Bacterial strains, biological and geographical origin, and culture conditions

Borrelia recurrentis A17 (human blood isolate, Addis Ababa, Ethiopia⁶⁴), B. recurrentis PAbJ (human blood isolate, Somalia²⁴), B. duttonii Ly (human blood isolate, Mvumi, Tanzania⁶⁴) and B. garinii G1 (CSF isolate, Germany⁶⁵) were cultured until mid-exponential phase (5 × 10⁷ cells per ml) at 33 °C in Barbour-Stoenner-Kelly (BSK-H) medium (Bio&SELL, Feucht, Germany) supplemented with 7.4% heat-inactivated rabbit serum (Merck, Darmstadt, Germany). All transformed B. garinii strains carrying the respective shuttle vectors were cultured in BSK-H medium supplemented with 100 μg/ml streptomycin (Merck). Escherichia coli strains NEB 5-alpha, BL21(DE3), BL21 Star (DE3) (New England Biolabs, Frankfurt, Germany) or M15 (Qiagen, Hilden, Germany) were used as hosts for propagation of respective vectors and for production of recombinant His-tagged proteins were grown at 37 °C in yeast tryptone (YT) broth containing 50 µg/ml ampicillin (Carl Roth GmbH, Karlsruhe, Germany) or 100 μg/ml streptomycin (Merck).

Human serum, proteins, and antibodies

NHS collected from healthy blood donors was initially tested for the presence of anti-Borrelia IgM and IgG antibodies by a commercially available ELISAs (Enzygnost Borreliosis/IgM and Enzygnost® Lyme link VlsE/IgG, Siemens Healthcare Diagnostics Products GmbH, Marburg, Germany)¹². Only sera considered to be negative were combined to form a serum pool. The total complement activity (CH50) of the serum pool was assessed by employing an ELISA-based assay (WiELISA) for measuring the activity of the classical pathway.

Complement components C1q, C2, C3, C3b, C4, C4b, C5, C9, FB, FH, FI, C4BP, and the C5b-6 complex were purchased from Complement Technology (Tyler, TX, USA). Human glu-plasminogen was purchased from Prolytix (Essex Junction, VT, USA). For the activation of plasminogen, urokinase plasminogen activator (uPA) and the chromogenic substrate D-Val-Leu-Lys p-nitroanilide dihydrochloride (S-2251) were used from Merck (Darmstadt, Germany). Polyclonal anti-plasminogen antibody was purchased from Acris Antibodies (Herford, Germany). Purified vitronectin were obtained from Merck.

The polyclonal anti-C5 antibody was obtained from Complement Technology (Tyler, Texas, USA). Polyclonal antisera raised against complement C3, C4, and FH were from Merck (Darmstadt, Germany) and polyclonal antisera against C1q, C5, FI, and FB as well as the neoepitope-specific monoclonal antibody against the C5b-9 complex were from Quidel (San Diego, USA). The polyclonal anti-plasminogen antibody was purchased from OriGene Technologies (Rockville, MD, USA). For the detection of His-tagged proteins, a mixture of monoclonal anti-His antibodies were used (GE Healthcare, Munich, Germany and Merck, Darmstadt, Germany). All horseradish peroxidase (HRP)-conjugated immunoglobulins were obtained from Agilent Technologies Denmark, Glostrup, Denmark. The monoclonal anti-HcpA and anti-CihC antibodies were described previously^15,16. The polyclonal rabbit anti-ChiB antibody was produced by a commercial provider (Eurogentec, Seraing, Belgium). For the detection of the periplasmatic FlaB protein, the monoclonal anti-FlaB Ab L41 1C11 was used⁶⁶. Proteinase K was purchased from Merck (Darmstadt, Germany) and Pefabloc SC was from Carl Roth (Karlsruhe, Germany).

Sequence and phylogenetic analysis

For comparative genomics, Borrelia genomes stored under BioProject PRJNA378726, as well as the B. recurrentis A1 genome stored under accession number NC_011244 were obtained from NCBI (Source data are provided as a Source data). All genome sequences were then uniformly annotated using Prokka 1.14.6. Searches for homologs were carried out based on resulting FASTA files using Blast p 2.9.0 and visualizations of genomic loci were created using CLC Genomics Workbench version 12.0.2. Lalign (EMBL-EBI, Hinxton, UK) was used for pairwise sequence alignment to calculate protein sequence identity and similarity.

