INTRODUCTION

Top Abstract Introduction Procedures Results Discussion References

Glucose is a major metabolic regulator of insulin secretion from the -cell of the pancreas. Intregral to this regulation is the capacity of the -cell to act as a "fuel sensor" detecting subtle changes in circulating blood nutrients (1, 2). Glucose, unlike hormones and neurotransmitters, does not bind to a receptor but is rapidly transported into the cell and metabolized, resulting in the generation of coupling factors and other mitochondrially derived signals (1, 3, 4). One important consequence of glucose metabolism is thought to be an increase in the ATP/ADP ratio, which brings about closure of ATP-sensitive potassium channels, resulting in cell depolarization, and the consequent opening of voltage-sensitive Ca²⁺ channels (5).

Although these initial consequences of glucose metabolism within the -cell have been well characterized and the distal components of the secretory process are being identified at a molecular level, the intermediate steps still remain poorly understood. One important link between the early stages of glucose metabolism and the exocytosis of insulin is a rise in [Ca²⁺]_i (5-7). Although it may also play a direct role in the docking or fusion of the secretory granules to the plasma membrane (8), there is much evidence to suggest that a major role of the rise in [Ca²⁺]_i is to activate Ca²⁺-dependent protein kinases (9). A number of kinases such as PKC¹ (10, 11), PKA (9, 12), and the CaM kinases (13) have been implicated in regulating insulin secretion. In general the substrates for these are still being identified, although one well defined substrate for PKC is the cytoskeleton-associated MARCKS (myristoylated alanine-rich kinase substrate) protein (14). The role of PKC, however, in nutrient-stimulated insulin secretion remains controversial (10, 11). The importance of CaM kinase is better documented, and it has been widely reported to mediate phosphorylation of a 55-kDa islet cell protein in response to nutrient stimulation (12). This protein has not yet been identified; however, tubulin, a cytoskeleton-associated protein, has been suggested as a possible candidate (15). More recently there has been direct evidence to show that MAP2, a microtubule-associated protein, is a major substrate for CaM kinase II in the pancreatic -cell (16). This suggests that proteins associated with the cytoskeleton may be important kinase substrates linking the rise in [Ca²⁺]_i to insulin secretion.

Cytoskeletal proteins have long been suspected to play a role in insulin secretion (17). Myosin is one of the major cytoskeletal proteins in eukaryotic cells (18). It functions primarily as a motor protein and is therefore involved in a diverse range of cellular functions including cytokinesis and cellular movement and has also been postulated to be involved in secretion (19-22). Conventional, Type II myosin forms part of the diverse family of vertebrate myosin proteins. It is a hexameric complex of proteins that consists of two 200-kDa heavy chains noncovalently bound to two pairs of light chains (17-22 kDa), of which one pair is classified as essential and the other pair is classified as regulatory (20). Cloning of the Type II myosin heavy chain (MHC) has shown that there are three general classes: smooth muscle, cardiac muscle, and nonmuscle myosin (23). Isoforms of MHC also exist within these groups. Conventional smooth muscle myosin has two alternatively spliced isoforms, MHC 200 and MHC 204 (24, 25), whereas nonmuscle MHC is comprised of two genetically distinct isoforms referred to as MHC-A and MHC-B (26, 27).

Regulation of the myosin-actin complex is thought to occur predominantly via the phosphorylation of the regulatory light chains at a number of specific sites. It has been shown that MLC kinase phosphorylates MLC sequentially on serine 19 and threonine 18 (28). In mast cells, PKC phosphorylates MLC on serine 1 and serine 2 residues (21). Phosphorylation of MLC by MLC kinase is thought to help stabilize the three-dimensional structure of the myosin complex and increase the actin-activated ATPase activity (29), whereas phosphorylation of MLC with PKC appears to have no obvious effect on myosin contractile activity. Although a number of recent studies have shown that MHC can also undergo phosphorylation in protozoans (30, 31, 60), there are few reports of phosphorylation of this protein in vertebrates.

The aim of this study was to define changes in Ca²⁺-dependent protein phosphorylation upon nutrient stimulation of rat pancreatic islets and RINm5F cells. The latter are rat insulinoma cells that secrete insulin in response to glyceraldehyde via mechanisms similar to those initiated by glucose in native -cells (32). During the course of this investigation we found that there was an increase in both serine and threonine phosphorylation of MHC, which occurred secondarily to the rise in [Ca²⁺]_i. Our results raise the possibility that myosin function might be controlled through mechanisms in addition to light chain phosphorylation and suggest that phosphorylation of MHC itself may be an important link between the initial rise of [Ca²⁺]_i and the latter steps of nutrient-stimulated insulin secretion.

	EXPERIMENTAL PROCEDURES

Top Abstract Introduction Procedures Results Discussion References

Materials-- Reagents were of analytical grade and were obtained from Sigma, Calbiochem (Alexandria, NSW, Australia), BDH (MERCK Pty Limited, VIC, Australia), or Bio-Rad unless otherwise stated. Tissue culture supplies were obtained from Life Technologies, Inc. except the fetal and newborn calf serum, which was purchased from CSL Limited (Parkville, VIC, Australia). Anti-myosin (nonmuscle) antibody was from Biomedical Technologies Inc. (Stoughton, MA). Anti-phosphothreonine and anti-phosphoserine antibodies were purchased from Zymed Laboratories Inc. (San Francisco, CA). The myosin IIA and myosin IIB antibodies were a generous gift from Dr. Robert Adelstein (Laboratory of Molecular Cardiology, NIH, Bethesda, MD). Donkey anti-rabbit horseradish peroxidase-conjugated secondary antibody was obtained from Jacksons Immunoresearch Laboratory Inc. (West Grove, PA). Insulin radioimmunoassay kit was purchased from Linco Research Inc. (St. Charles, MO). Fura 2 (free acid) and Fura 2AM were obtained from Molecular Probes (Eugene, OR). Enhanced chemiluminescence reagents for immunoblotting were purchased from NEN Life Science Products. [³²P]Orthophosphoric acid was obtained from Amersham Life Science (Buckinghamshire, UK).

