Metabolic Control of Ca2+/Calmodulin-dependent Protein Kinase II (CaMKII)-mediated Caspase-2 Suppression by the B55β/Protein Phosphatase 2A (PP2A)*

Background: CaMKII autophosphorylates at Thr286 in cell-free lysates supplemented with glucose-6-phosphate (G6P). Results: The B55β and C subunits of PP2A interact with and dephosphorylate CaMKII at novel sites following addition of G6P. Conclusion: B55β/PP2A is critical for stimulating CaMKII in the presence of G6P. Significance: This work identifies a novel role for a specific phosphatase complex in controlling CaMKII. High levels of metabolic activity confer resistance to apoptosis. Caspase-2, an apoptotic initiator, can be suppressed by high levels of nutrient flux through the pentose phosphate pathway. This metabolic control is exerted via inhibitory phosphorylation of the caspase-2 prodomain by activated Ca2+/calmodulin-dependent protein kinase II (CaMKII). We show here that this activation of CaMKII depends, in part, on dephosphorylation of CaMKII at novel sites (Thr393/Ser395) and that this is mediated by metabolic activation of protein phosphatase 2A in complex with the B55β targeting subunit. This represents a novel locus of CaMKII control and also provides a mechanism contributing to metabolic control of apoptosis. These findings may have implications for metabolic control of the many CaMKII-controlled and protein phosphatase 2A-regulated physiological processes, because both enzymes appear to be responsive to alterations in glucose metabolized via the pentose phosphate pathway.

In the Xenopus egg extract system, caspase-2 has also been tied to metabolic control of apoptosis (9 -11). We have reported that caspase-2 is important for recapitulating apoptotic events in this system and that its activity can be modulated by controlling the metabolic status of the egg extracts. Specifically, incubation of extracts at room temperature reduced levels of pentose phosphate pathway (PPP)-generated 3 NADPH, and supplementation of extracts with NADPH or PPP stimulatory glucose-6-phosphate (G6P) greatly delayed caspase-2 activation and ensuing apoptotic events (9). Biochemical analyses revealed that metabolic inhibition of caspase-2 was caused by inhibitory phosphorylation within the caspase-2 prodomain at Ser 135 (Xenopus numbering). Using kinase inhibitors and immunodepletions, we found that this phosphorylation was catalyzed by the Ca 2ϩ /calmodulin (CaM)-dependent protein kinase II (CaMKII) and that CaMKII activity was elevated following G6P or NADPH treatment of extracts (9).
Four highly similar isoforms exist of CaMKII, which is an important mediator of many Ca 2ϩ -induced signaling pathways (12)(13)(14)(15). Each isoform contains a catalytic domain near the N terminus, an autoregulatory domain, and a C-terminal association domain (16). When inactive, pseudosubstrate sequences bind and inhibit the catalytic domains (17). Ca 2ϩ /CaM binding disrupts catalytic and autoinhibitory domain interaction, activating the kinase and allowing access to an autophosphorylation site (Thr 286 , ␣ isoform) (18). Once activated, within the holoenzyme, one subunit phosphorylates an adjacent subunit at Thr 286 when both are bound to Ca 2ϩ /CaM (19). Once phosphorylated on Thr 286 , the Ca 2ϩ /CaM off-rate drops over 1000fold, stabilizing CaMKII activity (20). Therefore, the autophosphorylation of Thr 286 can be used as an indicator of CaMKII activation. Following Ca 2ϩ /CaM dissociation, Thr(P) 286 CaM-KII remains active, and further autophosphorylation occurs at Thr 305 , Thr 306 , and Ser 314 (21,22).
Recently, the Nutt laboratory reported that CoA, generated in Xenopus egg extracts in the presence of abundant nutrients, binds to and activates CaMKII (23). We show here that nutrientdriven CaMKII activation additionally requires release of a "brake." Specifically, we identify two novel sites of CaMKII phosphorylation (Thr 393 /Ser 395 on the Xenopus ␥ isoform L subunit and Thr 371 /Ser 373 on the human homolog) located within the association domain, whose phosphorylation falls in * This work was supported, in whole or in part, by National Institutes of Health Grant RO1 GM080333. 1 To whom correspondence may be addressed. the presence of high G6P levels. Dephosphorylation of these sites, catalyzed by protein phosphatase 2A (PP2A), is necessary (albeit not sufficient) for metabolic activation of CaMKII. In addition, nutrient-driven PP2A targeting to CaMKII is driven by metabolically regulated interaction of CaMKII with the PP2A targeting subunit B55␤. Furthermore, this mechanism of CaMKII regulation is conserved in mammalian cells. Together, these findings provide insight into metabolic control of apoptosis and define a new mechanism for controlling CaMKII, a protein critical for cell signaling in response to multiple stimuli.

