Borreliella burgdorferi B31 strain B314 (46) ( Table 1 ) was grown to mid-log phase in BSK-II and 6% normal rabbit serum (Pel-Freez Biologicals, Rogers, AR) under conventional conditions (1% CO₂, 32°C, pH 7.6). Genetically transformed B. burgdorferi strains were grown with kanamycin at 300 μg/mL. Borrelia miyamotoi strain FR64b was grown in MKP-F media (24) under the same conditions listed above for B. burgdorferi. Escherichia coli strain NEB-5α (New England Biolabs) was grown in Luria broth media under aerobic conditions at 37°C and, as appropriate, with kanamycin at 50 μg/mL. E. coli strain BL21 (DE3) (Thermo Fisher) was grown in Terrific Broth (Fisher Bioreagents) with kanamycin (50 μg/mL) at 37°C until an OD₆₀₀ between 0.6-0.8 was achieved, then moved to 16°C overnight for production of recombinant Fbp proteins.

The following constructs for B. miyamotoi FR64b fbpA (UNIPROT#: AHH06014.1) and fbpB (UNIPROT#: AHH05635.1) were subcloned into pT7HMT or a modification of pT7HMT encoding maltose binding protein (MBP) (48). All plasmid constructs are listed in Table 1 . E. coli codon-optimized DNA fragments were purchased from IDT Technologies gBlock Gene Fragment Service or produced using PCR amplification templated from previously cloned constructs using methodology and constructs previously described (32–34, 48): FbpA-C residues 222-371; FbpA-C-R264A-K343A; FbpB-C residues 264-432; FbpB-Δ5C residues 264-427; FbpA-N residues 152-201; and FbpB-N residues 207-248. Oligonucleotides used are listed in Table 2 . Residue numbering were based on UNIPROT numbering. The expression constructs for B. burgdorferi BBK32-C and C1r-CCP2-SP originated from previous studies (32, 33).

Expression constructs for B. miyamotoi fbpA and fbpB were assembled using the NEBBuilderHiFi DNA Assembly Cloning Kit as described in (34). The B. miyamotoi fbpA and fbpB genes were transcriptionally linked with the bbk32 promoter from B. burgdorferi to promote equal expression in B314 relative to bbk32. The R264A and K343A mutations in fbpA were constructed sequentially by synthesizing a 194 bp fragment of fbpA from nucleotide 918 to 1112 containing the R264A mutation, and then once the R264A construct was obtained, with a 254 bp fragment from nucleotide 1112 to 1366 containing the K343A mutation. Both DNA fragments were synthesized by IDT gBlocks using the native B. miyamotoi strain FR64b fbpA sequences. The fragments were each assembled into the fbpA gene using primers that added overhangs complementing the fbpA gene adjacent to the synthesized fragments. These fragments were then assembled into the fbpA-expressing vector using NEBuilderHiFi DNA Assembly Cloning Kit. All plasmids were sequenced to ensure the desired mutations were obtained and that no additional mutations were introduced during the assembly process.

Transformation of strain B314 with the plasmid constructs pAP8, pAP11, and pAP13 was done as previously described (33, 49).

Recombinant proteins were expressed and purified as in (32, 33). Briefly, after elution of the Ni column using the Elution buffer (20 mM Tris (pH 8.0), 500 mM NaCl, 500 mM Imidazole), the proteins were exchanged into 20 mM Tris (pH 8.0), 500 mM NaCl, 10 mM Imidazole using a Desalting 26/10 column (GE Healthcare). The His-tag was then removed by incubation with the tobacco etch virus (TEV) protease and 5 mM β-mercaptoethanol overnight at room temperature. The TEV digested proteins were separated from the His-tag by passing over a 5 mL HisTrap-FF column on the FPLC with the captured flowthrough further purified using a HiLoad Superdex 75 PG gel filtration column (GE Healthcare). The single peak after gel filtration was examined using SDS-PAGE analysis, then pooled and exchanged into HBS buffer (10 mM HEPES [pH 7.3], 140 mM NaCl). The C1r-CCP2-SP construct was purified according to previously published protocols with an added final gel filtration chromatography step using a HiLoad Superdex 75 PG column (GE Healthcare) and exchanged into HBS buffer (33, 34, 50, 51). Purified full-length complement proteins of both the zymogen and enzymatically active forms of C1r were obtained from Complement Technology, Inc. (Tyler, TX). Purified full-length human fibronectin was purchased from Millipore Sigma. Sequence alignments of the proteins were generated using Clustal alignment server from EMBL-EBI.

