Initiation of recA+-dependent recombination in Escherichia coli (λ)

Initiation of recA+-dependent recombination in Escherichia coli (λ)

J. Mol. Biol. (1977) 117, 159-174 Initiation of recA+-dependent Recombination Escherichia coli (1) in II. Specificity in the Induction of Recombi...

1MB Sizes 7 Downloads 31 Views

J. Mol. Biol. (1977)

117, 159-174

Initiation

of recA+-dependent Recombination Escherichia coli (1)

in

II. Specificity in the Induction of Recombination and Strand Cutting in Undamaged Covalent Circular Bacteriophage 186 and Lambda DNA Molecules in Phage-infected Cells PETER Ross AND PAUL HOWARD-FLANDERS Department of Molecular Biophysics and Biochemistry and Department of Therapeutic Radiology, Yale University New Haven, Conn. 06520> U.S.A. (Received 31 May lY77) Covalent circular X DNA molecules produced in Escherichia coli (h) host cells by infection with labeled X bacteriophages are cut following superinfection with h phages damaged by exposure to psoralen and 360 nm light. This cutting of undamaged covalent circular molecules is referred to as “cutting in truns”, and could be a step in damage-induced recombinat,ion (Ross & Howard-Flanders, 1977). Similar experiments performed with the temperate phage 186, which is not homologous with phage X, showed cutting in tram and damage-induced recombination to occur in homoimmune crosses with phage 186 also. Double lysogens carrying both h and 186 prophages were used in a test for specificity in cutting in tram- and in damage-induced recombination. The double lysogens were infected with 3H-labeled 186 and 32P-labeled h phages. When these doubly infected lysogens containing covalent circular phage DNA molecules of both types were superinfected with psoralen-damaged 186 phages and incubated, the covalent circular 186 DNA was cut, while h DNA remained intact. Similarly, superinfection with damaged X phages caused /\, but not 186, DNA to be cut. Evidently, cutting in tram was specific to the covalent circular DNA homologous to the DNA of the damaged phages. Homoimmune phageprophage genetic crosses were performed in the double lysogenic host infected with genetically marked X and 186 phages. Damage-induced recombination was observed in this system only between the damaged phage DNA and the homologous prophage, none being detected between other homolog pairs present in the same cell. This result makes it unlikely that the damaged phage DNA induces a general state of enhanced strand cutting and genetic recombination affecting all homolog pairs present in the host cell. The simplest interpretation of the specificity in cutting and in recombination is as follows. When they have been incised, the damaged pllage DNA molecules are able to pair directly with their urldamaged covalent circular homologs. The latter molecules are cut in a, rerA + -dependent reaction bJ a recombination endonucleascx that cuts the int,art member of the paired homologs.

1. Introduction When Escherichia COG (A) cells are infected with h bacteriophages carrying appropriate genetic markers, genetic exchanges occur between phage and prophage at a low frequency. The frequency of recombination between closely spaced markers in this phage-prophage cross can be increased more than 50-fold by treating the phages 159

160

P. ROSS

AKD

P. HOWARD-FLANDERS

with psoralen and light before infection (Lin et al., 1977). Furthermore, if E. coli (h) cells are infected with labeled A phages, the covalent circular X DNA molecules so produced are cut if the cells are superinfected with psoralen-damaged phages. This cutting of covalent circular molecules is referred to as “cutting in tram” (Ross $ Howard-Flanders, 1977). Both effects, the increase in the frequency of phage-prophage recombination, and the cutting in tram, occur 0nl.v in excision-proficient (uvrA + ), recombination-proficient (recA + ) cells, and require the presence of psoralen damage in the DNA of the superinfecting phages. Cutting in tram is of interest, because it could arise from the cutting and joinin, v of DNA molecules. which is thought to underlie genetic recombination. We have further investigated damage-induced recombination and cutting iu tram of phage DNA molecules in homoimmune lysogenic cells of E. co&. These experiments were to determine whether cutting in tram and damage-induced genetic recombination are initiated by direct interaction between damaged DNA molecules and their homologs, or whether damaged DNA stimulates cutting and recombination between DNA molecules present in the same cell even if they are not homologous to the damaged DNA. In these tests, we used phage 186 as well as phage h systems. Phage 186 is related structurally and genetically to phage P2 (Baldwin et al.. 1966; Bertani t Bertani, 1971; Younghusband & Inman, 1974; Bradley et al., 1975). Phage 186 contains a non-permuted, double-stranded linear DNA duplex of molecular weight 2.0~ lo7 with cohesive ends (Baldwin et al., 1966; Wang, 1967; Skalka $ Hanson, 1972). Phages h and 186 are capable of simultaneous growth in t,he sarnca cell, producing single bursts containing phages of both types. Moreover, double lysogens carrying both A and 186 prophages can be constructed (Baldwin et al., 1966 : Mandel 6 Kornreich, 1972). However, there is little. if any, base sequence homology shared by h and 186 as tests for ba,se-pairing between single DNA strands of the trio phages have yielded negative results (Skalka & Hanson, 1972), and efforts to detect the formation of 186-h hybrid recombinant phages were unsuccessful (,Jacob 85 Wollman, 1961). The experiments to be described tested for cutting in covalent circular A DNA and for h phage-prophage recombination in response to superinfect’ion with pjoralendamaged 186 phages. Reciprocal phage-prophage crosses were performed in which the roles of the h and 186 phages were reversed. The detection of cutting in trans and damage-induced recombination between non-homologous phages would indicate that base-pair homology was not required, and would be consist,ent with indirect mechanisms in which the levels of repair or recombination enzymes were altered in response to the presence in the cells of the psoralen-damaged phage DNA. On t,he other hand, a failure to detect either cutting in bans, or damage-induced recombination, would be consistent with the need for homologous pairing early in these processes. The 186 phage system was found to behave similarly to the phage X system in these experiments. However, our results show that psoralen-damaged 186 phages failed to induce either cutting in tram or phage-prophage recombination in the h system, and that psoralen-damaged h phages were without effect in the reciprocal tests on the phage 186 systems. Accordingly, we conclude that indirect mechanisms are unlikely. The simplest interpretation is tha,t complementary base-pairing occurred between the damaged and undamaged DNA before the undamaged second duplex was cut and then joined to form a recombinant.

