Cross-Regulation of Intracellular cGMP and cAMP in ... - Science Direct

Molecular Cell Biology Research Communications 4, 10 –14 (2000) doi:10.1006/mcbr.2000.0249, available online at http://www.idealibrary.com on

Cross-Regulation of Intracellular cGMP and cAMP in Cultured Human Corpus Cavernosum Smooth Muscle Cells Noel N. Kim,* Yue-hua Huang,* Robert B. Moreland,* Sandra S. Kwak,* Irwin Goldstein,* and Abdulmaged Traish* ,† *Department of Urology and †Department of Biochemistry, Boston University School of Medicine, Boston, Massachusetts 02118

Received August 29, 2000

nosum smooth muscle relaxation in vitro and in vivo (1– 4). Nitric oxide (NO), a nonadrenergic, noncholinergic neurotransmitter and autocoid, plays a critical role in attenuating penile cavernosal smooth muscle contraction and inducing smooth muscle relaxation (1– 4). Activation of neuronal (Type I) and endothelial (Type III) nitric oxide synthases results in the conversion of L-arginine and molecular oxygen to citrulline and NO. NO freely diffuses into smooth muscle cells and binds to the heme component of soluble guanylyl cyclase, stimulating cyclic guanosine monophosphate (cGMP) synthesis. Binding of cGMP to cGMP-dependent protein kinases (PKG) or cGMP-activated ion channels results in reduction of intracellular calcium, via calcium sequestration and extrusion, and activation of myosin light chain phosphatases, causing diminution of smooth muscle contractility and enhancing penile erection. The endogenous vasodilator prostaglandin E (PGE) may also play an important role in regulating cavernosal smooth muscle tone. Activation of specific G-protein coupled PGE receptors (EP) on the smooth muscle cell stimulates adenylyl cyclase and the production of cAMP (5– 8). The role of cAMP-dependent protein kinases (PKA) in vascular smooth muscle remains unclear relative to the regulation of intracellular calcium levels. However, cAMP has been shown to activate PKG and the calcium lowering effects of forskolin are not observed in the absence of PKG (9 –12). Thus, PKG has been proposed to be a common mediator of smooth muscle relaxation in response to both cAMP and cGMP. Intracellular cyclic nucleotide concentrations are determined by the activities of their respective cyclases or phosphodiesterases (PDE’s). Termination of signal transduction by hydrolysis of cGMP or cAMP depends on the expression of the specific phosphodiesterase iso-

The goal of this study was to assess the potential cross-regulation of cyclic nucleotides in human corpus cavernosum (HCC). Incubation of primary cultures of HCC smooth muscle cells with either the NO donor sodium nitroprusside (SNP, 10 ␮M) or the phosphodiesterase type 5 (PDE 5) inhibitor sildenafil (50 nM) produced little or no changes in the intracellular cGMP levels. Incubation with both SNP and sildenafil produced marked increases in cGMP. Interestingly, incubation of cells with 10 ␮M of forskolin or PGE 1 produced significant enhancement of cGMP accumulation. These increases were not further enhanced by the addition of SNP and sildenafil. Kinetic analyses of cGMP hydrolysis by PDE 5 showed that high concentrations of cAMP reversibly inhibited the enzyme with a K i of 258 ⴞ 54 ␮M. The increase in cGMP levels in response to cAMP generating agents is not due to assay artifact since cAMP did not cross-react with cGMP antibody. Our data suggest that cAMP up-regulates intracellular levels of cGMP, in part, by inhibition of PDE 5. We also noted that cGMP down-regulates cAMP synthesis via a mechanism requiring G-protein coupling of adenylyl cyclase. These observations may have important implications in the utility of pharmacotherapeutic agents targeting cyclic nucleotide metabolism for the treatment of erectile dysfunction. © 2000 Academic Press

The tone of the vascular smooth muscle in penile arteries and erectile tissue (corpus cavernosum) regulates the state of penile tumescence. The contractility of the smooth muscle within penile erectile tissue is modulated by various neurotransmitters and locally produced vasoactive substances (1). Endogenous factors or exogenous compounds that increase either intracellular cAMP or cGMP levels induce corpus caver1522-4724/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.