Surface plasmon resonance (SPR) was performed on a Biacore T200 instrument set to 25°C with a flowrate of 30 μL/min. The running buffer was HBS-T (10 mM HEPES (pH 7.3), 140 mM NaCl, and 0.005% Tween 20) or HBS-T containing 5 mM CaCl₂. Proteins of interest were immobilized on CMD200 sensor chips (Xantec Bioanalytics) via standard amine coupling as before (32–34). Immobilization quantities were as follows: human plasma fibronectin over MBP-FbpA-C (1015.3 RU), MBP-FbpB-C (1074.1 RU); C1r-CCP2-SP over FbpA-C (1061.2 RU), FbpB-C (877.7 RU), FbpA-C-R264A-K343A (FbpA DA-C) (1591.8 RU); C1r zymogen or active C1r over FbpA-C (852.4 RU), FbpB-C (1116.6 RU). C1r-CCP2-SP over Fbps was performed in multi-cycle using a twofold dilution series of the C1r-CCP2-SP ranging from 0-100 nM. Data was obtained in triplicate with each cycle having an association time of 2 min, a dissociation time of 3 min, and three 1 min injections of regeneration buffer (0.1 M Glycine [pH 2.0], 2.5 M NaCl). Equilibrium dissociation constants (K _D) were obtained from the resulting kinetic fits generated by the Biacore T200 Evaluation Software (GE Healthcare). Zymogen or active C1r binding data for FbpA-C and FbpB-C was obtained in single cycle using a fivefold dilution series of either C1r zymogen or active C1r ranging from 0-100 nM. Data was obtained in triplicate with each cycle having an association time of 5 min, a final dissociation time of 1 hour, and two 1 min regeneration injections with regeneration buffer (0.1 M Glycine [pH 2.0], 2.5 M NaCl). Equilibrium dissociation constants (K _D) were obtained from the resulting kinetic fits generated by the Biacore T200 Evaluation Software (GE Healthcare). Human plasma fibronectin was buffer exchanged into HBS-T and then injected over MBP-FbpA-N or MBP-FbpB-N. Data was obtained in multicycle using a twofold dilution series ranging from 0-500 nM. Data was obtained in triplicate with each cycle having an association time of 2 min, a dissociation time of 3 min, and two 1 min injections of regeneration buffer (10 mM CAPS [pH 10.0]). Steady-state affinity fits were performed using Biacore T200 Evaluation Software.

SPR was performed using conditions described for the direct binding experiments. FbpA-C (935.4 RU), FbpB-C (1082.9 RU), and FbpB-Δ5C (1115.4 RU) were immobilized on a new CMD200 sensor chip (Xantec Bioanalytics) via standard amine coupling as before (32–34). 25 nM C1r-CCP2-SP alone, or with 25 nM FbpA-C, FbpB-B, FbpA DA-C, or FbpB-Δ5C were injected over the Fbp biosensors. Injections were performed in triplicate with each cycle having an association time of 2 min, a dissociation time of 3 min, and two 1 min injections of regeneration buffer (0.1 M Glycine [pH 2.0], 2.5 M NaCl). The binding response just prior to injection stop was used to determine differences for the ability of the soluble Fbp proteins to compete with immobilized Fbps for C1r-CCP2-SP binding.

Polyclonal antibodies reactive with FbpA and FbpB were generated by separately immunizing female C57BL/6 mice intradermally with 25 µg either FbpA-C or FbpB-C in an equal volume of PBS and TiterMax^® Gold Adjuvant (Sigma Aldrich) as outlined (52). Two weeks after the initial immunization, mice were boosted with 25 µg of the same protein. Two weeks after boosting, mice were euthanized and blood was immediately isolated by exsangunation, allowed to clot, and serum removed after low speed centrifugation. Reactivity and specificity of each serum sample to B. miyamotoi FbpA and FbpB was evaluated by Western Blot using purified recombinant protein. All animal work was reviewed and approved by the Texas A&M University Institutional Animal Care and Use Committee (protocol number 2019-0422).