To recognize canonical promoter motifs within the chi gene cluster, the YAPP Eukaryotic Core Promoter Predictor (https://www.bioinformatics.org) and Promotech (R. Chevez-Guardado & L. Peña-Castillo, 2021, https://doi.org/10.1186/s13059-021-02514-9) were used. The YAPP tool is created for TATA boxes, initiator elements (INR), downstream core element (DPE) in upstream eukaryotic but also for prokaryotic promoter sequences. YAPP algorithm calculates matrix similarity score for matches with consensus sequences to qualify as promoter elements. True positive predictions display score values that tend to be around 0.5 or higher when using Promotech, a machine-learning-based method.

For phylogenetic analysis, sequences of all respective proteins were aligned using Clustal Omega multiple sequence alignment (https://www.ebi.ac.uk/jdispatcher/msa/clustalo). Visualization of the tree was drawn using iTOL (https://itol.embl.de/). Sequence analysis of the inserted DNA fragments was performed by using the CLC Sequence Viewer 8.0 (QIAGEN Aarhus A/S, Denmark) and SnapGene Viewer version 7 (GSL Biotech LLC, San Diego, CA, USA).

Crystallization, data collection, and processing

For crystallization of ChiB, the protein was concentrated in 200 mM NaCl and 50 mM Tris (pH 8.0) to 10 mg/ml. Crystals were grown at 22 °C in sitting drops with the vapor diffusion technique, using a Honeybee 961 crystallization robot. In the drop, we mixed 0.2 µl protein solution with 0.2 µl reservoir solution (1.7 M ammonium citrate, 5 mM DTT). The used heavy atom derivative crystal was produced by soaking a crystal with 1 mM ethylmercury phosphate for 3 h. Before data collection, crystals were soaked in mother liquor with a final concentration of 25% glycerol. ChiA crystals were retained using the seed bead method. The initial crystals (protein concentration 10 mg/ml) were grown in 20% PEG400, 16% PEG4000, 70 mM MgCl₂, 100 mM Tris (pH 8.5) at 4 °C in sitting drops with the vapor diffusion technique, using a Honeybee 961 crystallization robot (Genomic Solutions). The seed crystals were grown at 22 °C in hanging drops. In the drop we mixed 0.2 µl protein solution with 0.2 µl reservoir solution. The reservoir contained 16% PEG4000, 20% PEG400, 50 mM MgCl, 100 mM Tris (pH 8.5). Before data collection, the crystals were soaked in 25% PEG4000, 25% PEG400, 50 mM MgCl₂, 100 mM Tris (pH 8.5), and 10% glycerol. Diffraction data for all crystals were collected at X10SA (detector: Pilatus or EIGER2 16 M) of the Swiss Light Source in Villigen, Switzerland.

The data were collected at 100 °K and processed with XDS⁶⁷. The tetragonal crystals of native ChiB diffracted up to 1.5 Å resolution and obeyed P4₂2₁2 space group symmetry with one monomer in the asymmetric unit. ChiA crystallized in a monoclinic space group (C2) with two monomers in the asymmetric unit and diffracted up to 2.7 Å. During refinement, we omitted 10% (ChiA) or 8% (ChiB) of the reflections, which were used for calculation of an R_free value.

Structure determination

The three-dimensional structure of ChiB was solved through a combination of molecular replacement and anomalous dispersion techniques. A search model was constructed using SWISS-MODEL⁶⁸ through homology modeling. The outer surface protein BBA66 from Borrelia burgdorferi (2yn7)²⁹, sharing a sequence identity of 14% with ChiB, was employed as a template for this purpose. However, owing to the limited sequence similarity, this approach did not yield a solution.