Isolation and Incubation of Islets-- Islets were isolated from 230-270-g male Wistar rats by ductal infusion of collagenase and cultured as described previously (33). Batches of 200-300 islets were picked and washed with a modified KRB containing 5 mM NaHCO₃, 1 mM CaCl₂, 2.8 mM glucose, 10 mM Hepes (pH 7.4), and 0.5% (w/v) BSA. The islets were preincubated in 0.2 ml BSA-free KRB for 10 min at 37 °C. Stimulating solutions (0.8 ml) were then added, and the islets were incubated at 37 °C for the time periods as stated. The incubation was terminated by gentle centrifugation and rapid removal of the supernatant. Islets were resuspended in 50 µl of RIPA buffer (1% (w/v) deoxycholate, 1% (v/v) Triton X-100, 0.1% (w/v) SDS, 150 mM NaCl, 1 mM EDTA, 2 mM sodium orthovanadate in 50 mM Tris-HCl (pH 7.4)). Extracts were then sonicated 3 × 10 s using a Branson model 250 sonifier and microtip at power setting 1 and 10% duty cycle.

Cell Culture-- RINm5F cells were maintained in T-75 tissue culture flasks in RPMI 1640 supplemented with 2 mM glutamine, 10% heat-inactivated fetal calf serum, 100 units/ml penicillin and 100 µg/ml of streptomycin in 5% CO₂:95% air at 37 °C. For studies with cell suspensions, monolayers were treated with trypsin/EDTA solution. Cells were then allowed to recover in a spinner medium (containing the same components as the culturing medium with the addition of 10 mM Hepes and 1% newborn calf serum instead of 10% fetal calf serum) for 3 h at 37 °C. RBL-2H3 and COS-7 cells were maintained and used as adherent monolayers as described previously (34, 35).

Measuring Insulin Secretion-- Cells were harvested from the spinner medium and resuspended in KRB with 0.1% (w/v) BSA at 1 × 10⁶ cells/ml. 1-ml aliquots were placed into 1.5-ml centrifuge tubes and preincubated at 37 °C. After 15 min, 800 µl of supernatant was removed and pooled to enable measurement of background secretion. Stimulating solutions (800 µl of 1.2× concentrate) were added, and the cells were incubated at 37 °C for a further 15 min. Incubations were terminated by placing the samples on ice. Cells were then pelleted by centrifugation at 13,000 × g at 4 °C, and the resulting supernatant was removed to determine insulin content using a radioimmunoassay kit with rat insulin as standard.

Intracellular Ca²⁺ Measurements-- The fluorescence dye Fura 2AM (1 µM) was added 30 min prior to removing cells from the spinner medium. Cells were then washed three times in spinner medium and resuspended at 5 × 10⁶ cells/ml. Aliquots of 0.4 ml (2 × 10⁶ cells) were centrifuged (90 × g for 5 min), and the pellets were resuspended in 2 ml of KRB containing 0.1% (w/v) BSA and 200 µM sulfinpyrazone. Cells were activated by the addition of glyceraldehyde (10 mM) or KCl (30 mM), and the fluorescence was measured at 37 °C on a HITACHI F-4010 fluorescence spectrophotometer using an excitation wavelength of 340 nm and an emission wavelength of 505 nm.

Permeabilization with Digitonin-- Cells were harvested from spinner medium as described above and resuspended in a potassium glutamate buffer (buffer A: 140 mM potassium glutamate, 5 mM NaCl, 2 mM MgSO₄, and 20 mM Hepes (pH 7.0)) at a concentration of 5 × 10⁶ cells/ml (36). Aliquots (2 ml) were preincubated at 37 °C for 10 min. Digitonin (final concentration, 10 µM) was added to the cells, and they were incubated for a further 5 min. Permeabilization was terminated with the addition of 30 ml of ice-cold buffer A. Cells were then recovered after gentle centrifugation at 60 × g for 10 min and resuspended at 1.5 × 10⁷ cells/ml in buffer A with the addition of 5 mM Mg-ATP. Efficiency of cell permeabilization was measured using Trypan blue and was routinely 60-70% permeabilized. A 200-µl (3 × 10⁶ cells) aliquot of cells was then incubated on ice for 10 min with Ca²⁺/EGTA buffer. Defined free Ca²⁺ concentrations over a range of 10⁷ to 10⁵ M were obtained by mixing 10 mM Ca²⁺/EGTA with 10 mM Ca²⁺ according to previously calculated ratios (36) and as verified by fluorescence spectroscopy with Fura 2-free acid. Residual nonpermeabilized cells are not responsive to Ca²⁺ over this range. After preincubation with the specific Ca²⁺ buffer (and kinase inhibitors where stated) permeabilized cells were then placed at 37 °C for a further 5 min. The reaction was terminated by centrifugation of the cells at 13,000 × g. The pellet was then resuspended in 50 µl of RIPA buffer, sonicated, and subjected to SDS-PAGE as described below.

