Function of the Borrelia burgdorferi FtsH Homolog Is Essential for Viability both In Vitro and In Vivo and Independent of HflK/C

ABSTRACT In many bacteria, the FtsH protease and its modulators, HflK and HflC, form a large protein complex that contributes to both membrane protein quality control and regulation of the cellular response to environmental stress. Both activities are crucial to the Lyme disease pathogen Borrelia burgdorferi, which depends on membrane functions, such as motility, protein transport, and cell signaling, to respond to rapid changes in its environment. Using an inducible system, we demonstrate that FtsH production is essential for both mouse and tick infectivity and for in vitro growth of B. burgdorferi. FtsH depletion in B. burgdorferi cells resulted in membrane deformation and cell death. Overproduction of the protease did not have any detectable adverse effects on B. burgdorferi growth in vitro, suggesting that excess FtsH does not proteolytically overwhelm its substrates. In contrast, we did not observe any phenotype for cells lacking the protease modulators HflK and HflC (ΔHflK/C), although we examined morphology, growth rate, growth under stress conditions, and the complete mouse-tick infectious cycle. Our results demonstrate that FtsH provides an essential function in the life cycle of the obligate pathogen B. burgdorferi but that HflK and HflC do not detectably affect FtsH function.

lambda phage regulator contributing to the decision between a lysogenic and a lytic life cycle (11). LpxC, an enzyme necessary for lipid A biosynthesis, is proteolytically controlled by FtsH. Depletion of cellular FtsH levels creates an imbalance between the synthesis of lipopolysaccharides and phospholipids and results in cell death, therefore making FtsH essential to E. coli (4). Although the full range of FtsH substrates remains unknown, several membrane protein targets of FtsH have been identified and include SecY, a subunit of the protein translocation machinery (12), KdtA, a transferase involved in oligosaccharide biosynthesis (13), and YccA, a membrane protein of unknown function (14).
Although the regulation and substrate range of FtsH have not been fully elucidated, even in E. coli, two modulators of the protease have been identified. The cytoplasmic membrane proteins HflK and HflC (here, HflK/C) form a high-molecular-weight complex with the protease, referred to as the FtsH holoenzyme (12,15). Several reports indicate that HflK/C acts as a negative regulator of membrane substrates, selectively allowing specific membrane proteins access to the active site chamber of FtsH (11,12,14). Although these results are complex and not completely understood, evidence suggests that entry of cytoplasmic substrates into the FtsH protease may occur through a separate process (14). Unlike FtsH, HflK and HflC are not essential to E. coli and can be inactivated, with the main observable phenotype being a high frequency of lysogenization by phage (12).
All three proteins FtsH, HflK, and HflC, have chromosomally encoded homologs in Borrelia burgdorferi, the Lyme disease spirochete (16). B. burgdorferi is an obligate zoonotic parasite with limited metabolic capabilities which cycles between Ixodes ticks and vertebrate hosts. To survive in such disparate host environments, the spirochete differentially synthesizes various membrane proteins for sensing external stimuli, nutrient acquisition, immune evasion, and as adhesins. In addition, B. burgdorferi produces and maintains periplasmic flagella for motility, which are required for tick transmission and to disseminate and persist in the vertebrate host (17,18). All of these functions are either embedded in or exported across the cytoplasmic membrane; therefore, membrane quality control is likely to be crucial to B. burgdorferi survival. Although the mechanism(s) by which B. burgdorferi maintains the proper functioning of cytoplasmic membrane proteins has not been examined, a library of transposon mutants identified insertions in the hflK and hflC homologs (bb0203 and bb0204, respectively) that resulted in reduced murine infectivity (19). Interestingly, only one transposon insertion at the terminus of the ftsH homolog (bb0789) was isolated, and the authors postulated that the dearth of insertions in and around the bb0789 locus indicated that the encoded FtsH homolog confers an essential function to B. burgdorferi. Recently, Drecktrah and colleagues observed that ftsH expression in B. burgdorferi was controlled by B. burgdorderi Rel (Rel Bbu ) in response to nutrient starvation, whereas hflK and hflC transcript levels were not significantly affected (20).