To address this issue, we soaked the ChiB crystals with ethylmercury phosphate, varying the duration and concentration of the soak. Only one dataset exhibited an anomalous signal, detectable up to 4 Å. Merging of the native and heavy atom data sets of ChiB was performed using XDS⁶⁷. The asymmetric unit of the tetragonal crystals contains a single monomer. Notably, ChiB has only one methionine and one cysteine residue, indicating a single heavy atom site. Utilizing AutoSol, implemented in the program package Phenix⁶⁹, the native data and anomalous scattering data from the ethylmercury phosphate dataset (sad peak, 1 site) yielded a Bayesian correlation coefficient of 16% and an R_free of 54%. Despite these poor values, three helices were visible in the electron density calculated from the results of the heavy atom dataset. We then successfully manually fitted a partial model (150 amino acids) based on 2yn7 to this helix density. The R_free value of this new solution was 41%. Subsequent application of autobuild (implemented in the program package Phenix)⁶⁹ to the high-resolution dataset (1.5 Å) resulted in a CC of 82% and an R_free of 26%. After several rounds of manual rebuilding and subsequent refinement, we revealed a model of ChiB, comprising residues E24 to N231 and G237 to Q284 with a final R_free of 21% and good stereochemistry. In addition to several solvent molecules, the initial Fo–Fc electron density map revealed clear density that could not be satisfactorily explained by buffer components present during purification and crystallization (Tris, citrate, glycerol, DTT) and was most consistent with a phospholipid-like molecule. The density contained a prominent peak consistent with a phosphate group and extended into two elongated hydrophobic chains with additional density features compatible with carbonyl groups. Based on its size and geometry, the ligand was modeled as LPP and refined with standard stereochemical restraints. As no lipid was introduced during purification or crystallization, the most plausible explanation is co-purification of a bacterial phospholipid during heterologous expression in E. coli.

The sequence identity between ChiA and ChiB is only 38.7%. However, the structure of ChiA could be solved by using the refined model of ChiB as a template for molecular replacement. The final ChiA model includes residues D29 to S213 and H222 to I265 and was refined to an R_free value of 30%. The final statistics of ChiA and ChiB are summarized in Table S2.

The PHENIX program suite served for reflection phasing and structure refinement⁶⁹. The interactive graphics program Coot⁷⁰ was used for model building, the superpositions of the structures were made using SSM Superposition⁷¹, which is implemented in the Coot program package or with routines included in the UCSF Chimera package⁷¹. Molecular graphics images were produced using the UCSF Chimera package³⁵.

Modeling of ChiA, ChiB, ChiC, ChiD, and ChiE by AlphaFold2

We used the protein sequences of ChiA, ChiB, ChiC, ChiD, and ChiE from B. recurrentis A1 (GenBank accession number NC_011244 and Source data file) to perform structural predictions using the multimer model of Alphafold2. A comparative analysis between the predicted rank-ordered PDB structures and the structures obtained using the Amber relax option revealed only minor differences. Therefore, we selected the best ranked amber-relaxed structures. To assess which regions of the predicted structures are trustworthy, the heatmaps obtained from Alphafold2 were analyzed using UCSF ChimeraX⁷². In addition, a structural comparison was made between the X-ray structures of ChiA and ChiB and the corresponding AlphaFold2 structures. All structures were obtained using the AlphaFold2 program⁷³ available at the MPI for Medical Research in Heidelberg, Germany; monomers were modeled with version v2.2.0 using the multimode option, and dimers were modeled with version v2.2.0 and v2.3.1.

Determination of free thiol groups using Ellman’s reagent (DTNB assay)

To quantify accessible thiol groups, 200 µl of purified Chi proteins (0.2–6 mg/ml) were incubated with 5 mM reducing agent (DTT or β-mercaptoethanol) for 30 min at 37 °C. Meanwhile, ZebaSpin 2 ml desalting columns (Thermo Scientific) were prepared according to the manufacturer’s instructions. Columns were placed in 15 ml collection tubes and centrifuged three times at 1000 × g for 2 min at 4 °C with 1 ml of 150 mM potassium phosphate buffer (pH 8.0) per wash. After equilibration, 200 µl of reduced protein sample mixed with 40 µl of phosphate buffer (pH 8.0) was applied to the column and centrifuged at 1000 × g for 2 min at 4 °C. The eluate represented the desalted, reduced protein.

For DTNB assays, 20–60 µl of this eluate was transferred to a 96-well half-area plate in duplicates. Wells were adjusted to 80 µl total volume with phosphate buffer (pH 8.0) and DTNB added to a final concentration of 400 µM. Absorbance was recorded at 412 nm immediately and at 1, 2, 5, 10, and 20 min using a Tecan plate reader. Protein concentrations were determined by Bradford assay on remaining sample. To assess spontaneous oxidation, the remaining reduced protein was stored at 4 °C for 4–5 days, then processed again as above for DTNB reactivity (Table S3).

