|Home | About | Journals | Submit | Contact Us | Français|
Our understanding of leptospiral pathogenesis, which remains poorly understood, depends on reliable genetic tools for functional analysis of genes in pathogenic strains. In this study, we report the first demonstration of conjugation between Escherichia coli and Leptospira spp. by using RP4 derivative conjugative plasmids. The DNA transfer described here was due to authentic conjugation, as shown by the requirement for cell-to-cell contact and the resistance of DNA transfers to the addition of DNase I. Transposition via conjugation of a plasmid delivering Himar1 yielded frequencies ranging from 1 × 10−6 to 8.5 × 10−8 transconjugants/recipient cell in the saprophyte L. biflexa and the pathogen L. interrogans, respectively. Analysis of mutants indicated that transposition occurs randomly, and at single sites in the genome of these strains, allowing the utilization of this system to generate libraries of transposon mutants.
To date, there are no data for natural competence of or conjugation for Leptospira spp. The isolation of host-specific bacteriophages in the saprophyte Leptospira biflexa led to the discovery of the temperate bacteriophage LE1 (18). Analysis of the LE1 host range revealed that this bacteriophage is able to infect a few saprophytic strains but not those belonging to pathogens. Bacteriophage LE1, the only leptophage sequenced so far, presents a double-stranded DNA genome of approximately 74 kb (1). It has been demonstrated that LE1 replicates as a circular plasmid in L. biflexa (17). This fact, together with the identification of the LE1 replication origin, made this bacteriophage a valuable tool for development of replicative vectors. In Leptospira, the successful introduction of DNA has proved possible only following electroporation. The saprophyte L. biflexa can be transformed at a high frequency with plasmids based on the LE1 replication origin, with kanamycin or spectinomycin resistance used as a selectable marker (1, 17). However, LE1-based plasmids do not replicate in pathogenic strains. Gene transfer in pathogens was demonstrated only by random insertion of the Himar1 transposon, and it occurred at very low frequencies (2). Over the years, there has been no report of successful targeted gene inactivation in pathogenic strains. The use of genetic approaches to study the biology and pathogenesis of Leptospira remains therefore significantly limited by inefficient techniques for introducing DNA.
In the early 1970s, several self-transmissible resistance plasmids were isolated from antibiotic-resistant bacteria (7). These plasmids are members of E. coli incompatibility group P (IncP1α) (14) and are able to transfer in a wide variety of bacterial species. Derivatives of plasmids that contain the oriT of RK2/RP4 were transferred by conjugative mobilization from E. coli to different eubacteria, including alphaproteobacteria (3), gammaproteobacteria (3, 10), Firmicutes (13), and Actinobacteria (12), as well as yeasts (6) and mammalian cells (20). Since bacterial conjugation between different genera is possible, we wanted to determine whether E. coli can conjugate to leptospira, which belong to the class Spirochaetes, by using the RP4/RK2 plasmid system.
A vector called pCjSpLe94 for conjugative transfer from E. coli β2163 (3) to the saprophyte Leptospira biflexa serovar Patoc strain Patoc 1 (Institut Pasteur) was constructed, and it included (i) a conditional R6K origin of replication, which is recognized by a specific replication initiator protein (π) encoded by E. coli β2163 (5), (ii) the oriT of RP4, which can be efficiently transferred using the chromosomally encoded RP4 conjugation machinery, (iii) the replication origin from the leptophage LE1, which contains the rep and partition genes (17), (iv) a spectinomycin resistance cassette, and (v) a kanamycin resistance cassette (Fig. (Fig.11).
Cells of the donor strain E. coli β2163, which is auxotrophic for diaminopimelate (DAP), harboring the plasmid pCjSpLe94, and the recipient L. biflexa strain were grown in EMJH (4, 8) liquid medium supplemented with 0.3 mM DAP at 30°C to an optical density at 420 nm of 0.3 (corresponding to 2 × 108 E. coli cells/ml and 4 × 108 L. biflexa cells/ml). Mating experiments were performed with a donor/recipient ratio of 1/10; 0.5 ml and 4.5 ml of the donor and recipient strains, respectively, were mixed and concentrated on a cellulose-acetate filter (0.1-μm pore size, 25-mm diameter; Millipore) which was incubated for 20 h at 30°C on EMJH solid medium supplemented with DAP. Cells on the filter were resuspended by shaking in 5 ml EMJH liquid medium and plated at appropriate dilutions on EMJH and EMJH supplemented with 25 μg/ml spectinomycin. The viable count of the donor strain was determined by spreading the cells on LB agar plates supplemented with DAP. For L. biflexa, plates were incubated for 10 days, after which colonies were counted. The frequency of DNA transfer was assessed by determining the number of resistant transformants per parent or per recipient. Since the growth of the E. coli β2163 dap-negative mutant is dependent on exogenously added DAP, the absence of DAP in the EMJH medium provides an efficient counterselection against the donor strain. We obtained a frequency of 4.2 × 10−4 L. biflexa spectinomycin-resistant (Spcr) colonies per donor cell (or 3.2 × 10−5 L. biflexa Spcr colonies per recipient cell). The authenticity of transconjugants was established by preparing plasmid DNA preparations from Spcr cultures, using them to transform E. coli, and then isolating plasmid DNA from the resultant transformants and confirming their restriction profiles (data not shown). No resistant transconjugants were obtained when E. coli cells carrying pSW29TSp (Fig. (Fig.1),1), which is not able to replicate in L. biflexa, were used in mating experiments (Table (Table1).1). To examine genetic exchange in broth, E. coli and L. biflexa cells were incubated together overnight. No L. biflexa-resistant transconjugants were encountered. The development of transconjugants during incubation with E. coli could also result from cell lysis, leading to the release of free DNA. However, the addition of DNase I to cells did not prevent the development of transconjugants. In addition, in the presence of pCjSpLe94 alone, no DNA transfer was detected in L. biflexa (Table (Table1).1). After RP4-mediated filter mating, cells were resuspended in 10 mM Tris-HCl and samples were processed for electron microscopy as previously described (15). We observed the pili produced by the E. coli RP4 conjugation machinery and, more interestingly, intimate junctions between the L. biflexa and E. coli membranes (data not shown). These attachment zones between cell membranes resemble the conjugational junctions described by Samuels et al. (19). Our experiments therefore suggest that the DNA transfer from E. coli to L. biflexa occurs by a conjugative mechanism requiring close cell-to-cell contact between the donor and recipient cells. Conditions for optimal conjugal plasmid transfer were established by using E. coli β2163 harboring the replicative vector pCjSpLe94 as the donor and L. biflexa as a recipient. Various parameters affecting conjugation, such as the time of filter mating and the proportion of cells of the donor and recipient strains in filter mating assays, were investigated (data not shown). For optimal conjugation, filter mating assays were carried out overnight and employed a donor-to-recipient ratio of 1/10.