To investigate the contributions of the FtsH, HflK, and HflC homologs to the life cycle of B. burgdorferi, we constructed mutants of all three genes and characterized the strains both in vitro and in vivo. Surprisingly, our results did not identify any phenotype for the ⌬HflK/C double mutant, although we examined morphology, growth rate, growth under stress conditions, and the complete mouse-tick infectious cycle. In contrast, however, FtsH was shown to be essential to B. burgdorferi survival, both in vitro and in vivo. Using an inducible system to control ftsH expression, B. burgdorferi cells depleted of FtsH exhibited arrested cell growth and morphological defects during in vitro cultivation. Mouse-tick infection studies with the mutant strain demonstrated that FtsH is also required for survival in both the murine host and the tick vector. These studies indicate that the B. burgdorferi FtsH homolog provides an essential function both in vitro and in vivo, independently of the homologs of the HflK/C modulators. In vitro phenotypes of the HflK/C mutant are similar to that of the WT. We hypothesized that FtsH function may be essential to B. burgdorferi, based on the lack of transposon insertions observed in a comprehensive signature-tagged mutagenesis library (19). Therefore, we initially focused on the modulators of FtsH substrate specificity, HflK and HflC, homologs of which are widespread among spirochetes. Using an allelic exchange vector, we generated a deletion mutant that lacked both hflK and hflC (Fig. 1A). The ⌬hflK/C mutant was complemented (hflK/Ccomp), and the genetic structures of both strains were confirmed by Southern blot and immunoblot analyses ( Fig. 1B) (data not shown). Surprisingly, and despite extensive in vitro characterization, no distinguishing phenotypes were observed for the ⌬hflK/C strain compared to the wild-type (WT) or complemented strains. The deletion mutant's growth rate, protein profile (determined by Coomassie-stained SDS-PAGE, and immunoblots reacted with infected mouse sera), plating efficiency, length of time for colony formation on solid medium, and morphology (as assessed by electron microscopy) were similar to those of the WT (Fig. 1C) (data not shown).

BB0789
In a further attempt to characterize the contribution of HflK/C to B. burgdorferi physiology, we assessed the response of the mutant, complemented, and WT strains to the environmental stresses of heat shock and osmotic shock. Previous studies demonstrated that B. burgdorferi responds to heat shock by differential production of specific proteins when the cultivation temperature is shifted from 35°C to 39 or 40°C (22)(23)(24). However, when we subjected the WT strain to either a 1-or 4-h incubation at 40°C, HflK levels remained unchanged compared to the protein levels from cells cultured at 35°C (Fig. 1D), nor did we see any difference in the relative ability of the mutant to survive a 1-h temperature shift, as assessed by plating for viable spirochetes (data not shown). Finally, we examined the response of the ⌬hflK/C mutant relative to WT and hflK/C-comp strains to osmotic stress by ex-posing cells in different growth phases to 1 N NaCl and then determining cell viability (Fig. 1E). The number of viable spirochetes decreased after salt treatment for each strain at both the log and stationary phases. However, no significant differences in cell viability were detected among B. burgdorferi strains at either log phase or stationary phase.
The ⌬hflK/C mutant persists throughout the B. burgdorferi infectious cycle. Although the ⌬hflK/C mutant lacked a distinguishing phenotype in vitro, HflK and HflC function might potentially be important in vivo. Therefore, we assessed the ability of the mutant, complemented, and WT strains to complete the mouse-tick infectious cycle. However, no defect was observed in the ability to infect mice, either by needle inoculation or via tick bite (Table 1). Further, all three strains infected larval ticks and persisted through the molt at equivalent levels when Ixodes scapularis larvae were fed on infected mice, as determined by crushing fed ticks, plating, and counting CFU (Fig. 2). This result was independently confirmed by artificial infection of I. scapularis with all three strains and assessment of spirochete loads at both the larval and nymphal stages (data not shown) (25). Immunoblot analysis of HflK synthesis by different B. burgdorferi strains. Antiserum raised against B. burgdorferi HflK reacts with a protein of the appropriate size (36 kDa [arrowhead]) from whole-cell lysates of WT and complemented strains, but not with a lysate of the ⌬hflK/C mutant. (C) In vitro growth rate of the ⌬hflK/C mutant compared to the WT and hflK/C-comp strains. The mean and standard deviation (SD) are displayed. (D) Immunoblot analysis of cell lysates from WT B. burgdorferi exposed to a 40°C heat shock treatment for 1 h or maintained at the standard growth temperature (35°C). FlaB levels are shown to demonstrate equivalent