Purification of His6-tagged proteins

To purify His₆-tagged proteins, E. coli cells carrying the appropriate plasmid were grown in YT-broth supplemented with 50 µg/ml ampicillin (Merck, Darmstadt, Germany) to an OD₆₀₀ of 0.5. Thereafter, protein production was induced by adding 0.2 mM IPTG for 4 h at room temperature and cells were then harvested by centrifugation. The sedimented cells were stored at −80 °C until use and then resuspended in lysis buffer (50 mM NaH₂PO₄, 300 mM NaCl, 10 mM imidazole, pH 6.8) supplemented with 1 mg/ml lysozyme (Merck, Darmstadt, Germany) and incubated for 30 min on ice. E. coli cells were then disrupted by homogenization using a MiCCRA D-9 disperser (Art Prozess- & Labortechnik GmbH; Heitersheim, Germany) following sonification six times for 30 s each with a Sonifier 450 (Branson Ultrasonics, Danbury, CT). After centrifugation, the supernatant was passed through a 0.45 µm filter and proteins were purified by affinity chromatography using NEBExpress Ni resin (New England Biolabs, Frankfurt, Germany) with increasing imidazole concentrations of 50 to 300 mM. For buffer exchange, 50 mM Tris (pH 8.0) combined with an ultrafiltration centrifugal device (cut off 10,000, Pierce) was used. The purity and size of recombinant proteins were then analyzed by subjecting 20 µl of the column eluates on a 10% Tris/Tricine SDS-PAGE following silver staining⁴⁰. The protein concentration of each protein was determined by employing the Pierce BCA protein assay kit (Thermo Fisher Scientific, Rockford, IL, USA).

For the determination of free thiol groups (see below), E. coli cells producing Chi proteins were resuspended in lysis buffer containing 0.5 mM DDT. After incubation on ice for 30 min, cells were disrupted by homogenisation and sonication. Following centrifugation, proteins were purified by IMAC in the presence of 0.5 mM DTT with increasing imidazole concentrations of 50 to 300 mM. To obtain reduced proteins, fractions collected were concentrated in the presence of 50 mM Tris (pH 8.0) containing 0.5 mM DDT by using ultrafiltration centrifugal devices and proteins were then stored at −20 °C before use.

SDS-PAGE and Western blot analysis

Purified His₆-tagged proteins or whole cell lysates were separated to 10% Tris/Tricine SDS-PAGE under reducing conditions and transferred to nitrocellulose membranes¹⁸. Briefly, the membranes were blocked with 5% nonfat dry milk in TBS containing 0.1% Tween 20 (TBS-T). After three wash steps with TBS-T, membranes were incubated with appropriate antibodies followed by horseradish peroxidase-conjugated anti-mouse or anti-rabbit immunoglobulins. Protein-antigen complexes were detected by tetramethylbenzidine as substrate. Images of the gels and nitrocellulose membranes were processed by using a GS-900 calibrated densitometer (Bio-Rad, Hercules, CA, USA) and the Image Lab version 6.1 (Bio-Rad).

Enzyme-linked immunosorbent assay

To detect binding of complement components or plasminogen, microtiter plates (Nunc MaxiSorp, Thermo Fisher Scientific) were coated with 100 µl of purified His₆-tagged proteins (5 µg/ml) or BSA (5 µg/ml) in PBS at 4 °C overnight⁷⁴. Between every incubation step, wells were washed three times with PBS containing 0.05% (v/v) Tween 20 (PBS-T). After blocking with Blocking Buffer III BSA (AppliChem, Darmstadt, Germany) or with PBS containing 0.2% gelatine (w/v) (AppliChem, Darmstadt, Germany), complement components (5 µg/ml each) or glu-plasminogen (10 µg/ml) in PBS was added. Binding of complement components or glu-plasminogen were then assessed by utilizing specific primary antibodies (dilution 1:1000). Following incubation for 1 h at RT, HRP-conjugated anti-goat or anti-mouse IgG (dilution 1:1000) were added and protein complexes were visualized using o-phenylenediamine (Merck, Darmstadt, Germany). The absorbance was read at 490 nm employing the PowerWave HT spectrophotometer (Bio-Tek Instruments, Winooski, VT, USA).

To determine dose-dependency, borrelial His₆-tagged proteins were immobilized (5 µg/ml) and incubated with increasing amounts of the respective complement components or glu-plasminogen. The antigen-antibody complexes were detected by using appropriate anti-complement antibodies as described above.

