Structure-Guided Identification of a Nonhuman Morbillivirus with Zoonotic Potential

Abstract

Morbilliviruses infect a broad range of mammalian hosts, including ruminants, carnivores, and humans. The recent eradication of rinderpest virus (RPV) and the active campaigns for eradication of the human-specific measles virus (MeV) have raised significant concerns that the remaining morbilliviruses may emerge in so-called vacated ecological niches. Seeking to assess the zoonotic potential of nonhuman morbilliviruses within human populations, we found that peste des petits ruminants virus (PPRV)-the small-ruminant morbillivirus-is restricted at the point of entry into human cells due to deficient interactions with human SLAMF1-the immune cell receptor for morbilliviruses. Using a structure-guided approach, we characterized a single amino acid change, mapping to the receptor-binding domain in the PPRV hemagglutinin (H) protein, which overcomes this restriction. The same mutation allowed escape from some cross-protective, human patient, anti-MeV antibodies, raising concerns that PPRV is a pathogen with zoonotic potential. Analysis of natural variation within human and ovine SLAMF1 also identified polymorphisms that could correlate with disease resistance. Finally, the mechanistic nature of the PPRV restriction was also investigated, identifying charge incompatibility and steric hindrance between PPRV H and human SLAMF1 proteins. Importantly, this research was performed entirely using surrogate virus entry assays, negating the requirement for in situ derivation of a human-tropic PPRV and illustrating alternative strategies for identifying gain-of-function mutations in viral pathogens.IMPORTANCE A significant proportion of viral pandemics occur following zoonotic transmission events, where animal-associated viruses jump species into human populations. In order to provide forewarnings of the emergence of these viruses, it is necessary to develop a better understanding of what determines virus host range, often at the genetic and structural levels. In this study, we demonstrated that the small-ruminant morbillivirus, a close relative of measles, is unable to use human receptors to enter cells; however, a change of a single amino acid in the virus is sufficient to overcome this restriction. This information will be important for monitoring this virus's evolution in the field. Of note, this study was undertaken in vitro, without generation of a fully infectious virus with this phenotype.

Keywords: PPRV; host range; measles; morbillivirus; paramyxovirus; zoonoses.

FIG 1

PPRV glycoprotein-mediated fusion is restricted with the human SLAMF1 receptor. (A) The established mammalian morbilliviruses (phylogenetic tree assembled from complete genome sequences using Vector Nti; 0.1 scale bar; nucleotide substitutions). FmoPV, feline morbillivirus. (B) MeV cell-cell fusion is efficient when either hSLAM (black) (left panel) or oSLAM (gray) (middle panel) is present on target cells; however, PPRV is restricted to efficient fusion with oSLAM only. Minimal fusion is seen with HEK293 cells alone (right panel). Raw Renilla luciferase assay data are shown (RLU, relative light units). Noncognate virus-receptor interactions are indicated with patterned shading. (C) The current distribution of PPRV according to the World Organization for Animal Health (OIE) and the Food and Agriculture Organization for the United Nations (FAO). The countries highlighted in red have ongoing (or previously had) outbreaks of PPRV. Additional representative isolates of PPRV or genetic lineages I to IV from the indicated countries (Senegal, Benin, Kenya, and Ethiopia) were also used in this study to confirm the hSLAM restriction. The amino acid sequences of the PPRV H RBD at site 3 (aa 191 to 195) from these additional isolates (RAVTR and RTVTR) are shown. (D and E) All PPRV H proteins tested showed similar restrictions with respect to hSLAM-mediated cell-cell fusion. Results are expressed relative to cognate virus-host interactions, i.e., MeV-hSLAM (D) or PPRV (Turkey; lineage IV)-oSLAM (E). Noncognate virus-receptor interactions are indicated with patterned shading. (F) Western blot analysis of the various PPRV H proteins, from distinct lineages, expressed in the cell-cell fusion assays; both monomeric H and dimeric H were detected using an antibody targeting the cytoplasmic tail of H (H-cyt). Graphs denote the mean activity levels determined for >4 biological replicates, with error bars denoting standard deviations. Statistical analysis: one-way analysis of variance (ANOVA) with Dunnett’s multiple-comparison test (*, P = <0.05; **, P = <0.01; ***, P = <0.005; ****, P = <0.001; ns, nonsignificant).

