cell retraction (actin/microfilaments/cytoskeleton/calmodulin)

Transcription

1 Proc. Nati. Acad. Sci. USA Vol. 87, pp , January 1990 Medical Sciences Involvement of myosin light-chain kinase in endothelial cell retraction (actin/microfilaments/cytoskeleton/calmodulin) ROBERT B. WYSOLMERSKI* AND DAVID LAGUNOFF Department of Pathology, Saint Louis University School of Medicine, 1402 South Grand Boulevard, Saint Louis, MO Communicated by Earl P. Benditt, September 26, 1989 ABSTRACT Permeabilized bovine pulmonary artery endothelial cell monolayers were used to investigate the mechanism of endothelial cell retraction. Postconfluent endothelial cells permeabilized with saponin retracted upon exposure to ATP and Ca2+. Retraction was accompanied by thiophosphorylation of 19,000-Da myosin light chains when adenosine 5'-[y-["SSthioltrlphosphate was included in the dium. Both retraction and thiophosphorylation of myosin liht chains exhibited a graded quantitative dependence on Ca2+. When permeabilized monolayers were extracted in buffer D containing 100 mm KCI and 30 mm MgCI2 for 30 min, the cells failed to retract upon exposure to ATP and Ca2+, and no thiophosphorylation of myosin light chains occurred. The ability both to retract and to thiophosphorylate myosin light chains was restored by the addition to the permeabilized, extracted cells of myosin light-chain kinase and calmodulin together but not by either alone. These studies indicate that endothelial cell retraction, as does smooth muscle contraction, depends on myosin light-chain kinase phosphorylation of myosin light chains. A major function of the endothelial cell (EC) is to serve as a barrier to fluid and solute flux across the blood vessel wall. Breakdown of this barrier leads to increased permeability and the development of edema. Since the classic report by Majno and Palade (1) describing opening of intracellularjunctions on exposure to histamine, there has been considerable interest in the mechanism underlying this effect. Two broad possibilities have been entertained, change in the junctions themselves and intrinsic contractile activity of the ECs. In situ, lining blood vessels and in postconfluent cultures, ECs are flattened polygonal cells without obvious polarity, possessing a complex actin distribution with a dense peripheral band, stress fibers, and paranuclear filamentous array (2). Exposure of ECs in culture to histamine (3, 4), ethchlorvynol (2, 3), or cytochalasin B (3, 5) induces a reversible retraction of confluent ECs, leaving gaps between the cells, which remain attached to one another by thin processes. Intracellular ATP and extracellular Ca2' have been shown to be essential for the retraction induced by these agents, and the retraction is accompanied by extensive changes in actin filament distribution (3). These several observations favor the contraction hypothesis for intercellular gap formation. ECs, like other eukaryotic nonmuscle cells, contain myosin, actin, and associated proteins involved in cellular motile activities. For our studies, we have used the control of the contractile apparatus in smooth muscle as a model for EC contraction. Phosphorylation of the 20-kDa light chains of myosin is essential for smooth muscle contraction (6, 7). The phosphorylation is catalyzed by the Ca2", calmodulindependent enzyme myosin light-chain kinase (MLCK), which transfers the y-phosphate from ATP to myosin. The The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C