Borrelia burgdorferi Resistance to a Major Skin Antimicrobial Peptide Is Independent of Outer Surface Lipoprotein Content

Abstract

We hypothesize a potential role for Borrelia burgdorferi OspC in innate immune evasion at the initial stage of mammalian infection. We demonstrate that B. burgdorferi is resistant to high levels (>200 μg/ml) of cathelicidin and that this antimicrobial peptide exhibits limited binding to the spirochetal outer membrane, irrespective of OspC or other abundant surface lipoproteins. We conclude that the essential role of OspC is unrelated to resistance to this component of innate immunity.

Borrelia burgdorferi, a spirochete and the causal organism of Lyme disease, is naturally transmitted to mammals through the bite of infected Ixodes ticks (5, 8). A significant change in B. burgdorferi gene expression accompanies transmission between these diverse environments. This was first described for the inverse relationship between two abundant outer surface proteins of B. burgdorferi, in which synthesis of OspA declines and that of OspC increases during tick feeding (41). We and others have demonstrated the essential nature of OspC for colonization of the murine host (23, 35, 42, 45, 47, 49). These findings suggest a critical role for OspC in evasion of host innate immunity immediately after transmission (47). However, the essential contribution of OspC to early mammalian infection by B. burgdorferi remains undefined.

Microorganisms induce a variety of responses from the skin epithelial cells of their hosts, including the production of antimicrobial peptides, which are recognized as integral components of the innate immune system (20, 22). Defensins and cathelicidins comprise two major families of cationic antimicrobial peptides secreted by human and other mammalian skin neutrophils (20). Mouse neutrophils lack α defensins (14, 24), but about 30 cathelicidin members have been identified in various mammalian species, including mice (21, 50). These small, cationic, amphipathic molecules are primarily stored as inactive propeptides in the secretory granules of skin neutrophils. The mature bioactive peptides assume an α-helical structure in solution and preferentially interact with negatively charged cell surface components of a broad spectrum of bacteria and fungi, in which they disrupt cell membrane integrity (6, 9, 12, 20, 34). The importance of the sole murine cathelicidin, known as mCRAMP (mouse cathelin-related antimicrobial peptide) (19, 36), to innate host defense is well established, and mCRAMP has been shown to provide protection against bacterial skin infections in mice (33).

Resistance of B. burgdorferi to cathelicidin.

Treatment of Lyme borreliosis with antibiotics is generally successful, but there are rare instances of resistance (26), and several B. burgdorferi genes have been identified with potential roles in resistance to antibiotics (7, 10, 18, 40). However, potential mechanisms employed by the spirochete to evade the innate host response are not well understood yet. It has been demonstrated that unlike many other bacterial pathogens, B. burgdorferi is highly resistant to cathelicidin-derived peptides (27, 39), consistent with the spirochete's ability to persistently colonize the skin, where CRAMP is present. Sambri et al. (39) suggest that the resistance of B. burgdorferi to antimicrobial peptides may derive from the spirochete's lack of lipopolysaccharide, a negatively charged membrane component to which cationic peptides typically bind (25, 43). However, the B. burgdorferi outer membrane contains abundant lipoproteins with exposed charged residues that could mediate or repel cathelicidin binding, such as OspC (13, 30, 31), which is made by B. burgdorferi during the initial phase of mammalian infection, when the spirochete would encounter antimicrobial peptides in the skin. Although OspC is a basic protein with an isoelectric point of ∼9.0 and a net positive charge, the three-dimensional structure of OspC indicates the presence of a surface region with a strong negative electrostatic potential that would project away from the positively charged, membrane-proximal region (13, 30). This negatively charged, exposed surface of OspC is postulated to be important for binding to unidentified positively charged host molecules or ligands (13, 30). We hypothesize that as an abundant surface lipoprotein with limited membrane contact, OspC could shield the spirochete from lytic components of innate defense like cathelicidin by binding and sequestering them, thus preventing access to the cell membrane. This potential role of OspC in resistance is consistent with the rapid clearance from skin of mutant spirochetes that lack OspC (45).

