Selective recognition of cyclic RGD peptides of NMR defined conformation by alpha IIb beta 3, alpha V beta 3, and alpha 5 beta 1 integrins

Binding of fibrinogen to platelet integrin alpha IIb beta 3 in solution as monitored by tracer sedimentation equilibrium

Fibrinogen showed essentially no binding (KD > 1 mM) to platelet alpha IIb beta 3 integrin in solution in the presence of Triton or octylglucoside above critical micellar concentrations. Under these conditions the integrin was an alpha beta monomer. After removal of the detergent from the Triton containing buffer (25 mM Tris/HCl;, 150 mM NaCl, 1 mM CaCl2, 1 mM MgCl2, pH 7.4) the integrin formed aggregates with hexamers as the most prominent species, as demonstrated by analytical ultracentrifugation and electron microscopy. Tracer sedimentation equilibrium experiments indicate that fibrinogen binds to the integrin aggregates, but with a surprisingly large KD (at least 3 microM). This value is 10- to 100-fold higher than values determined by solid phase assays or with integrins reconstituted onto lipid bilayers.

A paradigm for drug discovery using a conformation from the crystal structure of a presentation scaffold

We describe a structural validation of the use of presentation scaffolds for control and elucidation of bioactive conformations of peptides. The protein REI-RGD34--produced by inserting the sequence RIPRGDMP into the CDR1 loop region of the immunoglobulin VL domain REI--strongly inhibits fibrinogen binding to the integrins alpha IIb beta 3 and alpha V beta 3. In the X-ray crystal structure of their protein at 2.4 A resolution, the RGD-containing loop exhibits defined electron density that is consistent with models for the bioactive conformations of ligands of these receptors based on previous small-molecule studies. Furthermore, a search of a small-molecule database with conformational information derived from the structure of REI-RGD34 identified constrained peptides and peptidomimetics known to be antagonists of the platelet receptor alpha IIb beta 3.

Definition of two angiogenic pathways by distinct alpha v integrins

Angiogenesis depends on cytokines and vascular cell adhesion events. Two cytokine-dependent pathways of angiogenesis were shown to exist and were defined by their dependency on distinct vascular cell integrins. In vivo angiogenesis in corneal or chorioallantoic membrane models induced by basic fibroblast growth factor or by tumor necrosis factor-alpha depended on alpha v beta 3, whereas angiogenesis initiated by vascular endothelial growth factor, transforming growth factor-alpha, or phorbol ester depended on alpha v beta 5. Antibody to each integrin selectively blocked one of these pathways, and a cyclic peptide antagonist of both integrins blocked angiogenesis stimulated by each cytokine tested. These pathways are further distinguished by their sensitivity to calphostin C, an inhibitor of protein kinase C that blocked angiogenesis potentiated by alpha v beta 5 but not by alpha v beta 3.

Structural comparison of two strains of foot-and-mouth disease virus subtype O1 and a laboratory antigenic variant, G67

BACKGROUND

Foot-and-mouth disease viruses (FMDVs) are members of the picornavirus family and cause an economically important disease of cloven-hoofed animals. To understand the structural basis of antigenic variation in FMDV, we have determined the structures of two viruses closely related to strain O1BFS whose structure is known.

RESULTS

The two new structure are, like O1BFS, both serotype O viruses. The first, O1 Kaüfbeuren (O1K), is a field isolate dating from an outbreak of FMD in Europe in the 1960s. The second, called G67, is a quadruple mutant of O1K, generated in the laboratory, that bears point mutations conferring resistance to neutralizing by monoclonal antibodies, specific for each of the four major antigenic sites defined previously. The availability of the three related virus structures permits a detailed analysis of the way amino acid substitutions influence antigenicity. Structural changes are seen to be limited, in general, to the substituted side chain. For example, the GH loop of VP1, a highly antigenic and mobile protuberance which becomes ordered only under reducing conditions, was essentially indistinguishable in the three viruses despite the accumulation of up to four changes within its 15-residue sequence. At one of the other antigenic sites, however, changes between the two field strains did perturb both side-chain and main-chain structures in the vicinity.

CONCLUSIONS

The conservation of conformation of the GH loop of VP1 adds to the evidence implicating an integrin as the cellular receptor for FMDV, since this loop contains a conserved RGD (Arg-Gly-Asp) sequence structurally similar to the same tripeptide in some other integrin-binding proteins. Structural changes required for the virus to escape neutralization by monoclonal antibodies are generally small. The more extensive type of structural change exhibited by the field isolates probably reflects differing selective pressures operating in vivo and in vitro.

