PLASMID

27, 169- I76 ( 1992)

Mobilization

and Location of the Genetic Determinant of Chloraknphenicol Resistance from Lactobacillus plantarum caTC2R

C.A~,*D.COLLINS-THOMPSON,~C.DUNCAN,~ANDM.E.STILES* *Department of Food Science, University of Alberta, Edmonton, Alberta, Canada T6G 2P5; and tDepartment of Environmental Biology, University of Guelph, Guelph. Ontario, Canada NlG 2 WI Received June 6, 199I ; revised October 15, 1991 The mobilization of a nonconjugative plasmid (pCaT) that mediates chloramphenicol resistance in Lactobacillusplantarum caTC2R was achieved by comobilization with the conjugative plasmid pAMpI. The conjugation studies conhrmed that the 8.5-kb pCaT in L. plantarum caTC2R contains the gene responsible for chloramphenicol resistanceand that the plasmid has several unique restriction sites which make it useful for genetic studies in Carnobacterium spp. Cloning studies showed that the gene responsible for chloramphenicol resistanceis located in the 2.6-kb EcoRV-SalI region of pCaT. This was confirmed by probing the 3.0-kb BglII fragment of pCaT with a biotin-labeled 1.6-kb BstEII-HpaII fragment from the streptococcal-derived plasmid pVA797(Cm’). Expression of chloramphenicol resistance in Camobacterium as well as in other Lactobacillus species was achieved by electrotransformation using donor DNA from pCaT. Q 1992 Academic Press, Inc.

The lactic acid bacteria associatedwith vacuum or modified atmosphere packaging with elevated levels of CO, have become of increased interest with the widespread use of these preservative packaging systems for meats. The lactic microflora that develops on meats is diverse (Shaw and Harding, 1984), consisting of Lactobacillus and Leuconostoc spp. including the newly named genus, Carnobacterium (Collins et al., 1987). Camobacteria are of interest because they are nonaciduric, heterofermentative bacteria which in our experience do not produce large amounts of carbon dioxide (Ahn and Stiles, 1990a). Some strains of Carnobacterium piscicola produce bacteriocins active against a relatively wide range of lactic acid bacteria, and also Enterococcus and Listeria spp. (Ahn and Stiles, 1990b). Our efforts to clone the bacteriocin gene from C. piscicola LV 17 using vector DNA available to us were not successful (unpublished data). Lack of an appropriate marker or genetic delivery system has been a major problem for strain improvement of meat starter cultures by genetic technology (Chassy, 1987; McKay and Baldwin, 1990);

hence we required a system, possibly more closely related to our bacteria, to clone the bacteriocin gene from our test organism. Lactobacillus plantarum caTC2R isolated from raw, ground pork showed chloramphenicol resistance dependent upon chloramphenicol acetyltransferase (CAT)’ activity and mediated by an 8.5-kb plasmid (Jewel1 and Collins-Thompson, 1989). The possible use of this 8.5-kb plasmid for a chloramphenicol resistance marker or as vector DNA is of interest. Further study of this plasmid in L. plantarum caTC2R was hindered by the presence of two other residential plasmids. Attempts to separatethe 8.5-kb plasmid by electrotransformation of a mixture of all three plasmids into a plasmidless strain of C. piscicola LV 17 with selection for chloramphenico1 resistance was unsuccessful. Conjugal transfer of the 8.5-kb plasmid was also unsuccessful. To circumvent this problem, the con’ Abbreviations used: CAT, chloramphenicol acctyltransferasc; MLS, macrolide-lincosamide-streptogramin B; Em, erythromycin; Cm, chloramphenicol; MIC, minimum inhibitory concentration.

169

0 147-619X/92 $5.00 Cupyrigbr 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.

