Histo-Blood Group Antigen Presentation Is Critical ... - ACS Publications

Subscriber access provided by University of Newcastle, Australia

Article

Histo-blood group antigen presentation is critical for norovirus VLP binding to glycosphingolipids in model membranes Waqas Nasir, Martin Frank, Angelika Kunze, Marta Bally, Francisco Parra, Per-Georg Nyholm, Fredrik Höök, and Göran Larson ACS Chem. Biol., Just Accepted Manuscript • DOI: 10.1021/acschembio.7b00152 • Publication Date (Web): 15 Mar 2017 Downloaded from http://pubs.acs.org on March 17, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Chemical Biology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Chemical Biology

Histo-blood group antigen presentation is critical for norovirus VLP binding to glycosphingolipids in model membranes Waqas Nasira, Martin Frankb, Angelika Kunzec, Marta Ballyc, Francisco Parrad, PerGeorg Nyholmb,e, Fredrik Höökc, Göran Larsona,*

a

Department of Clinical Chemistry and Transfusion Medicine, Sahlgrenska Academy, University of

Gothenburg, Gothenburg, Sweden b

Biognos AB. Generatorsgatan 1, P.O. Box 8963, 40274 Gothenburg, Sweden

c

Department of Applied Physics, Chalmers University of Technology, S-412 96 Gothenburg, Sweden

d

Instituto Universitario de Biotecnología de Asturias, Departamento de Bioquimíca y Biología

Molecular, Universidad de Oviedo, 33006 Oviedo, Spain e

Department of Medical Biochemistry and Cell Biology, University of Gothenburg, Gothenburg,

Sweden *Corresponding Author Göran Larson, Department of Clinical Chemistry and Transfusion Medicine, Sahlgrenska Academy, University of Gothenburg, Bruna Stråket 16, SE-413 45 Gothenburg, Sweden Email: [email protected] Phone: +46706250216

Keywords Norovirus, glycosphingolipids, epitope presentation, histo-blood group antigens, total internal reflection fluorescence microscopy, membrane, molecular dynamics

1 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 35

Abstract Virus entry depends on biomolecular recognition at the surface of cell membranes. In case of glycolipid receptors, these events are expected to be influenced by how the glycan epitope close to the membrane is presented to the virus. This presentation of membrane associated glycans is more restricted as compared to glycans in solution, particularly because of orientational constraints imposed on the glycolipid through its lateral interactions with other membrane lipids and proteins. We have developed and employed total internal reflection fluorescence microscopy (TIRFM) based binding assay and a scheme for molecular dynamics (MD) membrane simulations to investigate the consequences of various glycan presentation effects. The system studied was histo-blood group antigen (HBGA) epitopes of membrane bound glycosphingolipids (GSLs) derived from small intestinal epithelium of humans (type-1 chain) and dogs (type-2 chain) interacting with GII.4 norovirus-like particles (NVLPs). Our experimental results showed strong binding to all lipid linked type-1 chain HBGAs but no or only weak binding to the corresponding type-2 chain HBGAs. This is in contrast to results derived from STD experiments with free HBGAs in solution were binding was observed for Lewis x. The MD data suggest that the strong binding to type-1 chain glycolipids was due to the well exposed (1,2)-linked α-L-Fucp and (1,4)-linked α-L-Fucp residues while the weaker or no binding to type-2 chain HBGAs was due to the very restricted accessibility of the (1,3)linked α-L-Fucp residue when the glycolipid is embedded in a phospholipid membrane. Our results contribute not only to a general understanding of protein-carbohydrate interactions on model membrane surfaces, particularly in the context of virus binding, but also suggest a possible role of human intestinal GSLs as potential receptors for norovirus uptake.


Page 3 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Introduction All living cells are surrounded by a glycan coat. Virus-host interactions occurring at the surface of the target cell membrane mark the beginning of viral infection as well as the departure of newly formed virions. Many non-enveloped viruses, e.g. SV40 and rotaviruses, recognize carbohydrates of glycolipids and glycoproteins on the surface of their target cells, which may mediate such interactions. Similar interactions appear to be true for human noroviruses (NV) which are small (diameter = ~40 nm) non-enveloped RNA viruses with T=3 icosahedral symmetry. These viruses appear as global and local outbreaks of severe gastroenteritis and represent major threats to various community settings owing to their highly contagious nature1, 2. GII.4 NV strains are estimated to cause 60-80 % of all NV outbreaks3. Currently, there is no vaccine available for noroviruses, although the first generation of vaccines using so-called virus-like particles (VLPs) are in early clinical trials. The relatively low pace of research progress in norovirus field has been due to the lack of culturing methods for virus propagation in vitro. Consequently, norovirus research has been limited to epidemiological studies, genetic characterization of virus strains and model studies using recombinantly expressed VLPs and their interactions with host cell glycans. Such binding studies involving GII noroviruses have shown that the highly multivalent NV VLPs (carrying 180 carbohydrate binding sites per particle) specifically recognize certain histo-blood group antigens (HBGAs). These HBGAs, are carbohydrate determinants found at the non-reducing ends of the glycan moieties of glycolipids and/or glycoproteins, belonging to either the type-1 (precursor: Galβ1,3GlcNAcβ) or the type-2 (precursor: Galβ1,4GlcNAcβ) chain structures. In all the HBGAs recognized by GII VLPs a fucose residue (α-L-Fucp) was found to be the common nominator4-7. Fucose can be added to both type-1 and type-2 chain precursors (Figure 1, Table 1) in specific linkages by the action of α1,2-, α1,3 - or α1,4fucosyltransferase enzymes. For e.g. the human secretor gene (FUT2) encoded enzyme, i.e.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 35

