integration of the I/S plasmid is dependent upon the presence of at least 50 base pairs (bp) of sequence homology between the plasmid and the infecting phage chromosome, and the integration


TABLE 2

Effects of T4 mutations on plasmid integration ~p

Non-origin plasmid b Origin plasmid b

Phage mutant' Ints/pfu (re]) ' SD Av pfu/mle Expts Ints/pfu (rel) SD Av pfu/ml Expts

WT 4.3 X 1 0 - ~ (1) 1.7 X 1 0 - ~ 8.2 X lo9 12 3.2 X ( 1 ) 9.3 X lop4 4.1 X 1olo 7 u v s x 2.8 X 1 0 - ~ (0.007) 7.4 X 1.3 X lo9 3 9.6 X (0.3) ND 2.4 X lo9 2 UVSY 4.9 X 1 0 - ~ (0.011) 4.7 X lop5 1.8 X lo9 9 1.2 X (0.38) 8.9 X 1 0 - ~ 1.6 X lo9

3.5 X 1 0 - ~ (0.08) 1.2 X lop4 2.2 X 10' 5 6.8 X (0.21) 2.0 X lo-: 3.8 X lo8 3

39 9.7 X 10- (2.3) ND 1.3 X lo9 2 1.0 X lo-' (3.1) ND 4.6 X lo9 2

46 6

59 1.2 X 1 0 - ~ (0.28) 8.4 X 1 0 - ~ 6.7 X 10' 4 2.9 X lo-: (0.91) 6.7 X 10- 6.0 X 10' 3 4 9 7.0 X 10-t (0.16) 2.7 X 5.0 X 10' 3 6.7 X 10- (0.21) ND 1.4 X lo9 2

' Additional mutations present in T4 HE1 1 (24A1 denA denB-rZZ deletion). bNon origin plasmid (pHE1105) and origin plasmid (pHE111) were each resident in host stain AB1; see Figure 1 for plasmid maps. Numerical average of the int/pfu values measured in each experiment; relative values (rel) compared to wild-type control are in parentheses. Standard deviation of the int/pfu values measured in each experiment; ND, not determined.

e Numerical average of the pfu/ml values measured in each experiment.

frequency increases with length of homology until a pla- teau is reached at approximately 500 bp (also see G ~ L D

BERG 1966; BAUTZ and BAUTZ 1967; DRAKE 1967; SINGER et al. 1982). The I/S assay is not ideal for analyzing the proteins necessary for plasmid-phage recombination, because most available T4 mutants carry amber mutations that could be suppressed by the supF gene of the plasmid. We therefore developed an integration assay based on a different selectable marker, the essential late gene 24 (Figure 1A). Gene 24 encodes the vertex protein of the phage head and plays no known role in phage recombination or replication (EISERLING 1983). Gene 24 was deleted from the T4 chromosome, yielding a 24A1 phage that is normally grown in the presence of a gp24 providing (supply) plasmid (ENGMAN and KREUZER

1993). The gp24 supply plasmid (pKK032) has no homology with the genome of the 24A1 phage, and therefore the supply plasmid rarely integrates into the phage genome. Phage which integrate a gene 24containing selection plasmid that does have homology to the phage genome are capable of growth on a plasmid-free host and thus can be easily quantitated. To conduct the assays, cells containing the gene 24 selection plasmid are infected with the phage of interest. The resulting lysate is titered on a host that carries a gp24 supply plasmid to measure total plaque-forming units (pfu) and on a plasmid-free host to measure integrants (ints) . Integra- tion frequency is then expressed as ints/pfu.

Two selection plasmids, pHEl105 and pHE111, were constructed for these assays (Figure 1B). Both are pBR322 derivatives containing a 2.449-kb EcoRI fragment of T4 DNA that includes the promoter and entire coding region of gene 24. As with the supply plasmid, the chromosomes of 24A1 phage strains are not homologous to the gene 24 insert fragment. Plasmid pHE111 contains a T4 replication origin [ ori(?4)] and is referred to as an origin plasmid, while pHE1105 contains no known T4 origin and is referred to as a non- origin plasmid. The T4 replication origin in pHEll1 maps within an essential gene ( ? 4 ) in the T4 genome,

and therefore plasmid integration into this gene would not yield viable progeny. To provide a target for homologous recombination between the infecting 24A1 phage and the gene 24 plasmids, both plasmids contain a 700-bp fragment of the nonessential phagefrd (dihy- drofolate reductase) gene. The plasmid integration described below is dependent on the frd DNA insert; plasmids that lack the frd insert but are otherwise identical to pHE111 and pHE1105 resulted in integration frequencies of about lo-' int/pfu (data not shown).