EXPERIMENTAL PROCEDURES
Preparation of Xenopus Egg Extracts and Nutrient Treatment-Xenopus egg extracts were prepared as previously described (24). G6P was prepared as a 1 M solution in water. Extracts were prepared at 4°C, treated with G6P at a final concentration of 20 mM, and incubated at room temperature.
Cell Culture and Nutrient Treatment-HEK 293T cells were grown in DMEM with 10% FBS medium at 37°C. Before nutrient treatment, cells were starved with glucose-free DMEM with 10% dialyzed FBS medium containing no D-glucose and sodium pyruvate at 37°C for 12 h and then treated with or without 25 mM D-glucose (Sigma) for another 12 h. Cells were lysed in 50 mM Tris, pH 7.5, 150 mM NaCl, 1 mM DTT, and 1% Nonidet P-40 with 5 g/ml aprotonin/leupeptin and 100 M PMSF and phosphatase inhibitors (PhosSTOP Phosphatase Inhibitor Mixture Tablets from Roche, 20ϫ) on ice.
For immunoprecipitation, two micrograms of anti-CaMKII␣ (Sigma) or Mouse control IgG were incubated with 20 l of Dynabead protein G (Invitrogen) slurry overnight at 4°C. 100 l of Xenopus egg extracts treated with or without G6P were incubated with beads for 1 h at 4°C. Beads were washed four times with wash buffer (described below) and eluted with SDS-PAGE sample buffer. Samples were resolved by SDS-PAGE for immunoblotting.
For immunodepletion, 10 g of PPP2R2B antibody (0.25 g/l) or rabbit control IgG were incubated with 100 l of Dynabead protein A (Invitrogen) overnight at 4°C. Beads were washed and divided into three equal parts. 90 l of Xenopus egg extracts were incubated with beads for 1 h at 4°C. This step was repeated three times.
Phospho-antibody Purification-The following peptides were synthesized to generate and purify phospho-antibody against Thr 393  . Sera were run through a column of non-phospho-peptide conjugated with UltraLink Biosupport (Thermo Scientific) and were then purified on a column of phospho-peptide.
Gel Filtration Chromatography-Xenopus egg extracts were treated with G6P or water and co-treated with okadaic acid or DMSO. After incubation at room temperature for 4 h, the extracts were centrifuged at 200,000 ϫ g and were fractionated on Superose 6 (GE Healthcare Life Sciences).
Nano-Flow Liquid Chromatography Electrospray Ionization Tandem Mass Spectrometry (LC-MS/MS) Analysis-Xenopus egg extracts were treated with G6P or water, and CaM-Sepharose was added (Agilent Technologies). After incubation, the CaM-Sepharose beads were collected by centrifugation and washed four times with 1ϫ egg lysis buffer containing 500 mM NaCl and 0.5% Triton X-100 and then eluted with RapiGest SF Surfactant (Waters).
Following buffer exchange into 50 mM ammonium bicarbonate, pH 8.0, samples were subjected to a standardized reduction/alkylation procedure followed by overnight trypsin digestion according to the in-solution tryptic digestion protocol established by the Duke Center for Genomic and Computational Biology. Peptides were then either analyzed directly by LC-MS/MS analysis or subjected to a phosphopeptide enrichment with a 200 l of TiO 2 Protea Tip (Protea Bio) as per the manufacturer's recommended protocol. Samples were then subjected to LC-MS/MS analysis using a Waters NanoAquity UPLC equipped with a 1.7-m BEH130 C18 75-m inner diameter ϫ 250-mm reversed phase column employing a 90-min gradient at 300 nl/min from 5% acetonitrile, 0.1% formic acid to 40% acetonitrile, 0.1% formic acid. Eluting peptides were analyzed on a Thermo LTQ-Orbitrap XL mass spectrometer set to acquire a precursor MS scan in the Orbitrap from m/z 400 -2000 with r ϭ 60,000 at m/z 400 and a target AGC setting of 1e6 ions. MS/MS spectra of the five most abundant precursor ions were acquired either in the Orbitrap with r ϭ 7500 at m/z with a target AGC setting of 2e5 ions for nonenriched samples or in the ion trap with a target AGC setting of 1e3 for enriched samples. Max fill times were set to 1000 ms for full MS scans and either 500 ms for Orbitrap MS/MS scans or 250 ms for ion trap MS/MS scans with minimum MS/MS triggering thresholds of 5000 counts. For all experiments, fragmentation occurred in the LTQ linear ion trap with a CID energy setting of 35%, and a dynamic exclusion of 60 s was employed for previously fragmented precursor ions.
