Mol Gen Genet (1992) 236:60-64 © Springer-Verlag 1992

Intramolecular homologous recombination in Bacillus subtilis 168 Juan C. Alonso, Gerhild Liider, and Thomas A. Trautner Max-Planck-Institut fiir molekulare Genetik, Ihnestrasse 73, W-1000 Berlin 33, FRG Received July 2, 1992 / Accepted July 10, 1992

Summary. Plasmid resolution from a phage: :plasmid chimera was used to measure directly intramolecular recombination in Bacillus subtilis. The system is based on a sigma-replicating plasmid (pC194) cloned into a dispensable region of the lytic bacteriophage SPP1. The plasmid, which confers chloramphenicol resistance, is resolved when SPPI:: pC194 phages infect B. subtilis cells, provided the chimera carries a functional, intact copy of the plasmid repH gene. Intramolecular homologous recombination was independent of the RecA and RecL-RecR functions, but dependent on RecF, RecB, RecG, RecP, RecH and AddAB functions. These results are consistent with the hypothesis that B. subtilis has multiple pathways for genetic recombination and allow us to tentatively place the recB and recG genes into a new epistatic group e. Key words: rec mutants - Plasmid pC194 - Bacteriophage SPP1 - RepH protein

Introduction At present 11 different genes (recA, recB, recD, recF, recG, recH, recL, recR, recP, addA and addB) are known to be associated with D N A repair and recombination in Bacillus subtilis (Mazza and Gallizi 1989). The RecA protein controls and/or contributes to all postreplicational repair pathways (Ceglowski et al. 1990). Recently, the B. subtilis rec genes, other than recA, have been classified into three groups (~, 13and y) according to the recombinational pathway in which they operate (Alonso et al. 1991). Group a includes the recB, recD, recG, recF, recL and recR genes. The ability of indirect suppressors of recF mutations to suppress lesions in recL and recR supports this grouping (Alonso and Liider 1991). Some reservations, however, still pertain to the inclusion of the recB, recD and recG in group a. Group [3 comprises the addA and addB genes. The addA and addB genes encode Correspondence to: J. Alonso

different subunits of the multifunctional enzyme AddABC or ExoV (Shemyakin et al. 1979; Kooistra and Venema 1991). The AddABC protein is equivalent to RecBCD in Escherichia coli (Telander-Muskavitch and Linn 1981). Group Y includes the recP and recH genes (Alonso et al. 1991). It has been shown that sigma-type replication of a plasmid integrated into the B. subtilis chromosome markedly stimulates homologous intramolecular recombination (20- to 450-fold) between D N A sequences close to the integration site (Noiret et al. 1987). Furthermore, it has been documented that activation of sigma-type plasmid replication increased 3- to 7-fold the number of copies (DNA amplification) of sequences in the vicinity of the integrated plasmid. However, when the plasmid is integrated between duplicated regions only a long array of directly repeated amplified units was observed upon activation (Petit et al. 1992). In both conditions the replicated material has to be subsequently resolved via intramolecular recombination to give autonomously replicating plasmids. With an experimental design conceptually related to that of Noiret et al. (1987) we have extended our analysis of B. subtilis recombination functions. We have integrated a plasmid, which replicates in the sigma mode (pC194), into a dispensable region of a lytic bacteriophage (SPP1) genome, to generate SPPI:: pC194 chimeras. When a chimeric phage infects B. subtilis cells it can go through a successful infective cycle (generating phage progeny) or an abortive infective cycle (due to plasmid resolution). In the latter case, replication of the integrated plasmid replicon should lead to amplification of all DNA sequences in the vicinity of the integrated site (Noiret et al. 1987; Petit et al. 1992). Subsequently, the amplified material is resolved via homologous intramolecular recombination within the amplified region.

Materials and methods Bacterial strains, phages and plasmids. All B. subtilis strains used are listed in Table 1. All are isogenic with

61

Results and discussion

Table 1. B. subtilis strains Single mutant genotype

Strain

rec + recA4 ~ recB2 reeD41 recF15 recG40 reeH342 recL16 recR13 b recP149 addA5 addB72

(YB886) (YB1015) (YB 1290) (BG121) (BGI29) (BG123) (BG 119) (BG107) (BG127) (BG101) (BG125) (BG 126)

