Disruption of the Borrelia burgdorferi gac Gene, Encoding the Naturally Synthesized GyrA C-Terminal Domain

ABSTRACT The C-terminal domain of the A subunit of DNA gyrase, which we term Gac, is naturally synthesized in Borrelia burgdorferi as an abundant DNA-binding protein. Full-length GyrA, which includes the C-terminal domain, is also synthesized by the spirochete and functions as a subunit of DNA gyrase. We have disrupted synthesis of Gac as an independent protein and demonstrated that it is not essential for growth in a coumarin-resistant background. We detected no alterations in DNA maintenance, condensation, or topology in B. burgdorferi lacking this small DNA-binding protein.

which confer high-level resistance to coumermycin A 1 (D. S. Samuels, B. J. Kimmel, D. C. Criswell, C. F. Garon, W. M. Huang, and C. H. Eggers, unpublished data). The amplification product was adenylated and cloned into plasmid PCR 2.1-TOPO (Invitrogen) according to the manufacturer's instructions, creating plasmid pTAKO1 (Fig. 1A). Primer gyrA 1521R/GACKO was a mutagenic primer that introduced silent mutations into the Shine-Dalgarno sequence to decrease ribosome binding, mutated an ATG Met codon (position 499 of GyrA) (which is predicted to be the translational start codon) to a CTT Leu codon to prevent translation of Gac, and mutated an ATG Met codon (position 503 of GyrA) to an ATT Ile codon to prevent translation from initiating downstream (Fig.  1C). The introduced mutations correspond to residues found at the homologous sites in E. coli GyrA (15).
Plasmid pTAKO2 was constructed by PCR amplifying a 900-bp fragment encoding gac from B31-NGR, using primers gyrA 1462F/GACKO and gyrA 2362R. Primer gyrA 1462F/ GACKO was a mutagenic primer that introduced the complementary mutations as described above for primer gyrA 1521R/ GACKO. The amplification product was purified, the 5Ј ends were adenylated, and the product was cloned as described above, creating plasmid pTAKO2 (Fig. 1A). The mutagenic plasmid for disrupting synthesis of Gac (pGACKO) was constructed by ligating the approximately 800-bp SpeI fragment from plasmid pTAKO2 into the SpeI sites in plasmid pTAKO1 (Fig. 1B).
B. burgdorferi B31 was transformed with pGACKO and was plated in solid medium containing 0.5 g of coumermycin A 1 ml Ϫ1 as previously described (11). A successful recombination event resulted from a double crossover or a single crossover and branch migration spanning the mutations conferring coumarin resistance in the 5Ј region of gyrB and the mutations in the gac gene (approximately 4 kb). Cou r transformants were screened for gyrA mutations by PCR analysis with primers gyrA 1492F/GACKOSC and gyrA 2362R as previously described (10). Primer gyrA 1492F/GACKOSC is complementary to the mutated sequence, but only 15 of the 18 nucleotides in the primer are complementary to the wild-type sequence, including a noncomplementary nucleotide at the 3Ј end. These two primers amplify a ϳ900-bp fragment from the mutated sequence, but they fail to amplify the same product from wildtype gyrA with a 56°C annealing temperature (data not shown).
Three of 400 Cou r colonies contained the introduced mutations. The frequency of recombination was comparable to or better than the directed insertion of gyrB r into cp26 (1,10,16,18). An increase in recombination efficiency may be due to the need for only a single crossover, followed by branch migration, rather than a double crossover event. However, the efficiency remains low, possibly because of the large size of the fragment required to recombine. The presence of mutations in two clones, CKO-1 and CKO-2, was confirmed by DNA sequencing (Fig. 1C). Clone CKO-1 contained all of the introduced mutations. Clone CKO-2 contained the mutations that changed the two Met residues, but lacked the mutations in the Shine-Dalgarno sequence.
