|Home | About | Journals | Submit | Contact Us | Français|
Lyme disease is caused by the tick-borne spirochete, Borrelia burgdorferi. It has been documented that B. burgdorferi form aggregates within ticks and during in vitro culture. However, Borrelia aggregates remain poorly characterized, and their functional significance is unknown. Here we have characterized Borrelia aggregates using microscopy and flow cytometry. Borrelia aggregation was temperature, pH, and growth phase dependent. Environmental conditions (high temperature, low pH, and high cell density) favorable for aggregation were similar to the conditions that increased the expression of B. burgdoferi genes, such as outer surface protein C (ospC), that are regulated by the RpoN/RpoS sigma factors. Experiments were conducted to determine if there is a relationship between aggregation and gene regulation through the RpoN/RpoS pathway. ospC Transcript levels were similar between aggregates and free cells. Moreover, no differences were observed in aggregate formation when null mutants of rpoS, rpoN, or ospC were compared to wild-type spirochetes. These results indicated that, despite the similar external signals that promoted aggregation and the RpoN/RpoS pathway, the two processes were not linked at the molecular level. The methods developed here to study B. burgdorferi aggregates will be useful for further studies on spirochete aggregates.
Borrelia burgdorferi is a tick-borne spirochete that causes Lyme disease (Burgdorfer et al. 1982). Various investigators have reported that Borrelia make large aggregates and micro colonies within ticks and in culture (Burgdorfer et al. 1982, Barbour and Hayes 1986, Burgdorfer et al. 1989, Ewing et al. 1994, de Silva and Fikrig 1995, Sadziene et al. 1995, Shi et al. 1998, Gilmore and Piesman 2000, Shang et al. 2000, Tilly et al. 2000, Bugrysheva et al. 2002, Blevins et al. 2007). In ticks, aggregates were observed during and after the blood meal when the spirochetes grow to high densities in the tick's gut. In the gut Borrelia formed aggregates that were localized in the intercellular spaces between epithelial cells and within microvillar brush border cells of the gut (Burgdorfer et al. 1989).
Bacterial aggregation is involved in transmission, host colonization, exchange of genetic information, and protection from adverse environmental conditions (Kjaergaard et al. 2000, Wimpenny and Colasanti 2004, Prentice and Rahalison 2007). In Yersinia pestis, clumping of bacteria plays an important role in transmission from fleas to animals (Prentice and Rahalison 2007). Borrelia form aggregates during tick blood feeding, when they also get transmitted from tick vector to mammalian host. During this process, many B. burgdorferi genes required for transmission and host infection are expressed by a pathway regulated by the alternative sigma factors RpoN and RpoS. For example, the outer surface protein C (ospC) gene, which is required for transmission and host infection, is expressed by the RpoN/RpoS pathway. The timing of ospC expression within feeding ticks coincides with the formation of aggregates, indicating a possible relationship between aggregation and the RpoN/RpoS pathway. Here we report on studies that were done to characterize B. burgdorferi aggregates and to determine if aggregation was linked to the RpoN/RpoS pathway. Although the external signals for aggregation and activation of the RpoN/RpoS pathway are similar, our results demonstrate that the two processes are not linked at the molecular level.
A low passage culture of B. burgdorferi strain B31 (originally isolated from a tick in Shelter Island, NY) (Burgdorfer et al. 1982) was provided to us by the Centers for Disease Control and Prevention (Fort Collins, CO). The stock was cloned on solid BSK-H medium and named B31-C1. The clone was expanded in liquid BSK-H medium and frozen in aliquots that were used for this study. Strain 297 was obtained from Dr. Kayla Hagman at the University of Texas Southwestern Medical Center (Dallas, TX). Strain B31A3 and mutants ospC7, ospC7/ospC+4, and A3rpoS were obtained from Dr. Patricia A. Rosa, NIAID, Rocky Mountain Laboratories (Hamilton, MT) (Tilly et al. 2006), and strains A3-Gm, A3ntrA-Gm, ntrA-comp, and ntrA-VC were obtained from Dr. Frank Gherardini, NIAID, Rocky Mountain Laboratories (Fisher et al. 2005). Strains B31-CGFP and B31-FGFP were obtained from Dr. James Carroll, University of Pittsburgh (Pittsburgh, PA) (Carroll et al. 2003). B31-CGFP is B31 clone A3 harboring plasmid pBSVΦ(ospCp-gfp), and B31-FGFP is B31 clone A3 harboring plasmid pBSVΦ(flaBp-gfp). Plasmid pBSVΦ(ospCp-gfp) has gfp gene under control of ospC promoter, whereas in pBSVΦ(flaBp-gfp), gfp is under control of flaB promoter. Cultures were grown in BSK-H complete medium (Sigma, St. Louis, MO). The culture was grown at 35°C until the cells reached a density of 1×105 cells/mL. Cells were then transferred into fresh BSK-H complete medium for experiments. All cultures were vortexed before harvesting cells for counting or other analysis. Growth curves were generated for bacteria grown at 35°C by observing the bacteria by dark field microscopy at 20×and counting the organisms with a Petroff Hausser chamber (Hauser Scientific, Horsham, PA). Pictures of free cells and aggregates were taken using bright field microscopy at 20×. B. burgdorferi was cultured under different temperature and pH conditions. For these experiments, Borrelia culture was grown at 35°C until the cells reached a density of 1×105 cells/mL and subsequently transferred into fresh BSK-H complete medium at 35°C and pH 7.0 or 23°C and pH 8.0, buffered with 25mM HEPES. At equivalent cell densities of 5–7×107cells/mL, Borrelia were analyzed from both cultures.