Immunoblot Analysis-- Cells were incubated and stimulated in BSA-free KRB as for the insulin secretion studies. At the end of the incubation period, 1 × 10⁶ to 3 × 10⁶ cells were centrifuged, supernatant was removed, and cells were then lysed by the addition of 50 µl of RIPA buffer and sonicated as described in the islet experiments. A 10-15-µl aliquot of sample was then denatured in Laemmli sample buffer (37) for 5 min at 100 °C prior to SDS-PAGE. Proteins were transferred to nitrocellulose membranes, and transfer efficiency was checked by staining of the membrane with Ponceau Red. The membrane was then blocked for 1 h in Tris-buffered saline containing 0.05% (v/v) Tween 20 and 5% (w/v) nonfat milk powder. Incubation with the primary antibodies was for 2 h at room temperature followed by a 1-h incubation with horseradish peroxidase-conjugated secondary antibody. Bands were visualized using immunoblot chemiluminescence reagents. In the competition experiments the phosphothreonine antibody was preincubated for 2 h at room temperature with differing (phospho)amino acids (1 mM) and then used for immunoblotting as described above. When required, membranes were stripped of antibodies using a buffer containing 2% (w/v) SDS, 100 mM -mercaptoethanol and 62.5 mM Tris (pH 6.7) at 50 °C for 30 min.

Immunoprecipitation from RINm5F Cells-- Cells were seeded into a 6-well plate at 3 × 10⁶ cells/well and cultured overnight at 37 °C. Attached RINm5F cells were used in the immunoprecipitation experiments because they showed a similar increase in threonine phosphorylation of all proteins upon stimulation with KCl (data not shown) but were easier to manipulate under these circumstances than those in suspension. Cells were washed twice with 3 ml of KRB and preincubated in 2 ml of KRB at 37 °C for 10 min. Medium (1.6 ml) was then removed, and cells were activated with the addition of 1.6 ml of stimulating solutions. The incubation was terminated by placing the cells on ice and removing the medium. Lysis buffer (450 µl) containing 1% (v/v) Nonidet P-40, 100 mM sodium pyrophosphate, 250 mM NaCl, 50 mM NaF, 5 mM EGTA, 0.1 mM phenylmethylsulfonyl fluoride, 10 µg/ml leupeptin, 15 mM -mercaptoethanol, and 20 mM Tris-HCl (pH 7.9) (38) was added, and the plates were left on ice for 15 min. Cells were scraped off, and the lysate was sedimented at 100,000 × g for 5 min at 4 °C. The supernatant fraction was precleared with pansorbin (a standardized Staphylococcus aureus cell suspension) at one-fifth the final lysate volume for 30 min at 4 °C. The pansorbin was sedimented, and the cell lysate was then incubated with 20 µl of anti-myosin IgG (nonmuscle) or 20 µl (10 µg) of anti-phosphothreonine on a nutator at 4 °C for 2 h. A further 90 µl of pansorbin was added for an additional hour of incubation. The mixture was then sedimented by centrifugation at 13,000 × g at 4 °C. The resulting pellet was washed three times. The first wash was with lysis buffer as described above; the second was in a 1:1 dilution of lysis buffer with standard phosphate-buffered saline (pH 7.4); and the third was in phosphate-buffered saline alone. The pellet was then resuspended in 75 µl of preheated (90 °C) Laemmli sample buffer and heated for a further 5 min at 90 °C. After centrifugation, the supernatant was transferred to a fresh 1.5-ml tube, and a sample (10 µl) was separated on a 10% SDS-polyacrylamide gel. The MHC could be detected once the gel was stained with Coomassie Blue. Relative protein levels were determined by scanning the MHC band on a Molecular Dynamics Densitometer PDSI (Sunnyvale, CA), thereby allowing for correction of protein loading in subsequent immunoblots or phosphoamino acid analysis.

[³²P]Orthophosphate Labeling of RINm5F Cells-- Cells were cultured and seeded in 6-well plates as described above. On the day of the experiment, medium was removed, and cells were gently washed twice with a phosphate-free salt solution (119 mM NaCl, 5 mM KCl, 5.6 mM glucose, 0.4 mM MgCl₂, 1 mM CaCl₂, 0.1% (w/v) BSA, 4 mM glutamine, and 25 mM PIPES-NaOH (pH 7.2) (38). Cells were then labeled with [³²P]orthophosphoric acid (200 µCi/well) for 2 h at 37 °C. Radioactive supernatant was removed, and cells were washed once with 3 ml of warmed phosphate-free buffer and once with 3 ml of KRB. Cells were preincubated in 2.5 ml of KRB for 10 min at 37 °C. Cells were then stimulated for 5 min, and the incubation was stopped by aspirating the stimulating solutions and placing cells on ice. Lysis buffer was added, and immunoprecipitation was carried out as above.

Phosphopeptide Mapping-- The resulting pellet was resuspended in 10 µl of electrophoresis buffer containing acetic acid:formic acid:deionized water (15:5:80). Samples were then spotted onto a Silica Gel 60 TLC plate. The peptides were separated by electrophoresis at 1000 V for approximately 60 min at 4 °C in the first dimension using Orange G and Acid Fuchsin as marker dyes. Plates were dried, and peptides were then subjected to ascending chromatography in a solution containing n-butyl alcohol:pyridine:acetic acid:formic acid:deionized water (127.5:22.5:45:15:90) for 6 h. Plates were again dried, and radioactive peptides were identified by autoradiography.

