Glycan recognition in globally dominant human rotaviruses

Abstract

Rotaviruses (RVs) cause life-threatening diarrhea in infants and children worldwide. Recent biochemical and epidemiological studies underscore the importance of histo-blood group antigens (HBGA) as both cell attachment and susceptibility factors for the globally dominant P[4], P[6], and P[8] genotypes of human RVs. How these genotypes interact with HBGA is not known. Here, our crystal structures of P[4] and a neonate-specific P[6] VP8*s alone and in complex with H-type I HBGA reveal a unique glycan binding site that is conserved in the globally dominant genotypes and allows for the binding of ABH HBGAs, consistent with their prevalence. Remarkably, the VP8* of P[6] RVs isolated from neonates displays subtle structural changes in this binding site that may restrict its ability to bind branched glycans. This provides a structural basis for the age-restricted tropism of some P[6] RVs as developmentally regulated unbranched glycans are more abundant in the neonatal gut.

Fig. 1

Schematic diagram of type I HBGA synthesis shows the structures of HBGAs in secretor and non-secretor individuals

Fig. 2

VP8* structure of a neonatal human rotavirus strain P[6] RV3 and structural comparison with VP8*s of other HRVs in complex with HBGAs. a Ribbon representation of the P[6] RV3 VP8* structure displays a galectin-like fold with the two twisted β-sheets in yellow and green, respectively. The β-hairpin and the C-terminal α-helix and are colored in orange and purple, respectively. b Structural superposition of P[6] VP8* (gray) with HRV P[14] VP8* (light blue) in complex with A-type HBGA (PDB ID: 4DRV) and HRV P[11] VP8* (tan) in complex with LNT (PDB ID: 4YFZ). The width of the cleft between the β-sheets in P[6] and P[11] VP8*s is wider than that in P[14] VP8*, as indicated by red and black arrows, respectively. The glycans bound on VP8* are represented with sticks. Close-up views of the A-type HBGA (c) and LNT (d) binding sites show the amino acid changes in P[6] VP8* disallow the glycan binding at these sites. The interacting residues are shown in stick model and labeled. The proteins are colored as in b

Fig. 3

Structure of P[6] RV3 VP8* in complex with H type I pentasaccharide. a Structure of P[6] RV3 VP8* is shown in gray ribbon, with the bound glycan in stick model. The glycan residues are labeled. b Superimposition of P[6] VP8* apo structure (cyan) and the liganded P[6] VP8* (gray) shows the structural change of VP8* upon binding to H type I. The changes of sidechain orientations are indicated by black arrows. c Structural alignment of the structure of P[6] RV3 VP8* (gray) in complex with H type I and P[4] DS-1 VP8* (green) apo structure (PDB ID: 2AEN) shows how the Gal-GlcNAc moieties can bind to the conserved the amino acids in P[4] VP8*, and how the changes in P[4] DS-1 lead to loss of interaction with the Glc at the reducing end of H-type I glycan. d Structural alignment of the structure of P[6] RV3 VP8* (gray) in complex with H type I and P[8] Wa VP8* (orange) apo structure (PDB ID: 2DWR). The structures are shown as in c

Fig. 4

Structure-based sequence alignment of VP8*s in Chimera. The residues known to bind glycans are indicated with colored shade. The glycan binding amino acids that are conserved or non-conserved in the prevalent strains (P[6], P[4], P[8], and P[19]) are denoted with black and blue shades, respectively. The precursor glycan binding sites in P[11] RVs are labeled with light red shade. The A-type HBGA binding site in P[14] RV, and the sialic acid binding residues in P[3] and P[7] RVs are denoted with green shade. The PDB IDs for each structure are shown by the sequences

Fig. 5

Structure of P[4] Indian VP8* in complex with H type I pentasaccharide. a Structure of P[4] VP8* is shown in pink ribbon with the bound glycan in stick model. The glycan residues are labeled. b–d Superimposition of the structures of P[4] VP8* (pink)/H-type I and P[6] VP8* (gray)/H-type I. b VP8* in surface representation shows the glycan binding pockets in P[6] and P[8]. c, d The glycan binding residues in P[4] and P[6] are shown in stick model in two views. The rotation of the glycosylic bond between GlcNAc and Gal is indicated with red arrows in c. The interactions between the side chain of H169 in P[6] VP8* and the O6 atom of Gal4, and that of the main chain carbonyl oxygen of Y169 in P[4] VP8* with the oxygen atom O6 of the Gal4 moiety are indicated by red arrows in d

Fig. 6

H-type I glycan inhibits infectivity of the prevalent HRVs. PAA-conjugated H-type I (1 mg/ml) significantly reduced the infectivity of P[4], P[6], and P[8] HRVs but not of a P[3] animal rotavirus (ARV, RRV strain). Each bar represents mean % infectivity, with no glycan treatment considered to be 100%. All assays were carried out a minimum of two times, with triplicates within each experiment. Error bars represent standard error of the mean. P-values < 0.05 were considered statistically significant [analysis of variance (ANOVA) with Sidak’s correction for multiple comparisons]

Fig. 7

Summary of mutations at the glycan binding site in the prevalent strains. a The schematic diagram of glycan recognition in VP8*s of P[6], P[4], P[8], and P[19] genotypes that belong to the P[II] genogroup shows the common residues (black) interact with the Fuc-Gal-GlcNAc moieties, and the sequence changes (blue) lead to distinct recognition of the reducing end of the glycan in VP8*s. The molecular interactions between H type I glycan and P[4]/neonatal P[6] VP8*s are indicated by blue, red and black lines. The corresponding residues in P[6], P[8], and P[19] VP8*s are shown in the lateral columns. b Sequence alignment of VP8*s of the P[II] genogroup including those of porcine rotavirus (PRV) P[6], neonatal and non-neonatal P[6] HRV. The conserved residues that recognize the type I precursor are indicated black shades, and other glycan binding residues are indicated by blue shade. The mutations at residue 169 are highlighted with green and red shades

Fig. 8

Structural basis for resistance to rotavirus infection in non-secretors. P[6] VP8* in complex with H type I pentasaccharide is superimposed onto the structure of VP4 spike (PDB ID: 4V7Q) of P[3] animal strain RRV. The black arrow indicated how the glycan is extended from the C1 atom of Glc residue, and how the virus spike access host cell membrane. A close-up view of the P[6] VP8* with H-type I is shown in the inlet. The addition of a Lewis fucose via an α1,4 linkage to the GlcNAc moiety is indicated by a red hexagon, suggesting that Lewis fucose would clash with VP8*. The addition of GalNAc or Gal at the non-reducing end of HBGA is noted by a green hexagon, which is projecting away from VP8* without making steric hindrance