10

Vol. 4, No. 1, 2000

MOLECULAR CELL BIOLOGY RESEARCH COMMUNICATIONS

50 nM of sildenafil, 15 min prior to the addition of other agents. Incubations were terminated by quickly aspirating the media and adding 1.5 ml of ice-cold 1 N perchloric acid. Cells were scraped and all material was collected in microfuge tubes and processed for radioimmunoassay, as described below.

forms in the target tissue. The predominant phosphodiesterases in penile corpus cavernosum tissue are types 2, 3 and 5, with PDE 5 being responsible for most of the cGMP specific hydrolytic activity (13, 14). PDE 2 is non-specific and can break down both cAMP and cGMP, whereas PDE 3 is preferential for cAMP (9, 15, 16). Cyclic GMP enhances the cAMP hydrolytic activity of PDE 2 but inhibits cAMP hydrolysis by PDE 3 (9, 15, 16). The cross-activation of cyclic nucleotide-dependent protein kinases and the presence of multiple phosphodiesterases and their overlapping regulation serves as a means of signal integration and regulation amongst many different endocrine, paracrine and autocrine substances. However, this type of regulation remains largely undefined in penile corpus cavernosum tissue. Given the importance of cyclic nucleotide signalling in this tissue, we initiated this study to investigate the potential cross-regulation of cGMP and cAMP pathways in human corpus cavernosum smooth muscle cells.

Cellular extract preparation and determination of intracellular cyclic nucleotide levels. Complete cell disruption was accomplished by repeatedly passing samples through a 23-gauge needle. Samples were centrifuged at 3000g for 20 min at 4°C. Pellets were solubilized in 1 ml of 1 N NaOH at 90°C for 30 min and used to determine total cellular protein by Lowry assay. The supernatants were combined with 0.9 ml of 2N KOH in clean tubes, left on ice for 10 min and centrifuged to remove the resulting precipitate. The deproteinized, neutralized supernatant was removed and stored at ⫺20°C for later analysis of cyclic nucleotides (cAMP and cGMP). Cyclic nucleotide levels were determined by a commercially available radioimmunoassay (RIA; Biomedical Technologies, Inc., Stoughton, MA). Samples were thawed and diluted with buffer (supplied with RIA kits) 4- to 10-fold for cGMP or 100to 1000-fold for cAMP determinations. Typically, 50 ␮l aliquots of the standards and unknowns were assayed in a final incubation volume of 150 ␮l and processed according to the protocols supplied by the manufacturer.

MATERIALS AND METHODS Materials. Sildenafil (1-[4-ethoxy-3-(6,7-dihydro-1methyl-7-oxo-3-propyl-1H-pyrazolo [4,3-d] pyrimidin5-yl) phenylsulphonyl]-4-methyl-piperazine), a selective PDE type 5 inhibitor, was provided by Pfizer. Forskolin and phenylmethylsulfonyl fluoride (PMSF) were purchased from Sigma Chemical Co. (St. Louis, MO). Prostaglandin E 1 was purchased from Cayman Chemical Co. (Ann Arbor, MI). Forskolin stock solutions were made in DMSO and prostaglandin E 1 stock solutions were made in 100% ethanol. [ 3H]cGMP was obtained from Amersham Corp. (Arlington Heights, IL). Tissue culture media and antibiotics were purchased from GIBCO-BRL (Grand Island, NY) and fetal bovine serum was purchased from Summit Biotechnologies (Fort Collins, CO). All other general reagents were of analytical grade from commercial sources.

Preparation of human corpus cavernosum smooth muscle cell extracts for phosphodiesterase assay. Cells were washed with ice cold physiological salt solution, scraped and the cells were collected by centrifugation at 500g, for 10 min at 2°C. Cell pellets were washed in phosphate buffered saline and homogenized using a glass/glass homogenizer in 20 mM HEPES containing 1 mM EDTA, 1 mM PMSF, 250 mM sucrose, pH 7.20, at 2°C. The total cell extract was centrifuged at 105,000g for 60 min at 2°C. The high-speed supernatant was removed, and either used immediately or stored frozen at ⫺75°C until experimentation. Assay of phosphodiesterase activity. Phosphodiesterase activity was assessed as previously described (14). Aliquots (50 ␮l, in triplicate) of cytosol were incubated at 30°C for 30 min in the presence of unlabeled cGMP (0.5 or 1 ␮M) and [ 3H]cGMP (50 nM) in 40 mM MOPS buffer, containing 1 mM EDTA, pH 7.20 at 4°C, 0.8 mM EGTA, 5 mM Mg acetate, 0.2 mg/ml BSA with or without the indicated concentrations of sildenafil or cAMP, in a final volume of 250 ␮l. Parallel incubations were made in the absence of cytosol as control (blank). The reactions were terminated by incubation at 100°C for 1 min to inactivate the enzyme and all samples were incubated with 2.5 ␮l of 10 mg/ml Crotalus atrox venom to hydrolyze GMP to guanine for 10 min at 30°C. Deionized water (0.5 ml) was added to each sample and the total incubation mixture was chromatographed on a column of DEAE Sephadex A-25 (0.75 ml