fucosyltransferase-2 (Fuc-T2), adds α-L-Fucp in (1,2)-linkage to the terminal Galβ residue of preferentially type-1 chain precursors thereby producing H type 1 antigen in saliva and mucosa. Importantly, the genetic lack of an active α1,2-fucosyltransferase, as found in nonsecretors (FUT2 -/-), has been shown to cause innate resistance to most but not all norovirus strains appearing in outbreaks8. A landmark in norovirus research was recently achieved when various human norovirus strains, including GII.4, were shown to replicate successfully in stem cell-derived human intestinal enteroid (HIE) cultures9. Interestingly, FUT2-/- HIE cultures did not result in successful GII.4 norovirus replication. HBGAs are indeed presented on the surface of human intestinal cells as glycosphingolipids (GSLs), mainly but not exclusively of type-1 chain structures10, and on glycoproteins containing mainly type-2 chain HBGAs11. The GSLs (Figure 1) are comprised of a hydrophobic ceramide part, which serves as the lipophilic membrane anchor, and a hydrophilic saccharide moiety which is structurally diverse related to the genetic polymorphism of the ABO, Lewis (FUT3) and Secretor (FUT2) genes. The limited degrees of freedom of the ceramide-saccharide glycosidic linkage and the lateral interactions with neighboring lipids12,

13

and proteins are considered to restrict the orientational and

potentially also the conformational flexibility of membrane associated HBGAs as compared to the free HBGAs in solution. Whether these restrictions have implications on glycan recognition by viral proteins is poorly understood. However, studies have indicated that the differences in orientation of the saccharide moiety of a membrane glycolipid may affect bacterial adhesin receptor availability14 and alter receptor specificity15. In another study it was shown that the availability of a glycolipid receptor for toxin binding may be affected through cholesterol induced changes of glycolipid head group orientation 16. We here describe for the first time in the context of virus binding, how restrictions of glycan orientation and flexibility in model lipid membranes may affect the availability of


Page 5 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


membrane associated HBGA ligands of glycosphingolipids for virus recognition. Specifically, we report the binding patterns of GII.4 NVLPs with membrane bound HBGAs using purified GSLs from human meconium samples (type-1 chain) or canine intestinal epithelia (type-2 chain). We compare the results with the reported binding of the same GII.4 NVLPs with the corresponding oligosaccharides in solution17.

Analysis of the total internal reflection

fluorescence microscopy (TIRFM) experiments and nanoseconds to microseconds molecular dynamics (MD) simulations are combined to describe and explain the NVLP binding characteristics to the lipid linked HBGA structures at the model membrane surface. Our results suggest that not only the structure of a glycan ligand, but also its presentation in the membrane is equally critical for successful recognition by the viral protein. Although our model experimental system and simulation methods lack the complexity of a real cellular membrane we argue that our state of the art techniques and results represent an important step in the elucidation of presentation effects at membrane surfaces. Specifically, the current findings are of relevance for understanding the possible role of GSLs as receptors/attachment factors for NVs on the surface of intestinal enterocytes.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 35

Results and discussion This report elucidates the importance of the structures of the glycan binding epitopes and the importance of their presentation at the membrane surface in relation to virus binding. We address the impact of the orientational restrictions of membrane associated glycans on the binding of NV VLPs to glycosphingolipids using TIRFM binding assay

18

and MD

simulations. Analysis of the interactions of GSL containing vesicles with NVLPs using TIRFM The TIRFM methodology (Figure 2) was used to obtain binding data for different GSLcontaining vesicles revealing that all type-1 structures were recognized by the Ast 6139 GII.4 VLPs although with different kinetics. Among the type-2 structures only Le-y, but neither Lex nor ALe-y containing vesicles showed attachment to the membrane bound VLPs. Based on the shorter (3 secs) or longer (30 secs) residence times of the VLP-bound vesicles representing stronger and weaker VLP-vesicle interactions respectively, two different sets of association (Figure 3a) and dissociation kinetics (Figure 3b and 3c) were deduced from the measurements. In both cases Le-b and H-1 GSLs exhibited the highest rates of association (Figure 3a and Table 2) whereas, the vesicles containing ALe-b, A-1 and B-1 showed relatively slower attachment kinetics. Interestingly, the vesicles with ALe-b and B-1 showed the slowest rate of release and the highest number of irreversibly bound vesicles which is indicative of strongest binding for both time windows. The vesicles with Le-a demonstrated weak binding (smallest number of irreversibly bound vesicles) evident from fastest depletion of bound vesicles (Figure 3b). For negative controls and for GSL vesicles, which did not recognize the VLPs, the number of associated vesicles was extremely low and not enough to deduce any dissociation kinetics. A detailed treatment of dissociation behavior of the GSLcontaining vesicles and the corresponding activation energy of detachment is reported


Page 7 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


elsewhere19. The control experiments performed without VLPs or with VLPs but without GSLs incorporated in the lipid vesicles did not result in any detectable binding and thus no kinetics was deducible (Figure 3a). Chromatogram binding assay (CBA) CBA is a general surface-based technique in which glycosphingolipids are chromatographed over a thin-layer plate and assayed for binding to proteins of interest including bacteria and viruses20, 21. The CBA results are presented in Figure 4. All type-1 chain HBGAs showed binding to the VLPs whereas, for type-2 chain structure only Le-y was recognized by the VLPs. No bands were observed for neo-lactotetra (Galβ4GlcNAcβ3Galβ4Glcβ-Cer), the type2 chain GSL precursor, used as a negative GSL control. Similarly, control experiments were performed in the absence of VLPs which did not result in any unspecific NV-antibody binding to the GSLs. In agreement with the results from the TIRFM analysis, Le-y was the only type-2 chain HBGA recognized by the NVLPs. The faintest band observed was for Le-a GSL and the most intense band was observed for B-1 (Figure 4). Clearly, NVLPs did not show any attachment to Le-x or to ALe-y GSLs neither in the CBA assay nor in the TIRFM assay (Figure 3a and 4). This was in contrast to the previous STD NMR report showing binding of the same VLP with soluble HBGAs, where significant saturation transfer for (1,3)-linked α-L-Fucp in Le-x was observed in the presence of VLPs17. This strongly suggested that the presentation of the (1,3)-linked α-L-Fucp in Le-x GSL in membranes does not allow for VLP recognition. This led us to perform the membrane MD simulations and thorough structural analyses on all the GSLs considered in the present study in order to further investigate the possible presentation differences between the α-L-Fucp residues of type-1 and type-2 chain HBGAs on membrane surfaces.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 35

