Two point mutations in protocadherin-1 disrupt hantavirus recognition and afford protection against lethal infection

Abstract

Andes virus (ANDV) and Sin Nombre virus (SNV) are the etiologic agents of severe hantavirus cardiopulmonary syndrome (HCPS) in the Americas for which no FDA-approved countermeasures are available. Protocadherin-1 (PCDH1), a cadherin-superfamily protein recently identified as a critical host factor for ANDV and SNV, represents a new antiviral target; however, its precise role remains to be elucidated. Here, we use computational and experimental approaches to delineate the binding surface of the hantavirus glycoprotein complex on PCDH1's first extracellular cadherin repeat domain. Strikingly, a single amino acid residue in this PCDH1 surface influences the host species-specificity of SNV glycoprotein-PCDH1 interaction and cell entry. Mutation of this and a neighboring residue substantially protects Syrian hamsters from pulmonary disease and death caused by ANDV. We conclude that PCDH1 is a bona fide entry receptor for ANDV and SNV whose direct interaction with hantavirus glycoproteins could be targeted to develop new interventions against HCPS.

Fig. 1. Residue F83 in PCDH1 is a key determinant of Sin Nombre virus infection.

a Viral titer of rVSVs expressing G, ANDV Gn/Gc, or SNV Gn/Gc in human or mouse primary lung endothelial cells. Means ± SD: n = 6 wells of infected cells examined over three independent experiments. b Surface expression of endogenous PCDH1 in primary mouse lung microvascular endothelial cells (MLMECs) used in (a). Cells were immunostained with PCDH1-specific monoclonal antibody (mAb) 3305 or a negative control mAb (Ctrl.) Scale bar, 20 µm. c Infectivity of rVSVs bearing HTNV, ANDV, or SNV Gn/Gc in primary MLMECs expressing flag-tagged, wild-type (WT) or mouse-variant (F83L) human PCDH1. rVSV infectivities are expressed as fold change relative to that in non-complemented cells (set to one). Means ± SD: n = 35 wells of infected cells examined over three independent experiments. d Cells described in (c) were immunostained with an anti-flag antibody to detect total human PCDH1 expression in transfected MLMECs. Scale bar, 20 µm. e Alignment of human PCDH1-EC1 amino acid sequences with a selection of rodent and primate species. All of the residues within EC1 which deviate from the consensus are shown and highlighted. Residues that specifically deviate from human EC1 are indicated [green, experimentally tested SNV-susceptible hosts; purple, experimentally unknown; orange, residues in EC1 that are different between human and other species with no known link to susceptibility]. Alignments generated by Clustal Omega. f The capacity of U2OS PCDH1-KO cells expressing the indicated PCDH1 variants to support hantavirus Gn/Gc-dependent entry. Cells were exposed to rVSVs bearing the indicated Gn/Gc proteins or authentic ANDV, SNV, or HTNV. “None” indicates no PCDH1 expression. The infectivity of each virus was normalized to that obtained in U2OS PCDH1-KO cells complemented with WT PCDH1. rVSV means ± SD: n = 8 wells of infected cells examined over three independent experiments. ANDV/SNV/HTNV means ± SD: one experiment examining n = 3 (ANDV and SNV) or n = 10 (HTNV) wells of infected cells. Infectious units (a) were compared by unpaired, two-tailed t test with Welch’s correction. Infectivities (c, f) were compared by one-way ANOVA with Dunnett’s correction for multiple comparisons; ns > 0.05, **P < 0.01. Source data are provided as a Source Data file.

