Crystal structures of the insulin receptor substrate-1 (IRS-1) phospho
tyrosine-binding (PTB) domain, alone and complexed with the juxtamembr
ane region of the insulin receptor, show how this domain recognizes ph
osphorylated ''NPXY'' sequence motifs. The domain is a 7-stranded beta
sandwich capped by a C-terminal helix. The insulin receptor phosphope
ptide fills an L-shaped cleft on the domain. The N-terminal residues o
f the bound peptide form an additional strand in the beta sandwich, st
abilized by contacts with the C-terminal helix. These interactions exp
lain why IRS-1 binds to the insulin receptor but not to NPXpY motifs i
n growth factor receptors. The PTB domains of IRS-1 and Shc share a co
mmon fold with pleckstrin homology domains. Overall, ligand binding by
IRS-1 and Shc PTB domains is similar, but residues critical for phosp
hotyrosine recognition are not conserved.