FIG 2

Sequence variation within RBD site 3 may determine virus host range. (A) Multiple alignments of various morbillivirus H proteins, focusing on the 4 sites that constitute the RBD. The surrounding sequences are shown for clarity, although the exact amino acid positions and sequences are indicated (relative to the MeV isolate from Dublin). Of note, a number of specific amino acid residues were found in more than one site (Y524 and F552). The PPRV reference sequence used in this study is PPRV AJ849636. The lapinized RPV strain is AB547190. (B) Structure of measles virus H (MeV H) in complex with marmoset SLAMF1 (maSLAM) receptor (PDB 3ALX) (24). (Top) The complex is shown as ribbons. (Bottom) The interaction interfaces between MeV and maSLAM are shown as molecular surfaces that are labeled 1 to 4 and colored as defined by Hashiguchi et al. (24). (C) Multiple alignments of various morbillivirus H proteins, focusing on the 4 sites that constitute the RBD. The surrounding sequences are shown for clarity, although the exact amino acid positions and sequences are highlighted. Conservation within the separate sites was analyzed using the WebLogo online server. The overall variation detected (given as percent identity) within each site is also indicated. The sequences used for this analysis are provided in Data Set S1 together with the relative amino acid positions. (D) MeV H:maSLAM interaction region 3 comprises an interaction between adjacent anti-parallel β-sheets. The ribbon representation is colored as described for panel B, with selected residues shown as sticks (MeV H carbon atoms in pink, maSLAM carbon atoms in red) and selected backbone hydrogen bonds shown (yellow dots). (E) Morbillivirus interspecies amino acid variability within site 3 (aa 191 to 195; N-C amino acid termini are indicated) is high (comparative sequence alignment is indicated at the top of the panel); however, intraspecies variability is markedly lower (analysis of a complete spectrum of circulating MeV and PPRV genotypes is indicated at the bottom of the panel—see Data Set S1).

FIG 3

Minor changes to the PPRV H RBD overcome species-specific restrictions. (A) A chimeric PPRV H containing the site 3 amino acid sequence of MeV (RTVTR—PTTIR) is no longer restricted by hSLAM activity in the cell-cell fusion assay. Results are expressed relative to wild-type (WT) PPRV H and hSLAM interactions (see Fig. 1B). (B) Western blot analysis of the WT and mutant PPRV H expressed in effector cells shows equivalent expression levels of both monomeric and dimeric H protein. (C) MeV chimeric mutations within PPRV H site 3 have a neutral effect on oSLAM-dependent fusion. Results are expressed relative to WT PPRV H and oSLAM interactions. In all panels, graphs denote the mean activity from biological replicates, with error bars denoting standard deviations. Statistical analysis: one-way ANOVA with Dunnett’s multiple comparisons tests (*, P = <0.05; **, P = <0.01; ***, P = <0.005; ****, P = <0.001; ns, nonsignificant).

FIG 4

Restriction and GOF phenotypes are recapitulated in a separate model of viral entry. (A) Equivalent levels of PPRV H hSLAM restriction were observed using a PPRV-pseudotyped (PP) HIV-1-based entry assay. The graph denotes the mean activity from >4 biological replicates, with error bars denoting standard deviations. Noncognate virus-receptor interactions are indicated with patterned shading. Statistical analysis: Two-way ANOVA with Sidak’s multiple-comparison test (****, P = <0.001; ns, not significant). NE, nonenveloped control pseudotypes. (B) PPRV PPs bearing PPRV^MeV H protein chimeras overcome the hSLAM restriction, specifically, the full site 3 chimera (PTTIR) and the single R191P mutation, while having minimal effects on oSLAM-mediated fusion. Results are expressed as percent change relative to PPRV WT H (unmutated) PP entry. Graphs denote the mean activity from >4 biological replicates, with error bars denoting standard deviations. Noncognate virus-receptor interactions are indicated with patterned shading. Statistical analysis: one-way ANOVA with Dunnett’s multiple-comparison test (*, P = <0.05; **, P = <0.01; ***, P = <0.005; ****, P = <0.001; ns, nonsignificant). (C) A PPRV minigenome, supported by trans-expression of PPRV N, P, and L proteins, is functional in human A549 cells. Experiments were performed in triplicate. Error bars denote standard deviations; luciferase assays were normalized against untransfected cells.