170

AHN

jugative broad-host-range plasmid PAM/~ 1 with macrolide-lincosamide-streptogramin B (MLS) resistance was selectedas a mobilizing agent (Clewell et al., 1974). pAMP 1 has a broad host range in lactic acid bacteria and it has been used successfully to mobilize nonconjugative plasmids in various species(Gasson, 1990), including proteinase plasmids in Lactococcus lactis subsp. lactis UC 3 17 (Hayes et al., 1990). This study reports the successful mobilization of the 8.5kb chloramphenicol resistance plasmid (pCaT) from L. plantarum caTC2R to various Carnobacterium spp. by pAM@lassociated mobilization technique, followed by intensive restriction mapping of the plasmid and location of the chloramphenicol resistance gene. Expression of chloramphenico1resistancein other speciesis also reported.

ET AL.

Conjugation experiments. The experimental scheme for two-stage conjugation to mobilize the 8.5-kb plasmid by PAM/~1 is illustrated in Fig. 1. A 0.22-pm filter membrane (Nucleopore Corp., Pleasantview, CA) was used for filter mating. Conjugation mixtures were incubated at 30°C for 18 h under aerobic conditions. Transconjugants were screened by fermentation of appropriate carbohydrates and by plasmid profiles. Plasmid extraction. The method described by Ahn and Stiles ( 1990b) was used for miniscalepreparation of plasmids. Large-scale extraction of plasmid DNA was done by scaling-up (X 100) of the miniprep technique. If necessary,the plasmid DNA was further purified by CsCl density gradient centrifugation (Sambrook et al., 1989). Gel electrophoresis was done on 0.8% agarose gel (Pharmacia LKB Biotechnology 5-75 182, Uppsala, SweMATERIALS AND METHODS den) in TAE buffer (40 mM Tris-acetate, 1 Bacterial strains and culture media. BactemM EDTA) at 5 V/cm for 3 h. rial strains used in this study are listed in TaSeparation of pCaT and expression in ble 1. L. plantarum caTC2R containing the other strains. Separation of the plasmids was plasmid pCaT that mediates chloramphenicarried out by electrotransformation of the co1 (Cm) resistance was isolated from meat plasmid DNA extracted from the second(Jewel1and Collins-Thompson, 1989). Streptococcus sanguis Challis DL125 containing stage transconjugant containing both the the erythromycin (Em)-resistant, conjugative plasmids pAM@1 and pCaT into C. piscicola plasmid pAMPI was used as the donor for LVl7.C2 and screening for Em’ and Cm’ this plasmid. A plasmidless strain of C. pisci- electrotransformants. Electrotransformation cola LVl7.C2 (Ahn and Stiles, 1990b) was was done with a Gene Pulser (Bio-Rad Laboused as a recipient of pAM/31 and (or) pCaT ratories, Richmond, CA) according to the the 8.5-kb Cm’ plasmid from L. plantarum manufacturer’s protocol in Hepes-buffered caTC2R. Carnobacterium divergens LV 13 sucrose electroporation buffer (HEB; 272 was used as the indicator strain for bacterio- mM sucrose, 1 mM MgClz, 7 mM Hepes, pH cinogenic (Bat+) strains. Escherichia coli 7.4) with a single pulse at 25 @Fdand 250 BHB 2600 containing pMG36e (Van de ohms resistance at 12.5 kV/cm. The 8.5-kb Guchte et al., 1989) was obtained from G. (Cmr) plasmid from the transformant was furVenema (Department of Genetics, Univer- ther purified by CsCl density gradient ultrasity of Groningen, Haren, The Netherlands). centrifugation and electrotransformed into Stock cultures were stored in cooked meat C. piscicola UAL26, C. divergens LV 13, Leumedium (Difco Laboratories, Inc., Detroit, conostoc mesenteroides UAL60, LactobacilMI) at 4°C and subcultured in APT broth lus casei ATCC 393, and L. plantarum NC8. (Difco) at least twice before use in experi- Cm’ transformants of each strain were ments. LB broth (Sambrook et al., 1989) was checked for their plasmid profiles and miniused for growing the E. coli strains. Where mum inhibitory concentration (MIC) of appropriate, antibiotics were added to the chloramphenicol in APT agar. Restriction analysis. Purified 8.5-kb plasAPT broth to ensure antibiotic resistance of mid DNA was digested with selected restricthe strains.

