Outer Surface Protein C Is a Dissemination-Facilitating Factor of Borrelia burgdorferi during Mammalian Infection

Background The Lyme disease spirochete Borrelia burgdorferi dramatically upregulates outer surface protein C (OspC) in response to fresh bloodmeal during transmission from the tick vector to a mammal, and abundantly produces the antigen during early infection. As OspC is an effective immune target, to evade the immune system B. burgdorferi downregulates the antigen once the anti-OspC humoral response has developed, suggesting an important role for OspC during early infection. Methodology/Principal Findings In this study, a borrelial mutant producing an OspC antigen with a 5-amino-acid deletion was generated. The deletion didn't significantly increase the 50% infectious dose or reduce the tissue bacterial burden during infection of the murine host, indicating that the truncated OspC can effectively protect B. burgdorferi against innate elimination. However, the deletion greatly impaired the ability of B. burgdorferi to disseminate to remote tissues after inoculation into mice. Conclusions/Significance The study indicates that OspC plays an important role in dissemination of B. burgdorferi during mammalian infection.


Introduction
Coordinating production of outer surface proteins (Osps) is crucial for the pathogenic strategy of the Lyme disease spirochete, Borrelia burgdorferi [1]. In engorged and unfed ticks, B. burgdorferi abundantly produces OspA and OspB, but no OspC [2,3]. In response to a fresh bloodmeal, the pathogen dramatically upregulates OspC and other RpoS-dependent genes to prepare itself for infection of mammals [4][5][6]. B. burgdorferi maintains high OspC synthesis during early mammalian infection [7][8][9]. However, OspC is not only a strong immunogen, but also an effective target of protective immunity; its production ultimately induces a robust humoral response that imposes tremendous pressure on the pathogen [10,11]. To effectively evade the adaptive immune response and establish persistent infection, B. burgdorferi must downregulate ospC after the specific humoral response has developed [12], suggesting an early role for OspC in mammalian infection.
Inactivation of the ospC gene completely abolishes infectivity of B. burgdorferi [5]; however, the resulting mutant is able to persist in mammalian tissues once the initial requirement for OspC is overcome via introduction of an unstable ospC copy, which is eventually lost under the immune selection pressure during infection of immunocompetent mice, leading to a conclusion that OspC is required exclusively for initial mammalian infection [4,13,14]. However, this initial requirement for OspC can be overridden by either increasing expression of another Osp [15], or simply by adapting ospC mutants in mammalian hosts [16]. The nature of adaptation is to alter gene expression, and B. burgdorferi indeed undergoes dramatic changes in its surface lipoprotein expression during mammalian infection [8,9]. Most notably, the downregulation of OspC in response to the development of the anti-OspC humoral response occurs concurrently with the upregulation of both VlsE and BBF01 [7]. Although remaining to be investigated, the host adaptation process most likely provides OspC-deficient spirochetes with both time and environment that enable the upregulation of other Osps, such as VlsE and BBF01, to occur during the course of disappearance of unstable ospC copies. The ability of an Osp to replace OspC in initial mammalian infection highlights a redundant function of the Osps, which is to protect B. burgdorferi against innate immune elimination [15]. However, increasing expression of an Osp fails to fully restore ospC mutants with expected dissemination ability [15], leading us to hypothesize that OspC is a dissemination-facilitating factor.
As described by some investigators, the protective and dissemination-promoting functions of OspC are more like two sides of the same coin [17], highlighting the challenge to dissect them. In this study, fortunately, we were able to generate a truncated OspC, which effectively protected B. burgdorferi against innate elimination, as measured by the 50% infectious dose (ID 50 ) and tissue bacterial loads during murine infection, but failed to efficiently promote dissemination of B. burgdorferi to remote tissues. The study allowed us to conclude that OspC is a disseminationfacilitating factor of B. burgdorferi.