recA

+-DEPENDENT

CUTTING

IN

DNA

161

MOLECULES

2. Materials and Methods (a) Phagee and bacteria Phages X ~I.857 P3 and h ~1857 P80 were from R. P. Boyce, phages 186, 186 CRY, by J. B. Egan. The muta186 cIts 4-7, 186 cIta 12-19, 186 vir and 186 turn were provided tions P3, P80, 4-7 and 12-19 are suppressed by supE44 in strain C600. Gene 4 is necessary for formation of the tail of phage 186, while gene 12 is required for the head (J. B. Egan, personal communication). Like phage /\, wild-type prophage 186 can be induced to vegetative growth by DNA damaging agents, or by heating. Phage 186 carrying the turn mutation is resistant to induction by mitomycin C (Woods & Egan, 1974). Other phages and bacteria were described in the preceding paper (Ross & Howard-Flanders, 1977). (b) Media L Broth, LCG medium, X diluent, X buffer, KM, and KM + 10 media were described in the preceding paper (Ross & Howard-Flanders, 1977). KM + 10 + rha medium is KM + 10 supplemented with 5 ml of 20% rhamnose/l. 186 adsorption buffer contains : 0.01 ivr-Tris.HCl (pH 7*2), 0.01 M-MgCl,, 0.1 M-NaCl and 0.005 M-CaCl,. Agar for plating phage h contained (per liter of water): 5.0 g Bacto tryptone, 8.0 g Bacto peptone, 1.0 g NaCl, and 15.0 g Bacto agar. For plating phage 186, the agar was made 5 mM in CaCl, and 1% in NaCl before pouring. Top agar contained 0.6% (w/v) Bacto agar. For plating phage 186, it was made 1% in NaCl before pouring.

(c) Preparation

of phage

Preparation of A phage was as described (Ross & Howard-Flanders, 1977). Preparation of phage 186: phage 186 cite was prepared by heat induction of E. coli (186 &a). Cells were grown in LCGBO medium to 5 x lOa/ml at 30°C, transferred to 42°C for 20 min, and aerated at 40°C until lysis. 3H-labeled phage 186 was prepared by thermal induction of E. coli (186 cIts). Cells were grown in LCG medium containing 6 pg thymine/ml, to 5 x lO*/ml, centrifuged and resuspended in fresh medium containing 1 pg cold thymine/ml, held at 42°C for 20 min, supplemented with 3 &i [3H]thymidine/ml (New England Nuclear; spec. act. 1.2 Ci/mmol), and aerated at 40°C until lysis occurred. Lysates were shaken with 0.6 ml chloroform/l and 2.5 mg DNAase/ml. Yields of 5 x log to lOlo phages/ml were obtained. The phages were centrifuged in a Beckman SW27 rotor for 2 h, at 25,000 revs/min, through C&l step gradients and the band (at density about 1.4) was collected and dialyzed against X diluent. Unlabeled phages were stored at 2 x lo”, and 3H-labeled phages at 3 x lOlo phages/ml and lo5 cts/min per ml in X diluent which was made O.lyA in bovine serum albumin. (d) Treatment with psoralen and light Treatment of CsCl-purified phages was as described in the preceding paper (Ross & Howard-Flanders, 1977), except that phage 186 was resuspended in 186 adsorption buffer. In the genetic experiments described in Fig. 3, phages were irradiated in crude L broth lysates using about twice the dose of 360 nm light required to inactivate CsClpurified phages to the same extent.

3. Results (a) Circularization

of undamaged

and of psoralen-damaged

phage 186

DNA in E. coli (186) and test for cutting in trans We first asked whether phage 186 would form covalent circles upon infection of E. co& (186) and whether the fraction of covalent circular molecules would be affected by psoralen damage, as previously found for X phages (Ross t Howard-Flanders, 1977).

These

experiments

were

performed

in lysogens

carrying

the

non-inducible

prophage 186 turn. Exponentially growing E. coli (186 turn) lysogens were infected with 3H-labeled 186 phages, incubated in growth medium, lysed, and sedimented 11

C600

NH4807

NH4818

(c)

(d)

(e)

(186 turn)

(186 tunt)

turn)

3 3

type

WCA -

UWA -

Wild 3 3 3 3

10 10 10 10

Multiplicity of infection with 3H labeled 186 cIts phages

and psoralen

type

UWA -

Wild

Relevant phenotype

mutations

I

i+

Psoralen

-

0 2 0 2

kJ/m* 360 nm light,

30 30 30 30 30 30 + + f -t + +

-

:

