Efficient System for Directed Integration into theLactobacillus acidophilus and Lactobacillus gasseri Chromosomes via Homologous Recombination

In order to increase the host range of this system to include thermophilic lactobacilli, we sought to combine the useful features of the Ori⁺ vectors with a more suitable pWV01-based helper plasmid. Previous experiments in our laboratory have indicated that the wild-type pWV01 replicon, while not completely temperature sensitive, is moderately unstable at temperatures of >42°C in thermophilic lactobacilli, similar to what has been reported for the related plasmid, pE194, in Bacillus subtilis (14). While not suitable for performing traditional temperature-sensitive integration experiments, this feature made it possible to adapt the two-plasmid lactococcal integration strategy for use in Lactobacillus. A helper plasmid, based on pGK12, was created to provide repA in transfor the replication of pORI28-based plasmids (7). This helper plasmid, pTRK669, retains the pWV01 replicon including the origin of replication and repAC genes as well as thecat gene (Table 1). The erythromycin resistance (Em^r) marker was removed in order to make the new plasmid compatible with pORI28, which harbors its own Em^r gene. Plasmids pGK12 and pTRK669 were transformed into L. acidophilus, where their stability was monitored during growth at 37 and 43°C in the absence of antibiotic selection (Fig. 1). The results indicate that both plasmids are stable at 37°C but not at 43°C. An almost 4-log reduction of plasmid-containing cells was observed at 43°C over the course of 30 generations. Similar results were obtained usingL. gasseri cells (data not shown). The ability of pTRK669 to supply RepA for the replication of pORI28 was tested by transferring pORI28 into L. acidophilus and L. gasseri cells with or without pTRK669. Em^r clones could only be recovered from cells which also carried pTRK669, indicating that replication of pORI28 was supported by pTRK669.

In order to verify the functionality of the integration system, plasmid pTRK670 was designed to disrupt the L. acidophilus lacLgene, encoding β-galactosidase. The primers bgalF (5′-GACTGGATCCTGCCGAACGAGCCATGTATG-3′) and bgalR (5′-GACTGAATTCCCGGCATAAGATTCGTTTCC-3′), based on the previously reported lacL gene of L. acidophilus JCM 1229 (GenBank no. AB004867 ), were used to amplify a 945-bp internal region of lacL from L. acidophilus NCFM. This fragment was then cloned via theBamHI/EcoRI sites (underlined) into pORI28, an Ori⁺ RepA⁻ integration plasmid (4). DNA sequencing confirmed that the sequence of the NCFM fragment was 100% identical to the JCM1229 lacLsequence. Plasmid pTRK670 was then introduced by electroporation intoL. acidophilus(pTRK669). One Em^r, chloramphenicol-resistant (Cm^r) transformant, carrying both plasmids, was propagated overnight once at 37°C in MRS broth plus 5 μg of erythromycin and 5 μg of chloramphenicol per ml. The culture was then transferred three times at 43°C (1% inoculum) in MRS broth plus 5 μg of erythromycin/ml (ca. 30 generations). After this enrichment, the culture was plated with a Whitley Automatic Spiral Plater (Don Whitley Scientific Limited, West Yorkshire, England) on MRS–X-Gal (5-bromo-4-chloro-3-indolyl-β-d- galactopyranoside)–galactose plates with either erythromycin, chloramphenicol, or no antibiotic, and incubated for 48 h at 37°C. The results of the plating indicated that the final culture contained ∼10⁷Em^r cells per ml and <10⁻⁷ Cm^r cells per ml. All of the recovered colonies were white, indicating the loss of β-galactosidase activity. Conversely, a control culture, which had been enriched at 37°C instead of at 43°C, still contained a large percentage of LacL⁺, Cm^r cells. One Em^r Cm^s clone, designated NCK1398 was chosen for further confirmation. To confirm the loss of β-galactosidase activity, ONPG (o-nitrophenyl-β-d-galactopyranoside) assays (11) were performed on log-phase cultures ofL. acidophilus NCFM and NCK1398. In order to induce β-galactosidase production, cultures were grown with galactose as the sole carbon source. Upon analysis, no measurable β-galactosidase activity could be detected from NCK1398 compared to 2,652 ± 167 U from NCFM. In order to demonstrate that the disruption of β-galactosidase was the result of integration of pTRK670 in thelacL gene, Southern hybridizations were performed using the 945-bp lacL fragment as a probe (Fig.2). The lacL probe hybridized to an EcoRI fragment of ca. 11.5 kb in L. acidophilus NCFM. In NCK1398, this band disappeared and was replaced by junction fragments of ca. 5.3 and 8.8 kb due to the presence of a single EcoRI site present in the pTRK670 plasmid sequence. In addition to the two junction fragments, another band was observed that corresponded to the amplified copies of the 2.6-kb pTRK670 sequence. Further confirmation was obtained by hybridization to HindIII digestions because noHindIII sites were present within the pTRK670 sequence. The lacL probe hybridized to an ca. 3.2-kb fragment inL. acidophilus NCFM. In NCK1398, this band disappeared and was replaced with bands of ca. 5.8, 8.4, and 11.0 kb, corresponding to insertions of one, two, and three plasmid copies, respectively. These results were identical in each of the clones tested, indicating that each individual clone gave rise to a mixed population of cells harboring different numbers of plasmid copies. This variation in plasmid copy number has been observed in other lactic acid bacteria (9, 10). Leenhouts et al. (8) reportedL. lactis clones that had generated one, two, or three chromosomal plasmid copies after the insertion of a pORI-type plasmid. This amplification of plasmid sequences has generally been attributed to recombinatory activity between the flanking DNA regions of homology that result from campbell-like integration and is influenced by a number of factors, including the nature and location of the insertion event.