Raw LC-MS/MS data files were processed in Mascot distiller (Matrix Science) and then submitted to independent Mascot searches (Matrix Science) against an Trembl database (v 40.14 Xenopus laevis taxonomy, 12,530 forward entries) containing both forward and reverse entries of each protein. Search tolerances for LTQ-Orbitrap XL data were 5 ppm for precursor ions and 0.02 Da for Orbitrap product ions or 0.8 Da for ion trap product ions using full trypsin specificity. Carbamidomethylation (ϩ57.0214 Da on Cys) was set as a fixed modification, whereas oxidation (ϩ15.9949 Da on Met) and phosphorylation (ϩ79.9663 Da on Ser, Thr, and Tyr) were considered variable modifications. All searched spectra were imported into Scaffold (v4.0, Proteome Software), and mascot ion scoring thresholds of 26 (p 0.05 Mascot identity score was 14.0) were set to achieve a false discovery rate of 0.0%. Probability of correct phosphorylation modification localization to a specific Ser, Thr, or Tyr residue was performed by submitting each MS/MS spectrum to the AScore algorithm with AScores of Ͼ15 or Ͼ19, indicating a Ͼ90% or Ͼ99% probability of correct localization, respectively.
Relative quantitation was performed in Skyline (v1.4.1, University of Washington) by applying the full MS precursor extracted ion chromatogram function to integrate and measure peak areas (area under the curve) of each identified phosphopeptides. Reported area under the curve measurements for each phosphopeptide were the sum of the monoisotopic peak extracted ion chromatogram, as well as the second and third isotopomer extracted ion chromatogram. To adjust for slight variations in starting CaMKII prior to TiO 2 enrichment, the average area under the curve of three nonphosphorylatable (i.e. did not contain a STY) peptides from the nonenriched LC-MS samples were used to generate a correction factor.
Kinase Assay-40 l of glutathione-Sepharose containing 4 g of recombinant GST tagged caspase-2 prodomain or GSTonly fusion proteins were incubated in B55␤-depleted or undepleted egg extracts together with 20 Ci of [␥-32 P]ATP, treated by 20 mM G6P or water, at 30°C for 1 h. Samples were washed, eluted with SDS-PAGE sample buffer, and resolved by SDS-PAGE for autoradiography.

RESULTS
CaMKII Activity Is under Metabolic Control-As we reported previously, treatment of Xenopus egg extracts with G6P should elevate the kinase activity of CaMKII (9). We first tested the autophosphorylation of Thr 286 , an indicator of CaMKII activation. By directly treating egg extracts with or without G6P, we discovered that G6P increased phosphorylation of Thr 286 on CaMKII, a known site of CaMKII autophosphorylation ( Fig. 1A; note that the multiple phosphorylated bands are FIGURE 1. CaMKII activation is sensitive to metabolic status. A, Xenopus egg extracts were treated with or without G6P for 0. 5 h at room temperature and analyzed for CaMKII Thr 286 autophosphorylation using a Thr(P) 286 antibody. Note that the middle and bottom panels are from two different films as the CaMKII antibody recognized CaMKII␣ (ϳ50 kDa) much more strongly than the other isoforms, so detection of this isoform and the others required very different exposures. B, GST tagged Xenopus caspase-2 prodomain or GST bound to glutathione-Sepharose was incubated with Xenopus egg extracts supplemented with [␥-32 P]ATP and treated with or without G6P. Samples were resolved by SDS-PAGE and detected by autoradiography. CB, Coomassie Blue. C, glucose-starved 293T cells treated with or without glucose (25 mM) were lysed and analyzed for CaMKII Thr 286 autophosphorylation. D, glucose-starved 293T cells treated with or without glucose and with or without dehydroepiandrosterone (DHEA) were lysed and analyzed for CaMKII Thr 286 autophosphorylation.