Non-origin plasmid integration: To determine the phage protein requirements for plasmid integration, amber or deletion mutations in genes of interest were crossed into T4 strain HE11 (genotype: 24A1 denA denE-rZZ deletion) and the resulting derivatives were tested for plasmid integration frequencies. The results of integration assays with host cells containing the non- origin plasmid (pHE1105) are presented in the left half of Table 2. A mutation in gene uusX or UVSY reduced non-origin plasmid integration by about 100-fold. The large reductions caused by the uusXand uusYmutations contrast with the very modest decreases (e.g., two- to fourfold) in phage-phage recombination caused by mutations in these genes (IIAMLEIT and BERGER 1975; CUNNING HAM and BERGER 1977; our unpublished data). These data therefore suggest that integration of a non-origin plasmid occurs via only a subset of T4 recombination pathways, and that phage-phage recombination uses at least one additional pathway (i.e., a UvsXY-independent pathway).

A mutation in gene 46 also decreased the integration frequency of the non-origin plasmid, but only by a factor of about 12, and the gene 59 and 49 mutations caused four- and sixfold reductions, respectively. The products of genes 46, 59, and 49 apparently play stimulatory though nonessential roles in the formation of integrants with the non-origin plasmid. Interestingly, a mutation in gene ?9caused an approximately twofold stimulation of plasmid integration. A mutation in gene 32 was lethal under these conditions, and therefore could not be analyzed for plasmid integration.

Plasmid Integration into T4 Genome

TABLE 3

Effects of additional mutations on plasmid integration into 24' uvsl" phage

987

Non-origin plasmid Origin plasmid Phage

mutant' Ints/pfu (rel) SD Av pfu/mle Expts Ints/pfu (rel) SD Av pfu/ml Expts

uus Yf 4.9 X 1 0 - ~ (1.0) 4.7 X 1 0 - ~ 1.8 X lo9 9 1.2 X 1 0 - ~ (1.0) 8.9 X 1.6 X lo9 6

U U S Y 59 6.9 X (1.4) ND 4.4 x lox 2 2.2 X 1 0 - ~ (1.8) 7.9 X 2.9 X lox 3 U U S Y 4 9 9.9 X 1 0 - ~ (2.0) ND 1.7 X lo9 1 4.7 X 1 0 - ~ (0.39) 2.8 X 1 0 - ~ 6.7 X lo9 3

UVSY UUSX 1.4 X (0.29) ND 3.1 X lo9 1 7.1 X 1 0 - ~ (0.59) 2.9 X 1 0 - ~ 3.6 X lo9 3

' Additional mutations present in T4 HE11 (24A1 denA denB-rII deletion). b A s in Table 2. 'Numerical average of the int/pfu values measured in each experiment; relative values (rel) compared to UUSY single mutant are in

parentheses. Standard deviation of the int/pfu values measured in each experiment; ND, not determined.

e Numerical average of the pfu/ml values measured in each experiment. /The UUSY data was shown in Table 2; it is presented again for ease of comparison.

Origindependent plasmid integration: T4 replication origins map to recombination hotspots in the phage genome and are necessary and probably sufficient for activating the recombination hotspots (KREUZER and ALBERTS 1985; YAP and KREUZER 1991). We therefore tested the effect of including a 0.25-kb fragment containing a T4 replication origin [ori(?4)] in the integration plasmid. As described above, the T4 origin is from an essential phage gene and therefore plasmid integration into this gene would not yield viable progeny.

The presence of the origin fragment in the frd- containing plasmid had little noticeable effect on plasmid integration in the wild-type infection, but stimulated integration about 30-fold in the uvsX- and uvsY- mutant infections (Table 2). This marked stimulation suggests the existence of a recombination pathway that is independent of UvsX and UvsY proteins but dependent on the presence of a T4 replication origin in the test plasmid. Although the presence of the origin fragment stimulated plasmid integration during infections by the uvsXand uvsY mutants, the total amount of origin plasmid integration in either of these mutants was about threefold less than in the wild-type infection. The sim- plest interpretation is that the origin plasmid can integrate by either of the two pathways suggested above: the UvsXYdependent pathway (accounting for about ?4 of plasmid integration in wild-type infections) and the origindependent but UvsXY-independent pathway (accounting for about '/s of plasmid integration in wild-type infections). According to this view, the mutation in uvsX or uvsY essentially eliminates the only pathway available to the non-origin plasmid and eliminates one of two pathways available to the origin plasmid.

While the origin fragment markedly stimulated plasmid integration in the uvsX- or uvsY-mutant infection, little or no stimulation was detected in the other mutant infections. However, the small or nonexistent stimulation in these infections is difficult to interpret, primarily because each of these mutations allows much more integration of the non-origin plasmid than do the uvsXor

uvsY mutations. To further explore the requirements for origindependent integration, we tested a series of double mutants each containing the uvsYA1 mutation. According to the above interpretation, the presence of the uusY mutation in the double mutants should elimi- nate the UvsXYdependent integration pathway so that the effect of the second mutation on origindependent plasmid integration could be easily assessed. We at- tempted to introduce amber mutations in each of the following genes into the uvsYA1 genetic background: uvsX, 46,59, ?9, and 49. The uvsYA1 46double mutant produced no detectable burst in the non-amber- suppressing host, indicating that the combination of these two mutations is lethal (also see WAKEM and EBI- s u m 1981). The lethality of this double mutant combination unfortunately prevented a test of plasmid integration frequencies. We also were unable to measure integration in a uvsYA1 39 double mutant, which we failed to isolate for unknown reasons (perhaps involving lethality). We tested each of the three viable double mutants in integration assays with both the non-origin and origin plasmids (Table 3). As expected from the single mutant analyses above,

