A poorly hydrolyzed substrate, N6-etheno cyclic AMP, was used to assay high concentrations (0.2 - 0.6 MuM) of cyclic nucleotide phosphodiesterase (PDE) permitting direct comparison of activity with changes in physical properties of the enzyme. Although the interaction constant for this substrate (2-3mM) was 100-fold higher than that for cAMP, the regulatory properties of the enzyme were comparable (e.g., Ka for Mg2+, Ki for spermine, degree of stimulation by calmodulin (CaM). When the Ca2+-dependence of enzyme activation was compared with that for interaction with dansyl-calmodulin (D-CaM) using identical experimental samples, less Ca2+ was required for interaction than for stimulation of activity; this suggested sequential steps in the mechanism of PDE activation by CaM. Immunocytochemical studies in rat brain indicated that specific changes in the distribution of PDE in cerebellum occurred after pharmacologic lesions of the inferior olivary nucleus, while that of calcineurin (CN) did not. Since this treatment affects excitatory innervation of Purkinje cells, it is possible that such input pathways may modulate, transynaptically, the local expression of PDE. Using overlay procedures, CN has been identified as the predominant CaM-binding protein in isolated spleen cells and cultured PC-12 cells; smaller amounts of cytoskeletal CaM-binding proteins (caldesmon, spectrin) have also been found. In some instances, there were changes in the amounts of these proteins with differentiation, suggesting a role for Ca2+-dependent protein dephosphorylation and/or cytoskeletal modification during cellular activation. Expression vector immunoscreening procedures were optimized to permit isolation of putative cDNA clones for PDE and CN using a lambda GT-aa rat brain library. Lysogens of these clones were produced and the fusion proteins analyzed for immunoreactivity against affinity-purified CN and PDE antibodies, and against monoclonal anti-beta galactosidase antibody.