Peptide ligands for integrin alpha v beta 3 selected from random phage display libraries

The integrin alpha v beta 3 binds promiscuously to cell-adhesive proteins: vitronectin, fibronectin, and several others containing the RGD motif. We have explored molecular recognition by alpha v beta 3 through selection of ligands from large random libraries of peptides displayed on phage. Ligands bound by alpha beta 3 consisted primarily of RGD peptides; however, these peptides showed considerable heterogeneity with respect to the identities of amino acids flanking RGD. The tolerance of alpha v beta 3 for RGD peptides of diverse composition is consistent with its role in vivo as a versatile receptor for RGD-containing extracellular matrix proteins. Peptide ligands for alpha v beta 3 also included a novel binding sequence, identical to a tetrapeptide found in vitronectin, which is a candidate for a synergistic site in this adhesive protein that may act in concert with RGD to promote molecular recognition.

Integrin alpha v beta 3 antagonists promote tumor regression by inducing apoptosis of angiogenic blood vessels

A single intravascular injection of a cyclic peptide or monoclonal antibody antagonist of integrin alpha v beta 3 disrupts ongoing angiogenesis on the chick chorioallantoic membrane (CAM). This leads to the rapid regression of histologically distinct human tumors transplanted onto the CAM. Induction of angiogenesis by a tumor or cytokine promotes vascular cell entry into the cell cycle and expression of integrin alpha v beta 3. After angiogenesis is initiated, antagonists of this integrin induce apoptosis of the proliferative angiogenic vascular cells, leaving preexisting quiescent blood vessels unaffected. We demonstrate therefore that ligation of integrin alpha v beta 3 is required for the survival and maturation of newly forming blood vessels, an event essential for the proliferation of tumors.

Comparison of disintegrins with limited variation in the RGD loop in their binding to purified integrins alpha IIb beta 3, alpha V beta 3 and alpha 5 beta 1 and in cell adhesion inhibition

The inhibitory capacities of six different disintegrins and one related neurotoxin analogue for the binding of RGD-dependent integrins to either fibrinogen, vitronectin or fibronectin were compared in solid phase assays. Echistatin and flavoridin were the most active inhibitors for alpha V beta 3 and alpha 5 beta 1 integrins and moderately exceeded the activity of the natural protein ligands. The same disintegrins together with eristostatin, bitistatin and barbourin were also very potent inhibitors of fibrinogen binding to alpha IIb beta 3 integrin. For all three integrins, albolabrin showed the lowest affinity, but it still clearly exceeded that of synthetic GRGDS. However, assay conditions may determine these relative affinities, as shown for the alpha IIb beta 3 and alpha V beta 3 integrins when used either in immobilized or soluble form. For alpha IIb beta 3, however, a close correlation was found between KD values determined in platelet binding assays and the concentrations required for half maximal inhibition of three disintegrins. The inhibiting capacity of disintegrins in assays with purified integrins also correlated reasonably well with their inhibition of cell attachment to RGD-dependent protein substrates. However, sequence differences in the RGD loops of the various disintegrins may not fully account for the 20-100-fold difference in their binding capacities. This was particularly evident for echistatin and albolabrin, which differ in this region only by two conservative substitutions but have considerably different inhibitory activities. More remote regions of the disintegrins and alignment of disulfide bridges are therefore likely to contribute to their affinity and selectivity.

Pharmacophore refinement of gpIIb/IIIa antagonists based on comparative studies of antiadhesive cyclic and acyclic RGD peptides

Structurally guided design approaches to low-molecular-weight platelet aggregation antagonists addressing the platelet-associated heterodimeric cell surface receptor gpIIb/IIIa rely on comparative studies of an ensemble of conformationally and biologically characterized compounds, since no high-resolution structure of the receptor system is available. We report a classical indirect and comparative pharmacophore refinement approach based on a series of small cyclic Arg-Gly-Asp (RGD) peptides as gpIIb/IIIa-fibrinogen interaction antagonists. These peptides have previously been investigated as potent and selective tumor cell adhesion inhibitors. The definition of geometrical descriptors classifying the RGD peptide conformations and their subsequent analysis over selected RGD- and RXD-containing protein structures allows for a correlation of distinct structural features for platelet aggregation inhibition. An almost parallel alignment of the Arg and Asp side chains was identified by a vector analysis as being present in all active cyclic hexa- and pentapeptides. This orientation is induced mainly by the constraint of backbone cyclization and is not of any covalent tripeptide-inherent origin, which was rationalized by a 500 ps high-energy MD simulation of a sequentially related linear model peptide. The incorporation of the recognition tripeptide Arg-Gly-Asp into the cyclic peptide templates acted as a filter mechanism, restricting the otherwise free torsional relation of both side chains to a parallel orientation. Based on the derived results, several detailed features of the receptor binding site could be deduced in terms of receptor complementarity. These findings should govern the design of next-generation compounds with enhanced activities. Furthermore, the complementary stereochemical characteristics of the substrate can be used as boundary conditions for pseudoreceptor modelling studies that are capable of reconstructing a hypothetical binding pocket, qualitatively resembling the steric and electronic demands of gpIIb/IIIa. It is interesting to note that these features provide clear differentiation to requirements for inhibition of alpha v beta 3 substrate binding. This can account for the extremely high selectivity and activity of some of our constrained peptides for either the alpha IIb beta 3 or the alpha v beta 3 receptor.