Far Western overlays were carried out as described in (32, 35). Briefly, B. burgdorferi strain B314 derivatives were cultured in BSKII media supplemented with 300 μg/mL kanamycin. Cultures were harvested and B. burgdorferi whole-cell lysates were generated for B314 strains producing FbpA, FbpB, FbpA DA, or from the pBBE22luc vector-only control. 2.5 x10⁷ whole cell equivalents were resolved by SDS-PAGE with 0.625 μg of recombinant Fbp-C and MBP-Fbp-N proteins used as controls for C1r and fibronectin, respectively. The SDS-PAGE resolved proteins were transferred to a PVDF membrane and then blocked overnight in 5% non-fat milk. The blots were washed and incubated with 20 μg of zymogen or active C1r for 1 hour. PVDF membranes were probed for C1r binding using a goat antibody to human C1r (R&D Systems) at a 1:3,000 dilution followed by a rabbit anti-goat antibody conjugated to HRP (Invitrogen) at a 1:5,000 dilution. Similar experimentation was done after incubation with 20 μg of human fibronectin with subsequent binding determined using a mouse monoclonal antibody specific for human fibronectin conjugated to HRP (Santa Cruz Biotechnology) at a 1:1,000 dilution. All blots were then visualized using the SuperSignal™ West Pico Plus chemiluminescent kit (Thermo Scientific). Control Western blots were performed as above with 2.5 x10⁷ whole cell equivalents to monitor Fbp expression and to serve as loading controls. To probe for FbpA and FbpB expression, mouse antibodies against FbpA or FbpB were used at a 1:5,000 dilution, respectively. To examine the flagellar protein FlaB, a mouse monoclonal antibody specific for FlaB (US Biologicals, Inc.) was used at a 1:20,000 dilution. Detection of membrane bound immune complexes was facilitated with a goat anti-mouse immunoglobulin HRP conjugate diluted 1:10,000 (Invitrogen). Visualization was the same as above. All Far Western blots were performed in duplicate.

FbpA-C was concentrated to 12.8 mg/mL total protein in a buffer of (10 mM HEPES [pH 7.3], 50 mM NaCl). Crystals were obtained by vapor diffusion of sitting drops at 20°C by mixing 1 µL of protein with 1 µL of precipitant solution (0.1 M Tris [pH 8.5], 2 M ammonium sulfate). Large crystal plate clusters generally appeared after 9-10 d. Sequential rounds of micro-seeding produced samples of sufficient size and quality for harvesting. Samples were cryopreserved in precipitant solution supplemented with an additional 20% glycerol.

Crystallization trials using FbpB-C failed to produce diffraction quality crystals. Constructs removing residues from the N-terminus and C-terminus were screened resulting in identification of crystallization conditions for a construct of FbpB where five residues were removed from the C-terminus (i.e., FbpB-Δ5C). FbpB-Δ5C was concentrated to 19.1 mg/mL and 8.69 mg/mL in 10 mM Tris (pH 8.0), 50 mM NaCl buffer. Crystals of FbpB-Δ5C were obtained by vapor diffusion of sitting drops at 4°C for both concentrations. Initial crystals grew in a condition containing 0.2M MgCl₂, 0.1M Tris (pH 8.5), 30% PEG 4,000 and an optimized condition was identified by additive screening 4 mg/mL in 0.2 M MgCl₂, 0.1 M Tris (pH 8.5), 26.25% PEG 4,000, 35 mM spermidine followed by microseeding. Crystals were harvested and cryoprotected by supplementing the crystallization buffer with 10% glycerol.

Monochromatic X-ray diffraction data were collected at 0.975Å (FbpA-C) or 1.00Å (FbpB-Δ5C) wavelength using beamline 22-ID of the Advanced Photon Source (Argonne National Laboratory). Diffraction data were integrated, scaled, and reduced using the HKL2000 software suite (53) and assessed for data quality (54) ( Table 3 ). FbpA-C crystals grew in the space group P21 and contained two copies in the asymmetric unit which were related by twofold non-crystallographic symmetry. FbpB-Δ5C crystals grew in space group C2 and contained a single copy of FbpB-Δ5C in the asymmetric unit. Initial phases for FbpA-C were obtained by molecular replacement using a single copy of a poly-alanine model of BBK32-C (PBD:6N1L) generated with the program CHAINSAW via the CCP4 software package (55). Initial phases for FbpB-Δ5C were obtained by molecular replacement using a FbpA-C polyalanine model that was created using CHAINSAW from CCP4 (55). Following an initial round of refinement, models for FbpA-C and FbpB-Δ5C were completed by a combination of automated chain tracing using PHENIX.AUTOBUILD (56–58) and manual building into electron density maps using COOT (59). The final model was completed upon iterative cycles of manual rebuilding and refinement using PHENIX.REFINE (56–58). The N-terminal cloning artifact sequence for FbpA-C (Residues G-S-T-G-S) and FbpB-Δ5C, and residues 264-268, 355, 401-403, and 427 of FbpB-Δ5C were not modeled in the final refined structure due to poor electron density. Refined coordinates and structure factors have been deposited in the Protein Data Bank, Research Collaboratory for Structural Bioinformatics, Rutgers University (www.rcsb.org/) under PDB ID codes 7RPR for FbpA-C and 7RPS for FbpB-Δ5C. A description of crystal cell constants, diffraction data quality, and properties of the final models for FbpA-C and FbpB-Δ5C can be found in Table 3 . Representations of the protein structures and electron density maps were generated by PyMol (www.pymol.org/).

