Homology modeling.

The sequences for SUB and BM95 were modeled using Modeller (program for comparative protein structure modelling by satisfactions of spatial restraints, https://salilab.org/modeller/; accessed November 2022) by selecting the best structures from the alignments based on structure and loop modeling. Models were then submitted to model validation tools (MolProbity, http://molprobity.biochem.duke.edu and PDBsum Generate, https://bio.tools/pdbsum_generate; accessed in December 2022). Ramachandran two-dimensional (2-D) plots and scoring functions were applied to the best models according to Discrete Optimized Protein Energy (DOPE) results (S1 Fig). The structures were minimized using software suite for high-performance molecular dynamics and output analysis (GROMACS v.2021.3, https://www.gromacs.org; accessed in December 2022) using the steepest descent methodology until forces converged below 0.001 kJ mol⁻¹ nm⁻¹.

Docking.

Following model validations, the structures were prepared for protein-protein docking. We used a two-fold approximation of the problem by running both blind and guided docking indicating the attracting residues in BM95 to be D-268, C-269, R-270, V-271, Q-272, K-273, G-274, T-275, V-276, L-277, C-278, E-279, C-280, P-281, W-282, N-283, Q-284, H-285, L-286, V-287, G-288, D-289, T-290, C-291, I-292, S-293, D-294, C-295, V-296, D-297, K-298, K-299, C-300, H-301, E-302, E-303, F-304, M-305, D-306, C-307, and the interacting residues in SUB to be D-138, T-139, F-140, V-141, K-142, F-143, T-144, Y-145, D-146. We then applied two docking servers to the task (High Ambiguity Driven protein-protein DOCKing, HADDOCK v.2.2, https://milou.science.uu.nl/services/HADDOCK2.2/ and protein-protein docking, ClusPro v.2.0, https://cluspro.org/; accessed in December 2022). The results were selected based on the most favorable interacting energies (S2 Fig). The top cluster is the most reliable according to HADDOCK. Its Z-score indicates how many standard deviations from the average this cluster is located in terms of score (the more negative the better).

Protein-protein interface interactions analysis.

To investigate the protein-protein interface interactions, server for analysis of protein interactions in macromolecular assembly (PIMA) was employed (http://caps.ncbs.res.in/pima/; accessed in December 2022). The 2-D representation of the found interactions was generated with ligand-protein interaction diagrams (LigPlot⁺ v.2.2, https://www.ebi.ac.uk/thornton-srv/software/LigPlus/; accessed in December 2022). Briefly, PIMA calculates various energy terms (electrostatic energy, Van der Waals VDW energy, and total stabilization energy) in kJ/mol and number of interactions (hydrogen bonds, VDW pairs, and salt bridges) along with number of interacting residues on the protein-protein interface.

Molecular dynamics.

Molecular dynamics analyses were conducted using the simulation package GROMACS v.2021.3 for the BM95-SUB complex. The ff99SB force field was employed in all the simulations. The TIP3P water model was also used in the simulations. The system was placed in the center of the dodecahedron box, and the minimum distance between the system and the box boundaries was set to 1 nm. Integration of the equations of motion was done at a time step of 2 fs with full periodic boundary conditions (PBC) applied along the three Cartesian directions. The systems were energy-minimized using the conjugate gradient (CG) method, with 1 × 10⁶ (kJ/mol and kJ/mol x nm for energy difference and root mean square (RMS) force, respectively) convergence criteria. Then, a 200 ns constant temperature and volume (NVT) production run (2 ns for wetting calculations) at 300 K was performed using a Berendsen thermostat algorithm with a damping constant of 0.1 ps. During the simulations, a 1.0 nm cut off for Lennard-Jones (LJ) and Coulomb interactions was applied and the particle mesh-Ewald method was used for long-range electrostatics. The LINCS method was used as a constraint algorithm [33]. All the images were generated with Visual Molecular Dynamics (VMD, https://www.ks.uiuc.edu/Research/vmd/) and Xmgrace (https://plasma-gate.weizmann.ac.il/Grace/) (accessed in December 2022).