Phosphoamino Acid Analysis-- The pellet was resuspended in 200 µl of 6 M HCl and incubated at 110 °C for 2.5 h. The hydrolysate was lyophilized, and the pellet was resuspended in 10 µl of first dimension electrophoresis buffer (pH 1.9) (formic acid (88% w/v):glacial acetic acid:deionized water (1:3.1:35.9)). Samples (5-10 µl) were spotted onto cellulose TLC plates using 2 µl each of phosphothreonine, phosphoserine, and phosphotyrosine (1 mg/ml) as standards. Green marker dye (5 mg/ml -dinitrophenyl-lysine and 1 mg/ml xylene cyanol FF) was used to measure electrophoresis progress. Two-dimensional electrophoresis was carried out on a Pharmacia Biotech Multiphor II. The first dimension was run at 1000 V for 65 min in buffer (pH 1.9). The second dimension was run at 1000 V for 30 min in buffer (pH 3.6) (glacial acetic acid:pyridine:deionized water (10:1:189)). Phosphoamino acid standards were detected by spraying the TLC plate with ninhydrin. The TLC plate was dried, and the radioactivity of the spots corresponding to the relative phosphoamino acids was analyzed by phosphorimaging on a Molecular Dynamics PhosphorImager 445 SI. Duplicate samples (5 µl) were also subjected to two-dimensional electrophoresis (as described above), and the radioactive phosphoamino acids were scraped from the plates and counted on a Beckman LS6000 scintillation counter in 1 ml of deionized water.

	RESULTS

Top Abstract Introduction Procedures Results Discussion References

Protein Phosphorylation in RINm5F Cells-- The initial aim of this study was to investigate changes in threonine and serine phosphorylation in lysates of RINm5F cells that had been stimulated with the nutrient secretagogue glyceraldehyde. Commercial grade serine and threonine antibodies were used to study changes in protein phosphorylation. The serine antibody detected phosphorylated proteins only poorly (Fig. 1A). The threonine antibody proved to be more efficient. Stimulation of RINm5F cells with 10 mM glyceraldehyde resulted in an increase in threonine phosphorylation of a protein that comigrated with the 200-kDa marker (Fig. 1B). Although a 66-kDa protein showed higher basal levels of threonine phosphorylation, it was the 200 kDa protein that predominantly increased in threonine phosphorylation upon stimulation of these cells. Competition experiments demonstrated that the phosphothreonine antibody was specific for phosphothreonine residues because preincubation with 1 mM phosphothreonine completely abolished the immunoreactivity of the 200-kDa protein (Fig. 1C, lane 2). In contrast, clearly detectable signals were still obtained after preincubation with 1 mM phosphoserine or threonine itself (Fig. 1C, lanes 3 and 4).

View larger version (50K): [in this window] [in a new window]   Fig. 1.   Detection of phosphoproteins in lysates of RINm5F cells using anti-phosphoserine (A) or anti-phosphothreonine (B) antibodies and determining specificity of the anti-phosphothreonine antibody (C). RINm5F cell lysates (~106 cells) that had been stimulated with (+) or without () 10 mM glyceraldehyde (GLY) were subjected to 10% SDS-PAGE as described under "Experimental Procedures." Duplicate samples were used for the two immunoblots, which are representative of five individual experiments. For competition experiments, RINm5F cell lysates were immunoblotted with anti-phosphothreonine antibody that had been preincubated with buffer, 1 mM phosphothreonine, phosphoserine, or threonine itself (lanes 1-4, respectively).

Time Course of Threonine Phosphorylation-- To extend the relevance of this response to nontumoral cells, pancreatic islets were stimulated with glucose (16.7 mM), and threonine phosphorylation was determined over a series of time intervals (Fig. 2A). In islets there was also an increase in threonine phosphorylation of a 200-kDa protein. This response was rapid in onset, reached a maximum by 5 min, and declined toward basal levels by 15 min. Fig. 2B shows the similar time courses undertaken in RINm5F cells using glyceraldehyde (10 mM) and a depolarizing agent KCl (30 mM), both of which promote Ca²⁺ influx through voltage-gated ion channels. Stimulation with KCl resulted in a transient increase in threonine phosphorylation very similar to that seen with glucose in islets, except that it approached its peak within 1 min. The glyceraldehyde response showed a slightly different pattern, being slower in onset, peaking at 15 min, and remaining elevated up to 30 min. In RINm5F cells, stimulation with glyceraldehyde and KCl resulted in an increase in [Ca²⁺]_i. In parallel experiments using the Ca²⁺ indicator Fura 2AM, we found that the peak rise in [Ca²⁺]_i after KCl stimulation (10-15 s) preceded that for glyceraldehyde (1-2 min) (data not shown). The peaks in [Ca²⁺]_i therefore preceded or were coincident with the maximal increases in threonine phosphorylation of the 200-kDa protein seen for each of these stimuli. In all instances KCl gave greater maximal increases in threonine phosphorylation than glyceraldehyde (5-fold versus 3-fold), which corresponds to their relative abilities to raise [Ca²⁺]_i (data not shown).