Epitope presentation analysis of HBGAs in lipid membranes by MD simulations and ‘capsid/membrane clashing analysis’ The spatial orientations (“presentation”) of α-L-Fucp residues present in membraneembedded HBGA glycolipids were analyzed based on membrane MD simulations. For H-1, which is our reference glycolipid and represents in principle all the type-1 chain structures, the (1,2)-linked α-L-Fucp residue was located for most of the simulation time of 400 ns about 15 Å (Figure 5b, right hand side, top) above the ‘membrane surface’. This surface was approximated by the xz plane going through the glycosidic oxygen (named OL) that links the lipid moiety with the innermost glucose (Glc) residue of the carbohydrate chain. The angle β, describing the orientation of the α-L-Fucp residue, fluctuated about 15° around a value of 30°, meaning that the α-L-Fucp ring plane is orientated almost parallel to the membrane surface (Figure 5b, right hand side, bottom, SI Figure S1). Analysis of the bond vectors OL-C1 (Glc) and C4-O4 (Fuc) (angles α and β, Figure 5c, left hand side) revealed that there is no direct correlation between the orientation of the Glc to the ceramide linkage and the orientation of the terminal Fuc. This indicates that the presentation of the terminal α-L-Fucp is influenced by the conformational flexibility of the glycosidic linkages and that the carbohydrate moiety, although with limited degrees of freedom, does not behave like a ‘rigid body’. Additionally, the proximal Glc residue is found only partly exposed to the solvent and does not extend significantly beyond the polar layer of the membrane, which is in agreement with NMR studies22. In order to check which fractions of the MD snapshots sampled would be suitable for attachment of Norovirus capsid particles, a complete VLP was modeled (based on crystal structures 1IHM and 2OBT) and a ‘clashing analysis’ (Figure 5) was performed for all MD snapshots as described in SI Supplementary Methods. In brief, complete virus-capsid/GSL complexes were built using superimpositions of each α-L-Fucp residue present in all


Page 9 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


glycolipids for all snapshots, sampled during the membrane simulations, onto the crystallographic data of the α-L-Fucp in the VA387/B-Tri complex (Pdb id: 2OBT). Those snapshots that did not lead to significant clashing of the capsid with the membrane surface, were then counted as suitable for binding. Based on the results of the capsid/membrane clashing analysis (Figure 5c, right hand side) the type-1 structures including ALe-b, Le-b and H-1 gave the highest percentages of snapshots with accessible α-L-Fucp residues and allowed for binding both to (1,2)- and to (1,4)-linked α-L-Fucp residues. Among the type-2 structures, i.e. ALe-y, Le-y and Le-x, only Le-y remained accessible for binding and only through the (1,2)-linked α-L-Fucp residue. The (1,3)-linked α-L-Fucp residue of the type-2 chain Lewis structures was essentially not accessible without clashes and can therefore not support binding to membrane-bound glycans. The lack of measurable VLP binding to Le-x, as found in the TIRFM assay, may therefore be explained by this “clashing analysis”. Microsecond MD simulations of free HBGAs and ‘capsid/membrane linking analysis’ The explicit simulation of GSLs in lipid membranes with water is limited by the fact that sampling a conformational equilibration of such large molecular systems may take a long time, far above the time scales that are currently feasible using standard all-atom membrane MD simulations. We developed an alternative and independent modeling strategy in which the oligosaccharide was simulated with water for several micro-seconds and then we checked which percentage of sugar conformations sampled were geometrically suitable to ‘link’ the virus capsid with the membrane surface (‘capsid/membrane linking analysis’). The method is described in detail in SI supplementary methods. In brief, if one assumes, that in the membrane environment, the orientational preferences of the Glc at the reducing end is the same for all HBGAs, then the presentation of the α-L-Fucp epitope is solely determined by internal differences between the HBGAs themselves. Therefore, the accessible conformational space of the free HBGAs was sampled in water solution for 5 µs. For each HBGA, a



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 35

conformational ensemble consisting of 0.5 million snapshots respectively was further used in order to check whether the HBGAs are able to adopt conformations that are suitable to link the α-L-Fucp binding site of the NVLP with a correctly oriented lipid in the membrane. It was found, by analyzing the ‘non-clashing’ poses from the capsid/membrane clashing analysis of the H-1/VLP complex (Figure 5b, SI Figure S2), that for successful binding, the α-L-Fucp needs to be located >7 Å above the lipid linking atom OL and the α-L-Fucp ring plane has to be about parallel to the membrane with a maximum deviation of about 20°. Otherwise the atoms of the virus capsid would clash significantly into the membrane. The increase in clashes with deviation from the parallel orientation follows very steep curves (SI Figure S2), meaning that there is very limited tolerance. With this knowledge at hand the α-L-Fucp residues of the various HBGAs were then aligned in the xz plane, meaning parallel to virtual membrane surface, (Figure 6a-b) and it was checked in which fraction of the MD snapshots the oxygen at the reducing end (“OL”) has ycoordinates < -7 Å (Figure 6b, right hand side and SI Figure S3). The results are summarized in Figure 6c. In these conformations of the HBGA (visualized by the green shading in Figure 6c), the oxygen at the reducing end protrudes sufficiently from the VLP surface allowing for the construction of a 3D model of the capsid/HGBA-lipid/membrane system without atomic clashes. It can be seen (Figure 6c, and SI Figure S3) that for all (1,2)- and (1,4)-linked α-LFucp residues, the above criteria could be fulfilled for all of the HBGAs. However, in the case of (1,3)-linked α-L-Fucp residues of the type 2 chain Lewis structures, which have the reducing end OL located in the y-range [-5, 5], such a geometry would have the membrane clash with the VLP atoms. The internal linkage graph of Le-x therefore appears as not compatible with the geometrical requirements for a successful attachment of the VLP to the membrane-embedded HBGA. For ALe-y containing vesicles, no experimental binding was observed with the VLPs which is in good agreement with our limited membrane simulation


