Interaction between Salt-inducible Kinase 2 and Protein Phosphatase 2A Regulates the Activity of Calcium/Calmodulin-dependent Protein Kinase I and Protein Phosphatase Methylesterase-1*

Background: SIK2 is the only AMPK family kinase that interacts with PP2A, with a hitherto unknown functional consequences. Results: The interaction between SIK2 and PP2A preserves both enzyme activities and regulates CaMKI and PME-1 negatively. Conclusion: The SIK2·PP2A complex is a critical regulator of calcium/calmodulin-mediated activation of CaMKI and PME-1. Significance: SIK2·PP2A may play pivotal roles in regulating cell proliferation and stress response. Salt-inducible kinase 2 (SIK2) is the only AMP-activated kinase (AMPK) family member known to interact with protein phosphatase 2 (PP2A). However, the functional aspects of this complex are largely unknown. Here we report that the SIK2·PP2A complex preserves both kinase and phosphatase activities. In this capacity, SIK2 attenuates the association of the PP2A repressor, the protein phosphatase methylesterase-1 (PME-1), thus preserving the methylation status of the PP2A catalytic subunit. Furthermore, the SIK2·PP2A holoenzyme complex dephosphorylates and inactivates Ca2+/calmodulin-dependent protein kinase I (CaMKI), an upstream kinase for phosphorylating PME-1/Ser15. The functionally antagonistic SIK2·PP2A and CaMKI and PME-1 networks thus constitute a negative feedback loop that modulates the phosphatase activity of PP2A. Depletion of SIK2 led to disruption of the SIK2·PP2A complex, activation of CaMKI, and downstream effects, including phosphorylation of HDAC5/Ser259, sequestration of HDAC5 in the cytoplasm, and activation of myocyte-specific enhancer factor 2C (MEF2C)-mediated gene expression. These results suggest that the SIK2·PP2A complex functions in the regulation of MEF2C-dependent transcription. Furthermore, this study suggests that the tightly linked regulatory loop comprised of the SIK2·PP2A and CaMKI and PME-1 networks may function in fine-tuning cell proliferation and stress response.

Salt-inducible kinase 2 (SIK2) is the only AMP-activated kinase (AMPK) family member known to interact with protein phosphatase 2 (PP2A). However, the functional aspects of this complex are largely unknown. Here we report that the SIK2⅐PP2A complex preserves both kinase and phosphatase activities. In this capacity, SIK2 attenuates the association of the PP2A repressor, the protein phosphatase methylesterase-1 (PME-1), thus preserving the methylation status of the PP2A catalytic subunit. Furthermore, the SIK2⅐PP2A holoenzyme complex dephosphorylates and inactivates Ca 2؉ /calmodulin-dependent protein kinase I (CaMKI), an upstream kinase for phosphorylating PME-1/Ser 15 . The functionally antagonistic SIK2⅐PP2A and CaMKI and PME-1 networks thus constitute a negative feedback loop that modulates the phosphatase activity of PP2A. Depletion of SIK2 led to disruption of the SIK2⅐PP2A complex, activation of CaMKI, and downstream effects, including phosphorylation of HDAC5/Ser 259 , sequestration of HDAC5 in the cytoplasm, and activation of myocyte-specific enhancer factor 2C (MEF2C)-mediated gene expression. These results suggest that the SIK2⅐PP2A complex functions in the regulation of MEF2C-dependent transcription. Furthermore, this study suggests that the tightly linked regulatory loop comprised of the SIK2⅐PP2A and CaMKI and PME-1 networks may function in fine-tuning cell proliferation and stress response. SIK2 2 is a serine/threonine protein kinase belonging to the AMP-activated protein kinase (AMPK) superfamily. AMPK and its family members are believed to function in sensing energy state and mediating stress response. SIK2 is known to attenuate insulin signaling by phosphorylating IRS-1/Ser 789 (1).