Double mutant genotype recF15

addA5

NA BG145 BG 151 BG 139 NA BG141 BG 137 BG159 BG161 BG185 BGI43 ND

NA BG163 BG 167 BG 171 BG143 BG175 BG 177 BG179 BG161 BG185 NA BG 189

The isogenic background of all strains is trpC2 metB5 amyE sigB xin-1 attSP~. Strain BG189 by construction is resistant to phleomycin. The strains have been previously described (Kupsch et al., 1989 and references therein). a This allele was known previously as recE4 b This allele was formerly termed recM13. NA, not applicable; ND, not done

strain YB886. The bacteria were grown and maintained in TY medium (Biswal et al. 1967). All strains were identical with respect to their response to SPP1 infection. The various SPPl::pC194 chimeras (SPPlv20, SPPlv32 and SPPlv40) have been reported previously (Deichelbohrer et al. 1985; Alonso et al. 1986). All contain, (at the unique BamHI restriction site of phage SPPlv (Heilmann and Reeve 1982), a copy of pC194, linearized with MboI. The chimeras are depicted in Fig. 1. The plasmids used were pC194 (Horinouchi and Weisblum 1982) and pBT68 (Alonso and Trautner 1985). Intramolecular recombinational assay. Phage stocks were prepared as previously described (Deichelbohrer et al. 1985). Plasmid resolution was accomplished by infecting about 2.0 x 10s exponentially growing cells at a multiplicity of infection (moi) of 3 with the chimeric phages. Five minutes later the unadsorbed phages were inactivated with SPP1 antiserum essentially as previously described (Canosi et al. 1982) and the infected cells were plated on chloramphenicol-containing plates. Transfection and transformation followed the protocol of Rott1/inder and Trautner (1970). Antibiotic-resistant resolvants or transformants were selected on TY plates containing 5 gg/ml of either chloramphenicol (Cm) or tetracycline (Tc). Other biochemical techniques. Plasmid and phage DNA were prepared on preparative and analytical scales as previously described (Deichelbohrer et al. 1985). Restriction endonucleases and DNA modification enzymes were purchased from Boehringer (Mannheim, Germany). All enzymes were used as recommended by the supplier.

Structure of the SPP1 ." ."pC194 chimeric phages Plasmid pC194 is a high copy number plasmid that replicates in the sigma mode (Gruss and Ehrlich 1989). pC194 DNA, which has two MboI sites (coordinates 1 and 2001, Horinouchi and Weisblum 1982), was linearized by partial MboI digestion and ligated into the unique BamHI site of SPPlv (Heilmann and Reeve 1982). Competent cells were transfected with the ligation mixture. Plaques generated by phages containing pC194 DNA were identified and phages from three such plaques were purified and further analyzed. Their insert structure was determined by electron microscopic heteroduplex analysis and restriction enzyme digestion. As shown in Fig. 1, in all phages pC194 DNA is inserted in the same orientation with respect to the SPP1 genome. The chimeric phages SPPlv20 and SPPlv32 contain the entire pC194 plasmid in both possible permutations. Phage SPPlv40 was considered to be an in vivo deletion derivative of SPPlv32, in which about 0.6 kb of plasmid DNA between the ClaI site at nucleotide 2278 and the MboI site at nucleotide 2907 has been lost [nucleotide coordinates are given according to Horinouchi and Weisblum (1982), incorporating the latest amendments (Ballester et al. 1989)]. All chimeras contain the chloramphenicol acetyltransferase gene (cat). The repH gene, which codes for the positive trans-acting product required for initiation of plasmid replication (Alonso and Tailor 1987) is intact in SPPlv20, but disrupted in SPPlv32. In SPPlv40 the 3' terminal region of the repH gene is deleted. Recombination in a wild-type genetic background When plasmid-free cells were infected with SPPlv20, Cm ~ cells were observed among the survivors of the infection (Table 2). The Cm r phenotype was correlated with the appearance, in such cells, of a novel plasmid (p1948) with a molecular weight higher than that of pC194. More than 99% of the plasmids originating from Table 2. Resolution frequency of various chimeric phage stocks in the presence and absence of a repH gene product Phage

Presence of pBT68 in recipient cells

Frequency of Cm r cells among survivors

SPPlv20

+

1.1 x 10 .2 4.6 x 10 -~

SPPlv32

+

< 1.0 x 10 -sa 1.9 x 10 -3

SPPIv40

+

< 1.0 x 10 -s~ 3.3 x 10 -4

Equal aliquots of YB886 cells at a concentration of about 1.0 x 108 per ml and phage at a concentration of 1.0 x l0 s plaque-forming units/ml were mixed, pBT68 is a pC194-derived plasmid which confers resistance to tetracycline. Condition under which surviving cells were concentrated 10-fold

