The Muscle-specific Calmodulin-dependent Protein Kinase Assembles with the Glycolytic Enzyme Complex at the Sarcoplasmic Reticulum and Modulates the Activity of Glyceraldehyde-3-phosphate Dehydrogenase in a Ca2+/Calmodulin-dependent Manner*

The skeletal muscle specific Ca2+/calmodulin-dependent protein kinase (CaMKIIβM) is localized to the sarcoplasmic reticulum (SR) by an anchoring protein, αKAP, but its function remains to be defined. Protein interactions of CaMKIIβM indicated that it exists in complex with enzymes involved in glycolysis at the SR membrane. The kinase was found to complex with glycogen phosphorylase, glycogen debranching enzyme, glyceraldehyde-3-phosphate dehydrogenase (GAPDH), and creatine kinase in the SR membrane. CaMKIIβM was also found to assemble with aldolase A, GAPDH, enolase, lactate dehydrogenase, creatine kinase, pyruvate kinase, and phosphorylase b kinase from the cytosolic fraction. The interacting proteins were substrates of CaMKIIβM, and their phosphorylation was enhanced in a Ca2+- and calmodulin (CaM)-dependent manner. The CaMKIIβM could directly phosphorylate GAPDH and markedly increase (∼3.4-fold) its activity in a Ca2+/CaM-dependent manner. These data suggest that the muscle CaMKIIβM isoform may serve to assemble the glycogen-mobilizing and glycolytic enzymes at the SR membrane and specifically modulate the activity of GAPDH in response to calcium signaling. Thus, the activation of CaMKIIβM in response to calcium signaling would serve to modulate GAPDH and thereby ATP and NADH levels at the SR membrane, which in turn will regulate calcium transport processes.

Free calcium (Ca 2ϩ ) regulates diverse cellular functions by acting as an intracellular second messenger. A large part of these cellular functions are mediated by CaM, 1 which is the ubiquitous intracellular Ca 2ϩ receptor. The Ca 2ϩ ⅐CaM complex allosterically activates numerous proteins, including Ca 2ϩ /CaMdependent protein kinase II (CaMKII) (1). CaMKII is a multifunctional enzyme that is highly expressed in brain and muscle. The kinase is believed to serve important roles in synaptic transmission (2,3), gene transcription (4,5), cell growth (6), and control of excitation-contraction coupling (7)(8)(9).
The subcellular distribution of CaMKII indicates cytosolic and membrane localizations in different tissues (10). In skeletal muscle, an isoform of CaMKII is targeted to the SR membrane by a non-kinase protein, ␣KAP (11,12). Different studies have been conducted to determine whether membrane-bound CaM kinase could phosphorylate different substrate proteins by virtue of its proximity effects and thereby regulate SR function. Although the calcium release channel/ryanodine receptor (RyR) and the calcium pump/Ca 2ϩ -ATPase in skeletal muscle SR were shown to be substrates of CaMKII␤ (13,14), there does not appear to be any clear effects on the regulation of functional activity of these proteins induced by such phosphorylation (7,8,(15)(16)(17). Moreover, there is clear evidence that the RyR and calcium pump are regulated by local ATP, Ca 2ϩ , and CaM through direct ligand binding (15)(16)(17). In this regard, both Ca 2ϩ and CaM are present at the SR, and the level of ATP is believed to be tightly controlled through a membrane-bound glycolytic machinery involving phosphorylase b, creatine kinase, and GAPDH (18 -20). Studies show that membrane-associated GAPDH is able to support an ATP-regenerating system at the SR, which is tightly coupled to the calcium transport function (18 -20). How the glycolytic machinery is targeted to the SR membrane and whether this local ATP-generating system could be regulated in response to the calcium signal and muscle contraction remains to be defined. Recent studies also suggest that the local concentration of NADH can regulate the calcium transport activities of the SR (21). How modulations in NADH concentrations at the SR membrane can be achieved remains unknown.
