All experiments were done according to the animal use protocol approved by Texas A&M University Institutional Animal Care and Use Committee (IACUC) (AUP 2018-001 and 2020-0089) that meets all federal requirements, as defined in the Animal Welfare Act (AWA), the Public Health Service Policy (PHS), and the Humane Care and Use of Laboratory Animals.

The IxsS41(accession #EEC14235.1 and XP_002402368.4) sequence was obtained from GenBank. Cognizant of the fact that there are some errors in sequences that have been deposited in GenBank, nested 3′ prime rapid amplification of cDNA ends (RACE) primers (Forward: 5′GGTCGGTGGCGTTGTCTCGGGAGGTA3′ [used with the company provided Universal Primer A Mix (UPM)] and Forward Nested: 5′ATGAAGACTCTGGCAGCATTCCTGTC3′) [used with the company provided Universal Primer Short] were designed based on the EEC14235 sequence in GenBank. We designed long PCR primers to match the high annealing temperatures of universal primers in the kit. The RACE template was synthesized from available fed adult ticks using the SMARTer Clontech 3′ RACE kit according to instructions by the manufacturer (Takara Bio, San Jose, CA, USA). The primary and nested IxsS41 specific forward primers were used with universal primers in the Clontech kit. The nested PCR product was cloned into pGEMT cloning vector (Promega Corporation, Madison, WI, USA) and processed for DNA Sanger sequencing using T7 and SP6 promoter primers.

Using the sequence analysis application MacVector (North Carolina, USA), the translated amino acid sequence of IxsS41 that was cloned in this study was compared to EEC14235.1 and XP_002402368.4 amino acid sequences from GenBank. To gain insight into the relationship of IxsS41 with other tick serpins in GenBank, 22 tick proteins that showed at least 50% amino acid identity to IxsS41 were downloaded alongside an outlier Homo sapiens alpha-1-antitrypsin precursor (NP_001121178.1). These serpin sequences were used to construct the phylogeny tree by the neighbor-joining method in MacVector with parameters set to detect absolute differences and bootstrap values at 1,000 replications. The 22 serpin sequences that were downloaded include Dermacentor andersoni (n = 1), Dermacentor silvarum (n = 1), Ixodes persulcatus (n = 2), Ixodes ricinus (n = 3), Ixodes scapularis (n = 13), Rhipicephalus microplus (n = 1), and Rhipicephalus sanguineus (n = 1). Additionally, N-glycosylation and O-glycosylation site prediction servers, NetnGlyc-1.0, and NetOGlyc-4.0 (DTU Health Technology, Lyngby, Denmark) were used to predict N- and O-linked glycosylation sites of rIxsS41.

The expression plasmid for mature rIxsS41 containing a hexa-histidine tag at the C-terminal end was custom synthesized and cloned into the pPICZαA plasmid (Biomatik, Cambridge, Canada). Expression in Pichia pastoris X-33 cells and affinity purification of rIxsS41 were done as published (Tirloni et al., 2019; Kim et al., 2020a). The pPICZαA plasmid and X-33 cell expression system secretes the recombinant protein into culture media, and the expressed rIxsS41 was precipitated by ammonium sulfate saturation. For affinity purification, precipitates were resuspended and dialyzed against column binding buffer (20 mM Tris-HCl, 50 mM NaCl, and 5 mM imidazole, pH 7.4) and the expression of rIxsS41 was confirmed by Western blot using mouse anti-hexa-histidine tag 1:5,000 diluted (GenScript Biotech, Piscataway, NJ, USA) and silver-stained following manufacturer instructions (Pierce Silver Stain Kit, ThermoFisher Scientific, Waltham, MA). The rIxsS41 was subsequently purified under native conditions using a HiTrap Chelating HP column (Cytiva Life Sciences, Marlborough, MA, USA) and Tris-HCl-Imidazole buffer as published (Tirloni et al., 2019; Kim et al., 2020a). Buffer exchange into Tris-HCl buffer (20 mM Tris-HCl and 150 mM NaCl, pH 7.4) and volume reduction of affinity-purified rIxsS41 were done with a Pall Corporation Microsep Advance Centrifugal Device (30-kDa molecular weight cutoff) (Pall Corporation, Port Washington, NY, USA). The rIxsS41 concentration was quantified using the Pierce BCA Protein Assay Kit (ThermoFisher Scientific).

