Mechanistic insights into the structure-based design of a CspZ-targeting Lyme disease vaccine

Abstract

Borrelia burgdorferi (Bb) causes Lyme disease (LD), one of the most common vector-borne diseases in the Northern Hemisphere. Here, we solve the crystal structure of a mutated Bb vaccine antigen, CspZ-YA that lacks the ability to bind to host complement factor H (FH). We generate point mutants of CspZ-YA and identify CspZ-YA_I183Y and CspZ-YA_C187S to trigger more robust bactericidal responses. Compared to CspZ-YA, these CspZ-YA mutants require a lower immunization frequency to protect mice from LD-associated inflammation and bacterial colonization. Antigenicity of wild-type and mutant CspZ-YA proteins are similar, as measured using sera from infected people or immunized female mice. Structural comparison of CspZ-YA with CspZ-YA_I183Y and CspZ-YA_C187S shows enhanced interactions of two helices adjacent to the FH-binding sites in the mutants, consistent with their elevated thermostability. In line with these findings, protective CspZ-YA monoclonal antibodies show increased binding to CspZ-YA at a physiological temperature (37 °C). In summary, this proof-of-concept study applies structural vaccinology to enhance intramolecular interactions for the long-term stability of a Bb antigen while maintaining its protective epitopes, thus promoting LD vaccine development.

Fig. 1. The X-ray structure of CspZ-YA and the mutagenesis of amino acid residues in CspZ-YA by structure-based vaccine design.

A The crystal structure of CspZ-YA (gray; PDB ID 9F1V) is superimposed with the structure of B. burgdorferi B31 CspZ (blue) from the CspZ/SCR6-7 (gold) complex (PDB ID 9F7I; RMSD 0.9 Å) and B. burgdorferi B31 CspZ (green) from the CspZ/SCR7 (red) complex (PDB ID 6ATG; RMSD 0.74 Å). A portion of the electron density map from the newly solved structure of CspZ-YA is shown in Fig. S8A. All nine α-helices of CspZ are labeled (αA-αI). The inlet figures show the loop region between helices H and I, highlighting residues Y207 and Y211 in CspZ from B. burgdorferi strain B31 and the mutated residues A207 and A211 in CspZ-YA. Residues K212 and K213 found in the loop region in CspZ and in the extended helix I in CspZ-YA are shown. The interaction between residues R206 and E186 in CspZ is further indicated. The structure is presented from top and side views. B Design landscape of CspZ-YA (PDB ID 9F1V) is shown as a ribbon diagram with the side chains of the mutated amino acid residues shown as spheres. Insets highlight the position and side chains of selected stabilizing mutations. Side chains in each inset are shown as dark red sticks with sulfur atoms in yellow, nitrogen atoms in blue, and oxygen atoms in red.

Fig. 2. Mice immunized twice and three times with CspZ-YA_C187S or CspZ-YA_I183Y had sera with more robust levels of borreliacidal activity than CspZ-YA-vaccinated mice.

A The schematic diagram shows pre-adolescent C3H/HeN receiving one inoculation with PBS (control) or the TitierMax Gold (TMG) at the indicated timeframe and frequency, followed by the infection. B–G Sera were collected at 14 dpli from pre-adolescent C3H/HeN mice immunized B, C once, D, E twice, or F, G three times. These mice were immunized with PBS (control) or untagged CspZ-YA or histidine-tagged CspZ-YA (His-CspZ-YA), or their mutant proteins. These sera were diluted as indicated, and mixed with guinea pig complement and B. burgdorferi B31-A3 for 24 h. Surviving spirochetes were quantified from three fields of view microscopically in three independent experiments. A, C, E The survival percentage was derived from the proportion of serum-treated to untreated spirochetes. The data shown are the mean ± SEM of the survival percentage from three replicates in one representative experiment. B, D, F The BA₅₀ value, representing the dilution rate that effectively killed 50% of spirochetes, was obtained from curve-fitting and extrapolation of (A, C, E). Data shown are the geometric mean ± geometric standard deviation of BA₅₀ value from n = 6 mice immunized once or twice or n = 5 mice immunized three times with CspZ-YA, CspZ-YA_C187S, or His-CspZ-YA, or n = 5 mice inoculated with each of other proteins or PBS and also shown in Table S1. (“NK”), no killing. Statistical significance (p < 0.05, Kruskal–Wallis test with the two-stage step-up method of Benjamini, Krieger, and Yekutieli) of differences in borreliacidal titers between groups are indicated (“#”). E CspZ-YA: CspZ-YA_C187S p = 0.0005, His-CspZ-YA: His-CspZ-YA_I183Y p = 0.0032. G CspZ-YA: CspZ-YA_C187S p = 0.0068, His-CspZ-YA: His-CspZ-YA_I183Y p = 0.0109. Source data are provided as a Source Data file.