Far-UV spectra were collected to assess the secondary structure for FbpA-C, FbpB-C, FbpA DA-C, and FbpB-Δ5C. Samples were diluted to 10 μM in 10 mM Na₃PO₄ based on previous protocols (60). Spectra for a buffer control were used to subtract background noise. Spectra were collected across a wavelength range of 180-300 nm, at 120 nm min^-1, using 1 nm step, 0.5 sec response, and 1 nm bandwidth. Data was collected using a Chirascan V100 instrument with a square small volume quartz cuvette with a path length of 0.05 cm (Applied Photophysics, UK).

Analytical gel filtration was performed as before (34). 100 μg of FbpB-C and FbpB-Δ5C in a total volume of 500 μL were injected onto a Superdex 75 10/300 GL column (GE Healthcare) at a flowrate of 0.5 mL/min. The running buffer used was 10 mM HEPES (pH 7.3), 140 mM NaCl.

A colorimetric assay was used to monitor dose-dependent inhibition of C1r-CCP2-SP enzyme activity based off (33, 34). 1 μM Fbps-C were incubated with 15 nM C1r-CCP2-SP enzyme and 300 μM of the serine protease substrate Z-Gly-Arg-thiobenzyl (MP Biomedical) in HBS buffer (10 mM HEPES [pH 7.3] and 140 mM NaCl). Enzyme cleavage of Z-Gly-Arg-thiobenzyl was determined by incubating the reaction with 100 μM DTNB (5,5’-Dithiobis-[2-nitrobenzoic acid], Ellman’s reagent). The EnSight Multimode Plate Reader (PerkinElmer) was used to read the absorbance values for the reaction at 412 nm. Data were obtained in triplicate and normalized with 15 nM C1r-CCP2-SP lacking inhibitor as a positive control that represented the 100% amount of Z-Gly-Arg-thiobenzyl cleaved by C1r-CCP2-SP in 30 min at 25°C. Using a DTNB and Z-Gly-Arg-thiobenzyl only negative control allowed for subtraction of any non-specific colorimetric change.

An ELISA-based assay that utilizes IgM as a specific activator of the classical pathway as described (32–34, 50, 61), was used to determine the pathway’s downstream effect while in the presence of the Fbps. Data were obtained in triplicate and normalized using no inhibitor with 2% normal human serum (Innovative Research) as a positive control that represented the 100% amount of C4b deposition in 1 hour at 37°C. Non-specific interactions were subtracted out using a negative control with no serum. An EnSight Multimode Plate Reader (PerkinElmer) was used to read the absorbance values at 450 nm for the reaction.

A hemolysis assay that uses antibody-opsonized sheep red blood cells to specifically activate the classical pathway was performed as previously described (32, 33, 62) to determine if B. miyamotoi Fbp-C proteins protected erythrocytes from complement-mediated hemolysis. A serial dilution series of Fbp-C proteins was incubated with buffer-exchanged pre-sensitized sheep erythrocytes (Complement Technology) and 2% normal human serum (Innovative Research). After a 1 hour incubation at 37°C the reaction was clarified via centrifugation at 500 x g and the absorbance read using an EnSight Multimode Plate Reader (PerkinElmer) at 412 nm. Data were obtained in triplicate normalized using a no inhibitor with serum positive control that represented the 100% cell lysis at 37°C in 1 hour. Background hemolysis was subtracted out using a negative control with only buffer and red blood cells.

The interaction of B. miyamotoi Fbp proteins with C1r in human serum was measured using an ELISA-type binding assay based on (63). Fbp proteins were immobilized overnight onto high-binding 96-well plates (Greiner Bio-One) at 0.2 mg/mL in 100 mM NaCO₃ (pH 9.6). A twofold serial dilution of normal human serum (Innovative Research) ranging from 0.0012-10% serum in CP Assay buffer (10 mM HEPES [pH 7.3], 140 mM NaCl, 2 mM CaCl₂, 0.5 mM MgCl₂, 0.1% gelatin) was incubated with immobilized Fbp-C proteins at 37°C for 1 hour. The amount of serum C1r bound to immobilized Fbps was found using a goat antibody to human C1r (R&D Systems) diluted 1:2,000 and immune complexes detected with rabbit anti-goat Ig conjugated to HRP (Invitrogen) at a 1:3,000 dilution. Data were obtained in at least duplicate and read at 450 nm using the EnSight Multimode Plate Reader (PerkinElmer) with non-specific interactions subtracted out using a negative control with no serum.

