While
cdk5 was originally identified in vertebrates based on its homology to the cell cycle regulated
cdc2, further study demonstrated a function not in the cell cycle but instead in post-mitotic neurons. Mouse knockouts of
cdk5 and its activator,
p35, result in defects in the pattern of neuronal migration in the cortex (1,2). Studies in cultured rat neurons demonstrated that the
cdk5/p35 kinase activity can promote neurite outgrowth (3). To further study its role in neuronal function, we decided to study the C. elegans homolog of
cdk5.C. elegans contains a single homolog each of
cdk5 and the
p35 activator. Full length translational GFP reporter constructs for
cdk-5 and
p35 are expressed in the cytoplasm of most neurons beginning around the two-fold stage of embryogenesis and continuing into adulthood. To study
cdk-5 function, we identified a mutant from a deletion library, and the C. elegans Knockout Consortium identified an additional mutant. Both alleles delete a significant portion of the coding region and are predicted to be null.
cdk-5 mutants are viable and show no obvious defects in movement or body morphology. We detected no defects in cell migration or axon outgrowth.
cdk-5 mutants, however, are resistant to paralyzation by aldicarb, a drug that leads to increased levels of acetylcholine in the synaptic cleft. Conversely, worms that overexpress
cdk-5 are hypersensitive to aldicarb.
cdk-5 is unlikely to be acting in the muscle, because mutants show a normal response to levamisole, an acteylcholine receptor agonist. Consistent with this result, preliminary results indicate that
cdk-5 functions cell autonomously in the cholinergic motor neurons. Together, these results demonstrate a role for
cdk-5 in synaptic function. One in vitro target of the
cdk5 kinase is the mammalian homolog of UNC-18, a neuronal protein that binds to the closed form of syntaxin. This phosphorylation releases UNC-18 from syntaxin (4). One model proposed that this release allows syntaxin to adopt the open form and participate in exocytosis. We are testing this model by analyzing the potential interactions among
cdk-5,
unc-18, and syntaxin in vivo. The model predicts that a phosphomimetic version of UNC-18 should phenocopy the
cdk-5 overexpression phenotype. The model also predicts that an open form of syntaxin will bypass the
cdk-5 aldicarb phenotype. (1) Gilmore EC, et al., J Neurosci, 18:6370-7 (1998), (2) Chae T, et al., Neuron, 18:29-42 (1997), (3) Nikolic M, et al., Genes Dev, 10:816-25 (1996), (4) Fletcher AI, et al., J Biol Chem, 274:4027-35 (1999)