Here, we examined the functional significance of the SR membrane-bound CaMKII␤ M . We have utilized a proteomicsbased approach to identify potential binding partners of the muscle-specific CaM kinase. This approach has identified several enzymes of the glycolytic pathway, including GAPDH, pyruvate kinase, and LDH, as well as enzymes involved in glycogen mobilization, such as phosphorylase b kinase and the debranching enzyme. Importantly, our results reveal that CaMKII␤ M can phosphorylate these associated proteins and specifically increase the activity of GAPDH in a Ca 2ϩ /CaM-dependent manner. In view of the previous studies that implicate a role for local ATP and NADH in the regulation of SR function and excitation-contraction coupling, our results suggest that the membrane-bound CaMKII␤ M may be important for targeting the glycolytic machinery to the SR and modulate the local levels of NADH and ATP in response to calcium signaling in skeletal muscle.

EXPERIMENTAL PROCEDURES
Cloning of Muscle-specific CaM Kinases and Its Expression-An oligo(dT)-primed lambda Zap II cDNA library from skeletal muscle (Stratagene) was screened with a 500-bp DNA probe encompassing the coding region of the variable domain of rat brain CaMKII. The probe was generated by reverse transcription-PCR using rat brain RNA and ␤ subunit-specific primers (5Ј-CTG AAG AAG TTC AAT GCA AGG AGG-3Ј or 5Ј-GCA GCA GGT TCT CAA AGT AGA ATC-3Ј) and labeled with [ 32 P]dCTP using a random prime labeling kit from Amersham Biosciences. Hybridization was performed at 65°C in 10% polyethylene glycol, 1.5ϫ saline/sodium phosphate/EDTA, and 7% SDS for 12 h, then followed by four washes in 1-0.2ϫ SSC, 0.1% SDS at 50°C. The membranes were exposed to Biomax MR films, and cDNAs identified on tertiary screening were isolated and subcloned into pBluescript SK by in vivo excision as per Stratagene. Positive clones were categorized by restriction endonuclease mapping and size analysis. Overlapping clones were sequenced on both the strands by an automated ABI sequencer using M13F and M13R primers and sequences analyzed with Sequaid II (University of Kansas) and BLAST.
Full-length skeletal muscle CaMK␤II M as well as the brain isoforms were cloned in-frame with the glutathione S-transferase (GST) fusion protein in pGEX 3x⌬B vector in EcoRI site. Escherichia coli DH5␣ containing recombinant proteins were induced with 0.1 mM isopropyl-1-thio-␤-D-galactopyranoside. The GST fusion proteins were isolated on glutathione-Sepharose 4B beads (Amersham Biosciences) following sonication in PBST containing 1% Nonidet P-40.
Isolation of Cytosolic and SR Membrane Proteins-The preparation of SR vesicles was based on previously described protocols (11,22). Freshly dissected back muscle from rabbits was washed in cold phosphate-buffered saline, trimmed of fat and connective tissue, cut into cubes, and either frozen in liquid N 2 and stored at Ϫ110°C or freshly processed. The fresh or thawed muscle cubes were suspended in homogenization buffer (50 mM Tris, pH 7.4, 300 mM sucrose, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, and 10 g/ml pepstatin; in 5 ml/g tissue). The muscle was homogenized in a Waring blender with 5ϫ 20-s high speed bursts. The homogenate was centrifuged at 2600 ϫ g for 10 min, and the supernatant was filtered through cheesecloth and centrifuged at 10,000 ϫ g for 10 min. The supernatant so obtained was centrifuged at 186,000 ϫ g for 1 h, cytosolic supernatant was saved, and SR membrane preparation was re-suspended in buffer (10 mM Tris, pH 7.4, 400 mM sucrose, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, and 10 g/ml pepstatin) and dispersed with 15 strokes of a Dounce homogenizer. SR proteins were aliquoted and stored at Ϫ110°C for further use. Protein quantification of each fraction was performed according to Bradford (23).
A GAPDH-catalyzed reaction was assayed at 25°C by monitoring the reduction of NAD ϩ at 340 nm in an assay mixture containing 25 mM Tris buffer (pH 7.0), 0.1 mM NAD ϩ , and 0.1 mM glyceraldehyde 3-phosphate (GAP). The reaction was initiated by adding an appropriately diluted enzyme solution in the above reaction mixture. For reactions studying CaM kinase-dependent phosphorylation of GAPDH with CaMKII␤ M , the reaction was started by addition of GAP.