Other studies also showed that the insulin-activated SIK2 serves as a repressor of hepatic gluconeogenesis via phosphorylation and translocation of the CREB co-activator TORC2 from the nucleus to the cytoplasm (2,3). In mice, SIK2 was shown to regulate the melanogenesis by modulating the nuclear translocation of TORC1 (4). In adipocytes, SIK2 down-regulates the expression of lipogenic genes, PGC-1␣ and UCP-1 suggesting an involvement in metabolic regulation of adipose tissue (5). Moreover, SIK2 was shown to down-regulate the carbohydrate-responsive element-binding protein (ChREBP)mediated lipogenesis in hepatocytes through the inhibitory phosphorylation of p300/Ser 89 and to prevent steatosis in mice (6). SIK2 may play important roles in cell proliferation, as demonstrated by growth inhibition and cell death of ovarian cancer cells when SIK2 was down-regulated (7). A decreased level of SIK2 after cerebral ischemia may mediate the neuronal survival pathway via its phosphorylation of CREB co-activator TORC1 (8). Furthermore, our recent results revealed that reversible acetylation of SIK2 at Lys 53 regulates autophagy when the proteasome is inhibited (9). We have also uncovered a novel function of SIK2 in ER-associated protein degradation via its interaction with p97/VCP (10).
Protein phosphatase 2A (PP2A) is a multifunctional serine/ threonine phosphatase essential for cellular homeostasis via regulating various signal transduction pathways and fundamental cellular activities, such as cellular metabolism, cell cycle progression, DNA replication, transcription, translation, and apoptosis (11)(12)(13). Deregulation of PP2A may be responsible for several pathological conditions, such as Alzheimer disease and cancer (14 -16). PP2A holoenzyme is a heterotrimer composed of a heterodimeric core of catalytic C and structural A subunits and a regulatory B subunit. The B subunit is responsible for the substrate specificity and subcellular localization. There are more than 20 different B subunits encoded by the human genome, and they can be grouped into four different families annotated as B/B55/PR55, BЈ/B56/PR61, BЉ/PR72, and Bٞ/PR93/PR110, all of which share the same binding site on the core A subunit (11)(12)(13). Moreover, many of them undergo alternative splicing to generate different variants, further expanding the diversity of PP2A holoenzyme. Mechanisms governing the formation of heterotrimeric holoenzyme are important for maintaining its protein stability. Knockdown of either the A or C subunit accelerates the turnover of the other PP2A subunits in Drosophila S2 cells (17,18). Additionally, mammalian PP2A C and most B subunits are stable only when they complex with the A subunit (19,20). Some posttranslational modifications are known to influence PP2A holoenzyme formation or stability, such as phosphorylation of PP2Ac at Thr 304 and Tyr 307 (21,22). In addition to regulation by phosphorylation, reversible methylation at the C-terminal leucine of the PP2Ac subunit provides another mechanism to regulate PP2A; carboxymethylation of Leu 309 was carried out by S-adenosylmethionine-dependent leucine carboxyl methyltransferase 1 and removed by PME-1 (23)(24)(25). Carboxymethylation of PP2Ac was shown to facilitate binding of the B/PR55 family (26,27). The crystal structure of the PP2A holoenzyme showed a highly conserved C-terminal tail ( 304 TPDYFL 309 ) of the C subunit that lies in the interface between the A subunit and BЈ/PR61␥, further linking its modification to B subunit recruitment (28). PME-1 serves as a negative regulator of PP2A and can access only the A-C core and not the holoenzyme (29). The crystallographic structure of the PME-1⅐PP2Ac complex revealed that PME-1 mediates its inhibition partially by its demethylation activity on the PP2Ac C-terminal tail and by occupying the active site of PP2A (30). A catalytically inactive PP2Ac/H59Q mutant exhibits increased affinity for PME-1 and forms a stable complex in vivo (29). Furthermore, PME-1 gene disruption causes a perinatal lethality in mice (31). In glioma cells, PME-1 was shown to support ERK pathway signaling at a point upstream of Raf but downstream of PKC (32).
SIK2 is the only member of the AMPK family that can interact with PP2A (2); however, the functional impact of SIK2⅐PP2A interaction remains unknown. In this report, we showed that interaction between SIK2 and PP2A is important for preserving PP2A phosphatase activity by excluding the association of PME-1. We also discovered that there exists cross-regulation between CaMKI⅐PME-1 and SIK2⅐PP2A. The activity of CaMKI is inversely correlated to the level of SIK2dependent PP2A activity (i.e. SIK2⅐PP2A complex). When the CaMKI activity is elevated, it phosphorylates PME-1 at Ser 15 . Activated CaMKI negatively regulates SIK2, resulting in its degradation (8). Conversely, phosphorylated CaMKI/Thr 177 and PME-1/Ser 15 are substrates of PP2A. Both SIK2 and activated CaMKI could target HDAC5 for export to the cytoplasm and sequestration by 14-3-3, resulting in MEF2C-mediated transcription. Together, our present results suggest that cross-regulation between the SIK2⅐PP2A complex and CaMKI plays important physiological roles in coping with stress conditions or growth factor signaling cascades.