GENETIC DETERMINANT

OF CHLORAMPHENICOL

171

RESISTANCE

TABLE 1 BACTERIALSTRAINS Strain Streptococcussanguis Challis DL I25 Lactococcus lactis Lactobacillus plantarum caTC2R

Phenotype* (plasmid)

Source

Em’ (pAMpI)

LeBlanc and Lee ( 1984)

Cm’ (pVA797)

TRKb

Cm’ mal-

lac-

Jewel1and Collins-Thompson (1989) Conjugal transfer of pAM@l AWS* Chassy and Flickinger ( 1987)

Cm” mal+ Cm” mal+

Ahn and Stiles ( 1990b) Ahn and Stiles (1990a)

Cm’ mal+

BGSb

Plasmidless Cm’ (pC194)

Meat isolate Bacillus Genetic Stock Culture Collection Van de Guchte et al. (1989)

caTC2R.pAM NC8 Lactobacillus casei ATCC 393 Carnobacterium piscicola LV17C2 UAL26 Carnobacterium divergens LV13 Leuconostoc mesenteroides UAMO Bacillus subtilis IE 17

Cm’ mal- Em’ (pAMPI)

Escherichia coli BHB 2600

Em’ (pMG36e)

’ Cm, chloramphenicol; Em, erythromycin; lac, lactose; mal, maltose; s, sensitive; r, resistant. ’ Strains obtained from T. R. KIaenhammer (TRK) and A. W. Shrago (AWS), North Carolina State University, Raleigh, NC, and from Dr. B. G. Shaw (BGS), Institute of Food Research, Langford, Bristol, UK.

tion enzymes (Boehringer-Mannheim Canada, Laval, Quebec) using procedures specified by the supplier and mapped by procedures described by Sambrook et al. (1989). Calibration of fragment sizes was done by multiple regression analysis (Rochelle et al., 1985) using the HindI digest of bacteriophage X DNA (Boehringer-Mannheim) as mobility standards. Cloning. Cloning of the DNA fragment encoding chloramphenicol resistance in the 8.5kb plasmid was done by the procedure described by Sambrook et al. (1989). The 8.5kb plasmid cut by various sets of restriction enzymes was ligated with pMG36e vector DNA which was cut by the same set of restriction enzymes. T4 DNA ligase was obtained from BRL (Bethesda Research Laboratories, Life Technologies, Inc., Gaithersburg, MD) and used as recommended by the supplier. Ligation buffer without ethylene glycol was used. The ligation mixture was incubated at

16°C for 6 h, and the DNA was introduced into plasmidless C. piscicola strains LV17.C2 or UAL26 by electrotransformation as described earlier. Transformants were screened on APT agar plates containing 8 pg of chloramphenicol and 10 pg of erythromycin per milliliter of agar. DNA hybridization. DNA hybridization experiments were performed using the blotting procedure described by Sambrook et al. (1989). The probe DNA (1.6 kb BstEII/ HpaII fragment from the lactococcal plasmid pVA797) was labeled by nick translation with biotinylated dUTP using the Blu-gene kit (BRL). This was hybridized against the 3.0-kb BgIII fragment of pCaT using procedures specified by the supplier. RESULTS

The first-stage conjugation to transfer pAMfl1 from 5’. sanguisChallis DLl25 to L.