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Time course of stimulated threonine phosphorylation of the 200-kDa protein in islets (A) or RINm5F cells (B). Time courses were performed as described under "Experimental Procedures." For both islet and RINm5F cells, lysates were subjected to 7.5% SDS-PAGE and immunoblotted with the anti-phosphothreonine antibody. Immunoreactive bands were analyzed densitometrically. A represents the time course of threonine phosphorylation of the 200-kDa protein in rat islets in response to 16.7 mM glucose over 15 min. Points represent the means of three experiments ± S.D. expressed as a percentage of the 5 min (maximal) response, after subtraction of the zero time value. B represents the time course in RIN5mF cells that had been stimulated with 10 mM glyceraldehyde (GLY) or 30 mM KCl over 30 min. Points represent the means of two experiments ± S.D. expressed as percentages of the maximal 1 min or 15 min responses for KCl and glyceraldehyde, respectively.

Ca²⁺ Dependence of Threonine Phosphorylation-- The relationship between the time courses of phosphorylation and rises in [Ca²⁺]_i suggested that the increase in threonine phosphorylation may be Ca²⁺-dependent. This was further substantiated in experiments in which intact RINm5F cells were stimulated with KCl and glyceraldehyde in the presence of verapamil (a Ca²⁺ channel antagonist). This agent caused a significant decrease in stimulated threonine phosphorylation of the 200-kDa protein in response to KCl and glyceraldehyde (Fig. 3A), indicating that an influx of extracellular Ca²⁺ was important for the initiation of this response. To further confirm the Ca²⁺ dependence, RINm5F cells were permeabilized with 10 µM digitonin and incubated in a range of Ca²⁺ buffers reflecting the expected [Ca²⁺]_i range from basal state (0.1 µM) to maximally stimulated (10 µM) (36). As shown in Fig. 3B, a small increase in [Ca²⁺]_i from 0.1 to 0.15 µM was sufficient to increase threonine phosphorylation of the 200-kDa protein. This response was also concentration-dependent with further increases in [Ca²⁺]_i resulting in increases in threonine phosphorylation of the protein. This further supported the hypothesis that the response was tightly linked to the [Ca²⁺]_i.

View larger version (18K): [in this window] [in a new window]   Fig. 3.   Ca2+ dependence of threonine phosphorylation of the 200-kDa protein. Intact RINm5F cells were stimulated with 10 mM glyceraldehyde (GLY) and 30 mM KCl with (+Verap) or without (Verap) verapamil, and lysates (~1 × 106 cells) were subjected to 7.5% SDS-PAGE. Anti-phosphothreonine immunoblots were analyzed densitometrically. A represents the fold increase in threonine phosphorylation over basal. Results shown are the means ± S.D. of two experiments. B shows a Ca2+ dose response of digitonin-permeabilized RINm5F cells. RINm5F cells were permeabilized and incubated in a range of Ca2+ buffers as described under "Experimental Procedures." Lysates were subjected to 7.5% SDS-PAGE and immunoblotted with anti-phosphothreonine, and a representative immunoblot is shown as an inset. Immunoreactive bands were analyzed densitometrically, and the results are expressed as fold increase in threonine phosphorylation over basal [Ca2+]i (0.1 µM) with each point representing the mean ± S.D. of two experiments.

Protein Kinases Potentially Responsible for Threonine Phosphorylation of the 200-kDa Protein-- We next sought to establish which kinase(s) was involved in the phosphorylation of this 200-kDa substrate. The CaM kinase II family of kinases, some of which are well expressed in insulin-secreting cells (39), can be inhibited with KN-93. However, a potential problem with the use of this compound is that it can also inhibit L-type Ca²⁺ channels (40). To overcome this, experiments were performed in digitonin-permeabilized cells. However, in three experiments, KN-93 (0.1-10 µM) did not alter the threonine phosphorylation of the 200-kDa protein at either basal (0.1 µM) or stimulated (1.0 µM) Ca²⁺ (data not shown).

View larger version (23K): [in this window] [in a new window]   Fig. 4.   Effects on threonine phosphorylation of the 200-kDa protein by activators of PKCs and PKAs. Intact RINm5F cells were incubated in the presence of KCl (K), 30 µM forskolin (F), or 100 nM TPA (T) as described under "Experimental Procedures." Lysates (1 × 106 cells) were subjected to 7.5% SDS-PAGE and immunoblotted with anti-phosphothreonine antibody. The (A). B represents the level of insulin secretion (fold/basal) for the above activators (means ± S.D. of two experiments).

Identification of the 200-kDa Protein-- An obvious candidate for the 200-kDa phosphoprotein would be MHC. To test this possibility, MHC from control and KCl-stimulated RINm5F cell lysates was immunoprecipitated using a polyclonal nonmuscle myosin antibody. Fig. 5A represents the immunoblot using the same nonmuscle myosin antibody. The antibody was strongly immunoreactive with a protein comigrating with the 200-kDa marker. The protein that was immunoprecipitated also increased in threonine phosphorylation in KCl-stimulated cells as shown in Fig. 5B.

View larger version (44K): [in this window] [in a new window]   Fig. 5.   Immunoprecipitation of MHC from stimulated RINm5F cells lysates. MHC was immunoprecipitated from RINm5F cell lysates that had (+) or had not () been stimulated with 30 mM KCl using the nonmuscle myosin antibody as described under "Experimental Procedures." Immunoprecipitates were subjected to 7.5% SDS-PAGE and immunoblotted with the nonmuscle myosin antibody (A). The membrane was stripped of antibody using the method previously described and immunoblotted with the anti-phosphothreonine antibody (B).

