Recent work on this project has provided new insights into the role of the proline-directed kinase, Cdk5, in regulating cell-cell and cell-matrix adhesion. Our previous studies had indicated that two distinct mechanisms of action are involved in Cdk5-dependent regulation of adhesion and migration: one involving Cdk5 kinase activity and a second, which is independent of kinase activity, but requires an intact phosphorylation site at Cdk5(Y15). Recent evidence has linked both of these mechanisms to Cdk5-dependent regulation of Src. We have found that cSrc is phosphorylated by Cdk5 in normal lens epithelial cells at a site in the unique region of the cSrc N-terminus. Phosphorylation at this site then targets cSrc for ubiquitylation by the E3 ubiquitin ligase, Cullin5. Loss of Cullin5, Cdk5 kinase activity, or site-specific mutagenesis of the phosphorylation site in cSrc, inhibits ubiquitin-dependent degradation of the active form of cSrc, allowing it to accumulate 1.5 to 2.0 fold. To test whether the physiological significance of this increase in intracellular Src, we have examined stress fiber formation and contraction in spreading cells. Under normal circumstances, Cdk5 activity is sharply elevated 1-2 hours after cells are plated on an extracellular matrix. In contrast, Src activity, which is elevated immediately after plating, declines to low levels after 1-2 hours. Inhibiting Cdk5 activity at this time elevates Src activity and increases phosphorylation of the Src substrate, p190RhoGAP. This, in turn, inhibits Rho activity and Rho-dependent phosphorylation of myosin regulatory light chain (MRLC), causing stress fiber disassembly and weakening cellular attachment to the extracellular matrix. Blocking Src activity prevents the effects of inhibiting Cdk5 on Rho activity, myosin phosphorylation, or stress fibers, indicating that the Rho-dependent cytoskeletal changes are the result of increased Src activity. These findings indicate that Cdk5 is a key regulator of Src activity in spreading cells and provide a mechanism for previously observed effects of Cdk5 inhibitors in regulating cell adhesion and migration. Since our data indicate that Cdk5-dependent regulation of Src activity is physiologically important, we have begun to investigate its possible involvement in disease. Interestingly, constitutive phosphorylation of Src by Cdk5 is found in a number of tumor cell lines, including those derived from retinoblastomas. We have, therefore, begun to test for mutations affecting the Cdk5-Src-Cullin5 signaling pathway in retinoblastoma cells. While Cdk5 kinase activity regulates Src by targeting the active Src for degradation, other work in this laboratory has shown that the adaptor function of Cdk5 is involved in yet another aspect of Src regulation: activation of cSrc at nascent focal adhesions. Suppressing Cdk5 expression with siRNA or overexpressing mutated Cdk5 proteins that can not be phosphorylated on Y15 prevents the activation of cSrc normally seen immediately after plating. Cdk5 siRNA does not interfere with cell attachment to extracellular matrix via integrins or with binding of focal adhesion kinase (FAK) to the integrin cytoplasmic tail;but binding of cSrc to the integrin-FAK complex is blocked, preventing focal adhesion maturation. These findings suggest that the adaptor function of Cdk5 is required either to transport Src to sites of focal adhesion formation or to facilitate its binding to autophosphorylated FAK at these sites. Our work has also implicated Cdk5 in regulating cell-cell adhesion. We had previously found that pharmacological inhibitors of Cdk5 reduced cell-cell adhesion and promoted degradation of E-cadherin. To extend our study of Cdk5s role in cell-cell adhesion, we generated a stable human corneal epithelial cell line with very low levels of endogenous Cdk5 using lentiviral mediated gene transfer to suppress Cdk5 expression with small hairpin RNA (shRNA). Analysis of this line (shHCLE) has confirmed that lack of Cdk5 expression impairs junctional stability, leading to internalization and degradation of E-cadherin. This is correlated with overexpression of p120-catenin, a protein which normally binds the juxtamembrane region of E-cadherin and stabilizes junctions. Ongoing studies are investigating the mechanism of these effects. In the past fiscal year we have completed a study of Notch signaling during lens fiber cell differentiation. While previous studies of the developing lens had shown that Notch signaling regulates differentiation of lens fiber cells by maintaining a proliferating precursor pool in the anterior epithelium, it was not clear whether Notch signaling was also required after the onset of fiber cell differentiation. To explore this question, we used rat lens epithelial explants undergoing FGF-2 dependent differentiation in vitro. The results showed that FGF induced Notch2 signaling (as judged by the appearance of activated Notch2 Intracellular Domain (N2ICD)) and expression of the Notch ligand, Jag1, within 12-24 hours. These changes were accompanied by induction of the Notch effector, Hes5, upregulation of N-cadherin, and downregulation of E-cadherin, a cadherin switch characteristic of fiber cell differentiation. Induction of Jag1 was efficiently blocked by U0126, a specific inhibitor of MAPK/ERK signaling, indicating a requirement for signaling through this pathway downstream of the FGF receptor. Moreover, other growth factors that activate MAPK/ERK signaling (EGF, PDGF, IGF) did not induce Jag1. Inhibition of Notch signaling using gamma secretase inhibitors DAPT and L-685, 458 or anti-Jag1 antibody suppressed FGF-dependent expression of Jag1 demonstrating that Jag1 expression is regulated by Notch-dependent lateral induction. Inhibition of Notch signaling also reduced expression of N-cadherin and the cyclin dependent kinase inhibitor, p57Kip2. Since upregulation of these proteins marks the onset of fiber differentiation in mammalian lenses, their regulation by Notch directly implicates Notch signaling in secondary fiber cell differentiation. Interestingly, N-cadherin and p57Kip2 are not upregulated by Notch signaling in the absence of FGF: on the contrary, Notch signaling represses the p57Kip2 promoter and is required to prevent lens epithelial cells from exiting the cell cycle .Thus, these results not only demonstrate that Notch-mediated lateral induction of Jag1 is an essential component of FGF-dependent lens fiber cell differentiation, but also show that signaling downstream of the FGF receptor interacts with the Notch pathway to alter the effect of Notch signaling on p57Kip2 gene expression. Since our in vitro studies of Notch signaling in the lens had implicated Notch 2 in lens differentiation, we have extended our study of this pathway by generating conditional knockouts of the Notch 2 gene in the lens, corneal epithelium, and other ectodermally-derived eye tissues using the Le-Cre promoter. Preliminary analyses of the conditional knockouts show a severe lens phenotype, with developmental abnormalities of the iris, ciliary body, and anterior chamber. Further analysis of this phenotype is in progress. In addition, we have initiated a collaboration with Dr. Nadean Brown to generate mice with conditional deletions of both Notch1 and Notch2, to determine the extent to which these receptors have overlapping and distinct functions in the eye. Studies of the corneal phenotype of PdLim2 knockout mice, which were reported as part of this project in the past are now considered as part of EY000259 (Molecular Biology of the Cornea).
