Proc. Natl. Acad. Sci. USA Vol. 92, pp. 2244-2248, March 1995 Genetics

A 7-base-pair sequence protects DNA from exonucleolytic degradation in Lactococcus lactis (RecBCD/exonuclease/Chi/rolling-circle plasmid) INDRANIL BISWAS*, EMMANUELLE MAGUIN, S. DUSKO EHRLICH, AND ALEXANDRA GRUSSt Laboratoire de Genetique Microbienne, Institut National de la Recherche Agronomique, Domaine de Vilvert, 78352 Jouy en Josas, France Communicated by Frank W Stahl, University of Oregon, Eugene, OR, December 1, 1994

ABSTRACT Linear DNA molecules are subject to degra- hybridization (11, 24). HMW also accumulates in recBCD+ E. dation by various exonucleases in vivo unless their ends are coli on the condition that the plasmid contains a Chi site; the protected. It has been demonstrated that a specific 8-bp presence of Chi protects HMW from RecBCD degradation sequence, 5'-GCTGGTGG-3', named Chi, can protect linear (11). HMW accumulation is thus a sensitive indicator of the double-stranded DNA from the major exonu- presence of Chi on the plasmid (P. Dabert, S. Sourice, and clease RecBCD. Chi protects linear replication products of A.G., unpublished data). RC plasmids are widespread, and rolling-circle plasmids from RecBCD degradation in vivo, in their replication has been well characterized (for review, see agreement with observations in vitro. A unique 7-bp sequence, ref. 22). HMW accumulation by insertion of foreign DNA 5'-GCGCGTG-3', is shown to protect similar replication prod- inserts has been observed in several organisms (22) and, by ucts from degradation in Lactococcus lactis strains but not in analogy with E. coli, may indicate that a Chi site exists on the more distantly related Gram-positive . The properties of insert. In this work, we assume this to be the case. this sequence inL. lactis correspond to those ofa Chi site. Linear By using HMW as criterion, we found that the sequence plasmid replication products have been detected in numerous 5'-GCGCGTG-3' is necessary and sufficient to block degra- prokaryotes, suggesting the widespread existence of short spe- dation oflinear DNA inL. lactis. We predict thatExo V activity cies-specific sequences that preserve linear DNA from extensive in other Gram-positive bacteria will also be attenuated by short degradation by host cell exonucleases. specific DNA sequences.J

Double-stranded (ds) DNA ends, which exist permanently or MATERIALS AND METHODS transiently in the cell, are susceptible to degradation by cellular exonuclease (Exo). Linear , such as eukaryotic and Bacterial Growth Conditions and Transformations. E. coli some prokaryotic chromosomes, plasmids, and many viruses, TG1, supE hsdS thi(lac-proAB) F'(traD6proAB+lacIq lacZmlS) are often capped with specific sequences or proteins that (25) was grown in Luria broth and plasmid transformations appear to protect them from degradation (1-4). Certain were performed by published protocol (25). Lactococcal Escherichia coli bacteriophages do not have protected dsDNA strains used were L. lactis subsp. lactis dairy strain IL1403 (26), ends but avoid degradation by synthesizing proteins that L. lactis subsp. lactis biovar diacetylactis IL4599 (supplied by P. inhibit the major cellular Exo, known as ExoV or RecBCD Renault, Institut National de la Recherche Agronomique), L. (5-7). However, dsDNA breaks, created by cellular endonucle- lactis subsp. cremoris MG1363 (27), L. lactis subsp. lactis ases or DNA damaging agents, are substrates for ExoV and can vegetal strain NCDO 2118, Lactococcus hordnaie NCDO 2181, be lethal to the cell if not repaired. In vivo and in vitro studies Lactococcus raffinolactis NCDO 2950, and Lactococcus garviae show that an octanucleotide sequence (8) known as Chi has a NCDO 2155 (NCDO strains from Reading Laboratory, Ag- role in protecting E. coli DNA from ExoV (9-15). Degrada- riculture and Food Research Laboratory, Reading, England). tion initiating from a DNA break site is attenuated by a Chi Cultures were grown in M17 medium (28) supplemented with encounter, which may be the first step in DNA repair via 0.5% glucose. Enterococcus faecalis JH 2-2 and Streptococcus (refs. 16 and 17; for review, see thermophilus IL1704 (supplied by P. Renault) were grown in ref. 18). brain heart infusion broth (Difco/BRL). Gram-positive bac- RecBCD appears to be essential for DNA housekeeping and teria were electrotransformed as described (29). Erythromycin repair. RecBCD analogues have been found in numerous bac- was used at 100 ,ug/ml for E. coli and 5 ,tg/ml for other teria (19, 20) and may even exist in yeast (21). It has been bacteria. Crude protein extracts used to test Exo activities were proposed that control of degradative activity on self DNA is prepared from E. coli V66 (argA21 recF143 hisG4 met rpsL31 mediated by Chi-like sequences interspersed on the (ref. galK2 xyl-5 A- F-) and V186 (A[thyA-argA]232 IN[rrnD- 15 and Philippe Noirot, personal communication). To test this, we rrnE]1 A- F-) (30) and L. lactis IL1403. searched for short sequences that protect linear DNA from Plasmids and Plasmid Constructions. The RC vector degradation in Gram-positive bacterium Lactococcus lactis subsp. pG+host4, a broad-host-range RC plasmid that was modified lactis, which is taxonomically very distant from E. coli. in our laboratory for cloning purposes (previously known as We have used a replication product of rolling-circle (RC) pVE6004) (31), was used in all clonings. Random EcoRI-Sal plasmids as a linear DNA substrate for ExoV in vivo (ref. 11; I fragments (400-600 bp) of P. aeruginosa chromosomal DNA for review, see ref. 22). This substrate, presumably a o- form with a dsDNA tail (Fig. IA), is stabilized in the absence of Abbreviations: Exo, exonuclease; RC plasmid, rolling-circle plasmid; RecBCD, as shown in E. coli (11) and Bacillus subtilis (23). HMW, high molecular weight linear plasmid multimers; dsDNA, Replication of this intermediate would extend the dsDNA tail double-stranded DNA. to form tandem plasmid linear multimers of high molecular *Present address: Genetics and Biochemistry Branch, Building 10 weight (called HMW), which are readily detected on gels by R9D15, National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, MD 20892. tTo whom reprint requests should be addressed. The publication costs of this article were defrayed in part by page charge tThe sequences reported in this paper have been deposited in the payment. This article must therefore be hereby marked "advertisement" in GenBank data base (accession nos. L37763, L37764, L37765, L37766, accordance with 18 U.S.C. §1734 solely to indicate this fact. and L37767). 2244 Downloaded by guest on September 26, 2021 Genetics: Biswas et al. Proc. Natl. Acad Sci. USA 92 (1995) 2245 A Perkin-Elmer/Cetus Taq DNA polymerase in its buffer (1X) O GaRC plasmid replication intermediate with 250 pmol of a primer couple] and submitted to 30 cycles of 1 min of denaturation at 95°C, 2 min of annealing at 55°C, and 5 min of elongation at 70°C in a GeneAmp PCR System 9600 (Perkin-Elmer). The amplified plasmid DNA was treated with the restriction enzyme recognizing the primer tail, li- gated, and used to transform E. coli or L. lactis cells. Thirteen RecBCD + RecBCD - or Chi site on plasmid derivatives of insert 2, two of insert 5, and three of insert 8 (Fig. 1B) were so analyzed. Structures of plasmid inserts were verified by DNA sequencing. In some cases, the PCR tech- nique generated point mutations that were used to further map ciii". the ChiL site. HMW Oligonucleotide Linkers. Three oligonucleotide pairs were synthesized (Fig. 2A; Eurogentec) and cloned into pG+host4 B 1 2 3 4 5 6 7 8 9 polylinker sites, EcoRI-Sal I. Linker sequences were verified :.. ..:. by DNA sequencing after cloning. DNA Sequence Analysis. The dideoxynucleotide chain- ...... :: termination method was performed on both strands of plasmid c -HMW ...... DNA with the Taq Dye Cycle sequencing kit (Prism; Applied :',iilE': Biosystems) on a Perkin-Elmer PCR apparatus (GeneAmp ..,h ...... 9600) and analyzed on a model 370A automatic DNA se- ...... ::...... quencer (Applied Biosystems). .: Detection of HMW. Whole-cell DNA minilysates were pre- *. :. .::: pared from cultures grown to mid- to late-exponential phase _..._.:..... and electrophoresed on 0.6% agarose gels as described (24)...... HMW DNA, which comigrates with chromosomal DNA, was detected by Southern blot hybridization of gels (25) by using nick-translated plasmid pG+host4 32P-labeled probes (kits from Boehringer Mannheim).