Culture of human penile corpus cavernosum smooth muscle cells. Primary cultures of human penile corpus cavernosum smooth muscle cells were established by seeding tissue explants in 6-well plates with D-MEM (supplemented with 10% fetal bovine serum, penicillin, streptomycin and nystatin), as previously described (17). Cells were propagated in 75 cm 2 flasks (Costar, Cambridge, MA) and used at early passage (P1-P3) for all experiments. Treatment of human corpus cavernosum smooth muscle cells with vasoactive agents. Cells were grown to confluence in 6-well plates and media was replaced with fresh media 24 h prior to treatment. Cells were treated for 5 min at 37°C with 10 ␮M of sodium nitroprusside, prostaglandin E 1, forskolin or a combination of these agents. In some wells, cells were treated with 11

Vol. 4, No. 1, 2000

MOLECULAR CELL BIOLOGY RESEARCH COMMUNICATIONS

FIG. 2. Specificity of cGMP radioimmunoassay. Increasing concentrations of cAMP (0.0025– 0.500 pmol) were assayed using a cGMP radioimmunoassay kit.

FIG. 1. Effect of vasodilator agents on cGMP production. Human corpus cavernosum smooth muscle cells were incubated with the phosphodiesterase type 5 inhibitor sildenafil citrate (SIL; 50 nM), sodium nitroprusside (NP; 10 ␮M), forskolin (FSK; 10 ␮M), prostaglandin E 1 (PGE; 10 ␮M) or a combination of these agents, as indicated. Cyclic GMP was quantified in de-proteinized cell extracts by a commercially available radioimmunoassay. Data are the average of 5 independent experiments.

cAMP cross-reacting with anti-cGMP antibody, we performed the cGMP radioimmunoassay with increasing amounts of cAMP. As shown in Fig. 2, cAMP (2.5–500 fmol) did not alter the binding of [ 125I]cGMP to the anti-cGMP antibody, whereas non-radioactive cGMP effectively inhibited immunoprecipitation of [125I]cGMP. Effect of cAMP on cGMP hydrolysis. To examine if cAMP can directly affect cGMP hydrolysis, we incubated human corpus cavernosum cell extracts with cGMP (0.5 and 1 ␮M) in the presence of varying concentrations of cAMP. Only sildenafil sensitive cGMP hydrolysis was assessed to isolate phosphodiesterase type 5 activity. As shown in Fig. 3A, kinetic analyses indicated that cAMP inhibits hydrolysis of cGMP with a mean K i value of 258 ⫾ 54 ␮M (n ⫽ 3). When cGMP hydrolysis was assessed in cell extracts with varying concentrations of sildenafil and a fixed concentration of cAMP (250 ␮M), the inhibition constant of sildenafil to phosphodiesterase type 5 (Fig. 3B) was not significantly altered, relative to previously published K i val-

bed volume), preequilibrated with 20 mM Tris–HCl, pH 7.5. Hydrolyzed guanine was eluted with 3 ml of 20 mM Tris–HCl buffer directly into scintillation vials. Samples were mixed with 10 ml of Liquiscint (National Diagnostics, Atlanta, GA) and the radioactivity was quantified by liquid scintillation spectroscopy. The radioactivity representing spontaneous cGMP hydrolysis determined from the control (blank) incubation was considered background activity and was subtracted from radioactivity determined in each sample. The concentration of cGMP hydrolyzed was determined for each sample, normalized for isotope dilution and protein concentration, and the corrected data were subjected to Dixon analyses. RESULTS Effects of guanylyl and adenylyl cyclase activators on intracellular levels of cGMP. Treatment of human corpus cavernosum smooth muscle cells with the nitric oxide donor sodium nitroprusside (10 ␮M) alone or the phosphodiesterase type 5 inhibitor sildenafil (50 nM) alone resulted in no significant changes in cGMP levels (Fig. 1). Nitroprusside, in the presence of sildenafil, elevated cGMP levels 2.8-fold. In the absence of phosphodiesterase inhibitors, forskolin (10 ␮M) alone or prostaglandin E 1 (10 ␮M) alone, stimulated cGMP levels 2.4 and 4.7 fold, respectively. When cells were treated with nitroprusside and forskolin, in the presence of 50 nM sildenafil, cGMP levels also increased 4.2-fold. Similar results were obtained when cells were treated with nitroprusside and PGE 1 in the presence of sildenafil. To rule out the possibility that the increase in cGMP levels induced by forskolin or PGE 1 was not due to