likely due to multiple isoforms and allelic variants in the pseudotetraploid X. laevis). Although the total CaMKII antibody used here recognized predominantly the ␣ isoform, with increased exposure time, additional CaMKII bands became evident, corresponding to multiple bands of Thr 286 phosphorylation. This suggests that the observed increase in Thr 286 phosphorylation following G6P treatment is likely generalizable to multiple CaMKII isoforms. Additionally, as we reported previously, the Xenopus caspase-2 prodomain (which we showed to be a CaMKII substrate (9)) added to egg extracts was more heavily phosphorylated in the presence of G6P, again consistent with the notion that the kinase activity of CaMKII is up-regulated by G6P (Fig. 1B). As also reported, CaMKII can physically bind the caspase-2 prodomain, which is stimulated by G6P treatment (25). Interestingly, an up-regulation of Thr 286 phosphorylation was also observed in human 293T cells after glucose starvation and resupplementation (Fig. 1C). As we reported in Xenopus egg extracts, the increased activity of CaMKII depended on the activity of the PPP; glucoseinduced Thr 286 phosphorylation was reduced by co-treatment of 293T cells with dehydroepiandrosterone, an allosteric inhibitor of glucose-6-phosphate dehydrogenase, suggesting that the regulation of Thr 286 phosphorylation by glucose is also through the PPP (Fig. 1D).
The Phosphorylation Status of CaMKII Is Altered by G6P Treatment of Egg Extracts-Although phosphorylation of Thr 286 on CaMKII is stabilized following G6P treatment (25), it was not determined whether other modifications of CaMKII, upstream of activation, are modulated by metabolism. To examine the status of CaMKII in G6P-treated egg extracts, we resolved G6P-treated and untreated extracts by gel filtration and examined the profile of CaMKII fractionation. As shown in Fig. 2A, the apparent molecular weight of the CaMKII complex was increased following G6P treatment. It is possible that this G6P-induced shift in the CaMKII fractionation profile is caused by the association of additional proteins with CaMKII, because even in the untreated extract, the molecular weight of CaMKII was above 600,000, consistent with the dodecameric (potentially active) form; note that the shift might also be caused by the incorporation of other, higher molecular weight CaMKII isoforms into the CaMKII holoenzyme.
To further investigate the mechanism of CaMKII activation, we looked for post-translational modifications of CaMKII induced by G6P treatment. Endogenous CaMKII was precipitated from extracts treated with either G6P or buffer using CaM-Sepharose. These precipitates were then analyzed by mass spectrometry (MS). The predominant isoform identified was CaMKII␥ L subunit. MS analysis identified several phosphorylation sites on CaMKII that were responsive to G6P addition. As expected, we observed an increase in phosphorylated Thr 287 (Thr 286 on CaMKII␣) phosphorylation (Fig. 2C). More importantly, as shown in Fig. 2 (D and E), we identified two sites on CaMKII, Thr 393 and Ser 395 , whose phosphorylation was decreased in response to G6P treatment. These data suggested that regulated dephosphorylation of these two sites could influence G6P-induced CaMKII activation.
Based on these observations, we also performed the gel filtration experiment described in Fig. 2A with or without the phosphatase inhibitor, okadaic acid. Although G6P treatment alone induced an upshift in the apparent molecular weight of CaMKII, as expected, co-treatment with the phosphatase inhibitor okadaic acid largely abrogated this up-shift, suggesting that some okadaic acid-inhibitable phosphatase(s) might be required for the observed G6P-induced CaMKII molecular weight upshift (Fig.  2B). Combined with the MS analysis result, these data suggested the possible involvement of some phosphatase(s) in G6P-induced CaMKII activation.