- sCm Preparation of Crude Cellular Protein Extracts. A modifi- cation of the method of Eichler and Lehman (33) was used. E. t _ s: - sod,oc coli or L. lactis cells were grown to an A595 of 0.8-1.2. Cells :.: ::.:. were harvested and washed in 50 mM Tris-HCl, pH 7.5/0.1 M NaCl, at room temperature. Volumes of cell suspension in 50 FIG. 1. (A) RC plasmid replication product is a substrate for ExoV. mM Tris HCl, pH 7.5/10% (wt/vol) sucrose were adjusted so A likely RecBCD substrate that is formed by RC plasmid replication that a 1:400 dilution gave an A595 of 0.5-0.6. The cell suspen- is a o-shaped molecule with dsDNA tail (Upper). This molecule would sions were stored in aliquots at -20°C. To prepare extracts, be susceptible to Exo-mediated degradation in E. coli (Lower Left) cell suspensions were thawed, and Tris-HCl (pH 7.5), dithio- (11). Degradation is prevented if a Chi site is present on the RC threitol, and lysozyme were added to final concentrations of 10 plasmid or if the strain is RecBCD-dsDNA-Exo-defective (Lower 1 were Right); in these cases, the tail of the form is extended, as visualized mM, 0.5 mM, and mg/ml, respectively. Suspensions by HMW accumulation after Southern blot hybridization of gels (24). incubated for 30 min at 30°C (required for efficient disruption (B) Insertion of certain foreign DNA fragments into the RC plasmid of L. lactis cells) and for 30 min at 4°C. Cells were disrupted prevents HMW DNA degradation in L. lactis. Pseudomonas aeruginosa by using glass beads in Shake-It-Baby apparatus (Biospec chromosomal DNA fragments (400-500 bp) were inserted into RC Products, Bartlesville, OK) for 3 min, and cell debris was plasmid pG+host4 and resultant plasmids were established in L. lactis. removed by centrifugation at 4°C. The supernatant corre- Minilysates of strains containing plasmids were subjected to gel sponding to crude extract was collected, and protein concen- electrophoresis and Southern blot hybridization (24) by using 32p- trations were determined to be between 5 and 15 mg/ml by the labeled pG+host4 probe. Lanes: 1-8, pG+host4 containing different Bradford method (34), by using the Bio-Rad protein assay kit. inserts; 9, pG+host4 without insert. Inserts of lanes 1, 2,5,6, and 8 were The conditions used sequenced. The 392-bp insert (lane 2) was subjected to deletion ATP-Dependent Exonuclease Assay. analysis. The DNA forms migrating between supercoiled monomer are modified from those of Eichler and Lehman (33). The (scm) and HMW correspond to dimeric (scd) and open-circle (oc) reaction mixture (300 1u) contained 50 mM Tris HCl (pH 8.5), plasmid DNA. 20 mM MgCl2, 0.67 mM dithiothreitol, bovine serum albumin at 1 mg/ml, 10 nM nucleotide equivalent of E. coli in vivo- (67% G+C) were isolated by using Geneclean II kits (Bio 101) labeled chromosomal DNA ([3H]DNA; specific activity, 2 x and ligated into pG+host4; transformants were first estab- 104 cpm/nM nucleotide equivalent), 75 ,tM ATP (where lished in E. coli TG1, then transferred into L. lactis subsp. lactis added), and 10 ul of cell extract. After incubation at 30°C for IL1403, and screened for HMW accumulation. Three HMW- 30 min, the mixture was chilled on ice and 200 Al of salmon positive (Fig. 1B, lanes 2, 5, and 8) and two HMW-negative sperm DNA (2 mg/ml) and 300 ,lI of 15% (wt/vol) trichlo- (Fig. 1B, lanes 1 and 6) inserts were