solely to indicate this fact. light-chain phosphorylation permits activation of myosin ATPase by actin and is thereby believed to initiate the contractile events resulting in the shortening and tension development of smooth muscle cells (6-8). Skinned preparations of smooth muscle in which the membrane barrier to influx of large molecules and charged small molecules has been destroyed have been important in analyzing the biochemical characteristics of the contractile process in these cells (6, 7, 9). Similar preparations of fibroblasts (10, 11) and epithelial cells (12, 13) have also been studied. We have developed a preparation of permeabilized ECs that has provided us with the opportunity to investigate the role of ATP, Ca2+, calmodulin, and MLCK in the initiation of EC retraction. EXPERIMENTAL PROCEDURES Cell Culture. The bovine pulmonary artery cell line developed by Del Vecchio and Smith (14) was obtained from American Type Culture Collection (CLL-209). Cells were grown in Eagle's minimal essential medium supplemented with 2 mm glutamine, 10o or 20%o fetal calf serum, penicillin at 50 units/ml, and streptomycin at 50 gg/ml. Cells were maintained at 370C in a humidified 5% C02/95% air atmosphere. Cells used in these studies were 7 days postconfluent (2). Cell Permeabilization. EC cultures were washed with Dulbecco's phosphate-buffered saline (DPBS) (ph 7.2) and permeabilized with 2 ml of buffer A (20 mm Pipes/10 mm imidazole/50 mm KCI/1 mm EGTA/1 mm MgSO4/0.2 mm dithiothreitol/5,g of aprotonin per ml/5 ug of leupeptin per ml/10,ug of soybean trypsin inhibitor per ml/0.1 mm phenylmethylsulfonyl fluoride/0.5 mm benzamidine, ph 6.5) containing 25,ug of saponin per ml, incubated at 37 C for 10 min, and washed with 2 ml of buffer A without saponin. Care was taken not to allow the permeabilized monolayers to dry. Permeabilized monolayers were stimulated to retract on the addition of 100,uM free Ca2' and 100,uM exogenous ATP in buffer B (50 mm KCI/25 mm Pipes/2 mm MgSO4/1 mm EGTA/0.2 mm dithiothreitol/5,g of aprotonin per ml/5,ug of leupeptin per mj/10,ug of soybean trypsin inhibitor per ml/0.1 mm phenylmethylsulfonyl fluoride/0.5 mm benzamidine, ph 7.0) for 10 min at 370C. Retraction was terminated by removing the buffer containing ATP and adding 3% formaldehyde in buffer A. The free Ca2' concentration of buffer B was calculated by the method of Bers (15). Rhodamine Phalloidin Staining of Permeabilized Monolayers. For actin cytochemistry, cells were grown in Coming dishes (35 x 10 mm). Monolayers were permeabilized and the cells were stimulated to retract as described above. The monolayers were fixed in freshly prepared 3% formaldehyde in buffer A for 30 min at 220C. Fixed monolayers were washed with buffer A and stained with rhodamine phalloidin (Mo- Abbreviations: EC, endothelial cell; MLCK, myosin light-chain kinase; ATP[y-35S], adenosine-5'-[y-[35s]thio]triphosphate. *To whom reprint requests should be addressed. 16