Cell Culture and Transfection-HEK293T cells were cultured in Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 10% fetal bovine serum (Biological Industries) and 100 units/ml penicillin and streptomycin (Invitrogen) at 37°C in a 5% CO 2 humidified incubator. TurboFect TM (Fermentas) transfection was carried out according to the manufacturer's instructions.
Immunoprecipitation and Western Blotting-Whole cell extracts (WCE) were prepared by lysing the cells in WCE buffer (0.2 M NaCl, 20 mM Tris-HCl, pH 7.5, and 0.1% Triton X-100) supplemented with a mixture of protease and phosphatase inhibitors (0.5 mM NaF, 1 mM leupeptin, 1 mM pepstatin A, 1 mM PMSF, 1 mM sodium butyrate, and 1 mM sodium orthovanadate) on ice for 20 min. WCE was incubated with M2 beads (Sigma) for 1 h at 4°C. After incubation, the immunoprecipitates were washed three times in WCE buffer, added SDS-PAGE sample buffer (30 mM Tris-HCl, pH 6.8, 3% SDS, 0.015% bromphenol blue, 1.5 M urea, and 8% glycerol), and separated by SDS-PAGE. After the electrophoresis, the proteins were transferred to PVDF membrane and blocked for 30 min in blocking solution (5% nonfat milk in PBS-T) at room temperature. The membrane was probed with antibody for 1 h at room temperature, washed three times in PBS-T (PBS containing 0.1% Tween 20). The membrane was incubated with an HRP-labeled secondary antibody and washed three times in PBS-T. The immunoblot was detected with enhanced chemiluminescence (PerkinElmer Life Sciences).
Fractionation of SIK2-containing Complex by Superdex 200 FPLC-FLAG-SIK2-containing whole cell extracts (0.2 ml/1 mg) were subjected to Superdex 200 column (1.4 ϫ 40 cm) FPLC. A fraction of 0.5 ml was collected and precipitated with 5% TCA, rinsed with acetone, dried, and dissolved in SDS sample buffer for Western blot analysis.
Protein Purification and in Vitro Phosphatase and Kinase Assay-FLAG-tagged CaMKI and PME-1 were immunoprecipitated with M2 beads and washed with high salt and detergent buffer (TNT: 20 mM Tris-HCl pH 7.5, 0.5 M NaCl, and 0.5% Triton X-100) supplemented with 50 mM NaF and 20 M okadaic acid, to remove the associated phosphatases and other interacting proteins. Recombinant substrates were eluted by the addition of 2 g of FLAG peptide in 20-l reaction mixtures and incubated for 30 min at room temperature. The amounts and purity of substrate was verified by Coomassie Blue staining of SDS gel.
Phosphatase activity assay was performed following the protocol described below. Wild type PP2A and catalytically inactive (H59Q) mutant were immunoprecipitated with anti-HA beads. Washed beads were resuspended in assay buffer (20 mM Tris-HCl, pH 7.5, 1 mM DTT, 2 mM EGTA, 10 mM MgCl 2 , and 0.1 mg/ml BSA) and incubated with substrate at 30°C for 30 min. The reaction was terminated by the addition of SDS-sample buffer and subjected to immunoblot analysis. An assay for SIK2-associated PP2A activity was performed similarly; FLAG-SIK2-WT and FLAG-SIK2-KD mutant-associated PP2A were immunoprecipitated by using M2 beads instead. SIK2⅐PP2A complex was prepared by immunoprecipitating the HA-tagged PP2A-associated SIK2 with anti-HA beads, and an SIK2 kinase assay was performed as described previously (9).
FLAG-tagged PME-1 was immunoprecipitated with M2 beads and washed with TNT buffer. To elute the substrate, 2 g of FLAG peptide was added to the 30-l immunoprecipitates and incubated for 30 min at room temperature. The purity of the substrate was verified by Coomassie Blue staining of SDS gel. Recombinant FLAG-CaMKI was also purified similarly. CaMKI was preactivated with 10 mM CaCl 2 during lysis of cells. The CaMKI was used for a kinase assay immediately after immunoprecipitation.
For the CaMKI kinase assay, wild type CaMKI and dominant negative (K49E) mutant were assayed at 30°C by incubation in the following reaction mixtures: 50 mM Tris-HCl, pH 7.5, 1 mM CaCl 2 , 0.5 mM DTT, 0.5 mg/ml BSA, 0.2 mM ATP, and 1 g of PME-1 substrate. The reaction was terminated by SDS-sample buffer and subjected to Western blot analysis.