Western analysis confirmed that Gac synthesis was successfully disrupted in both clones ( Fig. 2A and B). Whole-cell lysates from B. burgdorferi strains B31 (wild type), B31-NGR, CKO-1, and CKO-2 were analyzed by using an anti-Gac antiserum (5). The antiserum recognized the full-length 91-kDa GyrA protein in all the whole-cell lysates, but it failed to detect Gac in strains CKO-1 and CKO-2. B. burgdorferi B31-NGR is isogenic to CKO-1 and CKO-2 with respect to the mutations in gyrB. The successful disruption of Gac synthesis in CKO-2 indicates that the mutations in the Shine-Dalgarno sequence were not critical for preventing translation.
The Gac-deficient strains grow in both liquid culture and solid medium. They exhibit a growth rate and cellular morphology similar to B31 and B31-NGR ( Fig. 3 and data not shown). We have hypothesized that Gac may function in the replication, compaction, and maintenance of linear replicons, or in other aspects of linear DNA metabolism, in B. burgdorferi (5). However, ethidium bromide staining of DNA extracts fractionated by agarose gel electrophoresis indicate that linear and circular DNA molecules are maintained in CKO-1 and CKO-2 (Fig. 4). Gross DNA structure and morphology in strains  CKO-1 and CKO-2 is also unchanged, as examined by 4Ј,6diamidino-2-phenylindole (DAPI) staining (data not shown).
No differences in gene expression of the GyrA C-terminal domain-deficient strains compared to strain B31-NGR were detected by Coomassie brilliant blue-stained sodium dodecyl sulfate-polyacrylamide gels of whole-cell lysates ( Fig. 2A).
The absence of a distinguishable phenotype in the gac mutants suggested that the mutation may be suppressed by an increase in expression of genes encoding other small DNAbinding proteins. We examined strains CKO-1 and CKO-2 for an increase in expression of the hbb gene, which encodes Hbb, the B. burgdorferi HU/IHF homolog (17). Whole-cell lysates were examined by Western analysis using a polyclonal antiserum raised against a synthetic peptide (KRKGRLNARN PQTGEA) designed with predicted antigenicity (MacVector; Oxford Molecular). Hbb was present at low levels in all strains examined, and its synthesis was unchanged in the gac mutants (Fig. 2C), supporting previous biochemical evidence that Hbb may play a limited role in B. burgdorferi DNA metabolism (5,17).
The ability to disrupt synthesis of Gac demonstrates that the protein is not essential. Based on the HU-like activity of the protein (5), this finding is perhaps not surprising. E. coli strains lacking both subunits of HU are viable (7). Major abnormalities of these E. coli strains include slow doubling times, poor plasmid maintenance, and the inability to support bacteriophage Mu growth (4,8,19). Some of the observed phenotypes of HU-deficient E. coli are unstable and are compensated by the accumulation of suppressor mutations (4,6). In E. coli, mutations that map to gyrB (and confer resistance to the coumarin antibiotic novobiocin) suppress an HU deficiency (6). The data presented here indicate that B. burgdorferi strains lacking Gac do not exhibit any phenotypic differences in DNA metabolism compared to isogenic strains. However, based on the HU-like activity of this protein, the phenotype may be suppressed by the presence of gyrB r , which confers resistance to the coumarin antibiotic coumermycin A 1 .
We thank Kit Tilly, Karl Drlica, and Tasha Knight for thoughtful and critical reading of the manuscript; Stuart Hill, Joe Hinnebusch, and Tim Gsell for useful discussions; and the late Joan Strange for DNA sequencing and peptide synthesis.
Work in our laboratory is supported by grants from the National Science Foundation (MCB-9722408), the National Institutes of Health (AI41559 and AI39695), MONTS (NSF EPSCoR), the Arthritis Foun- FIG. 3. Effect of Gac on growth of B. burgdorferi. Growth was assayed essentially as described previously (12). Cultures of B. burgdorferi strains B31 (F), B31-NGR (E), CKO-1 (‚), and CKO-2 () were inoculated with the same number of cells into BSK-H medium and were grown for 6 days. One milliliter of each culture was removed each day, cells were pelleted and resuspended in 1 ml of Dulbecco's phosphate-buffered saline, and the optical density at 600 nm was determined. Plotted are the means of five independent experiments (four for days 4 and 6); error bars represent standard deviations.