-

-

2 0 2

0

0 1

from homoimmune

Psoralen

phage 186 DNA

Multiplicity of superinfection with unlabeled 186 cIts phages

damage on the yield of covalent circular tests for cutting in trans

1

25 2 24 17 26 13 24 22 27 25

lysogens and

These expcrimentn test the proportion of labrlcd 186 phage DNA in the form of covalent circular DNA molecules following t,he infection of E. co&i (186) host cells with W-labeled 186 phagw damaged by exposure to psoralen and 360 nm light. They also test for cutting in Irccns in labeled covalent circular 186 DS.1 following superinfcction with damaged unlabeled 186 phages in homoimmune lysogens. 32P-labeled 186 phages infecting wild type and uw.%- host (a), (b) Yields of covalent rrcular DXA molecules formed by control and psoralen-damaged cells. The results show cutting in the damaged phage DNA to occur in wild type but not in uvrAhost cells. (c), (d), (e) Tests for cutting in trn,ls in H3-labeled covalent. circular 186 DNB carried by wild typ?, wuriland recAlysogcnic host cells, resulting from superinfection with control and psoralen-damaged 186 phagrs. The results show cutting in ~rrt?ts to occur in the phage 186 system in wild type, but not in wvrilor recC hosts. 10 T rha mrtlium at 37°C to 3 x 10s to 5 x 10’ cclls/n~l. 2 < lo9 cells were withdrawn, cent.rifuged, resuspended in 1.0 ml Lysugens were grown in KJI 186 adsorption medium, and 3H-labelrd 186 c1t.s pheges were added. The mixture was held at 37°C for 30 min, at which time 10 ml KM A 10 - rha medium were added and the mixture aerated at 37°C for 20 min. In the experiments shown in lines (a) and (b), the cells were washed and lysed immediately. In the experiments of linen (c), (d) and (e), the ccl1 suspension was divided in two and crntrifuged. The pellet was resuspended in 0.5 ml 186 adsorption medium. and 2 x IO’O psoralen-treated phage were added, which had bren kept in the dark or exposed to 360 nm light. The mixtures were held at 37°C for 30 min, when 5 ml KM + 10 1 rha medium were added to each tube. Then, each tube was iterated at 37°C for 45 min and the cells were washed in IO mwTris.HCl (pH 8.1). Lysis and centrifugation through alkaline sucrose were as described in the preceding paper (Ross & Howard-Flanders, 1977).

(IS6

NH4807

(b)

(186 lum)

C600(1861um)

(a)

Host lysogen

Effect of host repair

TABLE

rec.4 +-DEPENDENT

CUTTING

IN

DNA

1FB

MOLECULES

through alkaline sucrose gradients. About 25% of the infecting label was recovered in a peak sedimenting three to four times more rapidly in alkali than the rest of the 3H label, in the position expected of duplex covalent circular phage 186 DNA (Table l(a)). Similar results have been obtained with the related phage P2 (Lindqvist, 1971). When 186 phages were exposed to psoralen and 2 kJ/m2 of 360 nm light prior to infection, a normal yield of rapidly sedimenting DNA was recovered from E. coli uvrA- (186), but no rapidly sedimenting radioactivity was obtained from wild-type lysogens (Table l(aj and: (b)). Evidently, psoralen-damaged DNA from phage 186. like that of phage X, is injected forming the normal yield of covalent circles. As with phage h, the intracellular 186 DNA containing psoralen damages is apparently cut by the uw~A+ endonuclease, as covalent circles were not obtained from wild-type hosts at this dose level. Sinre the psoralen-damaged phage 186 could inject its DNA normally, we were able to test whether superinfection with psoralen-damaged 186 phages would cause cutting of covalent circular phage 186 DNA (cutting in trans). Wild type, uvrAand rclcA- strains of E. coli lysogenic for prophage 186 turn were infected with 3Hlabeled 186 phages. When superinfected with undamaged 186 phages, the fraction of label sedimenting rapidly was about 250/b. When the unlabeled 186 phages were damaged with psoralen and light before infection, however, the fraction of radioactive label sedimenting rapidly remained about 25 “,6 in the uvrA- and recA lysogens, but was reduced to 13% in the wild-type lysogens (Table l(c), (d) and (e)). In other experiments (data not shown), it was found that precircularized phage 186 DNA was cut when the wild-type 186 lysogenic hosts were superinfected with psoralendamaged 186 phages. Evidently, cutting in tran,s of covalent circular phage 186 DNA occurred in response to superinfection with psoralen-damaged phage 186. Both l~vr + and recA+ were required, as previously found with phage A. (b) A test for the specificity of cutting in trans in cells containing 3H-labeled phage 186 and 32P-.labeled phage h DNA

both

Since the DNAs of phages X and 186 share no detectable base sequence homology (Skalka & Hanson, 1972), and since phages h and 186 appear to behave in a similar manner under the conditions of the cutting experiment just described, it was feasible to test between two possibilities. First, cutting in trans might be detected in any covalent circular DNA present in cells at the time of superinfection with damaged phage DNA, regardless of homology. In this case, cutting in tram might be due to changes in the level of cellular repair enzymes or of the efficiency of repair of spont’aneous nicks in response to the presence of damaged DNA. Second, cutting in trans might be specific to homologs of the damaged phages. so that damage in superinfecting X phages would cause cutting in trans of X covalent circles, but not of 186 covalent circles even though both were present in the same cell at the same time. Similarly, damaged 186 would cause cutting in tratu of 186, but not of h DNA. This second result would be expected if the damaged and undamaged strands paired before cutting in trans occurred. It was first necessary to determine in the covalent circular form would da,maged phage h or with undamaged

whether intracellular phage h and 186 DNAs be affected by superinfection either with unphage 186.