Immunofluorescence Staining-Cells grown on coverslips were fixed in 4% formaldehyde for 20 min and permeabilized with 0.5% Triton X-100 for 5 min. After blocking with 1% BSA, the indicated primary antibodies and fluorescent secondary antibodies (Alexa 488-conjugated goat anti-rabbit IgG, Alexa 594-conjugated goat anti-rabbit IgG, and Alexa 594-conjugated goat anti-mouse IgG; Invitrogen) were incubated for 1 h at room temperature. Nuclei were visualized by Hoechst staining. Slides were then washed and mounted with coverslips. Fluorescent images were acquired by using an inverted confocal microscopy (LSM-510, Zeiss, Thornwood, NY).
Statistical Analysis-Values from three independent experiments were presented as mean Ϯ S.D. To determine the statistical difference between mean values, the two-tailed Student's t test was applied, and a p value of Ͻ0.05 was considered as statistically significant.

RESULTS
Complex Formation between SIK2 and PP2A-SIK2 is the only member of the AMPK family that can interact with PP2A (33). Our previous results have shown that interaction between SIK2 and p97/VCP regulates ER-associated protein degradation (10). Interestingly, when PP2Ac was overexpressed, the p97/VCP level in both wild type SIK2 (WT) and kinase-dead mutant SIK2 (KD) immunoprecipitates was reduced, implying that PP2A and p97/VCP may exist in two distinct SIK2-containing complexes (Fig. 1A). To further confirm this possibility, the ectopically expressed SIK2 from HEK293T lysates was subjected to Superdex-200 FPLC fractionation. Differential fractionation of p97/VCP and PP2A (with elution peaks at fractions 25 and 31, respectively; Fig. 1B) suggests that PP2A and p97/ VCP form distinct complexes with SIK2. Additionally, an immunoprecipitation assay showed that the interaction between SIK2 and PP2A did not require the kinase activity of SIK2 because both SIK2-WT and SIK2-KD could immunoprecipitate comparable levels of PP2Ac (Fig. 1A).
To determine which subunit of the heterotrimeric PP2A is responsible for interaction with SIK2, we tested whether catalytically inactive PP2Ac/H59Q mutant could interact with SIK2. H59Q mutant reportedly associates with adaptor A subunit but is unable to recruit B subunit (29). Such mutation also led to the loss of SIK2⅐PP2Ac interaction (Fig. 1C). To further test whether SIK2 interacts with PP2A in a phosphatase activity-dependent manner, extracts were prepared from HEK293T cells transfected with epitope-tagged SIK2 and PP2Ac in the presence of okadaic acid and immunoprecipitated with epitope-specific antibodies. The results showed that okadaic acid treatment disrupted the complex formation between SIK2 and PP2A (Fig. 1D). Next, to confirm the interaction between SIK2 and PP2A holoenzyme, we overexpressed SIK2 and the regulatory subunit B/PR55 in HEK293T cells and performed immunoprecipitation. We subsequently observed that SIK2 and B/PR55-containing PP2A holoenzyme co-existed in the same immunocomplex but lost such association in the presence of okadaic acid (Fig. 1E). On the contrary, SIK2 and p97/VCP complex formation was not affected by okadaic acid (Fig. 1, D and E). Together, these results indicate that SIK2 interacts with the catalytically active PP2A holoenzyme.
The Phosphatase Activity Is Preserved in SIK2⅐PP2A Complex by Excluding PME-1 from Associating with PP2A-To address the functional consequence of the SIK2⅐PP2A interaction, we performed further biochemical experiments for the SIK2⅐PP2A complex. When the level of SIK2 was reduced by shRNA-mediated knockdown, the levels of total and carboxymethylated PP2Ac also dropped ( Fig. 2A). However, in cells with overex-pressed SIK2-WT or SIK2-KD, the levels of PP2Ac and carboxymethylated PP2Ac remained unchanged (Fig. 2B). Because carboxymethylation of PP2Ac/Leu 309 facilitates PP2A holoenzyme formation and is a hallmark of active PP2A (26), these results suggest that PP2A expression and activity may be preserved in the SIK2⅐PP2A complex. Further comparison of the PP2Ac/L309Me levels from total and SIK2-immunoprecipitated PP2Ac revealed that SIK2-associated PP2Ac was enriched in the methylated form (Fig. 2C). These results thus suggest that the SIK2⅐PP2A complex could protect PP2A from the endogenous inhibitor, PME-1. To clarify the relationship between SIK2⅐PP2A complex and PME-1, the SIK2⅐PP2A Regulates CaMKI and PME-1 JULY 25, 2014 • VOLUME 289 • NUMBER 30 endogenous inhibitor of PP2A, we performed immunoprecipitation experiments on HEK293T cell extracts co-expressing FLAG-SIK2 and HA-PP2Ac. Our data showed that PME-1 was excluded from the SIK2⅐PP2A complex, whereas PP2A and PME-1 formed a distinct complex (Fig. 2D). These results thus suggest that the SIK2⅐PP2A could protect PP2A activity from the inhibition of PME-1.