Aldolase assay was carried out by measuring the change in absorbance at 340 nm (A 340 /min) in a coupled assay as described previously (26). Aldolase was diluted in 50 mM triethanolamine-HCl (pH 7.4) and added to a cuvette containing 10 mM EDTA, 0.16 mM NADH, and 10 g/ml GAPDH/triose phosphate isomerase in 50 mM triethanolamine buffer (pH 7.4). Assays (1 ml) were performed in triplicate at 30°C, and the reaction was initiated by addition of substrate.
Pyruvate kinase activity was determined in an LDH-coupled assay system by measuring the A 340 /min resulting from the oxidation of NADH as described by Imamura and Tanaka (27). The assay mixture (1 ml) contained 0.1 M KCl, 5 mM MgCl 2 , 2 mM ADP, 0.17 mM NADH, 2 mM phosphoenol pyruvate and 2 units of LDH in 50 mM Tris-HCl buffer (pH 7.4) at 30°C. The reaction was initiated by addition of phosphoenol pyruvate.
LDH activity was measured (28) by pyruvate reduction in the presence of NADH with NAD ϩ formation in a 1-ml reaction mixture containing 50 mM potassium phosphate buffer (pH 7.5), 0.18 mM NADH, 0.6 mM sodium pyruvate, and appropriately diluted enzyme at 25°C. The production of NAD ϩ was followed by a decrease in absorbance at 340 nm.
Glycogen phosphorylase activity was measured as described by Helmreich and Cori (29). The assay mixture contained 50 mM potassium phosphate (pH 7.1), 1 mM MgCl 2 , 0.2% glycogen, 50 nM glucose 1,6-diphosphate, 0.6 mM NADP, 4 units of phosphoglucomutase, and 0.8 unit of glucose-6-phosphate dehydrogenase. In a series of experiments the effects of CaMKII␤ M phosphorylation on the activity of this enzyme were determined in the absence of AMP in the above buffer.
GST Pull-down Assays-GST pull-down assays were carried out by immobilizing GST and GST-tagged CaMKII␤ M to glutathione-Sepharose 4B beads. Immobilized fusion proteins (12 g) were incubated with SR membrane or cytosolic proteins (500 g) in 1 ml of TBS (50 mM Tris, 100 mM NaCl, pH 7.4, plus 0.1% Triton X-100 with 1 mM free Ca 2ϩ ) buffer for 2 h at 4°C. The salt concentrations in the buffer used were similar to that used by others (30). Beads were washed four times in TBS. Proteins were visualized with Coomassie Blue or silver staining and identified by immunoblotting with antibodies using enhanced chemiluminescence (ECL).
Phosphorylation of Substrate Proteins-Phosphorylation of target proteins of CaMKII␤ M was analyzed by a GST pull-down-coupled phosphorylation reaction. GST pull-down proteins as obtained in the pulldown assay were subjected to a Ca 2ϩ /CaM-dependent phosphorylation reaction at 30°C in MOPS-NaOH buffer (50 mM, pH 7.4), 10 mM magnesium chloride, 0.4 mM [␥-32 P]ATP (0.4 ϫ 10 Ϫ5 cpm), 40 g/ml calmodulin, 5 mM dithiothreitol, and 100 M [Ca 2ϩ ] Free (ϩ) or 5 mM EGTA (Ϫ) for 15 min. After the incubation, the reaction was terminated by adding sample buffer to the reaction mixture, and proteins were resolved on SDS-PAGE and further detected on Kodak BioMax MR film.
Phosphorylation of each individual target protein was also analyzed by incubating GST-CaMKII␤ M (0.03 mg/ml) with purified protein (0.5 mg/ml) at 4°C for 1 h followed by a phosphorylation reaction as described above. At the end of incubation, aliquots of the reaction mixture were applied on Whatman 3MM filter papers, washed with 5% trichloroacetic acid, and dried. Radioactivity was determined by liquid scintillation spectrometry. Control samples contained only the protein kinase sample in the reaction.