164

P. ROSS AND

P. HOWARD-FLANDERS

Exponentially growing E. coli (186 turn) (h ind) double lysogens were infected with 3H-labeled 186 and 32P-labeled /\ phages. The cells were divided into six samples that were supplemented with unlabeled 186 or /\ phages which had been treated with psoralen and either kept dark or irradiated with 360 nm light. Cells were aerated in growth medium, lysed and centrifuged through alkaline sucrose. The results of these experiments are shown in Figure 1. When cells were superinfected with ten h or with twenty 186 phages per cell, that had been exposed to psoralen and kept dark, the fraction of 32P radioactivity (continuous line) sedimenting rapidly was about 72% and that of 3H (broken line) was about 32 “;O (Fig. l(a) and (d)). The peak of 3H radioactivity did not sediment as far as the peak of 32P radioactivity because the mass of 186 DNA is 20x lo6 molecular weight, or about 2/3 of that of h DNA. The fraction of 186 DNA sedimenting rapidly is typically smaller than that found for X DNA. Evidently, superinfection with undamaged phage h or 186 had no discernible effect on the fraction of preinfecting phage DNA in covalent circular form (Table 1). The experiment of Figure 1 was also designed to test whether cutting in trans of phage h could be induced by superinfection with damaged phage 186, and whether cutting in trans of phage 186 could be induced by damaged phage X. When the psoralen-treated h phages were exposed to 360 nm light before infection, the fraction of 32P radioactivity in the rapidly sedimenting peak decreased from 72% to 35%, indicating that the superinfecting damaged h DNA caused the undamaged h DNA to be cut in trans (Fig. l(b) and (c)). The fraction of 3H radioactivity sedimenting rapidly, however, increased from 31% to 35% at the highest dose of light. Thus, the fraction of 186 DNA in the covalent circular form remained undiminished despite the presence of damaged X DNA capable of inducing cuts in X covalent circles. Figure l(e) and (f) shows the results of the reciprocal experiment’: when lysogens with containing both 3H-labeled 186 and 32P-labeled h phages were superinfected psoralen-damaged 186 phages, the 186 DNA was cut, while the h DNA remained intact. In subsequent experiments, a more complete dose dependence of cutting was obtained (Fig. 2). It took about 1.6 times as much light to produce cutting in tram of 50% of the 186 covalent circles as was needed to cause cutting of 500/A of the phage h circles. This may be a reflection of the ratio of the molecular weights of the two phages. An experiment was also performed to test the possibility of a trivial explanation for the specificity of cutting in tram observed in the experiments of Figures 1 and 2. The apparent specificity could have arisen if there were two subpopulations of the double lysogens, one able to adsorb h but not 186 phages, and the other able to adsorb 186 but not h phages. To test for this possibility, exponentially growing E. coli (h ind) cells (lysogens of strain CSOO) were infected with 32P-labeled phage h ~71 and with 3H-labeled phage 186 cIts, and washed twice. A sample was removed and counted for 3H and 32P radioactivity. The remaining cells were aerated at 37°C in growth medium, to permit lytic growth of the 186 phages. Following a drop in 0.D.650 “,,, from 0.9 to 0.2, approximately 89% of the 3H radioactivity and 9324 of the 32P radioactivity was released from the pellet into the supernatant. This result indicates that the cells were about equally susceptible to infection by 186 and h phages. We therefore conclude that the specificity seen in the experiments on cutting in trans was not due to specificities in phage attachment.

rec.4

+-DEPENDENT

CUTTING

IN

DNA

MOLECULES

165

I

cl

:b)

I.2

2.4

kJ/n?

kJ/m’

. : .

600

t

:;

0

Froci~on

IO

no.

FIG. 1. Profiles of radioactivity from E. coli (A) (186) cells infected with both 3H-labeled phage 186 and 32P-labeled phage h following sedimentation through alkaline sucrose gradients, from right, to left. This experiment tested whether cutting in trans was induced specifically in covalent, circles homologous to the particular phage used for superinfection. Superinfection with h phages damaged by treatment with psoralen and light caused a dose-dependent reduction in the proportion of 3ZP-labeled h DNA sedimenting rapidly, but no reduction in that of the 3Hlabeled 186 DNA. Conversely, superinfection with damaged 186 phages caused a reduction in t)he proportion of 3H-labeled 186 DNA sedimenting rapidly, but had no effect on the distribution of the 32P-labeled h DNA. (a) to (c) Superinfection with control and psoralen-damaged X phages. (d) to (f) Superinfection with control and psoralen-damaged 186 phages. -a---, 32P-labeled X DNA; -- A-A--, 3H-labeled 186 DNA. 6 x 109 cells of E. coli (A ind) (186 turn), a double lysogen of strain C600, growing exponentially in KM + 10 + rha medium, were harvested and resuspended at, 4 x 10y cells/ml in /\ buffer. 32P-labeled L phages were added (m.o.i. 3) and were allowed to adsorb during a 10.min incubation at 37°C. Most phages adsorbed within this time. Next, an equal volume of 186 adsorption medium containing 3H-labeled 186 phages (m.o.i. 3) was added, and these phages were allowed 30 min at 37°C to adsorb. The infected cells were diluted l/l0 in fresh KM + 10 rha medium and were aerated 30 min at 37°C to permit circularization of the labeled DNAs. The cells were next divided into 6 portions and harvested by centrifugation. Three portions were resuspended at 2 x 10’ cells/ml in h buffer, and were infected at a multiplicity of 10 h phages per cell that had been treated with psoralen and variously irradiated. The remaining 3 portions were resuspended at 2 x 108/ml in 186 adsorption buffer and infected with 186 phages (m.o.i. 20) that had been treated with psoralen and irradiated with 360 nm light for various times. The infection mixtures were held 30 min at 37°C for adsorption, diluted l/10 in KM + 10 + rha medium, and aerated 46 min at 37°C. Lysis and centrifugation in alkaline sucrose gradients were as previously described (Ross & Howard-Flanders, 1977). Gradients were collected as 16 fractions of 0.33 ml each into 0.5 ml of 0.15 M-HCl and counted in 5 ml Formula 963 scintillat,ion fluid (New England Nuclear). m.o.i., multiplicity of infection.

166

P. ROSS

AND

P. HOWARD-FLANDERS

I

I

I

I

I

4

2

0 kJ/m’

(a)

(bl

FIG. 2. Per cent s2P-labeled X DNA and 3H-labeled phage 186 DNA sedimenting rapidly from the experiment of Fig. 1, plotted against the time the superinfected phages were exposed to psoralen and 360 nm light. The data show that only those covalent circular molecules homologous to the superinfecting phages were cut. (a) Superinfection with psoralen-treated 186 phages. (b) Superinfection with psoralen-treated h phages. The damaged phages in this experiment were samples of those used in the experiment of Fig. 4. (0, 0) 32P-labeled h DNA; (a, A) 3H-labeled 186 DNA.