sequenced. Inserts roacetic acid were added. Acid-insoluble material was re- allowing HMW accumulation were shortened by restriction moved by centrifugation for 15 min at 13,000 rpm in a Sigma enzymes. Further deletions and point substitutions on sub- microcentrifuge. Radioactivity of the supernatant was mea- clones were generated by inverted PCR (ref. 32; see below). sured in a scintillation counter for 2 min. One unit is defined Deletion by PCR. Two oligonucleotides (18- to 22-mers; as the amount of enzyme that produces 1 nM of acid-soluble from Eurogentec, Seraing, Belgium), complementary to the nucleotide under the above conditions. DNA flanking the region to be deleted, were used to prime the plasmid for the inverted PCR (32). Each oligonucleotide was RESULTS AND DISCUSSION designed with a tail containing a restriction site, allowing subsequent restriction and ligation (sequences are available on ExoVActivity inL. lactis. Total protein extracts ofL. lactis were request from the authors). One hundred nanograms of plasmid prepared and tested for ATP-dependent ExoV activities (Ta- DNA was added to a PCR mixture of 100 ,ul [2.5 units of ble 1). As expected, E. coli recBCD- extracts showed no Downloaded by guest on September 26, 2021 2246 Genetics: Biswas et aL Proc. NatL Acad Sci USA 92 (1995)

B LO L2 Ll Lm -....

A L2 5 'AATTCGCATGCGCGTG 3 -HMW 3 'GCGTACGCGCACAGCT 5' L1 5 'AATTCGCATGCCACGCGCG 3' 3 'GCGTACGGTGCGCGCAGCT 5 5 3 L AATTCGCATOcCGCGTG - 3 'GCGTACgGCGCACAGCT 5 scsd,oc _ WF

_ ~~scm

FIG. 2. A 7-nt sequence protects HMW from degradation. (A) Two linkers (L2 and LV) containing the 7-bp sequence presumed to protect HMW DNA (in boldface type) in both orientations or an altered sequence (Lm) in which the 5' guanosine of L2 is replaced with a cytidine (base in lowercase type) were cloned into the EcoRI-Sal I sites of pG+host4. Linkers were verified by sequencing after cloning. L2 is oriented for recognition by a RecBCD molecule entering through the tail of the o-form replication intermediate formed during plus-strand replication. Ll is in the opposite orientation. (B) HMW profile of plasmids containing no linker (LO), or linkers L2, LI, or Lm, analyzed as described in Fig. 1B. ATP-dependent ExoV activity. E. coli recBCD+ and L. lactis plasmid replication involves two steps that could produce extracts showed similar ratios of ExoV activities with ATP, RecBCD substrates: Plus-strand replication generates 0r-form compared to extracts without added ATP. Lower specific intermediates that could provide a dsDNA tail for RecBCD to activity of ExoV found in L. lactis extracts may be due to enter (22). During minus-strand replication (opposite polari- species differences or to loss of activity during the extraction ty), conversion of a single-stranded DNA circle to dsDNA may procedure. These results show that L. lactis has ATP- give rise to a c-form substrate for RecBCD if the 5' starting dependent ExoV activity. end of newly synthesized DNA is displaced from its template Identification of the L. lactis Chi-Like Attenuation Site. We (36). In E. coli, the polar effect of Chi on RC plasmids suggests searched in L. lactis for any sequences that protect HMW from that the majority of RecBCD substrate is formed during degradation. Random 400-bp to 500-bp P. aeruginosa DNA plus-strand synthesis (11). It is possible that in L. lactis, this (67% G+C