Materials and Methods
Previously generated strains and constructs used in the current study The B. burgdorferi B31 clone 13A, the ospC mutant (DospC), and the complemented clones DospC/FL/1 and DospC/FL/2 were generated previously [18]. The TA cloning vector pNCO1T was constructed in a previous study [19]. The shuttle vector pBBE22 was a gift from S. Norris [20]. The features of these clones and constructs are summarized in Table 1.
Construction of pBBE22-ospCnt5 and pBBE22-ospCnt10 As illustrated in Figure 1, to efficiently generate an N-terminus deletion, a 1057-bp fragment covering the ospC region and the upand down-stream sequences was amplified from B. burgdorferi B31 genomic DNA by PCR using primers P1F and P1R (Table 2) and cloned into the TA cloning vector pNCO1T [19], creating an intermediate vector designated pNCO1T-ospC. One large amplicon was generated by inverse PCR, using pNCO1T-ospC as a template and primers P2F and P2R (Table 2). After digestion with SapI and subsequent purification, the amplicon was circularized via ligation and then digested with BamHI and XbaI to release ospCnt5. This fragment, encoding an OspC variant with Nterminal 5-amino acid (AA) deletion, was cloned into pBBE22 after the vector was digested with BamHI and XbaI. The resulting construct was designated pBBE22-ospCnt5. The insert and its flanking regions within pBBE22 were sequenced to ensure the construct was as designed.
The same strategy was used to generate pBBE22-ospCnt10. Briefly, pNCO1T-ospC was amplified by using primers P3F and P3R ( Table 2). The resultant PCR product was digested, circularized, re-digested, and finally cloned into pBBE22, as described above for constructing pBBE22-ospCnt5.

Transformation of B. burgdorferi and selection of transformants
Constructs were electroporated into DospC; resulting transformants were screened and analyzed for plasmid content as described previously [21]. Restoration of OspC production was verified using immunoblots probed with a mixture of FlaB and OspC MAbs, as described in an earlier study [18].
Indirect immunofluorescence assay B. burgdorferi was grown to late-log phase (10 8 cells per ml) in BSK-H complete medium at 33uC. Approximately 2610 7 cells were harvested from 0.2 ml of culture by centrifugation at 10,0006 g for 10 min, gently suspended in 100 ml PBS containing OspC MAb, and incubated for 1 hour at room temperature. After 2 washes with excess volumes of PBS by centrifugation at 16,0006 g for 5 min, spirochetes were resuspended in 100 ml of PBS containing 1.0 mg of fluorescein isothiocyanate-conjugated goat anti-mouse IgG (Pierce Chemical Company, Rockford, IL), incubated for 1 hour, washed twice with PBS by centrifugation, resuspended in 50 ml PBS, applied to microscopic slides, and analyzed using Axio Imager (Carl Zeiss Microimaging, Inc., Thornwood, NY).

Infection study
Groups of three C3H SCID mice (ages 4-6 week; provided by the Division of Laboratory Animal Medicine at Louisiana State University, Baton Rouge, LA) were inoculated with 10 5 spirochetes. Mice were sacrificed 1 month post-inoculation; heart, tibiotarsal joint and skin specimens were harvested for spirochete culture in BSK-H complete medium. Cultures were examined for growing spirochetes under a darkfield microscope every week for up to 3 weeks. All animal procedures described here and below were approved by the Institutional Animal Care and Use Committee at Louisiana State University (LSU Protocol No. 07-019).

Quick clearance study
Groups of four C3H SCID mice each received two intradermal/ subcutaneous inoculations of 10 5 spirochetes. The two inoculation sites were at least 2 cm apart. Two animals from each group were euthanized at 24 or 48 hours later; inoculation site skin specimens were harvested for spirochete culture in BSK-H medium.

Determination of ID 50 values
Spirochetes were grown to late-log phase (10 8 cells per ml) in BSK-H complete medium at 33uC and 10-fold serially diluted with fresh medium. C3H SCID mice each received one single intradermal/ subcutaneous injection of 100 ml of spirochetal suspension (containing 10 1 to 10 4 organisms). Mice were euthanized one month postinoculation; heart, tibiotarsal joint, and skin (not from inoculation site) specimens were harvested for bacterial isolation. The ID 50 value was calculated as described by Reed and Muench [22].