Fig. 3. Immunizing twice with CspZ-YA_C187S or CspZ-YA_I183Y but not CspZ-YA protected mice from seroconversion and borrelial tissue colonization.

Pre-adolescent C3H/HeN mice inoculated with PBS or indicated OspA or CspZ-YA proteins or mutant protein (A–F) once, (G–L) twice, or (M–R) three times, followed by infection using nymphs carrying B. burgdorferi B31-A3 in the fashion described in Fig. 2A. Mice inoculated with PBS that are not fed on by nymphs were included as an uninfected control group (uninfect.). B, H, N Seropositivity was determined by measuring the levels of IgG against C6 peptides in the sera of those mice at 42 days post-last immunization using ELISA. The mouse was considered seropositive if that mouse had IgG levels against C6 peptides greater than the threshold, the mean plus 1.5-fold standard deviation of the IgG levels against C6 peptides from the PBS-inoculated, uninfected mice (dotted line). The number of mice in each group with anti-C6 IgG levels greater than the threshold (seropositive) is shown. Data shown are the geometric mean ± geometric standard deviation of the titers of anti-C6 IgG from n = 6 mice immunized once or twice or n = 5 mice immunized with CspZ-YA or CspZ-YA_C187S, or n = 5 mice immunized with each of other proteins or n = 5 uninfected mice. Statistical significances (p < 0.05, Kruskal–Wallis test with the two-stage step-up method of Benjamini, Krieger, and Yekutieli) of differences in IgG titers relative to (*) uninfected mice are presented. A, C–F, B, H–L, M, and O–R B. burgdorferi (Bb) burdens at A, G, M nymphs after when feeding to repletion or C, I, O the tick feeding site (“Bite Site”), D, J, P bladder, E, K, Q heart, and F, L, R knees, were quantitatively measured at 42 days post-last immunization, shown as the number of Bb per 100 ng total DNA. Data shown are the geometric mean ± geometric standard deviation of the spirochete burdens from n = 6 nymphs feeding on OspA-immunized mice, n = 7 nymphs feeding on mice immunized with other proteins, n = 6 mice immunized once or twice, or n = 5 mice immunized three times with CspZ-YA or CspZ-YA_C187S, or n = 5 mice immunized with each of other proteins, or n = 5 uninfected mice. Asterisks indicate the statistical significance (p < 0.05, Kruskal–Wallis test with the two-stage step-up method of Benjamini, Krieger, and Yekutieli) of differences in bacterial burdens relative to uninfected mice. B Uninfect.: PBS, OspA, CspZ-YA, CspZ-YA_C187S, p = 0.0028, 0.0016, 0.0329, 0.0047. C Uninfect.: PBS, OspA, CspZ-YA, CspZ-YA_C187S, p = 0.0018, 0.0207, 0.026, 0.0001. D Uninfect.: PBS, OspA, CspZ-YA, CspZ-YA_C187S, p = 0.0035, 0.0132, 0.0123, 0.0017. (E) Uninfect.: PBS, OspA, CspZ-YA, CspZ-YA_C187S, p = 0.0184, 0.0031, 0.0022, 0.0082. F Uninfect.: PBS, OspA, CspZ-YA, CspZ-YA_C187S, p = 0.0457, 0.0074, 0.0035, 0.0008. H Uninfect.: PBS, OspA, CspZ-YA, His-CspZ-YA, p = 0.002, 0.0063, 0.002, 0.003. I Uninfect.: PBS, OspA, CspZ-YA, His-CspZ-YA, p = 0.009, 0.0211, 0.0081, 0.0035. J Uninfect.: PBS, OspA, CspZ-YA, His-CspZ-YA, p = 0.0003, 0.0088, 0.0339, 0.0004. K Uninfect.: PBS, OspA, CspZ-YA, His-CspZ-YA, p = 0.0002, 0.0491, 0.0475, 0.0035. L Uninfect.: PBS, OspA, CspZ-YA, His-CspZ-YA, p = 0.0009, 0.0443, 0.0473, 0.0002. N Uninfect.: PBS p = 0.0071. O Uninfect.: PBS p = 0.0001. P Uninfect.: PBS p = 0.0002. Q Uninfect.: PBS p = 0.0001. R Uninfect.:PBS p = 0.0001. Source data are provided as a Source Data file.

Fig. 4. Immunizing twice with CspZ-YA_C187S or CspZ-YA_I183Y but not CspZ-YA protected mice from Lyme disease-associated joint inflammation.