protein loads. (E) Cell viability of B. burgdorferi strains subjected to 1 N NaCl osmotic stress compared to untreated spirochetes. After treatment, cells were plated in solid BSK medium, and the viability ratio was determined by counting CFU of treated compared to untreated spirochetes. SD bars are shown; no significant difference in cell viability was detected among B. burgdorferi strains at either the log phase (one-way analysis of variance [ANOVA], P ϭ 0.2850) or stationary phase (one-way ANOVA, P ϭ 0.1478). Subtle defects that might lower the competitive fitness of a strain may not be detected in the gross laboratory assessment of the mouse-tick infectious cycle described above. Two methods of assessing fitness differences between strains were also performed. First, the doses of in vitro-grown spirochetes required to infect 50% of inoculated mice (ID 50 ) were determined for the ⌬hflK/C mutant (3.16 ϫ 10 2 spirochetes) and for the WT (2.29 ϫ 10 2 spirochetes) (see Table S2 in the supplemental material). These values are not significantly different from each other and agreed with the previously determined ID 50 for WT B. burgdorferi strain A3 (26). Finally, mutant and WT strains were coinjected into mice at a 1:1 ratio to determine if the WT had a competitive advantage over the mutant. In three independent trials, each composed of 5 mice, neither strain consistently outperformed the other, as determined by sensitivity to the appropriate antibiotic (see Table S3 in the supplemental material). Further, larval I. scapularis ticks were artificially infected by immersion in Barbour-Stoenner-Kelly II (BSKII) medium containing both strains at approximately equivalent densities to determine if the ⌬hflK/C mutant had a reducedfitness phenotype in the vector. Both strains were present in fed larvae and survived through the molt and subsequent nymph feeding at approximately equal levels, spirochete numbers were quantified by plating crushed ticks on BSKII plates, and the data were analyzed with the Mann-Whitney test (see Fig. S2 in the supplemental material). The WT and ⌬hflK/C mutant were transmitted equally from the infected nymphs to naive mice, as assessed by isolation from mouse tissues (data not shown). These results indicate that inactivation of bb0203 (hflK) and bb0204 (hflC) does not detectably decrease the fitness of B. burgdorferi during tick colonization, persistence through the nymphal stage, or transmission to the murine host.
Generation and characterization of an inducible ftsH mutant. We attempted to delete bb0789 (the ftsH homolog) from the WT strain, B31-S9, in 6 independent transformations, but were unable to obtain a deletion mutant, supporting the hypothesis that FtsH function is essential to B. burgdorferi. Therefore, we generated an inducible ftsH mutant using the lac promoter/repressor system developed for B. burgdorferi (Fig. 3A) (27). Following the example of Gilbert et al., the lacI gene and a streptomycin resistance cassette were inserted into the bbe02 locus of the A3-68 WT strain, and the resulting strain was designated B31-68-LS (27). An allelic exchange construct that fused the inducible promoter, flacp, to ftsH and was linked to a gentamicin resistance marker was used to replace the endogenous ftsH open reading frame, and the resulting strain was designated the ftsH(in) strain. The structure of the inducible ftsH locus was confirmed by PCR (data not shown), and cell growth was shown to be dependent upon induction (Fig. 3B). Immunoblots confirmed LacI production and isopropyl-␤-D-thiogalactopyranoside (IPTG)controlled synthesis of FtsH (Fig. 3C).
The growth rate of the parental B31-68-LS strain was not affected by the addition of 2 mM IPTG (Fig. 3B). However, the growth rate of the ftsH(in) strain was dependent on the concentration of IPTG added to the BSKII medium, growing normally at concentrations ranging from 0.1 mM up to 10 mM ( Fig. 3B) (data not shown), but displaying a lower growth rate at or below IPTG concentrations of 0.01 mM. Immunoblot detection of FtsH showed that the ftsH(in) strain grown at concentrations of 1 mM IPTG and above produced significantly more FtsH than the parental B31-68-LS (WT) strain (Fig. 3C), but excess FtsH did not seem to adversely affect growth rate. When IPTG was depleted from the ftsH(in) culture, cell growth was arrested. Between 48 and 72 h post-IPTG depletion, FtsH levels were undetectable (Fig. 3C), spirochetes became nonmotile, and large membrane distortions developed (Fig. 4). By 72 h postdepletion, Ͼ99% of the cells were no longer viable. Strikingly, motile cells were again detected in these cultures around day 6 post-IPTG depletion, and a near-normal growth rate resumed (Fig. 3B). When examined at 10 days post-IPTG depletion, FtsH was again detectable (Fig. 3C), despite the absence of any inducing chemical. We designated this uncloned outgrowth genotype ftsH(res), for restored production of FtsH. These results were confirmed in three independent experiments.