Tissue spirochetal load analysis
C3H SCID mice each were intradermally/subcutaneously inoculated with 10 4 spirochetes. Animals were sacrificed one month post-inoculation; heart, joint, and skin specimens were harvested for DNA extraction. DNA was quantified for the copy numbers of flaB and murine actin genes by quantitative PCR (qPCR) as described previously [21]. The tissue spirochete burden was expressed as flaB DNA copies per 10 6 host cells (2610 6 actin DNA copies).

Dissemination studies
C3H SCID or BALB/c wild-type mice each were given a single intradermal/subcutaneous injection of 10 3 spirochetes and were euthanized at 1, 2, 3 and 4 weeks post-inoculation. Inoculation site and remote site skin, ear, heart and tiobiotarsal joint specimens were harvested for spirochete isolation as described previously [21]. Spirochetes were injected into the dermis of the chest, so the skin from the back was harvested as remote site.

Statistical analysis
A two-tailed Student t test was used to calculate a P value for each two groups. A P value #0.05 was considered to be significant.

Generation of B. burgdorferi producing OspC with either 5-or 10-AA deletion
The goal of the study was to dissect the dissemination-facilitating from protective function of OspC. Our strategy was to generate ospC   mutants that can be protected from innate elimination but are unable to efficiently disseminate via mutagenizing the ospC gene.
Based on X-ray analyses, OspC is a largely a-helical protein and may be a dimer with a characteristic central four-helical bundle formed by the association of the two longest helices, helices 1 and 5, from each subunit [23,24]. Each subunit consists of five a-helices, two b-sheets and six loops, in addition to the amino (N-) terminal 23-AA linker and the carboxyl (C-) terminal 14-AA stretch. Neither the N-terminal linker nor the C-terminal stretch contributes to the core three-dimensional structure of OspC; instead, the a-helices 1 and 5 bring the two terminal regions in close proximity, where the lipoprotein is anchored to the bacterial outer membrane through lipidation of the first cysteine residue of the N-terminal linker.
Within the core, several small secondary structures are present, including a-helix 4 (consisting of 8 AA), b-sheet 1 (5 AA) and 2 (5 AA). We first generated B. burgdorferi expressing OspC with each of these small units being deleted (data not shown). Unfortunately, none of the resulting mutants were infectious, suggesting that these sequences are critical for the functions of OspC. Next, we focused on the N-terminal sequence. One role of this sequence is to carry the sorting signal, which determines the surface location of OspC. In E. coli the so-called ''+2'' rule determines the cellular location of a lipoprotein. However, the sorting signal of B. burgdorferi can extend up to the +4 position [25]. To create N-terminal deletions, which can be successfully sorted to the outer surface, we intended to generate mutations starting at the +7 position. The construct pBBE22-ospCnt5 was generated as illustrated in Figure 1A.
The two constructs were electroporated into the ospC mutant, DospC, which was generated and characterized in our previous study [18]. Because DospC lacks lp25, the plasmid that carries the gene bbe22 coding for a nicotinamidase essential for survival of B. burgdorferi in the mammalian environment, the recombinant plasmid pBBE22, which harbors a copy of bbe22, was used as the shuttle vector [20]. Six and seven transformants were obtained receiving each construct; plasmid analyses led to the selection of four clones, DospC/DNt5/1, DospC/DNt5/2, DospC/DNt10/1 and DospC/DNt10/2. These clones shared the same plasmid content as DospC, which had lost lp25, lp5, lp21, lp56 and cp9 [18]. Expression of truncated OspC resulting from the introduced construct was confirmed by immunoblot analysis ( Figure 1B).
Neither the 5-nor the 10-AA deletion affects the surface location of OspC Because the deletion was generated close to the N-terminus, which may harbor the sorting signal, it is important to show that the truncated OspC can be successfully sorted to the outer surface of B. burgdorferi. To this end, we used indirect immunofluorescence to locate truncated OspC. As shown in Fig. S1, both truncated proteins showed similar fluorescence patterns as the wild-type control, indicating that neither deletion influenced the cellular location of OspC.