Pre-adolescent C3H/HeN mice inoculated with PBS or indicated OspA or CspZ-YA proteins or mutant protein were inoculated twice, followed by infection using nymphs carrying B. burgdorferi B31-A3 in the fashion described in Fig. 2A. Tibiotarsus joints at 42 days post-last immunization were collected to assess inflammation by staining these tissues using hematoxylin and eosin. Representative images from one mouse per group are shown. The top panels are lower-resolution images (joint, ×10 [bar, 160 µm]); the bottom panels are higher-resolution images (joint, 2 × 20 [bar, 80 µm]) of selected areas (highlighted in top panels). Arrows indicate infiltration of immune cells. (Inset figure) To quantify inflammation of joint tissues, at least ten random sections of tibiotarsus joints from each mouse were scored on a scale of 0–3 for the levels of inflammation. Data shown are the mean inflammation score ± standard deviation of the inflammatory scores from n = 5 mice immunized with OspA, His-CspZ-YA, or His-CspZ-YA_I183Y, n = 6 mice inoculated with PBS, CspZ-YA, or CspZ_C187S, or n = 7 uninfected mice. Asterisks indicate the statistical significance (p < 0.05, Kruskal–Wallis test with the two-stage step-up method of Benjamini, Krieger, and Yekutieli) of differences in inflammation relative to uninfected mice. All images were not cropped. (Inset figure) Uninfect.: PBS, OspA, CspZ-YA, His-CspZ-YA p = 0.0039, 0.015, 0.0004, 0.0003. Source data are provided as a Source Data file.

Fig. 5. Human and mouse CspZ antibodies recognized CspZ-YA_C187S or CspZ-YA_I183Y at indistinguishable levels from CspZ-YA.

A–D Patient serum samples (“Two tier positive”; Positive in Lyme disease two tier test) were to determine their levels of recognition to histidine-tagged CspZ-YA (His-CspZ-YA) or CspZ-YA_I183Y (His-CspZ-YA_I183Y), or untagged CspZ-YA_C187S or using ELISA. The serum samples from humans residing in non-endemic area of Lyme disease (Neg. ctrl.) were included as control. A Data shown are the geometric mean ± geometric SD of levels of recognition from n = 36 patient serum samples or n = 10 negative control serum samples. Statistical significance (p < 0.05, Kruskal–Wallis test with the two-stage step-up method of Benjamini, Krieger, and Yekutieli) of differences in levels of recognition by groups are indicated (“#”). E–H Sera from His-CspZ-YA or -CspZ-YA_I183Y (I183Y)-, or untagged CspZ-YA- or CspZ-YAC187S (C187S)-immunized C3H/HeN mice (twice immunization, Fig. 2A) were collected at 14dpli. PBS-inoculated mice were included as control. E The levels of serum recognition by indicated CspZ-YA proteins were measured by ELISA. Data shown are the geometric mean ± geometric SD of levels of recognition from serum samples of n = 6 mice immunized with CspZ-YA or CspZ-YA_C187S or n = 5 of mice inoculated with other proteins or PBS. Statistical significance (p < 0.05, Kruskal–Wallis test with the two-stage step-up method of Benjamini, Krieger, and Yekutieli) of differences in levels of recognition by groups are indicated (“#”). For B–D human or F–H mouse sera, the values representing the levels of recognition by B, F CspZ-YA vs. CspZ-YAC187S, C, G CspZ-YA vs. CspZ-YAI183Y, or D, H CspZ-YAC187S vs. CspZ-YAI183Y were plotted. The correlation of these values was determined using Spearman analysis and shown as R values and P values (p < 0.05, statistical significance). I Superimposed crystal structures of CspZ-YA (gray; PDB ID 9F1V), CspZ-YA_C187S (brown; PDB ID 9F21), and the predicted structure of CspZ-YA_I183Y (green). All nine α-helices are labeled (αA-αI). J Shown is the region in CspZ-YA, CspZ-YA_C187S, and CspZ-YA_I183Y where mutations were introduced. Residues associated with mutations are illustrated as thick bonds. A Neg. ctrl.:Two tier pos. for His-CspZ-YA p < 0.0001. Neg. ctrl.:Two tier pos. for CspZ-YA_C187S p < 0.0001. Neg. ctrl.:Two tier pos. for His-CspZ-YA_I183Y p < 0.0001. E PBS: CspZ-YA, CspZ-YA_C187S, His-CspZ-YA, His-CspZ-YA_I183Y for His-CspZ-YA p = 0.0359, 0.0030, 0.0028, 0.0089, PBS: CspZ-YA, CspZ-YA_C187S, His-CspZ-YA, His-CspZ-YA_I183Y for CspZ-YA_C187S p = 0.0182, 0.0002, 0.0031, 0.0079, PBS: CspZ-YA, CspZ-YA_C187S, His-CspZ-YA, His-CspZ-YA_I183Y for His-CspZ-YA_I183Y p = 0.0204, 0.0018, 0.0021, 0.0178. Source data are provided as a Source Data file.

