|Home | About | Journals | Submit | Contact Us | Français|
Borrelia burgdorferi, the causative agent of Lyme disease, exists in a complex enzootic cycle, transiting between its vector, Ixodes ticks, and a diverse range of vertebrate hosts. B. burgdorferi linear plasmid 38 (lp38) contains several genes that are differentially regulated in response to conditions mimicking the tick or mouse environments, suggesting that these plasmid-borne genes may encode proteins important for the B. burgdorferi infectious cycle. Some of these genes encode potential virulence factors, including hypothetical lipoproteins as well as a putative membrane transport system. To characterize the role of lp38 in the B. burgdorferi infectious cycle, we constructed a shuttle vector to selectively displace lp38 from the B. burgdorferi genome and analyzed the resulting clones to confirm the loss of lp38. We found that, in vitro, clones lacking lp38 were similar to isogenic wild-type bacteria, both in growth rate and in antigenic protein production. We analyzed these strains in an experimental mouse-tick infectious cycle, and our results demonstrate that clones lacking lp38 are fully infectious in a mouse, can efficiently colonize the tick vector, and are readily transmitted to a naive host.
Borrelia burgdorferi is the causative agent of Lyme disease, the most common tick-transmitted disease in the United States (52). B. burgdorferi transits between its vector, Ixodes ticks, and a diverse range of mammalian hosts (5, 11, 33, 34, 36, 41, 42). The unique B. burgdorferi genome, encoding factors required for survival in both vector and host, consists of a single linear chromosome and a large complement of linear and circular plasmids, ranging in size from 5 kb to 56 kb (2, 3, 15–17, 20, 21, 26, 46, 51, 53).
To successfully establish infection and undergo subsequent acquisition and transmission by its vector, B. burgdorferi must adapt to complex, alternating host environments. Numerous studies have demonstrated a direct link between B. burgdorferi plasmid content and the ability to establish infection and persist in the enzootic cycle. Loss of plasmid elements during in vitro passage or due to selective displacement of individual plasmids can result in a decrease or loss of infectivity in mice or of the ability to survive within the tick vector (23, 25, 28, 32, 37, 43, 49, 56, 59, 60, 66).
Several rigorous studies have analyzed the plasmid content of B. burgdorferi and identified key roles that several plasmid-borne genes play in establishing infection. However, several plasmids have not been characterized for their contribution to the B. burgdorferi infectious cycle, including the 38.8-kb linear plasmid, lp38, of the type strain B31. This plasmid is absent in some passaged laboratory strains and is missing in some human clinical and field isolates, suggesting that it does not carry any essential genes when grown under laboratory conditions (27, 61, 62). When lp38 is present in a strain, most of its sequences appear to be conserved, as assessed by genomic array hybridization (61).
Several lp38 genes are differentially regulated in vitro in response to exposure to blood or changes in temperature or pH or within dialysis membrane chambers implanted in rats (10, 39, 45, 64). These differentially regulated genes include bbj08 and bbj31, which encode uncharacterized theoretical lipoproteins, and bbj26 and bbj27, which encode a putative ABC membrane transport system. Eleven genes on lp38, including bbj24, bbj25, and bbj26, exhibit differential expression levels in a BosR deletion strain, suggesting a role in virulence or in the oxidative stress response (40). With a transposon mutagenesis screen, one study attempting to identify B. burgdorferi virulence determinants identified a clone that lacked only lp38 and contained a transposon insertion in circular plasmid 9 (cp9), which is dispensable for establishing infection in a mouse. Interestingly, this transposon mutant showed reduced infectivity in a mouse model, suggesting that the loss of genes on lp38 was responsible for the reduced-infectivity phenotype (9).
