Ceramide mediates sustained contraction of smooth muscle cells. C2 ceramide induced a rapid increase in Src kinase activity within 15 s, peaked at 1 min, and was sustained up to 8 min. Contraction and Src kinase activity were inhibited in cells incubated in Ca2+-free medium containing 2 mM EGTA and in cells preincubated with herbimycin A, a Src kinase inhibitor. Immunoblotting using a phosphospecific anti-Src (416Y) antibody showed a ceramide-induced increase in pp60src tyrosine phosphorylation. Immunoprecipitation using an anti-phosphotyrosine antibody followed by Western immunoblotting using a monoclonal IgG anti-phosphoinositide 3-kinase NH2 terminal-SH2 domain antibody showed a ceramide-induced increase in phosphoinositide 3-kinase (PI 3-kinase) tyrosine phosphorylation at a protein mass corresponding to 85 kDa, the regulatory subunit of PI 3-kinase, which contains the Src kinase binding site. PI 3-kinase phosphorylation was inhibited by herbimycin A and by the PI 3-kinase inhibitors wortmannin and LY-294002. Preincubation of cells with herbimycin A or PI 3-kinase inhibitors also resulted in an inhibition of mitogen-activated protein (MAP) kinase p42 and p44 activities as seen on Western blots. In summary, we found that 1) the maintenance of sustained contraction is dependent on extracellular Ca2+;2) ceramide activates a nonreceptor tyrosine kinase pathway through activation of pp60src and PI 3-kinase; and3) the converging signals are probably through activation of MAP kinase.
- rabbit colon
- signal transduction
- phosphoinositide 3-kinase
we have recently shown that the sustained contraction induced by the peptide agonist bombesin is accompanied by an increase in sphingolipid-derived ceramide (16). We have proposed a model by which receptor activation of protein kinase C (PKC) results in activation of sphingomyelinase, thus producing ceramide. Ceramide produced in the cell acts as an intracellular messenger. Ceramide induces a sustained contraction of smooth muscle cells through a pathway involving the activation of mitogen-activated protein (MAP) kinase (16). Thus ceramides could be important mediators of contraction and could account for the sustained contraction observed in circular smooth muscle cells from the rabbit rectosigmoid. We report here that in smooth muscle cells of the rectosigmoid1) ceramide-induced contraction is Ca2+ dependent, through which the sustained contraction induced by ceramide is inhibited in the absence of Ca2+ in the extracellular medium (0 mM Ca2+ and 2 mM EGTA);2) ceramide also seems to activate a nonreceptor tyrosine kinase pathway through activation of pp60src and phosphoinositide 3-kinase (PI 3-kinase); 3) ceramide activation of Src is also Ca2+dependent; and 4) this nonreceptor tyrosine kinase pathway seems to be a member of a cascade of kinases resulting in activation of MAP kinase.
MATERIALS AND METHODS
Chemicals were obtained from the following sources. C2 ceramide, C2 dihydro-ceramide, and C6 ceramide were from Matreya (Pleasant Gap, PA); wortmannin and soybean trypsin inhibitor (SBTI) from Sigma Chemical (St. Louis, MO); herbimycin A from Biomol (Plymouth, PA); LY-294002 from Calbiochem (San Diego, CA); DMEM from GIBCO (Grand Island, NY); collagenase (CLS type II) from Worthington Biochemical (Freehold, NJ); monoclonal IgG anti-Src (pp60src) antibody, monoclonal IgG anti-PI 3-kinase NH2terminal-SH2 domain antibody, and Src kinase assay kit from Upstate Biotechnology (Lake Placid, NY); monoclonal IgG anti-phosphotyrosine antibody (PY99) from Santa Cruz (Santa Cruz, CA); phosphospecific rabbit polyclonal IgG anti-MAP kinase antibody and anti-Src (P416Y) antibody (gift sample) from New England Biolabs (Beverly, MA); goat anti-mouse IgG (heavy and light chain)-horseradish peroxidase conjugate and protein assay reagent from Bio-Rad (Hercules, CA); goat anti-rabbit IgG (heavy and light chain) peroxidase labeled antibody from Kirkegaard and Perry Laboratories (Gaithersburg, MD); protein G-sepharose from Pharmacia Biotech (Uppsala, Sweden); [γ-32P]ATP and enhanced chemiluminescence detection reagents from Amersham (Arlington Heights, IL); and Cytoscint from ICN (Costa Mesa, CA). All other reagents were obtained from Sigma Chemical.