Fig. 6. The comparison of CspZ, CspZ-YA, CspZ-YA_C187S, and CspZ-YA_I183Y structures suggests the helix H–I interactions impacted by the C187S and I183Y mutagenesis.

The structures here were obtained from CspZ (PDB ID 9F7I), CspZ-YA (PDB ID 9F1V), CspZ-YA_C187S (PDB ID 9F21), and AlphaFold predicted structure of CspZ-YA_I183Y. A portion of the electron density map from the newly solved structure of CspZ-YA_C187S is shown in Fig. S8B. A, B Shown is the 2Fo-Fc electron density map contoured at 1σ of the region around C187 in (A) CspZ-YA and S187 in (B) CspZ-YA_C187S. The hydrogen bond formed between the water molecule with S187 and E214 were highlighted. (C–E) The crystal structures of C CspZ from B. burgdorferi B31, D superimposed CspZ-YA (gray) and CspZ-YA_C187S (brown) structures, and E the predicted structure of CspZ-YA_I183Y show the hydrophobic core accounting for helices G, H and I and the residues Y207, Y211, I183, and C187 in CspZ and the equivalent residues in CspZ-YA and CspZ-YA_I183Y.

Fig. 7. MD simulations of CspZ, CspZ-YA, CspZ-YA_C187S, and CspZ-YA_I183Y show flexibility of the loop between helices H and I impacted by C187S and I183Y mutagenesis.

A The structural drift of CspZ, CspZ-YA, CspZ-YA_C187S, and CspZ-YA_I183Y is shown over 300 ns at 300 K. B The RMSF measurements of Cα atoms for CspZ, CspZ-YA, CspZ-YA_C187S, and CspZ-YA_I183Y over 300 ns at 300 K are plotted against the residue number. Different colors (black, green, red) represent three simulation runs. C The crystal structures of wild-type CspZ (PDB ID 9F7I), CspZ-YA (PDB ID 9F1V), CspZ-YA_C187S (PDB ID 9F21), and the predicted structure of CspZ-YA_I183Y illustrate the five regions (R1-R5) with increased RMSF values. All nine α-helices of CspZ variants are labeled (αA-αI).

Fig. 8. CspZ-YA_C187S and CspZ-YA_I183Y maintained the recognition by protective CspZ IgGs at higher temperatures for a longer time.

A Untagged or histidine-tagged CspZ-YA or mutant proteins (10 µM) in PBS buffer were subjected to the thermal shift assays with the emission wavelength of 623 nm, as described in the materials and methods section. Shown are the fluorescence intensities of each of the CspZ-YA proteins under temperatures ranging from 25 to 99 °C from one representative experiment. The melting temperature (Tm) was extrapolated from the maximal positive derivative values of the fluorescence intensity (d(RFU)/dT) and shown in Table S3. B, C One µg indicated CspZ-YA proteins was incubated at 4 or 37 °C for 6- or 24-h prior to being coated on microtiter plate wells. The microtiter plate wells immobilized with each of these proteins before incubation (0-h) were included as unincubated control. The ability of the CspZ monoclonal IgG, B 1139c or C 1193c, to recognize each of these CspZ-YA proteins were determined using ELISA in the section “Accelerated stability study” in Materials and Methods. The work was performed on four independent experiments (one replicate per experiment). Data are expressed as the percent binding, derived by normalizing the levels of bound 1139c or 1193c from the wells coated with each of the CspZ-YA proteins in different incubating conditions to that in the unincubated control. Data shown are the mean ± standard deviation of the percent binding of 1139c or 1193c from n = 4 experiments. Statistical significance (p < 0.05, Kruskal–Wallis test with the two-stage step-up method of Benjamini, Krieger, and Yekutieli) of differences in percent binding between groups are indicated (“#”). B CspZ-YA 0 h: CspZ-YA 37 °C, 24 h p = 0.0019. His-CspZ-YA 0 h: His-CspZ-YA 37 °C, 24 h p = 0.0484. C CspZ-YA 0 h: CspZ-YA 37 °C, 24 h p = 0.0003. His-CspZ-YA 0 h: His-CspZ-YA 37 °C, 24 h p = 0.0028. Source data are provided as a Source Data file.