Complement inactivation assays

A modified ELISA-based approach (WiELISA) was applied to assess the inhibitory capacity of bacterial proteins on the alternative (AP), classical (CP), Lectin pathway (LP)⁷⁵. Briefly, Nunc MaxiSorp 96-well microtiter plates were coated with either LPS (10 µg/ml) (Hycult Biotech, Beutelsbach, Germany) for the AP, human IgM (3 µg/ml) (Merck, Darmstadt, Germany) for the CP, or mannan (100 µg/ml) (Merck, Darmstadt, Germany) for the LP at 4 °C overnight. Following three wash steps with TBS containing 0.5% (v/v) Tween 20 (TBS-T), the wells were blocked with PBS-T containing 1% BSA for 1 h at RT. NHS (15% for the AP, 1% for the CP, and 2% for the LP) was then pre-incubated with a final concentration of 4 µM (initial analyses) or increasing concentrations (0.5, 1, and 4 µM) (dose dependence analyses) of purified His₆-tagged proteins for 15 min at RT before being added to the wells to initiate complement activation of the respective pathway. After washing with TBS-T, a neoepitope-specific, monoclonal anti-C5b-9 antibody (1:500) was added to detect formation of the MAC as the final activation step of the cascade. Following incubation for 1 h at RT, wells were washed thoroughly with TBS-T and incubated with HRP-conjugated anti-mouse immunoglobulins (1:1000) at RT for 1 h. All reactions were developed applying tetramethylbenzidine as substrate.

In order to examine the inhibitory potential of the Chi proteins on the terminal pathway, a hemolytic assay was conducted⁷⁵. Briefly, sensitized sheep erythrocytes (1.5 × 10⁷ cells) (kindly provided by Dr. Michael Kirschfink (emer.), Institute of Immunology, Heidelberg, Germany) were pre-incubated with C5b-6 (1.5 μg/ml) for 10 min at RT. In parallel, complement C7 (2 μg/ml), C8 (0.4 μg/ml), and C9 (2 μg/ml) were pre-incubated with or without purified His₆-tagged proteins (0.5, 1, and 2 μM) for 5 min at RT. The pre-incubated proteins were then added to the C5b-6 coated sheep erythrocytes. Following incubation for 30 min at 37 °C, erythrocytes were sedimented by centrifugation and the supernatants were transferred to a microtiter plate. The hemolysis of the erythrocytes was then determined by measuring the absorbance of the supernatants at 414 nm.

Concerning the controls utilized, previously characterized His-tagged proteins originated from LD or RF borreliae as well as Acinetobacter baumanii were produced and purified using the same protocol as described above. These control proteins were selected according to their capability to inhibit the respective complement pathway as follows: BGA66 from B. bavariensis¹³ (AP inactivation), CihC from B. recurrentis¹⁸ and the C-terminal fragment of BBK32 (BBK32₂₀₅) from B. burgdorferi¹⁸ (CP inactivation), CipA from Acinetobacter baumanii⁷⁶ (LP inactivation), and CspA from B. burgdorferi¹¹ and CihC from B. recurrentis¹⁸ (TP inactivation). As negative controls, BtcA from B. turicatae³⁸ (AP inactivation), BDU1066 from B. recurrentis (CP inactivation), and Vsp1 from B. miyamotoi³⁷ (LP inactivation), and HcpA from B. recurrentis¹⁵ (TP inactivation) were chosen. In addition, vitronectin (Vn) was included in the cell-based hemolytic assay as a natural inhibitor of the TP as it binds to the preassembled Cb5-7, C5b-8, and C5b-9 complexes and thereby prevent MAC formation⁷⁷.

Determination of the inhibitory capacity of Chi proteins on C9 polymerization

To assess the inhibitory capacity of the purified borrelial proteins on C9 polymerization, increasing concentrations (final concentrations 0.005 µg/µl to 0.22 µg/µl or 0.2 µM to 9 µM) of the Chi proteins of B. recurrentis, CspA of B. burgdorferi (positive control), and BSA (negative control) was incubated with C9 (0.06 µg/µl or 0.9 µM) for 40 min at 37 °C^11,13. Thereafter, auto-polymerization of C9 was induced by adding 50 μM ZnCl₂ to each reaction mixture and incubated for 2 h at 37 °C. As additional controls, purified C9 was incubated with or without ZnCl₂. Reaction mixtures were then subjected to 8% Tris/Tricin-SDS gels and monomeric and polymeric C9 molecules were visualized by silver staining.