Resumption of FtsH production might theoretically result from a suppressor mutation that eliminates LacI activity or LacI protein or from a mutation in the 20-nucleotide lac operator sequence inserted into the normally constitutive flgB promoter (27). A deletion or a base mutation in the short operator sequence three B. burgdorferi strains. Larval ticks were allowed to feed on infected mice, and 8 to 10 days postfeeding, spirochete density was determined by macerating and plating individual ticks and enumerating CFU. Some fed larvae were allowed to molt to the nymphal stage and fed on naive mice, and nymphs were mechanically disrupted 10 days postfeeding and plated to determine spirochete CFU. Each point represents an individual tick, and the mean and upper standard deviation bars are shown (lower SD bars fall below the x axis). No significant difference among strains was detected using the Kruskal-Wallis test. The ratios below the graphs denote the number of ticks that had acquired B. burgdorferi relative to the number of ticks assessed for each strain.
Chu et al.
might decrease the affinity of the LacI repressor for the altered sequence, freeing the promoter to allow transcription and translation to occur, and thereby restoring FtsH production. We tested these hypotheses by examining LacI protein levels from cell lysates obtained at days 2 and 10 post-IPTG depletion. Immunoblots that reacted with an anti-LacI antibody detected constitutive levels of the repressor at both time points, indicating that the suppressor mutation had not reduced or eliminated LacI levels. Sequencing lacI from 7 clones isolated from the ftsH(res) outgrowth strain did not detect any nucleotide changes from the original lacI gene used to transform B. burgdorferi, providing further evidence that the nature of the suppressor mutation was not related to LacI activity or production. Therefore, we determined the lac operator sequence from 9 clones isolated from the ftsH(res) culture. All clones contained a single transversion that altered a G to a T in the lac operator sequence (Fig. 5), potentially decreasing the binding affinity of the LacI repressor for the altered sequence.
The requirement for FtsH function in vitro indicated that an essential substrate(s) is processed during growth in BSKII medium. B. burgdorferi cells lacking FtsH display membrane defects after 48 h and are not viable after 72 h, potentially identifying a window in which FtsH substrates might accumulate and be visible on protein gels. Therefore, we compared WT cell lysate to that of the ftsH(in) strain grown in the presence of 1 mM IPTG (overproduction of FtsH) and to that of the ftsH(in) strain incubated for 48 h in the absence of inducer (FtsH depleted). However, no discrete differences in protein profiles among samples were observed by either Coomassie-stained or silver-stained gels (data not shown).

FtsH is required for infection of both mice and ticks.
Although we demonstrated that FtsH is required for in vitro survival of B. burgdorferi, it is possible that other proteins may be able to compensate for the loss of this protease during infection. To assess the requirement for FtsH in vivo, we inoculated mice with the ftsH(in) strain cultured in vitro with 1 mM IPTG, to ensure FtsH production and cell viability of the inoculum. The ftsH(in) strain was unable to establish an infection in mice, nor did the strain survive long enough to elicit an immune response, although all mice inoculated with WT B. burgdorferi became infected (Table 2). We also tested the ftsH(res) outgrowth strain, which has IPTG-independent expression of FtsH. Interestingly, this strain successfully established infection in 4 of 5 mice and was reisolated from all tissues examined from the 4 infected mice.
Since the ftsH(in) strain was not able to infect mice, we assessed the strain's ability to infect and persist in Ixodes ticks by artificially infecting larvae by immersion in media containing either the WT or ftsH(in) strain. The immersion inocula were plated, and the WT and ftsH(in) strains were shown to be at similar concentrations: 5.6 ϫ 10 7 and 5.0 ϫ 10 7 spirochetes/ml, respectively. Artificially infected larvae were subsequently fed to repletion on naive mice, and spirochete burdens were again assessed directly by mechanical disruption of the ticks in BSK medium and plating in solid medium to enumerate viable colony-forming units. Eight of seventeen ticks (47%) exposed to WT B. burgdorferi were infected, Note that the immunoblot corresponding to the WT strain contained approximately 10-fold more cell lysate relative to the other immunoblots. The faint band below the FtsH protein is due to nonspecific binding of the antibody to an unrelated protein, as it is present at the same level at all IPTG concentrations. and the spirochete burden/infected larvae was 24,600 Ϯ 20,359 (mean Ϯ standard deviation [SD]; n ϭ 6). In contrast, none of 17 larvae exposed to the ftsH(in) strain contained viable spirochetes after the blood meal. An additional 10 ticks exposed to the ftsH(in) strain were assessed for infection after the molt to the nymphal stage, and again, no viable spirochetes were detected. Overall, the data indicate that FtsH provides an essential function necessary for B. burgdorferi viability throughout the mouse-tick infectious cycle.