Structural prediction and molecular docking.

The primary sequence amino-terminus for SUB was truncated (residues 1-113) to parallel the resolved carboxyl-terminus of Akirin2 [34]. The BM95 signal peptide (residues 1 – 19), as predicted by the SignalP server [35], was also truncated from the primary sequence. Both protein models were then individually constructed by the I-TASSER server (https://zhanggroup.org/I-TASSER/) [36]. The predicted I-TASSER tertiary structural pairings [pairings of two predicted tertiary structures for BM95 with five predicted for SUB), S3 Fig] were independently submitted to the HDOCK server (http://hdock.phys.hust.edu.cn/), a protein-protein docking algorithm that merges template-based modeling and ab initio free docking [37]. The docking parameters were kept blind (i.e., without indicating binding residues). The top ten resulting docked conformations for each predicted structural pairing (n = 100) were prepared and optimized using the Maestro software package (MacroModel Schrödinger Release 2022-1, Schrödinger, LLC, New York, NY, 2021) [38]. Firstly, the termini were capped, and the hydrogen atoms were replaced. Secondly, local minimizations were performed to remove any steric clashes. Thirdly, the hydrogen-bond networks were optimized with the Protein Preparation Wizard (https://www.schrodinger.com/science-articles/protein-preparation-wizard) [39]. Lastly, the prepared and optimized docked conformations were screened for residue contacts. All images were captured using the Maestro software package. All applications accessed in December 2022.

Molecular dynamics.

The extended BM95 termini from the two selected docked conformations were truncated (residues 1 – 165 and 496 – 550) and capped to minimize computational expenditure. The two docked complexes were then individually solvated in a 10 ų orthorhombic box with a TIP3P water model neutralized [40] and salted with 0.15 M NaCl. The CHARMM36 force field [41] was used to parameterize the docked complexes and ions (approximately 90000 atoms). All molecular dynamics (MD) simulations were conducted using a GPU-accelerated workstation with the Desmond software [42]. The protein model refinement designed by Zhu et al. [43] running 10  ×  5-ns MD simulations, was implemented using the Desmond default protocol. The Desmond protocol has several initial equilibration steps, and a final MD production step conducted under isotropic conditions with an NPT ensemble coupled with a Nose-Hoover thermostat [44] and a Martyna-Tobias-Klein barostat [45]. The temperature was set at 300 K with a RESPA [46] integrator at an inner time step of 2 fs. After refinement, a 500 ns MD was performed for analysis under the Desmond default protocol.

Amino acids contact analyses.

The two selected docked conformations were processed with PDBePISA [47] to analyze initial consistencies among residue interactions. The docked complexes and 500 ns MD trajectories were analyzed under the VMD software [48]. Residue interactions were inspected using the H-bond contacts, a VMD plugin. The distance cutoff was set at 3.5 Å with a 90 º angle threshold. Detailed information for residue pairs was included. The centroid of interacting residues with aromatic rings were calculated and the distances measured using the π-π stacking Tcl script (https://doi.org/10.5281/zenodo.6408973). The π-π stacking Tcl script also calculated the dihedral angle formed between the planes of the aromatic pairing.

Model generation.

The individual structures of BM95 (UniProt ID: E1B2Y1) and SUB (UniProt ID: A0A0U2T382) were obtained from using the AlphaFold website (https://alphafold.ebi.ac.uk). Four different BM95-SUB complexes were also generated with AlphaFold using the multimer model [49,50]. As an alternative approach, four complexes were obtained using the HADDOCK 2.4 webserver using the single protein structures generated using AlphaFold [51,52]. The initial structures of BM95 and SUB were first simulated using 1 𝜇s MD simulation, followed by the clustering in UCSF Chimera software (UCSF Chimera--a visualization system for exploratory research and analysis) [53]. The four most populous conformers of each protein were used to generate 16 possible structures of the BM95-SUB complexes using the HADDOCK server. Four complexes with the best HADDOCK scores were used for further MD simulations.

