Chemical dissection of the APC Repeat 3 multistep phosphorylation by the concerted action of protein kinases CK1 and GSK3.

A crucial event in machinery controlled by Wnt signaling is the association of beta-catenin with the adenomatous polyposis coli (APC) protein, which is essential for the degradation of beta-catenin and requires the multiple phosphorylation of APC at six serines (1501, 1503, 1504, 1505, 1507, and 1510) within its repeat three (R3) region. Such a phosphorylation is believed to occur by the concerted action of two protein kinases, CK1 and GSK3, but its mechanistic aspects are a matter of conjecture. Here, by combining the usage of variably phosphorylated peptides reproducing the APC R3 region and Edman degradation assisted localization of residues phosphorylated by individual kinases, we show that the process is initiated by CK1, able to phosphorylate S1510 and S1505, both specified by non-canonical determinants. Phosphorylation of S1505 primes subsequent phosphorylation of S1501 by GSK3. In turn, phospho-S1501 triggers the hierarchical phosphorylation of S1504 and S1507 by CK1. Once phosphorylated, S1507 primes the phosphorylation of both S1510 and S1503 by CK1 and GSK3, respectively, thus completing all six phosphorylation steps. Our data also rule out the intervention of CK2 despite the presence of a potential CK2 phosphoacceptor site, S1510LDE, in the R3 repeat. S1510 is entirely unaffected by CK2, while it is readily phosphorylated even in the unprimed peptide by CK1delta but not by CK1gamma. This discloses a novel motif significantly different from non-canonical sequences phosphorylated by CK1 in other proteins, which appears to be specifically recognized by the delta isoform of CK1.

The first armadillo repeat is involved in the recognition and regulation of beta-catenin phosphorylation by protein kinase CK1.

Multiple phosphorylation of beta-catenin by glycogen synthase kinase 3 (GSK3) in the Wnt pathway is primed by CK1 through phosphorylation of Ser-45, which lacks a typical CK1 canonical sequence. Synthetic peptides encompassing amino acids 38-64 of beta-catenin are phosphorylated by CK1 on Ser-45 with low affinity (K(m) approximately 1 mM), whereas intact beta-catenin is phosphorylated at Ser-45 with very high affinity (K(m) approximately 200 nM). Peptides extended to include a putative CK1 docking motif (FXXXF) at 70-74 positions or a F74AA mutation in full-length beta-catenin had no significant effect on CK1 phosphorylation efficiency. beta-Catenin C-terminal deletion mutants up to residue 181 maintained their high affinity, whereas removal of the 131-181 fragment, corresponding to the first armadillo repeat, was deleterious, resulting in a 50-fold increase in K(m) value. Implication of the first armadillo repeat in beta-catenin targeting by CK1 is supported in that the Y142E mutation, which mimics phosphorylation of Tyr-142 by tyrosine kinases and promotes dissociation of beta-catenin from alpha-catenin, further improves CK1 phosphorylation efficiency, lowering the K(m) value to <50 nM, approximating the physiological concentration of beta-catenin. In contrast, alpha-catenin, which interacts with the N-terminal region of beta-catenin, prevents Ser-45 phosphorylation of CK1 in a dose-dependent manner. Our data show that the integrity of the N-terminal region and the first armadillo repeat are necessary and sufficient for high-affinity phosphorylation by CK1 of Ser-45. They also suggest that beta-catenin association with alpha-catenin and beta-catenin phosphorylation by CK1 at Ser-45 are mutually exclusive.

Unique substrate specificity of anaplastic lymphoma kinase (ALK): development of phosphoacceptor peptides for the assay of ALK activity.

The anaplastic lymphoma kinase (ALK), whose constitutively active fusion proteins are responsible for 5-10% of non-Hodgkin's lymphomas, shares with the other members of the insulin receptor kinase (IRK) subfamily an activation loop (A-loop) with the triple tyrosine motif Y-x-x-x-Y-Y. However, the amino acid sequence of the ALK A-loop differs significantly from the sequences of both the IRK A-loop and the consensus A-loop for this kinase subfamily. A major difference is the presence of a unique "RAS" triplet between the first and second tyrosines of the ALK A-loop, which in IRK is replaced by "ETD". Here we show that a peptide reproducing the A-loop of ALK is readily phosphorylated by ALK, while a homologous IRK A-loop peptide is not unless its "ETD" triplet is substituted by "RAS". Phosphorylation occurs almost exclusively at the first tyrosine of the Y-x-x-x-Y-Y motif, as judged by Edman analysis of the phosphoradiolabeled product. Consequently, a peptide in which the first tyrosine had been replaced by phenylalanine (FYY) was almost unaffected by ALK. In contrast, a peptide in which the second and third tyrosines had been replaced by phenylalanine (YFF) was phosphorylated more rapidly than the parent peptide (YYY). A number of substitutions in the YFF peptide outlined the importance of Ile and Arg at positions n - 1 and n + 6 in addition to the central triplet, to ensure efficient phosphorylation by ALK. Such a peculiar substrate specificity allows the specific monitoring of ALK activity in crude extracts of NPM-ALK positive cells, using the YFF peptide, which is only marginally phosphorylated by a number of other tyrosine kinases.