Late log-phase (5×107 bacteria/mL) Borrelia culture was washed three times with PBS and fixed with 4% glutaraldehyde. Cell suspension was spread onto a poly-L-lysine–coated coverslip and incubated for 30min at room temperature. The coverslip was submerged in one change of PBS for 10min. Cells were then dehydrated by transferring through increasing concentrations of ethanol (30%, 50%, 75%, and anhydrous ethanol). Samples were transferred to the critical point dryer in 100% anhydrous ethanol. After the drying procedure, coverslips were mounted on 13mm aluminum planchettes using silver, and coverslips were coated with a 15–20nm thickness of gold palladium alloy (60:40) using a sputter coater or vacuum evaporator. Samples were observed using a Cambridge S200 scanning electron microscope.
To analyze the size and DNA content of Borrelia-free cells and aggregate populations, late log-phase (5×107 bacteria/mL) B. burgdorferi cultures grown at 35°C were incubated with Hoechst 33342 DNA stain (5μM) for 30min at room temperature. They were subsequently flow sorted to analyze forward scatter (FSC), side scatter (SSC), and Hoechst 33342 staining intensity. Flow cytometer data were acquired using Summit V.3.1 (Cytomation, Fort Collins, CO). A MoFlo Modular flow Cytometer (Cytomation) was used to sort the Borrelia-free cells and aggregates. Aggregates and free cells were collected based on their FSC and SSC values, where the mean values of (SSC, FSC) for free cells and aggregates were (39.11, 3.96) and (443.08, 102.65), respectively.
To analyze ospC promoter activity in single Borrelia cells, strain B31-CGFP was stained with Hoechst 33342 DNA stain (5μM) for 30min at room temperature. Cells were then analyzed on MoFlo Modular flow cytometer. Control strain B31-FGFP was subjected to identical treatment.
Flow-sorted cells were collected in lysis buffer containing PBS, 100mM DTT, RNAse inhibitor (RNAsin; Promega, Madison, WI), and DEPC water (Ambion, Austin, TX). Cells were then subjected to four freeze–thaw cycles. Genomic DNA in the cell lysates was removed using Turbo DNA-free kit (Ambion). Reverse transcription (RT) was performed using random primer (Invitrogen, Carlsbad, CA). Aliquotes from RT reaction mix were used for quantitative PCR using SYBR green master mix (Applied Biosystems, Foster City, CA). The flab cDNA was amplified using primers TTTCAGGGTCTCAAGCGTCT (forward) and TGTTGAGCTCCTTCCTGTTG (reverse), whereas ospC cDNA was amplified using primers GAAAGAGGTTGAAGCTTGC (forward) and ATTGCATAAGCTCCCGCTAA (reverse). Thermal cycles followed for amplification were 1 cycle at 95°C for 15min and 50 cycles at 95°C for 15s and 60°C for 1min using ABI Prism 7000 Sequence Detection System (Applied Biosystems). Melting curve was generated by following 95°C for 15s, 60°C for 20s, and 95°C for 15s. Relative quantitation of ospC transcripts was performed using comparative Ct method (ΔΔCt method) to calculate fold differences in ospC transcript between free cells and aggregates (Livak and Schmittgen 2001). Expression levels of ospC transcript (target) and flaB transcript (endogenous control) are evaluated. The levels of these amplicons in aggregates are compared with free cells (calibrator).