To better understand the role of this linear plasmid in the B. burgdorferi infectious cycle, we selectively displaced lp38 from two infectious B. burgdorferi strains. We displaced lp38 from B31-A3, a fully infectious wild-type clone of the sequenced B31 strain that has been characterized genetically, as well as phenotypically, in the experimental mouse-tick cycle (7, 19, 22, 25, 29, 63). We also utilized B31-A3-68Δbbe02, a fully infectious derivative of B31-A3 that lacks lp56 and contains an insertion/deletion on lp25 that disrupts bbe02, a restriction modification locus, thereby enhancing shuttle vector transformation efficiency approximately 1,000-fold (44). After confirming displacement of lp38 from these strains, we compared the in vitro-growth phenotypes of isogenic wild-type and lp38-deficient clones and determined their competence in the B. burgdorferi experimental mouse-tick cycle.
B. burgdorferi strains B31-A3 and B31-A3-68Δbbe02 have been described previously (19, 44). Briefly, the bbe02 locus on lp25 of B31-A3-68Δbbe02 was inactivated by insertion of the aadA cassette, which confers resistance to streptomycin and spectinomycin, and deletion of 1,250 bp (44); in this study, we designate this strain B31-S9. All B. burgdorferi strains were grown in Barbour-Stoenner-Kelly II (BSKII) medium as previously described (8) and plated in solid BSK medium. Gentamicin was used at 40 μg/ml and streptomycin at 50 μg/ml, where appropriate.
Figure 1 is a schematic diagram of the individual steps involved in shuttle vector construction. All primers are listed in Table 1 . Shuttle vector pBSV38G was constructed by PCR amplification with primers A and B and with pBSV2G as the template DNA, followed by SpeI digestion and ligation using T4 DNA ligase, creating plasmid pOG1. Plasmid pOG1 contains a ColE1 origin of replication for propagation in Escherichia coli, a multiple cloning site (MCS), and an flgB promoter driving expression of a gentamicin resistance cassette. Plasmid pOG1 was used as a backbone vector for introduction and testing of B. burgdorferi replication and segregation genes. Using primers C and D, the 3.6-kb region consisting of the genes putatively responsible for the autonomous replication of lp38, namely, bbj16, bbj17, bbj18, and bbj19, was PCR amplified from B31-A3 genomic DNA that had been purified using a Wizard total genomic DNA kit (Promega Corporation, Madison, WI) with KOD polymerase (EMD Biosciences, San Diego, CA). This product was cloned into PCR-XL-Topo (Invitrogen, Carlsbad, CA), creating PCR-XL-Aut38, and subsequently digested with SpeI. The pOG1 vector was digested with SpeI and XbaI, followed by ligation with the 3.6-kb fragment excised from PCR-XL-Aut38 (Fig. 1). All DNA manipulations were performed in E. coli Top10 (Invitrogen), and pertinent plasmid sequences were confirmed by DNA sequencing. All enzymes were from New England BioLabs (NEB; Ipswich, MA) unless otherwise noted.
B. burgdorferi strains B31-A3 and B31-S9 were grown to 5 × 107 spirochetes/ml and electroporated with 10 μg of pBSV38G as previously described (19, 48). Electroporated B. burgdorferi cells were immediately resuspended in 5 ml of BSKII medium and allowed to recover for 24 h before being plated in solid BSK medium containing gentamicin at 40 μg/ml and streptomycin at 50 μg/ml, where appropriate. Transformants were screened by PCR for the presence of aacC1, the gene encoding gentamicin resistance, and of lp38 by use of primer pairs E-F and G-H, respectively (Table 1). Transformants containing the shuttle vector and lacking lp38 were inoculated into 10 ml of BSKII medium containing gentamicin and grown to late exponential phase. Total genomic DNA was purified using a total genomic DNA purification kit (Qiagen, Valencia, CA) and used as a template in PCRs to determine plasmid content, as previously described (43).
B. burgdorferi strains were grown in 10 ml of BSKII medium to late exponential phase and harvested by centrifugation. Total genomic DNA was isolated using a Wizard total genomic DNA kit (Promega). Southern blot probes for aacC1 and bbj08 were labeled with digoxigenin (DIG) using a PCR DIG probe synthesis kit (Roche, Indianapolis, IN) with primer pairs E-F for aacC1 and G-H for bbj08 (Table 1). Total genomic DNA, along with a λ DNA-Mono Cut Mix (NEB) DNA ladder, was resolved on 0.7% agarose gels at 50 V overnight or run at 70 V for 36 h using field inversion gel electrophoresis (12) with a PPI-200 system from MJ Research (Hercules, CA) on program 2, as set by the manufacturer. The gels were depurinated and denatured, and the genomic DNA was transferred to a nylon membrane (GE Biosciences, Piscataway, NJ) using capillary transfer and UV cross-linked to the membrane with a Stratalinker 1800 (Stratagene, Indianapolis, IN). Southern blot probes were hybridized and washed as previously described (47, 55). Southern blot probes were detected using a DIG luminescence detection kit (Roche) and X-ray film.
