Characterisation of a GII-4 norovirus variant-specific surface-exposed site involved in antibody binding

Abstract

Background: The human noroviruses are a highly diverse group of viruses with a single-stranded RNA genome encoding a single major structural protein (VP1), which has a hypervariable domain (P2 domain) as the most exposed part of the virion. The noroviruses are classified on the basis of nucleotide sequence diversity in the VP1-encoding ORF2 gene, which divides the majority of human noroviruses into two genogroups (GI and GII). GII-4 noroviruses are the major aetiological agent of outbreaks of gastroenteritis around the world. During a winter season the diversity among the GII-4 noroviruses has been shown to fluctuate, driving the appearance of new virus variants in the population. We have previously shown that sequence data and in silico modelling experiments suggest there are two surface-exposed sites (site A and site B) in the hypervariable P2 domain. We predict these sites may form a functional variant-specific epitope that evolves under selective pressure from the host immune response and gives rise to antibody escape mutants.

Results: In this paper, we describe the construction of recombinant baculoviruses to express VLPs representing one pre-epidemic and one epidemic variant of GII-4 noroviruses, and the production of monoclonal antibodies against them. We use these novel reagents to provide evidence that site A and site B form a conformational, variant-specific, surface-exposed site on the GII-4 norovirus capsid that is involved in antibody binding.

Conclusion: As predicted by our earlier study, significant amino acid changes at site A and site B give rise to GII-4 norovirus epidemic variants that are antibody escape mutants.

Figure 1

GII-4 Norovirus Major Sturctural Protein. (a) Schematic representing the norovirus VP1 protein, highlighting the hypervariable P2 domain and putative epitopes Site A and Site B (as described in Allen et al., 2008). (b) Table showing amino acid variation at Site A and Site B in the period 1999-2006, as previously described in (Allen et al., 2008). Strains used in the work described here are highlighted. (c) Model of the norovirus VP1 P domain showing the location of Site A and Site B in the three-dimensional protein (structure from Cao et al., 2007). (d) Electron micrograph showing GII-4v2 VLPs purified from Sf9 cells, the morphology of which is representative for all VLPs described here. Magnification is 105 000× and VLPs are stained with 1.5% phosphotungstic acid. Scale bar is 100 nm.

Figure 2

Schematic representation of norovirus protein coding region of pRN16 constructs expressing wild-type and hybrid VLPs. Following construction of plasmids pRN-GII4v0 and pRN-GII4v2 expressing wild-type GII-4v0 and GII-4v2 VLPs, respectively, these plasmid constructs were modified by site directed mutagenesis at either putative epitope site A (nt 886-891, aa 296-298), or putative epitope site B (nt 1176-1182, aa 383-395) to generate plasmid constructs that expressed hybrid VLPs. The schematic shows a representation of the region of the plasmid encoding norovirus structural proteins (3'UTR and remainder of plasmid not shown for clarity). PCR mutagenesis was used to generate plasmid constructs that encoded an ORF2 identical to either GII-4v0 or GII-4v2, except at either site A or site B, which was modified to be as equivalent to that position in the heterologous variant. The resulting expressed VP1 protein was a hybrid of the two variants, and so the VLP formed from the hybrid VP1 was antigenically hybrid. Plasmid construct names are given on the left, whilst the hybrid VLPs are represented with VLP names next to them on the right. All GII-4v0 derived regions are shown in solid black, all GII-4v2 derived regions are shown in hatched lines.

Figure 3

Average percent reduction in binding of (a) anti-GII-4v0 and (b) anti-GII-4v2 mAbs to wild-type and mutant VLPs in a cross adsorption ELISA. As described in the Materials and Methods, each mAb was pre-incubated in a blocking step with the antigen indicated on the x-axis, before being transferred to a microtitre plate coated with antigen homologous to the mAbs being tested. Percent reduction in binding was then calculated using a PBS control. Cross absorption assays were repeated 3 times independently and the average data is presented here with bars showing the standard error of the mean. Cartoons representing the antigenic structure of the antigen (as described in Figure 2) are shown above the bars (with corresponding labels below the bars). All mAbs were used at 1:10 000 dilution, except mAbGII4v0.5, which was used at 1:1000.