(c) Specificity in the induction of recombination by psoralen-damaged 186 or A phages in homoimmune phqe-prophqe crosses The ability of psoralen damage in phage h DNA to increase the frequency of phage-prophage recombination in homoimmune crosses (Lin et al., 1977) might result from either of two mechanisms. First, damaged DNA undergoing repair might pair with intact homologous DNA. A direct interaction of this sort would enhance recombination only among phages and prophages sharing genetic homology with the damaged DNA. Second, infection with psoralen-damaged phages might cause a transient nicking of all DNAs in the cell, or by some other mechanism generally increase the frequency of recombination. In this case, all phage-prophage systems in the cell would manifest elevated recombination frequencies. An experiment was designed to test whether superinfection with damaged 186 phages would have any effect on the frequency of recombination in A phage-prophage crosses in the same host cells. Exponentially growing cells of the double lysogen E. coli supE44 (A ~1857 P80) (186 turn) were infected with undamaged or with psoralen-damaged h ~1857 P3 phages, and samples of the cells infected with undamaged phages were superinfected with 186 phages containing various amounts of

recA+

-DEPENDENT

CUTTING

IN

DNA

MOLECULES

167

psoralen damage. The multiply infected lysogens were grown through about ten generations at 30°C. Growth for an additional five generations had no effect on the fraction of h prophages that were P+ recombinants. Each culture was then assayed for the fraction of cells carrying h prophages that were P+ without risk of artifacts due to residual DNA from the infecting phages. The ratio of the number of plaques formed on permissive and non-permissive host cells lysogenic for 186 gave the and is plotted against the dose fraction of h prophages that were P+ recombinants, of light in Figure 3(a). Clearly, damage to the infecting X phages increased the fraction of PC recombinants among the X prophages, while equivalent damage to superinfecting 186 phages had no effect on this fraction. In the reciprocal experiment, it was asked whether superinfection with damaged phage X DNA could stimulate 186 phage-prophage recombination. The results of this experiment are plotted in Figure 3(b). Damage to the infecting phage 186 increased the fraction of wild-type recombinants by up to 50-fold, while damage t,o the superinfecting phage X had no significant effect on the fraction of wild-type 186 prophages. Evidently, damage stimulated recombination only between damaged DNAs and their homologs. ((1) Prophage reactivation

of psoralen-damaged homology

h and 186 phages: a test for

Prophage reactivation is said to occur if ultraviolet-irradiated or alkylated temperate bacteriophages form more plaques on lysogenic than on non-lysogenic host cells. Prophage reactivation requires recA + or red + . and is thought to depend upon genetic exchanges between phage and prophage (Hart t Ellison, 1970; Blanc0 & Devoret, 1973). This phenomenon has been observed only when the phage and prophage are at least partially homologous (Jacob & Wollman, 1953 ; Hart & Ellison, 1970). We have used prophage reactivation as a crude test for genetic homology between phages h and 186. Phages A and 186 were treated with psoralen and exposed to various doses of 360 nm light. The treated phages were then plated on lysogenic and non-lysogenic host cells. ,4s seen in Figure 4(a), the plaque-forming ability of irradiated phage 186 was similar on E. coli and on E. coli (X ind-), and it was marginally greater on a phage P2 lysogen. Similarly in Figure 4(b), it is seen that the plaque-forming ability of irradiated phage X was unaffected by prophage 186, but substantially increased by prophage h imm434. In similar experiments, survival of phage 186 vir was found to be 280/, on a 186 lysogen, at a dose reducing survival to lO;h on non-lysogens or on X lyxogens. Evidently, prophage reactivation occurred in host cells carrying a prophage known to share homology with the damaged phage. A 186 prophage did not reactivate psoralen-damaged phage h, nor did prophage h reactivate psoralendamaged phage 186. By these tests, therefore, phages h and 186 share no detectable genetic homology. It should be noted, however, that the sensitivity of prophage reactivation as a test to evaluate genetic homology is uncertain. For instance, reactivation of X vir by the defective prophage h dgal is substantially diminished relative to reactivation by a viable 434 prophage sharing even less homology with the damaged, infecting phage (Hart & Ellison, 1970 ; Simon et al., 1971). This does not, however, preclude the use of prophage reactivation for the present test, which, although crude, resembles and may be relevant to induced phage-prophage recombination.

168

P. ROSS

AND

P. HOWARD-FLANDEES

/

i

(Phage

X

Phoge 186

!=

2

hJ/d (0)

ib)

