Obtustatin and viperistatin represent the shortest known snake venom monomeric disintegrins. In the present study, we have produced recombinant full-length wild-type and site-directed mutants of obtustatin to assess the role of the K 21 TS 23 tripeptide and C-terminal residues for specific inhibition of the α 1 β 1 integrin. Thr 22 appeared to be the most critical residue for disintegrin activity, whereas substitution of the flanking lysine or serine residues for alanine resulted in a less pronounced decrease in the anti-α 1 β 1 integrin activity of the disintegrin. The triple mutant A 21 AA 23 was devoid of blocking activity towards α 1 β 1 integrin-mediated cell adhesion. The potency of recombinant KTS-disintegrins also depended on the residue C-terminally adjacent to the active motif. Substitution of Leu 24 of wild-type obtustatin for an alanine residue slightly decreased the inhibitory activity of the mutant, whereas an arginine residue in this position enhanced the potency of the mutant over wild-type obtustatin by 6-fold. In addition, the replacements L38V and P40Q may account for a further 25-fold increase in α 1 β 1 inhibitory potency of viperistatin over KTSR-obtustatin.

We report the isolation and amino acid sequences of six novel dimeric disintegrins from the venoms of Vipera lebetina obtusa (VLO), V. berus (VB), V. ammodytes (VA), Echis ocellatus (EO) and Echis multisquamatus (EMS). Disintegrins VLO4, VB7, VA6 and EO4 displayed the RGD motif and inhibited the adhesion of K562 cells, expressing the integrin α5β1 to immobilized fibronectin. A second group of dimeric disintegrins (VLO5 and EO5) had MLD and VGD motifs in their subunits and blocked the adhesion of the α4β1 integrin to vascular cell adhesion molecule 1 with high selectivity. On the other hand, disintegrin EMS11 inhibited both α5β1 and α4β1 integrins with almost the same degree of specificity. Comparison of the amino acid sequences of the dimeric disintegrins with those of other disintegrins by multiple-sequence alignment and phylogenetic analysis, in conjunction with current biochemical and genetic data, supports the view that the different disintegrin subfamilies evolved from a common ADAM ( a d isintegrin a nd m etalloproteinase-like) scaffold and that structural diversification occurred through disulphide bond engineering.

The disulphide-bond pattern of the heterodimeric disintegrin EMF-10, a potent and selective integrin α 5 β 1 antagonist from Eristocophis macmahoni venom, was established by combination of amino-acid analysis, N-terminal sequencing and collision-induced dissociation by nanoelectrospray ionization quadrupole ion-trap MS of fragments isolated by reversed-phase HPLC after degradation of EMF-10 with oxalic acid. Each EMF-10 subunit contains four intrachain disulphide bonds. Two interchain cystine residues join the EMF-10 polypeptides. The intrachain linkages are conserved in monomeric disintegrins. A molecular model of EMF-10 was built using averaged NMR co-ordinates of flavoridin as a template. The active hairpin loops of the EMF-10 subunits occupy opposite locations at the ends of an elongated disulphide-bond ladder. In the EMF-10 model the N-terminal polypeptide of EMF-10B is close to the RGD-loop of the EMF-10A subunit, suggesting that the N-terminal region of the B-subunit could potentially influence the biological activity of the A-subunit.

Two disintegrins with a high degree of amino acid sequence similarity, echistatin and eristostatin, showed a low level of interaction with Chinese hamster ovary (CHO) cells, but they bound to CHO cells transfected with α IIb β 3 genes (A5 cells) and to CHO cells transfected with α v β 3 genes (VNRC3 cells) in a reversible and saturable manner. Scatchard analysis revealed that eristostatin bound to 816000 sites per A5 cell ( K d 28 nM) and to 200000 sites ( K d 14 nM) per VNRC3 cell respectively. However, VNRC3 cells did not bind to immobilized eristostatin. Echistatin bound to 495000 sites ( K d 53 nM) per A5 cell and to 443000 sites ( K d 20 nM) per VNRC3 cell. As determined by flow cytometry, radiobinding assay and adhesion studies, binding of both disintegrins to A5 cells and resting platelets and binding of echistatin to VNRC3 cells resulted in the expression of ligand-induced binding sites (LIBS) on the β 3 subunit. Eristostatin inhibited, more strongly than echistatin, the binding of three monoclonal antibodies: OPG2 (RGD motif dependent), A2A9 (α IIb β 3 complex dependent) and 7E3 (α IIb β 3 and α v β 3 complex dependent) to A5 cells, to resting and to activated platelets and to purified α IIb β 3 . Experiments in which echistatin and eristostatin were used alone or in combination to inhibit the binding of 7E3 and OPG2 antibodies to resting platelets suggested that these two disintegrins bind to different but overlapping sites on α IIb β 3 integrin. Monoclonal antibody LM 609 and echistatin seemed to bind to different sites on α v β 3 integrin. However, echistatin inhibited binding of 7E3 antibody to VNRC3 cells and to purified α v β 3 , suggesting that α v β 3 and α IIb β 3 might share the same epitope to which both echistatin and 7E3 bind. Eristostatin had no effect in these systems, providing further evidence that it binds to a different epitope on α v β 3 .