Molecular dynamics simulations.

Molecular dynamics simulations were performed for BM95-SUB complexes in explicit water solvent, using a protocol similar to our previous studies [54,55]. Model preparation and simulations were performed using the AMBER v16 suite of programs for biomolecular simulations [56]. AMBER’s ff14SB force-fields were used for all simulations [57]. The MD simulations were performed and refined using NVIDIA graphical processing units (GPUs) and AMBER’s pmemd.cuda simulation engine using our lab protocols published previously [58,59]. We have verified the suitability of ff14SB for simulations at microsecond timescales for a number of proteins complexes [54,55]. Standard parameters from ff14SB force-field were used for the protein residues and nucleotides. SPC model was used for water [60,61]. The parameters for the counter-ions were used from AMBER, as available in the frcmod.ionsjc_spce and frcmod.ionslrcm_hfe_spce files.

Starting with the complex structures obtained from AlphaFold and HADDOCK servers, the missing hydrogen atoms were added with the AMBER’s tleap program. After processing the coordinates of the protein and substrate, all systems were neutralized by addition of counter-ions and the resulting system were solvated in a rectangular box of SPC/E water, with a 10 Å minimum distance between the protein and the edge of the periodic box. The prepared systems were equilibrated using a protocol described previously. The equilibrated systems were then used to run 1.0 μs of production MD under constant energy conditions (NVE ensemble). The use of NVE ensemble is preferred as it offers better computational stability and performance [62,63]. The production simulations were performed at a temperature of 300 K. As NVE ensemble was used for production runs, these values correspond to initial temperature at start of simulations. Temperature adjusting thermostat was not used in simulations; over the course of 1.0 μs simulations the temperature fluctuated around 300 K with RMS fluctuations between 2–4 K, which is typical for well equilibrated systems. A total of 1,000 conformational snapshots (stored every 1,000 ps) collected for each system was used for analysis.

Energy of interaction.

The energy for interactions was calculated as a sum of electrostatic and van der Waals energy between atom pairs. This protocol was previously developed to investigate other protein-substrate systems [64,65].

(1)

E_el is the electrostatic contribution, E_vdw is the van der Waals term and the summation runs over all atom pairs for the protein-substrate complex. The E_el and E_vdw terms were computed as follows,

and

(2)

where q_i are partial charges, and A_ij, B_ij are Lennard-Jones parameters. These parameters were obtained from the AMBER force field. A distance-dependent dielectric function was used:

(3)

B = ε_o-A; ε_o = 78.4 for water; A = -8.5525; λ= 0.003627 and k = 7.7839

All BM95 and SUB atom pairs were included in the calculations and resulting interaction energies were summed up per residue pair. The energies were calculated for 1,000 snapshots, every 1 ns, sampled during the full 1.0 μs simulation and were averaged over these 1,000 snapshots.

Data sonification.

The algorithm used to translate protein sequence into musical scores (Ms.) was previously described and validated [22,66,67]. Each codon either a constant base of a note on which one or two more notes may follow, or a base of two musical notes followed by one or two more. We provided a ternary structure (three beats of a quarter note) and a binary subdivision measure (3/4), where the base of a single note occupies the entire measure (UGC = B, dotted-half-note, and the base of two notes is expressed as a half-plus-quarter-note (CAA = EF). If the base is a single note accompanied by others, we always chose to extend the base in the first two beats of the measure as half-note when it had the succession of a single note (UCG = GD half-plus-quarter-note) or two notes (UCC = GED, half-plus-two-eighth notes). If the base is of two notes, a different metrical scheme was stablished for its expression, dotted-quarter-note-plus-eighth-note followed by a quarter-note (GCG = CDD) or by two eighth-notes (GCC = CDED). When the algorithm provided the same melodic formula for a double base (GGA = SL, GGC = SLDR) and single base (UCA = S; UCU = SL), the metric scheme half-plus-quarter-note in the first case and dotted-quarter-note-plus-dotted-quarter-note in the second case was selected. In this way, each codon has a unique rhythmic and melodic definition.