Plasmin(ogen) activation assay

Activation of glu-plasminogen by uPA and cleavage of the chromogenic substrate D-Val-Leu-Lys-p-nitroanilide dihydrochloride was entirely described previously^40,78. In brief, microtiter plates were immobilized with 100 µl of His₆-tagged proteins or BSA (5 ng/µl each) in PBS overnight following incubation with 10 ng/µl of glu-plasminogen for 1 h at room temperature. After three wash steps, each well was incubated with 0.3 µg/µl S-2251 in 50 mM Tris/HCl (pH 7.5), 300 mM NaCl, and 0.003% Triton X-100. Finally, 4 µl of 2.5 ng/µl urokinase plasminogen activator (uPA) were added to each well to activate protein-bound plasminogen. Further reactions containing 50 mM tranexamic acid, a lysine analogue with high affinity to the lysine binding sites of plasminogen to assess the role of lysines on binding of plasminogen to Chi proteins. As additional controls, reaction mixtures were prepared in which plasminogen or uPA were omitted. For long time measurement (24 h), microtiter plates were sealed, placed in an ELISA reader and incubated at 37 °C. The absorbance was measured every 30 min at 405 nm. To calculate the dissociation constant for the binding of plasminogen to Chi proteins, the non-linear regression model (four parameters dose-response curve) with a variable slope (Hill slope) was selected in GraphPad Prism 10.2.2.

C3b degradation by activated plasmin bound to Chi proteins

Chi proteins, BBA70 of B. burgdorferi, Vsp1 of B. recurrentis, and BSA (10 ng/µl each), respectively, were immobilized on microtiter plates in PBS overnight. After three wash steps, each well was incubated with glu-plasminogen (10 ng/µl) for 1 h at room temperature. Following washing, uPA (25 ng/µl) and C3b (20 ng/µl) were added to each well and reactions were incubated overnight at 37 °C. Samples were separated by Tris/Tricine-SDS-PAGE and transferred onto a nitrocellulose membrane. C3b cleavage products were detected by Western blotting employing a polyclonal anti-C3 antibody.

Generation of expression and shuttle vectors for the electroporation of spirochetes

The generation of vectors producing N-terminally His₆-tagged proteins used as controls for this study was previously described and includes CihC and HcpA of B. recurrentis A17^15,16, BtcA of B. turicatae³⁸, CbiA and Vsp1 of B. miyamotoi LB-2001^12,37, CspA and BBA70 of B. burgdorferi LW2^40,74, BGA66 of B. bavariensis PBi¹³, and CipA of Acinetobacter baumannii 19606⁷⁹. As a further control, a C-terminal fragment of BBK32 of B. burgdorferi B31 known to inhibit CP activation⁸⁰ was generated by PCR. First, the bbk32 gene was amplified using primers BBK32 Bam_FP and BBK32 Hind_RP (Table S5), and vector pMal-c/BBK32 as template (kindly provided by Yi-Pin Lin, Department of Infectious Diseases & Global Health, Cummings School of Veterinary Medicine, Tufts University, North Grafton, USA). The amplified DNA fragment was digested with appropriate restriction endonucleases and re-cloned into the expression vector pQE-30 Xa (Qiagen, Hilden, Germany). The resulting plasmid pQE-BBK32 served as template and oligonucleotides BBK32-205 BamHI and pQE-RP (Table S5) for a subsequent PCR amplification to engineer a His-tagged BBK32₂₀₅ fragment containing the C-terminal amino acids 205 to 356. After digestion, the DNA fragment was ligated into pQE-30 Xa and the resulting plasmid was transformed into E. coli BL21 Star (DE3) cells. In addition, BDU1066 located on lp165 of B. duttonii Ly and supposed to display similar complement-inhibitory functions as the factor H-binding CspA protein of B. burgdorferi B31²⁵ was also included in this study as a further control. The open reading frame (lacking the putative lipoprotein signal sequence) of the BDU1066 encoding gene of B. duttonii Ly was amplified by PCR using oligonucleotides Bre_1066_FP_Bam and Bre_1066_RP_Sal (Table S5). After digestion with BamHI and SalI, the DNA fragment was cloned into pQE-30 Xa. The resulting plasmid pQE-Bdu_1066 was then used to transform E. coli BL21 Star (DE3) cells. Plasmids isolated from selected clones were sequenced to ensure that no mutations were incorporated during PCR and the cloning procedure. Expression vectors encoding for N-terminally His₆-tagged ChiA, ChiB, ChiC, ChiD, and ChiE of B. recurrentis A17 were kindly provided by Reinhard Wallich (emer.), Institute of Immunology, University of Heidelberg, Germany. In addition, to increase the yield and purity of CihC, the encoding gene was re-cloned into the pET-16b expression vector (Merck, Darmstadt, Germany) by using primers CihC_Nde_FP and CihC_Bam_RP (Table S5). The resulting vector pET-CihC producing a N-terminally located His₆-tagged CihC protein was then transformed into E. coli M15 cells according to the manufacturer’s instructions.