Fig. 2. Residue F83 in PCDH1 is critical for Sin Nombre virus Gn/Gc: sEC1-2 binding.

a Diagram of direct binding ELISA comparing sEC1-2(WT) and sEC1-2(F83L) capture of rVSVs. b Direct binding ELISA. rVSVs expressing ANDV, SNV, or HTNV Gn/Gc were added to sEC1-2(WT) or sEC1-2(F83L) coated ELISA plates. Means ± SD: n = 4 wells examined over two independent experiments. A, absorbance. c Diagram depicting competition ELISA comparing sEC1-2(WT) and sEC1-2(F83L) as competitive reagents. d Competition ELISA. rVSVs expressing ANDV or SNV Gn/Gc were pre-incubated with sEC1-2(WT) or sEC1-2(F83L) before added to sEC1-2(WT) coated ELISA plates. The ELISA signal was normalized to that obtained without competing sEC1-2. Means ± SEM: n = 7 wells examined over three independent experiments (ANDV), n = 4 wells examined over two independent experiments (SNV). e Diagram depicting infection-inhibition assay comparing sEC1-2(WT) and sEC1-2(F83L) as inhibiting reagents. f Infection-inhibition assay using sEC1-2(WT) and sEC1-2(F83L) to block infection (MOI of 0.1) of rVSVs bearing ANDV or SNV Gn/Gc on primary human endothelial cells (HUVECs). The infectivity of each virus was normalized to that obtained without sEC1-2. Averages ± SD: n = 6 wells examined over two independent experiments. (sEC1-2, soluble extracellular cadherin domains 1 and 2). Figures (a, c, e) were created with BioRender.com. Source data are provided as a Source Data file.

Fig. 3. Structure-based interfacial prediction reveals a surface patch on PCDH1 EC1 that potentially drives the interaction with ANDV and SNV Gn/Gc.

a Schematic representation of PCDH1 and crystal structure of EC1 (PDB 6MGA) displaying two modeled conformations (green: “open conformation”, red: “closed conformation”) for the disordered, uncrystallized loop comprising of residues 80–89. Residue F83 is indicated in each predicted loop conformation. b EC1 crystal structure in the “open conformation” displaying the EC1 residues chosen for mutational screening, ranked and colored according to the number of supporting algorithms. Structure adapted from PDB 6MGA. c List of the EC1 residues chosen for mutational screening in (b) ranked according to the number of supporting algorithms (interface prediction column). The amino acid substitution(s) are listed for each residue. (EC1, extracellular cadherin domain). Rankings of the residues in the PCDH1 ectodomain for each of the five complimentary algorithms can be found in Supplementary Data 2.

Fig. 4. Binding capacity of mutant sEC1-2 to ANDV Gn/Gc.

a Diagram of competition ELISA depicting three different competition outcomes of mutant sEC1-2 proteins’ capacity to block rVSV-ANDV-Gn/Gc binding to sEC1-2(WT) coated wells. b Competition ELISA using WT and mutant sEC1-2 as competitive reagents to the binding of rVSV-ANDV-Gn/Gc to sEC1-2(WT) coated wells. The ELISA signal was normalized to that obtained without competing sEC1-2. Averages ± SEM: n = 6 wells of each dilution examined over three independent experiments, except for: sEC1-2(D102A, K104A, K104E, T105A, V144A, Q145A, D102R, D142A, G75R, Y81A, P84A, I140A) have n = 8 wells examined over four independent experiments [sEC1-2(Y81) has n = 7 wells for one dilution]. sEC1-2(WT) was used as a reference control for each experiment performed and has n = 32 wells of each dilution examined over 16 independent experiments. c Hierarchical clustering of WT and mutant sEC1-2 generated from sigmoidal curves of the competition ELISA data in (a). The dendrogram shows four clusters, separating the poor binders from the WT-like binders. The dotted line denotes the height at which the dendrogram is cut representing varying degrees of binding strength; WT-like binders (green), intermediate binders (WT-like-to-intermediate binders in dark orange, intermediate-to-poor binders in light orange), and poor binders (purple). The red to blue colorbar ranges from 0 to 100 which is determined by the minimum and maximum values observed in the heatmap. (sEC1-2, soluble extracellular cadherin domains 1 and 2). Figure (a) was created with BioRender.com. Source data are provided as a Source Data file.

