| Abstract: | Group A rotaviruses cause severe gastroenteritis in infants and young children worldwide, with P[II] genogroup rotaviruses (RVs) responsible for >90% of global cases. RVs have diverse host ranges in different human and animal populations determined by host histo-blood group antigen (HBGA) receptor polymorphism, but details governing diversity, host ranges, and species barriers remain elusive. In this study, crystal structures of complexes of the major P[II] genogroup P[4] and P[8] genotype RV VP8* receptor–binding domains together with Lewis epitope–containing LNDFH I glycans in combination with VP8* receptor-glycan ligand affinity measurements based on NMR titration experiments revealed the structural basis for RV genotype-specific switching between ββ and βα HBGA receptor–binding sites that determine RV host ranges. The data support the hypothesis that P[II] RV evolution progressed from animals to humans under the selection of type 1 HBGAs guided by stepwise host synthesis of type 1 ABH and Lewis HBGAs. The results help explain disease burden, species barriers, epidemiology, and limited efficacy of current RV vaccines in developing countries. The structural data has the potential to impact the design of future vaccine strategies against RV gastroenteritis.The major human rotaviruses (RVs), the P[8], P[4], and P[6] genotypes in the P[II] genogroup, are responsible for over 90% of human infections worldwide (1–3). Despite successes of the RotaTeq and Rotarix RV vaccines in many developed countries, their efficacy remains disappointingly poor in developing countries (4–6). Low efficacies of both vaccines in developing countries can be attributed to a lack of cross protection between P[8], which is more common in developed countries, and other P-type RVs, such as P[6] and P[11], that are less common in developed countries but more common in developing countries (7–13).Significant advances have been made in understanding RV evolution under the selection of stepwise synthesis of histo-blood group antigens (HBGAs) in humans. For example, P[II] RVs that mainly infect humans are thought to have originated from P[I] RVs with an animal host origin and evolved the ability to infect humans under selective pressure to bind polymorphic human HBGAs. This deduction is in agreement with a complete VP4 sequence phylogeny analysis that revealed that P[10]/P[12] in P[I] were genetically closer to P[19], P[6], and P[4]/P[8] in P[II] than other genotypes from other genogroups (14, 15). These observations led to the hypothesis that host ranges of P[II] genotypes for certain animal species and different human populations are dictated by the evolutionary stages of their HBGA receptors.P[19] appears to represent an early evolutionary branch of the P[II] genogroup since it recognizes type 1 precursor HBGAs and therefore commonly infects animals (porcine) but rarely humans. On the other hand, P[4] and P[8] appear to be more evolutionarily advanced since they have developed the ability to recognize more mature HBGA products that dominate in humans. P[6] appears to represent an intermediate stage of evolution close to P[19] that commonly infects both animals (porcine) and humans, likely because of its evolutionary status that allows it to recognize less mature type 1 HBGA precursor glycans shared between humans and animals (porcine). The deduced evolutionary path that enabled the transition from animal to human host, which is correlated with the emergence of the P[II] branch from the P[I] branch, may apply to other genotypes and genogroups and may be important for RV classification and epidemiology (16, 17).Evidence for HBGA-controlled RV host ranges and evolution is also available from structural analyses of genotype-specific interactions of RV VP8* domains with their glycan receptor ligands. For example, early structures showed that VP8* domains from animal and human RVs adopted similar galectin-like folds, and they recognize distinct HBGAs either through a ββ or βα site. However, our recent NMR spectroscopy–based docking and crystallographic studies showed that P[4], P[6], P[8], and P[19] VP8*s of P[II] interacted with H type 1 HBGA precursor using a common βα site (17–21), while P[8] VP8* bound Leb tetra-saccharide and Lewis epitope–containing hexa-saccharide (LNDFH I) in the ββ site (22).To elucidate the molecular basis for receptor-binding bias between βα- and ββ-binding sites, we characterized relative binding affinities of major P[II] RV VP8* domains for glycans representing different HBGA synthetic stages, including the Lewis epitope–containing LNDFH I, using NMR heteronuclear single quantum coherence spectroscopy (HSQC)-monitored titrations. The structural basis for the bias of ββ sites for Lewis epitope HBGAs and βα sites for HBGAs lacking the Lewis epitope was elucidated from crystal structures of P[4] and P[8] bound to LNDFH I and P[6] bound to Lacto-N-tetraose (LNT). Sequence- and structure-based analyses of differences in P[II] VP8* receptor–binding interfaces revealed molecular details responsible for receptor switching between genotype-specific ββ and βα HBGA–binding sites. Overall, the results provide strong evidence for HBGA-controlled P[II] RV evolution from an animal host origin that resulted in diverse genotypes infecting children in different populations and which may impact future strategies for RV disease control and prevention. |
