Borrelia burgdorferi Alters Its Gene Expression and Antigenic Profile in Response to CO₂ Levels

Strains and growth conditions.All strains of B. burgdorferi used in this study were grown in BSK-II medium supplemented with 6% normal rabbit serum (Pel-Freez Biologicals, Rogers, AR), which is referred to as complete BSK-II medium below. The CMRL-1066 medium (United States Biologicals, Swampscott, MA) used in BSK-II medium lacked any added bicarbonate. Low-passage, infectious, clonal B31 derivative MSK5 and clonal 297 derivatives AH130 (parent), AH210 (rpoS mutant), AH212 (rpoN mutant), and AH123 (rpoN mutant complemented by wild-type rpoN) were used in this study (26, 29). Microaerophilic cultures were grown statically at 32°C and pH 7.8 with 1% atmospheric CO₂ in complete BSK-II medium. The anaerobic culture conditions consisted of 5% CO₂, 3% H₂, pH 7.3, and 32°C in complete BSK-II medium. The level of oxygen was reduced 40-fold in the anaerobic complete BSK-II medium (0.087 ppm) compared to the microaerophilic medium (3.48 ppm), as determined using a DO-166 oxygen probe (Lazar Research Laboratories, Los Angeles, CA) (41). All cultures were inoculated at a density of 1× 10⁴ cells per ml and grown to a density of 5× 10⁷ cells per ml to acquire RNA or protein samples. For the pH study, microaerophilic cultures were grown in complete BSK-II medium at pH 7.3, and the pHs of anaerobic cultures with the same starting density were adjusted to 7.8. CO₂ was displaced from anaerobically grown cultures by treatment with nitrogen gas for 20 min at a pressure of 10 lb/in², which reduced the CO₂ level from 5,800 ppm to undetectable levels. Dissolved CO₂ levels were measured using a CO-35 probe (Lazar Research Laboratories, Los Angeles, CA). RNA and protein samples were taken from anaerobic cultures and anaerobic cultures lacking CO₂ at a density of 5× 10⁷ cells per ml. MSK5 was grown under anaerobic and microaerophilic conditions in modified BSK-II medium containing decreasing amounts of NaHCO₃, so that the medium contained 25 mM, 15 mM, 5 mM, 1 mM, or no added NaHCO₃; the highest concentration represented conventional BSK-II medium. The NaHCO₃-modified medium was supplemented with NaCl to maintain a constant osmotic balance. To examine the role of different CO₂ levels with a constant O₂ content, AH130 and MSK5 were also grown statically in the presence of atmospheric O₂ (3.48 ppm dissolved O₂) and 5% atmospheric CO₂ at 32°C or 37°C.

SDS-PAGE and immunoblotting.Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting using infection-derived and monospecific antisera as primary antibodies were conducted as previously described (41).

RNA isolation.Three independent cultures of B. burgdorferi strains MSK5 and AH130 were grown to the exponential growth phase (i.e., 5 ×10⁷ cells per ml), and total RNA was isolated from 1 ×10⁹ cells using a Versagene kit (Gentra Inc., Minneapolis, MN). RNA samples were treated with DNase I (Roche Inc., Indianapolis, IN) and Superase · In (Ambion Inc., Austin, TX) to eliminate contaminating DNA and inhibit RNase activity, respectively. Three independent RNA samples of each strain tested were pooled, and DNA contamination and crude RNA yield were examined by PCR and reverse transcription (RT)-PCR, respectively.

Quantitative RT-PCR.A defined set of genes was subjected to quantitative RT-PCR to ascertain whether the antigenic production observed was due to regulation at the transcriptional level. Oligonucleotide primers (Table 1) were designed with the Primer Express software (Perkin-Elmer Biosystems, Foster City, CA). Selected primer pairs were tested to confirm that they amplified a single product with a known size using genomic B. burgdorferi DNA as the template. Reverse transcription reactions were performed by combining TaqMan reverse transcription reagents (Applied Biosystems, Foster City, CA) with purified B. burgdorferi total RNA. A control reaction with a mixture lacking reverse transcriptase was performed for each primer set using total RNA from each B. burgdorferi strain to confirm that DNA was not present. Subsequently, the products from the reverse transcription reaction were subjected to real-time PCRs using an Applied Biosystems 7500 real-time PCR system. SYBR green PCRs were performed in triplicate, and each experiment was repeated in triplicate, resulting in nine data points for each gene of interest and for each B. burgdorferi strain tested. A constitutively expressed gene, flaB, which was not affected by any treatment tested in this study, was used for normalization as previously described (41). The levels of induction of genes induced during anaerobiosis with CO₂ compared to the levels observed during anaerobiosis without CO₂ were determined by the ΔΔC_t method as previously described (7, 41).