rich) fragments were cloned into an RC plasmid equilibrium is shifted and that minus-strand replication also and established in L. lactis. G+C-rich DNA was used, as it was generates substrates for ExoV. At present, the alternative found more likely to provoke accumulation of HMW in explanation, that ExoV recognizes the 7-bp sequence in either different microorganisms (24, 35). No HMW was detected in orientation, cannot be excluded. plasmid without inserts (Fig. 1B, lane 9). Inserts of three HMW Five of the seven positions of the lactococcal sequence are producer (lanes 2, 5, and 8) and two nonproducer (lanes 1 and 6) identical to the Chi sequence ofE. coli (Fig. 3). The structural plasmids were sequenced. Sequential left- or right-ward deletions that the L. lactis of one HMW-positive 392-bp insert (lane 2) resulted in the and functional similarities suggest identified identification of an 8-bp sequence, 5'-GCGCGTGC-3', that sequence corresponds to a Chi analogue, and we tentatively provokes HMW accumulation. A matching 7-bp sequence, miss- refer to it hereafter as ChiL. We expect that, as in E. coli, ChiL ing the 3'-cytidine, was found on the two other HMW-positive, has a stimulatory effect on homologous recombination when but not the two HMW-negative, clones. Insertion of this 7-bp present on a linear substrate; this prediction remains to be sequence into the RC plasmid resulted in HMW accumulation tested. The identification of ChiL and the fact that RC plasmids (Fig. 2). Alteration of the 5' guanosine (oligonucleotide Lm; Fig. with DNA inserts accumulate HMW in a wide range of 2) or the 3' guanosine (see below) abolished HMW accumulation. microorganisms (22, 35) suggest that cellular ExoVs are gen- We thus identified the 7-bp sequence 5'-GCGCGTG-3' as suf- erally accompanied by short specific sequences that attenuate ficient to protect linear DNA from degradation in L. lactis. their activities. Interestingly, in contrast with results in E. coli, orientation Uniqueness of the ChiL Sequence. The effect of point of the 7-bp sequence had little if any effect on HMW accu- substitutions on ChiL activity was examined. No HMW accu- mulation in L. lactis. Both orientations resulted in substantial mulated if any position of the sequence was altered (Fig. 4). As amounts ofHMW, although the oligonucleotide present on the the same sequence was present on the three HMW+ plasmids L2 linker sometimes gave the stronger HMW signal. RC analyzed and, in each case, a single nucleotide alteration in ChiL abolished HMW (data not shown), we conclude that the Table 1. ATP-dependent Exo activity fully active ChiL sequence is unique. However, it is possible Exo activity that substitutions other than those directly examined have Specific activity, partial ChiL activity, as reported for the E. coli Chi (37). Organism + ATP - ATP Protein, ,ug units/mg of protein Screening of protective DNA inserts did reveal low-level E. coli HMW producers (e.g., see Fig. 1B), suggesting that partially recBCD+ 35,910 14,820 120 8.78 active ChiL sites might exist. One insert that gave a weak HMW recBCD- 14,680 14,550 120 <0.05 L. lactis 7,965 3,345 90 2.57 5 'GCGCGTG 3 Lactococcus lactis II I I One unit of Exo activity is the amount of enzyme that produces 1 5 GCTGGTGG 3' Escherichia coli nM acid-soluble nucleotide under the above conditions. Exo activity was determined from two lysates, and the average cpm are presented. FIG. 3. Comparison of E. coli and L. lactis subsp. lactis Chi A 1 nM equivalent of nucleotide produces 2 x 104 cpm. sequences. Identities are indicated by bars. Downloaded by guest on September 26, 2021 Genetics: Biswas et al. Proc. Natl. Acad. Sci. USA 92 (1995) 2247