172

AHN ET AL.

s.sang&

LV17.C2 occurred with an efficiency of 2.0 1 X 10T6transconjugants per donor cell. Plasmid DNA profiles of the transconjugant cells confirmed that transfer of Cm resistance was due to cotransfer of the 8.5kb plasmid with pAM/31 (Fig. 2, lane 5). Transfer of the 8.5kb chloramphenicol resistance plasmid into C. L. $aiuown caTCZR.pAM piscicolu LV 17.C2 stabilized the plasmid Em’cdcefS C. piscida LV17.C2 EmsCmsCef’ which had shown an irregular pattern on gel electrophoresis as shown in Fig. 2 (lanes 2 to 4). Carbohydrate fermentation patterns of the transconjugants, including maltose, sucrose, and lactose in the presence of chloramphenicol, cephaloridine, and erythromycin as selective agents identified the origin of the transconjugants as C. piscicolu LV 17.C2. The efficiency of conjugation in the presence of DNase I was not changed, and controls with cell filtrates did not show any transconjugants on the screening media. The C. piscicolu LV 17.C2 transconjugant containing pAM@l and the 8.5-kb Cm’ plasElectrotransfomtation into C. piscicola LV17.C-2, mid was subcultured in APT broth without UAL26. C. divergem LV13 ETmvm’ the addition of Em for 50 generations. Em” PIG. 1. Experimental scheme for mobilization of 8.5 strains were not detected. However, when

alallis DL125 Em’ @Ah461)

kb Cm’ plasmid pCAT by pAMB1 and separation of pCAT from pAMj3 1by electrotransfonnation. Em, erythromycin; Cm, chloramphenicol; Cef, cephaloridine; r, resistant; s, sensitive.

plunturum caTC2R resulted in a conjugation efficiency (expressed as the number of Cm’ Em’ transconjugants per donor cell) of 4 X 1O-‘. The efficiency was not affected by treatment with DNase I. Control conjugation experiments with filtrates of donor or recipient cells with the respective cells did not reveal any transconjugants on the Cm-Em screening plates. The plasmid profile of the transconjugant designated as L. plantarum caTC2R.pAM showed the presence of 26.5-kb pAM@l and the three residential plasmids of the L. pluntarum strain (Fig. 2, lanes 2 to 4). All transconjugants grew but failed to ferment maltose, confirming that the origin of the transconjugants was L. plantarum caTC2R. The second-stageconjugation between L. plantarum caTC2R.pAM and C. piscicola

12345678

35.e ei -4

PAM

4.0

FIG. 2. Plasmid profiles of transconjugants and electrotransformants. Lane 1, mobility standard E. co/i V5 17; lane 2, L. plantarum caTC2R, lane 3, S. sanguis Challis DL 125 containing pAM@1; lane 4, L. plunturum caTC2R containing pAM#Il (caTC2R.pAM); lane 5, C. piscicola LV 1l.C2 containing pAM@1and pCaT; lanes 6 to 8, C. piscicola LV 17 .C2, C. piscicola UAL26, and C. divergem LV13 containing pCaT, respectively. Molecular weights of mobility standard are expressed in MDa.

GENETIC DETERMINANT

OF CHLORAMPHENICOL

173

RESISTANCE

TABLE 2 EFFICIENCYOF ELEC~ROTRANSFORMA~ON OF pCaT INTO VARIOUSSTRAINSOF Curnobacterium AND Luctobucihs AND MINIMUM INHIBITORYCONCENTRATION (MIC) OF CHLORAMPHENICOL IN EACH TRANSFORMANT Strain

Time constant’ bs)

C. piscicola LV 17.C2 C. piscicola UAL26 C. divergens LV 13 L. plantarum NC8 L. casei ATCC 393

2.3 2.2 2.0 2.8 2.3

Efficiency (transformants/pg DNA) 5.9 x 2.8 x 1.1 x 2.3 x 5.1 x

lo4 105 10’ 10’ 102

Wml

MIC of apar) 64 32 64 30 20

a Time constant for electrotransformation expressed in ms.