Initial experiments were done to characterize and quantify aggregate formation at different stages of growth in culture. B. burgdorferi strain B31-C1 was grown at 35°C in BSK-H complete medium and observed by light and dark field microscopy and scanning electron microscopy. At early log phase (4.3×106 cells/mL), almost all the spirochetes were present as individual cells (Fig. 1). As the cell density of the culture increased, the number of aggregates and the average size of each aggregate increased (Figs. 1 and 2A–D). A similar aggregation profile was also observed for B. burgdorferi strain RJ297 (data not shown). The earliest aggregates observed consisted of 2–6 spirochetes attached to each (Fig. 2 A, C). Most of the spirochetes in the aggregates appeared motile. When the culture reached a density 5×107 bacteria/mL, the average aggregate had greater than 20 bacteria (Fig. 2B, D). The center of each aggregate often appeared to have material without a clearly defined shape. Scanning electron microscopy demonstrated that this material was present toward the center of the aggregate and possibly holding the aggregate together (Fig. 2C, enlarged area). Once the culture reached stationary phase, the aggregates persisted, but the number of aggregates dropped and the average size of each aggregate also decreased (Figs. 1 and and2E2E).
Experiments were conducted to determine if FACS could be used to separate aggregates and free cells. We predicted that owing to the large size of aggregates, they will have higher FSC and SSC than the free cells. When a late log-phase culture was analyzed for the FSC and SSC properties of the bacteria, the majority of cells had low FSC and intermediate SSC (Fig. 3A, sector R2). A smaller population showed increasingly higher FSC and SSC values. We confirmed that the bacterial populations with lower scatter values were free cells (Fig. 3A, sector R2), while the ones with higher scatter values were aggregates by observing the bacteria by dark field microscopy (data not shown) and fluorescence microscopy (Fig. 3A, sector R24, R25). Further, flow analysis of Borrelia culture stained with Hoechst 33342 (which binds to DNA) showed that the FSC values of aggregates corresponded to their sizes because there was a strong positive correlation between FSC and the DNA content in each event (Fig. 3B).
At equivalent cell densities of 5–7×107 cells/mL, Borrelia grown at 35°C and pH 7.0 showed significantly higher aggregation than cultures grown at 23°C and pH 8.0 (Fig. 4). Larger aggregates were conspicuously missing in the cultures grown at low temperatures. Both sets of culture tubes were buffered with 25mM HEPES, and at the end of the experiment no appreciable change in pH was observed from the starting pH (data not shown).
The environmental conditions (high temperature, low pH, and high cell density) favorable for aggregation were similar to the conditions that increased the expression of B. burgdoferi genes, such as ospC, that are regulated by the RpoN/RpoS sigma factors (Schwan et al. 1995, Carroll et al. 1999, Hubner et al. 2001, Yang et al. 2005). Experiments were done to compare ospC expression per cell in aggregated versus free spirochetes using quantitative RT-PCR. Spirochetes were sorted into free cells and aggregates. RNA was purified from each population; ospC and flaB transcript levels were estimated by quantitative RT-PCR (Table 1). The ratio of ospC to flaB transcript levels was compared for the two populations, and ospC mRNA levels were found to be similar in free cells and aggregates. The fact that transcription from the ospC promoter was similar in free cells and aggregated cells was confirmed using mutant B31-CGFP, transfected with green fluorescent protein fused to the OspC promoter. A culture of ospCp-GFP mutant was grown under conditions (35°C/cell density at 1×107 cells/mL) that favor ospC expression, and stained with Hoechst 33342 to stain DNA of Borrelia cells. During flow cytometry, GFP intensity and Hoecht 33342 staining intensity of free cells and aggregates were measured. Ratio of GFP/Hoechst staining intensities showed ospC promoter activity per cell in free cells and aggregates (Table 2). Aggregates showed only 1.8-fold more GFP expression per cell compared to free cells. Aggregates of flaBp-GFP mutant showed 1.4-fold more GFP expression compared to free cells (data not shown). Thus, cells within aggregates showed insignificant increase in ospC promoter activity compared to free cells.