G6P Treatment Increases Binding of PP2A to CaMKII-To determine whether a phosphatase is involved in CaMKII activation, CaM-Sepharose was used to pull down endogenous CaMKII from Xenopus egg extracts treated with or without G6P. Precipitates were immunoblotted for PP1, PP2A, and PP5. Although there was constitutive binding between PP1 and CaMKII, consistent with a previous report (26), this association did not appear to be regulated metabolically (Fig. 3A). Only PP2A exhibited increased interaction with CaM-Sepharose following G6P treatment. PP2A did not bind directly to CaM in the extract because prior removal of all CaM-associated proteins, including CaMKII (via CaM-Sepharose), did not affect total PP2A levels but prevented G6P-inducible binding of PP2A to CaM-Sepharose (Fig. 3, B and C). Furthermore, endogenous PP2A could be co-immunoprecipitated with anti-CaMKII antibody at greater levels in the presence of G6P ( Fig. 3D; note that G6P treatment appeared to increase background CaMKII immunoprecipitation in the presence of G6P, but the PP2A interaction was elevated specifically in the presence of G6P). Together, these data strongly suggest that CaMKII interacts with PP2A, and this interaction is up-regulated by G6P-stimulated metabolism.
B55␤ Targets PP2A to Regulate CaMKII Activation-Functional PP2A is a multiprotein complex containing a catalytic (C) subunit, a scaffolding (A) subunit, and a regulatory (B) subunit. The B subunit typically determines substrate specificity. Four different B subunit families have been identified: B (PR55), BЈ (B56), BЉ (PR72), and Bٞ (PR93). The B55 and B56 families have been implicated in cell proliferation/death in several settings (27,28). We identified a B subunit subtype in the PR55 family, B55␤ whose interaction with CaM-Sepharose was up-regulated upon G6P treatment. B56 isoforms were unaffected (Fig. 4A). In agreement, G6P could stimulate the interaction of recombinant GST-tagged B55␤ protein with multiple CaMKII isoforms (Fig.  4B). In addition, recombinant CaMKII␥ (the isoform identified in the MS analysis) was added into egg extracts. Although G6P treatment increased background protein-protein interactions (GST alone could pull down some CaMKII protein), we were still able to see that G6P could significantly stimulate the interaction between B55␤ and the recombinant CaMKII␥ protein (Fig. 4C). Taken together, these data suggest that a complex containing CaMKII, PP2A C subunit, A subunit, and B55␤, forms in the presence of abundant nutrients, potentially contributing to the metabolic activation of CaMKII, and that this increased protein complex formation stimulated by G6P seems to occur with multiple CaMKII isoforms.
To determine whether B55␤ was required for the observed increased Thr 286 phosphorylation of CaMKII following G6P treatment, we immunodepleted B55␤ from egg extracts and monitored G6P-induced alterations in CaMKII phosphorylation. As shown in Fig. 4D, B55␤ was largely depleted from egg DECEMBER 26, 2014 • VOLUME 289 • NUMBER 52 extracts without significant depletion of the PP2A C subunit. The depleted and undepleted extracts were treated with or without G6P, and the Thr 286 autophosphorylation of CaMKII was monitored. Compared with undepleted extracts, the depleted extracts showed significantly diminished Thr 286 autophosphorylation in the presence of G6P (Fig. 4E). In addition, by using the GST-tagged Xenopus caspase-2 prodomain as a substrate, we were able to monitor the ability of B55␤-depleted and undepleted extracts to phosphorylate caspase-2, treated with or without G6P. We found that the undepleted extracts could phosphorylate caspase-2 in the presence of G6P, whereas this phosphorylation was largely diminished by B55␤ depletion (Fig. 4F). Because Thr 286 autophosphorylation and caspase-2 phosphorylation were both indicators of CaMKII activation, both measures indicated diminished CaMKII activation in the B55␤depleted extracts, suggesting that B55␤ is critical for G6P-driven CaMKII activation in the Xenopus egg extract system. In addition, the fact that the immunodepletion of B55␤ diminished multiple Thr 286 autophosphorylation bands suggests that B55␤ can regulate multiple CaMKII isoforms (Fig. 4E). Moreover, when endogenous B55␤ was knocked down by B55␤ siRNA in 293T cells, the glucose-driven stimulation of CaMKII Thr 286 autophosphorylation was suppressed, suggesting that the role of B55␤ in regulating CaMKII is evolutionarily conserved (Fig. 4G).