To determine growth rate, strains B31-A3, B31-S9, B31-A3Δlp38/pBSV38G, and B31-S9Δlp38/pBSV38G were inoculated into BSKII medium from frozen stocks and grown to ~5 × 107 spirochetes/ml. The cultures were then diluted to 5 × 103 spirochetes/ml in 7 ml of BSKII medium in triplicate, and culture density was determined at 24-h time points using a Petroff-Hausser chamber and dark-field microscopy.
Antigenic profiles of strains B31-A3, B31-S9, B31-A3Δlp38/pBSV38G, and B31-S9Δlp38/pBSV38G were determined by inoculating strains into BSKII medium from frozen stocks and growing them to ~5 × 108 spirochetes/ml. The cultures were diluted to 5 × 106 spirochetes/ml in 5 ml of buffered BSKII medium in which the pH had been adjusted to pH 6.8 or pH 7.5 and incubated at either 25°C or 35°C to mimic the tick or mammalian environment, respectively. Equivalent numbers of spirochetes were harvested from each culture condition based on enumerations of spirochetes in a Petroff-Hausser chamber, resuspended in Laemmli loading buffer, and resolved on 12% SDS-PAGE gels (47). Then, either gels were stained with Coomassie brilliant blue or proteins were transferred to a nitrocellulose membrane for immunoblotting (1). Membranes were incubated in a 4% nonfat milk solution (Lab Scientific, Livingston, NJ) for 1 h at room temperature and transferred to a 1% solution of nonfat milk containing a polyclonal rabbit anti-OspD serum (1:1,000) (55), a mouse monoclonal flagellin antibody (H9724, 1:1,000) (4), or pooled sera from mice infected with B31-A3 (1:200), for 1 h at room temperature with shaking. Immunoblots were then incubated in Tris-buffered saline containing 0.1% Tween 20 with an appropriate peroxidase-conjugated secondary antibody. For detection of OspD, a goat anti-rabbit, peroxidase-conjugated antibody (1:1,000) (Sigma-Aldrich) was used. For detection of monoclonal H9724 or pooled mouse sera, a goat anti-mouse, peroxidase-conjugated antibody (1:10,000) (Sigma-Aldrich, St. Louis, MO) was used. Following 1 h of incubation at room temperature, immunoblots were developed using a SuperSignal West Pico chemiluminescent substrate kit (Thermo Scientific, Rockford, IL).
Mouse infection studies were carried out in accordance with guidelines of the National Institutes of Health. All animal work was done according to protocols approved by the Rocky Mountain Laboratories Animal Care and Use Committee. The Rocky Mountain Laboratories are accredited by the International Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC). Studies were done 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 retro-orbitally bled prior to injection, and an inoculum was prepared by enumerating spirochetes using a Petroff-Hausser chamber. A portion of the inoculum was plated in BSK medium, and numbers of CFU were determined to confirm the number of spirochetes in the inoculum (Table 2) and to determine the plating efficiency for each strain. Spirochetes were injected intraperitoneally (4 × 103 spirochetes) and subcutaneously (1 × 103 spirochetes), mice were bled 3 weeks postinoculation to assess seroreactivity to B. burgdorferi proteins, and ear punch biopsy specimens (3 mm) were taken and placed into BSKII medium to assess the presence of spirochetes in these tissues. Approximately 100 Ixodes scapularis larvae were subsequently fed to repletion on mice infected with B. burgdorferi strain B31-S9Δlp38/pBSV38G. A subset of fed larvae was mechanically disrupted and plated in solid BSK medium 10 days postfeeding to assess the acquisition of B. burgdorferi from infected mice and to enumerate spirochetes. The remaining fed I. scapularis larvae were allowed to molt to nymphs (approximately 10 weeks after larval feeding), and then 10 ticks/mouse were fed on 9 naive RML mice to assess the persistence of B. burgdorferi mutant strains in ticks and the transmission to mice.
