Molecular basis for shifted receptor recognition by an encephalitic arbovirus

Abstract

After decades of inactivity throughout the Americas, western equine encephalitis virus (WEEV) recently re-emerged in South America, causing a large-scale outbreak in humans and horses. WEEV binds protocadherin 10 (PCDH10) as a receptor; however, nonpathogenic strains no longer bind human or equine PCDH10 but retain the ability to bind avian receptors. Highly virulent WEEV strains can also bind the very low-density lipoprotein receptor (VLDLR) and apolipoprotein E receptor 2 (ApoER2) as alternative receptors. Here, by determining cryo-electron microscopy structures of WEEV strains isolated from 1941-2005 bound to mammalian receptors, we identify polymorphisms in the WEEV spike protein that explain shifts in receptor dependencies and that can allow nonpathogenic strains to infect primary cortical neurons. We predict the receptor dependencies of additional strains and of a related North American alphavirus. Our findings have implications for outbreak preparedness and enhance understanding of arbovirus neurovirulence through virus receptor binding patterns.

Figure 1.. Structural basis for WEEV recognition of human PCDH10.

(A) Schematic diagrams of PCDH10 and Flag-tagged constructs. (B) Partial phylogenetic tree and summary of infectivity assays with GFP-expressing RVPs for various WEEV strains and Highlands J virus in K562 cells stably expressing the indicated receptors. Strains newly tested in this study are indicated with black triangles. Other results were previously reported in Li et al. (C) Cryo-EM maps of human PCDH10_EC1–Fc bound to WEEV CBA87 VLP. WEEV E2 and E1 and PCDH10 EC1 are shown in purple, green, and pink, respectively. The 5-fold (i5), 3-fold (i3), and 2-fold (i2) icosahedral symmetry axes are indicated with a closed circle, triangle, and hexagon, respectively. The icosahedral threefold (i3) and quasi-three-fold (q3) spikes are circled and indicated. (D) Ribbon diagram of a single WEEV E2–E1 protomer and PCDH10 EC1 repeat fitted into its associated cryo-EM density map. The angle with which EC1 is inserted into the cleft relative to the threefold axis of spike protein trimer is indicated. E2 and E1 domains are indicated. Dashed lines indicate the position of the viral membrane. TM: transmembrane. (E) Surface representation of the PCDH10 EC1-bound WEEV CBA87 spike protein. One of the three PCDH10 EC1 molecules is shown in ribbon representation. Clusters of spike protein residues that interact with EC1 are shown in yellow. Details for the indicated interacting regions are shown in panels F to J. (F–J) PCDH10 EC1 interactions with the WEEV E2–E1 protomer (F and G) and E2’–E1’ protomer (H-J). WEEV E2 L149, a key polymorphic residue, is indicated with an asterisk. Hydrogen bonds and salt bridges are shown as dashed lines. (K) Interaction interface of human PCDH10 EC1 and EC4 in a crystal structure of the PCDH10 EC1–EC4 homodimer (PDB: 6VFW). (L) Infection of K562 cells stably expressing wild-type and mutant human PCDH10 EC1 constructs by GFP-expressing WEEV RVP (strain 71V1658). Infection was quantified by flow cytometry. See Figure S8 for flow cytometry gating strategy and cell surface receptor staining. Data are mean ± s.d. from three experiments performed in duplicates or triplicates (n=3) (L). One-way ANOVA with Dunnett’s multiple comparisons test, ****P<0.0001 compared to stalk-Flag (L).

Figure 2.. WEEV recognition of avian PCDH10.