Page 11 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


data. However, another relevant observation in this regard is related to a crystal structure reported for Le-y in complex with norovirus GII.9 VA207 P dimer23. We note that extension of GalNAc residue (the A epitope) on this Le-y HBGA conformation would induce strong clashes with the binding site residues of the VA387 crystal structure (SI Figure S4). This clashing would also explain the non-binding of ALe-y in our experiments using Ast6139 GII.4 VLPs. In summary, both the experimental and structural analysis based on membrane MD simulation data demonstrate that the presence of a binding epitope (e.g. Le-x) in a molecule may not be sufficient for virus attachment in a biological system. It has already been shown that the membrane environment (e.g. the presence of cholesterol16) can have an effect on the orientation of a glycolipid head group thereby hampering binding though intermolecular interactions. Our results show that also the underlying molecular structure of the GSL itself can have a profound effect on the presentation of the glycan epitope to a virus. This can lead to a situation where an epitope shows binding when the GSL (or its head group) is free in solution and not binding when the GSL is embedded in a membrane because it is not accessible. Our results clearly demonstrate that performing binding studies based on methods that use free molecules (e.g. STD NMR or X-ray crystallography) or where the binding epitope is chemically bonded in an artificial environment (e.g. glycan array) may lead to results that may not be transferable to a biological system. GII.4 NVLPs preferentially recognize type-1 chain HBGAs on membrane surface The first report stating that GII.4 norovirus-HBGA binding kinetics differ between type-1 and type-2 chain HBGA ABO(H) oligosaccharides, observed through ELISA and SPR studies, came in 200824. For r104 GII.4 noroviruses, more stable interactions were observed for type1 chain as compared to type-2 chain structures. Following the same trend for difucosylated



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 35

HBGA structures, we observed that Le-y-containing vesicles exhibited slower rate of association and faster release in comparison with the ones containing Le-b. The preference of alpha fucosylated type-1 over type-2 chain HBGAs in GII.4 VLP binding gives insights into the tissue tropism of noroviruses since the GSL expression of type-1 but not type-2 HBGAs correlate with the binding of recombinant NV VLPs to the gastroduodenal junction from human intestinal biopsies6. Furthermore, this preference also correlates with the tissue expression and substrate specificity of the α1,2-fucosyltransferase, encoded by FUT2 gene, which preferentially adds α-L-Fucp residues on type-1 (and type-3) chain precursors rather than on type-2 chain precursors25,

26

. Individuals carrying the non-secretor pheno- and

genotypes are thus presumed to be, even if not totally resistant to all strains, less susceptible to GII.4 NV infections7, 27, 28. The successful replication of several human noroviruses in stem cell-derived HIEs was reported recently9. In agreement with epidemiological data, the strict FUT2 gene dependence was observed for GII.4, but not GII.3 NV, strains. The epithelium of human small intestine is rich in fucosylated HBGAs, which for GSLs29 are mainly of the type-1 chain10 but for Nglycoproteins are mainly of the type-2 chain11. Thus our experimental and simulation data predict the role of type-1 GSLs as potential receptors for norovirus uptake in intestinal enterocytes. Glycosphingolipids have previously been shown to facilitate the virus/bacterial uptake through spontaneously formed membrane invaginations. For instance, it was reported that polyoma viruses, e.g. SV40, as well as bacterial toxins sharing pentameric protein scaffolds interact with glycolipids thereby inducing membrane invaginations30. For SV40 it has been demonstrated that GM1 molecules with longer hydrocarbon chains are needed to induce such invaginations31. Norovirus VLPs, also sharing a similar multimeric protein scaffold, have also been shown to produce such invaginations in giant unilamellar vesicles exposing native GSLs


Page 13 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


with relevant HBGAs and lipid compositions32. However, the final proof of GSLs being NV receptors will have to await experimental results using intestinal target cells and live virions. In conclusion, both the experimental and structural analyses based on membrane MD simulation data demonstrate that the structure and the presentation of the glycan epitope are critical in defining the role of membrane bound HBGAs as NV attachment factors or potential receptors. The reported data should be of more general relevance to understand GSL-protein and even GSL-GSL interactions33. In future studies it should be possible to use our experimental and computational set-up to study the presentation effects in more complex membrane systems34.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 35

Materials and methods Glycosphingolipids (GSLs) The GSLs used in this study included H type 1 (H-1), A type-1 (A-1), B type-1 (B-1), Lewis a (Le-a), Lewis b (Le-b), A Lewis b (ALe-b), Lewis x (Le-x), Lewis y (Le-y) and A Lewis y (ALe-y) (Table 1, Figure 1). The structurally characterized GSLs were purified from meconium of single individuals or from pooled meconium samples, from single dog intestines or pooled red blood cells of individuals with same ABO blood group35-37. Virus like particles (VLPs) The Spanish norovirus isolate Ast6139/01/sp38 was used to produce the VLPs using a baculovirus-expression system as described in detail elsewhere17. The glycan binding pattern of this strain has earlier been thoroughly investigated using STD-NMR17. Thin layer chromatography and VLP chromatogram binding assay Purified fractions of GSLs (2µg) were applied on aluminum backed silica gel 60 HPTLC plates (Merck, Germany) and chromatographed using chloroform:methanol:water (60:35:8 by volume). The plates were then dried and cut in half, with both halves containing the same set of GSLs. Anisaldehyde:sulphuric acid:acetic acid (97:1:2, by vol.) was sprayed on one of the sections followed by heating at 180 oC for 1-3 minutes to visualize GSLs. The other plate was plasticized using P28 polyisobutylmethacrylate (0.5 % in diethylether) by dipping it in solution for 60 seconds followed by drying20,