SEQUENCE the Chi of E. coli, and we suggest that recombination in the vicinity of Chi-like sequences may be stimulated by other

5' properties of these sequences. - - - A Role for Chi. Chi is present on the E. coli genome once C - - - - A - - - - - every 4 kb on the genome, about 8 times more frequently than - - T - -- - predicted from a random representation (15, 44). Preliminary - -- T - - - evidence based on sequenced L. lactis IL1403 genomic DNA - -- - T - - suggests that ChiL (86% G+C rich) is also overrepresented on

-- - - C- the L. lactis chromosome (38% G+C-rich). A contiguous - -- - - C 27-kb sequenced region, constituting the his-ilv operons (Gen------G - bank accession nos. M90760 and M90761 and P. Renault, personal communication), contains four ChiL sites; the total -~~A 42-kb sequenced IL1403 chromosomal DNA contains five FIG. 4. ChiL sequence is unique. The ChiL substitutions shown are ChiL sites (added accession nos. L27422, M90762, M87483, -- X75982, and X64887). The expected frequency of finding ChiL, present on certain RC plasmids~cused in this study and were tested for HMW accumulation: GI -* C substitution is present on Lm linker (Fig. based on the G+C content, is once in 34 kb for both orien-

-- 2); C2 -* A and C4 T substitutions are present on different tations. These higher than expected frequencies may reflect a subfragments of insert 2 (Fig. 1B) mutated by PCR; G3 -* T, G5 -> functional role of Chi in regulating ExoV activity on dsDNA T, and G7 -* A substitutions are present on pG+host4; G5 C ends generated during replication with or without DNA dam- substitution is present on a subfragment of insert 2; T6 -* C, T6 G, age (18, 45). Does the cell distinguish between self and foreign and G7 -* C substitutions are present on insert 1. In all cases, single-base-pair alterations in ChiL abolish HMW accumulation. linear DNA? RecBCD, with restriction-modification systems, could serve as a first line of defense against heterologous signal (data not shown) contained three ChiL-like sequences incoming DNA. We propose that the overrepresented Chi with substitutions that had not been examined. sequence present on self DNA and on DNA of closely related ChiL Prevents HMW Degradation Only in Closely Related organisms will protect these DNA species from degradation, Bacteria. The activity of ChiL was examined in several Gram- either after DNA damage or during DNA uptake. In contrast, positive species (Table 2). The ChiL-containing RC plasmids heterologous DNA contains the Chi site at only the predicted accumulate HMW in three subspecies of L. lactis, suggesting low frequency and would be degraded before further discrim- that ChiL is recognized among closely related organisms. Very ination by recombination processes can begin. Examples of weak HMW signals of ChiL-containing plasmids were detected "self' recognition come from Haemophilus influenzae, for in three lactococcal species distantly related to L. lactis; among which DNA uptake during competence is preferential for self these species, Chi may