Isolation of smooth muscle cells from rabbit rectosigmoid.
The internal anal sphincter, consisting of the most distal 3 mm of the circular muscle layer, ending at the junction of skin and mucosa, was removed by sharp dissection. A 5-cm length of the rectosigmoid orad to the junction was dissected and digested to yield isolated smooth muscle cells. Cells were isolated as previously described (1, 2, 21). The tissue was incubated for two successive 1-h periods at 31°C in 15 ml of HEPES buffer, pH 7.4, containing (in mM) 115 NaCl, 5.7 KCl, 2.0 KH2PO4, 24.6 HEPES, 1.9 CaCl2, 0.6 MgCl2, and 5.6 glucose containing 0.1% (wt/vol) collagenase (150 U/mg Worthington CLS type II), 0.01% (wt/vol) SBTI, and 0.184% (wt/vol) DMEM. At the end of the second enzymatic incubation period, the medium was filtered through 500-μm Nitex mesh. The partially digested tissue left on the filter was washed four times with 10 ml of collagenase-free buffer solution. Tissue was then transferred into 15 ml of fresh collagenase-free buffer solution, and cells were gently dispersed. After a hemocytometric cell count, the harvested cells were resuspended in collagenase-free HEPES buffer (pH 7.4) and diluted as needed. Each rectosigmoid yielded 10–20 × 106 cells.
Measurement of Src kinase activity.
Src kinase was measured by a radioenzyme assay using a Src kinase assay kit. The assay system is based on phosphorylation of a specific substrate peptide [p34cdc2-(6—20)] by the transfer of [γ-32P]ATP by Src kinase. Aliquots (1.0 ml each) of the smooth muscle cell suspension (2 × 106cells/ml) were incubated with reagents from 15 s to 20 min at 37°C. The incubation was stopped with 1 ml of chilled HEPES buffer. The suspensions were immediately centrifuged at 10,000 rpm for 50 s at 4°C in a Microfuge, and the supernatant was removed. The resultant pellet was resuspended in 30 μl chilled HEPES buffer (pH 7.4), containing 50 mM β-glycerophosphate, 25 mM NaF, 1% Triton X-100, 150 mM NaCl, 20 mM EGTA, 15 mM MgCl2 1 mM dithiothreitol (DTT), 25 μg/ml leupeptin, and 25 μg/ml aprotinin (extraction buffer). Each suspension (20–30 samples) was immediately frozen in liquid nitrogen and stored at −70°C overnight. The suspensions were thawed and sonicated for 30 s at 4°C. The sonicates were vortexed and allowed to settle for 10 min at 4°C and then centrifuged at 10,000 rpm for 15 min at 4°C. Each 10-μl sample of the supernatant (∼30 μg protein) was combined with 20 μl of substrate solution, which contained 10 μg Src substrate peptide [p34cdc2-(6—20): Lys-Val-Glu-Lys-Ile-Gly-Glu-Gly-Thr-Tyr-Gly-Val-Val-Tyr-Lys], 50 mM Tris ⋅ HCl, 62.5 mM MgCl2, 12.5 mM MnCl2, 1 mM EGTA, 0.125 mM sodium orthovanadate, and 1 mM DTT [assay dilution buffer (ADB), pH 7.2]. For the control reactions, 10 μl of the supernatant from corresponding cell extracts were combined with 20 μl ADB without Src substrate peptide. All procedures were performed at 4°C. Aliquots (10 μl) each of [γ-32P]ATP (2 μCi) (Amersham) dissolved in 2 × ADB, 500 μm ATP, and 75 mM MnCl2 were added to each sample at 30°C. After incubation for 10 min at 30°C, the reaction was stopped with 20 μl of ice-cold 40% TCA. All samples were placed on ice for 10 min and centrifuged at 10,000 rpm for 10 min to precipitate any extract proteins. Supernatant (25 μl from each tube; total vol, 60 μl) was removed, spotted onto separate p81 phosphocellulose papers, and mounted on a piece of aluminum foil. Each disk was placed in a glass scintillation vial containing 15 ml of 0.75% phosphoric acid. After being mixed for 30 min at room temperature, the washing reagent was decanted. Each paper was mixed with 10 ml of 0.75% phosphoric acid, followed by 10 ml acetone, for 10 min, respectively, at room temperature. Each reagent was decanted after each washing. Finally, 10 ml of scintillant were added and the radioactivity remaining on each binding paper was counted in a liquid scintillation counter. Nonspecific binding of [γ-32P]ATP to the binding paper without substrate was subtracted from each control sample that included the substrate. Src kinase activities were expressed as picomoles per minute per milligram of protein. The protein in each sample (same fractions as Src kinase measurements) was measured using a Bio-Rad protein assay system.