DISCUSSION
B. burgdorferi, an obligate parasite, transitions between arthropod vectors and mammalian hosts, environmental conditions that are substantially different from each other. For example, nutrients, cell density, pH, and temperature all shift between hosts, and many of these cues result in production of different proteins (28)(29)(30)(31)(32). To be a successful pathogen, B. burgdorferi must respond to these rapid environmental changes by controlling both protein turnover and proper functioning of membrane processes, such as protein and nutrient transport, environmental sensing/signaling, and flagellar motor function. In other bacteria, the FtsH protease contributes to these crucial cellular functions (reviewed in reference 33).
A homolog of FtsH has been identified in B. burgdorferi (16) (see Fig. S1 in the supplemental material), but has not been previously studied in this spirochete. Construction and use of an inducible FtsH strain demonstrated that FtsH is essential to B. burgdorferi survival during in vitro cultivation. When cellular levels of FtsH were depleted, occurring between 24 and 48 h after removal of inducer (Fig. 3C), growth was arrested, and within 72 h, membrane deformations were clearly visible (Fig. 4). Although substrates of the B. burgdorferi FtsH have not been identified, the morphology of FtsH-negative cells suggests that a substrate of this protease contributes to membrane architecture. The substrate(s) that requires FtsH processing for B. burgdorferi viability presents an intriguing area of research for future investigation.
Surprisingly, viable cells were again detected on day 6 post-  IPTG depletion. These cells were producing FtsH, when assessed by immunoblotting on day 10 (Fig. 3C), despite the absence of any inducer. Investigating the nature of the presumptive suppressor mutation in the ftsH(res) outgrowth strain, we identified a singlebase mutation (a transversion of G to T) in the lac operator sequence (Fig. 5). In studies of the wild-type Lac repressor/operator sequence, this single-base mutation at the central nucleotide reduced the LacI binding affinity for the operator by 99% (34). However, the operator sequence constructed by Gilbert et al. and used in this study contained the "ideal" lac operator sequence, which lacks the central nucleotide and is perfectly symmetrical (27). We were unable to find any studies that examined repressor binding affinities for mutations in the ideal lac operator, but the transversion identified here would reduce the number of bases conferring symmetry. This result reinforces the essential nature of the B. burgdorferi FtsH by demonstrating that only cells with a secondary mutation (i.e., allowing FtsH production) are viable.
In another assessment of the requirement for FtsH, mice injected with the ftsH(in) strain did not become infected or seroconvert, whereas both the WT and the ftsH(res) strain in the outgrowth culture were infectious in mice ( Table 2). The combined in vitro, murine, and tick infectivity data demonstrate that FtsH is essential to B. burgdorferi survival both in liquid culture and throughout the infectious cycle. Further, strict regulation of FtsH production does not seem to be required, at least under the conditions examined in this study, which seems surprising for a protein with an essential role in cell survival. Production of FtsH in the outgrowth culture was enough to restore infectivity, despite the fact that ftsH was no longer under the control of its native promoter. Equally surprising was the finding that overproduction of FtsH by increasing IPTG levels above 0.1 mM did not have any detectable adverse effects on B. burgdorferi growth in vitro, suggesting that excess FtsH does not proteolytically overwhelm its substrates.
In contrast to the FtsH results, we did not detect any phenotype for cells lacking HflK and HflC. In E. coli, these proteins act to modulate FtsH substrate recognition (11,12,14), and homologs of these accessory proteins are widespread among spirochetes, although not ubiquitous in prokaryotes. Recently, Toledo et al. demonstrated that FtsH, HflK, and HflC are present in membrane lipid rafts of B. burgdorferi (35), consistent with previous findings in E. coli, where these three proteins form a large, multimeric inner membrane complex (12,15). Potentially, HflK and HflC regulate the accessibility of minor membrane substrates to the FtsH protease, and the processing of these substrates is not significant enough to present a phenotype in the laboratory setting but in nature would provide a selective advantage. Our ID 50 results and mouse-tick competition studies indicate this is an unlikely scenario, but we are unable to eliminate this possibility. Alternatively, the B. burgdorferi HflK and HflC proteins may have evolved a function different from that of the E. coli homologs and perhaps separate from FtsH function, and we have not yet identified the conditions under which these proteins provide a beneficial function in B. burgdorferi.