The 10-but not the 5-AA deletion abolishes infectivity
Groups of three C3H SCID mice were inoculated with 10 5 spirochetes of the clone DospC/DNt5/1, DospC/DNt5/2, DospC/ DNt10/1, or DospC/DNt10/2. As a control, mice were also inoculated with the clone DospC/FL/1 or DospC/FL/2. The clones DospC/FL/1 and DospC/FL/2 were generated via introduction of a full-length ospC gene carried by the shuttle vector pBBE22 into DospC in our previous study [18]. All mice were sacrificed 1 month post-inoculation; heart, tibiotarsal joint and skin specimens were harvested for spirochete culture in BSK-H complete medium. As shown in Table 3, like the control clones, the DospC/DNt5/1 and DospC/DNt5/2 spirochetes were grown from each sample of all inoculated mice. In contrast, neither DospC/DNt10/1 nor DospC/DNt10/2 bacteria were recovered from any specimens, suggesting that the 10-AA deletion may completely abolish the functions of OspC.
The DospC/DNt10/1 and DospC/DNt10/2 bacteria might remain at the inoculation sites without dissemination. To rule out this possibility, groups of four SCID mice each received two intradermal/subcutaneous inoculations of the clones DospC/ DNt10/1, DospC/DNt10/2, DospC/FL/1 or DospC/FL/2. Animals were euthanized at 24 or 48 hours later. As a positive control, the DospC/FL/1 and DospC/FL/2 bacteria were consistently grown from each of the 16 inoculation sites from all eight inoculated mice (data not shown). In contrast, neither DospC/DNt10/1 nor DospC/ DNt10/2 spirochetes were recovered from any of the 8 sites harvested at each time point, confirming that the 10-AA deletion completely destroys the functions of OspC and leads to quick clearance of spirochetes in the murine host.

The 5-AA deletion does not affect the ID 50 value in SCID mice
Because the ID 50 value reflects the smallest number of organisms to sufficiently initiate an infection in 50% of inoculated individuals, it is the best measurable criterion to assess how a pathogen is protected against innate immune elimination, especially when this is measured in SCID mice. To examine whether the 5-AA deletion reduced the ability of OspC to protect B. burgdorferi against innate immunity, groups of three C3H SCID mice each received one single inoculation of 10 1 to 10 4 spirochetes The 5-AA deletion does not reduce spirochete burdens in joint or skin of SCID mice Next, the tissue bacterial load was used to reflect whether the truncated OspC can effectively protect bacteria against innate elimination. To examine the influence of the 5-AA deletion on the tissue bacterial load, subgroups of five SCID mice each received a single intradermal/subcutaneous inoculation of 10 4 spirochetes of the clone DospC/DNt5/1, DospC/DNt5/2, DospC/FL/1, or DospC/ FL/2. In 10 mice that were inoculated with the DospC/FL/1 or DospC/FL/2, joint swelling evolved around 10 days post-inoculation and developed into severe arthritis a week later (data not shown). In the remaining mice, joint swelling did not become apparent until 3 weeks post-inoculation and slowly developed after then. All mice were euthanized 1 month post-inoculation; DNA was extracted from heart, joint and skin specimens and quantified for bacterial burden. Although the DospC/DNt5 spirochete burden was 6.7-fold lower than that of the DospC/FL (P = 2.83610 210 ) in the heart tissue, there was no significant difference detected either in joint (P = 0.52) or skin tissue (P = 0.13) (Figure 2). The study indicated that the 5-AA deletion does not reduce the ability of OspC to protect B. burgdorferi against innate immunity.