2 Medical Sciences: Wysolmerski and Lagunoff lecular Probes) for visualization of F actin as outlined by Barak et al. (16). Thiophosphorylation. Permeabilized cultures were incubated with 100 ACi of adenosine 5'-[y-[35S]thio]triphosphate (ATP[y-35S]) (Amersham) (1 Ci = 37 GBq) in buffer A under varying conditions for 10 min at 370C, washed twice with buffer A and scraped up in 50 pl of SDS sample buffer (17), heated for 3 min at 100'C, and electrophoresed as described below. In some instances, myosin was immunoprecipitated following the method described by Kawamato and Adelstein (18). The washed cultures were flooded with 600 Al of buffer C (25 mm Tris'HCl/100 mm sodium pyrophosphate/100 mm NaF/250 mm NaCI/10 mm EGTA/5 mm EDTA/1% Nonidet P-40/0.2 mm phenylmethylsulfonyl fluoride/0.5 mm benzamidine/10,g of aprotonin per ml/10,ug of leupeptin per ml/10,g of soybean trypsin inhibitor per ml, ph 8.8) scraped up with a rubber policeman, made 10 mm with respect to disodium ATP, sonicated in a bath sonicator, and incubated on ice for 20 min. The insoluble particulate material was sedimented at 80,000 x g for 10 min in a Beckman TL-100 ultracentrifuge. The soluble EC extracts were incubated with an IgG fraction of rabbit antiserum raised against human platelet myosin and generously provided by Robert Adelstein (National Institutes of Health), for 2 hr at 4 C prior to the addition of prewashed protein A-Sepharose 4B. After an additional 2-hr incubation, the immune complexes bound to protein A were collected by centrifugation for 5 min at 12,800 x g. The pellets were washed twice with buffer C, centrifuged through a 10% sucrose cushion, and washed once in DPBS. The pellets were boiled in SDS sample buffer (17) and subjected to electrophoresis on a 5-15% polyacrylamide gradient. To establish that the 200-kDa protein identified by Coomassie blue staining was the myosin heavy chain, immunoprecipitated proteins were transferred onto nitrocellulose paper from 7% polyacrylamide gel, and the identity of the heavy chains was confirmed by using rabbit antibodies raised against human platelet myosin. SDS/Polyacrylamide Gel Electrophoresis. Gel electrophoresis was carried out in either 5-15% or % gradient vertical slab gels by the buffer system of Laemmli (17). Gels were stained with Coomassie blue, destained, and then incubated in EN3HANCE (DuPont) for 60 min. Gels were exposed to Kodak X-OMAT x-ray film and developed with a Kodak automatic processor. For estimation of molecular mass, either Bio-Rad high molecular mass protein standards or Pharmacia low molecular weight calibration standards were used. Purification of Proteins. Smooth muscle myosin light chains and MLCK were prepared from bovine uterus according to the procedures of Adelstein and Klee (19). For isolation of MLCK, we modified the isolation protocol by eliminating the gel filtration column and applying the dialyzed 40-60%o ammonium sulfate fraction to a column (2.5 x 30 cm) of DEAE-Sephacryl equilibrated with 20 mm Tris-HCl, ph 7.8/1 mm EGTA/1 mm EDTA/1 mm dithiothreitol/0.1 mm phenylmethylsulfonyl fluoride. The column was eluted at a flow rate of 50 ml/hr with a 1200-ml linear NaCl gradient ( mm) made up in equilibration buffer. Fractions that contained myosin kinase activity were pooled and enough CaC12 and MgSO4 were added to give a final free concentration of 0.5 mm Ca2' and 2 mm Mg2+. The pooled fractions were applied to a 12-ml column of calmodulin-sepharose (Sigma) equilibrated with 40 mm Tris HCl, ph 7.4/50 mm NaCI/0.5 mm CaC12/2 mm MgCI2/1 mm dithiothreitol. The column was washed with equilibration buffer containing 0.2 M NaCl. The kinase was then eluted with equilibration buffer containing 2 mm EGTA and 0.2 M NaCl. The peak ofmlck activity from the calmodulin affinity column was pooled, and Proc. Nati. Acad. Sci. USA 87 (1990) 17 CaCI2 was added to give a final free concentration of 0.5 mm CaC12. The kinase was then mixed with phenyl-sepharose preequilibrated with 40 mm Tris