To test this hypothesis, we compared the resistance of B. burgdorferi variants that differ in outer surface lipoprotein composition to mouse cathelicidin-related antimicrobial peptide (mCRAMP). The bacterial strains and plasmids used in this study are described in Table 1, and the relative amounts of OspC produced by the B. burgdorferi study strains are shown in Fig. 1A and C. We initially compared the survival, following incubation with mCRAMP, of three B. burgdorferi clones synthesizing or lacking OspC (A3, the ospCK1 strain, and the ospCK1+ospC strain). Briefly, mid-log phase B. burgdorferi cultures were washed and resuspended in 10 mM sodium phosphate buffer (pH 7.4) at a concentration of ∼10⁷ organisms/ml, and 10 μl of bacterial culture was added to duplicate wells of a 96-well polypropylene plate (Sigma-Aldrich, St. Louis, MO) containing mCRAMP (Axxora, San Diego, CA) (1 mg/ml in 0.01% acetic acid containing 0.2% bovine serum albumin) at various concentrations using Mueller-Hinton broth as the assay medium, in a total volume of 100 μl. All three strains were found to be highly resistant to killing at a wide range of antimicrobial peptide concentrations, irrespective of their OspC phenotype (Fig. 2A and B). This experiment was conducted at even higher concentrations of mCRAMP (300 to 500 μg/ml), but no killing was observed for any of these strains (data not shown). However, the cathelicidin-susceptible species Escherichia coli and Leptospira biflexa were found to be highly sensitive to mCRAMP when assayed under identical experimental conditions (Fig. 2C and D). A similar killing assay conducted with polymyxin B, another standard cationic antimicrobial agent, demonstrated resistance of B. burgdorferi to this compound as well (Fig. 2E). These results indicate that the resistance of B. burgdorferi to these antimicrobial peptides is not OspC dependent.

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Description or characteristics	Source or reference(s)
B. burgdorferi strainsa
B31-A3	Low-passage, infectious clonal derivative of wild-type strain B31	15
ospCK1 strain	ospC mutant; isogenic, noninfectious derivative of B31-A3	47
ospCK1+ospC strain	OspC-positive derivative of ospCK1 (ospCK1/pBSV2G-flaBp-ospC; constitutively expresses ospC from the flaB promoter)	44
AC2	ΔospAB-ΔospC double mutant strain; derived by transforming ospA mutant ospA1 (4) with plasmid pAB109b; ospA1 is an isogenic derivative of B31-A3	4; this study
B312	High-passage, noninfectious clone, lacking all linear plasmids except lp54; synthesizes OspAB and some OspC	38
B314	High-passage, noninfectious clone lacking all linear plasmids; synthesizes a high level of OspC but no OspAB or any other linear plasmid-encoded lipoproteins	38
E. coli Top10b	F−mcrA Δ(mrr-hsdRMS-mcrBC) φ80lacZΔM15 ΔlacΧ74 recA1 araD139 Δ(ara-leu)7697 galU	Invitrogen
	galK rpsL (Strr) endA1 nupG λ−
L. biflexa Patoc Ic		11
Plasmid pAB109b (ΔospC::flgBp-aadA)	Constructed by excising the flgBp-aadA fusion from shuttle vector pKFSS1 (17) by digestion with NotI and XhoI and replacing the flgBp-kan fusion in NotI+XhoI-digested pJK109b (47)	This study

^aB. burgdorferi strains were inoculated from frozen stocks and grown at 35°C with 2.5% CO₂ (37) in liquid BSK II medium (1, 29).

^bE. coli was cultured in LB at 37°C to mid-log phase.

^cL. biflexa was cultivated at 30°C to mid-log phase in the Leptospira medium EMJH (16, 28, 32).

FIG. 1.

Outer surface protein profiles of B. burgdorferi clones. (A) Coomassie-stained sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) of whole-cell lysates of B31 clones: A3, the ospCK1 strain, the ospCK1+ospC strain, AC2, B312, and B314. (B and C) Immunoblotting of the replicate SDS-PAGE gel with anti-OspA (H5332; 1:200 dilution) (3) and anti-FlaB (H9724; 1:200 dilution) (2) monoclonal antibodies (B) and anti-OspC rabbit polyclonal antiserum (1:1000 dilution) (45) and anti-FlaB (H9724, 1:200 dilution) (2) monoclonal antibody (C). Membranes were subsequently treated with respective peroxidase-conjugated anti-mouse immunoglobulin or anti-rabbit immunoglobulin sera (Sigma-Aldrich), and peroxidase activity was detected using Super Signal reagents (Thermo Scientific, Rockford, IL). Molecular mass standards are indicated. B. burgdorferi lysates were prepared and electrophoretically separated by SDS-PAGE as previously described (46) with 50 mM dithiothreitol (Boehringer Mannheim, Indianapolis, IN) substituted for 2-mercaptoethanol in Laemmli sample buffer when lysates were prepared for analysis of the OspA protein (4). *, OspA is deleted in strain AC2 and not present in B312, which lacks the lp54 plasmid. **, the wild-type study strain B31-A3 and its isogenic derivatives (the ospCKI strain, the ospCKI+ospC strain, and AC2) have a truncated version of the ospB gene and hence do not produce OspB (15).

FIG. 2.