62

',,, 2 C A C-G T-A A-T G-C G-C C-G C-G i

i

i

i

I

i

i

i

i

i

i

i

i

i

Fig. 1. Physical structure of pC194 and of

G-c G,-,c c-G c-G

ectt

T-A A-T G-C i

i

i

i

5' B SPPlv 20--1

Hi Be ~ I I

B

SPPlv32__l

P M i

Hi

M

C

B

,~ I

I

I

I

Hi M

C

B

I

I

I

B

SPPlvZ, O__ I

CHi I I

~1

3'

C

B

M

I

I

I M I

CHi II

PBa I I

M I

C Io° A ° .

infecting SPP1 : : pC194 chimeric phages contain pl948 (data not shown). Plasmid p1948 has been fully sequenced and contains a 2.7 kb DNA fragment derived from both ends of the SPP1 genome (see Chai et al. 1992). Hence, upon infection, the SPP1 genome must be circularized. Furthermore, the junctions between the pC194 and the SPP1 DNA involve a 9 bp homology region. Hence, in view of the absence of apparent duplicated regions we proposed that resolution events possibly involving these sequences occur prior to amplification of the DNA sequences in the vicinity of the integrated site (see Petit et al. 1992). Since SPP1 is a lytic phage, we must assume that initiation of plasmid and phage replication are incompatible. Hence, p1948 had become resolved following abortive infection of B. subtilis with SPPlv20, which yielded no progeny phage. No Cm' cells [and therefore no resolution of the chimeras to a stable plasmid] were observed following infection of plasmid-free cells with SPP lv32 or SPP l v40 form, occurred (Table 2). However, when the pC194 replication initiation-termination protein (RepH protein) is provided in trans by a resident plasmid (pBT68), Cm ' colonies were observed among the survivors of infection with SPPlv32 or SPPlv40 (see Table 2). Given that plasmids that share replication functions are incompatible, the true resolution frequency might be higher. This follows from the observation that plasmid resolution from SPP1 v20 occurs in the presence of pBT68 only with a frequency of about 2% of the value observed after infection of plasmid-free cells (Table 2).

B

B B

the insert region of pC194 in chimeric SPP1 : : pC194 phages. Heavy lines represent pC194 DNA and thin lines SPPlv DNA. On the circular map of pC194 the arrows correspond to open reading frames. On the upper right panel the direct repeat (thin arrow) within the inverted repeat DNA sequence flanking the pC194 insertion at the unique BarnHI site of SPPlv, and the target sites for HaeIII, HpaII and XrnaIII restriction sites are shown. Inserts start at coordinates 2001 (M) (SPPlv20) or 1 (M*) (SPPIv32, SPPlv40) (see text). Abbreviations in upper panel: cat, Cmr gene; repH, initiation replication protein; ori, leading strand replication origin; lower panel: B, B91I; Ba, BarnHI; C, ClaI; Hi, HindIIl; M, MboI; P, PvulI; A, deletion

From these data we conclude that the product of the repH gene is needed to obtain cells with resolved plas-

mids. At present, we cannot distinguish whether expression of the repH gene is required only to guarantee the autonomous persistence of a plasmid once it has been resolved or whether it is also involved in the amplification of the DNA substrate prior to resolution (see Petit et al. 1992). Since resolution frequencies were independent of the moi (in the range 0.1 phage to 3 phages per cell) and bearing in mind that SPP1 is a lytic phage, we conclude that the plasmid has to be resolved before initiation of the lytic cycle and that resolution is an intramolecular rather than intermolecular recombinational event. Although plasmid resolution resembles that observed with phages P22 (Orbach and Jackson 1982) and T4 (Mattson et al. 1983), the process observed there involves in vivo insertion and resolution via a Campbell-type mechanism involving long streches of homology shared between the plasmid and phage. Here, however, such long direct repeat sequences are absent. In fact, our sequencing data, obtained with different p1948 resolvants, indicate that a 9 bp repeat within a dispensable region of pC194 and in E c o R I f r a g m e n t 9 of SPP1 was used to excise p1948 from SPPlv20 (Chai et al. 1992). An open question in the generation of p1948 concerns the nature of the selective pressure leading to the almost exclusive use of a 9 bp short direct repeat sequence in resolution.