FIa. 3. The frequency of X P+ or of 186 4+12+ recombinant prophages from homoimmune crosses is shown as a function of the exposure of infecting phages to psoralen and 360 nm light. The results show that while psoralen damage in phages can increase the frequency of either h recombination by more than SO-fold, damaged 186 phages do not stimulate or 186 phage-prophage h phage-prophage recombination, and damaged h phages do not stimulate 186 phage-prophage recombination. (a) E. coli (X cl857 P80) (186 lunz), a double lysogen of strain C600, was grown aerobically in L broth at 32°C to 5 x 108/ml. 10s cells were transferred to each of 4 tubes, and 5 x lo8 cells to a fifth tube. To each of the first 4 tubes were added psoralen-treated, variously irradiated phage X 5). The fifth tube received untreated phage A ~I857 P3 (m.o.i. 5). The tubes ~1857 P3 (m.o.i. were held at 32°C for 20 min, when the volume of each tube was doubled by addition of 186 adsorption buffer. The contents of the fifth tube were divided among 4 other tubes to each of which were added phage 186 cIts (m.o.i. 10) which had been treated with psoralen and various amounts of 360 nm light. All cells were held at 32°C for 60 min. The efficiency of adsorption was tested in the tubes containing unirradiated phages. Adsorption of the 186 phages was 83%. Portions (0.1 ml) from each of the 8 tubes were diluted l/20 in L broth + 1 y0 NaCl and incubated overnight at 32°C. The cultures were adjusted to lOa cells/ml and plated on E. coli supE44 (186) Plates were incubated at 42% for 2 h to induce and on E. coli sup+ (186) m d‘ lea t or bacteria. the thermosensitive lysogens and incubated at 37°C overnight. The per cent X P+ recombinant prophages was determined from the ratio of the number of plaques on E. coli sup + (186) to that on E. coli sup E44 (186) host cells. (b) E. coli supE44 (X ind) (186 cIts 12.19) were grown in L broth + 1% NaCl with aeration at 28°C to 2 x 10s cells/ml. Portions containing lOa cells were removed to each of 5 tubes, while a sixth tube received lo9 lysogens. To each of the first 5 tubes were added phage 186 cIts 4-7 (m.o.i. 16), which had been treated with psoralen and variously irradiated with 360 nm light, while the sixth tube received untreated phage 186 cIts 4.7 (m.o.i. 15). Following incubation at 28°C for 30 min, t,he cells in the sixth tube were centrifuged and the supernatant was titered for unadsorbed phages. The pellet was resuspended in 1.0 ml X buffer, and 0.1 ml containing 10’ cells was transferred to each of 5 tubes. To these tubes were added A ~71 phages (m.o.i. 6) which had been exposed to psoralen and various doses of 360 nm light. Following 30 min further incubation, a sample from each of the 10 tubes was diluted l/l00 in L broth and incubated at 28°C overnight. Adsorption of the phage 186 was 75%, and adsorption of t)he I phages was 97%. The next day, the lysogens were diluted to 5x 10B/ml, plated on E. coli sup’ (h ind) and on E. cobi supE44 (A ind), induced by incubation at 42°C for 2 h and incubated overnight at 37°C. The was determined from the fraction fraction of lysogens carrying recombinant, 4 + 12 + 186 prophages forming plaques on the sup+ host cells as above.

rec.4

+-DEPENDENT

CUTTING

kJ/m2

IN

350nm

DNA

MOLECULES

169

hqhl (b)

(0)

pla. 4. Survival of the plaque-forming ability of psoralen-treated 186 and X phrtges as a function of 6xposure to 360 nm light. Survival of damaged phage 186 was not enhanced by the presence of a X prophage, nor was survival of damaged phage h enhanced in cells carrying prophage 186. These results are consistent, with the view that phage h and 186 share no genetic homology. Phages h c71 and 186 dt8 were treated with psoralen end either kept dark, or exposed to 360 nm light. The treated phages were then plated on lysogenic and non-lysogenic host cells and incubated overnight at 37°C.

As might be expected, the rates of inactivation on non-lysogenic hosts, are in the ratio of about the molecular weights of their DNAs.

of phages A and 186, when plated 1.8, which is close to the ratio of

4. Discussion The DNA from 186 and h phages was acted on in much the same way following the infection of homoimmune lysogenic host cells. About 30% of the DNA from untreated 186 phages infecting E. coli (186) was converted to the covalent circular form, as judged by the proportion of 186 DNA sedimenting rapidly in alkali (Fig. 1). The presence of psoralen damage in the 186 phage DNA did not interfere with injection and circularization, but the psoralen-damaged molecules were apparently cut in

170

P. ROSS

AND

P. HOWARD-FLANDERS

excision-proficient cells (Table l(a) and (b)) by the uvrA-controlled endonuclease activity (Braun & Grossman, 1974). These results are similar to those previously obtained with h phage in E. coli (h) (Ross & Howard-Flanders, 1977). When E. coli (186) cells carrying covalent circular 186 DNA were superinfected with psoralen-damaged 186 phages, the covalent circular DNA was cut (Table l(c)). Thus, cutting in tram was detected in phage 186 as well as phage h systems, and in both cases was dependent on the uvrA+ and recA+ functions in the host lysogen (Table l(d) and (e)). As stated in the Introduction, phages X and 186 differ in size, are structurally distinct’, and show no detectable base sequence homology nor ability to produce hybrid recombinants. The lack of detectable prophage reactivation between these phages provides further evidence for the absence of genetic homology (Fig. 4) and confirms the suitability of t,hese phages for studies on the interaction between non-homologous phages. Evidence that cutting in tram might result from an interaction involving specific pairing between damaged DNA with covalent circular homologs was obtained in experiments with the double lysogen E. coli (X) (186). If these cells were infected with h and then with 186 phages, each phage DNA was converted to the covalent circular form in normel yield. This multiply infected double lysogen was used to test whether damaged phage DNA causing cutting in tram in the DNA of one phage would also cause cutt)ing in the covalent circular DNA of the other phage type present in the same cells. When double lysogenic cells carrying both types of covalent circular DNA were superinfected with psoralen-damaged h phages, cutting was detected in the h but not the 186 circles (Figs 1 and 2). Similarly, superinfection with damaged 186 phages caused the 186 DNA circles to be cut, but did not affect those of h DNA. These results support an interpretat’ion in which base-pairing occurred between the damaged DNA molecules undergoing repair and covalent circular molecules before cutting in tram took place. They are not consistent with an alternative interpretation in which the psoralen-damaged phage DNA activated a nuclease that caused non-specific cutting in tram in all DNA molecules in the cell. A second test of specificity in psoralen damage-induced recombination was made by genetic methods. In this case, it was asked whether the increased frequency of recombination in phage-prophage crosses in homoimmune lysogenic host cells was specific to the damaged molecule and its homolog, or was of a more generalized nature. One possibility was that infection with the psoralen-damaged phage might cause changes in the levels of nucleases and other enzymes which would increase the frequency of genetic recombination between any pairs of homologs present in the same host cell. Again taking advantage of the lack of genetic homology between 186 and h phages, E. coEi (186) (h) double lysogens carrying genetically marked 186 prophages were infected with control and psoralen-damaged 186 phages. The presence of psoralen damage in the infecting 186 DNA caused a dose-dependent increase in the frequency of 186 phage-prophage recombination (Fig. 3(b)). This result extends previous findings using phage A to 186 phage-prophage recombination, and since phage 186 appears to lack a general recombination system (Mandel & Kornreich, 1972), supports the view that damage-stimulated recombination in homoimmune crosses was mediated by host functions. Damage-stimulated phage-prophage recombination in homoimmune lysogens appeared to be specific for the damaged phages and their homologs, because there was no effect upon the frequency of 186 phage-prophage recombination when double