To directly address the role of lp38 in the B. burgdorferi infectious cycle, we sought to create a clone lacking lp38 while retaining the parental complement of plasmids known to be required for virulence. Due to the complexity of the B. burgdorferi genome, it is essential to investigate plasmids individually to evaluate the phenotypic effect on the bacterium, thereby allowing any resultant phenotype to be traceable to a single genetic element. Previous work has demonstrated that B. burgdorferi plasmids can be selectively displaced by the introduction of an incompatible Borrelia shuttle vector (6, 18, 23, 28, 57, 58). Using this strategy, we constructed a shuttle vector, pBSV38G, containing the lp38 genes putatively responsible for plasmid retention and incompatibility in order to selectively displace lp38 from B. burgdorferi.
The putative plasmid partitioning and replication genes on lp38 were identified by Fraser et al. (21) as bbj16, bbj17, bbj18, and bbj19, belonging to the paralogous gene families 49, 32, 50, and 62, respectively. These genes were amplified by PCR and cloned into plasmid pOG1, which contains a multiple cloning site (MCS), a ColE1 origin of replication for propagation in E. coli, and a gentamicin resistance cassette for E. coli and B. burgdorferi (Fig. 1). The resulting plasmid, pBSV38G, was electroporated into two low-passage-number infectious clones, B31-A3 and B31-S9. Transformants were selected in solid BSK medium containing gentamicin and screened by PCR for the presence of the aacC1 gene, encoding gentamicin resistance, and of bbj09, located on lp38, as well as for the full parental complement of B. burgdorferi plasmids. All 108 clones screened by PCR were positive for the aacC1 gene, and all 108 had lost lp38. Some pBSV38G transformants had lost other B. burgdorferi plasmids, but most retained a full complement of plasmids except for lp38. Those clones positive for aacC1 and negative for lp38, while retaining all of the parental plasmids, were selected for further analysis.
The introduction of an incompatible shuttle vector into B. burgdorferi can result in several possible genetic outcomes, including coreplication of the B. burgdorferi shuttle vector with the endogenous plasmid, integration of the shuttle vector into the genome, or selective displacement of the targeted plasmid. Although PCR screening suggested that lp38 had been displaced, several clones were selected for further characterization by Southern blot analysis.
To confirm the presence and autonomous replication of pBSV38G in transformants, an aacC1-specific probe was hybridized to total genomic DNA isolated from selected transformants. Southern blot data demonstrated that the aacC1 probe hybridized with genomic DNA isolated from clones transformed with pBSV38G (Fig. 2 A). This blot also demonstrated that the electrophoretic mobility of pBSV38G from the transformants was similar to that of pBSV38G plasmid DNA isolated from E. coli, indicating that pBSV38G was replicating as an autonomous plasmid and had not integrated into the genome. The pBSV38G plasmid isolated from E. coli was readily detectable in both nicked (Fig. 2A, upper band) and supercoiled (Fig. 2A, lower band) forms, and the pBSV38G plasmid detected in the clones was mostly in the nicked form. We were able to detect pBSV38G in our B. burgdorferi clones in the supercoiled form but only after a much longer exposure (data not shown).
To confirm that lp38 had been selectively displaced, we also analyzed the genomic DNA from pBSV38G transformants with a probe specific for the bbj08 locus on lp38. This gene is located outside the region required for autonomous replication of lp38 and is not present on pBSV38G. A hybridizing band indicated the presence of bbj08 in the parental B31-A3 and the B31-S9 strain but not in transformants that contain pBSV38G (Fig. 2B), indicating that these clones no longer carry lp38. Taken together, these data demonstrate that pBSV38G is an autonomously replicating plasmid capable of displacing the endogenous lp38.