In silico modeling based on in music model.

In silico modeling based on results of in music model were constructed using Swiss-Model (https://swissmodel.expasy.org), I-TASSER-MTD (https://zhanggroup.org/I-TASSER-MTD/) and Fupred contact map-based domain partition (https://zhanggroup.org/FUpred/), all accessed in November 2022.

Sequence alignment.

Protein sequences for SUB (MT241515.1) and BM95 (AF150891.2) were aligned using the EMBOSS Needle optimal global alignment of two sequences using the Needleman-Wunsch algorithm (https://www.ebi.ac.uk/Tools/psa/).

Production of recombinant proteins and vaccine formulation.

Consensus protein epitope sequences selected based on different models as involved in SUB-BM95 interactions (SUB: 130-STKLAEQYDTFVKFTYDQIQKRFEGATPSYLS-161, BM95: 270-RVQKGTVLCECPWNQHLVGDTCISDCVDKKCHE-302) were chemically synthesized with Keyhole Limpet Hemocyanin (KLH) conjugation on C-terminal and more than 95% purity (Synpeptide Co., Ltd, Shanghai, China). Additionally, a quantum vaccinology approach was applied to combine SUB Q38 protective chimera with candidate BM95 consensus epitopes interacting with SUB and thus likely involved in protective immune response. The resulting chimeric antigen Q38-95 (NH2-MACATLKRTHDWDPLHSPNGRSPKPSPFGEVPPKSSPLESGSPSATPPASPTGLSPGGLLSPVRRDQPLFTFRQVGLICERMMKERESQIRDEYDHVLSAKLAEQYDTFVKFTYDQIQKRFEGATPSYLSgggsHKPFGSPSSPSSSAIAAAAAAAKRPSPFAEAVCPKQLTFNTGSRPDSPPSMVLFTFKQALREQYDAVLTNKLAEQYDAAAPSYLSgggsRVQKGTVLCECPWNQHLVGDTCISDCVDKKCHE-COOH; the position of GGGS linkers is shown in lowercase letters) was produced for vaccine formulation. Recombinant Q38-95 chimera was produced from a synthetic gene (GenScript, Rijswijk, Netherlands) optimized for codon usage and expression in E. coli using pET-21a (+) vector (cloning site NdeI/XhoI) and purified using Ni affinity chromatography using 1 ml HisTrap FF columns mounted on an AKTAgo–FPLC system (Cytiva, Marlborough, MA, USA) in the presence of 7 M urea lysis buffer [18]. The eluted fraction containing the purified proteins was dialyzed against 1000 volumes of PBS (137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, 1.8 mM KH₂PO₄), pH 7.4 for 12 h at 4 ºC. Recombinant proteins were then concentrated using an Amicon Ultra-15 ultrafiltration device (cut off 10 kDa) (Millipore-Merck, Darmstadt, Germany), and adjusted to 0.20 mg/ml. For vaccine formulation, recombinant proteins were adjuvated in Montanide ISA 50 V2 (Seppic, Paris, France).

Modelling Q38-95 and analysis of predicted B-cell, T-cell and conserved protective epitopes.

Modelling of chimeric Q38-95 antigen was conducted using Swiss-Model (https://swissmodel.expasy.org). Predicted T-cell epitopes (MHC I and MHC II) were determined using MHCPred v.2. (http://www.ddg-pharmfac.net/mhcpred/MHCPred/) with IC₅₀ value (nM) confidence of prediction max = 1. For continuous B-cell epitopes, prediction was conducted with Bcepred Prediction Server of continuous B-cell epitopes in antigenic sequences using physico-chemical properties (https://webs.iiitd.edu.in/raghava/bcepred/bcepred_submission.html) with Flexi parameter respective threshold = 2. The analysis of Q38-95 protective epitopes was conducted using UniProt CLUSTAL O(1.2.4) (https://www.uniprot.org, accessed on Feb 4, 2024) tools (a) peptide search (https://www.uniprot.org/peptide-search) for Q38 SAKLAEQYDTFVKFTYDQIQKRFEGATPSYLS and BM86/BM95 RVQKGTVLCECPWNQHLVGDTCISDCVDKKCHE consensus sequences, and (b) align (https://www.uniprot.org/align/) for SUB SAKLAEQYDTFVKFTYDQIQKRFEGATPSYLS in tick species. The Q38-95 self-interaction was modeled with AlphaFold3 and AlphaFold-multimer v2.3 using 20 recycles and 3 seeds (https://www.alphafoldserver.com).