FIG. 3. Inhibition of cGMP hydrolytic activity by cAMP. Cytosolic extracts from human corpus cavernosum smooth muscle cells were incubated with two different concentrations (0.5 and 1 ␮M) of [ 3H]cGMP and increasing concentrations of unlabeled cAMP (A). In parallel experiments, cell extracts were also incubated with varying concentrations of sildenafil in the presence of 250 ␮M cAMP (B). Specific hydrolysis of cGMP was assessed and the data were subjected to Dixon analyses. Shown is a representative experiment (n ⫽ 3). 12

Vol. 4, No. 1, 2000

MOLECULAR CELL BIOLOGY RESEARCH COMMUNICATIONS

manner via cyclic nucleotide-dependent protein kinase modulation of enzyme activity. In the present study, we have demonstrated that cAMP inhibits sildenafilsensitive cGMP hydrolysis in a competitive manner. This suggests that cAMP may increase cGMP levels by inhibiting PDE 5. The major PDE types (2, 3, and 5) present in human cavernosal tissue are not known to be regulated by cAMP. In conjunction with the lack of effect of cAMP on the inhibition constant for sildenafil, Dixon analyses of our data indicate that cAMP does not interact with PDE 5 in an allosteric fashion. While we have not directly investigated the possibility of PKA or PKG-mediated alterations in PDE 5 activity, Corbin et al. have recently demonstrated that phosphorylation of PDE 5 by cyclic nucleotide-dependent protein kinases actually enhances cGMP hydrolytic activity (19). This would argue against the possibility of protein kinase-mediated inhibition of cGMP hydrolysis. The effects of cAMP on guanylyl cyclase activity remain to be investigated. Thus, in corpus cavernosum smooth muscle, inhibition of cGMP hydrolysis may be a major, if not exclusive, mechanism by which cAMP elevates cGMP levels. We have also observed that progressive elevation of NO-stimulated cGMP levels through inhibition of PDE 5 results in decreased PGE 1-induced cAMP levels. However, in parallel experiments, PDE 5 inhibition caused an increase in forskolin-induced cAMP levels. This reciprocal regulation of cAMP by cGMP has been reported previously and mainly attributed to the influence of cGMP on PDE 3. In peripheral blood vessels, numerous lines of evidence suggest that cGMP inhibits PDE 3 to elevate cAMP and potentiate vasodilation in response to cAMP generating agents (9, 20, 21). In further support of this mechanism, nitrovasodilators have been observed to synergistically enhance the effects of cAMP on platelet aggregation and degranulation (22, 23). In addition, the effects of nitrovasodilators in both vascular smooth muscle and platelets were identical to the effects of cAMP phosphodiesterase inhibitors (21, 22). Recently, Stief et al. have shown that moderate to high concentrations of sildenafil (0.01–10 ␮M) cause elevations in cAMP levels in both penile cavernosal and cardiac atrial tissue (24). The authors also attributed this effect to the inhibition of PDE 3 by cGMP. This mechanism would explain our results with forskolin, but are contrary to the data obtained with PGE 1 treatment. This divergence in results suggests the existence of multiple pathways of regulation in which cAMP can be regulated by cGMP. Opposing effects of guanylyl cyclase stimulation on cAMP levels have been reported in platelets. Dickinson et al. observed that nitroprusside synergistically enhanced cAMP levels at low doses (1 nM) of prostacyclin but antagonized cAMP accumulation at high doses (20 nM) of prostacyclin (25). EHNA, a selective inhibitor of PDE 2, potentiated