origindependent integration in the uvsYA1 background was essentially unaffected by addition of a uvsX mutation. This result provides evidence that the same integration pathway is operative in uvsX and uvsY single-mutant infections and in double-mutant infections. The gene 59 mutation also had no inhibitory effect on origin-dependent plasmid integration, and the mutation in gene 49 apparently reduced the frequency of integrants by a small factor. These results therefore indicate that origindependent integration can occur in the absence of UvsX, UvsY, gp59, or gp49.

Origindependent plasmid integration occurs in a recA-mutant host: Origin-dependent plasmid integration is independent of the T4 strandexchange protein UvsX and its accessory protein UvsY. We therefore tested whether the host RecA protein participates in origindependent integration. The origin and non-origin test plasmids were introduced into isogenic recA+ and


TABLE 4

Plasmid inteption is independent of host recA gene

Non-origin plasmid Origin plasmid Host Phage

genotype a mutant Ints/pfu (rei) SD e AV pfu/mlf Ints/pfu (re11 SD Av pfu/ml

recA + WT 6.3 X 1 0 - ~ (1) 1.2 X 1 0 - ~ 1.8 X 1olo 4.3 X lop3 (1) 2.9 X 6.8 X 10" recA - WT 4.2 X lo-' (0.67) 9.8 X 1.0 X 10" 2.9 X (0.67) 2.3 X 3.1 X 10''

recA+ u v s x 7.6 X 1 0 - ~ (0.012) 9.0 X 1 0 - ~ 2.9 X lo9 6.4 X (0.15) 6.4 X 1 0 - ~ 3.1 X log recA - uusx 3.0 X 1 0 - ~ (0.005) 8.7 X 10" 1.9 X lo9 3.1 X 1 0 - ~ (0.072) 1.2 X lop5 1.9 X lo9 recA + UVSY 6.6 X (0,010) 4.2 X 2.5 X lo9 4.2 X (0.10) 2.7 X lop5 3.0 X log recA- UUSY 2.9 X (0.005) 2.8 X IO@ 1.4 X lo9 2.3 X (0.053) 2.7 X 1.4 X lo9

6 a Host strains were: recA+: JC5029; recA-: JC5088; see MATERM AND METHODS.

Non-origin plasmid, pHE1105; origin plasmid, pHE111; see Figure 1. Additional mutations present in T4 HE1 1 (24A1 denA denB-rII deletion). Numerical average of the int/pfu values measured in three experiments; relative values (rel) compared to wild-type control in recA+ host are

in parentheses. e Standard deviation of the int/pfu values measured in three experiments. /Numerical average of the pfu/ml values measured in three experiments.

recA56mutant strains, and plasmid integration frequencies were measured for the wild-type 24A1 phage and its uvsX amber mutant and uvsYA1 derivatives. All the measured integration frequencies were about %fold lower in the recA-mutant host than in the recA+ host (Table 4). We believe that this small reduction is due to some secondary effect related to the altered physiology of recA mutants, because: (1) the reduction was detected with either plasmid; (2) the reduction was detected with all three phage genotypes; and (3) all burst sizes were also reduced in the recA-mutant host.

As in the experiments above (which used a different bacterial host), the mutation in either uvsX or uvsY reduced the integration of the non-origin plasmid by about 100-fold in either the recA' or recA56 host (Table 4). These results thereby support the conclusion that nearly all integration of the non-origin plasmid requires the phage-encoded strand-exchange protein UvsX and its accessory protein UvsY. The main purpose of this experiment was to analyze origindependent plasmid integration. In the recA+ host, the presence of the T4 replication origin again stimulated plasmid integration in both the uvsX- and uvsY-mutant infections, although not as dramatically as in the previous recA+ bacterial host (approximately 8-fold us. 30-fold; compare Tables 4 and 2). We do not understand why the extent of stimulation was different in these two bacterial hosts. Nonetheless, the important conclusion is that plasmid integration was stimulated by the origin just as much in the uvsX- and uvsY-mutant infections of the recA56 host as in the recA+ host (Table 4). Thus, origindependent plasmid integration occurs in the absence of both the phage- and host-encoded DNA strand-exchange proteins.

DISCUSSION

The gene 24 integration assay: The gene 24 integration assay appears useful for studying recombination mechanisms during phage T4 infection. Mutations in

genes uvsXor uvsY reduce non-origin plasmid integration by about 100-fold (Tables 2 and 4), but reduce phage-phage recombination by only a few fold (HAMLETT

and BERGER 1975; CUNNINGHAM and BERGER 1977). Thus, non-origin plasmids apparently integrate into the phage chromosome only by a uvsXY-dependent pathway, while phage-phage recombination occurs by both uvsXY- dependent and -independent pathways. The effects of particular DNA segments on recombination can be readily tested using plasmid integration approaches. In this way, the addition of a fragment containing a T4 replication origin was found to activate plasmid integration in the absence of UvsX, UvsY or both (Tables 2-4).