Protein Identification by Mass Spectrometry-Protein bands were excised, destained, and then subjected to tryptic digestion to produce peptides in the range of 1000 -3000 Da. The peptides were concentrated and desalted using C 18 Zip-Tips TM (Millipore, Bedford, MA). MALDI-TOF mass spectrometry peptide mass mapping was performed using a Waters M@LDI and Q-TOF mass spectrometer equipped with a 337-nm nitrogen laser at an accuracy of 0.05-0.15 atomic mass unit by Dr. David Hyndman at the Queens University Facility.

Distribution of CaMKII␤ M in Cellular
Fractions-Subcellular distribution of muscle-specific CaMKII␤ M was carried out by Western blot analysis of SR and cytosolic fractions from fast twitch skeletal muscle with CaMKII␤ monoclonal antibody (Cb␤-1). This identified a CaMKII␤ M polypeptide of ϳ73 kDa in SR (Fig. 1, lane 2) but not the cytosolic fraction (Fig. 1, lane 1), consistent with that published previously (11). The recombinant kinase was expressed as a GST fusion protein, and the molecular mass of GST-conjugated CaMKII␤ M was 98 kDa with several truncated products of ϳ73, 64, 56, 41, and 30 kDa being also detected by CaMKII␤ antibody (Fig. 1, lane 3).
The functional activity of recombinant CaMKII␤ M was examined in an autocamtide-2 assay system (Upstate CaM kinase assay kit), and this indicated an increase of ϳ90-fold in phosphotransferase activity in the presence of calcium (100 M [Ca 2ϩ ] Free ) and CaM (2 M), which corresponds to ϳ4.86 pmol/min/g. Molecular Interactions of the Muscle-specific CaMKII␤ M -To determine the functional significance of the SR membranebound CaMKII␤ M , identification of the potential interacting proteins was carried out using GST pull-down assays with cytosolic and detergent-soluble SR fractions from skeletal muscle. Characterization of the interacting proteins with respect to molecular mass revealed that recombinant GST-CaMKII␤ M specifically binds polypeptides of ϳ55, 47, 44, 39, 36, and 32 kDa from the cytosol (Fig. 2A, lane 2) and of ϳ177, 97, 61, 44, and 36 kDa from the SR membrane fraction (Fig. 2B, lane 2). There was no binding of these polypeptides to GST alone ( Fig Fig. 2B, lane 1). The expressed kinase and its truncated products have been specifically labeled in parentheses in Fig. 2 (A, lane 1 and B, lane 1) and showed a similar migration pattern as seen in Western blot analysis with anti-CaMKII␤ antibody (see Fig. 1, lane 3). The 25-kDa GST protein visualized in the 10% SDS gel in panel A migrated out of the 7.5% gel in panel B, and a nonspecific polypeptide of ϳ50 kDa was visualized in each lane.
The Glycolytic Enzymes Are Interacting Proteins of CaMKII␤ M -To further identify the specifically bound polypeptides, MALDI-TOF was employed, and this revealed the interaction of CaMKII␤ M with enzymes of glycolytic pathway and glycogen metabolism (Table I). The interacting proteins from cytosolic fraction (Table I and Fig. 2A, lane 2) were determined to be LDH and enzymes of the Embden-Meyerhof-Parnas pathway, namely aldolase A, GAPDH, enolase, and pyruvate kinase, whereas those from the SR fraction (Table I and Fig. 2B, lane 2) revealed an identity with glycogen phosphorylase b, glycogen debranching enzyme, creatine kinase, and GAPDH. The data revealed that GAPDH and creatine kinase were present as interacting proteins in both cytosolic and SR membrane fractions. An interacting polypeptide of 61 kDa in the pulldown assay with the SR proteins (Fig. 2B, lane 2, asterisk) and another of 275 kDa (not shown) in pull-down assay with cytosolic proteins were determined to be CaMKII␤, indicating the same class of kinase-kinase interactions. In this regard, the CaM kinase family of enzymes is known to exist as oligomers, and these data are consistent with the self-assembly of these molecules into multisubunit enzyme complexes (1). Two additional polypeptides of 138 and 125 kDa from cytosolic fraction were also seen when the pull-down proteins were resolved on 7.5% SDS-PAGE (data not shown), which were identified by MALDI-TOF as phosphorylase b kinase ␣ and ␤ chains, respectively (Table I).