Fig. 5. Inhibition of ANDV Gn/Gc-mediated infection by mutant sEC1-2.

a WT and mutant sEC1-2 were tested on their ability to block rVSV-ANDV-Gn/Gc entry in primary human endothelial cells (HUVECs). The infectivity was normalized to that obtained without sEC1-2. Averages ± SEM: n = 6 wells of infected cells for each sEC1-2 dilution examined over three independent experiments [sEC1-2(T141A, L143A, V144A, Q145A, D85A, D142A, D142R, Y81A, F83A, D85A, I140A) have n = 7, sEC1-2(Y81A) has n = 6 for one dilution]. sEC1-2(WT) was used as a reference control and has n = 27 wells examined over 13 independent experiments. b Hierarchical clustering of WT and mutant sEC1-2 generated from sigmoidal curves of the infection-inhibition assay in (a). The dotted line denotes the height at which the dendrogram is cut to obtain three clusters representing varying degrees of inhibition of ANDV Gn/Gc-initiated infection; WT-like inhibition (green), intermediate inhibition (orange), and poor inhibition (purple). The red to blue colorbar ranges from 0 to 100 which is determined by the minimum and maximum values observed in the heatmap. c Area under the curve (AUC) for the binding activity of WT and mutant sEC1-2 to rVSV-ANDV-Gn/Gc, as determined by competition ELISA (see Fig. 4b), plotted against AUC values as determined by rVSV-ANDV-Gn/Gc infection-inhibition assay (a). The red line denotes the R squared value. A list of the sEC1-2 mutants that are classified as poor binders are listed to the right. d EC1 crystal structure in the “open conformation” displaying mutated residues representing three degrees of binding strength to ANDV Gn/Gc and inhibition of rVSV-ANDV-Gn/Gc infection. sEC1-2 mutants that bind and inhibit similarly to WT (WT-like binders), green; sEC1-2 mutants that display a mild reduction in binding and inhibition (intermediate binders), orange, and sEC1-2 mutants that display a strong reduction in binding and inhibition (poor binders), purple. Structure adapted from PDB 6MGA. Source data are provided as a Source Data file.

Fig. 6. Two key amino acids in PCDH1 mediate entry for SNV and ANDV.

a Relative infectivity of rVSVs bearing ANDV, SNV, or HTNV Gn/Gc on U2OS PCDH1-KO cells complemented with WT or mutant PCDH1. The infectivity of each virus was normalized to that obtained in U2OS PCDH1-KO cells complemented with WT PCDH1. Means ± SD: n = 9 wells of infected cells examined over three independent experiments (rVSV-ANDV-Gn/Gc infection on U2OS PCDH1-KO cells complemented with V86A had n = 8). Infectivities were compared by one-way ANOVA with Dunnett’s test for multiple comparisons. b Relative infectivity of authentic ANDV, SNV, or HTNV on the cell lines described in (a). The infectivity of each virus was normalized to that obtained in U2OS PCDH1-KO cells complemented with WT PCDH1. Means ± SD: n = 9 infected wells were examined over three independent experiments (HTNV infection on U2OS PCDH1-KO cells complemented with D85A and F83A/D85A had n = 6 wells of infected cells examined over two independent experiments). For ANDV and SNV infection on the control cell line, U2OS PCDH1-KO cells complemented with WT, n = 12 (SNV) and n = 18 (ANDV) wells of infected cells were examined over four independent experiments. Infectivities were compared by one-way ANOVA with Dunnett’s test for multiple comparisons. Source data are provided as a Source Data file.