Preliminary SDS-PAGE and silver staining analysis revealed that rIxsS41 was migrating as a diffuse band, suggesting that it was glycosylated. Thus, to confirm this, affinity-purified rIxsS41 was subjected to deglycosylation using NEB Protein Deglycosylation Mix II, which removes both N- and O-linked glycans under both denaturing and non-denaturing conditions following the company-provided protocol (NEB, Inswic, MA, USA).

Preliminary silver staining analysis of deglycosylated rIxsS41 (under denaturing and non-denaturing conditions) showed multiple bands. To confirm the identity of multiple bands, both deglycosylated and glycosylated versions of rIxsS41 (7 µg each) were resolved on a 7.5% SDS-PAGE gel and silver stained using the Pierce Silver Stain for Mass Spectrometry (ThermoFisher Scientific). Single bands were then excised and submitted to the University of Florida Interdisciplinary Center for Biotechnology Research (ICBR) Proteomics & Mass Spectrometry Facility for LC-MS/MS analysis using the in-gel digestion approach (45 mM dithiothreitol [DTT], alkylated with 100 mM chloroacetamide [2-CAA], and trypsin-Lys-C). Using the Mascot program (Matrix Science, London, United Kingdom; version 2.7.0), tandem mass spectra were searched against the database of Pichia pastoris proteins (16,167 entries) and the amino acid sequence of IxsS41. Protein identifications were accepted if they could be established at greater than 95.0% probability and contained at least 2 identified peptides. Protein probabilities were assigned by the Protein Prophet algorithm (Nesvizhskii et al., 2003) and those that contained similar peptides and could not be differentiated based on MS/MS analysis alone were grouped to satisfy the principles of parsimony.

To determine if native IxsS41 elicits an antibody response in rabbits, an empirically optimized rIxsS41 amount (2 µg) was resolved on a 10% SDS-PAGE gel and transferred onto an Immobilon-P PVDF Membrane (Sigma-Aldrich, Burlington, MA, USA). Sera of rabbits that were repeatedly infested (or fed on twice) with uninfected and B. burgdorferi-infected I. scapularis nymph ticks (Ibelli et al., 2014; Bakshi et al., 2018; Kim et al., 2021) diluted at 1:200 and 1:1,000 were used to probe PVDF membranes at 4°C overnight. The membranes were thoroughly washed in phosphate buffered saline (PBS)-Tween 20, 0.05% (T), and 1:2,000 diluted goat anti-rabbit IgG HRP-conjugated antibody (SouthernBiotech, Birmingham, AL, USA) was used as secondary antibody for 1 h at room temperature. To detect the positive signal, washed PVDF membranes were incubated with the SuperSignal West Dura Extended Duration Substrate (ThermoFisher Scientific) for 10 min at room temperature and signal was detected using ChemiDocXRS+ imager (Bio-Rad, Hercules, CA, USA).

For ELISA analysis, high-binding 96-well plates (Corning, NY, USA) were coated with increasing amounts of rIxsS41 (0.12–1.00 µg) in triplicates overnight at 4°C. The following day, the wells were washed three times with PBS-T and blocked with 5% skim milk powder in PBS-T for 1 h at room temperature. Subsequently, wells were incubated overnight at 4°C with serially diluted rabbit serum 1:200, 1:400, 1:600, 1:800, and 1:1,000. Preliminary Western blotting analysis revealed that binding intensity of pre-immune sera was diminished in diluted antibodies. Thus, pre-immune sera was only tested at 1:200. Following washing, wells were incubated with 1:2,000 diluted goat anti-rabbit IgG HRP-conjugated antibody (SouthernBiotech) for 1 h at room temperature. Following another round of washing, the 1-Step Ultra TMB-ELISA substrate (ThermoFisher Scientific) and 2N sulfuric acid (VWR, Radnor, PA, USA) were added and A_450nm was recorded with the Synergy H1 plate reader (BioTek Instruments Inc., Winooski, VA, USA).

The inhibitory activity of rIxsS41 was profiled against 18 mammalian serine proteases involved in host defense mechanisms against tick feeding. Excess amount of rIxsS41 (1 µM in 20 mM Tris-HCl, 50 mM NaCl, and Tween 0.1%, pH 7.4) was pre-incubated with each serine protease for 15 min at 37°C followed by the addition of its respective substrate (200 µM). Hydrolysis of the substrate was measured every 11 s for 15 min at 30°C at A _405nm using the Synergy H1 plate reader (BioTek Instruments Inc.) as previously described (Horvath et al., 2011; Kim et al., 2020a) and data were analyzed with published formulas in Horvath et al. (2011).