Several limitations of this plasmid-based recombination assay must be considered. First, plasmids may exist in a topological form and/or subcellular localization that differs from that of the phage chromosome. At least at the very start of infection, plasmid but not phage genomic DNA would be in a negative superhelical form, which could affect recombination capacities (see SINDEN and PETTIJOHN 1982; ALBRIGHT and GEIDUSCHEK 1986). Second, unreplicated plasmid DNA is not modified with the glucosyl and hydroxymethyl groups characteristic of native T4 DNA, and this might affect participation of plasmid DNA in phage-promoted recombination. For example, the origindependent pathway of plasmid integration could require a plasmid-borne origin only to generate modified plasmid DNA that acts as a substrate in a recombination pathway that requires modified DNA. Third, as already mentioned in the Introduction, it is not certain that plasmid integration occurs only by a simple reciprocal recombination event. Fourth, the assay is compromised by unexpected plasmid replication caused by the gene 24 DNA insert. The promoter region of gene 24 induces a small amount of plasmid replication that is independent of T4 recombination proteins (ENGMAN and KRFAJZER 1993), and therefore the "non-origin" plasmid used in this study replicates weakly.

Plasmid Integration into T4 Genome 989

This low level of replication may contribute to the background level of non-origin plasmid integration in the uvsY (and presumably U V S X ) mutant infections. This inference is based on an independent measure of plasmid integration using the T4 I/S system, which relies on a plasmid containing the supF selectable marker instead of gene 24. Integration of the non-origin supF- containing plasmid was reduced nearly lo5-fold by the uvsYmutation (unpublished data), rather than the 100- fold reduction detected with the non-origin gene 24- containing plasmid. Nevertheless, the small amount of replication of the gene 24containing plasmid does not prevent detection of a strong stimulation of integration by the origin-containing fragment.

Recombination-dependent DNA replication: The majority of DNA synthesis during T4 infection occurs at late times of infection and is dependent upon the products of T4 genes UVSX, UVSY, 46/47,59, and 39/52/60 (among others; for reviews, see MOSIG 1983; KREUZER and MORRICAL 1994). This late DNA replication was proposed to initiate via a unique mechanism in which recombination intermediates are converted into replication forks (LUDER and MOSIC 1982). In the first step of this model, a single-stranded 3' end (generated by the replication of linear DNA) invades an homologous duplex DNA to create a recombination intermediate. In the second step, the intermediate is converted into a replication fork, presumably utilizing the invading 3' end as a primer for leading-strand synthesis.

Plasmids bearing homology to the infecting phage chromosome, but no T4 origin sequence, are replicated during T4 infection by a process that mimics phage recombinationdependent replication (KREUZER et al. 1988b). The majority of the replicated plasmid product is in the form of long plasmid concatamers, arguing strongly that plasmid replication is an active process and not simply the result of passive replication after the integration of plasmids into the phage genome. Assuming that recombination-dependent plasmid replication involves the conversion of a phage-plasmid recombination intermediate into a replication fork, it should be possible to distinguish between phage proteins involved in the two steps outlined above. Mutations inactivating proteins required for the first step should abolish both recombination-dependent plasmid replication and plasmid-phage recombination, as do mutations in uvsx and uvsY (Table 2; KREUZER et al. 1988b). The products of these genes are therefore implicated in the formation of recombination intermediates during recombination- dependent replication. The assignment of these two proteins in the generation of a plasmid-phage recombination intermediate fits well with the biochemical functions of these proteins in the strandexchange reaction (see below). A mutation in gene 46 reduces both plasmid replication (KREUZER et al. 1988b) and plasmid integration (Table 2) by about 10-20-fold, suggesting

that gp46/47 also participates in formation of a plasmid- phage recombination intermediate. Mutations in genes whose products are required for the second step (the conversion of recombination intermediates into replication forks) should block plasmid replication but not plasmid-phage recombination. The gene 39 mutation greatly reduces recombination-dependent plasmid replication (KREUZER et al. 1988b) but not plasmid-phage recombination (Table 2). The T4 topoisomerase may therefore be involved in either the initiation of replication or the subsequent propagation of the fork (perhaps as a replication fork swivelase).

Two pathways of plasmid integration: Previous studies of T4 recombination pathways compared phage- phage recombination frequencies from singly and multiply mutant phage infections. In the most compre- hensive analysis, CUNNINGHAM and BERGER (1977) con- cluded that UvsX and UvsY act independently of gp46/ 47, while all four of these gene products act independently of gp59. Based on such analyses of gene product requirements, at least two or three different pathways for homologous recombination were inferred (HAMLEIT and BERGER 1975; CUNNINCHAM and BERGER 1977). Based on a consideration of the structures of recombination substrates and products, MOSIG (1994) ar- gued that T4 uses four recombination pathways, two of which predominate in a wild-type infection. The possible relationships between the pathways proposed from gene product requirements and those proposed by MOSIG (1994) are not clear.

