Structural basis for ligand recognition by integrins

Integrins, the major cell surface receptors mediating cell–extracellular matrix (ECM) adhesion, are central to the basic physiology underlying all multicellular organisms. As the complexity of animal body architecture increased, integrins were forced to acquire recognition capabilities toward the wide variety of ECM ligands and cell surface counter-receptors that emerged during evolution. Structural determination of the integrin–ligand complexes for both I domain-containing and non-I domain-containing integrins revealed two fundamentally different types of integrin-binding surfaces. In addition, recent advances in the biochemical and pharmacological characterization of the integrin–ligand interactions are beginning to reveal how integrins achieve specific recognition of wide variety of ligands using a small binding cleft at the subunit interface common to all integrins.

Introduction

Integrins are heterodimeric adhesion receptors composed of α and β subunits that link ECM to the cytoskeleton. Although they can be found in simple metazoan organisms such as sponges and corals [1], their numbers and variety have greatly expanded throughout the course of evolution. Humans have 18 α and 8 β subunits, assembled into 24 distinct heterodimers. Multiple α subunits can combine with single β subunits (and vice versa), giving rise to ‘combinatorial’ ligand specificity. The need to gain a mechanistic understanding of the ligand recognition exercised by integrins remains paramount, since basic cellular events such as cell adhesion, cell migration, anchorage-dependent cell survival, and cell–cell communication depend heavily on this type of molecular interaction. Integrins are also the focus of extensive clinical research, because of their crucial involvement in numerous disease states. However, the extent to which we understand the structural basis for these interaction varies among different integrin–ligand pairs, because of the limited number of high-resolution crystal structures thus far determined (Table 1). There are only five independent integrin–ligand complex structures available: three are for I domain integrins and two are for β3 integrins. The vast majority of integrin heterodimers, particularly those of β1 class, have not been crystallized, even in their apo-form. Although direct structural determination is the most straightforward way to gain insights into the molecular mechanisms underlying recognition, recent advances in the biochemical characterization of integrin–ligand interactions using purified recombinant proteins (as opposed to more conventional cell adhesion assays) have contributed to unraveling some of the unexpected binding modes used by certain integrin–ligand pairs. Efforts to develop ligand mimetic compounds that target therapeutic intervention have also provided structural insights into the ligand-binding pocket of integrins. A rich body of these biochemical and pharmacological studies can therefore be combined with the structures determined thus far to increase our mechanistic understanding of these interactions. In this review, I will focus on the general structure and specificity of ligand recognition as revealed by crystal structures and other recent biochemical findings on integrin–ligand interactions. For a detailed overview of integrin ectodomain structure and its conformational variations, readers are referred to a recent review by Luo and Springer [2], as well as one by Arnaout in this volume.