PP2A Dephosphorylates Ser 395 on CaMKII in the Presence of G6P-Because the MS analysis showed that Thr 393 and Ser 395
of CaMKII were dephosphorylated in the presence of G6P, we postulated that these sites might be targeted by B55␤-PP2A. To assess this, we generated phospho-antibodies recognizing each of these sites on the Xenopus CaMKII␥ isoform. Although we were unable to produce a high quality antibody specific for phosphorylated Thr 393 , we were successful with Ser(P) 395 . As shown in Fig. 5A, WT CaMKII␥ protein (SF9, baculovirus produced) exhibited basal phosphorylation of Ser 395 . However, the S395A mutation greatly diminished antibody recognition, attesting to its phospho-specificity. To confirm that the phosphorylation level of Ser 395 was down-regulated by G6P, we incubated Xenopus egg extract with or without G6P, precipitated endogenous CaMKII with CaM-Sepharose, and immunoblotted with the phospho-antibody. As shown in Fig. 5B, phosphorylation of Ser 395 was down-regulated in the presence of G6P (note that the ␥ isoform is shown in the loading control as the Ser 395 antibody was designed based on a ␥ phosphopeptide sequence). Additionally, when B55␤ was depleted from the Xenopus egg extract, the G6P effect was inhibited, strongly suggesting that Ser 395 was dephosphorylated by the B55␤-PP2A (Fig. 5C). Although the Ser 395 antibody was unable to recognize the CaMKII␣ isoform, because B55␤ immunodepletion could suppress the activation of all CaMKII isoforms, it is quite possible that Ser 344 (the equivalent residue to Ser 395 in the ␣ isoform) is also dephosphorylated by B55␤-PP2A.
The Phosphorylation Status of Thr 393 and Ser 395 Affects CaMKII Activation-To determine whether B55␤-mediated dephosphorylation of Ser 395 is critical for nutrient-driven activation of CaMKII, we first added CaMKII␥ WT, T393A/S395A mutant, and T393D/S395D mutant proteins into Xenopus egg extract, treated with or without G6P, and monitored Thr 287 (Thr 286 for CaMKII␣) autophosphorylation. As shown in Fig.  5D, after only 10 min of incubation with G6P, the T393A/ S395A mutant exhibited higher phosphorylation of Thr 287 , compared with WT CaMKII. Consistent with these data, G6Pinduced stimulation of CaMKII Thr 287 autophosphorylation was significantly dampened by the T393D/S395D mutations (which may not perfectly mimic phosphorylation) (Fig. 5E). In addition, an in vitro kinase assay was performed, incubating CaMKII␥ T393A/S395A or T393D/S395D with calmodulin and the Xenopus caspase-2 prodomain. We found that the CaMKII T393A/S395A was more able than the T393D/S395D to phosphorylate the caspase-2 prodomain, consistent with the idea that the dephosphorylation of Thr 393 and Ser 395 is important for CaMKII activation (Fig. 5F). Finally, CaMKII␥ WT, T393A/S395A, and T393D/S395D were added into egg extracts, treated with or without G6P, and pulled down using GST-tagged B55␤ to monitor the interaction between B55␤ and CaMKII proteins. We found that the T393A/S395A mutant exhibited stronger interactions and that the T393D/ S395D showed weaker interactions with B55␤, compared with the WT CaMKII, suggesting that the dephosphorylation of Thr 393 and Ser 395 might stabilize the binding of B55␤ to CaMKII (Fig. 5, G and H). Collectively, these data suggest that nutrient status is communicated to CaMKII in part through the B55␤/PP2A-mediated dephosphorylation of CaMKII, a novel locus for the control of CaMKII.

DISCUSSION
Although calcium is a central regulator of CaMKII, previous work from our lab has shown that the centrifugal removal of Ca 2ϩ stores from Xenopus egg extracts does not impede G6Pmediated activation of CaMKII, suggesting that increased available Ca 2ϩ does not underlie nutrient-dependent CaMKII activation. Studies of CaMKII regulation have largely focused on Thr 286 autophosphorylation after Ca 2ϩ /CaM binding. Indeed, G6P treatment of egg extract impairs PP1-mediated dephosphorylation of Thr 286 (25). However, this is unlikely to drive FIGURE 3. CaMKII-PP2A interactions are regulated by G6P. A, CaM-Sepharose was dipped into Xenopus egg extracts that had been treated with or without G6P for 0.5 h at room temperature and incubated for 1 h at 4°C. Beads were retrieved by centrifugation and analyzed for the presence of candidate phosphatases by immunoblotting. B, CaM-Sepharose was incubated with Xenopus egg extracts for 1 h at 4°C and then removed by centrifugation. This process was repeated three times. Depleted and undepleted extracts were analyzed by CaMKII or PP2A immunoblotting. C, CaM-Sepharose was incubated with either CaMKII-depleted or undepleted extracts described in B treated with or without G6P. Beads were retrieved by centrifugation and analyzed by PP2A C subunit immunoblotting. D, CaMKII antibody or control IgG coupled to protein G beads was dipped into Xenopus egg extracts treated with or without G6P, incubated for 1 h at 4°C, and retrieved for CaMKII or PP2A C immunoblotting.