View larger version (49K): [in this window] [in a new window]   Fig. 6.   Immunoprecipitation of MHC from 32P-labeled RINm5F cells. MHC from basal (lane B) or cells stimulated with glyceraldehyde (lane G) or KCl (lane K) was immunoprecipitated with the nonmuscle myosin antibody as described under "Experimental Procedures." An aliquot (10 µl) of immunoprecipitate was subjected to 7.5% SDS-PAGE. Before immunoblotting, the nitrocellulose membrane was analyzed via a PhosphorImager to detect increases in 32P-labeled MHC, and the image is shown in A. B represents the immunoblot of the samples using the anti-phosphothreonine antibody. This membrane was then stripped as described previously and immunoblotted with the nonmuscle myosin antibody, and the immunoblot is represented in C.

Phosphopeptide Mapping-- Phosphopeptide mapping was undertaken to determine whether the site of phosphorylation of MHC in RINm5F cells stimulated with glyceraldehyde or KCl was the same. The ³²P-labeled myosin was immunoprecipitated from control, glyceraldehyde- and KCl-stimulated cells and subjected to tryptic digestion. Peptides were separated via electrophoresis (E) and ascending chromatography (C). Representative autoradiograms are shown in Fig. 7. There is a basally phosphorylated peptide present (peptide-1) that increases in phosphorylation upon stimulation with both KCl and glyceraldehyde. In addition to this, one major (peptide-2) and two minor peptides (peptides 3 and 4) are also phosphorylated in stimulated cells but to a lesser degree. Importantly, under both conditions the phosphorylated peptides show a similar migratory pattern, suggesting that they are in fact identical and that the target sites of phosphorylation for these two stimulants is the same. However, it cannot be excluded that additional sites are phosphorylated but remain undetected because of limitations in the lower degree of incorporation of ³²P into RINm5F cells as opposed to other (e.g. RBL-2H3) cells.²

View larger version (59K): [in this window] [in a new window]   Fig. 7.   Phosphopeptide mapping of MHC from RINm5F cells. MHC was immunoprecipitated from RINm5F cells prelabeled with 32P and subjected to tryptic digestion. The resulting peptides were separated via electrophoresis (E) and ascending chromatography (C) as described under "Experimental Procedures." Shown is a representative phosphorimage showing migratory patterns of peptides from basal, KCl, and glyceraldehyde (GLY) stimulated RINm5F cells. Peptides are numbered 1-4 in order of intensity.

Phosphoamino Acid Analysis-- Phosphoamino acid analysis was undertaken to confirm the increase in threonine phosphorylation of MHC upon stimulation of RINm5F cells. The ³²P-labeled myosin was immunoprecipitated from control and KCl-stimulated RINm5F cells and subjected to tryptic digestion, amino acid hydrolysis, and two-dimensional phosphoamino acid analysis. Fig. 8 (left panel) is a representative result. From analysis of the phosphorimage we were able to measure semiquantitatively the changes in phosphothreonine and phosphoserine (Fig. 8, right panel). MHC displayed a high degree of serine phosphorylation under basal conditions, and this increased 2.2-fold upon stimulation with KCl. In contrast threonine phosphorylation was very low in unstimulated cells but increased markedly in response to KCl (~5.4-fold/basal). This densitometric analysis was confirmed by removal of the spots from the plate and quantification of the radioactivity present by Cerenkov counting (data not shown). These results, therefore, confirmed that MHC does undergo an increase in threonine phosphorylation and, in addition to the competition data (Fig. 1), validates that the anti-phosphothreonine antibody used above was specific for phosphothreonine residues. They also reveal an increase in serine phosphorylation that was undetectable with the anti-phosphoserine antibody (data not shown).

View larger version (25K): [in this window] [in a new window]   Fig. 8.   Phosphoamino acid analysis of MHC from RINm5F cells. MHC was immunoprecipitated from RINm5F cells prelabeled with 32P. Tryptic digestion and acid hydrolysis were performed as described under "Experimental Procedures." The left panel is an autoradiogram of the two-dimensional phosphoamino acid analysis for unstimulated RINm5F cells (i) or cells stimulated with 30 mM KCl (ii). The first dimension of the separation was run using pH 1.9 buffer, and the second was run with pH 3.5 buffer. Phosphothreonine and phosphoserine residues are labeled. The graph presented in the right panel is the mean of two experiments (± S.D.) and represents the changes in phosphorylation of threonine and serine residues as quantified densitometrically and is expressed relative to the phosphoserine and phosphothreonine contents of unstimulated cells.

Subtypes of Myosin in RINm5F Cells and Islets-- The polyclonal nonmuscle myosin antibody potentially immunoreacts with a number of different subtypes of myosin. Using antibodies generated against synthetic peptides of specific regions of myosin IIA and myosin IIB, we were able to establish the subtypes present in RINm5F and islet cells. Fig. 9 (A and B) represents an immunoblot obtained using these antibodies with four different cell types. We have confirmed that RBL-2H3 mast cells, as previously reported, contained only myosin IIA (41). COS-7 cells contained only myosin IIB. There was no cross-reactivity of the antibodies, demonstrating that they were specific for their respective myosin isoform. Importantly, RINm5F cells and islets were shown here to contain both the myosin IIA and myosin IIB isoforms.

