This study shows that the Drosophila insulin receptor autonomously controls cell and organ size, and that overexpression of a gene encoding an insulin-like peptide is sufficient to increase body size.

We report the cloning and primary structure of the Drosophila insulin receptor gene (inr), functional expression of the predicted polypeptide, and the isolation of mutations in the inr locus. Our data indicate that the structure and processing of the Drosophila insulin proreceptor are somewhat different from those of the mammalian insulin and IGF 1 receptor precursors. The INR proreceptor (M(r) 280 kDa) is processed proteolytically to generate an insulin‐binding alpha subunit (M(r) 120 kDa) and a beta subunit (M(r) 170 kDa) with protein tyrosine kinase domain. The INR beta 170 subunit contains a novel domain at the carboxyterminal side of the tyrosine kinase, in the form of a 60 kDa extension which contains multiple potential tyrosine autophosphorylation sites. This 60 kDa C‐terminal domain undergoes cell‐specific proteolytic cleavage which leads to the generation of a total of four polypeptides (alpha 120, beta 170, beta 90 and a free 60 kDa C‐terminus) from the inr gene. These subunits assemble into mature INR receptors with the structures alpha 2(beta 170)2 or alpha 2(beta 90)2. Mammalian insulin stimulates tyrosine phosphorylation of both types of beta subunits, which in turn allows the beta 170, but not the beta 90 subunit, to bind directly to p85 SH2 domains of PI‐3 kinase. It is likely that the two different isoforms of INR have different signaling potentials. Finally, we show that loss of function mutations in the inr gene, induced by either a P‐element insertion occurring within the predicted ORF, or by ethylmethane sulfonate treatment, render pleiotropic recessive phenotypes that lead to embryonic lethality. The activity of inr appears to be required in the embryonic epidermis and nervous system among others, since development of the cuticle, as well as the peripheral and central nervous systems are affected by inr mutations.

A set of 57 synthetic peptides encompassing the entire triplehelical domain of human collagen III was used to locate binding sites for the collagen-binding integrin ␣ 2 ␤ 1 . The capacity of the peptides to support Mg 2؉ -dependent binding of several integrin preparations was examined. Wild-type integrins (recombinant ␣ 2 I-domain, ␣ 2 ␤ 1 purified from platelet membranes, and recombinant soluble ␣ 2 ␤ 1 expressed as an ␣ 2 -Fos/␤ 1 -Jun heterodimer) bound well to only three peptides, two containing GXXGER motifs (GROGER and GMOGER, where O is hydroxyproline) and one containing two adjacent GXXGEN motifs (GLKGEN and GLOGEN). Two mutant ␣ 2 I-domains were tested: the inactive T221A mutant, which recognized no peptides, and the constitutively active E318W mutant, which bound a larger subset of peptides. Adhesion of activated human platelets to GER-containing peptides was greater than that of resting platelets, and HT1080 cells bound well to more of the peptides compared with platelets. Binding of cells and recombinant proteins was abolished by anti-␣ 2 monoclonal antibody 6F1 and by chelation of Mg 2؉ . We describe two novel high affinity integrinbinding motifs in human collagen III (GROGER and GLOGEN) and a third motif (GLKGEN) that displays intermediate activity. Each motif was verified using shorter synthetic peptides.