recrl

+-DEPENDENT

CUTTING

IN

DNA

MOLECULES

171

lysogens infected by undamaged 186 phages were superinfected with psoralendamaged X phages (Fig. 3(b)). Similarly, superinfection with damaged 186 phagex had no effect on the frequency of h phage-prophage recombination (Fig. 3(a)). In view of these results, it is unlikely that the psoralen-damaged phage DNA increased the phage-prophage recombination through the induction or activation of nucleases or other enzymes that initiate recombination at many points in the bacterial and phage DNA. The results of these tests throw some light on the sequence of reactions, particularly the sequence of cutting and pairing by which genetic recombination can be initiated. A cut-cut,-pair mechanism, for example, is one in which cuts are made in both duplexes before homologous pairing and the paired molecules are then covalently joined to form recombinants. Genetic recombination might be initiated by a pair-cut-cut, by a cut-pair-cut, or by a cut-cut-pair mechanism, as illustrated in Figure 5.

Dlsplocement

-Cut

Cut A ---

---

POli

Po1r I

1

z&c

IX

lel

i b )Whttehouse ( c 1 Holhday

‘6

\

Folr-Cdt-Cut

v

/

\ cut-Pair-Cut

cut-cut-Po1r PalrIng between homologs:

(f 1

m

FIG. 5. Possible mechanisms for the initiation of genetic recombination between 2 homologous DNA duplexes. (a) Pairing takes place before either duplex is cut. (b) Both duplexes are cut at nearby sites before pairing, as proposed by Whitehouse (1963). (c) (:uts made at similar positions in both duplexes before pairing, as proposed by Holliday (1964). (d) One strand is cut and displacement synthesis leads to pairing before the second duplex is cut, as proposed by Meselson & Raddiug (1975). (e) and (f) Mechanisms for the initiation of recombination following the incision of 1 duplex containing a crosslink. This is followed by pairing and cutting in the second duplex in response to pairing. Alternative forms of pairing are shown.

172

P. ROSS

AND

P. HOWARD-FLANDERS

Mechanisms for genetic recombination involving pairing between the homologs before strand cutting (pair-cut-cut) have not been considered plausible because pairing between uncut duplexes would not be energetically favorable. However, the discovery of omega protein (Wang, 1969) and of DNA gyrase (Gellert et al., 1976) may eliminate some of the objections to a pair-cut-cut mechanism. Cut-pair-cut mechanisms have been proposed for the initiation of genetic recombination in bacterial transformation (Fox, 1966), in recombinational repair (Rupp recombination & Howard-Flanders, 1968; Rupp et al., 1971), in ultraviolet-induced in +X phages (Benbow et al., 1975), in phage T4 (Mosig et aE., 1971), and other systems. The uptake of single-strand fragments of phage +X DNA that bind to covalent circular duplex 4X DNA from infected cells (Holloman et al., 1975) illustrates how base-pairing can occur between a single strand and an intact supercoiled duplex homolog. On the other hand, cut-cut-pair models for the initiation of recombination have been proposed in which both duplexes are cut before homologous pairing in fungi and phage T4 systems (Whitehouse, 1963; Holliday, 1964: Stahl, 1969; Broker, 1973; Watson, 1976). However, evidence for both duplexes having to be cut before complementary base-pairing occurs is st,ill lacking. and this feature of the latter five models remains unsupported. The finding that damage-induced phage-prophage recombination (Lin et al., 1977), and also cutting in tram in homoimmune lysogenic host cells, are both dependent upon uvrA+ and by inference the uvrA B endonuclease demonstrates the need for strands to be cut for the initiation of recA+-dependent recombination. Since the induced recombination as well as cutting in tram was detected only in the phage DNA molecules sharing genetic homology with the damaged phage DNA, it may be inferred that recombination in this system is initiated by the direct pairing of the incised DNA molecules with their homologs. In so far as cutting in tram may be a, step in damage-induced recombination, the failure to detect cutting in tram between the non-homologous phages h and 186 can be interpreted as supporting a cut-pair-cut mechanism. It is unlikely that a cut-cut-pair mechanism is obligatory for cutting in tram and for genetic recombination in E. coZi. The present results show there to be a mechanism in bacteria for initiating recombination, in which DNA containing nicks or gaps may pair with homologs, and in which pairing must take place before cutting can be detected in the intact duplex, possibly as illustrated in Figure 5(e) or (f). Cuts by the uvrA endonuclease in the psoralen-damaged duplex lead to cutting in homologous covalent circular molecules, while non-homologous molecules within the same host cell remain intact. Similarly, damaged phages initiate recombination with their homolog. but have no effect on other pairs of homologs present in the same host. It is not known how the incised DNA molecules are able to initiate pairing with undamaged homologs. In the Meselson & Radding (1975) model for meiotic recombination in fungi (Fig. 5(d)), it is proposed that recombination is initiated through a single-strand cut followed by displacement synthesis by a DNA polymerase adding to a free end, The strand so displaced then pairs with the homolog. However, the application of this model to the present results would be complicated by the presence in the template strand of the partially excised crosslink. In an alternative mechanism, displacement of the single strand might be accomplished by the DNA unwinding would then enzyme of E. coZi (Abdel-Monen et al., 1977), or in other ways. Initiation occur by the sequence shown in Figure 5(e). Yet another possibility is that exonuclease