Our studies suggest that plasmids can integrate into the T4 chromosome by two distinct pathways; one requires the phage genes uvsXand UVSY, while the other is independent of these two genes but requires a T4 replication origin insert fragment on the test plasmid. Both pathways involve homologous recombination in the f rd target region because plasmids lacking the frd insert fail to integrate (data not shown). Recombination of the non-origin plasmid with the phage genome is strongly dependent upon the products of phage genes UVSX and UVSY (Table 2), suggesting that a well- characterized strandexchange reaction is involved. The invasion of single-stranded DNA into homologous duplex DNA is efficiently catalyzed by a combination of highly purified UvsX, UvsY and gp32 (YONESAKI and MI- NAGAWA 1989; I - h m S and GRIFFITH 1989). A mutation in gene 46 reduced integration of the non-origin plasmid by about 12-fold, suggesting that the gp46/47 complex stimulates but is not required for integration. Based on the inferred exonuclease activity of gp46/47 ( MICKELSON and WIBERC 1981), this protein may generate a single- stranded DNA substrate for the strandexchange reaction catalyzed by UvsX, UvsY and gp32.

Our results also suggest a second pathway for plasmid integration when the test plasmid contains a T4 replication origin insert. The insert fragment containing

990 H. W. E. Kreuzer and K N. Kreuzer

ori(34) is quite small (0.25 kb) and contains no other known sequences involved in T4 DNA metabolism. The previously defined replication origin, comprising about half of this fragment (MENKENS and KREUZER 1988), is thus likely to be responsible for stimulating integration. Based on data from both single- and double-mutant tests, integration via the origindependent pathway does not require the products of uvsx uvsx 59 and 49 (although it might be stimulated by one or more of these proteins).

Because the origin-dependent plasmid integration was found to be independent of the T4encoded strandexchange protein UvsX and its accessory protein UvsY, we tested for a possible requirement for the host RecA protein. Origin-dependent integration was found to occur during a uvsX- or uvsY-mutant infection of a recA- host (Table 4), and therefore origin-dependent plasmid integration is not catalyzed by either of these two strandexchange proteins. Considering the lack of requirement for UvsX and RecA, we speculate that origin-dependent integration may proceed via a mechanism different from strand invasion. One interesting possibility is that origindependent integration involves single strand annealing promoted by gp32, which anneals homologous single strands in vitro (ALBERTS and FREY 1970). A T4 recombination pathway requiring gp32 (and gp46/47) and proceeding via a single strand annealing mechanism has previously been proposed, based upon electron micro- graphic studies of branched DNA structures isolated from replication-deficient phage infections (BROKER and LEHMAN 1971). The function of the replication origin in plasmid integration could be to provide single-stranded DNA regions for gp32-promoted annealing. Nonrepli- cating plasmid DNA might not provide single-stranded regions for gp32-promoted annealing, in part because its closed circular structure precludes the putative exonuclease action of gp46/47. Single-stranded regions could be generated as displacement loops during r e p lication initiation or as the result of the action of a helicase loaded at the origin. Alternatively, plasmid replication might generate a suitable substrate for exonuclease action by gp46/47, such as a double-strand break or the end of a rolling-circle replicative intermediate. The stimulation of plasmid integration by a phage origin is intriguing because T4 replication origins generate recombination hotspots in the phage chromosome (YAP and KREUZER 1991). At least some of the hotspot recombination occurs by a UvsY-independent mechanism (YAP and KREUZER 1991), and so the en- hanced recombination at the hotspots could be mechanistically related to the origindependent plasmid integration described here.

Other homologous recombination pathways that are independent of R e d or analogous strandexchange proteins have also been described. These include plasmid recombination by the RecE pathway of E. coli (FISHEL et al. 1981; LABAN and COHEN 1981; SYMINGTON

et al. 1985) and the A Red pathway ( SHULMAN et al. 1970; RADDINC 1970). Interestingly, the A Red pathway is independent of host R e d protein only when phage DNA replication is allowed (STAHL et al. 1985; STAHL and STAHL 1986); replication could thus play a similar role in the A Red pathway and in T4 origindependent plasmid integration. The RecE and Red pathways have recently been shown to utilize a second family of strandexchange proteins, mechanistically distinct from RecA or UvsX (HALL et al. 1993; HALL and KOLODNER 1994; also see KMIEC and HOLLOW 1981; MUNIYAPPA and RADDINC 1986). In vitro strand exchange by members of this second family proceeds as follows: (1) exonucleo- lytic degradation of the duplex to expose a 3' single- stranded end, (2) annealing of the eroded end with the complementary single-stranded molecule, and (3) strand exchange through the remaining duplex region. Therefore, as an alternative to the (gp32dependent) single strand annealing model proposed in the para- graph above, T4 origin-dependent plasmid integration could depend on a member of this second family of strand-exchange proteins. T4 origindependent integration presumably does not involve the host RecE pathway, because the recA strain used in our experiments (Table 4) does not contain an sbcA mutation that is necessary for induction of this pathway. Perhaps origindependent integration depends on a Tkncoded protein which is an unrecognized member of this second family of strandexchange proteins. In this context, gp46/47 seems like an excellent candidate, given its recognized role in T4 recombination and its putative exonuclease activity (see Introduction).