21

. The plate was left overnight at room

temperature and then immersed in blocking buffer (3% BSA, 0.05 % TWEEN-20, PBS pH 7.2) for an hour. The VLPs were then dissolved in the dilution buffer (0.5 % BSA, 0.05 % TWEEN-20, PBS pH 7.2), added onto the surface of the plate and incubated for 2 hours at final concentration of 2.5 µg/mL. The plate was washed 3 times in PBS with 0.05 %


Page 15 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


TWEEN-20 by gentle swirling. This was then followed by incubation with rabbit anti-NV antibody (a kind gift from Prof. Lennart Svensson, Linköping, Sweden) (1:500 in dilution buffer) for 1.5 hours at room temperature. Afterwards the plate was gently washed three times. Finally, the rabbit alkaline phosphatase conjugated antibody (1:500, Sigma Aldrich) was applied on the plate for 1 hour in dilution buffer at room temperature. After washing 3 times, immune-staining was performed using Sigma Fast BCIP/NBT tablet dissolved in water (10 mL). TIRFM binding assay Flat glass-bottom microtiter wells (96 well-plate) were used for the binding experiments. The chambers were cleaned with 3% Hellmanex III (by vol., Helma Analytics) in ionized water for at least 2 hours followed by washing with 300 mL of ionized water at least 7 times. PBS buffer (10 mM PBS, 1 mM CaCl2, pH 7.2) was used for the assay. Lipid bilayers were spontaneously formed on the glass surface from 100 nm vesicles containing 5 % (by weight) H type 1 GSL in POPC (50 µL, 0.1 mg/mL) incubated for 25 minutes. The chamber was then washed at least seven times with 200 µL of buffer without drying the surface. The wells were then incubated with Ast6139 GII.4 norovirus VLPs (10 µg/mL) for at least 20 minutes to establish VLP binding to GSLs in the supported lipid bilayer (SLB). After rinsing the VLPs in solution, POPC vesicles (100 nm; 50 µg/mL) were added to the well for 15-20 minutes to minimize the presence of defects in the lipid bilayer. Finally, fluorescently labeled vesicles were injected at the final concentration of 250 ng/mL. The total volume of the well was 100 µL at the time of measurement. The attachment and detachment events, recorded as images, were gathered from the TIRF microscope at least 30 minutes after the addition of fluorescent vesicles (Figure 2). Molecular Dynamics Simulations



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 35

Since, at present, it is not technically feasible to simulate a glycolipid membrane including the NV capsid in explicit solvent over a sufficiently long timescale two alternative strategies have been employed: The accessible conformational space (and potential restrictions) of the glycolipids embedded in a membrane has been explored by membrane simulations on the hundred ns timescale and the fraction of the α-L-Fucp epitopes suitable for binding of the VLP without resulting in clashes with the membrane has been analyzed (capsid/membrane clashing analysis). Independently, and in order to explore the accessible conformational space of the free HBGAs more completely, ten times longer trajectories (µs timescale) were calculated and used to check which fraction adopts a conformation that is suitable to ‘virtually link’ the α-L-Fucp residue located in the binding site of the VLP with a ceramide molecule still embedded in a membrane (capsid/membrane linking analysis). Molecular dynamics simulations were performed in explicit solvent using YASARA39, 40. The AMBER0341 force field was chosen in YASARA which also uses GLYCAM42 parameters for the carbohydrates and assigns GAFF parameters for the lipid residues43. Pressure control was performed by scaling the periodic simulation box in order to keep the water density at 0.997 g/L. Long-range Coulomb interactions are calculated using the Particle Mesh Ewald (PME) algorithm 44. Embedding of glycolipids into a POPE membrane, solvation, adding of ions and performing the membrane MD simulation was done using YANACONDA macros which are part of the YASARA package. The sampling time for the MDs of isolated HBGAs in explicit solvent was 5 microseconds. Snapshots were recorded every 10 picoseconds. For the membrane simulations snapshots were saved every 100 ps and 400 ns were sampled for H-1, 600 ns for Le-y, 460 ns for ALe-y, 200 ns for ALe-b and 600 ns for Le-x. Analysis of the trajectories, ‘capsid/membrane clashing’ and ‘capsid/membrane linking’ analyses (described in detail in supporting information) as well as scientific plotting was performed using Conformational Analysis Tools (www.md-simulations.de/CAT/). VMD was


Page 17 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


used for display of molecular structures45. Further details of the simulations are given in SI Supplementary Methods.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 35

Supporting Information Description of methods for “Lipid vesicles preparation”, “Image acquisition and TIRFM analysis”, “Capsid/membrane clashing analysis” and “Capsid/membrane linking analysis” together with supporting figures.

Acknowledgements This work was supported by grants from the Swedish Research Council (2013-8266 to G.L., 2014-5557 to F.H., 2012-5024 to M.B.), VINNOVA, Swedish Foundation for Strategic Research (KF10-0088 to G.L., F.H. [N. Lycke, principle investigator]) and Governmental Grants to the Sahlgrenska University Hospital (G.L.). Work at F.P. lab has been supported by grant GRUPIN14-099 from Principado de Asturias (Spain). We would like to dedicate this study to Professors K.-A. Karlsson and I. Pascher, University of Gothenburg, for their pioneering work on sphingolipid structures and membrane biology.