Statistical analyses.The real-time RT-PCR data from three independent experiments were analyzed using a resampling bootstrap procedure and the permutation two-sample test. The bootstrap distribution provides an accurate estimate of the lower and upper limits, respectively, of a 95% confidence interval around the true mean. For the permutation two-sample test, the distribution of the data tested the null hypothesis that CO₂ had no influence on gene expression. A P value of ≤0.01 was used. The data set for each gene was based on ΔΔC_t values (relative to flaB for B. burgdorferi grown with or without CO₂) to ensure that there were normal distributions for accurate probability estimates. All statistical tests and data resampling operations were performed with S-PLUS, version 7.02 (Insightful Corp., Seattle, WA).

Different antigenic compositions of B. burgdorferi under anaerobic and microaerophilic conditions.As B. burgdorferi moves through the disparate environments encountered in the arthropod vector and mammalian host, changes in temperature, pH, and other host factors modulate gene expression (2, 3, 7, 10, 16, 40, 45, 53, 55). Previous studies indicated that the redox environment of B. burgdorferi alters gene expression and protein synthesis in this spirochetal pathogen (4, 41). To address this question further, defined anaerobic and microaerophilic conditions were imposed to determine how dissolved gases, including CO₂, affect borrelial gene regulation and protein synthesis. B. burgdorferi was grown statically in complete BSK-II medium at 32°C for all conditions tested unless indicated otherwise. Microaerophilic growth conditions were maintained with 1% atmospheric CO₂, and the anaerobic environment was defined as 5% CO₂ and 3% H₂ atmospheric levels in a controlled anaerobic chamber. An oxygen electrode was used to measure dissolved oxygen levels in microaerophilic complete medium (3.48 ppm) and anaerobic complete medium (0.087 ppm), and the results indicated that the anaerobic culture conditions resulted in a 40-fold reduction in the dissolved oxygen level compared to the level in microaerophilic medium. All B. burgdorferi cells tested exhibited normal motility under both of the culture conditions mentioned above without a significant difference in growth (data not shown). The antigenic responses of B. burgdorferi to anaerobic and microaerophilic growth conditions were assessed by Western immunoblot analysis. AH130 and MSK5 samples from cultures grown under each condition were probed with serum from a patient with chronic Lyme disease or with infection-derived mouse serum (Fig. 1). Major antigenic differences for B. burgdorferi AH130 and MSK5 were observed throughout the immunoblot, and there was significantly greater synthesis of the antigens in the anaerobically grown cultures (Fig. 1).

• Open in new tab
• Download powerpoint

FIG. 1.

Immunoreactive proteins induced when B. burgdorferi AH130 and MSK5 are grown under anaerobic or microaerophilic conditions with constant CO₂ levels. Protein samples were probed with serum from a patient with chronic Lyme borreliosis (A) or with infection-derived mouse serum (B). Lanes 1 and 2 contained protein from the strain 297 clonal derivative AH130 grown anaerobically and microaerophilically, respectively. Lanes 3 and 4 contained protein from the strain B31 clonal derivative MSK5 grown anaerobically and microaerophilically, respectively. The numbers on the left indicate the molecular masses (in kilodaltons) of protein markers.