Additionally, the in silico protein-to-protein server PSOPIA (prediction server of protein-to-protein interactions; https://mizuguchilab.org/PSOPIA/) (Murakami and Mizuguchi, 2014) was used to predict interactions between IxsS41 and the 18 mammalian serine proteases. Interactions with an Averaged One-Dependence Estimators (AODE) of more than 0.75 were considered significant (Murakami and Mizuguchi, 2014).

The stoichiometry of inhibition (SI) of rIxsS41 was determined for proteases whose enzymatic activity was inhibited by more than 80%. Various concentrations of rIxsS41 were incubated with a constant concentration of each protease for 1 h at 37°C. Colorimetric substrates were used to measure residual enzymatic activity and data were plotted as previously described (Horvath et al., 2011; Kim et al., 2015; Tirloni et al., 2019; Kim et al., 2020a). Forthwith, the second-order rate constant (k _a) for the inhibition of chymase and cathepsin G (proteases whose SI values were below 7) by rIxsS41 was determined using the discontinuous method (Horvath et al., 2011). Increasing amounts of rIxsS41 (25–300 nM for chymase or 100–1600 nM for cathepsin G) were pre-incubated for 0–15 min at 37°C with constant amounts of chymase or cathepsin G to obtain the residual enzymatic activity. The k_obs (pseudo-first order constant) was obtained from the slope of a semi-log plot of the residual chymase or cathepsin G activity against incubation time. The best-fit line of k_obs values was plotted against different amounts of rIxsS41, thus producing the second-order rate constant k _a (Horvath et al., 2011; Kim et al., 2015; Tirloni et al., 2019; Kim et al., 2020a).

Complex formation between rIxsS41 and targeted proteases whose enzymatic activity was inhibited by more than 80% and an SI value below 7 was determined. Chymase and Cathepsin G (0.1 µg) were incubated with rIxsS41 at varying molar ratios (protease-to-rIxsS41 of 10–0.625:1) for 1 h at 37°C. The reaction was stopped by adding denaturing SDS-PAGE reducing sample buffer and heat denaturation at 95°C for 5 min. Denatured samples were resolved on a 10% SDS-PAGE gel and visualized by silver staining using the Pierce Silver Stain Kit (ThermoFisher Scientific) (Tirloni et al., 2019). The rIxsS41 and protease complex was detected at a high molecular weight approximately equal to the sum of the molecular size of rIxsS41 and the target proteases.

The effect of rIxsS41 on deposition of the membrane attack complex (MAC) via each of the three activation pathways of complement—Classic, Alternative, and Mannose Binding Lectin (MBL) systems—was tested using the WiesLab Complement System kit (Svar Life Science AB, Malmö, Sweden). The kit quantifies C5b-9 neoantigen formation with the use of specific alkaline phosphatase labeled antibodies. Affinity-purified rIxsS41 (4 µM) was pre-incubated with human serum, provided by the company (positive control), for 30 min at 37°C. Thereafter, treatments (negative control [NC, containing human serum], positive control [PC, containing human serum that was originally freeze dried], and rIxsS41 + PC) were added, in duplicate, to their respective wells and incubated at 37°C for 60 min, followed by the addition of conjugate and substrate. The A _405nm was obtained using the Synergy H1 plate reader following manufacturer instructions and MAC deposition calculated using the following formula:

Preliminary analysis revealed that rIxsS41 reduced deposition of MAC via the MBL and alternative complement activation pathways. Thus, to gauge insight into possible rIxsS41 targets in the complement system, the PSOPIA prediction server was utilized to evaluate interactions between IxsS41 and proteases involved in the complement pathways (MASP1-3, C1r, C1s, C2, factor B, factor D, and factor I) as described above. Additionally, the interaction between rIxsS41 and factor H was also evaluated as negative control given that factor H is not a protease.