Furthermore, we observed an up-regulation of activated phosphorylation of SIK2/Thr 175 level as well as down-regulation of repressive phosphorylation of SIK2/Ser 587 in the complex (Fig. 2E), which may indicate that SIK2 activity is also preserved in association with PP2A. To further confirm the kinase activity in the SIK2⅐PP2A complex, we performed co-transfection of HEK293T cells with SIK2-WT or SIK2-KD with HA-PP2Ac. SIK2⅐PP2A complex was prepared by immunoprecipitation of HA-PP2Ac and then assayed for kinase activity with GST-Syntide substrate. The results showed that the SIK2⅐PP2A complex is indeed active in phosphorylating this substrate (Fig. 2F). Taken together, these results imply that the PP2A holoenzyme-dependent SIK2⅐PP2A complex formation has the distinct function of preserving both SIK2 and PP2A activities.
Negative Regulation of CaMKI by PP2A-As a known substrate of PP2A, CaMKIV can be inactivated and dephosphorylated in vitro by PP2A (35). Characterization of phosphospecific antibody was shown in Fig. 3A. To address whether the functionally related CaMKI is negatively regulated, we next investigated whether PP2A is also involved in its inactivation. To this end, we found that when PP2A activity was inhibited by okadaic acid or under overexpression of PP2Ac/H59Q, the phosphorylated CaMKI/Thr 177 level increased (Fig. 3, B and C). Similarly, when PP2Ac was down-regulated by shRNA, the phosphorylated CaMKI/Thr 177 level was elevated (Fig. 3D). In vitro phosphatase assay using immunoprecipitated HA-tagged PP2A complex further demonstrated that CaMKI/Thr 177 is specifically dephosphorylated by PP2A (Fig. 3E). These findings there- fore indicate that PP2A is responsible for the dephosphorylation of CaMKI/Thr 177 .
Cross-regulation between SIK2⅐PP2A Complex and CaMKI-CaMKI is known to negatively regulate SIK2 through Thr 484 phosphorylation, which is linked to its degradation (8). Overexpression of constitutively active CaMKI indeed led to a reduction in the SIK2 protein level in HEK293T cells (Fig. 4A). To address whether there exists a cross-regulation between CaMKI and SIK2, we performed SIK2 knockdown along with CaMKI overexpression in HEK293T cells. Immunoblotting analysis showed that SIK2 knockdown enhanced the phosphorylation level of CaMKI/Thr 177 (Fig. 4B). However, there was no difference in this phosphorylation level upon SIK2-WT or SIK2-KD overexpression (Fig. 4C), suggesting that depletion of SIK2 resulted in CaMKI activation. To further elucidate how SIK2 may affect the activity of CaMKI, we focused on the functional characterization of SIK2 and PP2A complex. In vitro phosphatase activity of immunoprecipitated SIK2⅐PP2A complex demonstrated that the SIK2⅐PP2A complex serves as a negative regulator for CaMKI by dephosphorylation of CaMKI/ Thr 177 (Fig. 4D). Taken together, these data suggest that depletion of SIK2, with the consequence of SIK2⅐PP2A complex disruption, contributes to CaMKI activation.