Page 19 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


References [1] de Graaf, M., van Beek, J., and Koopmans, M. P. (2016) Human norovirus transmission and evolution in a changing world, Nat. Rev. Microbiol. 14, 421-433. [2] David, R. (2013) Norovirus strikes back, Nat. Rev. Microbiol. 11, 70-70. [3] Hoa Tran, T. N., Trainor, E., Nakagomi, T., Cunliffe, N. A., and Nakagomi, O. (2013) Molecular epidemiology of noroviruses associated with acute sporadic gastroenteritis in children: global distribution of genogroups, genotypes and GII.4 variants, J. Clin. Virol. 56, 185-193. [4] Rydell, G. E., Dahlin, A. B., Hook, F., and Larson, G. (2009) QCM-D studies of human norovirus VLPs binding to glycosphingolipids in supported lipid bilayers reveal strain-specific characteristics, Glycobiology 19, 1176-1184. [5] Huang, P., Farkas, T., Zhong, W., Tan, M., Thornton, S., Morrow, A. L., and Jiang, X. (2005) Norovirus and histo-blood group antigens: demonstration of a wide spectrum of strain specificities and classification of two major binding groups among multiple binding patterns, J. Virol. 79, 6714-6722. [6] Marionneau, S., Ruvoen, N., Le Moullac-Vaidye, B., Clement, M., Cailleau-Thomas, A., RuizPalacois, G., Huang, P., Jiang, X., and Le Pendu, J. (2002) Norwalk virus binds to histo-blood group antigens present on gastroduodenal epithelial cells of secretor individuals, Gastroenterology 122, 1967-1977. [7] Rydell, G. E., Kindberg, E., Larson, G., and Svensson, L. (2011) Susceptibility to winter vomiting disease: a sweet matter, Rev. Med. Virol. 21, 370-382. [8] Thorven, M., Grahn, A., Hedlund, K.-O., Johansson, H., Wahlfrid, C., Larson, G., and Svensson, L. (2005) A homozygous nonsense mutation (428G→ A) in the human secretor (FUT2) gene provides resistance to symptomatic norovirus (GGII) infections, Journal of virology 79, 15351-15355. [9] Ettayebi, K., Crawford, S. E., Murakami, K., Broughman, J. R., Karandikar, U., Tenge, V. R., Neill, F. H., Blutt, S. E., Zeng, X.-L., and Qu, L. (2016) Replication of human noroviruses in stem cell–derived human enteroids, Science, aaf5211.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 35

[10] Breimer, M. E. (1984) Tissue specificity of glycosphingolipids as expressed in pancreas and small intestine of blood group A and B human individuals, Arch. Biochem. Biophys. 228, 7185. [11] Finne, J., Breimer, M., Hansson, G., Karlsson, K., Leffler, H., Vliegenthart, J., and Van Halbeek, H. (1989) Novel polyfucosylated N-linked glycopeptides with blood group A, H, X, and Y determinants from human small intestinal epithelial cells, J. Biol. Chem. 264, 5720-5735. [12] Nyholm, P. G., and Pascher, I. (1993) Orientation of the saccharide chains of glycolipids at the membrane surface: conformational analysis of the glucose-ceramide and the glucose-glyceride linkages using molecular mechanics (MM3), Biochemistry 32, 1225-1234. [13] Nyholm, P. G., Samuelsson, B. E., Breimer, M., and Pascher, I. (1989) Conformational analysis of blood group A-active glycosphingolipids using HSEA-calculations. The possible significance of the core oligosaccharide chain for the presentation and recognition of the Adeterminant, J. Mol. Recognit. 2, 103-113. [14] Strömberg, N., Nyholm, P.-G., Pascher, I., and Normark, S. (1991) Saccharide orientation at the cell surface affects glycolipid receptor function, Proc. Natl. Acad. Sci. U. S. A. 88, 9340-9344. [15] Saito, S., Levery, S. B., Salyan, M. E., Goldberg, R. I., and Hakomori, S. (1994) Common tetrasaccharide epitope NeuAc alpha 2-->3Gal beta 1-->3(Neu-Ac alpha 2-->6)GalNAc, presented by different carrier glycosylceramides or O-linked peptides, is recognized by different antibodies and ligands having distinct specificities, J. Biol. Chem. 269, 5644-5652. [16] Lingwood, D., Binnington, B., Rog, T., Vattulainen, I., Grzybek, M., Coskun, U., Lingwood, C. A., and Simons, K. (2011) Cholesterol modulates glycolipid conformation and receptor activity, Nat. Chem. Biol. 7, 260-262. [17] Fiege, B., Rademacher, C., Cartmell, J., Kitov, P. I., Parra, F., and Peters, T. (2012) Molecular details of the recognition of blood group antigens by a human norovirus as determined by STD NMR spectroscopy, Angew. Chem. Int. Ed. Engl. 51, 928-932. [18] Bally, M., Gunnarsson, A., Svensson, L., Larson, G., Zhdanov, V. P., and Hook, F. (2011) Interaction of single viruslike particles with vesicles containing glycosphingolipids, Phys. Rev. Lett. 107, 188103.


Page 21 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


[19] Nasir, W., Bally, M., Zhdanov, V. P., Larson, G., and Hook, F. (2015) Interaction of Virus-Like Particles with Vesicles Containing Glycolipids: Kinetics of Detachment, J. Phys. Chem. B 119, 11466-11472. [20] Karlsson, K. A., and Stromberg, N. (1987) Overlay and solid-phase analysis of glycolipid receptors for bacteria and viruses, Methods Enzymol. 138, 220-232. [21] Hansson, G. C., Karlsson, K.-A., Larson, G., Strömberg, N., Thurin, J., Örvell, C., and Norrby, E. (1984) A novel approach to the study of glycolipid receptors for viruses, FEBS Lett. 170, 1518. [22] Demarco, M. L., Woods, R. J., Prestegard, J. H., and Tian, F. (2010) Presentation of membraneanchored glycosphingolipids determined from molecular dynamics simulations and NMR paramagnetic relaxation rate enhancement, J. Am. Chem. Soc. 132, 1334-1338. [23] Chen, Y., Tan, M., Xia, M., Hao, N., Zhang, X. C., Huang, P., Jiang, X., Li, X., and Rao, Z. (2011) Crystallography of a Lewis-binding norovirus, elucidation of strain-specificity to the polymorphic human histo-blood group antigens, PLoS Pathog. 7, e1002152. [24] Shirato, H., Ogawa, S., Ito, H., Sato, T., Kameyama, A., Narimatsu, H., Xiaofan, Z., Miyamura, T., Wakita, T., Ishii, K., and Takeda, N. (2008) Noroviruses distinguish between type 1 and type 2 histo-blood group antigens for binding, J. Virol. 82, 10756-10767. [25] Oriol, R., Pendu, J., and Mollicone, R. (1986) Genetics of ABO, H, Lewis, X and related antigens, Vox sang. 51, 161-171. [26] Lowe, J. B. (1993) 7 The blood group-specific human glycosyltransferases, Baillieres Clin. Haematol. 6, 465-492. [27] Kindberg, E., Akerlind, B., Johnsen, C., Knudsen, J. D., Heltberg, O., Larson, G., Bottiger, B., and Svensson, L. (2007) Host genetic resistance to symptomatic norovirus (GGII.4) infections in Denmark, J. Clin. Microbiol. 45, 2720-2722. [28] Larsson, M. M., Rydell, G. E., Grahn, A., Rodriguez-Diaz, J., Akerlind, B., Hutson, A. M., Estes, M. K., Larson, G., and Svensson, L. (2006) Antibody prevalence and titer to norovirus (genogroup II) correlate with secretor (FUT2) but not with ABO phenotype or Lewis (FUT3) genotype, J. Infect. Dis. 194, 1422-1427.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 35

