The topographic location of cyclic AMPdependent protein kinase has been studied in preparations of permeable and sealed membranes derived from human erythrocytes. Both the catalytic and cyclic AMP-binding components of protein kinase are localized on the inner, cytoplasmic, surface of the plasma membrane.3':5'-Cyclic AMP has been implicated in the regulation of a variety of membrane-associated functions including: contact inhibition (1, 2), hormone secretion (3), cellular permeability (4), and synaptic transmission (5, 6). It is possible that membrane-associated, cyclic AMP-dependent protein kinases could alter the structure and function of cell membranes by catalyzing the phosphorylation of specific polypeptides in plasma membranes (7-13).Plasma membranes derived from human erythrocytes contain greater than 70% of the total protein kinase activity found in these cells (14), and catalyze the cyclic AMPcontrolled phosphorylation of two endogenous polypeptide chains (9-11). These observations suggest that erythrocyte membranes provide an appropriate system for investigations on protein kinase-mediated alterations in the properties of membranes.Since the protein components of the erythrocyte membrane are asymmetrically positioned with respect to the cytoplasmic and exterior membrane surfaces (15), the biological role of cyclic AMP in membranes may be uniquely determined by the topographic distribution of the enzymes involved in the metabolism and action of the cyclic nucleotide. In the present experiments the spatial orientations of the catalytic and cyclic AMP-binding components of protein kinase were studied in permeable ghosts, sealed rightside-out ghosts, and sealed inside-out vesicles prepared according to the methods of Steck and Kant (16,17) and Bodemann and Passow (18). The disposition of protein kinase in the membrane was determined by comparing its accessibility in the various membrane preparations to the accessibility of enzymes known to be associated with either the cytoplasmic or the exterior surfaces of the membrane (15-17). (EC 1.6.4.3) by the procedure of Wallach and Kamat (22). Assays of enzymic and cyclic AMP-binding activities in salt-sealed ghosts were performed in saline buffered with 10 mM Tris -HCl, pH 7.4, to prevent lysis. Substitution of buffered saline for the standard buffers had no effect on the enzymic and binding activities in any of the membrane preparations. Inside-out vesicles and Mg2+-sealed ghosts were not lysed by incubation in the standard assay buffers.
MATERIALS AND METHODS
RESULTSThe data on the accessibility of cyclic AMP-dependent protein kinase and three reference enzymes are presented in Table 1. Acetylcholinesterase, a marker for the exterior surface of the erythrocyte membrane (23), is completely accessible in permeable ghosts and sealed, rightside-out membranes, but is relatively inaccessible in sealed, inside-out membranes. Conversely, glyceraldehyde-3-phosphate dehydrogenase, an indicator of the inner surface of the membrane (24), is completely available i...