Because attempts to perform transformation in pathogenic Leptospira with the L. biflexa-E. coli shuttle vector were unsuccessful (17), we have designed a Himar1-based transposon (2) that can be used to evaluate DNA transfer from E. coli to the pathogen L. interrogans serovar Lai strain Lai 56601. We delivered the Himar1 transposon on a suicide vector, pCjTKS1, that contains (i) the Himar1 transposon, which contains a kanamycin resistance cassette and the R6K origin of replication (5), (ii) the C9 hyperactive transposase (9), (iii) the oriT of RP4, and (iv) a spectinomycin resistance cassette (Fig. (Fig.1).1). The protocol used for the Himar1 delivery was identical to the protocol used for the transfer of the replicative vector to L. biflexa (see above).
We first tested the Himar1 transfer from E. coli to L. biflexa. As shown in Table Table2,2, L. biflexa-resistant colonies appeared at a mean frequency of 1 × 10−6 transconjugants per donor cell, 10-fold lower than that obtained with the replicative vector pCjSpLe94 (Table (Table2).2). The transposition frequency (number of colonies obtained with the suicide delivery vector/number of colonies transformed by the replicative vector) in cells that receive the delivery suicide vector is therefore highly efficient. By using this system, we are able to obtain large libraries of independent kanamycin-resistant mutants.
For L. interrogans, all plates were incubated for one month at 30°C, after which colonies were counted. The results showed that mating experiments with E. coli and L. interrogans with pCjTKS1 gave rise to resistant colonies at a frequency of approximately 8.5 × 10−8 per donor. As previously shown by electroporation with L. interrogans (2), the transposon insertions occurred at random, as determined by ligation-mediated PCR (16) of forty randomly selected mutants.
In this study, we used RP4-based broad-host-range plasmids to transfer DNA from a donor strain of E. coli to recipient strains belonging to both saprophytic and pathogenic Leptospira species. This is the first evidence, to our knowledge, of intergeneric conjugal transfer of plasmid DNA from E. coli to Leptospira spp. and, more generally, to spirochetes, thus extending the known host range of RP4-based plasmids.
To date, transformation in Leptospira spp. has been performed by electroporation (2, 17). In previous studies, we showed that electroporation of the replicative plasmid pGKlep4 gave an average of 5 × 104 to 10 × 104 L. biflexa transformants/μg of DNA (17). Transformation efficiencies of electroporation of suicide plasmids delivering the transposon Himar1 ranged up to 2 × 104 transformants/μg of DNA and 50 transformants/μg of DNA in L. biflexa and L. interrogans, respectively (11). In this study, we also obtained a high-frequency transfer with the RP4-based plasmid delivering Himar1 into L. biflexa (Table (Table2).2). Although the rate of conjugal transfer of Himar1 in L. interrogans serovar Lai is relatively low, we obtained hundreds of mutants from the recovery cultures after mating experiments, in comparison to a mean of 50 random mutants obtained by electroporation. In addition, results appeared more reproducible by conjugation (8.5 × 10−8 ± 2.4 × 10−8 transconjugants per recipient cell) than by electroporation (5.1 × 10−9 ± 4.1 × 10−9 electrotransformants per recipient cell). Inconsistent transformation efficiencies obtained by electroporation can be the result of impure reagents (contamination of plasmid DNA with salts or proteins), subtle variations in the technique, or the physiological state of the cells. In addition, the survival of leptospires was affected when they were transformed by electroporation. For the one saprophytic strain and three pathogenic strains tested, similar survival rates of 8 to 10% were obtained with field strength of 9 kV/cm, using 0.2-cm cuvettes.
In conclusion, conjugative plasmid can be mobilized from E. coli to L. biflexa or L. interrogans through the RP4-mediated conjugal transfer functions encoded by the chromosome of the E. coli donor strain. This approach should significantly enhance the ability of researchers to manipulate genes in pathogenic strains.
I am thankful to F. Le Roux and D. Mazel for the gift of plasmid pSW29T and their support and encouragement. I also thank C. Bourguignon for assistance. I am grateful to Evelyne Couture-Tosi for the electron microscopy analysis.
This work was supported by the French Ministry of Research/ANR Jeunes Chercheurs (grant no. 05-JCJC-0105-01).
Published ahead of print on 9 November 2007.