In vitro analysis of protein/peptide interactions.

The peptide/recombinant protein combinations produced were between (a) SUB and BM86 proteins, (b) BM95 (pBM95) and SUB (pSUB) peptides, (c) BM86 protein and pSUB, (d) BM86 protein and pBM95, (e) SUB protein and pSUB, (f) SUB protein and pBM95, (g) Q38-95 protein and pBM95, and (h) Q38-95 protein and pSUB. For antigen combinations, equal molar ratios of each protein or peptide were incubated on a Stuart SB3 tube rotator (Richmond Scientific Ltd., Chorley, UK) at 4 °C overnight.

ELISA using sera from immunized cattle.

Pooled sera from crossbred cows collected 15 days after the third immunization and before tick infestation were obtained from not immunized control cattle (n = 3) and cattle immunized with MSP1a-SUB-BM95 (SUB+BM95; n = 6), BM86 (n = 8) or SUB (n = 5) recombinant proteins and tested with antigen-specific indirect ELISA against the 8 peptide/recombinant protein combinations (a–h) described above and the individual recombinant proteins BM86 and SUB, pSUB and pBM95 and the Q38-95 chimeric antigen. The plates were coated with 0.1 μg per well and incubated overnight at 4 °C with gentle shaking. Plates were then washed once with 100 μl/well PBST and blocked with 100 μl/well of blocking solution (2.5% of skimmed milk in PBS) at room temperature (RT) with gentle agitation for 1 h. Plates were then washed three times with PBST, and bovine serum samples were added at 1:100 dilution in blocking solution and incubated for 1 h at 37 °C. Plates were washed three times with PBST and 1:10,000 dilution of rabbit anti-bovine IgG-HRP conjugate (Sigma) was added to each well. The plates were incubated for 1 h at RT with gentle shaking and after three washes with PBST, 100 μl/well of TMB One Solution (Promega, Madison, WI, USA) were added and incubated for 5–10 min at RT with agitation in dark. Reactions were stopped with 50 μl/well of H₂SO₄ and absorbance was measured in a spectrophotometer at O.D.450nm. Antibody titers were compared between the different peptide/protein interactions for each serum plate using a One-way ANOVA test with post-Tukey Honestly Significant Difference (HSD) (p = 0.05) using Real Statistics Resource Pack (Release 8.6.3) from Microsoft Excel.

Immunization of rabbits with Q38-95.

Two rabbits were subcutaneously immunized with 0.5 ml vaccine formulation containing 50 µg of the recombinant antigen Q38-95 in Montanide ISA 50 V2 (Seppic) per dose [23]. Serum samples were obtained from whole blood by centrifugation at 2,500 x g for 15 min in samples collected before the first immunization (T0), before the second immunization 22 days after the first immunization (T1), and 33 and 49 days after the first immunization (T2 and T3, respectively).

ELISA using sera from Q38-95-immunized rabbits.

