Biotecnologia Aplicada Elfos Scientiae ISSN: 0684-4551 Vol. 17, Num. 3, 2000, pp. 196

Biotecnología Aplicada, Volume 17, July-September 2000, p. 196

Modelling of Chimeric b-galactosidase Antigenic Fusion Proteins

N Boutonnet, M Prévost

Université Libre de Bruxelles, Ingénierie Biomoléculaire, CP165/64, 50 av. FD Roosevelt, B-1050 Bruxelles, Belgium; fax number: 00.32.26503606; E-mail: nboutonn@ulb.ac.be ; mprevost@ulb.ac.be

From a selection of papers from Biotecnología Habana`99 Congress.
November 28-December 3, 1999.

Code Number: BA00060

Introduction

A new principle for antibody detection has been proven to have a high potential for the development of faster analytical tests. lt has been shown that the insertion of antigenic peptides of the foot-and-mouth disease virus (FMDV) into specific sites of the bacterial enzyme b-galactosidase causes a severe reduction of the activity of this enzyme which can be recovered upon antibody binding to the inserted antigenic peptide [1]. In principle, the recovery of b-galactosidase activity upon antigen-antibody binding allows the detection of antibodies by simple colorimetric quantification of b-galactosidase activity in a homogeneous test system, involving only the antigenic fusion protein, a chromogenic substrate and a serum sample for testing.

The lack of three-dimensional structural information on the chimeric proteins renders their modelling requisite to understand the chimeric enzyme inactivation and reactivation upon antibody binding at the molecular level. For this purpose, we have determined models of chimeric b-galactosidase by comparative modelling and structure prediction techniques.

Materials and Methods

The 2D structures are predicted using the algorithm Fugue [2], while the 3D structures are calculated using the comparative modelling program Modeller5 [3].

Results and Discussion

Construction of the antigenic peptide at position 278 of the b-galactosidase has been performed since the construct is known to display high antigenicity and reactivation upon antibody binding to whole antibodies, but not to a single Fab fragment. It involves three main steps that will constitute a protocol to be used for other constructs.

1. To model the most probable conformations of the antigenic peptide to be inserted exploiting structural data either from X ray diffraction experiments or from secondary structure predictions. Several X ray structures of the whole viral proteins are available. The only structures, however, that show a well-defined conformation of the antigenic peptide are those of a mutant virus lacking one disulfide bridge (serotype O) and those of a short peptide (13 aa, serotype C) in interaction with SD6 antibody. Both experimental and prediction data agree on the local conformation: the segment before the highly conserved triplet Arg-Gly-Asp, which is involved in the recognition process, is in an extended conformation and the segment beyond adopts a helical conformation. However, the two experimental X ray structures show a relative orientation of the two segments that markedly differs. To model the peptide (27 aa) from serotype C with comparative modelling, only the serotype O structure is used as it displays a peptide sequence long enough to model the insertion even though it presents the drawback of a non-entirely homologous sequence (serotype O versus C).

2. To properly model the 3D structures of the peptide-inserted tetrameric b-galactosidase using b-galactosidase X ray structure, secondary structure predictions on the chimeric sequence and comparative modelling. Two- and four-residue lengths are chosen to anchor the peptide onto the enzyme so as to produce different potential orientations of the peptide at the enzyme surface. Overall 110 3D dimeric models are generated, 55 for each anchoring mode. A clustering procedure based on the computed root mean square deviation of the backbone atom coordinates is applied on the models. For example, with the 2-residue anchoring mode, 23 clusters are obtained for a rms cutoff of 1.5Å. As residues in the vicinity of the insertion site in the wild-type protein are known to interact with the active site of the facing monomer, conformational analysis of the selected models is performed to detect interactions between the inserted peptide in one monomer with active site residues of the other that could account for partial inactivation of the chimeric protein as experimentally observed. Molecular dynamics simulations are performed on the dimeric models to improve the conformational sampling of the inserted fragment accounting for its flexibility.

3. To model the interaction of the chimeric enzyme with the antibodies. The modelled dimers that display a solvent accessibility surface area large enough to bind the antibody are selected. Tetramers are built to test their ability to bind an antibody molecule.

The studies are currently extended to model other chimeric proteins carrying antigenic FMDV inserted at different positions. All models should help, in the one hand, to interpret and rationalize the large wealth of experimental data previously obtained. In the second hand, they should permit to guide new experiments on inserted-peptide enzymes and more particularly on new recombinant b-galactosidase HIV antigenic fusion proteins as enzymatic probes for anti-HIV antibody detection in HIV infected individuals.

References

Copyright Elfos Scientiae 2000