FIG 5

Random mutagenesis of PPRV H R191 identifies numerous GOF mutations. (A) Using the standard cell-cell fusion assay, random mutagenesis of the arginine (R) residue in PPRV H (R191) revealed that multiple amino acid changes are sufficient to overcome the PPRV-hSLAM restriction (diagonally striped bars). The effects of these amino acid changes on oSLAM-dependent fusion are also indicated (gray bars). All results are expressed relative to PPRV WT H (unmutated) interactions with hSLAM and oSLAM (left and right panels, respectively). (B) Similar results were observed using a subset of these mutants and the PPRV-pseudotyped (PP) HIV-1-based entry assay; all results are expressed as percent change relative to PPRV WT (unmutated) H PP entry. Left panel, hSLAM; right panel, oSLAM. (C) Various mutations at PPRV H amino acid position 191 do not affect the relative stability of this protein. kDa size markers refer to SDS-PAGE ladder positions. (D) Model of PPRV H in complex with maSLAM. The maSLAM surface is colored according to residue hydrophobicity, from white (polar) to green (hydrophobic). (E) H residue 191 interacts with conserved SLAM residue 131. In the maSLAM-plus-MeV H complex (top), the prolidyl ring of P191 forms a stacking interaction with the hydrophobic F131 side chain. The adjacent disulfide bond that connects the A and G strands (A^# and G^#) of the maSLAM Ig domain is highlighted. The presence of the bulky charged residue arginine at this location in PRRV H (bottom) would likely interfere with complex formation. Residues in MeV H:maSLAM interaction region 3 are colored as described for Fig. 2D. In all panels, graphs denote the mean activity from >4 biological replicates, with error bars denoting standard deviations. Statistical analysis: one-way ANOVA with Dunnett’s multiple-comparison test (*, P = <0.05; **, P = <0.01; ***, P = <0.005; ****, P = <0.001).

FIG 6

Minor variations in mammalian SLAMF1 proteins determine morbillivirus host range. (A) An amino acid alignment of the HBS of hSLAM and oSLAM highlights variations between the two proteins in sites 1 to 4. The asterisks (*) indicate that analysis of site 4 was extended to include S73 and N76. (B) Stepwise mutation of the hSLAM sequence to generate a site 3 or site 4 oSLAM chimera is sufficient to overcome the PPRV-hSLAM restriction. Results are expressed relative to WT PPRV H protein interactions with the unmutated WT hSLAM receptor. (C) Western blot analysis of the WT and mutant hSLAM sequences expressed in target cells (antibody, SLAM [N-19], Santa Cruz sc-1334). Graphs denote the mean activity from >4 biological replicates, with error bars denoting standard deviations. Statistical analysis: one-way ANOVA with Dunnett’s multiple-comparison test (*, P = <0.05; **, P = <0.01; ***, P = <0.005; ****, P = <0.001; ns, nonsignificant).