The production of individual borrelial antigens was examined under anaerobic and microaerophilic conditions (Fig. 2). FlaB synthesis was unchanged under the experimental conditions employed and was used as a control to demonstrate equivalent protein levels in samples. Proteins associated with oxidative stress, including BosR, a borrelial oxidative stress regulatory protein, and NapA, a Dps/Dpr homolog, produced more of these specific antigens during anaerobic growth than during microaerophilic growth for both strains of B. burgdorferi analyzed, AH130 and MSK5 (Fig. 2). Previous studies demonstrated that NapA levels decreased when the culture medium was pretreated with nitrogen gas displacement and with Oxyrase to deplete all dissolved gases and O₂, respectively (41). The results obtained here using an anaerobic chamber did not corroborate these findings, suggesting that the different culture conditions utilized accounted for the different NapA levels observed. The levels of the decorin binding adhesin, DbpA, and OspC increased under anaerobic conditions, similar to previously reported observations (41). In addition, the level of the borrelial RpoS sigma factor also increased when B. burgdorferi was grown under anaerobic conditions (Fig. 2).

• Open in new tab
• Download powerpoint

FIG. 2.

Antigens from B. burgdorferi strains AH130 and MSK5 are synthesized differentially when the organisms are cultivated under anaerobic (lanes A) and microaerophilic (lanes M) growth conditions. Protein lysates from each strain grown under each condition were resolved by SDS-PAGE, immobilized on polyvinylidene difluoride membranes, and probed with antiserum specific for the antigens indicated on the left. Constitutively synthesized FlaB was used as a control to demonstrate that there were equivalent amounts of protein in the samples.

Effect of pH on borrelial protein production.Increased levels of CO₂/bicarbonate decreased the pH of complete BSK-II medium, and since pH is known to affect B. burgdorferi gene expression and protein production, the appropriate adjustments were made to the medium to compensate for the difference (9, 10). The pH values of anaerobic and microaerophilic complete BSK-II media were 7.3 and 7.8, respectively. To assess if the observed changes in specific borrelial protein production could be attributed to the difference in pH values, MSK5 was also grown anaerobically in medium whose pH was adjusted to 7.8 and concurrently, the pH of microaerophilic medium was adjust to 7.3 to reflect the hydrogen ion concentration of the alternate growth conditions (Fig. 3). Under these conditions, DbpA, NapA, and BosR protein production did not change in response to pH compared to the protein profiles observed for microaerophilically or anaerobically grown B. burgdorferi, as shown in Fig. 2. If the induction was due to differences in pH, then one would expect that the level of each antigen would be significantly induced when organisms were grown microaerophilically at pH 7.3. Although the level of DbpA increased somewhat in cells grown microaerophilically at pH 7.3, the level never approached what was observed under conventional anaerobic conditions (Fig. 2). Furthermore, the levels of NapA and BosR were not affected by lowering the pH under microaerophilic conditions (Fig. 3). Taken together, these results suggest that the change in protein production previously observed during anaerobic growth (Fig. 2) is independent of pH.

• Open in new tab
• Download powerpoint

FIG. 3.

Modulation of the protein produced under the experimental conditions used is independent of the pH. A B. burgdorferi B31 derivative, MSK5, was grown under anaerobic conditions at pH 7.3 and 7.8 (lanes A) and under microaerophilic conditions at pH 7.3 and 7.8 (lanes M). Protein samples were probed with monospecific antisera to the borrelial proteins DbpA, NapA, and BosR.

Influence of CO₂ and bicarbonate on B. burgdorferi.The effect of CO₂ and bicarbonate levels on borrelial protein synthesis was analyzed, as the atmospheric level of CO₂ was 4% higher under the defined anaerobic growth conditions than under the microaerophilic growth conditions. Dissolved O₂ and CO₂ are present under both in vivo and in vitro growth conditions, and in this study nitrogen gas was used to indiscriminately displace both dissolved O₂ and CO₂ from the growth medium. A CO₂ probe was used to measure dissolved CO₂ levels in microaerophilic (1,700 ppm), anaerobic (5,800 ppm), and nitrogen gas-purged anaerobic (0 ppm) complete BSK-II media. To directly assess the effect of CO₂ on gene expression and protein production, anaerobic cultures were treated with nitrogen gas to completely displace the dissolved CO₂ (Fig. 4). Protein samples from the anaerobic, nitrogen gas-treated anaerobic, and microaerophilic cultures were probed with monospecific antisera to several borrelial proteins. As described above, constitutively synthesized FlaB was used as a control between samples and for the different treatments employed. The DbpA and OspC levels were greatly increased when CO₂ was present, suggesting that in addition to O₂, CO₂ levels modulate gene expression and protein production in B. burgdorferi (Fig. 4).