[29] Björk, S., Breimer, M., Hansson, G. C., Karlsson, K.-A., and Leffler, H. (1987) Structures of blood group glycosphingolipids of human small intestine. A relation between the expression of fucolipids of epithelial cells and the ABO, Le and Se phenotype of the donor, J. Biol. Chem. 262, 6758-6765. [30] Römer, W., Berland, L., Chambon, V., Gaus, K., Windschiegl, B., Tenza, D., Aly, M. R., Fraisier, V., Florent, J.-C., and Perrais, D. (2007) Shiga toxin induces tubular membrane invaginations for its uptake into cells, Nature 450, 670-675. [31] Ewers, H., Romer, W., Smith, A. E., Bacia, K., Dmitrieff, S., Chai, W., Mancini, R., Kartenbeck, J., Chambon, V., Berland, L., Oppenheim, A., Schwarzmann, G., Feizi, T., Schwille, P., Sens, P., Helenius, A., and Johannes, L. (2010) GM1 structure determines SV40-induced membrane invagination and infection, Nat. Cell. Biol. 12, 11-18; sup pp 11-12. [32] Rydell, G. E., Svensson, L., Larson, G., Johannes, L., and Romer, W. (2013) Human GII.4 norovirus VLP induces membrane invaginations on giant unilamellar vesicles containing secretor gene dependent alpha1,2-fucosylated glycosphingolipids, Biochim. Biophys. Acta. 1828, 1840-1845. [33] Kunze, A., Bally, M., Hook, F., and Larson, G. (2013) Equilibrium-fluctuation-analysis of single liposome binding events reveals how cholesterol and Ca2+ modulate glycosphingolipid transinteractions, Sci. Rep. 3, 1452. [34] Pace, H., Simonsson Nyström, L., Gunnarsson, A., Eck, E., Monson, C., Geschwindner, S., Snijder, A., and Höök, F. (2015) Preserved transmembrane protein mobility in polymersupported lipid bilayers derived from cell membranes, Analytical chemistry 87, 9194-9203. [35] Karlsson, K. A., and Larson, G. (1981) Molecular characterization of cell surface antigens of fetal tissue. Detailed analysis of glycosphingolipids of meconium of a human O Le(a--b+) secretor, J. Biol. Chem. 256, 3512-3524. [36] Karlsson, K. A. (1987) Preparation of total nonacid glycolipids for overlay analysis of receptors for bacteria and viruses and for other studies, Methods Enzymol. 138, 212-220.


Page 23 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


[37] McKibbin, J. M., Spencer, W. A., Smith, E. L., Mansson, J. E., Karlsson, K. A., Samuelsson, B. E., Li, Y. T., and Li, S. C. (1982) Lewis blood group fucolipids and their isomers from human and canine intestine, J. Biol. Chem. 257, 755-760. [38] Ng, K. K., Pendas-Franco, N., Rojo, J., Boga, J. A., Machin, A., Alonso, J. M., and Parra, F. (2004) Crystal structure of norwalk virus polymerase reveals the carboxyl terminus in the active site cleft, J. Biol. Chem. 279, 16638-16645. [39] Krieger, E., and Vriend, G. (2014) YASARA View - molecular graphics for all devices - from smartphones to workstations, Bioinformatics 30, 2981-2982. [40] Krieger, E., and Vriend, G. (2015) New ways to boost molecular dynamics simulations, J. Comput. Chem. 36, 996-1007. [41] Duan, Y., Wu, C., Chowdhury, S., Lee, M. C., Xiong, G., Zhang, W., Yang, R., Cieplak, P., Luo, R., Lee, T., Caldwell, J., Wang, J., and Kollman, P. (2003) A point-charge force field for molecular mechanics simulations of proteins based on condensed-phase quantum mechanical calculations, J. Comput. Chem. 24, 1999-2012. [42] Woods, R. J., Dwek, R. A., Edge, C. J., and Fraser-Reid, B. (1995) Molecular mechanical and molecular dynamic simulations of glycoproteins and oligosaccharides. 1. GLYCAM_93 parameter development, J. Phys. Chem. 99, 3832-3846. [43] Jakalian, A., Jack, D. B., and Bayly, C. I. (2002) Fast, efficient generation of high-quality atomic charges. AM1-BCC model: II. Parameterization and validation, J. Comput. Chem. 23, 16231641. [44] Essmann, U., Perera, L., Berkowitz, M. L., Darden, T., Lee, H., and Pedersen, L. G. (1995) A smooth particle mesh Ewald method, J. Chem. Phys. 103, 8577-8593. [45] Hsin, J., Arkhipov, A., Yin, Y., Stone, J. E., and Schulten, K. (2008) Using VMD: an introductory tutorial, Curr. Protoc. Bioinformatics Chapter 5, Unit 5 7.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 35

Table 1 Glycosphingolipids used in this study.