Immunoprecipitation using a monoclonal anti-phosphotyrosine antibody.
Isolated smooth muscle cells were counted on a hemocytometer and diluted with HEPES buffer as needed. Cells were then treated with reagents for indicated periods. After treatment, the cells were washed with buffer A [in mM: 150 NaCl, 16 Na2HPO4, and 4 NaH2PO4, pH 7.4 (PBS) containing 1 mM Na3VO4]. The cells were then disrupted by sonication inbuffer B (1 mM Na3VO4, 1 mM NaF, 2 mM phenylmethylsulfonyl fluoride, 5 mM EDTA, 1 mM Na4MoO4, 1 mM DTT, 20 mM NaH2PO4, 20 mM Na2HPO4, 20 mM Na4P2O7.10 H2O, 50 μl/ml DNase/RNase, 10 μg/ml aprotinin, 10 μg/ml leupeptin, 10 μg/ml pepstatin A, and 10 μg/ml antipain, pH 7.4) and centrifuged for 15 min at 14,000g. Protein G-sepharose was washed two times with buffer B to make a 50% suspension. Lysate containing 200–500 μg protein in a total of 500 μl ofbuffer B was precleared with 50 μl of protein G-sepharose bead slurry by rocking at 4°C for 30 min. The mixture was spun at 14,000 g for 5 min at 4°C, and 1 μg of monoclonal anti-phosphotyrosine antibody (PY99) was added to the resultant supernatant. The mixture was rocked at 4°C for 1 h or overnight followed by the addition of 50 μl of protein G-sepharose bead slurry. The mixture was further rocked at 4°C for 2 h and spun at 14,000 gfor 5 min, and the supernatant was aspirated off. The pellet was washed three times with buffer A and resuspended in 25 μl of 2× Laemmli sample buffer and boiled for 5 min.
Western immunoblotting using a polyclonal phosphospecific anti-Src (416Y) antibody, a polyclonal phosphospecific anti-MAP kinase antibody, or a monoclonal IgG anti-PI 3-kinase NH2terminal-SH2 domain antibody.
Lysates (100 μg protein) or immunoprecipitates were subjected to SDS-PAGE (7.5% or 12.5%) and electrophoretically transferred to a polyvinylidene fluoride membrane. Immunoblotting was performed using either a mouse monoclonal anti-Src antibody (1:500), a polyclonal phosphospecific anti-Src (416Y) antibody (1:500), an anti-PI 3-kinase antibody (1:500), or a polyclonal phosphospecific MAP kinase antibody (1:1,000) as a primary antibody. Then the membrane was reacted with peroxidase-conjugated goat anti-mouse IgG (1:5,000 dilution) for anti-Src and anti-PI 3-kinase antibodies and a peroxidase-conjugated goat anti-rabbit (1:2,500 or 1:1,000) for anti-Src (P416Y) antibody and anti-MAP kinase antibody, respectively, for 1 h at 24°C. The enzymes on the membrane were visualized with luminescent substrates.
Measurement of contraction.
Aliquots consisting of 2.5 × 104 cells in 0.5 ml of medium were added to 0.1 ml of a solution containing the test agents. The agents were agonists or combinations of agonists and antagonists. In kinetic experiments the reaction was interrupted at various intervals (30 s to 4 min) by the addition of 0.1 ml of acrolein at a final concentration of 1% (vol/vol). Individual cell length was measured by computerized image micrometry. The average length of cells in the control state or after addition of test agents was obtained from 50 cells encountered randomly in successive microscopic fields. The contractile response is defined as the decrease in the average length of the 50 cells and is expressed as the absolute change or the percent change from control length (1).