• Open in new tab
• Download powerpoint

FIG. 4.

Displacement of dissolved CO₂ modulates antigen synthesis in B. burgdorferi. AH130 and MSK5 cultures were grown anaerobically (0.087 ppm O₂, 5% CO₂) (lanes A), anaerobically with CO₂ removed by N₂ gas displacement (0.087 ppm O₂, no CO₂) (lanes N), and microaerophilically (3.48 ppm O₂, 1% CO₂) (lanes M). Protein lysates from each strain grown under each condition were resolved by SDS-PAGE, immobilized on polyvinylidene difluoride membranes, and probed with antisera specific for the antigens indicated on the left. Constitutive synthesis of FlaB was used as a control to demonstrate that there were equivalent amounts of protein in the samples.

Dissolved CO₂ levels also affected the synthesis of borrelial RpoS and BosR, an alternate sigma factor and redox regulatory protein, respectively. Specifically, RpoS and BosR were induced greatly when B. burgdorferi was grown anaerobically in the presence of 5% CO₂ compared to the synthesis in cells grown anaerobically without CO₂ or in microaerophilically grown B. burgdorferi (Fig. 4). Note that BosR, DbpA, and OspC, although not readily detectable (Fig. 4), were synthesized in microaerophilically grown B. burgdorferi if the exposure time of the blot was increased (data not shown). This suggests that the production of some borrelial proteins is influenced by dissolved CO₂ levels, as well as dissolved oxygen levels. The enhanced production of DbpA and OspC is consistent with the increased synthesis of RpoS since previous studies have demonstrated that borrelial RpoS is required for expression of dbpA and ospC (26, 53). Taken together, these results suggest that CO₂ is an additional signal that is integrated by B. burgdorferi to modulate gene expression.

Since RpoS, OspC, and DbpA levels were increased by anaerobiosis, we examined whether the increases were dependent on the well-characterized borrelial RpoN-RpoS regulatory system (8, 19, 26, 53, 55). To assess this, a B. burgdorferi rpoN mutant, an rpoS mutant, and the complement of the rpoN mutant were grown under microaerophilic, anaerobic, and nitrogen-treated anaerobic conditions. In the presence of CO₂, the synthesis of RpoS, DbpA, and OspC was significantly reduced in the absence of RpoN, but the levels reverted to wild-type levels in the rpoN complemented strain (data not shown), indicating that increased production of RpoS, OspC, and DbpA is dependent on a functional RpoN and apparently does not involve other regulatory loci, at least under these experimental conditions.

Previous results demonstrated that NapA levels were reduced when B. burgdorferi was grown in BSK-II medium purged with nitrogen gas compared to the levels in microaerophilically grown samples (41). In the previous study, the decrease in the NapA level was attributed to a decrease in the dissolved oxygen level. In this study, the levels of NapA were evaluated when cells were grown in an anaerobic chamber containing 5% CO₂. Surprisingly, appreciably more NapA was produced by cells under these conditions than by cells grown anaerobically without CO₂ (Fig. 5). Furthermore, the reduction in the NapA level was apparently dependent on a decrease in the CO₂ level independent of oxygen since the levels of NapA produced anaerobically with CO₂ were slightly greater than the levels found in microaerophilically grown B. burgdorferi (Fig. 2 and 5). In fact, the level of NapA production was maximal when O₂ was missing and CO₂ was present, followed by microaerophilic growth; anaerobic growth without CO₂ resulted in the smallest amount of NapA for both strain MSK5 and strain AH130 (Fig. 5). This hierarchy of regulation is consistent with the amounts of dissolved CO₂ available in the samples; that is, the larger the amount of CO₂ in the sample, the larger the amount of NapA produced. These results indicate that CO₂/bicarbonate levels dramatically influence NapA protein production in B. burgdorferi and suggest that dissolved oxygen plays a minor role in the regulation of napA.

• Open in new tab
• Download powerpoint

FIG. 5.