In closing, it is interesting to note that the gene product requirements for integration of the origin- containing plasmid closely parallel gene product requirements that have previously been deduced for phage-phage recombination (compare data in Table 2 to values quoted in the Introduction). The optimis- tic interpretation is that the two plasmid integration pathways that operate with the origin-containing plasmid mimic two major pathways of phage-phage recombination. Further work should test the validity of this interpretation, and explore additional mechanistic details of the two plasmid integration pathways.

We thank CATHY LAURIE and DON BURDICK for helpful discussions. We thank the reviewers of this paper for their substantial efforts toward improving the manuscript. This work was supported by Grant GM34622 from the National Institutes of Health. H.W.E.K. was sup ported by National Research Service Awards 5T32 GM07754 and 5T32 CAO9111 from the National Institutes of Health. This work was done during the tenure of an Established Investigatorship from the Ameri- can Heart Association to K.N.K.

LITERATURE CITED

ALBERTS, B. M., and L. FREY, 1970 T4 bacteriophage gene 32: a struc- tural protein in the replication and recombination of DNA. Nature 227: 1313-1318.

Plasmid Integration into T4 Genome 99 1

ALBRIGHT, L. M., and E. P. GEIDUSCHEK, 1983 Site-specific cleavage of T4 DNA associated with the absence of gene 46 product function. J. Virol. 47: 77-88.

ALBRIGHT, L. M., and E. P. GEIDUSCHEK, 1986 Topoisomerization of plasmid DNA in Escherichia coli infected with bacteriophage T4. J. Mol. Biol. 190 329-341.

BAUTZ, F. A,, and E. K. F. BAUTZ, 1967 Transformation in phage T4: minimal recognition length between donor and recipient DNA. Genetics 57: 887-895.

BENZER, S., 1955 Fine structure of a genetic region in bacteriophage. Proc. Natl. Acad. Sci. USA 41: 344-354.

BERGER, H., A. J. WARREN and K. E. FRY, 1969 Variations in genetic recombination due to amber mutations in T4D bacteriophage. J. Virol. 3: 171-175.

BRESCHKIN, A., and G. MOSIG, 1977 Multiple interactions of a DNA- binding protein i n vivo. I. Gene 32 mutations of phage T4 inactivate different steps in DNA replication and recombination. J.

BROKER, T. R., and I. R. LEHMAN, 1971 Branched DNA molecules: Intermediates in T4 recombination. J. Mol. Biol. 6 0 131-149.

CLARK, A. J., 1967 The beginning of a genetic analysis of recombination proficiency. J. Cell. Physiol. 70 (Suppl. 1): 165- 180.

CUNNINGHAM, R. P., and H. BERGER, 1977 Mutations affecting genetic recombination in bacteriophage T4D. I. Pathway analysis. Virol-

DANNENBERG, R., and G. MOSIC, 1981 Semiconservative DNA replication is initiated at a single site in recombinationdeficient gene 32 mutants of bacteriophage T4. J. Virol. 4 0 890-900.

DRAKE, J. W., 1967 The length of the homologous pairing region for genetic recombination in bacteriophage T4. Proc. Natl. Acad. Sci. USA 58: 962-966.

EISERLING, F. A., 1983 T4 Structure and Initiation of Infection, pp. 11-24 in Bacteriophage T4, edited by C. K. MATHEWS, E. M. KUTTER, G. MOSIG and P. B. BERGET. American Society for Microbiology, Washington, D.C.

ENGMAN, H. W., and K. N. KREUZER, 1993 Deletion of the essential gene 24 from the bacteriophage T4 genome. Gene 123: 69-74.

FISHEL, R. S., A. A. JAMES and R. KOLODNER, 1981 RecA-independent general genetic recombination ofplasmids. Nature 294: 184-186.

FORMOSA, T., and B. M. ALBERTS, 1986 Purification and characterization of the T4 bacteriophage uvsX protein. J. Biol. Chem. 261: 6107-6118.

FUJISAWA, H., T. YONESAIU and T. MINAGAWA, 1985 Sequence of the T4 recombination gene, UVSX, and its comparison with that of the r e d gene of Escherichia coli. Nucleic Acids Res. 1 3 7473-7481.

GOLDBERG, E. B., 1966 The amount of DNA between genetic markers in phage T4. Proc. Natl. Acad. Sci. USA 56: 1457-1463.

GRIFTITH, J., and T. FORMOSA, 1985 The uvsX protein of bacteriophage T4 arranges single-stranded and double-stranded DNA into similar helical nucleoprotein filaments. J. Biol. Chem. 260: 4484- 4491.