plasmid DNA from this strain was electrotransformed into plasmidless C. pisciculu LV17.C2 (time constant 2.2 ms at 12.5 kV/ cm, 25 PF, and 200 ohms), all electrotransformants that appeared on the chloramphenicol-containing screening media were erythromycin-sensitive. Plasmid profiles of the transformants confirmed that only the Cm’ plasmid had been transformed (Fig. 2, lane 6). The 8.5-kb plasmid in these transformants was prepared by large-scale extraction and purified by CsCl gradient ultracentrifugation. The purified plasmid was again electrotransformed into C. piscicola LV 17.C2 as an efficiency control, and into C. piscicolu UAL26, C. divergens LV13, Leuc. mesenteroides UAL60, L. plantarum NC8, and L. casei ATCC 393. The efficiency of each transformation under the same conditions of electroporation is shown in Table 2. High efficiencies of lo4 to lo5 transformants/pg DNA were achieved for the strains of C. piscicola, but low frequencies of 10’ to 102/pg DNA were achieved for other organisms. No transformants were detected for Leuc. mesenteroides UAL 60. Plasmid profiles of the transformants showed that the 8.5-kb plasmid is present intact in Camobacterium species (Fig. 2, lanes 7 and 8). The presence of the 8.5-kb plasmid rendered the transformants chloramphenicol-resistant up to 32 to 64 pg/ml in strains of Carnobacterium and 20 to 30 pg/ml in strains of Lactobacillus (Table 2) when tested on APT agar.

Plasmid profiles of L. casei transformants also showed the presence of the 8.5-kb plasmid (Fig. 3). The restriction enzyme digests with BglII and HpaII (Fig. 3, lanes 2 and 3) were similar to those of the donor plasmid pCaT (Fig. 3 lanes 6 and 7). The size of the plasmid was confirmed by the Hind111digests

oc &5kb

FIG. 3. Restriction enzyme analysis of plasmid DNA in transformed L. plantarum NCS/L. casei ATCC 393 and donor strain L. pIantanrm caTC2R. Lane 1, transformed L. casei/L. plantarum with 8.5 kb plasmid; lanes 2 and 3, Bg/II and HpaII digests of transformed 8.5 kb plasmid, respectively; lane 4, Hi&III digest of bacteriophage X DNA; lane 5, pCaT plasmid from L. pluntnrum caTC2R; and lanes 6 and 7, HpaII and Bg&I digests of L. planturum caTC2R plasmid, respectively. OC, open circular form of 8.5-kb plasmid.

174

AHN ET AL.

was further refined by digesting the plasmid with BstEII/HpaII to obtain a 1.6-kb fragment which contains the CAT gene. This fragment was biotin-labeled by nick translation and probed against the two BgnI fragments of the 8.5-kb plasmid. The results showed that the 3.0-kb fragment containing the EcoRV-SalI region was homologous to the pVA797 fragment (Fig. 5). The lack of known restriction sites on the 2.6-kb EcoRV-SalI fragment prevented a more precise location of the CAT gene on pCaT. FIG. 4. Restriction map of 8.5-kb Cm’ plasmid pCaT. Hind111 site is arbitrarily assigned as zero. The heavy black bar indicates the 2.6-kb EcoRV-WI1 region that contains the CAT gene.

DISCUSSION

The separation of the nonconjugative 8.5kb plasmid associated with chloramphenicol resistance in L. plunturum caTC2R (Jewell and Collins-Thompson, 1989) from the other two residential plasmids was achieved by introduction of the conjugative plasmid pAM/31 into L. plunturum caTC2R in the first-stage conjugation and by mobilization of the 8.5-kb plasmid together with pAMB1 in the second-stageconjugation into the plasmidless strain of C. piscicolu LV 17.C2. Ex-