To further confirm lp38 displacement by pBSV38G, we analyzed the production of outer surface lipoprotein D (OspD). The ospD gene, bbj09, located on lp38, is highly upregulated in the tick yet dispensable for successful completion of the experimental B. burgdorferi infectious cycle (35, 55). We analyzed the B31-A3/pBSV38G and B31-S9/pBSV38G transformants and their parental strains by immunoblotting with antiserum that recognizes OspD. Those strains that contain pBSV38G no longer produce OspD protein, in contrast to their parental strains (Fig. 3). An antibody that recognizes flagellin was included as a loading control. These results, together with the Southern blotting results, made it clear that lp38 had been successfully displaced by pBSV38G in both low-passage-number B31-A3 and B31-S9 clones transformed with pBSV38G.
Having established that lp38 had been selectively displaced in these transformants, we next screened for phenotypic changes in lp38-deficient clones. We first compared the growth rates of B31-A3Δlp38/pBSV38G and B31-S9Δlp38/pBSV38G to those of their parental strains in BSKII medium. We found no significant difference in the growth rates of clones lacking lp38 and of their parental strains (Fig. 4) or any difference in the plating efficiencies of these strains, which were all greater than 85% (data not shown). We also did not observe any morphological differences in strains lacking lp38 when they were viewed by dark-field microscopy (data not shown).
During in vitro growth, B. burgdorferi responds to changes in pH and temperature, with concomitant changes in gene expression and protein profile (10, 13, 14, 45, 50, 54, 64). These changes in gene expression are thought to reflect adaptations that occur while B. burgdorferi resides in the tick vector or mammalian host. To detect changes in gene expression that might be affected by the loss of lp38, we incubated strains B31-A3, B31-S9, B31-A3Δlp38/pBSV38G, and B31-S9Δlp38/pBSV38G in BSKII medium at pH 6.8 to mimic tick-like conditions and at pH 7.5 to mimic mammal-like conditions. We analyzed the protein lysates of these strains by Coomassie brilliant blue staining but were unable to detect any significant differences (Fig. 5 A). We also analyzed the expression of OspD using OspD antiserum. As expected, we detected OspD in wild-type strains grown at pH 6.8 or pH 7.5 but not in strains lacking lp38 (data not shown). To further compare these protein lysates, we used immunoblotting and pooled sera from mice infected with wild-type B31-A3 to determine the antigenic proteins synthesized by spirochetes grown at pH 6.8 and pH 7.5. As anticipated, distinct differences were present in antigenic proteins synthesized by spirochetes grown at pH 6.8 versus pH 7.5 (Fig. 5b), including a group of proteins in the range of 5 to 25 kDa. We also incubated spirochetes at pH 7.5 at 25°C and analyzed the protein lysates of those strains for changes in antigenic proteins (data not shown). Although previous microarray data suggest differential regulation of lp38 genes in response to changes in pH and temperature, we did not detect any significant differences by Coomassie brilliant blue staining or immunoblot analysis between isogenic clones lacking lp38 and those containing lp38 (Fig. 5 and data not shown). These data suggest that there are no major antigenic proteins encoded by lp38 that are differentially expressed at pH 6.8 versus pH 7.5 or at 35°C versus 25°C in vitro. This also demonstrates that the loss of lp38 does not affect the ability of B. burgdorferi to alter expression of some antigenic proteins in response to changes in pH or temperature.
We next used these strains to analyze the role of lp38 in a mouse model of infection. Groups of four mice per strain were challenged with B31-A3, B31-S9, or B31-A3Δlp38/pBSV38G or challenged individually with both independent B31-S9Δlp38/pBSV38G clones by needle inoculation (Table 2). Infectious status was assessed 3 weeks postinoculation by analyzing the sera of the inoculated mice for seroconversion to B. burgdorferi whole-cell lysates and by attempted reisolation of spirochetes from ear punch biopsy specimens. This initial characterization was then followed by attempted reisolation of spirochetes at 3 weeks postinoculation from the ear, bladder, and ankle joint to assess dissemination. We found that 11 out of the 12 mice inoculated with a strain lacking lp38 developed a serological response, and spirochetes were successfully reisolated from all of the tissues tested (Table 2). Spirochetes were reisolated from only 1 out of 4 mice inoculated with the wild-type B31-A3 strain, which may be due to the lower number of spirochetes in the A3 inoculum (Table 2). It is clear, however, that clones lacking lp38 in both the B31-A3 and B31-S9 backgrounds are fully proficient to establish infection, persist in a mouse, and disseminate to distant tissues when mice are infected by this route, indicating that lp38 is not required for these processes.