Sera from immunized rabbits were tested using antigen-specific indirect ELISA plates coated with the recombinant proteins SUB, BM86 and Q38-95. Plates were coated with 0.1 μg antigen per well and incubated overnight at 4 ºC with gentle shaking. Then, plates were washed with 100 μl/well of PBST, followed by blocking with 100 μl/well of blocking solution (2.5% of skimmed milk in PBS) at RT with gentle agitation for 1 h. Plates were then washed three times with the washing solution, and rabbit serum samples were added at 1:100 dilution in blocking solution and incubated for 1 h at 37 °C. After three washes with the washing solution, a 1:10,000 dilution of anti-rabbit IgG peroxidase from goats (Sigma) was added to each well and incubated for 1 h at RT with gentle shaking. Plates were washed three times with the washing solution, and 100 μl/well of TMB One Solution (Promega) was added and incubated for 5–10 min at RT with agitation in the dark. The reaction was stopped by adding 50 μl/well of H₂SO₄, and the absorbance was measured in a spectrophotometer at O.D._{450 nm}. Antibody titers were analyzed by comparing the mean of the duplicate values between T3 and T0 by Student’s t-test with unequal variance and for T3-T0 by one-way ANOVA with post-hoc Tukey HSD (https://astatsa.com/OneWay_Anova_with_TukeyHSD/) (p = 0.05; n = 2 for each group).

Antibody-mediated inhibition of SUB-BM95 protein-protein interactions.

Experiments were conducted using sera from rabbits and cattle immunized with Q38-95 and controls. For protein-protein interactions using rabbit sera, equal molar ratios of each protein and anti-Q38-95 antibodies were incubated overnight at 4 °C on a Stuart SB3 tube rotator (Richmond Scientific Ltd.) as described above for in vitro analysis of protein/peptide interactions. The quality of anti-Q38-95 antibodies was assessed by comparing its specificity in plates coated with Q38-95 or E. coli protein extract after antigen purification by Ni affinity chromatography to discard unspecific binding. Ten μg of each interaction were loaded onto a 12% SDS-polyacrylamide pre-cast gel (Genscript Corporation, Piscataway, NJ, USA) and transferred to a nitrocellulose membrane. Membranes were blocked with 5% bovine serum albumin (BSA) (Sigma-Aldrich, St. Louis, MI, USA) for 1.5 h at RT and washed three times with tris-buffered saline (TBS)T (50 mM Tris-Cl, pH 7.5, 150 mM NaCl, 0.5% Tween 20). Purified IgG antibodies from rabbits immunized with recombinant SUB or BM86 were used as primary antibodies at 1:500 and 1:200 dilutions in TBS, respectively. Membranes were incubated overnight at 4 °C and then washed four times with TBST. Then, membranes were incubated with an anti-rabbit IgG-horseradish peroxidase (HRP) conjugate (Sigma-Aldrich) diluted 1:1000 in blocking solution (TBS with 3% BSA). Membranes were washed four times with TBST and finally developed with ECL western blotting substrate kit (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s recommendations. For cattle sera, a quantitative analysis was conducted by ELISA as described for rabbits but using sera from immunized cattle in the proof-of-concept study described below. The ELISA plates were coated with no protein (negative control), 0.1 μg SUB (positive control) and 0.1 μg BM86 per well and incubated overnight at 4 °C with gentle shaking. Plates were then incubated with none or an equal molar concentration of pSUB (6 ng) and anti-Q38-95 antibodies (0 to 0.2 μg). Antibody titers against SUB were compared between the different treatments by one-way ANOVA with post-hoc Tukey HSD (https://astatsa.com/OneWay_Anova_with_TukeyHSD/) (p = 0.05; n = 9 for each treatment).

Vaccine efficacy of recombinant Q38-95 antigen.