Synthesis of NapA is affected by CO₂ levels. NapA production was assessed by immunoblot analysis of protein lysates from strains AH130 and MSK5 grown either anaerobically (0.087 ppm O₂, 5% CO₂) (lanes A), anaerobically with CO₂ removed by N₂ gas displacement (0.087 ppm O₂, no CO₂) (lanes N), or microaerophilically (3.48 ppm O₂, 1% CO₂) (lanes M).

Bacteria can sense CO₂/bicarbonate via adenylyl cyclase, which results in increased cAMP levels that, in turn, modulate gene expression (22, 52). Conventional microaerophilic growth conditions consist of atmospheric CO₂ and the 25 mM NaHCO₃ in BSK-II medium. To examine the effect of bicarbonate levels in anaerobically (5% CO₂) and microaerophilically (1% CO₂) grown B. burgdorferi, the concentration of sodium bicarbonate was decreased in the borrelial growth medium while the overall salt content was controlled (Fig. 6). MSK5 was grown anaerobically or microaerophilically with additional NaHCO₃ added to the medium at a concentration of 25 mM, 15 mM, 5 mM, or 1 mM or with no additional NaHCO₃ in addition to the equilibrium that was already present between bicarbonate and dissolved CO₂. Under these conditions, changes in antigen production were evaluated (Fig. 6). FlaB was not affected by bicarbonate levels and served as a control for equal protein loading (Fig. 6). Decreasing the concentration of sodium bicarbonate did not influence the synthesis of BosR, DbpA, OspC, RpoS, or NapA (Fig. 6). The synthesis of these antigens was consistent with the synthesis observed for anaerobic and microaerophilic growth conditions (Fig. 2, 4, and 5), and the results indicated that the level of production observed is not appreciably affected by the addition of bicarbonate (Fig. 6).

• Open in new tab
• Download powerpoint

FIG. 6.

Antigen synthesis is influenced by CO₂ rather than by bicarbonate levels. A strain B31 derivative, MSK5, was grown with 25 mM (conventional concentration), 15 mM, 5 mM, 1 mM, and no added NaHCO₃ under both anaerobic and microaerophilic conditions. Protein lysates from each strain grown under each condition were resolved by SDS-PAGE, immobilized on polyvinylidene difluoride membranes, and probed with antisera specific for the antigens indicated on the left. Constitutive synthesis of FlaB was used as a control to demonstrate that there were equivalent amounts of protein in the samples. Note that synthesis of the antigens tested in the microaerophilically grown samples generated protein species that were detected following longer exposure times.

To confirm the finding that 5% CO₂ altered borrelial protein synthesis and to examine the role of oxygen in this process, AH130 and MSK5 were grown microaerophilically in the presence of 5% CO₂ with atmospheric O₂ (3.48 ppm dissolved O₂) at 32°C and 37°C (Fig. 7). Then protein production was evaluated compared to the production under conventional microaerophilic growth conditions (1% CO₂, atmospheric O₂, 32°C), and FlaB antibody was used to control for equivalent protein loading. As in previous experiments, there was increased synthesis of BosR, NapA, DbpA, OspC, RpoS, and BBK32 in the presence of 5% CO₂ compared to the synthesis in the presence of 1% CO₂ when the oxygen levels remained constant. It is important to note that BosR, NapA, DbpA, OspC, and BBK32 were detectable in the samples grown in the presence of 1% CO₂ at 32°C when the blot was exposed for a longer time, indicating that there was a great difference in protein production when CO₂ was limiting (Fig. 7). For RpoS, there was apparently more synthesis under microaerophilic conditions than observed previously (Fig. 4) due to the longer exposure times required to see antigen in any of the samples tested. Regardless, the overall trend (i.e., more RpoS as the level of CO₂ increased) was consistent. These results suggest that CO₂ levels modulate the production of borrelial proteins that are believed to be important for adaptation in the mammalian host environment.

• Open in new tab
• Download powerpoint

FIG. 7.