(A) Summary of the EC₅₀ values measured in ELISAs with human (Hs) PCDH10_EC1–Fc or P. domesticus (Pd) PCDH10_EC1–Fc on immobilized WEEV VLPs for the indicated strains. See Figure S2F for additional information. (B) Cryo-EM map of PdPCDH10_EC1–Fc bound to WEEV Imperial 181 VLP. E2, E1, and PCDH10 are colored purple, green, and pink, respectively. (C) Superposition of the cryo-EM structure of PdPCDH10 EC1 (cyan) bound to the WEEV Imperial 181 spike protein and HsPCDH10 EC1 (pink) bound to the CBA87 spike protein. PCDH10 EC1 is shown as a ribbon diagram and the spike protein subunits are shown in dark (E2) and light (E1) gray. (D–I) Side-by-side comparison of the contact residues of PdPCDH10-bound Imperial 181 (D, F, H) and HsPCDH10-bound CBA87 (E, G, I). Residues that participate in interactions are indicated, with polar contacts shown as gray dashed lines. Dashed lines shown in black indicate the closest distances between PCDH10 residue L74 and atoms on the WEEV spike protein for the protomers shown (H, I). The residue at E2 position 149, a key polymorphic residue in the PCDH10 interface, is highlighted with an asterisk. IMP181: Imperial 181. (J) Partial sequence alignment of WEEV E2 proteins, generated using ESPript3. The key polymorphic residue at position 149 is indicated with a star. Light gray background indicates residues that are completely conserved. Boxed residues indicate positions where a single majority residue or multiple chemically similar residues are present. These residues are shown in dark blue. Glutamine at E2 position 149 (indicated with a star) is colored pink. See also Figures S9A and S10 for additional information. (K) K562 cells stably expressing human MXRA8, VLDLR or PCDH10 were infected with GFP-expressing wild-type or L149Q mutant Fleming RVPs. Infection was quantified by flow cytometry. Data are mean ± s.d. from two experiments performed in duplicates or triplicates (n=3) (K). Two-way ANOVA with Dunnett’s multiple comparisons test, ****P<0.0001 (K).

Figure 3.. VLDLR recognition by ancestral WEEV strains.

(A) Summary of the results of infectivity studies with GFP-expressing RVPs for the indicated alphaviruses with K562 cells expressing VLDLR single LA repeat constructs, with entry quantified using flow cytometry. Infectivity with WEEV McMillan is reported here (see Figures S9A–C) and results for EEEV, SFV, and SINV were previously reported. “Entry” indicates that the construct mediates statistically significant (p < 0.05) RVP infection when compared to a control construct that lacks a ligand-binding domain (LBD). (B) Cryo-EM map VLDLR_LBD–Fc bound to WEEV McMillan VLP. E2, E1, and VLDLR LA1 and LA2 are colored in purple, green, light yellow, and dark yellow, respectively. The 5-fold (i5), 3-fold (i3), and 2-fold (i2) icosahedral symmetry axes are indicated respectively with solid circle, triangle, and hexagon. The icosahedral threefold spike (i3) and quasi-threefold spike (q3) are circled and labeled. (C) Ribbon diagram of a single WEEV E2–E1 protomer and LA1–2 from the VLDLR LBD fitted into the cryo-EM density map. E2 and E1 domains are indicated. Dashed lines indicate the position of the viral membrane. TM: transmembrane. (D) Surface representation of the VLDLR-bound WEEV CBA87 spike protein. One of the three VLDLR LA1–2 segments are shown in ribbon representation. (E and F) Contacts between WEEV McMillan and VLDLR LA1 (E) or LA2 (F). Polar contacts are shown as dashed lines. Ca²⁺ ions are shown as green spheres. Polymorphic residues that contact VLDLR are highlighted with asterisks. (G) K562 cells expressing human MXRA8, VLDLR, or PCDH10 were infected by GFP-expressing wild-type or mutant WEEV McMillan RVPs. (H) K562 cells stably expressing human MXRA8, VLDLR, or ApoER2 were infected with GFP-expressing wild-type or E1 K227A (site 1) + E2 K190A (site 2) mutant WEEV Fleming RVPs. Infection was quantified by flow cytometry. (I) Top view of the WEEV trimeric spike showing three potential binding sites on WEEV E2 and E1 for VLDLR LA repeats. K181 (site 1, cyan) on the E2 glycoprotein and K227 (site 2, yellow) on the E1 glycoprotein are critical contact residues with VLDLR in WEEV McMillan. A lysine at position 81 (site 3, pink) in WEEV Fleming E2 likely binds VLDLR LA repeats. (J) K562 cells stably expressing human MXRA8, VLDLR, or ApoER2 were infected with GFP-expressing wild-type or E2 K81E (site 3) mutant WEEV Fleming RVPs. Infection was quantified by flow cytometry. Data are mean ± s.d. from three experiments performed in triplicates (n=3) (G, H, J). Two-way ANOVA with Dunnett’s multiple comparisons test, ****P<0.0001 (G, H, I).