HALL, S. D., and R. D. KOLODNER, 1994 Homologous pairing and strand exchange promoted by the Escherichia coli RecT protein. Proc. Natl. Acad. Sci. USA 91: 3205-3209.

HALL, S. D., M. F. KANE and R. D. KOLODNER, 1993 Identification and characterization of the Escherichia coli RecT protein, a protein encoded by the recE region that promotes renaturation of h e mologous single-stranded DNA. J. Bacteriol. 175: 277-287.

HAMLETT, N. V., and H. BERGER, 1975 Mutations altering genetic recombination and repair of DNA in bacteriophage T4. Virology.

HARRIS, L. D., and J. D. GRIFFITH, 1989 UvsYprotein of bacteriophage T4 is an accessory protein for i n vitro catalysis of strand exchange. J. Mol. Biol. 206 19-27.

HINTON, D. M., and N. G. NOSSAL, 1986 Cloning of the bacteriophage T4 UVSX gene and purification and characterization of the T4 UVSX recombination protein. J. Biol. Chem. 261: 5663-5673.

HOSODA, J., E. MATHEWS and B. JANSEN, 1971 Role of genes 46 and 47 in bacteriophage T4 reproduction. J. Virol. 8: 372-387.

JENSCH, F., and B. KEMPER, 1986 Endonuclease VI1 resolves Y-junctions in branched DNA i n vitro. EMBO J. 5: 181-189.

KEMPER, B., and D. T. BROWN, 1976 Function of gene 49 of bat- teriophage T4. 11. Analysis of intracellular development and

Mol. BIOI. 112: 279-294.

ogy. 80: 67-82.

63: 539-567.

the structure of very fast-sedimenting DNA. J. Virol. 18:

KMIEC, E., and W. K. HOLLOW, 1981 p protein of bacteriophage A promotes renaturation of DNA. J. Biol. Chem. 256: 12636-12638.

KODADEK, T., D.-C. GAN and K. STEMKE-HALE, 1989 The phage T4 UVSY recombination protein stabilizes presynaptic filaments. J. Biol. Chem. 264: 16451-16457.

KREUZER, K. N., and B. M. ALBERTS, 1985 A defective phage system reveals bacteriophage T4 replication origins that coincide with recombination hot spots. Proc. Natl. Acad. Sci. USA 82: 3345-3349.

KREUZER, K. N., and B. M. ALBERTS, 1986 Characterization of a defective phage system for the analysis of bacteriophage T4 DNA replication origins. J. Mol. Biol. 188: 185-198.

KREUZER, R N., and W. M. HUANC, 1983 T4 DNA topoisomerase, pp. 90-96 in Bacteriophage T4: edited by C. K. MATHEWS, E. M. KUT- TER, G. MOSIG and P. B. BERGET. American Society for Microbiol- ogy, Washington, D.C.

KREUZER, K. N., and S. W. MORRICAL, 1994 Initiation of DNA replication, pp. 28-42 in Molecular Biology of Bacteriophage T4: edited by J. KARAM. American Society for Microbiology, Washington, D.C.

KREUZER, K. N., H. W. ENGMAN and W. Y. YAP, 1988a Tertiary initiation of replication in bacteriophage T4. Deletion of the overlapping uvsY promoter/replication origin from the phage genome. J. Biol. Chem. 263: 11348-11357.

KREUZER, K. N., W. Y. YAP, A. E. MENKENS and H. W. ENGMAN, 1988b Recombination- dependent replication of plasmids during bacteriophage T4 infection. J. Biol. Chem. 263: 11366-11373.

LABAN, A,, and A. COHEN, 1981 Interplasmidic and intraplasmidic recombination in Escherichia coli K-12. Mol. Gen. Genet. 184:

LEUNG, D., M. T. BEHME and K. EBISUZAKI, 1975 Effect of DNA delay mutations of bacteriophage T4 on genetic recombination. J. Virol. 16: 203-205.

LUDER, A., and G. MOSIG, 1982 Two alternative mechanisms for initiation of DNA replication forks in bacteriophage T 4 priming by RNA polymerase and by recombination. Proc. Natl. Acad. Sci.

MATTSON, T., G. VAN HOUWE, A. BOLLE and R. EPSTEIN, 1983a Fate of cloned bacteriophage T4 DNA after phage T4 infection of clone- bearing cells. J. Mol. Biol. 169: 343-355.

MATTSON, T., G. VAN H o r n and R. EPSTEIN, 1983b Recombination between bacteriophage T4 and plasmid pBR322 molecules containing cloned T4 DNA. J. Mol. Biol. 169: 357-379.

MENKENS, A. E., and R N. KREUZER, 1988 Deletion analysis of bacteriophage T4 tertiary origins. A promoter sequence is required for a rifampicin-resistant replication origin. J. Biol. Chem. 263: 11358-11365.

MICKELSON, C., and J. S. WIBERG, 1981 Membrane-associated DNase activity controlled by genes 46 and 47 of bacteriophage T4D and elevated DNase activity associated with the T4 das mutation. J. Virol. 40: 65-77.

