Absence of functional insulin receptor substrate-3 (IRS-3) gene in humans

Aim/hypothesis. Insulin receptor substrate (IRS) proteins play important roles in insulin action and pancreatic beta-cell function. At least four mammalian IRS molecules have been identified. Although genes and cDNAs encoding human IRS-1, IRS-2, and IRS-4 have been cloned, IRS-3 has been identified only in rodents. Thus, we have attempted to clone the human IRS-3 gene. Methods. Insulin-stimulated rat or human adipocytes were subjected to Western blot analysis to assess IRS-3 tyrosine phosphorylation. Human liver and adipose cDNA libraries were screened in an effort to clone IRS-3 cDNA. A PCR-based approach was designed to amplify IRS-3 cDNA. Reverse transcription PCR was carried out using mRNA from adipose tissue, liver, and skeletal muscle as templates in combination with an in silico screen using mouse IRS-1, IRS-2 and IRS-3 in a tblastn search of the draft public human genome. Results. In human adipocytes we did not detect a M-r 60 000 phosphoprotein corresponding to IRS-3, whereas in rat adipocytes IRS-3 protein and insulin-stimulated tyrosine phosphorylation was readily observed. None of the molecular approaches provided evidence for a functional IRS-3 gene in human tissue. Two deletions in human IRS-3 gene were identified using bioinformatics. The human IRS-3 gene product is predicted to lack a phosphotyrosine binding domain and also the sequence corresponding amino acid 353-407 of murine IRS-3. The contiguous sequence of genomic DNA between these two homologous regions does not have the coding information for human IRS-3. Conclusion/interpretation. In silico screening of the human IRS-3 genome region, combined with further biological and molecular validation, provides evidence against a functional IRS-3 in humans.