CaMKII activation because this phosphorylation is a result of activation. It has also been reported that CoA generation is increased downstream of G6P addition to egg extracts (though the mechanism underlying this increase is not clear) and that CoA can directly bind to and activate CaMKII. We have now identified an additional metabolically regulated break to CaMKII activation that must be lifted for G6P to robustly activate CaMKII via dephosphorylation of Thr 393 /Ser 395 .
In addition to Thr 393 /Ser 395 phosphorylation, MS analysis also identified several previously uncharacterized sites with increased phosphorylation in the presence of G6P: Ser 311 , Ser 326 , Ser 333 , and Thr 421 (data not shown). Although we have not yet investigated these sites, they may contribute to full CaMKII activation. The T393D/S395D mutant of CaMKII is less potently activated by G6P than WT CaMKII, consistent with a requirement for phosphorylation of additional sites for full activation. Furthermore, the T393A/S395A mutant is not spontaneously active, so there must be additional metabolic input for full activation, likely CoA (the high levels of exogenous CoA used in the published experiments may have forced activation of CaMKII despite the brake (23)).
Effects of G6P-stimulated Dephosphorylation on CaMKII-Interestingly, most of the sites whose phosphorylation was altered by nutrient status are located in the association domain of CaMKII, responsible for self-association (29,30). However, our gel filtration results suggest that oligomerization itself is not affected by phosphorylation. Rather, G6P appears to shift the molecular weight of the full holoenzyme, suggesting association of additional factors. Importantly, the molecular weight shift was largely abrogated by treatment with okadaic acid, consistent with the idea that dephosphorylation of Thr 393 /Ser 395 might be required. It is attractive to speculate that Thr 393 / Ser 395 dephosphorylation might allow binding of additional regulatory factors to the CaMKII association domain. Interestingly, one variant of CaMKII␥ (CaMKII␥G-2) contains within its association domain a targeting sequence essential for ERK FIGURE 4. B55␤ regulates CaMKII activation. A, CaM-Sepharose was incubated with Xenopus egg extracts treated with or without G6P and incubated for 1 h at 4°C. Beads were retrieved by centrifugation and analyzed by immunoblotting for PP2A regulatory subunits. B, GST-B55␤ or GST bound to glutathione-Sepharose was incubated for 1 h at 4°C with Xenopus egg extracts treated with or without G6P. Beads were retrieved by centrifugation and analyzed for CaMKII or B55␤ immunoblotting. Note that the top and middle panels are from two different films with different exposures as the CaMKII antibody recognized CaMKII␣ much more strongly than the other isoforms. C, Xenopus CaMKII␥ WT protein expressed from baculoviral vectors in SF9 cells was added into Xenopus egg extracts for 0.5 h at room temperature, and then the extracts were treated with or without G6P for 0.5 h at room temperature and incubated with GST-B55␤ or GST bound to glutathione-Sepharose for 1 h at 4°C. The beads were retrieved by centrifugation and analyzed for CaMKII or B55␤ immunoblotting. D, B55␤ antibody or control rabbit IgG bound to Dynabead-linked protein A was incubated with Xenopus egg extracts for 1 h at 4°C and removed with a magnet (repeated three times). Extracts were analyzed by B55␤ and PP2A C subunit immunoblotting. E, the B55␤-depleted or undepleted extract was treated with or without G6P for 0.5 h and analyzed by pT286 immunoblotting. F, GST tagged Xenopus caspase-2 prodomain or GST bound to glutathione-Sepharose was incubated with B55␤-depleted or undepleted Xenopus egg extracts supplemented with [␥-32 P]ATP and treated with or without G6P. The samples were resolved by SDS-PAGE and detected by autoradiography. CB, Coomassie Blue. G, B55␤-specific or control siRNA-treated 293T cells were glucose-starved for 12 h and then incubated with or without 25 mM glucose. The lysates were analyzed by immunoblotting for pT286, CaMKII, and B55␤. activation and contractility of smooth muscle cells, suggesting that the association domain can perform roles other than selfassociation (31).
