Mike Snyder, director of the Stanford Center for Genomics and Personalized Medicine, and his colleagues have applied a peptide-based approach for the high-throughput determination of kinase consensus phosphorylation motifs across the Saccharomyces kinome. Though they weren't able to determine the binding specificities for all yeast kinases, Snyder says that their comprehensive analysis of 111 out of the 122 total kinases "produced results" for 61 of them, representing more than half of those investigated. He also says that his team's phosphorylation studies in yeast transfer readily to studies of the human kinome.
"A lot of these kinases are very homologous to human kinases," Snyder says. "In fact, kinases as a family are more conserved than your average yeast protein. We take the results, just in general, to translate over."
Using an approach orthogonal to that proposed by Snyder and his collaborators in a 2004 Nature Methods paper, the team screened 111 kinases using peptide arrays. For kinases they couldn't obtain with substantial yield from yeast, the team used alternative mammalian and bacterial cell expression systems.
Sixty-one of the kinases displayed preferential phosphorylation at specific sites; Yck1 and Cka1 proved to be highly sequence-specific, requiring particular amino acid residues at several positions, whereas Cak1 and Rad53 displayed less selectivity — they required no residues at the phosphoacceptor substrate. The remaining kinases, the authors write, lie between these extremes; a finding they've visualized in their specificity quotient heat map.
Snyder says that these initial investigations opened up two novel possibilities for the team. First, he says, since the 3D structures of generic kinases and substrates are known, the team was able to predict which residues on the kinase aid in the recognition of residues at the substrate site.
"I call this the 'kinase code,'" he says. "The cool result is we can go in, and in one case we change the residue on the kinase — change its recognition title on the substrate in a predictable fashion — so this, then, is really working out the rules of how kinases recognize their substrates. And that's obviously a very generic issue, whether you're talking yeast, or humans, or anything."
The team was also able to predict additional substrate targets and they "validated a number of them," Snyder says. "That general scheme seems to work pretty well and so now people are [predicting substrates] for human kinases as well."