HCI, ph 7.4/50 mm NaCI/ 0.5 mm CaCl2/1 mm dithiothreitol for 1 hr at 40C. The phenyl-sepharose (Sigma) suspension was centrifuged at 1000 x g for 10 min, and the soluble MLCK was removed and dialyzed overnight against three 6-liter changes of 20 mm Tris HCI, ph 7.4/50 mm NaCI/1 mm EGTA/1 mm dithiothreitol. The purified kinase was stored at -70'C in 0.1-ml aliquots. Assay for MLCK Activity. MLCK activity was determined by incubation of 0.1 ml of column fractions with 0.1 ml of 25 mm Tris-HCI, ph 7.3/5 mm MgC12/50 mm KCI/0.2 mm CaC12 excess over the EGTA and EDTA present/5,g of calmodulin (Biomedical Technologies, Stoughton, MA) per ml/0.5 mg of isolated myosin light chains per ml/ly-32pjatp sufficient to give 300,000 cpm per tube. Column fractions were incubated at 25 C for 10 min and the assay was terminated by addition of 50%o trichloroacetic acid/109o sodium pyrophosphate to give a final concentration of 10% trichloroacetic acid and 2% sodium pyrophosphate. Samples were heated to 90 C for 20 min, cooled on ice and filtered through Schliecher and Schuell 0.45-,m filters, washed with 5% trichloroacetic acid/1% sodium pyrophosphate, dried, and counted in a Beckman LS1801 scintillation counter to assess phosphate incorporation into myosin light chains. The activity measured was demonstrated to be Ca2+ dependent. Cell Extraction and Reconstitution With Purified MLCK. Permeabilized EC monolayers were extracted in buffer D (40 mm Tris-HCI/100 mm KCI/30 mm MgCI2/1 mm dithiothreitol/5 mm EGTA/1 mm EDTA/5,ug of aprotonin per ml/5,ug of leupeptin per ml/10,g of soybean trypsin inhibitor per ml/0.2 mm phenylmethylsulfonyl fluoride/0.5 mm benzamidine, ph 7.5) for 30 min to remove MLCK. Extracted monolayers were reconstituted with purified MLCK and calmodulin in buffer B as described below for 10 min prior to the addition of ATP or ATP[y-35S]. RESULTS Cultures permeabilized with 25,ug of saponin per ml exhibited a complex pattern of actin filaments identifiable by rhodamine phalloidin staining and closely resembling that in intact cells. The dense peripheral band of actin was present but had lost some of its definition. Stress fibers of the cells, like the complex arrangement of paranuclear filamentous actin, were little affected by permeabilization. Exposure of monolayers to either Ca2" or ATP alone (Fig. 1) resulted in neither cell retraction nor rearrangement of F actin. Permeabilized EC monolayers exposed to both 100 AM Na-ATP and 100,uM free Ca2+ for 10 min in buffer B containing 2 mm Mg2+ and then fixed and stained with rhodamine phalloidin exhibited extensive cell retraction. The actin stress fibers and the paranuclear array were seen to have contracted centrally, leaving only a few strands radiating outward (Fig. 2). Parallel changes occur in intact monolayer exposed to platelet activating factor (Fig. 3). Having demonstrated that permeabilized EC monolayers were responsive to ATP and Ca2, we sought to determine whether stimulation of EC retraction was accompanied by phosphorylation of myosin light chains. Fig. 4 shows the thiophosphorylation pattern from a SDS/polyacrylamide gel analysis of immunoprecipitated EC myosin. EC cultures incubated with ATP[y-35S] resulted in thiophosphorylation of myosin light chains in a Ca2"-dependent manner. No thiophosphorylation of the 19-kDa myosin light chains occurred in the presence of EGTA (lane 1), and the extent of thiophosphorylation decreased with decreasing Ca2" concentration (lanes 2-4).