Contraction in the Absence of Extracellular Ca2+
C2 ceramide, the permeable short-chain ceramide (10−7M), induced contraction of smooth muscle cells that peaked at 30 s (36.7 ± 3.2% decrease in cell length from control) (Fig.1). The contraction was sustained at a plateau (39 ± 1.8%) at 4 min. In cells incubated in Ca2+-free medium containing 2 mM EGTA, the initial and sustained contractions induced by ceramide were inhibited at 30 s (7.5 ± 3.4% decrease in cell length) and 4 min (8.6 ± 1.8% decrease in cell length), indicating that ceramide-induced contraction is dependent on Ca2+ from extracellular sources.
Ceramide-Induced Src Kinase Activities
We next investigated whether C2ceramide had an effect on Src kinase activity, through performing a radioenzyme assay with a Src kinase assay kit. Phosphotransferase activity in the basal state was 3.79 ± 0.90 pmol ⋅ min−1 ⋅ mg protein−1(n = 12) (Fig.2 A). C2 ceramide (10−7 M) induced a rapid increase in kinase activity within 15 s after stimulation (6.54 ± 0.81 pmol ⋅ min−1 ⋅ mg protein−1;n = 12,P < 0.05). It reached a maximal peak of 12.23 ± 3.63 pmol ⋅ min−1 ⋅ mg protein−1(n = 12,P < 0.05) at 1 min and remained elevated at 6.60 ± 1.38 pmol ⋅ min−1 ⋅ mg protein−1(n = 12) at 4 min after stimulation. Src kinase activity levels returned to basal level (2.82 ± 0.55 pmol ⋅ min−1 ⋅ mg protein−1;n = 6) within 20 min after stimulation. In addition, C2-dihydro-ceramide as a negative control did not increase Src kinase activities over basal (Table1). On the other hand, Src kinase activities increased twofold over the basal level in response to stimulation with the longer chain C6 ceramide for 1 min (5.93 ± 0.45 pmol ⋅ min−1 ⋅ mg protein−1;n = 3). C2 ceramide activation of Src kinase was also Ca2+ dependent. Stimulation of the cells with ceramide in the presence of Ca2+-free medium containing 2 mM EGTA resulted in a similar, but transient, increase at 15 s (6.52 ± 1.91 pmol ⋅ min−1 ⋅ mg protein−1;n = 4) (Fig.2 A). The increase of Src kinase activity in cells incubated in Ca2+-free medium containing 2 mM EGTA could not be sustained up to 30 s or thereafter. The data seem to suggest that ceramide-induced Src kinase activity in smooth muscle cells might be biphasic. The initial activation at 15 s was independent of extracellular Ca2+ for activation, whereas the sustained second activation was dependent on extracellular Ca2+ for activity.
To test whether Src kinase activation was specific, we tested the effect of C2 ceramide on Src kinase activity in the presence of the Src kinase inhibitor herbimycin A (19). Preincubation of smooth muscle cells with 3 μM herbimycin A for 5 min followed by ceramide stimulation resulted in inhibition of Src kinase activity. Basal Src kinase activity in the presence of herbimycin A was 6.24 ± 2.38 pmol ⋅ min−1 ⋅ mg protein−1(n = 8) (Fig.2 B). There was no appreciable change in Src kinase activity up to 8 min after stimulation with ceramide in the presence of herbimycin A. On the other hand, the PI 3-kinase inhibitors wortmannin (3 μM) and LY-294002 (1 μM) (12) had no effect on ceramide-induced Src kinase activation after 1 min of ceramide stimulation. The increase in Src kinase activity as measured by radioenzyme assay was corroborated by results obtained from immunoblotting studies using a phosphospecific anti-Src (416Y) antibody directed against the autophosphorylation site of Src. An increase in tyrosine phosphorylation of Tyr416corresponding to Src (60 kDa) was notable at 1 and 4 min stimulation with C2 ceramide (Fig.3 A). Compared with basal (11.06% of total volume), the intensity of the bands increased twofold (20.69% of total volume) at 1-min stimulation with ceramide as measured by densitometry. An increase in immunoreactivities of the 60-kDa band was inhibited in cells preincubated with the Src kinase inhibitor herbimycin A (3 μM) for 5 min in the presence of extracellular Ca2+ (Fig.3 A). In cells preincubated with Ca2+-free medium containing 2 mM EGTA, ceramide treatment decreased Src phosphorylation of Tyr416 at 1 and 4 min (Fig.3 A).