B. burgdorferi complement-resistant B314/BBK32 (pCD100) (Garcia et al., 2016) and complement-sensitive B314/pBBE22luc strains, kindly gifted by Dr. Jon T. Skare (TAMU College of Medicine), were cultured in BSK-II media at 32°C with 1% CO₂ until mid-log phase. Thereafter, BSK-II media was used to dilute cultures to a concentration of 1 × 10⁶ spirochetes/mL. To investigate the effect of rIxsS41 on B. burgdorferi viability throughout serum complement exposure, decreasing amounts (4, 2, and 1 µM) of rIxsS41 were pre-incubated with normal human serum or heat-inactivated normal human serum (Complement Technology, Inc., Tyler, TX, USA) in a microtiter plate for 30 min at 37°C. Thereafter, 85 µL of either B314/BBK32 or B314/pBBE22luc diluted to 1 × 10⁶ spirochetes/mL was added to its respective well and incubated for 1.5 h at 32°C with shaking at 100 rpm. Spirochete survival was scored under dark field microscopy on 10 chosen fields 1.5 h after the start of incubation and every 30 min thereafter. We focused on identifying spirochetes that showed signs of (1) immobilization and (2) membrane damage (bacteriolysis), which were then recorded as not viable as previously described (Kurtenbach et al., 1998; Xie et al., 2019; Booth et al., 2022).

Prompted by data that showed rIxsS41 inhibition of pro-inflammatory proteases, chymase, and cathepsin G, we sought to investigate rIxsS41 anti-inflammatory effect in vivo. The paw edema assay was performed using C48/80 as the inflammation agonist on retired female BALB/c mice as described (Tirloni et al., 2019; Kim et al., 2020a). The experiment was completed twice, each involving 12 mice divided into four treatment groups: saline, C48/80 (1 µg), rIxsS41 (25 µg), and mix (C48/80 [1 µg] and rIxsS41 [25 µg]). The left hind paw of each mouse was carefully pre-marked to the same height and the initial volume was measured using a digital plethysmometer (Harvard Apparatus Inc.). The treatments were respectively administered in 20 µL total volume using a 31-gauge (0.25 mm) 15/64 in (6 mm) 3/10 cc (0.3 mL) U100 needle. Paw swelling (inflammation) was measured by volume displacement using the digital plethysmometer for 8 h: every 30 min for the first 2 h and every hour thereafter. Mice were then humanely euthanized by CO₂ and cervical dislocation.

Three retired female BALB/c mice per treatment as described in the previous section (Paw Edema Assay) were injected with a 27-gauge needle on the left basal footpad. Given that the peak of inflammation during the paw edema assay took place 30–50 min after injection, mice were humanely euthanized as described before, 50 min after injection, and the skin injection site was immediately excised with a scalpel. The collected skin tissues were fixed in 4% paraformaldehyde for 13 h at room temperature. Post-fixation, the sections were thoroughly washed in 70% ethanol for submission to the Texas A&M School of Veterinary Medicine and Biomedical Sciences Core Histology Lab (RRID : SCR_022201) for routine processing and staining with hematoxylin & eosin (H&E) and toluidine blue (TB). The glass slides were scanned at 20× magnification using the Pannoramic Scan II by 3DHistec (Budapest, Hungary).

These scans were analyzed, in a blinded manner, by a board-certified pathologist, Dr. L. Garry Adams (VMBS), and uploaded to the Aiforia image processing platform (Aiforia Inc., Cambridge, MA, USA) to quantify mast cells; cytoplasmic granules were highlighted with the TB stain in order to confirm cell identity and neutrophils via deep learning convolutional neural networks (CNNs) and supervised learning (see Supplementary Material for validation of AI models) models that were developed and validated by Dr. Lindsey A Smith (Aiforia Technologies), Dr. Syeda Areeha Batool (Aiforia Technologies), and Dr. L. Garry Adams. The number of mast cells and PMNs was calculated by dividing the total mast cell count by the area (mm²); both values were provided by the AI.