63 Table 3. Frequency of plasmid resolution from the SPPlv20 chimera Epistatic group [3 ?+ 13 ?+ [3 ?+ 13 a+[3 a+13 T+ 13 7+ 13

Relevant genotype

Frequency of resolvants Epistatic among surviving cells group

Relevant genotype

rec+ recA4 addA5 addA5 addB72 addA5 recB2 addA5 recD41 addA5 recG40 addA5 reeL16 addA5 recR13 addA5 recH342 addA5 recP149 addA5

9.8 × 10.3 4.2 x 10-a 5.1 × 10.3 1.5 x 10.3 4.2 × 10 3 2.0 x t0 -3 4.3 × 10-3 4.6× 10-3 8.0 × 10-3 8.0 x 10-4 7.6 x 10-4

recA4 recA4 recFl5 recF15 addA5 reeF15 recB2 recF15 recD41 recF15 recG40 recF15 recL16 recF15 recR13 recF15 recH342 recF15 recP159 recF15

ct ~+~ ?+ c~ ?+ ct 9.+ (~. fl~-o, ~-[-~ 7+c~ 7+a

Frequency of resolvants among surviving cells 3.3 × l0 -3 2.5 × 10-3 4,8 x 10.3 4.8 × 10-4 < 1.0 x 10-6 1.8 × 10 3 8.8 x 10-4 4.5 x 10 3 5.7 x 10-~ 1.5 x 10-4 7.1 x 10-~

Equal aliquots of bacteria at 2.0 x 108 cells per ml and phage at 6.0 x 108 plaque-forming units/ml were mixed. Survival rate under these conditions usually ranged from 5.0 x 107 to 1 x 108 cells

Measurement o f intramolecular recombination in double rec- strains The recombination assay of p1948 resolution revealed a number of features. Firstly, when single rec- mutants were analyzed, the frequency of plasmid resolution did not change more than 3-fold relative to the rec + value (data not shown). This suggests that resolution is either phage-mediated - provided phage genes are expressed or that alternative pathways not defined by the single mutants used might be involved. Secondly, the product of the recA gene, which contributes to all intermolecular recombinational pathways and has a key role in the removal of D N A damage (Ceglowski et al. 1990; Alonso et al. 1991 ; Viret et al. 1991) is not required for plasmid resolution (Table 3). Thirdly, the products of the recB, recD, recG, recF, reeL and recR genes define the ~ recombinational pathway (Alonso et al. 1991; Alonso and Liider 1991). However, there is still some doubt as to whether the recB and recG genes really belong in group a (Alonso et al. 1991). We show here that plasmid resolution is only marginally affected in the double mutant reeF15 reeD41, reeF15 recR13 and reeF15 reeL16 strains, but is blocked in the recFl5 recB2 and markedly reduced in the reeF15 recG40 strains (Table 3). Based on these results we assume that the recB and recG functions belong to a different epistatic group from that which includes the reeF function. Lastly, severe inhibition o f intermolecular recombination is observed when mutations classified within the 7 group (recH342 or recP149) were placed into either the reeF15, recR1 (c~ group) or addA5 ([3 group) genetic backgrounds (Alonso et al. 1991). As revealed in Table 3, the products o f r e c H and recP genes are also required for intramolecular recombination. Thus, genetic analysis demonstrates that apart from functions derived from the chimeric phage, resolution also requires defined host ree products. It is clear from the results reported here that the B. subtilis ree genes provide overlapping activities that can compensate for one another in single ree mutants. Inactivation o f two compensating activities, e.g. represented by the recB and

recF functions, leads to a recombinational defect. F r o m the data presented in Table 3, we assume that mutations in the addA, r e c H or reeF genes do not affect the same pathway as mutations in recB or recG. On the basis of our previous results (Alonso et al. 1991) and the genetic data presented here, we can exclude the recB and perhaps the recG genes from the a epistatic group. Therefore, we provisionally place the recB and recG genes in a new epistatic group, the ~ group. We show here that the recR reeF (a + ~), addA r e e f ([3 + ~), recR addA (ct + ~3), r e c H addA (7 + [3) and recB addA (~ + [3) double m u t a n t strains have a high residual level o f intramolecular recombination (5-80 % relative to the wild-type value) compared with the reeB r e e f (~ + ~.) double mutant strain ( < 1 0 -4 relative to wild-type levels). As previously hypothesized, this may be explained by assuming that the ~, [3, 7 and 8 groups define different recombinational pathways and that the RecF function is required for more than one pathway (Alonso et al. 1991). Acknowledgements. We thank G. Morelli for performing DNA heteroduplex analyses and B. Behrens for cooperation during early stages of this work. This research was partially supported by Deutsche Forscliungsgemeinschaft, DFG (A1 284/1-1).