of bacteriophage X DNA (Fig. 3, lane 4). The 8.5kb plasmid was mapped with 12 kinds of restriction enzymes. As shown in Fig. 4, the plasmid has unique restriction sites for AvaI, EcoRI, EcoRV, HindII, HindIII, HpaII, &I, SalI, and XbaI, and two restriction sites for each of AccI, Bg/II, and SphI. The AccI and Hind11 sites at the 6.9-kb point and SulI and BglII sites at the 6.6-kb point from the Hind111site are too close to map their correct 1 3 345 12345 positions. The 5.9-kb fragment of PstI-Hi&III, the 6.6-kb fragment of SuZI-HindIII, the 4.0-kb fragment of PstI-&r/I, the 6.4-kb fragment of SphI, and the 5.7-kb fragment of EcoRIEcoRV were cloned into pMG36e, rendering transformants resistant to chloramphenicol up to 32 or 64 &ml. The 2.6-kb EcoRV,WI region of the pCaT plasmid is responsible for chloramphenicol resistance. The location of the chloramphenicol gene 1.6r was further confirmed by DNA hybridization experiments. In preliminary studies the staphylococcal plasmid pC 194 (Cmr) isolated from B. subtilis IE17 was biotin-labeled and probed against the 8.5-kb plasmid along FIG. 5. DNA hybridization studies with 8.5-kb plaswith appropriate controls. No hybridization mid with a 1.6-kb fragment from pVA797 as a probe. was detected between these two plasmids. 1, HindI digest of bacteriophage X DNA; lane 2, In an additional experiment, a biotin-labeled Lane BsfEII-HpuII digest of pVA797 and homology to probe probe made using the plasmid pVA797 (a de- (control); lane 3, pCaT (8.5 kb) (control); lane 4, rivative of pIPSOl) was shown to hybridize pVA797 (control); and lane 5, BgfiI fragment of pCaT with the 8.5-kb plasmid. The pVA797 probe and homology of the probe to the 3.0-kb BglII fragment.

GENETIC DETERMINANT

OF CHLORAMPHENICOL

pression of chloramphenicol resistance in a heterospeciessuch as Carnobacterium and in other Lactobacillus spp. proved that the 8.5kb plasmid mediates cbloramphenicol resistance and excluded the possibility of a concerted mode of action with the chromosome or other residential plasmids in L. plantarum. Furthermore, separation of the 8.5-kb plasmid from pAMj3 1 in transconjugants was easily achieved by separative electrotransformation, proving the applicability of the technique to remove pAM@l when used as a mobilizing agent for a nonconjugative plasmid far smaller than itself. The conjugal transfer efficiency of PAM@1 from S. sanguis Challis DL125 to L. plantarum caTC2R was similar to those reported by West and Warner (1985) between S. lactis 7 12 lac- (pAM@l ) and L. plantarum 340, 352, and 1752 and higher than those reported by Sasaki et al. (1988) between S. faecalis JH2-2 (PAM@1) and L. plantarum JCM 1149 on a Millipore filter membrane (0.45 pm). However, Sasaki et al. (1988) did not detect any transconjugants when they used a Nucleopore filter membrane (0.4 pm). They attributed this to the structural characteristics of the membrane. Unfortunately, there are no data for a 0.22-pm Nucleopore filter membrane for S. sanguis Challis as donor. The successful conjugation in our experiments with the 0.22-pm Nucleopore filter membrane is due either to the strain-specific characteristics or to procedural differences. Intergeneric transfer of pAMP1 from the first-stage transconjugant L. plantarum caTC2R (pAM/31) to Carnobacterium in the second-stage conjugation showed intactness of pAM@l conjugativity (tra region of the plasmid) even after the first-stage conjugation. West and Warner (1985) also reported conservation of conjugal capability of pAM@l when L. plantarum was used as the first host. However, the mechanism of cotransfer of the 8.5-kb plasmid with pAMP 1 is uncertain becauseco-integrate formation reported by Oultram et al. (1987) was not observed. Results of cloning experiments defined a 2.6-kb EcoRV-SalI region ofpCaT asrespon-