Mice were anesthetized with Ketamine/Xylazine (87.5/12. 5 mg/kg) cocktail via IP. The dorsal thoracic area was shaved with hair clippers and thoroughly cleansed with 70% ethanol. Three mice per treatment were intradermally injected using a 27G, ½″ needle with 1 × 10⁷ B. burgdorferi spirochetes strain 31 MSK₅ (Labandeira-Rey and Skare, 2001; Labandeira-Rey et al., 2003) in BSK-II mixed with or without rIxsS41 (25 µg), C48/80 (1 µg), or mix (rIxsS41 [25 µg] and C48/80 [1 µg]) (100 µL total volume). The relative number of spirochetes and the amount of IxsS41 secreted by ticks are unknown; however, a 2002 study showed that 2-mm skin biopsies from untreated LD patients with erythema migrans (EM) lesions had an average of 3,381 ± 544 spirochetes per specimen (Liveris et al., 2002). Therefore, for this experiment, we decided to use a high number of spirochetes to compensate for the high amount of IxsS41. We assumed that spirochetes injected at high numbers could survive the pro-inflammatory condition that was induced by C48/80.

Following needle inoculation, the tarsal joint swelling was measured using a caliper and ear skin tissue was sampled at 7 days post-inoculation (closer to the duration of tick feeding), followed by euthanasia at 14 days and a second round of tarsal joint measurement. Immediately after, necropsy was performed to harvest brain, heart, tibiotarsal joint, skin, and spleen. All organs were subjected to gDNA extraction using the DNeasy Blood & Tissue kit (Qiagen, Germantown, MD, USA), which was later used as template (60 ng of each organ) for quantitative (q) PCR analysis using published Bb flaB primers: F: 5′-TCTTTTCTCTGGTGAGGGAGCT-3′ and R: 5′-TCCTTCCTGTTGAACACCCTCT-3′ and murine β-actin primers: F: 5′-CAAGTCATCACTATTGGCAACGA-3′ and R: 5′-CCAAGAAGGAAGGCTGGAA AA (Van Laar et al., 2016) at 0.3 µM concentration in 50 µL of qPCR reactions using iTaq Universal SYBR Green Supermix (Bio-Rad, Hercules, CA, USA). Cultured Bb gDNA (19.6 ng/µL) and mouse gDNA (131.1 ng/µL) were 5-fold diluted and used as standards. The Bio-Rad CFX96 was used with the following conditions: one cycle of 50°C for 2 min and activation at 95°C for 10 min, followed by 40 cycles of denaturation at 95°C for 15 s, and annealing/extension at 60°C for 1 min. Quantification of the flaB gene target was done using the delta delta Ct method (2^−ΔΔCt) as described previously (Livak and Schmittgen, 2001).

All statistical analyses were conducted using Prism 9 (GraphPad Software Inc, San Diego, CA, USA). Unpaired t -test was used to determine the statistical significance of the rIxsS41 effect on complement MAC (2.8). Ordinary one-way ANOVA with Dunnett’s multiple comparisons test was used to determine the significance of rIxsS41 on B. burgdorferi survival in Serum Complement Sensitivity Assay (2.9), in Balb/C mice paw edema (2.10), and in co-inoculating rIxsS41 and B. burgdorferi with and without C48/80 (2.12). To verify the significance among mast cells and neutrophils that was quantified by AI (2.11), the two-way ANOVA with Tukey’s multiple comparisons test was conducted. A p-value ≤ 0.05 was considered statistically significant.

IxsS41 (EEC14235.1 or XP_002402368), which we previously described among 45 tick serpins (Mulenga et al., 2009), was found among tick saliva proteins that were abundantly secreted by I. scapularis nymphs infected with B. burgdorferi (Kim et al., 2021). In GenBank, the EEC14235.1 sequence is truncated at the C-terminal end (22 amino acid residues are missing). Here, we successfully used 3′ RACE to clone the full-length cDNA. Comparative nucleic acid and amino acid sequence analyses reveal that IxsS41 in this study is 98% and 96% identical to EEEC14235.1 and XP_002402368.4 in GenBank, respectively. Notably, the functional domain reactive center loop (RCL) of IxsS41 cloned in this study has one amino acid difference with I. scapularis serpin EEC14235.1 and XP_002402368 at position 12 (L-V) and an additional difference with the latter at position 16 (E-K) (not shown). It is very likely that the amino acid differences between IxsS41 (cloned in this study) and sequences in GenBank (EEC14235.1 and XP_002402368.4) are likely due to sequencing errors as we did not find any amplicons with K and V at positions 16 and 12 in the RCL out of 10 amplicons that were sequenced (not shown).