The Glycolytic Enzymes Are Substrates of CaMKII␤ M -Because the protein-protein interaction assays indicated a specific interaction with the molecules of the glycolysis pathway, we examined whether these polypeptides were phosphorylated by CaMKII␤ M . The kinase-associated proteins isolated in the GST pull-down assay were examined in a phosphorylation reaction in the absence and presence of Ca 2ϩ and CaM, and the autoradiogram showing 32 P incorporation is shown in Fig. 3. A number of the interacting proteins were phosphorylated in a Ca 2ϩ /CaM-dependent manner in cytosolic (Fig. 3, panel b) and SR membrane (Fig. 3, panel c) fractions of skeletal muscle tissue. These substrate proteins correspond to the polypeptides of ϳ55, 47, 44, 39, and 36 kDa from the cytosol (Fig. 3, panel b), which were identified as pyruvate kinase, enolase, creatine kinase, aldolase, and GAPDH, respectively, and the ϳ97-, 61-, and 36-kDa polypeptides from the SR membrane extract (Fig.  3, panel c), which correspond to glycogen phosphorylase b, CaMKII␤, and GAPDH, respectively. When these polypeptides were separated on SDS-PAGE, the phosphorylated protein bands appeared at approximately corresponding positions and order (Fig. 3, panels b and c), as found in GST pull-down experiments (compare with Fig. 2, A and B), suggesting the Ca 2ϩ /CaM-dependent 32 P incorporation into these interacting proteins was enhanced by CaMKII␤ M . Autophosphorylation of CaMKII␤ M fusion protein and its truncated products is seen as 98-, 73-, 64-, 56-, and 41-kDa bands (Fig. 3, panel a).
The phosphorylation of the substrate proteins was further confirmed by reconstituting the individual purified protein with purified recombinant CaMKII␤ M in the presence and absence of Ca 2ϩ and CaM. These data shown in Table II indicated that pyruvate kinase, LDH, and GAPDH were good substrates of CaMKII␤ M . These data are consistent with previous studies that indicate that GAPDH, LDH, and pyruvate kinase are associated with the SR membrane (21,31,32) and phosphorylated by the endogenous CaMKII (31, 32).
CaMKII␤ M Specifically Modulates the Activity of GAPDH-To determine the functional effects of CaMKII␤ M -induced phosphorylation on the substrate proteins, the activity of the various glycolytic enzymes was determined in the absence and presence of Ca 2ϩ and CaM in a reconstituted kinase reaction. The data in Table II indicate that there was no effect of phosphorylation on the enzyme activity of aldolase, pyruvate kinase, LDH, creatine kinase, and glycogen phosphorylase b, however, the activity of GAPDH was significantly increased (ϳ3.4-fold) by CaMKII␤ M in the presence of Ca 2ϩ /CaM.
Because the SR membrane-associated GAPDH activation is believed to directly support ATP-coupled Ca 2ϩ transport and the CaMKII␤ M is specifically targeted to the SR, we sought to determine whether GAPDH would reside in the membrane in direct association with the kinase. To assess this, cytosolic and detergent-soluble SR membrane proteins were first examined for GAPDH content by Western blot analysis with anti-GAPDH antibody, and this revealed the distribution of GAPDH in both cytosolic and SR membrane fractions (Fig. 4A). These data are consistent with the previous studies on GAPDH distribution in skeletal muscle (33). Furthermore, GST pull-down assays were performed using purified GST-CaMKII␤ M and GST alone from cytosolic and detergent-solubilized SR membranes. Fig. 4B (top  panel) indicates that the CaMKII␤ M fusion protein specifically binds a 36-kDa polypeptide that is recognized by the anti-GAPDH antibody from both cytosolic (Fig. 4B, lane 3) and SR membranes (Fig. 4B, lane 5). The GST alone did not exhibit binding of any immunoreactive polypeptide from cytosolic or SR membrane fractions (Fig. 4B, lanes 4 and 6) similar to that seen with the buffer control for the GST alone and the CaM kinase fusion protein (Fig. 4B, lanes 1 and 2). The immunoblot was stripped and stained with anti-GST antibody as a control for protein loading (Fig. 4B, bottom panel).
