|Home | About | Journals | Submit | Contact Us | Français|
We have demonstrated that rpsL, encoding the S12 protein of the small ribosomal subunit, can be used as a counterselectable marker in Borrelia burgdorferi, the causative agent of Lyme disease. Mutations in rpsL confer streptomycin resistance. Streptomycin susceptibility is dominant in an rpsL merodiploid, and streptomycin selects for the loss of wild-type rpsL carried in trans. This is the first description of a counterselectable marker in B. burgdorferi.
The approach for genetically manipulating Borrelia burgdorferi, a spirochete that causes Lyme borreliosis (2, 6, 23), has slowly matured over the past 15 years (17, 18). A number of selectable markers conferring antibiotic resistance have been developed and used to disrupt genes and maintain plasmids, including gyrB (Cour) (21), aphI (Kanr) (5), ermC (Ermr) (22), aadA (Spt/Strr) (9), and aacC1 (Genr) (8), which have enabled effective, if not efficient, molecular genetic tools. However, to date, no counterselectable marker has been available to select for the loss of a particular DNA sequence (16). We now demonstrate that susceptibility to streptomycin is dominant in a merodiploid carrying a wild-type rpsL allele and a streptomycin-resistant rpsL allele and that rpsL can function as a counterselectable marker in B. burgdorferi. This now provides a genetic tool, previously unavailable, for studying the biology of B. burgdorferi and the pathogenesis of Lyme disease.
In Escherichia coli, the rpsL gene encodes the S12 ribosomal protein of the 30S subunit. Streptomycin exerts its antimicrobial activity by binding to 16S rRNA, near the binding interface of S12, to inhibit protein synthesis by increasing translational errors through the recruitment of incorrect tRNAs (11, 12). Mutations in rpsL confer resistance in numerous bacteria, including E. coli (4, 10, 14) and the spirochete Leptospira biflexa (15). Streptomycin inhibits the growth of streptomycin-resistant rpsL mutants when wild-type (streptomycin-sensitive) rpsL is expressed in trans, indicating that the antibiotic susceptibility phenotype is dominant (13). Furthermore, growing merodiploid rpsL strains in the presence of streptomycin can select for the loss of wild-type rpsL, demonstrating the utility of this gene as a counterselectable marker (16).
We recently isolated B. burgdorferi mutants that were 10-fold more resistant to streptomycin than the parental strain, B31-A (7). One isolate contained a mutation in rpsL encoding the single amino acid substitution K88E in S12, the same residue that was found to be mutated in streptomycin-resistant L. biflexa (15) and E. coli (10). We constructed an rpsL merodiploid strain to test the feasibility of streptomycin susceptibility as a counterselectable marker in B. burgdorferi.
rpsL was amplified from B. burgdorferi strain B31-A3 genomic DNA by PCR using the primers rpsLU133F+SgrAI (5′-CGCCGGTGACTGGAACTGGTATGGGTC-3′) and rpsL375R+SgrAI (5′-CACCGGCGTTAAGCTTTAGGCTTTTTTGTTC-3′). The PCR product was cloned into pCR2.1-TOPO (Invitrogen) to yield pTArpsL and confirmed by direct sequencing. pTArpsL was digested with XhoI and SacI and ligated into the Borrelia shuttle vector pBSV2 (24), which had been digested with SalI and SacI, to yield pBSrpsL (Fig. (Fig.1A).1A). Low-passage B31-A was transformed with pBSrpsL, and the streptomycin-resistant strain DCSmR4 (7) was transformed with either pBSrpsL or the empty vector pBSV2, essentially as described previously (19). Cultures of transformants were diluted with Barbour-Stoenner-Kelly (BSK) II medium (1) containing kanamycin (200 μg/ml) in 96-well plates (25) to select for those containing pBSV2 or pBSrpsL harboring the kanamycin resistance gene aphI. Positive wells were screened for the presence of the plasmids by PCR using the primers PflgB5+MfeI (5′-CAATTGTACCCGAGCTTCAAGGAA-3′) and KanR 488R (5′-TCACTCGCATCAACCAAACC-3′) to detect aphI. Clones were chosen from 96-well plates that had fewer than 10 positive wells because the probability that a well was inoculated with a single cell is greater than 0.94 (J. M. Graham and D. S. Samuels, unpublished data).