To gain insight into the sequence relationship of IxsS41 to other tick serpins, we constructed a neighbor-joining phylogeny tree of IxsS41 and 22 other tick serpin sequences that were >50% identical with bootstrap set to 1,000 replications. The phylogeny tree reveals that IxsS41 is not redundant in I. scapularis but is conserved in Ixodes ricinus and I. pesulcatus. Pairwise alignment shows that IxsS41 amino acid sequence is up to 54% identical to other I. scapularis serpins ( Figure 1 ). However, IxsS41 amino acid sequence is 88% and 94% identical to I. persulcatus (KAG0425650.1) and I. ricinus (ABI94056.2), respectively, indicating high conservation in other Ixodes spp. ticks. Notably, pairwise amino acid sequence alignment shows that the RCL of IxsS41 (EEGTEAAAATGVVIVPYSLGP) has two amino acid differences at positions 13 (E-Y) and 9 (A-T) with I. ricinus ABI94056.2 RCL (EEGTVAAATTGVVIVPYSLGP) and one difference at position 13 (E-Y) with KAG0425650.1 RCL (EEGTVAAAATGVVIVPYSLGP) (SF1). Finally, in silico analysis predicted two N-linked potential sites at amino acids 88 (NSTL) and 249 (NLTI) and two O-linked potential sites at amino acids 76 and 79 (not shown).

Pilot expression showed that rIxsS41 was methanol induced within 24 h and increased in abundance over 5 days ( Figure 2A ). Subsequently, large-scale expression (1 L batches) was methanol induced for 5 days ( Figure 2B ), and affinity purified using Nickel affinity columns as published (Kim et al., 2015; Tirloni et al., 2019). As shown ( Figure 2B ; lane 1), the protein band of affinity-purified rIxsS41 was diffuse with molecular weight ranging between ~55 kDa and slightly under 250 kDa, which is higher than the calculated 42 kDa for IxsS41.

Treatment with deglycosylation enzyme mix to remove both N- and O-linked glycans confirmed that rIxsS41 is expressed as a glycosylated protein in P. pastoris as the molecular weight of the deglycosylated (under non-denaturing conditions [ Figure 2B ; lane 2; Supplementary Figure 1B ]) rIxsS41 was reduced to two major protein bands at ~50–55 kDa and three high-molecular-weight protein bands at 100 kDa, 130 kDa, and 250 kDa. However, when treated under denaturing conditions, we observed two additional predominant bands at 37 and 45 kDa ( Supplementary Figure 1C ). Thus, to gauge the purity of rIxsS41, we used in gel LC-MS/MS analysis to identify proteins in each of the bands ( Supplementary Figure 1 ). This analysis showed that rIxsS41 was the most predominant among the protein bands that were analyzed as determined by the normalized spectra abundance factor (NSAF) ( Supplementary Table 1 ).

Since IxsS41 was identified among tick saliva proteins that are highly secreted by Bb-infected nymphs (Kim et al., 2021), we sought to determine its immunogenicity. We found that IxsS41 is among tick saliva proteins that elicit a rabbit immune response to saliva proteins of Bb-infected nymphs ( Figure 3 ). Western blot analysis shows that binding intensities to rIxsS41 of immune sera of rabbits that were fed on by uninfected I. scapularis nymphs ( Figure 3B ) were weaker than immune sera of rabbits that were fed on by B. burgdorferi-infected ticks ( Figure 3C ). These findings were reproduced in our ELISA data ( Figure 3D ). As shown, A _450nm (optical density) levels for uninfected sera ELISA were lower than infected sera ELISA confirming that rabbits that were fed on by Bb-infected ticks had high antibody titer to rIxsS41. We would like to note that sera of rabbits before they were fed on by ticks (or pre-immune sera) was non-specifically binding to rIxsS41. However, non-specific binding was significantly diminished when the antibody was diluted 1,000-fold. It is also notable that at high primary antibody concentration (1:200 dilution), binding to deglycosylated rIxsS41 (blue arrows ↑) is much weaker (faint) when probing with pre-immune ( Figure 3A ) or uninfected rabbit serum ( Figure 3B ), compared to the strong (darker) band observed with Bb-infected rabbit serum ( Figure 3C ). Similarly, when the same sera type is used at a more diluted concentration (1:1,000) to probe deglycosylated rIxsS41 (red arrows ↑), binding to the rIxsS41 protein backbone is only observed when using Bb-infected rabbit serum. Considering that binding to the glycosylated version of rIxsS41 is observed across all three sera types, we suspect that glycans contributed to the nonspecific binding.