3 18 Medical Sciences: Wysolmerski and Lagunoff Proc. Natl. Acad. Sci. USA 87 (1990) FIG. 1. Fluorescence micrograph of F-actin filament distribution in control permeabilized EC monolayer stained with rhodamine phalloidin. Cells were incubated in buffer B containing 100 AM Na-ATP for 10 min. No EC retraction results in the absence of Ca2". An accumulation of F actin delineates the cell margins; only an occasional gap between adjacent cells is present within the monolayer. The stress fibers and paranuclear array remain intact. (x380.) To define the mechanism involved in EC retraction, we further investigated the role of MLCK. Cultures were extracted in buffer D for 30 min at 370C to remove MLCK activity. The extraction medium was assayed and found to contain MLCK activity (data not shown). Extraction of cultures resulted in only minor changes in the cytoskeleton. Neither cell rounding nor detachment from the culture dish was observed; however, occasional small gaps were evident between adjacent cells. Extracted monolayers showed some modest loss of F actin staining from the cell periphery, whereas the central microfilament bundles were unaltered. In addition, extraction with buffer D resulted in minimal loss of EC myosin (Fig. 5). When extracted monolayers were incubated with ATP, Ca2", and calmodulin for as long as 20 min, no retraction was evident (Fig. 6). Extracted EC reconstituted with purified MLCK and calmodulin for 10 min at 370C retracted when exposed to ATP and Ca2+ in buffer C (Fig. 7). The F-actin network assumed a collapsed pattern around the nucleus. Permeabilized, extracted ECs reconstituted with the following combinations-mlck and Ca2, MLCK and calmodulin, and calmodulin and Ca2+-and then incubated with ATP failed to retract. FIG. 3. Fluorescence micrograph of F-actin distribution in intact EC monolayers incubated with 20,M platelet activating factor. Severe EC retraction comparable to that seen in permeabilized EC monolayers treated with ATP and Ca2+ occurs within 15 min with loss of the prominent dense peripheral band of actin filament. The F-actin cytoskeleton retracts centrally, leaving cells attached by numerous slender filamentous processes. Neither cell rounding nor detachment from the culture dish was noted. (x380.) In parallel with the studies of reconstitution of retraction in extracted monolayers, we investigated thiophosphorylation of myosin light chains. Fig. 8A is a representative autoradiogram of the thiophosphorylation pattern of extracted and reconstituted monolayers. Cultures incubated with ATP[y- 35S] for 10 min showed no thiophosphorylation of myosin light chains after a 30-min extraction with buffer D (lane 4). Reconstitution of extracted monolayers with MLCK and FIG. 2. BC monolayer exposed to 100,uM Na-ATP, 100,uM Ca2+ free in buffer B for 10 min. Severe EC retraction of the stained F-actin cytoskeleton occurs within 10 min without cells detaching from the substratum. The actin filaments retract toward the nucleus. (x380.) FIG. 4. Autoradiogram of immunoprecipitated thiophosphorylated myosin light chains of permeabilized ECs. Lanes 2-4, thiophosphorylation of the 19-kDa myosin light chains in the presence of decreasing free Ca2+ concentration is shown. The apparent increase in thiophosphorylation of myosin light chain was confirmed by comparing the ratio of Coomassie blue-stained myosin heavy chains to thiophosphorylated light chains. Lanes: 1, 1 mm EGTA; 2, 100,M Ca2+; 3, 50 AM Ca2+; 4, 1,M free Ca2+. Numbers on right are kda.