The complete sequence of the L. gasseri ADH gusAgene, encoding β-glucuronidase, has been reported recently (13). A 777-bp internal region of the gusA gene was amplified by PCR using the primers GUS1F (5′-GACTTCTAGAACAGTTGACGAATACACAGAT-3′) and GUS1R (5′-GTGAATTCAGGCGATGAGAAGAAGATAATG-3′). This fragment was then cloned into pORI28 as anXbaI/EcoRI fragment using the sites added to the 5′ ends of the primers (underlined). The resulting plasmid, pTRK685, was then transferred by electroporation into L. gasseriADH(pTRK669), and the integration strategy was carried out as described above. The results were similar to those obtained for lacL; only Em^r Cm^s colonies were obtained, and disruption of the gusA gene was evidenced by the formation of only white colonies. One clone, designated NCK1423, was subjected to further testing. Confirmation of the GusA⁻ phenotype was obtained by performing β-glucuronidase assays (13) on stationary-phase L. gasseri ADH and NCK1423 cells. No measurable β-glucuronidase activity could be detected with NCK1423 compared to 97.2 ± 11 U in ADH. Integration of plasmid DNA into the gusA gene was confirmed by Southern hybridization (Fig.3). The results of the Southern hybridization revealed that integration had occurred within thegusA gene. However, unlike the previous integration intoL. acidophilus lacL, the number of plasmid DNA sequences integrated into the gusA gene appeared to be fixed at two copies.

To determine the stability of the integrated plasmids, NCK1398 and NCK1423 were propagated in MRS broth in the absence of antibiotics for ca. 50 generations and plated on media with the appropriate chromogenic substrate, i.e., either X-Gal or X-Gluc (5-bromo-4-chloro-3-indolyl-β-d-glucuronic acid). Loss of the integrated plasmid was assessed by restoration of the LacL⁺ or GusA⁺phenotypes, as indicated by the formation of blue colonies. After 50 generations, the percentages of revertants were 5.36% forlacL and 0.72% for gusA, corresponding to frequencies of 0.11% and 0.01% per generation, respectively. These results indicated that, while plasmid DNA is stably integrated in both strains, the integrated DNA is ca. 10-fold more stable in thegusA gene of L. gasseri NCK1423.

The advantages of the method reported here over other integration strategies for lactobacilli are that (i) it is not dependent on transformation efficiency, (ii) it allows growth of mutant strains at preferred growth temperatures, (iii) it allows efficient recovery of stable integrants, and (iv) it is likely to function across a broad range of species. While this study has focused on the use of a single integration event to insertionally inactivate functional genes, a number of other applications are envisioned. For example, this strategy can be easily modified to construct and integrate expression cassettes into preferred sites in the chromosome. However, the primary disadvantage of the current design is that plasmid DNA and antibiotic markers remain integrated in the chromosome, leading to potential complications in food-grade applications. Recently, the pORI system has been used to create unlabeled gene replacements in L. lactis and B. subtilis (7) and to create a food-grade multiple-copy integration system (6) that has been used to overexpress proteins in Lactococcus (5). It is expected that the protocol described here could be easily modified to facilitate similar strategies in lactobacilli.

This research was supported by The Southeast Dairy Foods Research Center; Dairy Management, Inc.; Rhodia, Inc.; and the North Carolina Dairy Foundation. W. Michael Russell was supported in part by a U.S. Department of Education GAANN Biotechnology Fellowship.

We thank Jan Kok, University of Groningen, for kindly providing pORI28 and E. coli EC1000. We also thank Evelyn Durmaz, Eric Altermann, Olivia McAuliffe, and Michael Callanan for their help and for critical review of the manuscript.