It is also important to note to that pre-immune and uninfected rabbit serum also non-specifically bound to some Bb proteins (marked by asterisks [*]): multiple other bands are observed when using Bb-infected rabbit serum.

Substrate hydrolysis assays of 18 mammalian serine proteases confirmed that rIxsS41 (1 µM) inhibited the enzymatic activity of 6 of the 18 tested proteases by nearly 100% ( Figure 4A ). These include chymase (34.2 nM), chymotrypsin (25.7 nM), cathepsin G (154. 4 nM), thrombin (19.8 nM), pancreatic trypsin (0.5 nM), and trypsin IV (12.5 nM). PSOPIA analysis revealed high interaction prediction scores between rIxsS41 and the aforementioned proteases: chymase (0.9590), chymotrypsin (0.9590), cathepsin G (0.9960), thrombin (0.9960), pancreatic trypsin (0.8573), and trypsin IV (0.8910) ( Figure 4B ). Interestingly, PSOPIA analysis revealed moderate to high interaction prediction scores for proteases that were not inhibited by rIxsS41 in our substrate hydrolysis assay ( Figure 4B [table insert]). PSOPIA analysis shows that rIxsS41 likely interact with neutrophil elastase (0.9960), proteinase 3 (0.9874), kallikrein (0.9590), tissue plasminogen activator ([TPA] 0.9222), factor XIa and pancreatic elastase (0.8910), tryptase (0.8710), plasmin and factor Xa (0.8573), and factor XIIa (0.7918). Consistently, low PSOPIA prediction scores were observed for activated protein C ([APC] 0.5333) and papain (0.5555), proteases that were not inhibited by rIxsS41 like factor H, our non-protease control ( Figure 4B ).

Since molar excess of rIxsS41 was used in substrate hydrolysis, there was the possibility that some of our observations were not physiologically relevant. To resolve this, we next used the stoichiometry of inhibition (SI) analysis to estimate the lowest amount of rIxsS41 required to inhibit one molecule of the target protease by 100%. This analysis estimated that approximately 2.8 and 2.2 molecules of rIxsS41 are needed to 100% inhibit one molecule of chymase and cathepsin G, respectively ( Figures 5A, B ). We observed SI of above 7 against chymotrypsin, thrombin, pancreatic Trypsin, and Trypsin IV (not shown).

A typical inhibitory serpin will form a heat- and SDS-stable complex with its target protease (Marijanovic et al., 2019). Consistently, rIxsS41 formed an irreversible heat- and SDS-stable complex with both chymase and cathepsin G ( Figures 5C, D ). We next used the discontinuous method (Horvath et al., 2011) to estimate the rate of rIxsS41 inhibition (k _a) against both chymase and cathepsin G. This analysis estimated that rIxsS41 inhibited chymase and cathepsin G at a k _a of 6.6 × 10³ ±1.7 × 10³ M⁻¹ s⁻¹ and a k _a of 2.4 × 10³ ±1.7 × 10³ M⁻¹ s⁻¹, respectively ( Figures 5E, F ).

Although rIxsS41 inhibited blood clotting factor IIa (thrombin), it had no effect against blood clotting (not shown). In the complement assay, rIxsS41 reduced deposition of the MAC via the Alternative (47% reduction p = 0.0027) and MBL (55%, p = 0.0042) pathways but not the Classical pathway ( Figure 6A ). Consistently, PSOPIA prediction analysis revealed high interaction scores for IxsS41 against proteases in MBL and Alternative complement pathways: MASP1 (0.8573), MASP2 (0.7640), MASP3 (0.8573), C1r (0.8573), C1s (0.8054), C2 (0.8054), factor D (0.9590), and factor I (0.8054) ( Figure 6B ).

Since the MBL pathway of complement is important to B. burgdorferi clearance (Sajanti et al., 2015), we next evaluated if rIxsS41 was protective against killing of serum complement-sensitive B. burgdorferi strain (B314/pBBE22luc [luc]) by complement ( Figure 6C ). Consistent with reduced deposition of the MAC, rIxsS41 dose-dependently protected the complement-sensitive B. burgdorferi strain from killing by complement. Albeit not statistically significant at the 1.5-h time point, the survival rate of the serum sensitive luc strain treated with rIxsS41 was higher than that of the complement-sensitive strain. Likewise, comparable to the resistant strain, the survival rate of the complement-sensitive luc strain incubated with 4 and 2 µM of rIxsS41 was significantly higher than the luc strain incubated with PBS at p = 0.0207, p = 0.0158, or p = 0.0011 at the 2.0-, 2.5-, and 3.0-h time points, respectively. The survival rate of serum-sensitive luc strain treated with 1 µM of rIxsS41 was apparently higher than the sensitive strain but not statistically significant at all time points.