4 Medical Sciences: Wysolmerski and Lagunoff FIG. 5. Radioimmunoblot of immunoprecipitated EC myosin. Immunoprecipitated myosin from 1.5 x 106 ECs was electrophoresed on 7% polyacrylamide gels electrophoretically transferred to nitrocellulose paper and incubated with rabbit anti-human platelet myosin antibodies. Nitrocellulose blots were then incubated with goat anti-rabbit 1251-labeled IgG to visualize myosin heavy chains. Lanes: 1, immunoprecipitated myosin from intact EC monolayers; 2, myosin immunoprecipitated from saponin permeabilized monolayers; 3, myosin immunoprecipitated from EC monolayer extracted with buffer D for 30 min. Extraction procedures used to extract MLCK did not result in significant loss of endogenous EC myosin. calmodulin reestablished the thiophosphorylation of the 19- kda myosin light chains in the presence of calcium (lane 5). Extracted monolayers incubated with either MLCK and Ca2+ or MLCK and calmodulin did not thiophosphorylate myosin light chains (lanes 6 and 7). To establish that the 19-kDa protein thiophosphorylated was indeed myosin light chains, we analyzed immunoprecipitated myosin from both extracted and reconstituted monolayers. Figure 8B shows the thiophosphorylation pattern of immunoprecipitated myosin from extracted (lane 1) and reconstituted (lane 2) monolayers. Cultures extracted with buffer D exhibited minimal thiophosphorylation of myosin light chains, while reconstitution with MLCK, Ca2+, and calmodulin reestablished the thiophosphorylation of the 19- kda myosin light chains (lane 2). DISCUSSION In 1969 Majno et al. (20), proposed that active contraction of endothelium involving cytoskeletal elements was responsible for the permeability edema induced by histamine. Our own previous studies have shown that agents that cause increased vascular permeability in vivo induce a rearrangement of the actin filaments of the EC cytoskeleton in vitro with retraction of the cells (2). We have extended these studies to show that EC retraction and accompanying cytoskeletal changes require permissive levels of ATP (3). A considerable body of FIG. 6. F-actin distribution in MLCK extracted monolayer. Cells were extracted for 30 min in buffer D and then exposed to 100,uM ATP/100,uM Ca2/5,ug of calmodulin per ml/for 10 min. There is some loss of peripheral F actin compared to that in unextracted permeabilized ECs, but there is no evidence of retraction or substantial loss of F actin. (x380.) _w-.~~~~.4 ~~~~97 Proc. Natl. Acad. Sci. USA 87 (1990) 19 FIG. 7. Fluorescence micrograph of MLCK reconstituted monolayer. Retraction is reestablished when extracted monolayers are incubated with 1,uM MLCK/100,M free Ca2+/5,ug of calmodulin per ml and are then exposed to 100,uM ATP in buffer B. There is severe EC retraction and F-actin filaments have contracted down around the nucleus. No cells have detached from the substratum during this process. (x380.) work has accumulated that suggests that in nonmuscle cells as in smooth muscle, myosin light-chain phosphorylation 67 * O.30. _ 4~~~~ _ A B 1 2 FIG. 8. (A) Autoradiogram of thiophosphorylated EC proteins. Lane 1, permeabilized monolayer incubated in the presence of 1 mm EGTA/100,uCi of ATP[y-35S]. No thiophosphorylation of myosin light chains occurs in the absence of Ca2+. Lane 2, ECs incubated in the presence of 100,uM free Ca2+/100,uCi of ATP[y-35S] result in thiophosphorylation of the 19-kDa myosin light chains (arrowhead). Monolayers extracted with buffer D for 15 min (lane 3) or for 30 min (lane 4) lose the ability to thiophosphorylate myosin light chains upon exposure to ATP[y-35S], Ca2+, and calmodulin. Extracted ECs incubated with 1 AM MLCK/100,M free Ca2/5A,ug of calmodulin per ml (lane 5) result in thiophosphorylation of endogenous myosin light chains upon exposure to ATP[y-35S]. Extracted monolayers reconstituted with MLCK/calmodulin (lane 6) or MLCK/Ca2+ but no calmodulin (lane 7) alone and then exposed to ATP[y-35S] fail to thiophosphorylate myosin light chains. Numbers on right are kda. (B) The protein thiophosphorylated by the addition of exogenous MLCK to permeabilized MLCK extracted monolayers was identified as myosin light chains by immunoprecipitation. In the absence of endogenous MLCK, there is only weak thiophosphorylation of myosin light chains (lane 1). Monolayers were reconstituted with 1,uM MLCK/ATPEy-35S]/100 A4M free Ca2+/5,ug ofcalmodulin per ml exhibited strong thiophosphorylation of the identified myosin light chains (lane 2).