We first tested the hypothesis that streptomycin susceptibility is dominant in an rpsL merodiploid. Each strain was initially grown in liquid BSK II medium containing only kanamycin to maintain pBSV2 or pBSrpsL. Borrelia spirochetes were then counted using a Petroff-Hausser counting chamber, and 1,000 spirochetes were plated in semisolid BSK medium (19) containing (i) no antibiotics, (ii) kanamycin (200 μg/ml), (iii) streptomycin (50 μg/ml), or (iv) kanamycin plus streptomycin. Plates were incubated at 37°C and 5% CO2 for 2 weeks before enumeration. The merodiploid strain DCSmR/pBSrpsL did not form colonies in the presence of both kanamycin and streptomycin, suggesting that streptomycin susceptibility is dominant (Table (Table1).1). A few DCSmR/pBSrpsL colonies formed in the presence of streptomycin alone (Table (Table1).1). Similar results were confirmed using a second DCSmR/pBSrpsL clone (data not shown). The frequency of loss of pBSrpsL (0.04%) is much lower than we have observed with another counterselectable marker system using fluoroquinolones and parC (~3%) (18); the higher rate of plasmid loss in the parC mutants may be due to a partitioning defect. However, the pBSV2 backbone vector is stable, with 100% retention after ~90 generations without selection (24).
Streptomycin should select for streptomycin-resistant Borrelia spirochetes that have lost pBSrpsL, which confers susceptibility, and that carry only the genomic rpsL allele conferring streptomycin resistance (Fig. (Fig.1B).1B). To confirm that kanamycin-resistant colonies retained pBSrpsL and streptomycin-resistant colonies lost pBSrpsL, DNA from individual colonies was analyzed by xenodiagnosis and PCR. In each experiment five randomly selected colonies from each plate were grown in liquid BSK II medium (containing either kanamycin or streptomycin) and total genomic DNA was isolated as previously described (20). For xenodiagnosis, chemically competent E. coli DH5α cells were transformed with genomic DNA and plated on lysogeny broth (3) plates containing kanamycin to select for transformants carrying pBSrpsL. DNA from kanamycin-resistant DCSmR/pBSrpsL Borrelia yielded kanamycin-resistant E. coli colonies in all cases (15/15), suggesting that pBSrpsL was present, while no colonies were observed using DNA from streptomycin-resistant DCSmR/pBSrpsL (0/15).
To confirm these results, total genomic DNA was used to screen for the presence of pBSrpsL by PCR analysis using primers PflgB5+MfeI and KanR 488R. All 15 kanamycin-resistant DCSmR/pBSrpsL clones were positive for pBSrpsL by PCR, while all 15 streptomycin-resistant DCSmR/pBSrpsL clones were negative. Although the absence of a genetic element cannot be proven, these data, taken together, suggest that streptomycin can be used to select for the loss of the wild-type rpsL allele, conferring streptomycin susceptibility, in a merodiploid (DCSmR/pBSrpsL). Thus, rpsL can function as a counterselectable marker and should be a useful molecular tool for genetic experiments that require the loss of a DNA sequence. We are currently isolating rpsL mutations in low-passage infectious strains, and we are attempting to apply the counterselectable marker system for use in the animal model of Lyme borreliosis.
We thank Steve Lodmell and Meghan Lybecker for thoughtful and critical reading of the manuscript; Mathieu Picardeau (who suggested developing rpsL as a counterselectable marker), Jon Graham (who performed the statistical analysis of clones), and Meghan Lybecker for valuable discussions; Patty McIntire (Murdock DNA Sequencing Facility) for DNA sequencing; Philip Stewart for providing pBSV2; and Laura Hall for excellent technical assistance.
J.M.D. is a student at Hellgate High School.
This work was supported by a grant from the National Institutes of Health (AI051486) to D.S.S.
Published ahead of print on 4 December 2009.