Unit: Structural Immunology
Director: BENTLEY Graham
Our research objectives focus on the structural biology of proteins from pathogenic agents. Although structure determination by X-ray crystallography is the principal activity of the unit, we complement our structural studies with other techniques, such as immunochemistry and surface plasmon resonance, to increase our understanding of structure-function relationships. Some of our projects concern proteins implicated in the development of vaccines and medicaments. Much of our interest centres on surface antigens of the malaria parasite Plasmodium that are vaccine candidates. A more detailed account of our projects is available on our web site: www.pasteur.fr/recherche/unites/ImmStr/accueil.html
Apical Membrane Antigen 1: (G. Bentley, M.L. Chesne-Seck, G. Faure, S. Igonet, F. Saul, B. Vulliez-Le Normand in collaboration with A. Thomas, Biomedial Primate Research Centre, Rijswijk, The Netherlands, and M. Blackman, National Institute of Medical Research, Mill Hill, U.K.)
Apical Membrane Antigen 1 (AMA1) is a type I transmembrane protein that plays an essential although poorly characterized role in host cell invasion. Nonetheless, experimental data are consistent with the protein having a receptor-binding function. In particular, antibodies raised against the ectoplasmic region of AMA1 can inhibit parasite invasion in animal model systems of malaria, and this part of the antigen is a vaccine candidate currently in clinical trials.
The crystal structure of the ectoplasmic region of AMA1 from the species P. vivax has been solved and refined at 1.8 Å resolution. Although structural folds had not been previously attributed to the AMA1 domains, the crystal structure shows that both Domains I and II are based on the PAN motif (Plasminogen, Apple, Nematode). PAN modules are often found in proteins with diverse adhesion functions, binding to protein or carbohydrate receptors. The structure of PvAMA1 is thus consistent with the putative receptor-binding role of the protein.
A model of the AMA1 structure from P. falicparum was derived by homology in order to examine the 3-dimensional distribution of polymorphic residues in this species. Polymorphic sites in PfAMA1 are distributed over Domains I, II and III, with Domain I containing an important fraction of the polymorphism. The distribution, however, is highly biased to one side of the molecule. This suggests a preferential exposure of this side of the antigen to the exterior on the parasite surface and/or that functional constraints are acting on the less polymorphic face of the protein. Polymorphic data on PvAMA1 is essentially restricted to Domain I, but a trend similar to that in PfAMA1 is observed.
The epitope of the invasion-inhibitory mAb 4G2, which is specific for PfAMA1, was mapped by mutagenesis studies (M. Blackman, N.I.M.R., Mill Hill, U.K.). The region recognised by this mAb was localised to the base of a 40-residue loop on Domain II of the PfAMA1 model, which is disordered in the PvAMA1 crystal structure. Interestingly, the Domain II loop does not contain any known PfAMA1 polymorphic sites and has also been shown to carry a Th-cell epitope in the human immune response; this could have important consequences in optimising AMA1 as a vaccine candidate.
Crystals of a complex formed between PvAMA1 and the Fab fragment from the specific monoclonal antibody, F8.12.19, have been obtained and the structure has been solved and refined to 2.6 Å resolution. The epitope is localised on Domain III and the conserved nature of the interactions between antibody and antigen that are found in the crystal structure explain the observed cross-reactivity of mAb F8.12.19 with schizonts of P. falciparum, P. berghei, P. knowlesi and P. cynomolgi: antibody contacts are almost entirely to main-chain atoms and conserved side chains of AMA1. Kinetic studies of PvAMA1 and PfAMA1 binding to mAb F8.12.19 were made by surface plasmon resonance measurements, giving an apparent dissociation constant, KD (koff/kon) of 1 nM for PvAMA1 and 100 nM for PfAMA1.
P. falciparum Erythrocyte Membrane Protein 1 (PfEMP1): (C. Badaut, G.A. Bentley, S. Igonet, Hélène Souchon in collaboration with O. Puijalon, Immunologie Moléculaire des Parasites, I.P., M. Klinkert, Bernhard Nocht Institute, Hamburg, P. Deloron, Institut de Recherche pour le Développement, Université Paris V)
Inside the infected erythrocyte, P. falciparum expresses PfEMP1, a virulence factor that is transported to, and presented on, the surface of the host cell. PfEMP1 variants form family of adhesins that confer upon infected erythrocytes the capacity to auto-agglutinate, to adhere to non-infected red blood cells (rosetting) or to sequester in the vascular beds of divers tissues (sequestration). These adhesion phenomena can lead to severe malaria. PfEMP1 is encoded by the var (variable) gene family, present in about 60 copies (depending on the parasite strain); however, only one variant is presented on the erythrocyte surface at any given time, thus giving different parasite isolates their particular receptor specificity for agglutination, rosetting or sequestration. PfEMP1 molecules are organised in a modular fashion, with the extra-cellular region composed mainly of DBL (Duffy-Binding Like) and CIDR (cysteine-rich Interdomain region) domains belonging to different classes. We are studying the structural and functional properties of different variants of PfEMP1.
Our interest centres on two variants of PfEMP1, one implicated in rosetting (collaboration with O. Puijalon) and the other in placental malaria (with M. Klinkert and P. Deloron). Because PfEMP1 are large proteins (250-350 KDa), only individual domains of these molecules are expressed for structural and functional studies.
Placental malaria arises from sequestration of infected erythrocytes in the placenta. Parasite isolates causing placental malaria express PfEMP1 molecules that show binding-specificity to chondroitin sulphate A (CSA) present on certain proteoglycans, such as CSPG and thrombomodulin, in this tissue. CSA specificity appears to be due, in part, to a subgroup of the domains DBL class γ. We have expressed the DBL3γ domain of the 732 placental isolate in large quantity using the insect cell/baculovirus system. This domain is specifically recognised by CSA on CSPG obtained from placental preparations, and antibodies raised against the recombinant protein can block binding of infected erythrocytes to CSA. Moreover, sera obtained from women with placental malaria specifically recognise the recombinant domain. The protein thus has functional properties consistent with a role in placental sequestration.
Keywords: structural biology, X-ray crystallography, antigenic recognition, antibody structure, Plasmodium antigens