Further investigation to identify the substrates of FtsH should provide insight into the essential nature of this protease and may indicate specific targets that inhibit B. burgdorferi viability. Additional characterization of FtsH-HflK-HflC function will undoubtedly be useful in identifying interesting mechanisms of this spirochete's physiology and obligate pathogenic life cycle. Bacterial strains and growth conditions. The bacterial strains used in this study are listed in Table S1 in the supplemental material. B. burgdorferi strains were grown in liquid Barbour-Stoenner-Kelly II (BSKII) medium (38,39) supplemented with 6% rabbit serum (PelFreez Biologicals, Rogers, AZ) and appropriate antibiotics (kanamycin, 200 g/ml; streptomycin, 50 g/ml; gentamicin, 40 g/ml). Cloning vectors were propagated using E. coli strain TOP10 (Invitrogen, Carlsbad, CA) or DH5␣ (New England Biolabs, Ipswich, MA).

MATERIALS AND METHODS
Sequence analysis. MacVector version 14.0.3 (MacVector, Inc., Apex, NC) was used to align amino acid sequences with ClustalW. The conserved domains of BB0789, BB0203 and BB0204 were searched and analyzed in the CDD database (CDD v3.12), a conserved domain database for the functional annotation of proteins (http://www.ncbi.nlm.nih.gov/ Structure/cdd/wrpsb.cgi).
Generation of BB0203 and BB0789 antisera. Protein BB0203 lacking the predicted transmembrane helix domain (matching residues 1 to 35 of BB0203 protein) and a 16-amino-acid peptide matching residues 317 to 332 of BB0789 were commercially synthesized (GenScript, Piscataway, NJ). The 16-amino-acid peptide region is relatively well conserved among eubacteria and was originally described by Tomoyasu et al. (1). BB0203specific and BB0789-specific antisera were generated by immunizing 1-year-old female New Zealand White rabbits with the corresponding protein in the Rocky Mountain Laboratories Animal Unit.
Construction of mutant strains. All B. burgdorferi transformations were performed similarly: 10 g of plasmid DNA was electroporated into wild-type strain B31-S9 (or specified strains). All B. burgdorferi transformants described were confirmed by PCR and Southern blot analysis, and plasmid content was determined as previously described (40,41).
All primers used in this study are listed in Table S4 in the supplemental a The inocula for needle infection were 8 ϫ 10 3 spirochetes injected intraperitoneally and 2 ϫ 10 3 spirochetes injected subcutaneously. b Day 10 outgrowth was isolated from the ftsH(in) strain cultured in the absence of IPTG and designated the ftsH(res) strain. A spontaneous mutation in the lac operator sequence presumably allowed for resumption of FtsH production and cell growth.
material. The allelic exchange construct for generation of the ⌬hflK/C mutant (Fig. 1A) was constructed by PCR amplification of a 561-bp upstream region of bb0203 (hflK) and a 619-bp downstream region of bb0204 (hflC) from B31-S9 genomic DNA. The DNA fragments were cloned flanking the kanamycin resistance cassette driven by the B. burgdorferi flgB promoter (42) in pCR-XL-TOPO (Invitrogen). The resulting vector, designated pTAKO.bb0203/0204, was confirmed by sequencing and digestion with the appropriate restriction enzymes. The mutant ⌬hflK/C strain was complemented by inserting hflK and hflC into the native chromosomal locus adjacent to the deletion mutation (Fig. 1A). In brief, a DNA fragment containing hflK and hflC with 232 bp of the putative native promoter region was PCR amplified and sequenced to confirm nucleotide fidelity. This fragment was cloned into the allelic exchange vector (described above) with the gentamicin resistance cassette driven by the flaB promoter (43), creating vector pTAComple.bb0203/ 0204, and inserted after the kanamycin resistance cassette.
We attempted to isolate a deletion mutant of ftsH using an approach similar to that described above for the ⌬hflK/C mutant. However, after 6 independent electroporations, we were unable to obtain a deletion mutant of ftsH, although this approach worked successfully for obtaining the HflK/C mutant. Therefore, we focused on constructing an FtsH-inducible strain.