Regulation of Thr 393 /Ser 395 Phosphorylation-Our data demonstrate a previously unknown role for the B55␤ targeting subunit of PP2A in regulating CaMKII. Often, the B subunit not only targets PP2A to the correct substrate(s) but also serves as a key locus of regulation. For example, Chk1-mediated phosphorylation B56 can inhibit PP2A-mediated dephosphorylation and activation of the Cdc25 mitotic regulator during DNA damage checkpoint signaling (27). Although it is possible that the CamKII-B55␤ association is regulated at the level of CaMKII modification, we speculate that B55␤ may be differentially modified in a nutrient-dependent manner to alter targeting of PP2A to CaMKII. Such a modification could be sensitive to CoA levels.
Thr 393 and Ser 395 are phosphorylated in the egg extract before any treatment. Therefore, Xenopus egg extracts must contain a kinase(s) directed against these sites. Because Thr 393 / Ser 395 phosphorylation decreases prior to CaMKII activation, these residues are clearly not autophosphorylation sites. The relevant kinase is not currently known but may provide an additional locus of metabolic regulation if its activity is high in the presence of abundant nutrients and low upon their depletion. Future experiments will be directed toward the identification and characterization of kinase(s) directed toward these sites.
Metabolic Regulation of CaMKII in Other Signaling Pathways-CaMKII is ubiquitously expressed and has a diverse array of substrates. For example, CaMKII␥ regulates the contractility of smooth muscle cells, and reduced contractile force has been observed to be associated with altered metabolism. Although this decrease in force has been associated with impaired Ca 2ϩ flux, metabolically controlled dephosphorylation of Thr 393 and Ser 395 might also control the activation of CaMKII␥ in smooth muscle cells. It will be interesting to determine whether these sites play a  395 immunoblotting. B, CaM-Sepharose was incubated with Xenopus egg extracts treated with or without G6P, retrieved by centrifugation, and analyzed by Ser(P) 395 immunoblotting. C, B55␤-depleted or undepleted Xenopus egg extracts with the addition of recombinant CaMKII␥ were treated with or without G6P. Recombinant CaMKII␥ was precipitated with CaM-Sepharose. The beads were retrieved by centrifugation and analyzed for Ser(P) 395 . D, CaMKII␥ WT or T393A/S395A mutant was added into Xenopus egg extracts for 0.5 h at room temperature, and then the extracts were treated with or without G6P and analyzed for pT287. E, CaMKII␥ WT or T393D/S395D mutant was treated and processed as in D. F, GST tagged Xenopus caspase-2 prodomain or GST bound to glutathione-Sepharose was incubated with Xenopus CaMKII␥ T393A/S395A or T393D/S395D mutants, calmodulin, and [␥-32 P] ATP at 30°C for 0.5 h. Samples were resolved by SDS-PAGE and detected by autoradiography. CB, Coomassie Blue. G, Xenopus CaMKII␥ WT or T393A/S395A mutant proteins were added into Xenopus egg extracts for 0.5 h at room temperature, and then the extracts were treated with or without G6P for 0.5 h at room temperature and incubated with GST-B55␤ or GST bound to glutathione-Sepharose for 1 h at 4°C. The beads were retrieved by centrifugation and analyzed for CaMKII or B55␤ immunoblotting. H, Xenopus CaMKII␥ WT or T393D/S395D mutant proteins were added into Xenopus egg extracts for 0.5 h at room temperature, and then the extracts were treated with or without G6P for 0.5 h at room temperature and incubated with GST-B55␤ or GST bound to glutathione-Sepharose for 1 h at 4°C. Beads were retrieved by centrifugation and analyzed for CaMKII or B55␤ immunoblotting. similar role in regulating the highly abundant CaMKII␣ isoform in neurons, which is important for long term potentiation. Similarly, B55␤ may control CaMKII in other physiological paradigms of regulation or metabolic control, warranting further investigation in other physiological settings.