FIG 7

Host variation in oSLAM is not associated with altered PPRV-glycoprotein mediated fusion. (A) Amino acid variants encoded within the ovine SLAMF1 gene locus were characterized in a diverse set of 171 sheep with available genome sequences. From these data, a rooted maximum parsimony phylogenetic tree of the haplotype-phased protein variants was constructed. Each node in the tree represents a different, naturally occurring protein isoform; the isoforms differ by single amino acids. The areas of the circles are proportional to the variant frequencies (see Data Set S1). (B) Model of MeV H (white ribbons) in complex with oSLAM (light blue molecular surface), showing that the ovine SNPs (pink sticks) lie outside the likely HBS (residues equivalent to maSLAM HBS are colored as described for Fig. 2B). (C) Variation in oSLAM does not markedly affect the cell-cell fusion potential of PPRV F and H proteins. Results are expressed relative to the dominant variant of oSLAM (V1; S₉R₈₅H₉₁D₉₄C₂₇₂). (D) Western blot analysis of the eight variant oSLAMs (V1 to V8) expressed in target cells (antibody: His tag). The graph denotes the mean activity from >4 biological replicates, with error bars denoting standard deviations. Statistical analysis: one-way ANOVA with Dunnett’s multiple-comparison test (*, P = <0.05; **, P = <0.01; ***, P = <0.005; ****, P = <0.001).

FIG 8

Minor-frequency nsSNPs within the hSLAM HBS are associated with altered MeV-glycoprotein mediated fusion. (A) Allele frequencies of human SLAMF1 nsSNPs within the HBS of hSLAM. (B) Residues equivalent to nsSNPs found within the publicly available EXAC database and the hSLAM HBS are shown on the molecular surface of maSLAM (HBS regions are colored as described for Fig. 2B). (C) Variation within the hSLAM HBS reduced MeV-induced cell-cell fusion. Results are expressed relative to the WT hSLAM amino acid sequence (UniProt accession no. Q13291). (D) Western blot analysis of the three variant hSLAMs (and of the dominant sequence [WT]) expressed in target cells. The graph denotes the mean activity from >4 biological replicates, with error bars denoting standard deviations. Statistical analysis: one-way ANOVA with Dunnett’s multiple-comparison test (*, P = <0.05; **, P = <0.01; ***, P = <0.005; ****, P = <0.001).

FIG 9

Cross-protective neutralization is affected by gain-of-function mutations within the PPRV H RBD. (A and B) Neutralization of WT and R191P-bearing PPRV PPs by sera from goats (PPRV specific; left panel; n = 10) (A) and humans (MeV specific, right panel; n = 8) (B). EC₉₀ titers (color-matched by serum) are shown; error bars denote standard errors of the means. Antibody titers were calculated by interpolating the point at which there was a 90% reduction in luciferase activity (90% neutralization or 90% inhibitory concentration [EC₉₀]). Statistical analysis was performed using the nonparametric Wilcoxon matched-pair signed rank test (*, P =<0.05). (C) In certain human serum samples, e.g., 377 and 584 (red and green squares, respectively), R191P confers a nAb escape phenotype to PPRV PPs. Surrogate VNT titrations were performed with individual MeV-specific human sera to calculate cross-protective titers against WT and R191P PPRV pseudotypes. Titrations were performed in triplicate on nonrestricted HEK293 canine SLAM cells, with error bars denoting standard errors of the means. The coloring of the lines within the graphs matches the sera used in the summary EC₉₀ panel (i.e., panel B).

FIG 10

PPRV has zoonotic potential in human populations. (A) Factors influencing the emergence of morbilliviruses in atypical hosts. (B) The chance of PPRV emergence in humans. Blue line, the rate of spillovers of a variant of PPRV, capable of human-to-human transmission, into a human population in which a fraction n of individuals have cross-protective nAbs, relative to the equivalent rate in a population where no individuals have cross-protective nAbs [z(n)z(0)]; black line, following a single spillover event, the probability of a major outbreak driven by human-to-human transmission [p(n)]; red line, relative rates of major outbreaks in human populations, compared to a population where no individuals have cross-protective nAbs [m(n)m(0)]. Refer to Materials and Methods for detailed formulas. We assumed that a variant of PPRV capable of human-to-human transmission would have an R₀ value of 6.85 in human populations with no cross-protective nAbs. This represents the estimated basic reproduction number of PPRV in Afghan (Bulkhi) sheep in Pakistan (60).