The B. burgdorferi IPTG-inducible promoter system constructed by Gilbert et al. (27) was used to generate an inducible ftsH (bb0789) mutant. Constructs pTAflacp and pBBE02::lacI-Strep r were generously provided by Dan Drecktrah and Scott Samuels (see Table S1 in the supplemental material). Plasmid pBBE02::lacI-Strep r was used to construct a LacIproducing strain in an infectious background by transformation into B. burgdorferi clone B31 A3-68; the resulting strain was designated B31-68-LS (see Table S1). LacI production in B31-68-LS was confirmed by immunoblot analysis with an antibody raised against LacI (Rockland Immunochemicals, Pottstown, PA). The vector for generating an inducible bb0789 mutant was constructed by amplifying 646 nucleotides of the 3= region of bb0788 from wild-type genomic DNA and cloning it upstream of the inducible promoter in plasmid pTAflacp. The bb0789 coding region was amplified and cloned downstream of the flac promoter to generate plasmid pTAbb0788-flacp-bb0789. The downstream flanking region (bb0790) and selectable marker were assembled in pCR-XL-TOPO. First, the flaBp-gentamicin resistance cassette was amplified and cloned into pCR-XL-TOPO. Subsequently, 672 nucleotides of the 5= region of bb0790 was inserted downstream of the gentamicin marker to produce plasmid pTAflaBp-gent-bb0790. The bb0788-flacp-bb0789 fragment from pTAbb0788-flacp-bb0789 was subcloned into plasmid pTAflaBp-gent-bb0790 to generate an allelic exchange vector, pTAindu.bb0789. B31-68-LS was electroporated with pTAindu.bb0789 and selected on solid medium for gentamicin resistance in the presence of 2 mM IPTG to generate the inducible ftsH(in) strain (Fig. 3A). Transformants were confirmed by amplifying the gentamicin resistance cassette and the flacp promoter upstream of bb0789.
The IPTG-dependent growth of the ftsH(in) strain was assessed by culture in the presence and absence of IPTG. Spirochetes grown in medium supplemented with 2 mM IPTG and appropriate antibiotics were harvested and washed twice in BSK-H medium (Sigma-Aldrich, Atlanta, GA). Cells were resuspended in fresh BSKII medium to a density of 1 ϫ 10 5 spirochetes/ml and supplemented with IPTG to final concentrations of 0.01, 0.1, 2, 5, or 10 mM or without IPTG. Growth of the parental strain B31-68-LS was also assessed with either 2 mM IPTG or without IPTG. Cultures were grown in triplicate for statistical accuracy, and the entire growth analysis was repeated independently. Cell densities were monitored every 24 h by dark-field microscopy using a Petroff-Hausser chamber.
To confirm IPTG-dependent regulation of BB0789 in the ftsH(in) strain, cell lysates from the growth assays were assessed by immunoblot analysis using an antibody raised against an FtsH peptide (described above). Spirochetes inoculated with 0.01, 0.1, 2, or 10 mM IPTG were harvested when cell densities reached approximately 1 ϫ 10 8 spirochetes/ ml. Cell lysates were prepared for immunoblot analysis using rabbit anti-BB0789 antiserum (1:1,000 dilution) and anti-LacI (rabbit) antibody (1: 5,000 dilution).
Experimental mouse-tick infection studies. Mouse infection studies were performed with 6-to 8-week-old female RML mice, an outbred strain of Swiss-Webster mice reared at the Rocky Mountain Laboratories breeding facility. Mice were inoculated intraperitoneally (4 ϫ 10 3 spirochetes) and subcutaneously (1 ϫ 10 3 spirochetes). The number of injected spirochetes was confirmed by plating an aliquot of the inoculum and determining CFU. Approximately 20 B. burgdorferi colonies per strain were screened by PCR to confirm the presence of plasmids lp25, lp28-1, and lp36, which are somewhat unstable in vitro but required for B. burgdorferi infectivity (41,44). Mouse infection was assessed at 3 weeks postinjection by mouse seroconversion to B. burgdorferi antigens and isolation of spirochetes from mouse tissues (ear, bladder, and ankle joint) in liquid culture.
Approximately 100 to 200 naive I. scapularis larvae were fed to repletion on each infected mouse. Acquisition of B. burgdorferi by larval ticks was assessed 8 to 10 days postfeeding, and spirochete load was enumerated by mechanical disruption and plating. Basically, individual ticks were placed in a 1.5-ml Eppendorf tube containing 0.5 ml BSKII medium and crushed with a sterile, disposable pestle. After disruption of the tick, an additional 0.5 ml of medium was added, and aliquots were plated in solid BSK medium with appropriate antibiotics. Where relevant, the remaining I. scapularis larvae were allowed to molt to nymphs, and cohorts of 5 to 10 nymphs were fed on individual naive mice to assess B. burgdorferi transmission. Fed nymphs were mechanically disrupted 10 days postfeeding and plated in solid BSK medium to determine spirochete loads.