B. burgdorferi strain B314 was grown to mid-log phase and harvested by centrifugation at 4,500 x g for 10 minutes at 4°C, and washed once with PBS. The cell pellet was resuspended in 0.8 mL of either PBS alone, or PBS with proteinase K to a final concentration of 200 μg/mL, and incubated at 20°C for 40 min. Reactions were terminated by the addition of phenylmethylsulfonyl fluoride (PMSF) to a final concentration of 1 mM. Cells were centrifuged (4,500 x g for 10 min at 4°C), washed twice with PBS, and resuspended in sample buffer. Samples corresponding to 5x10⁷ whole cell equivalents were run on SDS-PAGE gel, transferred to PVDF membranes and immunoblotted with mouse polyclonal antibodies to FbpA and FbpB or monoclonal antibodies directed against BBK32 or borrelial FlaB (US Biologicals Inc.) and imaged with chemiluminescence as previously described (33).

Assays were performed as in (33, 34). Briefly, B. burgdorferi strain B314 producing BBK32 (pCD100), B. miyamotoi FbpA (pAP8) or FbpA-DA (pAP13), and B. miyamotoi FbpB (pAP11) (see Table 1 ), as well as the vector-only control B314 pBBE22luc, were grown at 32°C in 1% CO₂ to early- to mid-log phase and diluted in complete BSK-II media to a final cell density of 1x10⁶ cells/mL. B. burgdorferi cells were then incubated in 15% normal human serum (Complement Technology) or heat inactivated serum (55°C for 30 min.), with thermal inactivation of complement proteins as a positive control for survival, for 1.5 hours at 37°C under gentle agitation. Cell survival was assessed by dark field microscopy based on cell motility and overt membrane damage or lysis.

The ability of exogenous B. miyamotoi FbpA-C, FbpA DA-C, and FbpB-C to rescue serum-sensitive B. burgdorferi B314 pBBE22luc was measured similarly to the serum sensitivity assay. Exogenous complement inhibitory proteins were added at a final concentration of 48 nM, 240 nM, and 1.2 μM, approximately five-fold less than, equal to, and five-fold greater than the concentration of C1r in the human serum sample used. Survival was compared to that of B314 pBBE22luc with no exogenous protein addition, and of B314 pCD100, which produces BBK32, as described for the serum sensitivity assays.

Data is shown as the average of the replicates with 95% confidence intervals. Half-maximal inhibitory concentrations (IC₅₀) values for the CP inhibitory ELISA and CP hemolysis assays were determined using a non-linear variable slope regression fit. The half-maximal effective concentration (EC₅₀) for the serum C1r binding ELISA was determined using a non-linear one-site total regression fit. One-way ANOVA was used for the C1r enzyme and competition SPR experiments to determine differences among the means, followed by a post-hoc Tukey analysis for pairwise comparison of each Fbp protein. For serum sensitivity and B314 “rescue” assays, two-way ANOVA with a Šidák correction for multiple comparisons was used. All statistical analysis was performed using GraphPad Prism version 9.3.

B. burgdorferi BBK32 interacts with multiple host proteins through non-overlapping protein domains. The N-terminus of BBK32 (BBK32-N) binds to fibronectin (35–37, 40, 64–70), while the C-terminal domain (BBK32-C) interacts with complement serine protease C1r (32, 33). Relapsing fever spirochetes encode a combination of up to three separate BBK32 orthologs termed fibronectin binding proteins (Fbps) (41) ( Figure 1A ). Fbp genes organize into three protein families denoted FbpA, FbpB, and FbpC and vary in their sequence identity to BBK32 between 56-62% (FbpAs), 25-30% (FbpBs), and 22-27% (FbpCs) ( Figure S1A ). A BLASTp search of the B. miyamotoi strain FR64b genomic sequence, using B. burgdorferi strain B31 BBK32 as a query, identified only two orthologous sequences; one belonging to the FbpA family and one to the FbpB family ( Figure 1A ). Unlike B. miyamotoi FbpA, FbpB lacks significant amino acid sequence homology for a key motif that is important for BBK32’s interaction with fibronectin ( Figure 1B ) (i.e., gelatin-binding domain sequence [GBD]) (69). Similarly, while a critical residue involved in the BBK32/C1r interaction is conserved in both FbpA and FbpB (i.e. BBK32-R248/FbpA-R264/FbpB-R309), another important ‘hot-spot’ lysine residue is conserved only in FbpA (BBK32-K327/FbpA-K343) ( Figure 1B ) (34).

We hypothesized that the sequence variation in predicted functional sites of FbpB, and to a lesser degree FbpA, may result in altered function of B. miyamotoi Fbps relative to BBK32. To test this, we first began by evaluating the direct interaction of each protein with human fibronectin. We expressed and purified recombinant maltose-binding protein (MBP) fusion proteins that encoded the predicted fibronectin-binding site based on alignment to BBK32 (termed MBP-FbpA-N and MBP-FbpB-N) ( Figure 1B , green box). Surface plasmon resonance (SPR) experiments were carried out by immobilizing each fusion protein on the surface of an SPR sensorchip and evaluating purified human fibronectin as an analyte in a dose-dependent manner. MBP-FbpA-N bound dose-dependently to soluble human fibronectin (K _D = 170 nM) ( Figures 2A , C ), whereas MBP-FbpB-N exhibited no detectable interaction with fibronectin at concentrations up to 500 nM ( Figures 2B , D ).