Name

Abbreviation

Structural formulae

H type 1-Cer

H-1

Fucα2Galβ3GlcNAcβ3Galβ4Glcβ-Cer

B type 1-Cer

B-1

Galα3(Fucα2)Galβ3GlcNAcβ3Galβ4Glcβ-Cer

A type 1-Cer

A-1

GalNAcα3(Fucα2)Galβ3GlcNAcβ3Galβ4Glcβ-Cer

Lewis a-Cer

Le-a

Galβ3(Fucα4)GlcNAcβ3Galβ4Glcβ-Cer

Lewis b-Cer

Le-b

Fucα2Galβ3(Fucα4)GlcNAcβ3Galβ4Glcβ-Cer

A Lewis b-Cer

ALe-b

Neo lactotetra-Cer

nLC4

Lewis x-Cer

Le-x

Galβ4(Fucα3)GlcNAcβ3Galβ4Glcβ-Cer

Lewis y-Cer

Le-y

Fucα2Galβ4(Fucα3)GlcNAcβ3Galβ4Glcβ-Cer

A Lewis y-Cer

ALe-y

GalNAcα3 (Fucα2)Galβ3(Fucα4)GlcNAcβ3Galβ4Glcβ-Cer

Galβ4GlcNAcβ3Galβ4Glcβ-Cer

GalNAcα3(Fucα2)Galβ4(Fucα3)GlcNAcβ3Galβ4Glcβ-Cer


Page 25 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Table 2 Reaction rate constant (kon) values for the vesicles containing GSLs considered in the present study. The coverage of the VLPs for calculations is estimated to be 10 VLPs per µm-2 based on the experiments performed using quartz crystal microbalance with dissipation monitoring19. The kon values for both time windows with 3 or 30 seconds of minimum residence time (τmin) are shown separately.

Ligands

kon (M-1s-1) (τmin=3 sec)

kon (M-1s-1) (τmin=30 sec)

H-1

(3.79±1.2) X 106

(0.79±0.15) X 106

B-1

(2.61±0.6) X 106

(0.36±0.03) X 106

A-1

(1.91±0.5) X 106

(0.41±0.04) X 106

Le-a

(3.52±0.5) X 106

(0.55±0.02) X 106

Le-b

(3.65±0.2) X 106

(1.3±0.1) X 106

ALe-b

(1.47±0.3) X 106

(0.54±0.03) X 106

Le-x

(0.25±0.3) X 106

(0.03±0.0) X 106

Le-y

(3.07±0.3) X 106

(0.51±0.03) X 106

ALe-y

(.08±0.01) X 106

(0.01±0.0) X 106



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 35

Figure legends Figure 1 Structures of type-1 (ALe-b; left) and type-2 (ALe-y; right) chain HBGAs of GSLs. All α-LFucp residues are shown in different colors based on types and linkages. An imaginary membrane plane (passing through OL atom of the Glc residue) is shown as a dotted line. The rotation of membrane bound GSLs is indicated by circled arrows whereas straight arrows show movement in xz plane and bent arrows indicate movements in the internal torsion angles of the glycosidic linkages giving the conformational space of the glycans. Figure 2 Principle of TIRFM based binding assay. The fluorescently labeled vesicles, which contain the GSLs of interest, are used to detect the vesicle-VLP interactions (a). The (1,2)-linked α-LFucp of ALe-b GSL was superimposed onto the (1,2)-linked α-L-Fucp of the B-tri saccharide into the GII.4 VA387 VLP modeled as described in supplementary methods with domains P1 red, P2 blue and shell yellow (b). Bound VLPs appear as spots in TIRF images (c) where each spot represents vesicle-VLP complex. The association and dissociation kinetics are recorded based on the residence time analysis (d). Figure 3 The association (a), dissociation (b) and mean vesicle coverage (c) analyses of GSLcontaining vesicles binding to the NVLPs attached to H-1 GSL in POPC bilayers summarized for two minimum residence times (τmin) of 3 (left) and 30 seconds (right) respectively. Vesicles with Le-b or H-1 GSL demonstrate the fastest attachment kinetics whereas, the coverage analysis and detachments kinetics data shows that the vesicles containing ALe-b or


Page 27 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


B-1 GSL most strongly bind the Ast6139 VLPs. No attachment was observed for Le-x or ALe-y GSL vesicles. Figure 4 TLC-CBA of Ast6139 VLPs binding to GSLs. Left panel is stained with anisaldehyde reagent to visualize all individual GSLs. The right panel shows the Ast6139 VLPs binding activity to specific GSLs, in complete agreement with the TIRFM data on binding kinetics. Figure 5 Summary of simulation methodology and results from “capsid/membrane clashing analysis”. In a) the membrane is shown in cyan with the GSL in ball representation (α-L-Fucp shown in yellow). Part of VLP is shown as vdW spheres (shades of blue). Magenta dotted line represents the membrane plane which passes through the OL atom of the Glc residue. In b) the analysis of the parameters β (angle between C4-O4 bond vector in α-L-Fucp and the yaxis) and d (distance of α-L-Fucp from membrane surface) is shown for the 400 ns membrane MD simulation of H-1 GSL. In c) the left hand side plot shows the lack of correlation between the angles α (angle between OL-C1 bond vector in Glc and the y-axis) and β, i.e. the movement of α-L-Fucp plane and the orientation of ceramide-saccharide linkage. The right hand side plot shows the results from capsid/membrane clashing analysis for individual GSLs. The specific α-L-Fucp residue used to produce the fit for VLP-glycan complex (binding site α-L-Fucp) is written in parenthesis. The bars represent the percentage of non-clashing poses during the membrane simulation. Figure 6 Summary of simulation methodology and results from “Capsid/membrane linking analysis”. In a) the virtual lipid is shown as cyan sticks and the magenta dotted line represents the



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 35

membrane plane which passes through the OL atom of the Glc residue. In both a) and b) the OL atoms, taken from the H-1 MD simulation, are shown as grey dots. In the right hand side plot, in b), the majority of the data points can be seen to have a distance of at least 7 Å between the α-L-Fucp plane and the OL atom of Glc residue in the GSL. In c) the accessible range of

Histo-Blood Group Antigen Presentation Is Critical ... - ACS Publications

Recommend Documents