be conserved but not identical. Unre- sequences. In the former case, a specific 11-bp motif was lated genera S. thermophilus and Enterococcus faecalis appear identified that confers self-specificity (46); the sequence, like to have distinct Chi sequences. In the former host, HMW was Chi, is present every -4 kb on the chromosome rather than at detected in the absence of ChiL, suggesting that a different its expected low frequency of about once per chromosome sequence is recognized as Chi. In the latter host, no HMW (47). response was observed, regardless of ChiL. These results are in agreement with previous in vitro studies on the E. coli Chi site, We greatly appreciate the participation of T. H6g6 and S. Sourice showing that Chi recognition is limited to related host strains in DNA sequencing. We are grateful to P. Dabert for his participation in developing the model system and for extensive discussion of this (38). Thus, these results predict that Chi-like sites will differ work. We thank P. Duwat for his advice in PCR. P. Noirot, P. Duwat, between organisms. Recombination hot spots in eukaryotes S. Sourice, M. El Karoui, P. Langella, and Y. Le Loir offered many have been previously associated with the presence of E. coli suggestions and counsel on Chi. P. Renault provided us with unpub- Chi homologues (39-43). Chi-like sites have been reported on lished L. lactis sequence information and gave his insight on strain immunoglobulin genes (39), within the major human histo- relatedness in the course of this project, and he and M. C. Chopin compatibility complex DNA (40), on minisatellite DNA (41), kindly provided several of the non-E. coli strains used in this work. and adjacent to oncogene translocation break points (42, 43). Our data predict that a eukaryotic Chi site would differ from 1. Hinnebusch, J. & Barbour, A. G. (1991) J. Bacteriol. 173, 7233- 7239. Table 2. ChiL is recognized in a restricted host range 2. Zakian, V. (1989) Annu. Rev. Microbiol. 23, 579-604. 3. Oliver, D. B. & Goldberg, E. B. (1977)J. Moi. Biol. 116,877-881. ChiL sequence 4. Williams, J. G. K. & Radding, C. M. (1981).J. Virol. 39, 548-558. Species tested recognition 5. Zissler, J., Signer, E. R. & Schaefer, F. (1971) in The Bacterio- phage Lambda, ed. Hershey, A. D. (Cold Spring Harbor Lab. Subspecies 455. L. lactis subsp. lactis dairy Press, Plainview, NY), p. + 6. Murphy, K. C. (1991) J. Bacteriol. 173, 5808-5821. L. lactis subsp. lactis vegetal 7. Murphy, K. C. (1994) J. Biol. Chem. 269, 1-10. L. lactis subsp. lactis biovar diacetylactis 8. Smith, G. R., Kunes, S. M., Schultz, D. W., Taylor, A. & Triman, L. lactis subsp. cremoris K. L. (1981) Cell 24, 429-436. Species + 9. Lam, S. T., Stahl, M. M. & Stahl, F. W. (1974) Genetics 77, L. garviae 425-433. L. hordnaie 10. Ponticelli, A. S., Schultz, D. W., Taylor, A. F. & Smith, G. R. L. raffinolactis (1985) Cell 41, 145-151. Genera 11. Dabert, P., Ehrlich, S. D. & Gruss, A. (1992) Proc. Natl. Acad. Enterococcus faecalis Sci. USA 89, 12073-12077. 12. Thaler, D. S., Sampson, E., Siddiqi, I., Rosenberg, S. M., Thom- Streptococcus thermophilus ason, L. C., Stahl, F. W. & Stahl, M. M. (1989) Genome 31, Strain numbers are in Materials and Methods. pG+host4 derivatives 53-67. without or with LI or L2 were analyzed for HMW accumulation in 13. Stahl, F. W., Thomason, L. C., Siddiqi, I. & Stahl, M. M. (1990) different Gram-positive strains. Strains were evaluated as recognizing Genetics 126, 519-533. ChiL if plasmids accumulate HMW with but not without ChiL. +, ChiL 14. Dixon, D. A. & Kowalczykowski, S. C. (1993) Cell 73, 87-96. recognized; -, ChiL not recognized; +, faint ChiL-specific signal 15. Kuzminov, A., Schabtach, E. & Stahl, F. W. (1994) EMBO J. 13, observed. 2764-2776. Downloaded by guest on September 26, 2021 2248 Genetics: Biswas et al. Proc. Natl. Acad. Sci. USA 92 (1995)