Two critically important steps in the B. burgdorferi enzootic cycle are the acquisition of B. burgdorferi by its vector, Ixodes ticks, and the subsequent transmission to a mammalian host. To further analyze the role of lp38 in the B. burgdorferi infectious cycle, we analyzed the acquisition of B31-S9Δlp38/pBSV38G by feeding approximately 100 naive I. scapularis larvae to repletion on each of 6 mice infected with the B. burgdorferi B31-S9Δlp38/pBSV38G strain. Ten days postfeeding, a subset of the larvae was analyzed for viable spirochetes by mechanically disrupting fed ticks and then plating in solid BSK media. As shown in Fig. 6 A, larval ticks acquired spirochetes from 5 of the 6 mice that were infected with the B. burgdorferi B31-S9Δlp38/pBSV38G strain. The spirochete load in the fed larvae was highly variable, ranging from 8 to 2,800 spirochetes per tick. These values are similar to what has been reported for wild-type B31-A3, ranging from a few spirochetes to ~104 spirochetes per larval tick that had fed on an infected mouse (7, 29, 30). Although the spirochete load per tick was variable, spirochetes lacking lp38 were readily acquired by I. scapularis larvae that fed on infected mice.
Having established that the B31-S9Δlp38/pBSV38G strain was acquired by I. scapularis larvae, we next analyzed the ability of these ticks to transmit B. burgdorferi B31-S9Δlp38/pBSV38G and infect a naive host. To this end, the infected larvae were allowed to molt into nymphs, and then 10 ticks per mouse were fed to repletion on 9 naive mice. Ticks were mechanically disrupted and plated 10 days postfeeding to determine spirochete load and to confirm survival of the spirochetes through the tick molt. The number of spirochetes per nymph was variable and averaged between ~103 and 104 spirochetes per tick (Fig. 6B). These data are similar to the average range of 103 to 105 spirochetes per tick that has been reported for wild-type B. burgdorferi in infected I. scapularis nymphs (7, 29, 30, 55) and indicate that all 9 naive mice were fed upon by infected ticks. Three weeks after the tick feeding, mice were assessed for seroreactivity to B. burgdorferi proteins, and reisolation of spirochetes from the ear, bladder, and joint tissues was attempted. All mice (9/9) seroconverted, and spirochetes were reisolated from all of the tissues tested (27/27). These data indicate that B. burgdorferi strains lacking lp38 efficiently persist through the tick molt and replicate following exposure to a blood meal. These data also clearly demonstrate that spirochetes lacking lp38 are readily transmitted by I. scapularis nymphs and are fully infectious when inoculated by a natural route.
The role of extrachromosomal elements in virulence has been an area of intense research in B. burgdorferi, as well as in other pathogenic bacteria. In B. burgdorferi, several plasmids that are required to complete the infectious cycle have been identified. We have developed a B. burgdorferi shuttle vector, pBSV38G, using the genes putatively responsible for lp38 autonomous replication and here show that the shuttle vector autonomously replicates, is incompatible with lp38, and effectively displaces lp38 from low-passage-number infectious B. burgdorferi isolates. We have analyzed these lp38-deficient clones in vitro for phenotypic changes associated with the loss of this plasmid, as well as in vivo to determine the role of this plasmid in the infectious cycle.