RESISTANCE

175

sible for chloramphenicol resistance. For the gram-negative CAT gene, Alton and Vapnek (1979) reported a 1102-bp region of Tn9 which is responsible for chloramphenicol resistance in E. coli. The size of the structural gene itself was deduced as 657 bp from the primary structure data of the type I CAT enzyme (Shaw et al., 1979). For the gram-positive CAT gene, a staphylococcal chloramphenicol resistanceplasmid, pC 194,wascompletely sequenced, and a 1305-bp region which was responsible for the structural gene, promoter, and regulatory element for expression of chloramphenicol resistance was defined (Horinuchi and Weisblum, 1982). Compared with the data for other gram-positive and gram-negative CAT genes, the 2.6kb region of pCaT needs further study to narrow the region of the chloramphenicol resistance gene. The BstEII/HpaII 1.6-kb fragment of pVA797 was shown by Pepper et al. (1987) to carry most of the CAT genes. The DNA homology studies with the CAT gene from the streptococcal plasmid pVA797 raises interesting questions regarding the origin of the chloramphenicol plasmid in L. plantarum caTC2R. The results of this study suggestthat the origin of pCaT plasmid may have been Streptococcus. Evidence of genetic exchange by conjugation between Lactobacillus and Streptococcus has been demonstrated (Shrago and Dobrogosz, 1988; West and Warner, 1985). Such genetic exchange could easily occur in the intestinal tract of animals because both Lactobacillus and Streptococcus are naturally present. As suggested by Smith and Bums (1984), further DNA homology studies with other grampositive and gram-negative CAT geneswould be a good way to trace the origin of chloramphenicol resistance in L. plantarum. The high transformation efficiency and intact expression of pCaT in Carnobacterium and the presence of several unique restriction sites create a good opportunity to use this plasmid in genetic studies of Carnobacterium speciesin two respects: using chloramphenico1 resistance as a genetic marker in mixed culture studies, or developing the plasmid as vector DNA.

176

AHN ET AL.

ACKNOWLEDGMENTS Funds for this research were provided by NSERC operating grants for D.C-T. and M.E.S. The opportunity for collaborative work was facilitated by a Central Research Fund grant from the University of Alberta. This study forms part of a project funded by Alberta Agriculture, Farming for the Future research project to develop “Innovative Techniques for Preservation of Meats” (Project 87-O159).The authors acknowledge with thanks receipt of the vector plasmid pMG36e from Dr. G. Venema at University of Groningen, The Netherlandv, and cultures provided by Dr. T. R. Klaenhammer and Dr. A. W. Sbrago, North Carolina State University, Raleigh, North Carolina, and by Dr. B. G. Shaw, Institute of Food Research, Langford, Bristol, United Kingdom.

REFERENCES AHN, C., AND STILES,M. E. (199Oa).Antibacterial activity of lactic acid bacteria isolated from vacuum-packaged meats. J. Appl. Bacterial. 69,302-3 10. AHN, C., AND STILES,M. E. (1990b). Plasmid-associated bacteriocin production by a strain of Camobacterium piscicola from meat. Appl. Environ. Microbial. 56, 2503-25 10. ALTON, N. K., AND VAPNEK, D. (1979). Nucleotide sequence analysis of the chloramphenicol resistance transposon Tn9. Nature 282,864-869. CHASSY,B. M. (1987). Prospectsfor the genetic manipulation of lactobacilli. FEMS Microbial. Rev. 46,297312.

CHASSY,B. M., AND FLICKINGER,J. L. (1987). Transformation of Lactobacillus casei by electroporation. FEMS Microbial. Lett. 44, 173- 177. CLEWELL, D. B., YAGI, Y., DUNNY, G. M., AND SCHULTZ,S. K. (1974). Characterization of three plasmid deoxyribonucleic acid molecules in a strain of Streptococcusfaecalis: Identification of a p&mid determining erythromycin resistance. J. Bacterial. 117, 283-289.

LEBLANC, D. J., AND LEE, L. N. (1984). Physical and genetic analyses of streptococcat p&mid PAM@ and cloning of its replication region. J. Bacterioi. 157,445453.

MCKAY, L. L., AND BALDWIN, K. A. (1990). Applications for biotechnology: Present and future improvements in lactic acid bacteria. FEMS Microbial. Rev. 87, 3-14.