Role of PI 3-Kinase
We next investigated whether C2ceramide activated another kinase in the cascade of nonreceptor tyrosine kinases. Immunoprecipitation using an anti-phosphotyrosine antibody and subsequent detection of bands by immunoblotting with a monoclonal IgG anti-PI 3-kinase NH2 terminal-SH2 domain antibody revealed PI 3-kinase at a molecular mass of 85 kDa (regulatory subunit that contains Src kinase binding sites) (Fig.3 B). Tyrosine phosphorylation of PI 3-kinase could not be detected at the basal level. Phosphorylation of PI 3-kinase seemed to increase in response to ceramide after 30 s and 1 and 4 min of cell stimulation. The level of phosphorylation of PI 3-kinase was highest at 1 min stimulation with ceramide. Preincubation of the cells with two different PI 3-kinase inhibitors [wortmannin (3 μM) and LY-294002 (1 μM)], resulted in abolition of PI 3-kinase phosphorylation (Fig.3 B). Preincubation of the cells with the Src kinase inhibitor herbimycin A also abolished PI 3-kinase phosphorylation. These data, together with the inhibition of Src kinase activity by herbimycin A, but not by wortmannin and LY-294002, seem to indicate that PI 3-kinase is activated by ceramide and that it is downstream of Src kinase in the cascade of the protein tyrosine kinases.
Role of MAP Kinase
We have previously shown that C2ceramide activates MAP kinase (16). We further investigated the relationship between MAP kinase, Src kinase, and PI 3-kinase activation. We performed Western blotting using a phosphospecific MAP kinase antibody. This antibody detects p42 and p44 MAP kinase [extracellular signal-related kinase (ERK1 and 2)] only when catalytically activated by phosphorylation of Tyr204. This antibody does not cross-react with the corresponding phosphorylated tyrosine of either c-Jun NH2-terminal kinase/stress-activated protein kinase (SAPK) or p38 MAP kinase homologues. C2ceramide induced a phosphorylation of the corresponding MAP kinase band (Fig. 3 C). Preincubation of the cells with the Src kinase inhibitor herbimycin A (3 μM) for 5 min or with two different PI 3-kinase inhibitors [wortmannin (3 μM) and LY-294002 (1 μM)] for 5 min resulted in an inhibition of the corresponding MAP kinase p42 and p44 bands as seen on Western blots (Fig.3 C). The data thus seem to suggest that the activation of MAP kinase may be downstream of the Src kinase and PI 3-kinase cascade or is part of a different and parallel pathway (Fig. 4).
pp60src kinase activity has been shown to be phosphorylated and stimulated by agonists, such as CCK-8 in rat pancreatic acini (18) and ATP in bovine coronary arterial smooth muscle (7). In our study, we have investigated the effects of C2 ceramide, as an agonist, on pp60src kinase activity. Work peformed in our laboratory (16) has previously shown that ceramide is a second messenger molecule produced in smooth muscle cells and acts as an intracellular messenger. It induces a sustained contraction of smooth muscle cells, through a pathway that involves the activation of MAP kinase.
Our results provide evidence that pp60src kinase exists and functions in stimulus-contraction coupling in smooth muscle cells from the rectosigmoid. Src kinase activity as measured by radioenzyme assays remained at a basal level before stimulation with C2 ceramide. Ceramide induced an increase in the phosphotransferase activities of protein tyrosine kinase when a Src substrate peptide [p34cdc2-(6—20): KVEKIGEGT([32P])YGVVYK] was used (6, 11). In our system, C2 ceramide induced a ∼3.2-fold increase in Src kinase activities at 1-min stimulation. Another longer chain ceramide, C6 ceramide, at 1-min stimulation induced only about a twofold increase in Src kinase activities compared with basal, suggesting that C6 ceramide is less permeable to cells compared with C2 ceramide. On the other hand, C2-dihydro-ceramide as a negative control did not increase Src kinase activities over basal. Src kinase activity induced by C2 ceramide seemed to increase in a biphasic manner. The initial small peak at 15 s after stimulation was independent of extracellular Ca2+, and the second large increase at 1 min was dependent on the presence of Ca2+ in the extracellular medium. The results obtained from the enzyme assay of Src kinase were corroborated by results obtained from immunoblotting studies using a phosphospecific anti-Src (416Y) antibody that recognizes the autophosphorylation site of Src. There was an increase in Tyr416 phosphorylation of pp60src on stimulation with C2 ceramide, and an inhibition of ceramide-induced phosphorylation was observed in the presence of herbimycin A and the absence of extracellular Ca2+. Ceramide-induced activation of Src kinase requires extracellular Ca2+ for the sustained response. This observation is not surprising since extracellular Ca2+ has been shown to be required in the activation of the Src kinase pathways in pancreatic acinar cells (18). Furthermore, Src family kinases have been implicated in the control of receptor-operated Ca2+influx in various cell types (3, 8, 9, 14, 17, 18, 20). Note that protein tyrosine kinase activation required extracellular Ca2+ in intact cells (6, 11) and that Src kinase activation in microsomes was measured in the absence of Ca2+ (seeMethods). This suggests that the presence of extracellular Ca2+ is a necessary intermediate for activation of Src, but once Src is activated, it does not require Ca2+ for subsequent substrate phosphorylation. These results correlate with the contraction studies, in which ceramide-induced contraction reached a maximum at 30 s and was sustained up to 4 min. In the absence of Ca2+(Ca2+-free medium containing 2 mM EGTA) ceramide-induced contraction was inhibited at both 30 s and 4 min (79% and 78% respectively).