MIZUUCHI, K., B. KEMPER, J. HAE and R. A. WEISBERG, 1982 T4 endonuclease VI1 cleaves Holliday structures. Cell 29: 357-365.

MOSIG, G., 1983 Relationship of T4 DNA replication and recombination, pp. 120-130 in Bacteriophage T4; edited by C. K. MATHEWS, E. M. KUTTER, G. MOSIC and P. B. BERGET. American Society for Microbiology, Washington, D.C.

MOSIG, G., 1987 The essential role of recombination in phage T4 growth. Annu. Rev. Genet. 21: 347-371.

MOSIG, G., 1994 Homologous recombination in bacteriophage T4, pp. 54-82 in Molecular Biology of Bacteriophage T4, edited by J. KARAM. American Society for Microbiology, Washington, D.C.

MUFTI, S., and H. BERNSTEIN, 1974 The DNAdelay mutants of bacteriophage T4. J. Virol. 14 860-871.

MUNIYAPPA, R, and C. M. RADDING, 1986 The homologous recombination system of phage A. Pairing activities of p protein. J. Biol. Chem. 261: 7472-7478.

RADDING, C. M., 1970 The role of exonuclease and p protein of bacteriophage A in genetic recombination. I. Effects of red mutants on protein structure. J. Mol. Biol. 52: 491-499.

SELICK, H. E., K. N. KREUZER and B. M. ALBERTS, 1988 The bacteriophage T4 insertion/substitution vector system. A method for

Biol. Chem. 263: 11336-11347. introducing site-specific mutations into the virus chromosome. J.

1000-1015.

200-207.

USA 79: 1101-1105.


SHAH, D. B., 1976 Replication and recombination of gene 59 mutant of bacteriophage T4D. J. Virol. 17: 175-182.

SHAH, D. B., and H. BERGER, 1971 Replication of gene 46-47 amber mutants of bacteriophage T4D. J. Mol. Biol. 57: 17-34.

SHULMAN, M. J., L. M. HALLICK, H. ECHO= and E. R. SIGNER, 1970 Prop erties of recombinationdeficient mutants of bacteriophage lambda. J. Mol. Biol. 52: 501-520.

SINDEN, R. R., and D. E. PETTIJOHN, 1982 Torsional tension in intracellular bacteriophage T4 DNA. Evidence that a linear DNA duplex can be supercoiled in v ivo . J. Mol. Biol. 162: 659-677.

SINGER, B. S., L. GOLD, P. GAUSS and D. H. DOHERTY, 1982 Determi- nation of the amount of homology required for recombination in bacteriophage T4. Cell 31: 25-33.

SOLARO, P. C., K. BIRKENKAMP, P. PFEIFFER and B. KEMPER, 1993 Endo- nuclease VI1 of phage T4 triggers mismatch correction in v i tro. J. Mol. Biol. 230: 868-877.

STAHL, F. W., and M. M. STAHL, 1986 DNA synthesis at the site of a Red-mediated exchange in phage A. Genetics 113: 1-12.

STAHL, F. W., I. KOBAMHI and M. M. STAHL, 1985 In phage lambda, cos is a recombinator in the red pathway. J. Mol. Biol. 181: 199-209.

SWINGTON, L. S., P. MORRISON and R. KOLODNER, 1985 Intramo- lecular recombination of linear DNA catalyzed by the Es- cherichia coli Re& recombination system. J. Mol. Biol. 186: 515-525.

WAKEM, L. P., and K. EBISUZAIU, 1981 DNA repair-recombination functions in the DNA processing pathway of bacteriophage T4. Virology. 112: 472-479.

WARNER, H. R., D. P. SNUSTAD, S. E. JORGENSEN, and J. F. KOERNER, 1970 Isolation of bacteriophage T4 mutants defective in the ability to degrade host deoxyribonucleic acid. J. Virol. 5: 700-708.

WILLETTS, N. S., A. J. CLARK and B. Low, 1969 Genetic location of certain mutations conferring recombination deficiency in Es- cherichia coli. J. Bacteriol. 97: 244-249.

YAP, W. Y., and K. N. KREUZER, 1991 Recombination hotspots in bacteriophage T4 are dependent on replication origins. Proc. Natl. Acad. Sci. USA 88: 6043-6047.

YONESAKI, T., and T. MINAGAWA, 1985 T4 phage gene uvsX product catalyzes homologous DNA pairing. EMBO J. 4: 3321-3327.

YONESAKI, T., and T. MINAGAWA, 1989 Synergistic action of three recombination gene products of bacteriophage T4, uvsX, UVSY, and gene 32 proteins. J. Biol. Chem. 264: 7814-7820.

YONESAKI, T., Y. RYO, T. MINAGAWA and H. TAKAHASHI, 1985 Purifica- tion and some of the functions of the products of bacteriophage T4 recombination genes uvsX and UVSY. Eur. J. Biochem. 148: 127-134.

Communicating editor: G. R. SMITH

integration of plasmids into the bacteriophage t4 genome · integration of plasmids into the...

Documents