References Alonso JC, L/ider G (1991) Characterization of reeF suppressors in Bacillus subtilis. Biochemie 73: 277-280 Alonso JC, Tailor RH (1987) Plasmid pC194 replication and its control in Bacillus subtilis. Mol Gen Genet 210:476~84 Alonso JC, Trautner TA (1985) Cold sensitivity in the transfer of a plasmid with a deletion hot spot into recombination deficient Bacillus subtilis cells. Mol Gen Genet 198 : 437-440 Alonso JC, Liider G, Trautner TA (1986) Requirement for the formation of plasmid-transducing particles of Bacillus subtilis bacteriophage SPP1. EMBO J 5:3723-3728 Alonso JC, Ltider G, Tailor RH (1991) Characterization of Bacillus subtilis recombinational pathways. J Bacteriol 173:3977--3980 Ballester S, Lopez P, Espinosa M, Alonso JC, Lacks SA (1989) Plasmid structural instability associated with pC 194 replication funetions. J Bacteriol 171:2271-2277

64 Biswal N. Kleinschmidt HC, Spatz HC, Trautner TA (1967) Physical properties of the DNA of bacteriophage SP50. Mol Gen Genet 100:39-55 Canosi U, Morelli G, Trautner TA (1978) The relationship between molecular structure and transformation efficiency of some S. aureus plasmids isolated from B. subtilis. Mol Gen Genet 166: 259-267 Canosi U, Lfider G, Trautner TA (1982) SPPl-mediated plasmid transduction. J Virol 44:431-436 Ceglowski P, Lfider G, Alonso JC (1990) Genetic analysis of recE activities in Bacillus subtilis. Mol Gen Genet 222:441--445 Chai S, Bravo A, Liider G, Nedlin A, Trautner TA, Alonso JC (1992) Molecular analysis of the Bacillus subtilis bacteriophage SPP1 region encompassing genes 1 to 6. The products of gene 1 and gene 2 are required for pac cleavage. J Mol Biol 225: 87-102 Deichelbohrer I, Alonso JC, Lfider G, Trautner TA (1985) Plasmid transduction by Bacillus subtilis bacteriophage SPP1 : Effects of DNA homology between plasmid and bacteriophage. J Bacteriol 162:1238-1243 Gruss A, Erhlich SD (1989) The family of highly interrelated singlestranded deoxyribonucleic acid plasmids. Microbiol. Rev. 53: 231-241 Heilmann H, Reeve JN (1982) Construction and use of SPPlv, a viral cloning vector for Bacillus subtilis. Gene 17:91-100 Horinouchi S, Weisblum B (1982) Nucleotide sequence and functional map of pC 194, a plasmid that specifies inducible chloramphenicol resistance. J Bacteriol 150: 815-825 Kooistra J, Venema G (1991) Cloning, sequencing and expression of Bacillus subtilis genes involved in ATP-dependent nuclease synthesis. J Bacteriol 173:3644-3655

Kupsch J, Alonso JC, Trautner TA (1989) Analysis of structural and biological parameters affecting plasmid deletion formation in Bacillus subtilis. Mol Gen Genet 218:402-408 Mazza G, Galizzi A (1989) Revised genetics of DNA metabolism in Bacillus subtilis. Microbiologica 12: 157-179 Mattson T, Van Houwe G, Bolle A, Epstein R (1983) Recombination between bacteriophage T4 and plasmid pBR322 molecules containing cloned T4 DNA. J Mol Biol 170:35%379 Noiret P, Petit MA, Ehrlich SD (1987) Plasmid replication stimulates DNA recombination. J Mol Biol 196:39--48 Orbach MJ, Jackson EN (1982) Transfer of chimeric plasmids among Salmonella typhimurium strains by P22 transduction. J Bacteriol 149:985-994 Petit MA, Mesas JM, Noiret P, Morel-Deville F, Ehrlich SD (1992) Induction of DNA amplification in Bacillus subtilis chromosome. EMBO J 11:131%1326 Rottl/inder E, Trautner TA (1970) Genetic and transfection studies with B. subtilis phage SP50. I. Phage mutants with restricted growth on B. subtilis strain 168. Mol Gen Genet 108:4%60 Shemyakin MF, Grepachevsky AA, Chestukin AV (1979). Properties of Bacillus subtilis ATP-dependent deoxyribonuclease. Eur J Biochem 98: 417-423 Telander-Muskavitch KM, Linn S (1981) recBC-like enzymes: Exonuclease V deoxyribonucleases. In: The Enzymes (Boyer PD. ed), vol XIV part A, Academic Press, New York, pp 234-250 Viret JF, Bravo A, Alonso JC, (1991) Recombination-dependent concatemeric plasmid replication. Microbiol Rev 55:675-683

Communicated by H. B6hme