racA

+-DEPENDENT

CUTTING

IN

l)NA

MOLECULES

173

degradation following the incision of the crosslink might permit the single strand in the gap so formed to pair with an intact homologous duplex, as in Figure 5(f). To conclude, these experiments show there to be a mechanism for general recombination in E. coli (h) which is dependent upon host functions including recA+, and which is presumably initiated through a cut-pair-cut sequence of events. This process was initiated when psoralen-damaged h or 186 phage DNA was injected into host cells lysogenic for both h and 186 phages, and the infecting DNA was cut by the uvrA-controlled endonuclease. The incised phage DNA caused homologous covalent circular phage DNA to be cut, while non-homologous DNA was spared. Infection wit,h psoralen-damaged h or 186 phages also induced phage-prophage recombination in homoimmune lysogens. This recombination was specific to the damaged phage DNA and its homologous prophage, none being induced in ot’her phage-prophage homolog pairs present in the same cell. These results make it unlikely that recA+dependent recombination induced by infection with psoralen-damaged temperate phages could be due to an indirect mechanism acting on all homolog pairs present in the host cell. More likely, following incision, the damaged phage DN-4 molecules pair directly wit’h their homologs and cause them to he cut in a recA+-dependent reaction by a recombination endonuclease which cuts the undamaged member of the paired

homologs.

Tliis bvork was supported by the United GM 11014, AMK 69397 and GM 00711.

States

Public

Healt,h

Service

grants

CA 06519,

REFERENCES Abdel-Monen, M., Lauppe, H.-F., Kartenbeck, J., Durwald, H. & Hoffmann-Berling, H. (1977). J. MoZ. Biol. 110, 667-685. Baldwin, R. L., Barrand, P., Fritsch, A., Goldthwait. D. A. & Jacob, F. (1966). J. Mol. Biol. 17, 343-357. Benbow-, R. M., Zuccarelli, A. J. & Sinsheimer. R. L. (1975). Proc. Nat. Acad. Sci., U.S.A. 72, 235-239. Bertani. L. & Bertani, G. (1971). Advan. Genet. 16, 199~.237. Blanco, M. & Devoret, R. (1973). Mutat. Res. 17, 293-305. Bradley, C., Ling, 0. I’. & Egan, J. B. (1975). MoZ. Cert. Genet. 140, 123-135. Braun, A. & Grossman, L. (1974). Proc. Nut. Acad. Scl:., li.S.il. 71. 18381842. Broker, T. R. (1973). J. MOE. Biol. 81, 1-16. Fox, M. S. (1966). J. Gen. Physiol. 49, 1833196. Gellert, M., Mizuuchi, K., O’Dean, M. H. & Nash, H. A. (1976). Proc. Nat. Acad. Sci., U.S.A. 73, 3872-3876. Hart, M. G. R. & Ellison, J. (1970). J. Gen. Viral. 8, 197-208. Holliday, R. (1964). Genet. Res. Camb. 5, 282-304. Holloman, W. K., Wiegand, R., Hoessli, C. $ Radding, C. M. (1975). Proc. Nat. Acad. Sci., U.S.A. 72, 2394-2398. Jacob, F. & Wollman, E. (1953). Cold Spring Harbor Symp. Quant. BioZ. 18, 101-121. Jacob, F. $ Wollman, E. (1961). Sezuality and the Genetics of Bacteria, Academic Press, New York. Lin, P.-F., Bardwell, E. & Howard-Flanders, P. (1977). Proc. Nat. Acad. Sci., U.S.A. 74, 291-304. Lindyvist, B. (1971). Mol. Gen. Genet. 110, 551-556. Mandel, M. & Kornreich, B. (1972). Virology, 49, 300-301. Meselson, M. S. & Radding, C. M. (1975). Proc. Nat. Acad. Sci., U.S.A. 72, 3588361. Mosig, G., Ehring, R., Schliewen, W. & Bock, S. (1971). Mol. Gen. Genet. 113, 51-91. Ross, P. & Howard-Flanders, P. (1977). J. Mol. BioZ. 117, 137-158. Rupp, W. D. & Howard-Flanders, P. (1968). J. Mol. BioZ. 31. 291l304.

174 Rupp,

P. ROSS

AND

P. HOWARD-FLANDERS

W. D., Wilde, C. E., Reno, D. 8: Howard-Flanders, P. (1971). J. &lol. &oZ. 61, 25-44. Simon, M., Davis, R. W. & Davidson, N. (1971). In Z’he Racterioph,age Lambda, pp. 313 328, Cold Spring Harbor Laboratory Press, New York. Skalka, A. 85 Hanson, P. (1972). J. Viral. 9, 583-593. Stahl, I?. W. (1969). Genetics (Suppl. I), 61, I-13. Wang, J. C. (1967). J. Mol. Biol. 28, 403-411. Wang, J. C. (1969). J. Mol. Biol. 43, 263--272. Watson, J. D. (1976). In The Molecular Biology of the Gene, 3rd edit., pp. 264-267, Benjamin, Menlo Park, California. Whitehouse, H. L. K. (1963). Nature (London), 199, 1034~-1040. Woods, W. K. & Egan, J. B. (1974). J. rirol. 14, 1349-1356. Younghusband, H. B. & Inmall, R. H. (1974). Virology, 62, 530--538.