To exclude the possibility that FbpB may utilize a site outside of the predicted fibronectin-binding region to engage fibronectin, and to evaluate the interaction in the context of a full-length lipoprotein, we expressed intact fbpA and fbpB ectopically in B. burgdorferi strain B314. B. burgdorferi B314 is a serum-sensitive, high-passaged strain that has lost all linear plasmids including the 36 kilobase linear plasmid (lp36) that encodes bbk32 (46). Equivalent protein lysates from B. burgdorferi strain B314 derivatives were tested for their ability to produce native B. miyamotoi FbpA, a double alanine mutant of FbpA at residues 264 and 343—R264A and K343A—denoted FbpA-DA (discussed below), and FbpB, by SDS-PAGE and Western immunoblot analysis ( Figures S2A, B ). Whereas the antibody to FbpA recognized only the FbpA protein, antisera to FbpB recognized both FbpB and, to a lesser extent, FbpA ( Figure S2A ). We then tested these same samples using a Far Western approach with human fibronectin as the probe. Both FbpA and FbpA-DA in B314 bound human fibronectin whereas samples containing FbpB or the pBBE22luc shuttle vector control spirochetes exhibited no detectable binding ( Figure 2E ). Consistent with the SPR experiments ( Figures 2A –D ), recombinant MBP-FbpA-N fusion protein bound human fibronectin in this assay format, whereas the MBP-FbpB-N protein did not ( Figure 2E ). Analysis of total lysate samples by SDS-PAGE ( Figure S2B ) and FlaB protein by Western blot ( Figure 2E ) demonstrated equivalent protein levels were used for comparison. These results show that while B. miyamotoi FbpA retains BBK32-like fibronectin-binding activity, FbpB does not.

Next, we examined if B. miyamotoi FbpA and FbpB bound human C1r in a manner similar to B. burgdorferi BBK32 (32–34). BBK32 binds C1r via a C-terminal domain where residues R248 and K327 are critical in mediating the interaction with the S1 site of the C1r serine protease (SP) domain and the B-loop of C1r, respectively (34). Each of these key residues are conserved in FbpA (i.e., R264 and K343), whereas in FbpB, only the arginine is conserved (i.e., R309) ( Figure 1B ). We expressed and purified recombinant C-terminal truncation constructs of B. miyamotoi FbpA (FbpA-C), a site-directed double alanine mutant of FbpA-C (FbpA-R264A-K343A, denoted FbpA DA-C) ( Figure S2C ), and FbpB (FbpB-C), and assessed their direct interaction with a recombinant activated C1r construct encompassing the complement control protein 2 and serine protease domains (C1r-CCP2-SP). SPR binding assays showed that FbpA-C and FbpB-C bound with high affinity to C1r-CCP2-SP with K _D = 0.90 nM and 1.4 nM, respectively ( Figures 3A , B and Table 4 ), whereas FbpA DA-C exhibited no detectable C1r-CCP2-SP binding ( Figure 3C ).

Next, we determined if FbpA-C and FbpB-C could compete with one another for binding to C1r-CCP2-SP. Equimolar mixtures of C1r-CCP2-SP with soluble FbpA-C, FbpA DA-C or FbpB-C were injected over immobilized FbpA-C and FbpB-C. Significantly reduced binding signals were observed when soluble FbpA-C or FbpB-C was used, but not for FbpA DA-C, indicating that active FbpA and FbpB proteins are capable of competing with one another for binding to activated C1r ( Figures 3D–F ). These results suggest that both FbpA and FbpB bind to activated C1r in a manner similar to what has been reported for BBK32 and underscores the importance of the conserved R248 and K327 residues (BBK32 numbering) in the FbpA/C1r interaction (34). Interestingly, although BBK32-K327 is not conserved in FbpB ( Figure 1B ), this did not preclude high-affinity interaction with activated C1r-CCP2-SP.