To further assess the ability of the muscle-specific CaMKII␤ M and the soluble CaMKII isoform (brain-specific CaMKII␤ B ) to interact with GAPDH, we conducted GST pulldown assays with the SR fraction. Fig. 5A (top panel), shows a Western blot analysis with anti-GAPDH antibody, which revealed that both CaMKII␤ M (Fig. 5A, lane 3) and CaMKII␤ B (Fig. 5A, lane 5) specifically bind GAPDH. GST-alone constructs did not bind GAPDH as noted above (Fig. 5A, lanes 4  and 6). Neither CaMKII␤ M nor CaMKII␤ B showed any GAPDH binding when incubated with the TBS buffer control (Fig. 5A,  lanes 1 and 2). The immunoblot was stripped and stained with anti-GST antibody, which revealed identical protein loading for each lane in the assay (Fig. 5A, bottom panel). The data from these blots was quantified by densitometry (Fig. 5B) and revealed that ϳ7-fold higher amount of GAPDH was in complex  -associated complex in skeletal muscle Cytosolic and SR membrane proteins from skeletal muscle were incubated with GST alone and GST-CaMKII␤ M . Bound proteins were washed, eluted, and resolved in SDS-PAGE (7.5 or 10%) and finally stained with Coomassie. Protein bands were excised, destained, and subjected to tryptic digestion to produce peptides in the range of 1000 -3000 Da. The peptides were concentrated, and MALDI-TOF mass spectrometry peptide mass mapping was performed using a Waters mass spectrometer equipped with a 337-nm nitrogen laser at an accuracy of 0.05-0. 15  with the muscle-specific CaMKII␤ M compared with the brain isoform (Fig. 5B).
To further examine the association of endogenous CaMKII␤ M and GAPDH in skeletal muscle SR membranes, we performed immunoprecipitation assays with anti-CaMKII␤ monoclonal antibody (Cb␤-1) of detergent-solubilized SR extracts, which was analyzed for the presence of GAPDH. Fig. 5C  (lane 1) shows that anti-CaM kinase antibody can immunoprecipitate endogenous ϳ73-kDa CaMKII␤ M polypeptide from SR as analyzed by Western blot by CaMKII␤ antibody. The same IP complex also contained the ϳ36-kDa polypeptide that was recognized by anti-GAPDH antibody (Fig. 5C, lane 2) further implying a potential interaction between these two proteins. To assess whether CaMKII␤ M can directly interact with GAPDH, we carried out pull-down assays with purified GAPDH and CaMKII␤ M -GST and analyzed this association in Western blots with anti-GAPDH antibody (Fig. 5D, top panel). The CaMKII␤ M -GST was able to bind directly and effectively with GAPDH (Fig. 5D, lane 1) compared with GST alone (Fig. 5D,   lane 2), even when GST alone was present at about a ϳ10-fold higher concentration as seen by immunoblotting with anti-GST (Fig. 5D, bottom panel, lanes 1 and 2).