Increased CO₂ levels in the presence of atmospheric O₂ result in levels of antigen synthesis comparable to levels observed for anaerobically grown B. burgdorferi. AH130 and MSK5 were grown under conventional microaerophilic growth conditions (1% CO₂ with atmospheric O₂) and with increased levels of CO₂ and atmospheric O₂ (either 5% CO₂ at 32°C or 5% CO₂ at 37°C). Protein lysates from each strain grown under each condition used were resolved by SDS-PAGE, immobilized on polyvinylidene difluoride membranes, and probed with antisera specific for the antigens indicated on the left. Constitutive synthesis of FlaB was used as a control to demonstrate that there were equivalent amounts of protein in the samples. Note that synthesis of BosR, NapA, DbpA, OspC, and BBK32 was observed in the samples grown with 1% CO₂ at 32°C following longer exposure times.

To examine the effect of temperature on the response to CO₂, cultures were grown at 32°C with 1% CO₂ or 5% CO₂ and compared with borrelial cells grown at 37°C with 5% CO₂. In nearly all cases, an increase in the CO₂ level resulted in enhanced synthesis of borrelial proteins independent of the temperature (Fig. 7). For strain MSK5 (a B31 derivative), an increase in the temperature did not enhance synthesis of the antigens tested (Fig. 7). A variable response was observed for strain AH130 (a 297 derivative) compared to MSK5, particularly for OspC and BBK32, indicating that multiple factors affect certain strains differently to alter the synthesis of the protein species.

CO₂ regulated gene expression.To determine the method of CO₂ regulation, the results of quantitative real-time RT-PCR of total RNA were analyzed for select genes of B. burgdorferi strains AH130 and MSK5. Total RNA from these strains was isolated from three independent anaerobic and nitrogen-treated anaerobic cultures to allow a comparison of the effect of dissolved CO₂ on gene expression independent of dissolved O₂ (Fig. 8). An endogenous, constitutively expressed gene, flaB, was used as a normalization control. The results showed that there was a dramatic change at the transcriptional level for BBK32, dbpA, and ospC in the presence of CO₂. Specifically, the BBK32, dbpA, and ospC transcript levels were 17.4-fold, 18-fold, and 93.2-fold greater, respectively, when AH130 was grown anaerobically in the presence of dissolved CO₂ than when cells were grown anaerobically without detectable dissolved CO₂. The MSK5 transcript levels of BBK32, dbpA, and ospC were increased 3.82-fold, 4.41-fold, and 8.48-fold, respectively, in anaerobic samples containing CO₂. The increases in gene expression observed in strain AH130 (a strain 297 derivative) compared to strain MSK5 (a strain B31 derivative), particularly for ospC, are consistent with previously published observations for comparisons of expression in these strains (53) and suggest that while there are absolute expression differences between these strains, the trends observed are similar (Fig. 8).

• Open in new tab
• Download powerpoint

FIG. 8.

Quantitative RT-PCR indicates that B. burgdorferi modulates gene expression at the transcriptional level when it is grown anaerobically in the presence of CO₂. The data show the ratio of the level of transcripts produced under anaerobic conditions with CO₂ present to the level of transcripts produced under anaerobic conditions without CO₂ for the genes indicated at the bottom and are the averages for three experiments, each done in triplicate. The error bars indicate standard errors. The results for strains AH130 (a 297 derivative) and MSK5 (a B31 derivative) are indicated by open bars and solid bars, respectively. The data for all samples were normalized to the data for an endogenous control, flaB, whose transcription was not affected by either of the conditions used in this experiment. The asterisks indicate that differences are statistically significant at a P value of <0.01.

The increased expression in the presence of CO₂ of lipoprotein genes (i.e., ospC, dbpA, and BBK32) involved in the adherence of B. burgdorferi to host structures suggests that CO₂ may serve as a cue for mammalian adaptation and that the regulation observed is at the transcriptional level. In addition to BBK32, dbpA, and ospC, two genes associated with the oxidative stress response (bosR and sodA) were also evaluated by RT-PCR, but the transcription of neither was enhanced greatly when cells were grown anaerobically with or without dissolved CO₂. Specifically, a less-than-twofold change was observed for bosR and sodA, although for bosR in strain MSK5 and for sodA in strain AH130 the increase in transcript production observed when CO₂ was present was statistically significant (Fig. 8). However, for bosR, for which there was great induction of protein production when CO₂ was present (Fig. 2, 4, and 7), the regulation observed appeared to be not linked to transcript production and instead may have been at the translational or posttranslational level.