Role of OspC in cathelicidin resistance of B. burgdorferi. The wild-type strain A3, the ospCK1 mutant, and the constitutively OspC-expressing ospCK1+ospC strain were incubated with various concentrations of cathelicidin (A and B) and polymyxin B (E) at 35°C under 2.5% CO₂ for 3 h. Viable bacteria were enumerated by direct microscopic count of motile spirochetes (A and E) and by plating in solid BSK II medium (37) to determine the number of CFU (B). Bacterial survival following incubation with mCRAMP was calculated relative to bacterial survival in a parallel incubation without mCRAMP. Dose-dependent killing of E. coli and L. biflexa by mCRAMP was determined in a similar fashion at 37°C and 30°C, respectively, and surviving bacteria were quantitated either as CFU on LB agar plates (E. coli) (C) or by direct microscopic count of motile bacteria (L. biflexa) (D).

Surface lipoprotein contribution to cathelicidin resistance.

The in vitro-grown spirochetes tested in this initial assay continuously synthesized OspA, while OspC synthesis varied (Fig. 1). It was observed recently that when synthesis of OspA is manipulated to remain artificially high following inoculation into a mouse, OspA can restore mouse infectivity to an ospC mutant, albeit inefficiently (48). To determine whether the presence of OspA contributed to the observed resistance to cathelicidin and masked an otherwise critical role of OspC, we constructed and tested a mutant strain that lacks genes for ospAB and ospC and hence does not produce any of these lipoproteins (AC2) (Table 1 and Fig. 1). However, this strain still maintained a high level of resistance to cathelicidin (Fig. 3). We also found that despite lacking many lipoproteins, two highly attenuated B. burgdorferi clones, B312 and B314, were highly resistant to cathelicidin-mediated lysis (Fig. 3). We therefore conclude that physical properties of the outer membrane of B. burgdorferi, not the presence of specific lipoproteins, likely confer the spirochete's high level of resistance to these cationic antimicrobial peptides.

FIG. 3.

Outer surface lipoprotein content of B. burgdorferi and cathelicidin resistance. B312, B314, and the double mutant AC2 (lacking OspAB and OspC) were incubated with the indicated concentrations of mouse cathelicidin for 3 h. Bacterial survival was determined by direct microscopic count of motile spirochetes.

Cathelicidin binding.

The mechanism by which B. burgdorferi maintains resistance to cathelicidin and other antimicrobial peptides is largely unknown. To determine whether sensitivity to cathelicidin reflects peptide binding to the outer surface of the spirochete, we employed an indirect immunofluorescence assay to assess mCRAMP binding to B. burgdorferi and L. biflexa (which was used as a positive control). We observed limited mCRAMP binding to B. burgdorferi, in contrast to extensive mCRAMP binding to L. biflexa (Fig. 4). The highly attenuated clones B312 and B314, varying in OspAB and OspC synthesis, as well as strain AC2, lacking OspAB and OspC, similarly demonstrated very low levels of cathelicidin binding, consistent with their resistant phenotypes (Fig. 4C).

FIG. 4.

Relative CRAMP binding of B. burgdorferi and L. biflexa. B. burgdorferi (A) and L. biflexa (B) were fixed with 3% paraformaldehyde at 37°C for 10 min, washed, resuspended in phosphate-buffered saline, and incubated with CRAMP (50 μg/ml) for 3 h using Mueller-Hinton broth as the assay medium. Bacteria were subsequently incubated with polyclonal rabbit anti-CRAMP antibody (αmCRAMP) (Axxora) (1:100) and fluorescein isothiocyanate-labeled goat anti-rabbit immunoglobulin antiserum (Sigma) (1:100) for 30 min each at 37°C and counterstained with the membrane stain FM4-64FX (Invitrogen) for 10 min at 37°C. Analysis was done with an epifluorescence microscope (Eclipse 800; Nikon, Tokyo, Japan), and bacteria observed with tetramethyl rhodamine isocyanate and fluorescein isothiocyanate filters were photographed (DXM1200C; Nikon), counted, and compared. The lower panels show negative controls, where bacteria were incubated without CRAMP (w/o mCRAMP). Magnification, ×450. (C) Quantitation of the relative binding of CRAMP to different B. burgdorferi clones and L. biflexa. The number of bacteria binding CRAMP (αmCRAMP) was counted and compared with the number of total bacteria visualized (FM4-64FX) in the same field. Five independent fields were analyzed for each of the B. burgdorferi clones and L. biflexa, and the relative CRAMP binding was calculated from the average values.

Our data are in agreement with the previous observation that B. burgdorferi is highly resistant to the innate skin antimicrobial peptide cathelicidin (39), and the data presented here indicate that the critical early role of OspC during mammalian infection is unrelated to resistance to major antimicrobial peptides in the skin; hence, the role of OspC remains elusive. Our results support a model by which the observed resistance of B. burgdorferi to cathelicidin correlates with a lack of antimicrobial peptide binding to the spirochete membrane, and we suggest that this characteristic is consistent with the unique biophysical properties of the Borrelia outer membrane, independent of lipoprotein content.