Because CaMKII␤ M could directly bind GAPDH, we examined the kinetics of phosphorylation and activation of GAPDH by the kinase. In a phosphorylation assay of the CaMKII␤ M using purified GAPDH as a substrate, we found that increasing phosphorylation on activity of the glycolytic complex Phosphorylation of target proteins was studied in the presence of 100 M ͓Ca 2ϩ ͔ Free (ϩ) or 5 mM EGTA (Ϫ) and 2 M CaM at 30°C for 60 min (see legends of Fig. 3 also). At the end of incubation, aliquots of the reaction mixtures were applied on Whatman 3MM filter papers, washed with 5% trichloroacetic acid, and dried, and radioactivity was determined by liquid scintillation spectrometry. Data are mean Ϯ S.D. of at least three independent experiments. The enzyme activity determination methodology of each individual enzyme is described under "Experimental Procedures."   (Fig. 6, A and B, respectively). Using optimized concentrations of Ca 2ϩ and CaM, this corresponds to incorporation of 0.96 mol of 32 P per mol of GAPDH tetramer (Fig. 6C). Finally, this phosphorylation was accompanied by a 3.4-fold increase in enzyme activity of GAPDH (Fig. 6D). DISCUSSION Ca 2ϩ /CaM-dependent phosphorylation does not appear to regulate the RyR or the Calcium pump of the SR in fast twitch skeletal muscle (7,8,15,34,35). A CaMKII␤ isoform referred to as a muscle-specific CaMKII␤ M was shown to be targeted to the SR membrane in skeletal muscle by a non-kinase protein, ␣-KAP (11,12), but the role of this kinase in SR function remains unknown. In this study, we found that CaMKII␤ M specifically associates with the enzymes of glycolysis and glycogen metabolism. The MALDI-TOF identified glycogen phosphorylase, glycogen debranching enzyme, GAPDH, and creatine kinase as interacting proteins from the SR fraction and aldolase A, GAPDH, enolase, LDH, creatine kinase, pyruvate kinase, and phosphorylase b kinase from the cytosolic fraction. Previous studies have indicated that glycogen phosphorylase resides at the SR membrane in association with glycogen and may account for its mobilization to glucose (18,36). Our data suggest that CaMKII␤ M can complex with phosphorylase b kinase at the SR, and the later enzyme has previously been established to be regulated directly by Ca 2ϩ /CaM binding (36,37). Our data here also identify the glycogen debranching enzyme as a novel component at the SR, which would further assist in glycogen mobilization. Furthermore, the data indicate that CaMKII␤ M can interact with GAPDH to phosphorylate and modulate its activity in a Ca 2ϩ -and CaM-dependent manner. GAPDH was found to be distributed in muscle cytosol and the SR membrane, and ultrastructural localization studies have also revealed that GAPDH along with other glycolytic enzymes are bound to the cytoplasmic face of the SR membrane in skeletal muscle (33,38). A role for GAPDH together with 3-phosphoglycerate kinase (PGK) has been previously established in the production of ATP from GAP, NAD ϩ , P i , and ADP at the level of the SR membrane (19,20,39,40). In addition, a recent study demonstrated that synaptic vesicles were capable of generating local ATP via the membrane-associated GAPDH system, which could support the accumulation of the excitatory neurotransmitter glutamate even during significantly reduced global cellular ATP concentrations (41). It is notable that synaptic vesicles contain a CaMKII␤ isoform (42), which we found to associate with GAPDH but to a much less degree when compared with the muscle-specific isoform of the kinase. How- ever, these studies do imply that CaMKII isoforms may target GAPDH to distinct membrane systems such as the SR in muscle and synaptic vesicles in neurons to regulate local ATP production in response to changes in free calcium. Thus, these studies point to a previously unrecognized role for the multifunctional CaMKII in regulating membrane ATP through targeting and activation of glycolytic enzymes such as GAPDH in response to the calcium signal. Although, PGK and GAPDH normally exist in a complex and at the SR membrane (20,39,(41)(42)(43), we were unable to detect this in our assay conditions, because this association is fragile and disrupted in 0.1 M KCl (44).
Thus, collectively, the findings here place the glycogen mobilizing and glucose-metabolizing enzymes in a complex with CaMKII␤ M at the SR membrane, and we propose a model in which this complex could regulate SR function (Fig. 7). The assembly of glycolytic machinery with the CaMKII␤ M at the SR membrane would assist in the local regulation of ATP to support the Ca 2ϩ regulatory function of this membrane. Ca 2ϩ release from the SR would lead to the activation of phosphorylase b kinase through direct Ca 2ϩ /CaM binding, which would mobilize glycogen to glucose via the phosphorylase b and debranching enzyme cascade (36,37). The Ca 2ϩ /CaM would also activate CaMKII␤ M , and a marked up-regulation of GAPDH would result in enhanced ATP and NADH production at the SR membrane. ATP can then act as a ligand for the RyR and the Ca-ATPase. The NADH generated locally could also serve to regulate calcium release and uptake as reported previously (21,45,46). Thus, our data may be of some significance in muscle function, because changes in calcium and consequently the activity of CaMKII␤ M could allow for a novel mechanism for the regulation of SR function via increase in GAPDH activity targeted to this membrane system.