An analysis of the substrate specificity of insulin-stimulated protein kinase-1, a mammalian homologue of S6 kinase-II

The specificity determinants for insulin-stimulated protein kinase-I (ISPK-1) have been investigated with synthetic peptides based on naturally-occurring protein phosphoacceptor sequences. Peptides (Arg-Arg-Xaa-Ser-Xaa) that fulfill the consensus sequence for cyclic-AMP-dependent protein kinase (PK-A) are also phosphorylated readily by ISPK-1. The phosphorylation efficiency is improved by increasing the number of N-terminal arginine residues and by moving the arginyl cluster one residue further away from the serine, the nonapeptide (Arg)₄-Ala-Ala-Ser-Val-Ala being the best substrate among all the short peptides tested (K_m = 15 μM). Conversely, the substitution of either Thr for Ser or Lys for Arg is detrimental. Likewise, two flanking Pro residues and an Arg immediately N-terminal to the Ser act as negative specificity determinants. While the specificity of ISPK-1 shows several similarities to that of PK-A, including an absolute requirement for basic residues on the N-terminal side of the target Ser, it differs in several other respects including (1), the detrimental effect of a Lys for Arg substitution which is still compatible with some phosphorylation by ISPK-1, but not PK-A; (2), the presence of C-terminal acidic residues which are tolerated very well by ISPK-1, but are detrimental to PK-A; (3), the effect of substituting Phe for Val in the peptide Arg-Arg-Ala-Ser-Val-Ala, which improves the efficiency of phosphorylation by PK-A (lowering the K_m 4-fold), but has no effect on phosphorylation by ISPK-1. These differences in peptide substrate specificity may account in part for the different rates of phosphorylation of physiological substrates for ISPK-1 and PK-A, such as the G subunit of protein phosphatase-1.