Figure 4.. WEEV E2 protein polymorphisms that alter receptor recognition determine WEEV neurotropism.

(A) List of mutations generated for WEEV Imperial 181 and summary of K562 infectivity assay as performed in (C). (B) Side view of the WEEV spike highlighting mutated residues on the E2 glycoprotein. (C) K562 cells expressing MXRA8, VLDLR, ApoER2, or PCDH10 were infected with wild-type or mutant Imperial 181 RVPs. Infection was monitored by flow cytometry. (D) K562 cells expressing human PCDH10 (HsPCDH10), sparrow PCDH10 (PdMXRA8), or sparrow MXRA8 (PdMXRA8) were infected with the indicated GFP-expressing RVPs at various MOIs, as pre-determined of Vero E6 cells (see Methods for additional information). Infection was quantified by flow cytometry. (E) Primary murine cortical neurons were infected with GFP-expressing wild-type or mutant WEEV Imperial 181 RVPs in the presence of 316 μg ml⁻¹ PCDH10_EC1–Fc, 100 μg ml⁻¹ RAP, or an isotype control. Infection was monitored through a live cell imaging system. (F) Representative merged images of GFP and bright field are shown for the experiment in (E). Scale bars are 100 μm. Data are mean ± s.d. from three experiments performed in duplicates or triplicates (n=3) (C, D, E). One-way ANOVA with Dunnett’s multiple comparisons test, ****P<0.0001 (C, D). Two-way ANOVA with Šídák’s multiple comparisons test, ****P<0.0001 (E).

Figure 5.. Prediction of alphavirus receptor usage based on spike protein sequences.

(A) K562 cells expressing human MXRA8, VLDLR, ApoER2, or PCDH10 were infected with GFP-expressing RVPs for the indicated South American WEEV strains. (B) K562 cells expressing the indicated orthologs of PCDH10 or MXRA8 were infected with GFP-expressing RVPs for the indicated South American WEEV strains. (C) K562 cells expressing human MXRA8, VLDLR, ApoER2, or PCDH10 were infected with GFP-expressing WEEV BFS09997 (lineage B1) and WEEV EP6 (lineage B2) RVPs. (D) K562 cells expressing human PCDH10 (HsPCDH10), sparrow PCDH10 (PdMXRA8), or sparrow MXRA8 (PdMXRA8) were infected with GFP-expressing WEEV AG80–646 RVPs at the indicated MOIs, as determined on Vero E6 cells (see Methods for additional information). (E) Maximum likelihood phylogenetic tree of select alphaviruses (WEEV, highlands J virus (HJV), Sindbis virus (SINV), VEEV, Madariaga virus (MADV), EEEV, Chikungunya virus (CHIKV), O’nyong’nyong virus (ONNV), Mayaro virus (MAYV), Semliki Forest virus (SFV), Ross River virus (RRV), Getah virus (GETV)) using the coding sequences of the structural polyprotein genes. See Table S1 for strain information. (F) K562 cells expressing human MXRA8, VLDLR, ApoER2, or PCDH10 were infected with GFP-expressing HJV RVPs. (G) K562 cells expressing HsPCDH10, PdPCDH10, HsMXRA8, or PdMXRA8 were infected with GFP-expressing HJV RVPs. (H) K562 cells expressing PCDH10 orthologs were infected with GFP-expressing wild-type (WT) and HJV E2 A177K mutant RVPs. This experiment was performed with an MOI of 1 for WT and mutant RVPs measured on Vero E6 cells. Infection was quantified by flow cytometry. Data are mean ± s.d. from three experiments performed in duplicates or triplicates (n=3) (A, B, C, D, F, G, and H). Two-way ANOVA with Dunnett’s multiple comparisons test (A, B, C). Two-way ANOVA with Šídák’s multiple comparisons test (H). One-way ANOVA with Dunnett’s multiple comparisons test (F, G)