5 20 Medical Sciences: Wysolmerski and Lagunoff plays a major role in regulation of actin-myosin-based contraction. This regulation of myosin function is mediated by the Ca2l-activated, calmodulin-dependent enzyme MLCK (6, 7, 9). This is an obvious model for EC retraction; calcium-dependent phosphorylation of myosin could initiate the sequence of events modifying the actin network, which in turn could account for the formation of retractive gaps between cells and increased vascular permeability. In the present study, we performed a series of experiments to evaluate the role of myosin phosphorylation in EC retraction. To carry out these studies, we developed a permeabilized cell preparation using saponin. Since it was difficult to delineate cell margins in the permeabilized cells by phase-contrast microscopy, we assessed retraction in terms of the change in cytoskeletal F actin. Addition of ATP and Ca2+ in the presence of Mg2+ induced retraction of permeabilized EC cytoskeleton; the extent of retraction was dependent on both the Ca2+ and ATP concentrations. Labeling studies with ATP[y-35S] showed that numerous EC proteins were thiophosphorylated. A prominent band of 19-kDa was thiophosphorylated in a Ca2+-dependent manner. This band was positively identified as the myosin light chain by electrophoresis of immunoprecipitated myosin (Fig. 4). Cells in monolayers extracted with high salt concentration did not retract upon exposure to ATP and Ca2W, and thiophosphorylation of myosin light chains was correspondingly absent in these preparations. Addition of calmodulin did not restore either retraction or thiophosphorylation in the presence of ATP and Ca2+, nor did addition.of purified MLCK without calmodulin restore retraction or thiophosphorylation. Addition of MLCK and calmodulin together restored both retraction (Fig. 7) and thiophosphorylation (Fig. 8). This is the behavior predicted based on the events in smooth muscle contraction. Several nonmuscle cells exhibit a similar set of control mechanisms in which MLCK is believed to play a pivotal role in the initiation of contractile activity (6, 7, 9). We have demonstrated the presence of a MLCK controlled contractile mechanism in ECs; the next step is to prove that Proc. Natl. Acad. Sci. USA 87 (1990) this mechanism is the one invoked by histamine and similarly acting agents in intact cells. This work was supported by National Institutes of Health Grant Specialized Center of Research Adult Respiratory Failure HL Majno, G. & Palade, G. E. (1%1) J. Biophys. Biochem. Cytol. 11, Wysolmerski, R. B. & Lagunoff, D. (1985) Am. J. Pathol. 119, Wysolmerski, R. B. & Lagunoff, D. (1988) Am. J. Pathol. 132, Laposata, M., Dovnarsky, D. K. & Shen, H. S. (1982) Blood 62, Shasby, M. D., Shasby, S. S., Sullivan, J. M. & Peach, M. J. (1982) Circ. Res. 51, Sellers, J. R. & Adelstein, R. S. (1987) in The Enzymes, eds. Boyer, P. D. & Krebs, E. G. (Academic, Orlando, FL), Vol. 18, pp Stull, J. T., Nunnally, M. H., Mecknoff, C. H. (1986) in The Enzymes, eds. Boyer, P. D. & Krebs, E. G. (Academic, Orlando, FL), Vol. 17, pp Kargacin, G. J. & Fay, F. S. (1987) J. Gen. Physiol. 90, Kamm, K. E. & Stull, J. T. (1985) Annu. Rev. Pharmacol. Toxicol. 25, Holzapfel, G., Wehland, J. & Weber, K. (1983) Exp. Cell Res. 148, Masuda, H., Owaribe, K., Hayashi, H. & Hatano, S. (1984) Cell Motil. 4, Broschat, K. O., Stidwill, R. P. & Burgess, D. R. (1983) Cell 35, Porrello, K. & Burnside, B. (1984) J. Cell Biol. 98, Del Vecchio, P. J. & Smith, J. R. (1981) J. Cell. Physiol. 108, Bers, D. M. (1982) Am. J. Physiol. 242, C404-C Barak, L. S., Yocum, R. R., Nothnagel, E. A. & Webb, W. W. (1980) Proc. Natl. Acad. Sci. USA 77, Laemmli, U. K. (1970) Nature (London) 227, Kawamoto, S. & Adelstein, R. S. (1988) J. Biol. Chem. 263, Adelstein, R. S. & Klee, C. B. (1981) J. Biol. Chem. 256, Majno, G., Shea, S. M. & Leventhal, H. (1969) J. Cell Biol. 42,