FIG. 4. Effect of phosphodiesterase type 5 inhibition on cAMP production. Human corpus cavernosum smooth muscle cells were incubated with the NO donor sodium nitroprusside (SNP) and either forskolin (FSK) or prostaglandin E 1 (PGE 1) in the presence of increasing concentrations of the specific PDE type 5 inhibitor sildenafil. Cyclic AMP was quantified in deproteinized cell extracts. Data are the average of 4 independent experiments.

ues. These data suggest that cAMP acts as a competitive inhibitor of phosphodiesterase type 5. Effects of phosphodiesterase type 5 inhibition on intracellular levels of cAMP. To determine if elevation of cGMP affects cAMP levels in human corpus cavernosum smooth muscle, we treated cells with increasing concentrations of sildenafil and a single concentration of nitroprusside. Constant adenylyl cyclase activation was accomplished with either prostaglandin E 1 or forskolin. Under these conditions, sildenafil caused a dose-dependent decrease in cAMP levels with prostaglandin E 1 stimulation, whereas a dose-dependent increase in intracellular cAMP levels was observed with forskolin stimulation (Fig. 4). DISCUSSION Relaxation of corpus cavernosum trabecular smooth muscle can be modulated by neurotransmitters and vasoactive agents, which promote synthesis or prevent hydrolysis of cAMP and cGMP. However, little is known about the interactions between these two important intracellular messengers in penile cavernosal smooth muscle. Our data indicate that both indirect and direct activators of adenylyl cyclase (PGE 1 and forskolin) can cause increases in intracellular cGMP levels in human corpus cavernosum smooth muscle. Similar observations have been made in rat cerebral arteries and in porcine palmar veins (9, 18), suggesting that elevations in cGMP due to adenylyl cyclase stimulation may be a phenomenon common to vascular smooth muscle cells. Several potential mechanisms may account for this regulation. Increased cAMP synthesis may lead to stimulation of guanylyl cyclase or to inhibition of phosphodiesterases that hydrolyze cGMP. These alterations of cyclase or phosphodiesterase activity could occur via direct interaction with cAMP or in an indirect 13

Vol. 4, No. 1, 2000

MOLECULAR CELL BIOLOGY RESEARCH COMMUNICATIONS

the increase in cAMP with 1 nM prostacyclin and attenuated the inhibition of cAMP with 20 nM prostacyclin. In addition to the effects of cGMP on PDE 3, the authors concluded that cGMP may also limit cAMP accumulation by stimulating PDE 2. Thus, the influence of cGMP on cAMP may depend on the relative levels of each cyclic nucleotide and the sensitivity of the phosphodiesterases present. However, even this mechanism does not fully explain the effects of sildenafil on PGE 1-induced cAMP accumulation in corpus cavernosum smooth muscle cells. We have previously demonstrated that the levels of cAMP generated in response to PGE 1 treatment are similar to or slightly greater than those produced in response to forskolin treatment in cultured human cavernosal smooth muscle cells (6). Thus, if the degree of cAMP accumulation determines the overall manner in which cGMP modulates its levels, we would expect to observe similar and not divergent effects between PGE 1 and forskolin treatment in our studies. The distinction between direct activation of adenylyl cyclase by forskolin and receptor-mediated G-protein activation of adenylyl cyclase by PGE 1 suggests that the point of regulation by cGMP is more proximal to the receptor than other distal events that would be common to forskolin and PGE 1 signalling. We postulate an alternative mechanism in which increased cGMP may decrease intracellular cAMP levels through desensitization of prostaglandin E (EP) receptors. This may occur via phosphorylation or inactivation of the EP receptor or Gs protein, resulting in a reduction in adenylyl cyclase activity. Indeed, Wang et al. have recently reported that cGMP-activated PKG can phosphorylate the carboxyl terminus of thromboxane A 2 receptors and inhibit IP 3 signalling and intracellular calcium mobilization in platelets (26). Furthermore, PKG has also been shown to phosphorylate the IP 3 receptor in intact rat aorta (27, 28). Whether cGMP causes the phosphorylation of EP receptors to alter their binding or signalling capacity in corpus cavernosum smooth muscle cells remains to be seen, but serves as a plausible mechanism consistent with our data. The successful and widespread use of the PDE 5 inhibitor sildenafil (Viagra) for the treatment of male erectile dysfunction highlights the utility of targeting cyclic nucleotide metabolism to modulate the contractility of penile erectile tissue. The existence of crossregulation of cyclic nucleotide signalling in penile corpus cavernosum smooth muscle may have important implications for existing or new pharmacotherapeutic strategies for the treatment of male erectile dysfunction.