The Q38-95 E was evaluated in a proof-of-concept preliminary experiment with cross-bred Bos indicus x Bos taurus (local Ankole x Viking Jersey) cattle infested with R. decoloratus. Before the experiment, cattle were assessed for general health status, and selected cattle were those with no previous exposure to ticks and tick-borne diseases, particularly anaplasmosis, theileriosis and babesiosis as previously described [68]. Eight healthy male (n = 4) and female (n = 4) 7 to 10 months calves were selected and randomly distributed with gender balance between vaccinated and control groups. Animals were housed individually in arthropod-free, well-ventilated isolation pens and fed on fodder and 20% protein concentrate, receiving water ad libitum. Cattle were vaccinated with 100 µg Q38-95 in 1 ml vaccine formulation produced as described above. Control animals were treated with PBS in Montanide ISA 50 V2 (Seppic). Treatments were injected in the neck muscles using a 5 ml syringe and needle. The first dose was administered on day zero (T0) and the second dose on day 30 (T1). Pathogen-free colonies of R. decoloratus were prepared from ticks previously collected in Uganda’s agro-ecological zones as previously described [68]. Tick laid eggs were incubated in humidity chambers at 12 h light: 12 h dark photoperiod, 20 ˚C and 95% relative humidity for hatching. Fifteen days after hatching, the emerging larvae were used for challenge at day 45 (T2) with 300 R. decoloratus larvae per animal applied on the back and held in place by a glued bag with a zipper. Adult engorged female ticks dropping off the animal were collected daily at days 73 to 89, counted and weighed. All collected ticks were incubated and assessed for oviposition (egg mass/tick) and egg fertility (weight of larvae/tick). Vaccine efficacy, E (%) = 100 [l - (CRT x CRO x CRF)], where CRT, CRO and CRF are the reduction in the number of adult female ticks, oviposition and egg fertility compared with the control group was calculated as described above (Table 1) and previously reported [31,68].

Analysis of antibody response and expression of selected biomarkers in Q38-95 immunized cattle.

Blood samples were collected from vaccinated and control cattle at T0 (before first treatment), day 30 (T1) (before second treatment), day 45 (T2) before tick infestations and day 100 (T3) at the end of the trial. Antibody response against SUB, BM86 and Q38-95 were characterized by ELISA in cattle sera as described above in rabbits. Total RNA was extracted from blood samples collected at T0, T1 and T2 using TRI Reagent (Sigma-Aldrich, St. Louis, USA), following the manufacturer’s instructions and used for quantitative real-time polymerase chain reaction (qRT-PCR) analysis using the Quantitect SYBR Green RT-PCR Kit and SYBR Green Master Mix (Bio-Rad, Applied Biosystems, Waltham, USA) in a Rotor Gene Q thermocycler (Qiagen, Inc. Valencia, CA, USA) with forward (F, 5´-3´) and reverse (R, 5´-3´) primers and annealing temperature of the selected genes involved in immunity, complement component 3 (C3; NM_001040469.2, F: ATTGCCAGGTTCTTGTACGG and R: GTCACTGCCTGATTGCAAGA, 56°C), tumor necrosis factor alpha (tnf-α; AF348421, F: CCTCACCCACACCATCAG, and R: GCGATCTCCCTTCTCCAG, 54 °C), interleukin 1 beta (il-1β; NM174093, F: TCAGAATGGAAACCCTCTCTC and R: GCATGGATCAGACAACAGTG, 56 °C), interleukin 2 (il-2; NM180997, F: AAGTGAAGTCATTGCTGCTG and R: TGTCCATTGAATCCTTGATCTC, 54 °C), akirin 2 (akr2; NM_001110087.1, F: CATTTATGGGCTGCCTTGTT and R: TGCACAGCTTCTACCACGAC, 54 °C), and interleukin 12 (il-12; U11815.1, F: TCTGGACACTTCACCTGCTG and R: TGCACAGCTTCTACCACGAC, 58 °C) (33, 36). Each PCR reaction had two technical replicates and two negative controls per sample. The mRNA Ct values were normalized against Bos taurus ß-actin (AY141970.1, F: GGCCGAGCGGAAATCG and R: GCCATCTCCTGCTCGAAGTC, 52 °C) using the 2^− ΔΔCt method for relative quantification. Values were compared between control and vaccinated groups at T0, T1 and T2 by one-way ANOVA with post-hoc Tukey HSD (https://astatsa.com/OneWay_Anova_with_TukeyHSD/) (p = 0.05; n = 3 for each group). Antibody response against SUB, BM86 and Q38-95 were characterized by ELISA as described above in rabbits.