Prompted by findings that rIxsS41 inhibited chymase ( Figures 4 and 5 ), we investigated its effect on chymase-mediated paw edema that was induced by mast cell degranulator, C48/80 (Irman-Florjanc and Erjavec, 1983; Chatterjea et al., 2012). The paw edema assay in BALB/c mice validated that rIxsS41 is a functional anti-inflammatory protein in vivo using the paw edema assay in BALB/c mice ( Figure 7A ). In the absence of rIxsS41, C48/80 caused swelling of treated hind paws within 30–50 min after injection. However, co-injection of C48/80 (1 µg) with rIxsS41 (25 µg) significantly reduced swelling of treated paws by 85% (p = 0.0037). Peak volume displacement reading reduced from 0.07 (0.07± 0.0068) in C48/80 to 0.01 (0.01 ± 0.0033) in rIxsS41 + C48/80 injected paws at p = 0.0037 ( Figure 7A ). The results presented are an average of six mice per treatment.

We next sought to understand if the anti-inflammatory effect of rIxsS41 affected the profiles of pro-inflammatory cells in the inflamed site using Toluidine blue and H&E staining ( Figure 7B ). Quantification of mast cells and neutrophils with deep learning convolutional neural networks (CNNs) and supervised learning (Aubreville et al., 2020; Allier et al., 2022) revealed no significant differences in the number of mast cells or neutrophils per mm² in injected sites with any of the treatments. However, similar to saline (48.13 ± 17.73), mast cells per mm² in rIxsS41-injected sites (49.33 ± 13.38) were apparently lower than in normal skin (untreated control [65.96 ± 12.30]) and C48/80-injected (67.36 ± 12.62) and rIxsS41- and C48/80-injected sites (Mix [70.86 ± 16.75]). In case of neutrophils, untreated skin had apparently fewer cells per mm² (40.14 ± 3.834) than saline (56.33 ± 15.83), C48/80 (84.97 ± 9.407), rIxsS41 (80.61 ± 30.63), and Mix (75.16 ± 10.79) ( Figure 7C ).

Given the anti-inflammatory effects of rIxsS41 ( Figure 7A ), we investigated the effects of rIxsS41 on B. burgdorferi colonization of C3H/HeN mice. Figure 8 summarizes the effect of rIxsS41 and C48/80 on B. burgdorferi colonization of C3H/HeN organs. Compared to B. burgdorferi-only control, C48/80 (1 µg) enhanced B. burgdorferi colonization of ear skin at 1 week post-intradermal inoculation by 0.62 ± 0.22-fold and 1.32 ± 0.40-fold in joints (p = 0.02), and 0.40 ± 0.25-fold in brain at 2 weeks post-inoculation. Likewise, C3H/HeN that intradermally received rIxsS41 (25 µg) and B. burgdorferi had a high spirochete load by 3.50 ± 1.50-fold in liver while mice receiving the mixture rIxsS41(25 µg) and C48/80 (1 µg) had significantly higher spirochete load (2.12 ± 0.68-fold) in heart tissue (p = 0.01). No apparent or significant increase in B. burgdorferi colonization was seen in skin and spleen 2 weeks post-inoculation. It is also notable that the spirochete load was reduced by 0.50 ± 0.29-fold and 0.70 ± 0.26-fold in spleens of mice that received C48/80 and the mixture of rIxsS41 and C48/80, respectively.

Figure 8B summarizes tarsal joint swelling at 1 and 2 weeks after intradermally co-injecting mice with 1 × 10⁷ spirochetes mixed with or without rIxsS41 (25 µg), C48/80 (1 µg), or rIxsS41 and C48/80 mixture. As shown, at 1 week post-inoculation, tarsal swelling of control mice that had received spirochetes only was significantly lower than all other treatment groups. However, at 2 weeks post-infection, differences in tarsal swelling were modestly apparent but not statistically significant.