Immunoprecipitation using an anti-phosphotyrosine antibody followed by immunoblotting using an anti-Src (pp60src) antibody showed relatively small increases in the mass of total tyrosine phosphorylated Src on stimulation with ceramide (data not shown). This result suggests that the activity and total phosphorylation of Src are not parallel. The data suggest that the stoichiometry of dephosphorylation at the regulatory subunit (Tyr527) and phosphorylation at the catalytic subunit (Tyr416) may not be 1:1 before or after stimulation due to the possible involvement of other phosphorylation sites (e.g., Tyr216 and Tyr338). Inhibition of Src bands in the absence of extracellular Ca2+ and in the presence of herbimycin A during ceramide stimulation suggests that the autophosphorylation step of Tyr (from Tyr527 to Tyr416) may be Ca2+ and ATP dependent.
Our results also suggest that PI 3-kinase might be downstream of Src kinase in the signal transduction pathway. This deduction is made possible from the results obtained from radioenzyme assays to measure Src kinase activities and immunoprecipitation using an anti-phosphotyrosine antibody followed by Western blot analysis to detect PI 3-kinase phosphorylation. The effects of the Src kinase inhibitor herbimycin A and the PI 3-kinase inhibitors wortmannin and LY-294002 were tested on phosphorylation of PI 3-kinase. Although wortmannin and LY-294002 inhibited PI 3-kinase phosphorylation, these inhibitors did not have an inhibitory effect on Src kinase activity. On the other hand herbimycin A, an inhibitor of pp60src, inhibited Src kinase activity and PI 3-kinase phosphorylation, suggesting that PI 3-kinase is downstream of Src. This is likely since PI 3-kinase is a preferential substrate for pp60src in several cell types (4,5). Lin et al. (11) have shown that tyrosine phosphorylation of PI 3-kinase may be linked to PI 3-kinase upregulation. We propose that ceramide stimulates phosphotransferase activities of Src kinase in a Ca2+-dependent manner, which results in activation of PI 3-kinase to mediate smooth muscle cell contraction.
C2 ceramide induced a phosphorylation of the corresponding MAP kinase band, confirming our previous results (16). Preincubation of the cells with the Src kinase inhibitor herbimycin A for 5 min resulted in an inhibition of the corresponding MAP kinase p42 and p44 bands as seen on Western blots. The data thus seem to indicate that ceramide-induced activation of MAP kinase in rectosigmoid smooth muscle cells is parallel yet downstream of Src and PI 3-kinase activation.
The data seem to correlate with a model proposing a relationship between the tyrosine kinase pathway and the PKC pathway. Ceramide in smooth muscle cells seems to activate both pathways. Herbimycin A, a Src kinase inhibitor, reduced the phosphorylation of MAP kinase. The results suggest that the two kinase cascades are linked and probably have common downstream substrates that modulate smooth muscle contraction.
We thank Erin McDaid for preparing the manuscript and Dr. Michael Su for technical assistance.
Address for reprint requests: K. N. Bitar, Univ. of Michigan Medical School, 1150 W. Medical Center Dr., A520D, MSRB I, Ann Arbor, MI 48109-0656.
This study was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant RO1-DK-42876.
- Copyright © 1998 the American Physiological Society