View larger version (37K): [in this window] [in a new window]   Fig. 9.   Identification of myosin IIA and IIB subtypes present in RINm5F cells and rat islets and investigating which isoform is threonine-phosphorylated. Duplicate samples of lysates (equivalent to 1 × 106 cells) from RINm5F (RIN), pancreatic islets (ISLET), RBL-2H3, or COS-7 cells were subjected to 7.5% SDS-PAGE as described under "Experimental Procedures." Immunoblots using myosin IIA (A) and myosin IIB (B) are shown. MHC was immunoprecipitated with the anti-phosphothreonine antibody from cells stimulated with (+) or without () 30 mM KCl as described under "Experimental Procedures." The immunoprecipitate was then subjected to 7.5% SDS-PAGE after correction to ensure equal loading of total MHC protein. Immunoblots using myosin IIA (C) and myosin IIB (D) are shown and are representative of n = 3 experiments.

Identity of Threonine-phosphorylated Myosin Isoform-- To examine the threonine phosphorylation of the individual MHC isoforms, proteins were immunoprecipitated from control and KCl-stimulated RINm5F cells using the anti-phosphothreonine antibody. A sample of the immunoprecipitate was subjected to 10% SDS-PAGE and stained with Coomassie Blue. The protein loading was corrected according to the amount of total MHC present as assessed densitometrically. Immunoblots using the specific myosin IIA and myosin IIB antibodies are represented in Fig. 9 (C and D), respectively. These show that more myosin IIA, as compared with myosin IIB, is recruited to the immunoprecipitates following stimulation. Densitometric analysis of this and replicate experiments revealed a stimulated increase of 4.5 ± 1.0-fold over basal for myosin IIA as opposed to 2.3 ± 0.5-fold for myosin IIB. These results suggest that myosin IIA is the predominant threonine-phosphorylated species following stimulation.

	DISCUSSION

Top Abstract Introduction Procedures Results Discussion References

The initial steps of nutrient metabolism in the pancreatic -cell and the manner in which these instigate a rise in [Ca²⁺]_i have been broadly characterized (5-7). However, the Ca²⁺-dependent kinases that are activated upon the rise in [Ca²⁺]_i, the potential substrates on which they act, and the roles of these substrates in the latter part of the secretory process are still poorly understood. In this study we have provided evidence that MHC is a possible substrate for Ca²⁺-dependent kinases and is both serine- and threonine-phosphorylated upon stimulation of the rat insulinoma cell line RINm5F. MHC has never previously been shown to be a kinase substrate in insulin-secreting cells nor to our knowledge has threonine phosphorylation of nonmuscle MHC been demonstrated before in systems other than protozoa (31). Because of its novelty and because it was more amenable to further investigation via immunoblotting, the threonine response was characterized in more detail. The maximal increase in threonine phosphorylation occurs slightly later (glyceraldehyde) or in parallel to (KCl) the peak rise in [Ca²⁺]_i, but its onset certainly precedes that of insulin secretion, which lags several minutes after nutrient delivery (7, 43). Threonine phosphorylation was triggered by even small increases in [Ca²⁺]_i above 0.1 µM and, because of the very low phosphothreonine content of MHC under basal conditions, showed a greater increase (6-9-fold) than did serine phosphorylation (2-fold). For these reasons it is possible that threonine phosphorylation of MHC plays an important role in Ca²⁺-dependent and hence nutrient-stimulated insulin secretion. This is consistent with the finding that both the KCl and glyceraldehyde appeared to activate the same kinase in that identical MHC peptides were phosphorylated by the two Ca²⁺-dependent stimuli. However, these and other results do not preclude an additional or alternative involvement of serine phosphorylation of MHC. Results from the phosphoamino acid analysis suggest that basal phosphorylation is predominantly on serine, but this does increase upon stimulation. Peptide 1, phosphorylated under basal conditions, might therefore contain serine site(s), implying that one or more of the sites on peptides 2, 3, and 4 are threonine. Serine phosphorylation may be particularly important for PKC-dependent secretory processes, as previously suggested for RBL-2H3 cells (21), because phosphothreonine content of MHC was not augmented by PKC activation in RINm5F cells.

Irrespective of the potential role of threonine phosphorylation in Ca²⁺-dependent insulin secretion, it is apparent that these two processes do not correspond in a moment to moment basis, because threonine phosphorylation is transient in islets in response to glucose, whereas secretion would be expected to be ongoing. The short-lived nature of this response does not preclude MHC phosphorylation from having a possible important role in initiating secretion. This transient response was also seen in KCl-stimulated RINm5F cells, suggesting that a prolonged elevation of [Ca²⁺]_i might also regulate the dephosphorylation of the 200-kDa protein, possibly by activation of the Ca²⁺-dependent phosphatases (44, 45). The anti-phosphothreonine antibody used in these studies detects net changes in phosphorylation, suggesting that the transient response to glucose is indeed real. This might not necessarily be the case in metabolic labeling experiments, in which an apparent decrease in phosphorylation might be because of a decrease in the specific activity of the intracellular ATP pool, secondary to nutrient metabolism.