B. miyamotoi FbpA-C and FbpB-C were next assessed for the ability to inhibit the classical complement cascade due ostensibly to the inhibition of C1r. A classical pathway-specific ELISA-based assay was implemented with Fbp-based inhibitors and normal human serum (32–34). As expected, in light of the C1r-CCP2-SP binding data ( Figure 3 ), FbpA-C demonstrated potent inhibition of classical pathway-mediated complement activation (IC₅₀ = 23 nM), whereas the FbpA DA-C mutant lost all inhibitory activity ( Figure 4A and Table 4 ). FbpB-C blocked complement activation but, surprisingly, was ~80-fold less potent than FbpA (IC₅₀ = 1,900 nM) despite having similar high affinity for activated C1r-CCP2-SP. To confirm these activities, each protein was serially diluted and incubated with opsonized sheep erythrocytes and normal human serum. Consistent with the ELISA-based assay, FbpA, but not FbpA DA prevented hemolysis ( Figure 4B ). In this assay format, FbpB-C had even lower activity, showing no inhibition at concentrations up to 4µM.

While differences in the ability of B. miyamotoi FbpA and FbpB to bind fibronectin are likely attributed to low sequence conservation of critical fibronectin-binding motifs in FbpB ( Figure 1B ) (36, 37, 66, 69, 70), the basis for differences in the complement inhibitory activities of FbpA-C and FbpB-C were less clear. To determine if FbpA and FbpB have altered structures in their C1r-binding C-terminal regions, we solved the crystal structures of each protein. FbpA-C crystals diffracted to a limiting resolution of 1.9Å (R_{free (%)}/R_{work (%)} = 20.6/23.2) (PDB: 7RPR) ( Figure 5A , Figures S3A, B and Table 3 ). FbpB-C was recalcitrant to crystallization; however, an active FbpB-C construct lacking five C-terminal residues ( Figure S4 , referred to as FbpB-Δ5C hereafter), was crystallized and the structure resolved to a limiting resolution of 2.1Å (R_{free (%)}/R_{work (%)} = 22.5/25.8) (PDB: 7RPS) ( Figure 5B , Figures S3C, D and Table 3 ).

A structural overlay with BBK32-C (PDB: 6N1L) and B. miyamotoi FbpA-C and FbpB-Δ5C reveals that all three proteins have similar folds involving a core four-helix bundle (helices 1, 3, 4, and 5) with alpha helix 2 extending away from the core ( Figure 5C ). BBK32-C was previously shown to interact with activated C1r using two primary binding sites: i) the active site of C1r involving several BBK32 residues, including R248, and ii) the B-loop of C1r involving residues on the fifth alpha helix, including K327 (34). A structural overlay of FbpA-C, FbpB-Δ5C, and BBK32-C shows that the three proteins retain an overall similar structure around the conserved R248 residue (BBK32 numbering) ( Figure 5D ). However, we noted that the area near BBK32-K327 involves a minor structural alteration in FbpA-C and a more major alteration in FbpB-Δ5C ( Figure 5E ). In FbpA-C, the kinked portion of alpha helix 5 extends one turn longer than in BBK32, resulting in FbpA-C containing an additional surface-exposed residue on this helix ( Figure 5E ). More strikingly, the structure of FbpB-Δ5C has a significantly altered secondary structure in this region when compared to both FbpA-C and BBK32-C ( Figure 5E ). Specifically, FbpB-Δ5C shows a loss of secondary structure within alpha helix 5 that accommodates a loop insertion sequence (i.e., residues Q394-Q405), which are not conserved in both BBK32-C or FbpA-C. Furthermore, a proline residue (P397) in this sequence interrupts the alpha helical secondary structure in FbpB-Δ5C. A portion of this unique loop structure in FbpB has apparent increased disorder in the crystal structure, indicated by weak electron density surrounding residues 401-403 in this region of the protein. Collectively, these data show that FbpA and FbpB are structurally divergent from BBK32 at a functionally critical region of the protein on the fifth alpha helix ( Figures 5C –E ).

C1r is a serine protease subcomponent of the C1 complex which circulates in blood as an inactive zymogen, and is converted to an activated enzyme by autocatalysis when C1 binds to an activating ligand (71–75). This transition involves a conformational change in the SP domain of C1r resulting in a rearrangement of several loop structures, including the C1r B-loop (75) ( Figure S5 ). We have previously shown that BBK32 binds to both active and zymogen forms of C1r in a manner that relies on R248 and K327 (34). Given the involvement of B-loop residues in the BBK32-K327 binding site, the observed structural differences at this site in the Fbp crystal structures ( Figure 5 ), and the potential for different conformations of this loop between zymogen and active states of C1r ( Figure S5 ) (75), we hypothesized that FbpA and FbpB may have altered recognition properties for zymogen C1r.