The 5-AA deletion leads to slowed dissemination during infection of SCID mice
The ID 50 value was examined at one month post-inoculation. To determine whether the deletion caused deficiency in dissemination, we examined infection at different time points. Groups of 10 or 15 SCID mice each received a single intradermal/subcutaneous inoculation of the clone DospC/FL/1 or DospC/DNt5/1. Five animals from each group were euthanized at 1-week intervals; inoculation site and remote site skin, ear, heart, and joint specimens were harvested for spirochete isolation. As a positive control, the DospC/FL/1 bacteria were grown from all of the skin, joint and heart specimens but from none of the ear samples at first week; all sites became culture positive at 2 weeks after initial inoculation (Table 5). In contrast, the DospC/DNt5/1 spirochetes were grown only from inoculation sites at a week post-inoculation. Although the mutant disseminated to most of the joint, remote skin and heart specimens at 2 weeks, majority of ear tissues were not colonized until 3 weeks. These data indicated that OspC is important for efficient dissemination.

The 5-AA deletion causes more severely impaired dissemination during infection of immunocompetent mice
More severely impaired dissemination was noted during infection of immunocompetent mice. Groups of 24 BALB/c mice each received a single intradermal/subcutaneous inoculation of  the clone DospC/FL/1 or DospC/DNt5/1. Six animals from each group were euthanized at 1-week intervals; inoculation site and remote site skin, ear, heart, and joint specimens were harvested for spirochete isolation. As a positive control, the DospC/FL/1 bacteria were grown from all of the remote tissues but the ear at first week; all sites became culture positive at 2 weeks after initial inoculation (Table 6). In contrast, the DospC/DNt5/1 spirochetes were not grown from joint tissues until 3 weeks, and even more significantly, no heart, remote skin or ear tissues became culture positive during the period of the 4-week study. These data further highlight that OspC is required for efficient dissemination.