To further explore if aggregation and the RpoN/RpoS pathways were linked, experiments were done with null mutants of ntrA (coding for RpoN), rpoS, and ospC. The wild-type and ospC knockout strains aggregated to similar levels (WT 1.24% and ΔospC 1.91% of cells in aggregates), and the aggregates that formed in wild-type and ospC-disrupted strains were structurally similar (data not shown). RpoS and RpoN together and independently regulate OspC and a large number of other genes and cellular processes (Hubner et al. 2001, Yang et al. 2005). Experiments were conducted with rpoS and ntrA (coding for RpoN) knockout mutants to determine if these genes influenced aggregate formation. Both rpoS and ntrA mutants aggregated to similar levels as wild type (WT 1.24%, ΔntrA 2.03%, and ΔrpoS 2.04% of cells in aggregates), indicating that Borrelia aggregation was not dependent on genes expressed by RpoN or RpoS (data not shown).
Investigators studying Borrelia have long known that these organisms often aggregate, but this phenomenon has not been systematically studied. Given the growing appreciation that bacterial aggregation is of biological significance, Borrelia aggregation deserves further study. B. burgdorferi persists after antibiotic treatment in mice (Hodzic et al. 2008), and there are conflicting reports about the effectiveness of antibiotic therapy in human Lyme disease (Moody et al. 1994, Nowakowski et al. 1995, Wang et al. 1998). In Pseudomonas aeruginosa aggregation has been linked to persistence and antibiotic resistance in cystic fibrosis patients (Drenkard and Ausubel 2002). Further studies are needed to explore the role of Borrelia aggregates in persistence both during normal infection and after antibiotic treatment. Here we demonstrate that high temperature, low pH, and high cell density enhance Borrelia aggregate formation in vitro. Previous studies have demonstrated that high cell density, low pH, and high temperature cause expression of ospC and other Borrelia genes controlled by the global transcriptional regulators RpoS and RpoN (Schwan et al. 1995, Carroll et al. 1999, Elias et al. 2000, Yang et al. 2000, 2005, Hubner et al. 2001, Ramamoorthy and Scholl-Meeker 2001, Fisher et al. 2005, Burtnick et al. 2007, Smith et al. 2007). However, the results reported here demonstrate that spirochete aggregation is not linked to ospC or other genes regulated by the RpoN/RpoS pathway. These results corroborate an earlier report that showed that Borrelia mutant B313 strain that does not express OspA, B, C, or D formed microscopic and macroscopic aggregates (Sadziene et al. 1993, 1995). B313 strain of Borrelia has only 6 out of 21 plasmids. It lacks all plasmids except cp32-1, cp32-2, cp32-3, cp32-4, cp26, and lp17, indicating that gene responsible for Borrelia aggregation may be in either these plasmid(s) or in the genome.
It may be possible to correlate gene expression and protein production in an aggregate, although such an analysis would require estimation of live and dead cells, within an aggregate. For example, in live cells, ospC transcript and proportional OspC protein can be detected. However, in cells within aggregates that are not alive, amount of OspC protein may not be proportional to ospC transcript, owing to shorter half-life of mRNA than protein. Using GFP reporter system may be more accurate, as cells showing active transcription of ospC would synthesize both GFP and OspC protein.
We and others have used flow cytometry to analyze protein and gene expression by spirochetes grown in culture (Caimano et al. 2005, Bakker et al. 2007, Motaleb et al. 2007, Srivastava and de Silva 2008). Here we also demonstrate the utility of flow cytometry to compare the properties of individual or aggregated spirochetes. With flow cytometry it was also possible to separate aggregates of different size. While the current study has demonstrated that the RpoN/RpoS pathway is not linked to aggregation, the tools developed here can be applied to further characterize aggregates and to evaluate their biological significance. This method could also be used to analyze difference in expression of gene(s) between Borrelia populations with different physical characteristics such as nonmotile cystic forms and vegetative spirochetes, within the same culture (Murgia et al. 2002).
In summary, we have shown that high temperature, low pH, and high cell density enhance Borrelia aggregate formation in vitro. However, spirochete aggregation is not linked to ospC or other genes regulated by the RpoN/RpoS pathway.
This work was supported by Public Health Service Grant ROI AR47948 from the National Institute for Arthritis and Musculoskeletal and Skin Diseases. We thank Patricia Rosa, Frank Gherardini (NIAID), and Jay Carroll (University of Pittsburgh) for sending us Borrelia strains. Larry Arnold and Nancy Martin (UNC—Chapel Hill) assisted us with establishing the FACS assays for studying Borrelia. We also thank members of the de Silva laboratory for their advice.
Siddharth Y. Srivastava: No competing financial interests exist. Aravinda M. de Silva: No competing financial interests exist.