Norovirus GII.4 Virus-like Particles Recognize Galactosylceramides in Domains of Planar Supported Lipid Bilayers**

In-depth understanding of the cell-surface dependent processes leading to virus binding and infection of host cells, including the identification of new receptors mediating the initial steps, is of central importance for the development of new anti-viral therapies. One still poorly understood but highly contagious virus is the human norovirus, which causes acute gastroenteritis. It is a single-stranded non-enveloped RNA virus belonging to the Caliciviridae family and during recent years, new strains, particularly of the genotype GII.4, have emerged causing worldwide epidemics. However, detailed knowledge on 1) the target cell, 2) the mechanism of viral entry, and 3) the receptor(s) for cellular uptake are still lacking. Because there is presently no simple cell-culture model available for propagating human norovirus, experimental studies aimed at resolving (1) to (3) are heavily dependent on in vitro production of virus-like particles (VLPs), which, for the norovirus, form spontaneously when the major capsid protein (VP1) is recombinantly expressed. Such VLPs show a symmetric icosahedral shell assembled from 180 copies of the VP1 protein, with a morphology, antigenicity, and binding characteristics that reflect those of native viruses. VLP binding studies have revealed that most human noroviruses recognize, in a strain-specific manner, ABO, Lewis (FUT3), and secretor (FUT2) gene-dependent histo-blood group antigens (HBGAs), on glycoproteins and glycosphingolipids. VLPs also bind in a fucose-dependent manner to epithelial cells of duodenal tissue sections and saliva from secretors but not from non-secretors. This study was motivated by the fact that histopathological biopsies point to the upper small intestine as the likely target organ for this virus. Besides type 1 chain fucolipids, epithelial cells of the human small intestine contain large amounts of monoglycosylceramides (mainly b-galactosylceramide, (GalCer)). Moreover, although not previously explored in the context of norovirus infection, GalCer and other short-chain glycosphingolipids (GSLs) have been identified as facilitators of other viral infections, such as HIV. To explore if GalCer could potentially be a receptor for norovirus infection, thus acting alone or in combination with other HBGAs on glycolipids that have been identified to interact specifically with different norovirus strains, this work is focused on characterizing norovirus VLP binding to GalCer. GSLs are usually found together with cholesterol and sphingomyelin in phase-separated domains. In addition to the verified role of GSL domains in a variety of cellular processes, there is also increasing evidence that such domains can be sites for pathogen attack. Of particular relevance to this work, GalCer is believed to cluster into microdomains, as it has been isolated in the form of detergent-resistant membranes. Phase separation of GalCer is further supported by a variety of studies on model supported lipid bilayers (SLBs). Herein, we report the recognition of GalCer by norovirus VLPs, with particular focus on the dependence of this interaction on GSL presentation and packing, including domain formation. Protein–carbohydrate interactions between the VLPs from norovirus and GalCer were identified using the chromatogram binding assay (CBA), a well-established method for probing binding of proteins to GSLs. As shown in Figure 1, VLPs bound to reference H type 1 GSLs, typically present in the small intestine of an OLe(a-b-) secretor but not of an OLe(a+b-) non-secretor, a phenotype known to be genetically resistant to infection by most GII.4 norovirus strains. Interestingly, VLPs also bound to monoglycosylceramides, a mixture of GlcCer and GalCer, of smalland [*] M. Bally, F. Hççk Department of Applied Physics, Chalmers University of Technology SE-41133 Gçteborg (Sweden) E-mail: [email protected]