Discussion
Like typical Gram-negative bacteria, B. burgdorferi possesses inner and outer membranes, between which is a periplasmic space [26,27]. Gram-negative pathogens make a thick LPS coat to provide a broad array of crucial protection [28]. However, B. burgdorferi does not produce any LPS, but instead, abundantly expresses lipoproteins and anchors them to the outer membranous surface through lipidation [26,29,30]. This unique surface antigenic structure may facilitate the host-pathogen interactions and potentially contribute to the pathogenic strategy of B. burgdorferi. As protein is a direct product of gene expression, regulation of surface antigen expression can quickly result in modification of the surface antigenic architecture. B. burgdorferi indeed takes advantage of this and vigorously modifies its surface lipoprotein expression to constantly reformulate its overall surface antigenic architecture during its enzootic life cycle traveling between the tick vector and a mammal, and during the course of mammalian infection. It abundantly expresses OspA and OspB in the unfed tick, because both may play an important role in the tick vector [2,3,31,32]. A fresh bloodmeal induces the down-regulation of OspA/B and the up-regulation of OspC and other Osps, a process that prepares B. burgdorferi for infection of mammals [4][5][6]33]. Abundant OspC expression ultimately induces a robust early humoral response that imposes tremendous pressure on the pathogen [10,12]. To evade the specific humoral response and cause persistent infection, B. burgdorferi down-regulates OspC and dramatically upregulates other surface lipoproteins, including VlsE and BBF01 [7][8][9]34]. This well-defined ospC expression pattern suggests an early role in mammalian infection.
Recent studies showed that this initial requirement for OspC can be readily overridden by either increasing expression of another Osp [15], or by simply adapting ospC mutants in mammalian hosts [16]. Clearly, the adaptation process provides B. burgdorferi with both the environment and time to increase expression of other Osps, such as VlsE and BBF01, when OspC production is being reduced in response to development of an anti-OspC humoral response [7]. Taken together, these studies indicate that increasing expression of an Osp(s), either via genetic modification or adaptation, can overcome the initial requirement for OspC, leading us to hypothesize that the borrelial lipoproteins may have a common function, which is to protect B. burgdorferi from innate elimination. To survive in the mammalian host, B. burgdorferi must maintain a certain level of surface lipoprotein production. During early infection, B. burgdorferi abundantly produces OspC. In addition to this common function, the specific function of OspC is to facilitate dissemination, which can be only partially replaced by another Osp. In the current study, we were able to separate protective and dissemination-facilitating functions as we successfully generated a truncated OspC, which can effectively protect B. burgdorferi against innate elimination but fails to make spirochetes to disseminate as efficiently as the wild-type control.
Our previous study showed that increasing production of one of four randomly chosen surface lipoproteins can replace OspC to  Although the 5-AA deletion retains the full protective function of OspC, several other deletions, including that of the N-terminal 10-AA, a-helix 4 (8 AA), and b-sheet 1 (5 AA) and 2 (5 AA), completely destroys its protective function (Seemanapalli and Liang, unpublished data). This leads us to hypothesize that Osps may play a common structural role that is required for the protective function of the borrelial surface lipoproteins. Some deletions on OspC may cause this special structure to collapse and consequently abolish the protective function. After inoculated into murine skin, B. burgdorferi must first be able to evade initial innate immune elimination and gain a foothold. This ability is best reflected by the ID 50 value because this criterion measures the smallest number of organisms that is sufficient to escape initial innate elimination and establish an infection. In spite of large inocula, OspC-deficient spirochetes are quickly cleared after inoculation into murine skin [14,15]. Production of OspC with 5-AA deletion fully restored OspC-deficient B. burgdorferi with ID 50 values similar to the control, indicating that the truncated molecule retains the full protective function of OspC.
Although tissue colonization is an extremely complicated event carried out by an extracellular bacterial pathogen during infection and can be affected by many factors, one can envision the ability of the pathogen to evade immune clearance should be one of the major determinants. B. burgdorferi producing OspC with 5-AA deletion was able to generate similar bacterial loads in both joint and skin tissues of SCID mice as the control, indicating that the truncated OspC can effectively protect the pathogen against innate immune clearance at least in these tissues. Because many other determinants, such as the interactions of host cells and extracellular matrices with the pathogen mediated by its surface antigens, may significantly influence tissue colonization, the lower DospC/DNt5 spirochete load in the heart does not necessarily suggest that the truncated OspC cannot effectively provide protection against innate immune elimination. Alternatively, OspC may be an important factor contributing to the ability of B. burgdorferi to colonize the heart tissue. This explanation is also supported by previous studies showing that treatment with OspC antibody shuts off ospC expression and more effectively reduces the bacterial load in the heart than other tissues [7,11] and that OspC-deficient spirochetes with increased expression of another Osp register reduced bacterial loads in the heart but not other tissues of SCID mice [15].
Both ID 50 and tissue bacterial load data demonstrated that the 5-AA deletion does not affect the protective function of OspC, allowing us to dissect the protective from dissemination-facilitating functions. B. burgdorferi producing OspC with 5-AA deletion disseminated to remote tissues, especially to the heart and ear at a remarkably slow pace in immunodeficient mice. In immunocompetent mice, the defect in dissemination caused by deletion was much more severe. Four weeks after inoculation, the mutant was not able to colonize heart or ear tissues, in contrast to the control, which disseminated to all tissues examined within 2 weeks. The study also highlights the importance of quick dissemination during infection of immunocompetent animals. As infection induces immune responses, if B. burgdorferi fails to colonize remote tissues and establish a systemic infection before effective immune responses develop, its infectivity potential would be severely impaired. Figure S1 Deletion of N-terminus 5-AA or 10-AA does not affect the surface location of OspC. The DospC/FL/1, DospC/DNt5/1 or DospC/DNt10/1 spirochetes were grown to late log phase (10 8 per ml) in BSK-H medium at 33uC. Spirochetes were incubated with OspC MAb, washed in PBS, probed with FITC-conjugated goat anti-mouse antibody, washed again before being placed onto slides. Differential interference contrast (left panel) and immunofluorescence images (right panel) were taken from the same field. Imagines were taken at 6400 magnification. (PDF)