There is strong evidence for the role of CaM kinases as primary regulators of the insulin secretory response (13, 46, 47). A subtype of this family, CaM kinase II, has been reported to phosphorylate a number of cytoskeleton-associated proteins including synapsin I, an important protein involved in neurotransmitter release (48, 44) and MAP2, a microtubule protein (16). An isoform of CaM kinase II, ₂, has recently been detected in both rat islets and RINm5F cells and shown to be predominantly located in the cytoskeletal fraction and associated with the insulin secretory vesicles in -cells (39). This suggests that CaM kinase may be important in the regulation of insulin secretion and may possibly be the kinase that is responsible for threonine phosphorylation of MHC in stimulated RINm5F cells. However, inhibition of CaM kinase II with KN-93, which has similar actions to KN-63 in that it inhibits both the autophosphorylation of CaM kinase II and phosphorylation of substrates (47, 49), had no effect on threonine phosphorylation of MHC, suggesting that CaM kinase II was not involved. This would be consistent with the earlier findings that this enzyme appears to require higher concentrations of Ca²⁺ for activation than the very low levels capable of stimulating MHC phosphorylation in permeabilized RINm5F cells. However, it is possible that phosphorylation of MHC on the threonine residue may be carried out by another member of the CaM kinase family. Another possible candidate could be MHC kinase A, which belongs to a newly identified kinase family and has been shown to be responsible for the threonine phosphorylation of MHC in Dictyostelium (63). However, there is no evidence to date that this family of kinases is present in RINm5F cells. It should also be noted that although a direct activation of the responsible kinase by Ca²⁺ is the simplest explanation for our results, it cannot be excluded that Ca²⁺ is acting indirectly. Examples of indirect activation might include a phosphorylation cascade or changes in the cytoskeleton allowing MHC more ready access to the activated kinase. Identification of the kinase would probably resolve these issues and is currently being addressed.

Myosin can be classified as conventional (Type II) and unconventional (all other types) on the basis of the structure of its head region. Myosin is ubiquitously expressed, and a particular cell type expresses a variety of myosin subtypes (18). Moreover, it is becoming clear that the role of each subtype may be different, making it important to identify the myosins that may be involved in secretion. Recent work with the unconventional myosins, in particular Type V, has demonstrated that they have a function in vesicular movement (50, 51). However, there is also limited but growing evidence that the Type II myosin might play an important role in the regulation of secretion. Myosin Type II has been implicated in the regulation of neurotransmitter release because microinjection of an inhibitory myosin II antibody was found to retard synaptic transmission (22). Myosin II and the phosphorylation of its heavy and light chains is also thought to be important in the regulation of histamine release in rat mast cells (RBL-2H3) (21). More recently it has been shown in a mouse -cell line, that the inhibition of MLC kinase, which modulates the myosin protein complex, also inhibits insulin secretion (52). Myosin has also been reported to be important in both the fast and slow phases of secretion in sea urchin eggs. In this system 2,3-butanedione monoxime, an inhibitor of myosin ATPase activity (42, 54), significantly decreased secretion (53). We have found using RINm5F cells that the same compound also inhibited insulin secretion.² Although not conclusive, because 2,3-butanedione monoxime can also inhibit Ca²⁺ channels under some conditions (62), this finding is at least consistent with a role for myosin II in the insulin secretory process.

In nonmuscle cells myosin II exists as two isoforms, IIA and IIB, which are expressed in different ratios in a tissue-dependent manner (27). We have shown for the first time that RINm5F and rat islets contain both myosin IIA and IIB subtypes. It is now becoming apparent that these subtypes also display different subcellular distributions (55, 56), and it is possible that one of the factors that determines their localization is differential phosphorylation. Myosin IIB has an insert that varies in size depending on the species in which it is expressed. This insert in oocytes is known to be phosphorylated by Cdc2 kinase during meiosis, which is thought to regulate the location of this myosin subtype (57). Interestingly, in RINm5F cells it appears that myosin IIA is the subtype predominantly threonine-phosphorylated upon stimulation. This then raises the possibility that the two myosin subtypes are differentiallly regulated and might exert different functions. An alternative possibility, which remains to be tested, is that myosin IIB is serine-phosphorylated by Ca²⁺-dependent kinases, and this exerts the same functional consequences as phosphorylation of an equivalent threonine residue in myosin IIA.

One final question to address is how MHC phosphorylation might regulate insulin secretion. Although the role of regulatory light chains in controlling myosin function, at least at the level of ATPase activity, is well documented (23), it is now becoming apparent that regulation of this complex could also occur via phosphorylation of the heavy chain (21, 30, 31, 58-60). An important structural characteristic of MHC is that it consists of a globular, amino-terminal head domain, which contains the actin- and ATP-binding sites, and a carboxyl terminus that, when dimerized, forms a coiled-coil rod structure (58). From studies in the protozoans Dictyostelium and Acanthamoeba, it has been found that phosphorylation, in particular the location of this phosphorylation, on MHC has a number of consequences. In Acanthamoeba, MHC phosphorylation inhibits both the ATPase activity and filament formation in vitro (59). In Dictyostelium, phosphorylation of three threonine residues in the tail region of MHC alters only filament formation (31, 60). It has also been recently demonstrated that phosphorylation of the tail region of MHC in rabbit brain also disrupts filament formation (61). The conclusion emerging from these studies is that phosphorylation of MHC in the head region appears to affect actin binding and ATPase activity, thereby modifying its contractile activity, whereas phosphorylation on the tail region of MHC is thought to alter filament formation (30). Based on the protozoan models, it had been hypothesized that phosphorylation of vertebrate MHC in the tail region might result in filament instability and the breakdown of the cortical actin web, which then allows easier access of the granules to the plasma membrane (21). In addition to the evidence already presented, it has recently been demonstrated in Dictyostelium that the threonine phosphorylation on the tail of the MHC also plays a role in localizing the protein during cytokinesis (60), confirming that phosphorylation of MHC may be important in determining the locality of the protein within the cell. The phosphorylation of MHC is therefore potentially involved in regulation of the myosin complex, and hence secretion, via a number of mechanisms. Which, if any, of these mechanisms is actually involved remains to be determined.