To introduce deletions as well as single and double amino acid substitutions in ChiB, site-directed mutagenesis was conducted⁴⁰. Briefly, PCR was carried out for 18 cycles (95 °C for 15 s, 60 °C for 15 s and 72 °C for 90 s) using 50 ng/µl pQE-ChiB, 125 ng each of the oligonucleotides (Table S5), and 4 U PCRBIO VeriFi polymerase (PCR Biosystems, London, UK). Following incubation with 10 U DpnI (New England Biolabs, Frankfurt, Germany) to eliminate the remaining vector DNA, reactions were used to transform E. coli NEB 5-alpha cells. To generate truncated ChiB proteins lacking either the N- or the C-terminus or a ChiB protein lacking the loop region, PCR was performed with specific oligonucleotides (Table S5) and the PCRBIO HiFi polymerase (PCR Biosystems, London, UK) applying the identical conditions as stated above.

Shuttle vectors harboring the genes encoding for ChiA, ChiB, ChiC, ChiD or ChiE were generated by PCR amplification of the respective genes along with their potential promoters at the 5´end. Each gene was amplified by using genomic DNA from B. recurrentis A17 as template with oligonucleotides listed in Table S5. Following amplification and digestion with the respective endonucleases, each DNA fragment was cloned into the shuttle vector pKFSS1. Plasmids were prepared from presumptive E. coli clones with the Monarch plasmid kit (New England Biolabs, Frankfurt, Germany) and DNA inserts were Sanger sequenced by a commercial provider (Eurofins Genomics, Ebersberg, Germany).

Transformation and characterization of serum-sensitive B. garinii strains ectopically producing individual Chi protein of B. recurrentis A17

A high-passage, non-infectious B. garinii strain G1 selected as surrogate strain was grown in 100 ml BSK-H medium and harvested at mid-exponential phase (5 × 10⁷ to 1 × 10⁸ cells/ml). Electrocompetent cells were prepared as described previously with slight modifications⁷⁵. Briefly, 50 µl aliquots of competent B. garinii G1 cells were electroporated at 12.5 kV/cm in ice-cold 2-mm cuvettes with 20 µg of plasmid DNA. After electroporation, spirochetes were immediately transferred into 10 ml of BSK-H medium without antibiotics. Following incubation for 18 h at 33 °C, the cell suspension was further diluted by adding 90 ml BSK-H medium containing streptomycin (25 µg/ml) and 200 µl aliquots were seeded into 96-well cell culture plates. After six to eight weeks, streptomycin-resistant clones were macroscopically detected by a color change of the BSK-H medium. Individual clones selected were expanded in 1 ml of fresh BSK medium without antibiotic selection for 7 days, and then transferred into 10 ml of fresh BSK-H medium containing streptomycin (50 µg/ml). Selected clones were then further characterized by amplifying the inserted genes using primers M13 For and M13 Rev (Table S5). To confirm that no mutations were introduced during selection of positive clones, whole genomic DNA was isolated from the transformed spirochetes using the QIAamp DNA Mini kit (Qiagen, Hilden, Germany). The purified DNA containing the shuttle vectors were used for transformation of E. coli NEB 5-alpha cells (New England Biolabs, Frankfurt, Germany) and plasmids purified from selected clones were sequenced as mentioned above.

Gene expression analyses

The RNA from spirochetes grown at mid-logarithmic phase was extracted by using the RNAprotect Bacteria Reagent and the RNeasy Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer´s instructions. Thereafter, the purified RNA was treated twice with RNAse-free DNase, and proteins were removed by applying the Monarch RNA Cleanup kit (New England Biolabs, Frankfurt, Germany). The cDNA synthesis of the isolated RNA (1 µg) was then performed by using the LunaScript RT Supermix Kit (New England Biolabs, Frankfurt, Germany) following qPCR with 50 ng of cDNA and 10 µM of appropriate oligonucleotides (Table S5) according to the manufacturer´s instructions. Briefly, reactions were incubated initially at 95 °C for 60 s followed by 40 to 45 cycles of 95 °C (15 s), 60 °C (30 s), and 60 to 95 °C on a Roche Lightcycler 480 (Roche Diagnostics, Rotkreuz, Switzerland). Results were then analyzed using LinregPCR software⁸¹, and the relative values of the genes of interest (chiA, chiB, chiC, chiD, chiE, cihC, hcpA, flaB, 16S rRNA) were compared to the values of the control sample (cDNA synthesized from spirochetes carrying the empty shuttle vector) by employing the 2^ΔCt-method.