REFERENCES 1. Andersson, K-E., and Wagner, G. (1995) Physiol. Rev. 75, 191– 236. 2. Soderling, S. H., and Beavo, J. A. (2000) Curr. Opin. Cell Biol. 12, 174 –179. 3. Adams, M. A., Banting, J. D., Maurice, D. H., Morales, A., and Heaton, J. P. W. (1997) Int. J. Impot. Res. 9, 85–91. 4. Christ, G. J., Richards, S., and Winkler, A. (1997) Int. J. Impot. Res. 9, 69 – 84. 5. Hedlund, H., and Andersson, K-E. (1985) J. Urol. 134, 1245– 1250. 6. Traish, A. M., Moreland, R. B., Gallant, C., Huang, Y-H., and Goldstein, I. (1997) Recept. Signal Transduct. 7, 121–132. 7. Pierce, K. L., Gil, D. W., Woodward, D. F., and Regan, J. W. (1995) Trends Pharmacol. Sci. 16, 1–22. 8. Moreland, R. B., Kim, N. N., Nehra, A., Parulkar, B. G., and Traish, A. (2000) Int. J. Impot. Res., in press. 9. Pelligrino, D. A., and Wang, Q. (1998) Prog. Neurobiol. 56, 1–18. 10. Komalavilas, P., and Lincoln, T. M. (1996) J. Biol. Chem. 271, 21933–21938. 11. Lincoln, T. M., Komalavilas, P., and Cornwell, T. L. (1994) Hypertension 23, 1141–1147. 12. Lincoln, T. M., and Cornwell, T. L. (1991) Blood Vessels 28, 129 –137. 13. Boolell, M., Allen, M. J., Ballard, S. A., Gepi-Attee, S., Muirhead, G. J., Naylor, A. M., Osterloh, I. H., and Gingell, C. (1996) Int. J. Impot. Res. 8, 47–52. 14. Moreland, R. B., Goldstein, I., and Traish, A. (1998) Life Sci. 62, PL309 –PL318. 15. Beavo, J. A. (1995) Physiol. Rev. 75, 725–748. 16. Rybalkin, S. D., and Beavo, J. A. (1996) Biochem. Soc. Trans. 24, 1005–1009. 17. Moreland, R. B., Traish, A. M., McMillin, M. A., Smith, B., Goldstein, I., and Saenz de Tejada, I. (1995) J. Urol. 153, 826 – 834. 18. Wright, I. K., Amirchetty-Rao, S., and Kendall, D. A. (1994) Br. J. Pharmacol. 112, 1146 –1150. 19. Corbin, J. D., Turko, I. V., Beasley, A., and Francis, S. H. (2000) Eur. J. Biochem. 267, 2760 –2767. 20. Pyne, N. J., Arshavsky, V., and Lochhead, A. (1996) Biochem. Soc. Trans. 24, 1019 –1022. 21. Maurice, D. H., Crankshaw, D., and Haslam, R. J. (1991) Eur. J. Pharmacol. 192, 235–242. 22. Maurice, D. H., and Haslam, R. J. (1990) Mol. Pharmacol. 37, 671– 681. 23. Bowen, R., and Haslam, R. J. (1991) J. Cardiovasc. Pharmacol. 17, 424 – 433. ¨ ckert, S., Becker, A. J., Harringer, W., Truss, M. C., 24. Stief, C. G., U Forssman, W-G., and Jonas, U. (2000) Urology 55, 146 –150. 25. Dickinson, N. T., Jang, E. K., and Haslam, R. J. (1997) Biochem. J. 323, 371–377. 26. Wang, G. R., Zhu, Y., Halushka, P. V., Lincoln, T. M., and Mendelsohn, M. E. (1998) Proc. Natl. Acad. Sci. USA 95, 4888 – 4893. 27. Komalavilas, P., and Lincoln, T. M. (1996) J. Biol. Chem. 271, 21933–21938. 28. Komalavilas, P., and Lincoln, T. M. (1994) J. Biol. Chem. 269, 8701– 8707.

ACKNOWLEDGMENT This work was supported by a grant from the National Institute for Diabetes, Digestive, and Kidney Diseases (K01 DK02696).

14