In vitro, we were unable to detect differences in growth rate, protein profile, or production of antigenic proteins between strains lacking lp38 and their parental strains. Furthermore, our data indicate that both the B31-A3 and B31-S9 strains lacking lp38 are fully infectious in a mouse model by needle inoculation and that loss of lp38 does not affect the ability of spirochetes to disseminate from the inoculation site to distant tissues. Additionally, spirochetes lacking lp38 are readily acquired from the skin of infected mice by feeding I. scapularis larvae, survive through the molt, and can establish infection when transmitted by feeding nymphs. Our use of naturally infected ticks as a route of transmission also suggests that strains lacking lp38 are competent to establish infection at a low dose.
We find these results surprising for several reasons. The low sequence divergence of lp38 among strains suggests that these genes have been conserved for a functional purpose (61). In vivo, several lp38 genes are upregulated when B. burgdorferi resides in the skin of a mouse (31). Many of these genes are also differentially regulated in response to conditions mimicking those found in vivo, suggesting a possible role in survival within the tick vector or mammalian host (10, 31, 45, 64). Some of these genes, including the hypothetical lipoprotein genes bbj08 and bbj31 and components of a putative ABC transport system composed of bbj26 and bbj27, represent possible virulence determinants. We were, however, unable to detect differences in the antigenic protein profiles of wild-type strains from those lacking lp38. This suggests that either the upregulated lp38 genes do not produce sufficient protein to elicit an immune response in vivo, the proteins are not antigenic, or these conditions do not result in a protein expression profile by in vitro-grown spirochetes that adequately reflects what occurs in the tick or mammalian host environments. Although we could detect OspD in wild-type strains but not in strains lacking lp38, OspD does not elicit an immune response in infected mice and therefore is not detected by our assay (55). Furthermore, our results do not support the preliminary findings of Botkin et al. (9), in whose study a clone lacking lp38 exhibited a significant decrease in mouse infectivity. Our data clearly demonstrate that all of the genes on lp38 located outside the region required for autonomous replication are dispensable for B. burgdorferi to complete its infectious cycle.
These results raise interesting questions regarding lp38. If this plasmid is dispensable for completion of the B. burgdorferi infectious cycle, why has it been retained in nature? What role do lp38-borne genes play in the B. burgdorferi life cycle? This plasmid has been shown to be retained in B. burgdorferi after 25 passages at both 25°C and 35°C (24), and in a study analyzing the plasmid contents of 19 clonal isolates, lp38 was present in all of the strains tested (43). In a recent study analyzing 44 low-passage-number B. burgdorferi clones and 4,464 B. burgdorferi transformants, all 44 clones retained lp38 and only 1.1% of the transformants had lost lp38 (38). Human clinical and field isolates that lack lp38 have been identified, although it is unclear if these isolates originally carried lp38 and lost it during the course of isolation. Complicating this genotypic heterogeneity is the inability of some of these isolates to establish infection when reintroduced in a mouse model (65). This highlights the utility of selectively displacing an individual plasmid from a defined genetic background in order to link a phenotype with a specific genetic element.
We have examined a selected set of variables, including growth rate, protein and antigenic profile, the ability to infect a mouse by both artificial and natural routes of infection, and the ability to survive within a tick. We were unable to find a significant phenotype under any of the conditions tested. However, BSK medium used to grow B. burgdorferi is complex and may mask differences that would be detectable under more stringent growth conditions, perhaps in a more minimal medium. It is also possible that in a natural host, such as Peromyscus, the retention of lp38 might be required to establish and maintain infection. In addition, neither seroreactivity nor reisolation of spirochetes from various tissues adequately gauges the pathogenicity of a strain; they gauge only whether the spirochetes have elicited an immune response from the host and are proficient at disseminating to distant tissues. It is possible that within the ecology of the natural infectious cycle, which includes different host species, mixed infections, and varying physical environments, conditions under which spirochetes carrying lp38 have a selective advantage may exist. However, within the context of this experimental mouse-tick infectious cycle, we conclude that all the genes on lp38 can be added to the growing list of conserved and regulated yet expendable elements of the B. burgdorferi genome.
We thank Jean Celli, Anders Omsland, Tom Schwan, and all members of the Rosa lab for critical reading of the manuscript and Anita Mora and Austin Athman for assistance with figures.
This research was supported by the Intramural Research Program of the NIH, NIAID.
Published ahead of print on 27 June 2011.