The relative fitness of B. burgdorferi strains was determined in three independent mouse coinfection studies and also within ticks by artificial infection. The inoculum per mouse was an equal mixture of two strains (i.e., the ⌬hflK/C mutant to WT, or the complemented strain to WT) with 5 ϫ 10 3 or 1 ϫ 10 4 spirochetes. An aliquot of the inoculum was plated with or without antibiotic selection to determine the actual ratio of strains. Mouse infection was determined by seroreactivity and confirmed by spirochete isolation from mouse tissues (ear, bladder, and ankle joint) at 4, 8, or 10 weeks postinoculation. The ratio of the two strains in mouse tissues was determined by plating spirochetes in solid BSK medium (with or without antibiotic selection) and enumerating CFU and/or by PCR screening of the B. burgdorferi colonies from plates lacking antibiotics.
Artificial tick coinfection was previously described (25,26). Briefly, about 200 I. scapularis larvae were artificially infected by immersion in a B. burgdorferi culture. To determine the relative fitnesses of two strains, equal numbers of ⌬hflK/C and wild-type cells or the complemented strain and wild type at a combined density of~1 ϫ 10 8 spirochetes/ml were mixed together and used for immersion of ticks. The ratio of each strain in the mixture was determined by plating in solid medium, as described above. Each cohort of artificially infected larvae was fed to repletion on naive RML mice and allowed to molt to nymphs. The nymphs were subsequently fed to repletion on naive mice. The ratio of B. burgdorferi strains within ticks was determined by mechanically disrupting the ticks and plating with or without antibiotic selection.
Determination of ID 50 . The ID 50 s for B31-S9 and the ⌬hflK/C strain were assessed by inoculating mice with 10-fold serial dilutions from 10 5 to 10 2 spirochetes for each strain. Six mice were inoculated per dose. The actual number of injected viable spirochetes was determined by plating a portion of the inoculum and enumerating CFU. Mouse infection was assessed by B. burgdorferi isolation from mouse tissues at 4 weeks postinoculation. The ID 50 value for each strain was calculated according to the method of Reed and Muench (45).
In vitro phenotype analysis of the ⌬hflK/C mutant. To assess any in vitro phenotypes of ⌬hflK/C strain, the growth rate, morphology, and protein profile of the mutant strain were compared to those of the wild type, B31-S9, and the complemented strain. The growth rate was deter-mined by inoculating triplicate 5-ml BSKII cultures with 1 ϫ 10 5 spirochetes/ml and incubating the cultures at 35°C. Cell densities were monitored every 24 h by dark-field microscopy using a Petroff-Hausser chamber. Morphological characteristics of B. burgdorferi strains were observed by scanning electron microscopy, as previously described (46). The protein profiles and antigenic profiles of all three strains were determined by SDS-PAGE and immunoblot analysis using pooled sera from B. burgdorferi-infected mice as previously described (26,47).
Osmotic shock and heat stress assays were performed similarly to compare the response of the ⌬hflK/C strain to that of the WT and complemented strains. These three strains were individually inoculated in triplicate at an initial concentration of 1 ϫ 10 5 spirochetes/ml, and each was subdivided into 2 portions when cultures reached a cell density of approximately 5 ϫ 10 7 spirochetes/ml (mid-log phase) or 2 ϫ 10 8 spirochetes/ml (stationary phase). For the osmotic shock assay, one set of cultures was treated with 1 N NaCl for 10 min for mid-log-phase cultures or 40 min for stationary-phase cultures, similar to the protocol of Elias et al. (48). Control cultures were supplemented with an equivalent volume of BSK medium. After treatment, spirochetes were immediately plated in solid BSK medium, and CFU were subsequently counted to assess cell viability under salt stress by calculating the ratio of CFU with salt treatment relative to that without treatment. The heat shock assays were set up as described above, except that one set of B. burgdorferi cultures was incubated for 1 h at 40°C, a temperature demonstrated to induce a heat shock response in B. burgdorferi (22)(23)(24), and the other was kept at 35°C for 1 h as a control. After treatment, equivalent numbers of spirochetes were harvested, and cell lysates were prepared for immunoblot analysis using rabbit anti-BB0203 antiserum (1:2,000 dilution). The experiment was repeated with the heat shock treatment extended to 4 h.