Fig. 7. Monomeric and dimeric PCDH1 provide suitable entry receptors for ANDV.

a Crystal structure of the proposed anti-parallel EC1-4 trans-dimer. Structure is in the “open conformation” displaying residues representing three degrees of binding strength to ANDV Gn/Gc. Residues that when mutated bind similarly to WT (green); those that display a mild reduction in binding (orange); and those that display a strong reduction in binding (purple). The Gn/Gc binding site relative to the EC1:EC4 adhesive interface (dark blue) is indicated. An alternative view of the EC1:EC4 binding interface is shown to the right. Structure adapted from PDB 6MGA. b Competition ELISA using WT and mutant sEC1-2 as competitive reagents to the binding of rVSV-ANDV-Gn/Gc to WT sEC1-2 coated wells. Averages ± SD: n = 4 wells of each sEC1-2 dilution examined over two independent experiments [n = 3 for one of the dilutions of sEC1-2(E137R)]. c Relative infectivity of rVSVs bearing ANDV, SNV, or HTNV Gn/Gc on U2OS PCDH1-KO cells complemented with WT or ΔEC4 PCDH1. The infectivity of each virus was normalized to that obtained in U2OS PCDH1-KO cells complemented with WT PCDH1. Means ± SD: n = 9 wells of infected cells examined over three experiments. d Schematic representation of WT or mutant sEC1-4 proteins forming monomers or dimers. e Non-reduced and reduced purified WT and mutant sEC1-4 were separated on an SDS-polyacrylamide gel and visualized by Coomassie Brilliant Blue staining. kDa, kilodalton. f Non-reduced samples in (e) were run on a native-polyacrylamide gel and visualized as in (e). A representative gel from one experiment of two independent experiments is shown for (e) and (f). g ELISA detecting WT and mutant sEC1-4 coated plates, using an anti-Flag-HRP antibody. Mean ± SD: n = 4 wells examined over two independent experiments. (h) Capacity of rVSV-ANDV-Gn/Gc to bind to WT or mutant sEC1-4 coated plates. Done in parallel with (g). Mean ± SD: n = 4 wells of each viral particle dilution examined over two independent experiments. Infectivities (c) and ELISA signal (g) were compared by one-way ANOVA with Dunnett’s test for multiple comparisons. (sEC1-4, soluble extracellular cadherin domains 1–4). Source data are provided as a Source Data file.

Fig. 8. Two point mutations in PCDH1 confer protection of Syrian hamsters against a lethal ANDV challenge.

a Reference nucleotide and amino acid sequence of PCDH1-EC1 Syrian hamster (WT, above) and representative sequences and trace files of Syrian hamsters after CRISPR-Cas9 genome editing [PCDH1(F83A/D85R), lower left] and [PCDH1(10a.a.) lower right]. The nucleotides encoding the corresponding human PCDH1-EC1 Gn/Gc-interacting residues, are highlighted: F83 in purple and D85 in green along with the location of the single guide RNAs (KI, knock-in; sgRNA, single guide RNA). b Immunoblot detecting PCDH1 in lung tissue lysates from WT or CRISPR knock-in mutant Syrian hamsters. Antibody targets PCDH1’s cytoplasmic tail. kDa, kilodalton. A representative blot from a single experiment of two independent experiments is shown. Uncropped blots in Source Data. c Syrian hamster ANDV challenge. Groups of WT, PCDH1(F83A/D85R), and PCDH1(10a.a.) CRISPR knock-in mutant hamsters were inoculated intranasally with ANDV (2,000 PFU). Mortality was monitored and hamsters were euthanized on day 35 post-exposure. One experiment was performed, with n = 8 hamsters for each group. Data was analyzed using two-sided, log-rank Mantel–Cox test. d Lung sections from WT and PCDH1(F83A/D85R) hamsters were collected 15 days post ANDV exposure. Representative histochemical images indicate inflammation in pulmonary tissue (left), ANDV nucleoprotein (N) (middle, tan staining), and ANDV RNA (right, red staining, detected by in situ hybridization). Representative images from one experiment from one out of three hamsters from each group are shown. Scale bars represent 100 µm. Figure (a) includes an image from Flaticon.com. Source data are provided as a Source Data file.