To assess this, B. miyamotoi FbpA-C and FbpB-C were immobilized on an SPR chip and a five-fold dilution series of either zymogen or active forms of full-length purified C1r were injected over each surface. Interestingly, while FbpA-C was able to bind to both zymogen and active forms of C1r readily (C1r zymogen K _D = 0.85 ± 0.036 nM, active C1r K _D = 0.39 ± 0.049 nM), FbpB selectively interacted with active C1r only (active C1r K _D= 1.1 ± 0.042 nM) ( Figure 6A ). We next confirmed that the selective nature of FbpB was observed with full-length proteins expressed in B. burgdorferi B314 strains in an approach identical to that presented for fibronectin in Figure 2E . Whereas FbpA could recognize both zymogen and activated forms of purified C1r, FbpB preferentially bound activated C1r. The negative controls, FbpA DA and vector-only lysates, did not bind either form of C1r as expected ( Figure 6B ), although we note that FbpA DA retained fibronectin binding ( Figure 2E ).

We considered whether the apparent zymogen/active C1r selectivity differences between FbpA and FbpB could be observed in serum where C1r is in a zymogen form (71–74). To test this, we implemented a serum-based ELISA-type binding assay. B. miyamotoi FbpA-C, FbpB-C, and FbpA DA-C were immobilized onto an ELISA plate and incubated with a dilution series of normal human serum followed by detection of C1r binding using a C1r antibody. FbpA-C readily bound serum C1r (half-maximal effective concentration (EC₅₀) = 0.27% (v/v) serum concentration), while FbpB-C and FbpA DA-C did not ( Figure 6C ). However, consistent with the ability of FbpB-C to bind tightly to previously activated forms of C1r ( Figures 3 , 6A ), FbpB-C inhibited the enzymatic activity of pre-activated C1r-CCP2-SP enzyme similar to FbpA-C ( Figure 6D ). Collectively, these results show that, like BBK32, B. miyamotoi FbpA can bind and inhibit both the zymogen and active forms of C1r, whereas B. miyamotoi FbpB selectively interacts with and inhibits only the active form of the C1r protease.

Next, we assessed the protection B. miyamotoi FbpA and FbpB conferred to B. burgdorferi B314 when ectopically expressed on the cell surface. Since B. miyamotoi has not yet been genetically manipulated, we modeled these proteins’ surface expression and activity in B314 relative to a pBBE22luc vector-only control and cells producing BBK32, given that BBK32 confers serum resistance to strain B314 (32–34). As a prelude to the serum sensitivity assay, we first tested whether B. miyamotoi FbpA, FbpA DA or FbpB were surface exposed in B. burgdorferi strain B314 relative to a B314 BBK32 control, as well as expressed on the surface of B. miyamotoi strain FR64b, using the well-established proteinase K accessibility assay (32, 76). The results clearly show that FbpA, FbpA DA, FbpB, and BBK32 were all produced and are all protease-sensitive, indicating surface exposure in strain B314. However, only FbpA is detected on the surface of B. miyamotoi strain FR64b, potentially due to the in vitro growth conditions not being conducive for expression of fbpB ( Figure 7A ). Importantly, the subsurface endoflagellar protein, FlaB, was unaffected by the treatment in both B. miyamotoi strain FR64b and B. burgdorferi B314, indicating the preserved structural integrity of the borrelial cells tested ( Figure 7A ). As observed in Figure S2A , antibody to B. miyamotoi FbpB also recognized FbpA and FbpA DA, albeit to a lesser extent ( Figure 7A ).

We then exposed each B314 knock-in strain to normal human serum and assessed the viability of the cells. B. miyamotoi FbpA conferred levels of resistance comparable to BBK32, while the FbpA DA strain was as sensitive to serum killing as the vector-only control ( Figure 7B ). In contrast to the lower apparent inhibitory activity of recombinant FbpB-C ( Figure 4 ), B. burgdorferi strain B314 producing FbpB on the surface also conferred significant protection against human serum ( Figure 7B ). In order to more fully address differences in the resistance observed for FbpB in the prior assays, we incubated serum sensitive B314 pBBE22luc with differential levels of B. miyamotoi FbpA-C, FbpB-C, FbpA DA-C, and B. burgdorferi BBK32-C to ask whether these soluble components could provide resistance to complement-dependent killing with normal human serum and thus “rescue” this sensitive B. burgdorferi strain. Exogenous addition of the C-terminus of the B. miyamotoi Fbps and B. burgdorferi BBK32-C resulted in dose-dependent rescue for BBK32-C, FbpA-C, and FbpB-C ( Figure 7C ). However, the inhibitory effect was weaker for FbpB-C, with significant inhibition observed only at the 1.2 µM treatment, which is comparable to the IC₅₀ value observed for FbpB-C in the ELISA-based inhibition assay ( Figure 4 ). Consistent with its inability to inhibit complement, addition of FbpA DA-C was unable to rescue serum-based killing of B314 pBBE22luc cells at any of the protein concentrations used ( Figure 7C ). Collectively, these data demonstrate that ectopically-produced B. miyamotoi FbpA and FbpB confer resistance to human serum.