Serum protection assay

Protection from complement-mediated lysis mediated by recombinant proteins was assessed by pre-incubation of 25 µl NHS with Chi proteins, BGA66 of B. bavariensis PBi¹³ or BSA (10 µM each) for 15 min at 37 °C with gentle agitation. The pre-incubated serum samples were then adjusted to 100 µl with BSK-H medium. As additional reaction mixtures, native NHS (not pre-treated), heat-inactivated NHS and a Tris/HCl-buffer control were also included. In parallel, 1 × 10⁷ spirochetes of serum-sensitive B. garinii G1 were sedimented by centrifugation and resuspended in either the pre-treated serum samples or the controls. All reaction mixtures were then incubated for 4 h at 37 °C with gentle agitation. The percentage of motile and viable cells was determined by dark field microscopy after 4 h, respectively. Spirochetes in nine microscopy fields were counted by using Glasstic slides 10 (KOVA International Inc., CA, USA). At least three independent biological replicates were performed and ±SEM was determined by using GraphPad Prism version 7.

Serum bactericidal assay

Spirochetes grown at mid-logarithmic phase were sedimented by centrifugation and resuspended in 500 µl BSK-H medium. Reaction mixtures consisting of 75 µl highly viable spirochetes (1 × 10⁷) and 25 µl of NHS (30%) were incubated at 37 °C with gentle agitation. The percentage of motile cells was determined as described above.

In situ proteinase K treatment and immunofluorescence microscopy

To obtain complement-susceptible B. recurrentis A17 cells, a protease accessibility assay was performed. Spirochetes (6 × 10⁶ cells) were incubated with or without proteinase K (200 µg/ml) for 40 min at RT and the proteolytic activity was then terminated by adding Pefabloc SC (3 mM). Following sedimentation, cells were carefully washed twice with PBS and resuspended in 100 µl PBS containing 1% BSA (PBSA). Proteinase K-treated spirochetes were either lysated by sonication for Western blot analysis or incubated for 30 min at 33 °C with either 50 µl NHS or 50 µl hiNHS. After sedimentation, spirochetes were diluted 1:20 in PBSA and aliquots of 12 µl were spotted on diagnostic slides (Waldemar Knittel Glasbearbeitungs GmbH, Braunschweig, Germany). Slides were allowed to air-dried overnight and thereafter incubated for 10 min at RT with 40% glyoxal solution. After fixation, slides were incubated for 1 h at 33 °C in a humidified chamber with either a polyclonal anti-C3 antibody (dilution of 1:1000) or a monoclonal C5b-9 antibody (dilution of 1:50), respectively. Following four washes with PBS, the slides were incubated for 1 h at 33 °C with 1:2000 dilutions of Alexa 488-conjugated secondary antibodies (Life Technologies, Carlsbad, CA, USA). To stain Borrelia DNA, slides were washed four times with PBS and incubated with 40 µl of a DAPI solution (2 µg/ml) for 10 min at 4 °C. After mounting with fluorescence mounting medium (Dako), complement components deposited on the spirochetal surface were visualized by using an Axio Imager M2 fluorescence microscope (Zeiss, Oberkochen, Germany) equipped with a Spot RT3 camera (Visitron Systems, Puchheim, Germany). Fluorescent images were acquired by using 63x objective (Zeiss Plan-Apochromat) and processed and analyzed using Visiview (Visitron Systems GmbH, Puchheim, Germany).

Statistical analysis

Statistical analyses were performed using one-way ANOVA followed by Bonferroni’s post hoc test, using GraphPad Prism (GraphPad Software, San Diego, CA, USA). A p value of p p p p 

Ethics declarations

Collection of blood samples and consent documents was approved by the ethics committee at the University Hospital of Frankfurt (control numbers 160/10 and 222/14), Goethe University of Frankfurt am Main. All healthy blood donors provided written informed consent in accordance with the Declaration of Helsinki.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.