VIRUS-ANTIBODY INTERACTIONS VISUALISED BY COMBINING CRYO-ELECTRON MICROSCOPY AND X-RAY DATA

Elizabeth Hewat

Institut de Biologie Structurale Jean-Pierre Ebel, 41 av. des Martyrs, 38027 Grenoble, France hewat@ibs.fr

Keywords: cryo-electron microscopy, X-ray crystallography, virus/antibody complexes

Cryo-electron microscopy has been very useful in the study of the interactions of monoclonal antibodies (MAbs) with several viruses [e.g. 1, 2, 3]. These structural studies provide information not only about the neutralisation of viruses by MAbs but also about the interactions and properties of MAbs in general, their flexibility and modes of attachment. From the low resolution structure of a virus/antibody complex one can determine if the antibody is bound to the virus monovalently or bivalently; the occupancy of the different binding sites; steric effects on binding; whether the receptor binding site is obstructed; if the virus has under gone large scale structural modifications, etc. For example, we have recently shown that the spanning distance for bivalent MAb attachment can be as little as 60Å [2]; and in a study of rabbit haemorrhagic disease virus (RHDV) complexed with a strongly neutralising MAb it was found that each MAb may be attached to any pair of the three neighbouring binding sites none of which are related by a 2-fold axis [4].

These studies of virus/antibody complexes have been greatly enhanced by combining the results from cryo-electron microscopy and X-ray crystallography which give complementary data on macromolecular complexes. The low resolution structure of the complex determined by cryo-electron microscopy (20-30Å) has been used to position the atomic resolution X-ray structures of the constituent molecules with a precision of about 4Å, thus giving a first approximation to the structure of the interface. It will be recalled that to determine a protein structure by X-ray crystallography to resolution of say 2Å, the known structures of amino acid residues are placed in the electron density map with a precision of better than 0.1 Å. By analogy the known structures of the molecular components may be placed in a 20 to 30Å map of the complex with a precision of the order of 4Å. In general to date a rigid body fit of the component molecules has been performed. However, the recent improvements in resolution (to approx. 7Å) obtained using cryo-electron microscopes fitted with a field emission gun (FEG) and more powerful computers for averaging thousands of particle images, will allow a relaxation of this constraint and a better model of the interfaces to be determined.

We have employed cryo-electron microscopy and 3-D reconstruction techniques combined with X-ray crystallographic data to study the structure of several virus/antibody complexes including RHDV, rotavirus and two picornaviruses, namely human rhinovirus serotype 2 (HRV2) and foot-and-mouth disease virus type C (FMDV C) . In particular, HRV2 complexed with two weakly neutralising MAbs 8F5 [2] and 3B10 [5], and foot and mouth disease virus type C (FMDV C) complexed with the Fabs of strongly neutralising MAbs SD6 [3] and 4C4 [6]. Three of these MAbs (8F5, SD6 and 4C4) bind not only to the virus, but also to a synthetic peptide corresponding to the contiguous epitope on the virus.

Human rhinoviruses, the major cause of the common cold, exhibit vast antigenic variation with over 100 serotypes currently identified. In the case of HRV2, the MAb 8F5 is seen to be bound bivalently across the icosahedral 2-fold axis despite the very short distance of 60Å between the symmetry-related epitopes. The footprints of 8F5 and 3B10 are very similar. They are largely on VP2 but also cover a viral protein 3 (VP3) loop. However, the Fab 3B10 is bound in an orientation, inclined at approximately 45° to the surface of the virus capsid, which is only compatible with monovalent binding of the antibody. The canyon around the 5-fold axis (the putative receptor binding site) is not directly obstructed by the bound Fabs in either complex.

Although FMDV is closely related structurally to HRVs the surface is comparatively smooth with only one long loop, the G-H loop, projecting from viral protein 1 (VP1). This highly flexible loop is not resolved in the X-ray structure of the native FMDVs and contains the RGD sequence required for binding to the cellular receptor molecule, an integrin. It is of considerable immunogenic importance since the binding sites for neutralising antibodies are generally located on loops which decorate the viral surface. The MAbs SD6 and 4C4 bind monovalently to this G-H loop with comparable affinity and have similar neutralising efficiency. It is remarkable that, MAb SD6 exhibits 100% occupancy and a well defined position while the 4C4 Fab appears to interact almost exclusively with the G-H loop of VP1 making no other contacts with the viral capsid, thus the loop retains it flexibility. These remarkable differences in binding do not apparently affect the neutralising efficiency of these antibodies. We conclude that steric inhibition of the receptor binding site is the major mechanism by which both antibodies neutralise FMDV infectivity.

From these results it appears that a MAb which binds strongly to the receptor binding site will be a strong neutraliser. The cooperativity of bivalent mAb binding can give an affinity constant of 100-1000 higher than for the Fab alone, however, bivalent binding is neither necessary (e.g. SD6) nor sufficient (e.g. 8F5) to ensure strong neutralisation. It is the affinity (or avidity) of the antibody which matters.

1. Smith TJ, Olson NH, Cheng RH, Chase ES and Baker TS (1993) PNAS, 90, 7015-7018.
2. Hewat EA and Blaas D (1996) EMBO J, 15,1515-1523)
3. Hewat, E.A., Verdaguer, N., Fita, I., Blakemore, W., Brookes, S., King, A., Newman, J., Domingo, E., Mateu, M.G. and Stuart, D.,EMBO J. (1997) 16, 1492-1500
4. Thouvenin, E., Laurent,S., Madelaine, M-F., Rasschaert, D., Vautherot, J-F. and Hewat, E.A., J Mol. Biol. (1997) 270, 238-246.
5. Hewat E. A., Marlovitz T. and Blaas D. (1998) J. Virol., (1998) 72, in press
6. Verdaguer, N., Schoehn, G., Ochoa, W.F., Fita I., Brookes, S., King, A., Domingo, E., Mateu, M.G., Stuart, D.I. and Hewat, E.A., submitted