OULTRAM, J. D., DAVIES,A., AND YOUNG, M. (1987). Conjugal transfer of a small plasmid from Bacillus subtilis to Clostridium acetobutylicum by cointegrate formation with plasmid pAM@l. FEMS Microbial. Lett. 42, 113-l 19. PEPPER,K., LE BOUGU~NEC,C., DE C&P&DES,G., COLMAR, I., AND HORAUD, T. (1987). Dissemination of a plasm&borne chloramphenicoi resistance gene in streptococcal and enterococcal clinical isolates. In “Streptococcal Genetics” (J. J. Ferretti and R. Curtiss, III, Eds.), pp. 79-82. American Society for Microbiology, Washington, DC. ROCHELLE,P. A., FRY, J. C., DAY, M. J., AND BALE, M. J. (1985). An accurate method for estimating sizes of small and large plasmids and DNA fragments by gel electrophoresis. J. Gen. Microbial. 132, 53-59. SAMBROOK,J., FRITSCH, E. F., AND MANIATIS, T. (1989). “Molecular Cloning: A Laboratory Manual.” Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. SASAKI,Y., TAKETOMO,N., AND SASAKI,T. (1988). Factors affecting transfer frequency of pAM@l from Streptococcusfaecalis to Lactobacillus plantarum. J. Bacterial. 170, 5939-5942. SHAW,B. G., AND HARDING, C. D. (1984). A numerical taxonomic study of lactic acid bacteria from vacuumpacked beef, pork, lamb and bacon. J. Appl. Bacterial. 56,25-40. SHAW,W. V., PACHMAN,L. C., BURLEIGH,B. D., DELL, A., MORRIS,H. R., AND HARTLEY, B. S. (1979). Primary structure of a chloramphenicol acetyl transferase specified by R plasmids. Nature 282, 870-872. SHRAGO,A. W., AND DOBROGOSZ, W. J. (1988). Conjugal transfer of group B streptococcal plasmids and comobilization of E. coli-Streptococcus shuttle plasmids to Lactobacillus plantarum. Appl. Environ. Microbial.

COLLINS,M. D., FARROW,J. A. E., PHILLIPS,B. A., FERusu, S., AND JONES,D. (1987). Classification of Lactobacillus divergens, Lactobacillus piscicola, and wrne catalase-negative, asporogenous, rod-shaped bacteria from poultry in a new genus, Carnobacterium. Int. J. 52,574-576. Syst. Bacterial. 37, 310-3 16. SMITH, A. L., AND BURNS, J. L. (1984). Resistance to GASSON, M. J. (1990). In vivo genetic systems in lactic chloramphenicol and fusidic acid. In “Antimicrobial acid bacteria. FEMS Microbial. Rev. 87,43-60. Drug Resistance” (L. E. Bryan, Ed.), pp. 293-3 15.AcaHAYES,F., CAPLICE,E., MCSWEENEY,A., FITZGERALD, demic Press,New York. G. F., AND DALY, C. (1990). pAMj31-associated mobilization of proteinase plasmids from Lactococcus lac- VAN DE GUCHTE, M., VAN DER VOSSEN,J. M. B. M., KOK, J., AND VENEMA,G. (1989). Construction of a tis subsp. lactis UC317 and L. lactis subsp. cremoris UC205. Appl. Environ. Microbial. 56, 195-201. lactococcal expression vector: Expression of hen egg HORINOUCHI,S., AND WEISBLUM,B. (1982). Nucleotide white lysozyme in Lactococcus lactis subsp. lactis. sequenceand functional map of pC194, a plasmid that Appl. Environ. Microbial. 55,224-228. specifiesinducible chloramphenicol resistance.J. Bac- WEST,C. A., AND WARNER,P. P. (1985). Plasmid proteriol. 150, 815-825. files and transfer of ptasmid-encoded antibiotic resisJEWELL,B., AND COLLINS-THOMPSON,D. L. (1989). tance in Lactobacillus plantarum. Appl. Environ. MiCharacterization of cbloramphenicol resistance in crobial. 50, 1319- 1321. Lactobacillus plantarum caTC2. Curr. Microbial. 19, Communicated by Saleem A. Khan 343-346.