16. Dower, N. A. & Stahl, F. W. (1981) Proc. Natl. Acad. Sci. USA 78, 31. Maguin, E., Duwat, P., Hege, T., Ehrlich, S. D. & Gruss, A. 7033-7037. (1992) J. Bacteriol. 174, 5633-5638. 17. Stahl, F. W., Craseman, J. M. & Stahl, M. M. (1975) J. Mol. Biol. 32. Hemsley, A., Arnheim, N., Toney, M. D., Cortopassi, G. & Galas, 94, 203-212. D. J. (1989) Nucleic Acids Res. 17, 6545-6551. 18. Myers, R. S. & Stahl, F. W. (1994) Annu. Rev. Genet. 28, 49-70. 33. Eichler, D. C. & Lehman, I. R. (1977) J. Biol. Chem. 252, 19. Telander Muskavitch, K. M. & Linn, S. (1981) in The Enzymes, 499-503. ed. Boyer, P. C. (Academic, New York), Vol. 14, pp. 233-250. 34. Bradford, M. M. (1976) Anal. Biochem. 72, 248-254. 20. Kooistra, J., Haijema, B. J. & Venema, G. (1993) Moi. Microbiol. 35. Gruss, A. & Ehrlich, S. D. (1989) Microbiol. Rev. 53, 231-241. 7, 915-923. 36. Mok, M. & Marians, K. (1987) J. Biol. Chem. 262, 2304-2309. 21. Fraser, M. J., Koa, H. & Chow, T. Y.-K. (1990) J. Bacteriol. 172, 37. Cheng, K. C. & Smith, G. R. (1987) J. Mol. Biol. 194, 747-750. 507-510. 38. McKittrick, N. H. & Smith, G. R. (1987) J. Mol. Biol. 210, 22. Janniere, L., Gruss, A. & Ehrlich, S. D. (1993) in Bacillus subtilis 485-495. and Other Gram-Positive Bacteria: Biochemistry, Physiology, and 39. Kenter, A. L. & Birshtein, B. K. (1981) Nature (London) 293, Molecular Biology, eds. Hoch, J. A. & Losick, R. (Am. Soc. 402-404. Microbiol., Washington, DC), pp. 625-644. 23. Viret, J.-F. & Alonso, J. C. (1987) Nucleic Acids Res. 15, 6349- 40. Wu, S., Saunders, T. L. & Bach, F. H. (1986) Nature (London) 6367. 324, 676-679. 24. Dabert, P., Ehrlich, S. D. & Gruss, A. (1992) J. Bacteriol. 174, 41. Jeffreys, A. J., Wilson, V. & Thein, S. L. (1985) Nature (London) 173-178. 314, 67-73. 25. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) Molecular 42. Krowczynska, A. M., Rudders, R. A. & Krontiris, T. G. (1990) Cloning: A Laboratory Manual (Cold Spring Harbor Lab. Press, Nucleic Acids Res. 18, 1121-1127. Plainview, NY), 2nd Ed. 43. Wyatt, R. T., Rudders, R. A., Zelenetz, A., Delellis, R. A. & 26. Chopin, A., Chopin, M.-C., Moillo-Batt, A. & Langella, P. (1984) Krontiris, T. G. (1992) J. Exp. Med. 175, 1575-1588. Plasmid 11, 260-263. 44. Burland, V., Plunkett, G., III, Daniels, D. L. & Blattner, F. R. 27. Gasson, M. J. (1983) J. Bacteriol. 154, 1-9. (1993) Genomics 16, 551-561. 28. Terzaghi, B. E. & Sandine, W. E. (1975) Appl. Microbiol. 29, 45. Louam, J.-M., Louarn, J., Francois, V. & Patte, J. (1991) J. 807-813. Bacteriol. 173, 5097-5104. 29. Holo, H. & Nes, I. F. (1989) Appl. Environ. Microbiol. 55, 46. Danner, D. B., Deich, R. A., Sisco, K. L. & Smith, H. 0. (1980) 3119-3123. Gene 11, 311-318. 30. Amundsen, S. K., Neiman, A. M., Thibodeaux, S. M. & Smith, 47. Sisco, K. L. & Smith, H. 0. (1979) Proc. Natl. Acad. Sci. USA 76, G. R. (1990) Genetics 126, 25-40. 972-976. Downloaded by guest on September 26, 2021