CaMKI Is the Upstream Kinase for Phosphorylation of PME-1/Ser 15 -We have established that down-regulation of SIK2 resulted in decrease of PP2Ac and PP2Ac/L309Me levels ( Fig. 2A). We further noticed that the amino acid sequence surrounding the PME-1/Ser 15 bears similarity to the consensus target of AMPK, CaMKI/IV, SIK1, or SIK2 ( 10 LGRLPpSR-PPL 19 ) (Fig. 5A). However, the increased phosphorylated PME-1/Ser 15 level in response to SIK2 knockdown excluded the possibility that it is a target of SIK2 (Fig. 5, B and C). Besides, the phosphorylated PME-1/Ser 15 level was also elevated under okadaic acid treatment (Fig. 5D), consistent with the results of both the PP2Ac knockdown (Fig. 5E) and PP2Ac/H59Q overexpression experiments (Fig. 5F). To further assess whether the phosphatase activity in SIK2⅐PP2A complex is responsible for the dephosphorylation of PME-1/Ser 15 , we performed in vitro phosphatase assay using immunoprecipitated SIK2⅐PP2A and purified recombinant PME1. The results subsequently revealed that phosphorylated PME-1/Ser 15 is dephosphorylated by SIK2⅐PP2A (Fig. 5G). Together with Fig. 4D, these results indi- cate that the elevated phosphorylation levels of both CaMKI/ Thr 177 and PME-1/Ser 15 are inversely correlated with the PP2A activity associated with SIK2⅐PP2A complex. Next, to substantiate whether PME-1/Ser 15 is reciprocally targeted by CaMKImediated phosphorylation, we performed an in vitro kinase assay using purified recombinant CaMKI and PME-1. Our results indeed showed that PME-1 could be phosphorylated in vitro by active CaMKI (Fig. 6A). In addition, this phosphorylation could be suppressed by the addition of the CaMK inhibitor, KN-93 (Fig. 6B), but elevated by overexpression of the constitutively active CaMKI (Fig. 6C). Collectively, these data are in line with the scenario that PME-1 serves as downstream regulator of SIK2⅐PP2A-CaMKI signaling, and CaMKI is the upstream kinase responsible for PME-1/Ser 15 phosphorylation.

HDAC5-mediated Transcriptional Inhibition of MEF2C Is Regulated by SIK2⅐PP2A-Previous studies demonstrated that
CaMK family kinases contribute to MEF-dependent transcriptional regulation (36). In this capacity, class IIa HDACs (e.g. HDAC4 and -5) are phosphorylated by CaMK and sequestered to the cytoplasm by binding to 14-3-3 (37). To investigate whether the SIK2⅐PP2A-CaMKI signal cascade participates in modulation of MEF-dependent transcription, we probed the phosphorylation of HDAC5/Ser 259 in the context of SIK2 knockdown. Our results showed that depletion of SIK2 led to increased phosphorylation level of HDAC5/Ser 259 (Fig. 7A). Immunofluorescence staining also revealed a greater extent of HDAC5 cytoplasmic localization in the knockdown cells (Fig.  7B). Consistent with these changes, a co-immunoprecipitation experiment showed increased association of 14-3-3 with HDAC5 when SIK2 was depleted (Fig. 7C). To address the functional consequence of such regulation in a transcriptional context, a 3ϫMEF binding site-containing promoter reporter assay was performed. The reporter activity was induced by depletion of SIK2 (Fig. 7D). Previous studies also showed that HDAC5/ Ser 259 could be phosphorylated by AMPK and SIK1 (38 -40), we examined whether SIK2 may also target HDAC5/Ser 259 . Under SIK2 overexpression, the MEF2C-mediated transcription was activated comparably with that of SIK2 depletion (Fig.  7E). These results therefore suggest that SIK2 can phosphorylate HDAC5/Ser 259 and trigger its sequestration in the cytoplasm, a function similar to depletion of SIK2-activated CaMKI.
Because the kinase activity is preserved in the SIK2⅐PP2A complex (Fig. 2E), we addressed whether the SIK2 in that complex could phosphorylate HDAC5/Ser 259 . Overexpression of SIK2-WT seems to have a dominant effect over the presence of recombinant PP2A-WT or PP2A/H59Q mutant (Fig. 7F) on the phosphorylation of HDAC5/Ser 259 . These results suggest that SIK2 (as shown by the absence of complex formation by SIK2⅐PP2A/H59Q) per se or in the context of the SIK2⅐PP2A-WT complex is responsible for HDAC5/Ser 259 phosphorylation, and activation of MEF2C-dependent transcription (Fig. 8).

DISCUSSION
This study reported the interaction between SIK2 and PP2A that preserves the kinase and phosphatase activities of SIK2 and PP2A, respectively. These findings extended our previous results that demonstrated a link of SIK2 to p97/VCP and the process of ER-associated protein degradation (2) and thus broadened the functional implications of SIK2 in cells. Our present results further showed that the PP2A activity is preserved in the SIK2⅐PP2A complex and serves as a negative regulator for the activation of the CaMKI and PME-1 network. Interestingly, knockdown of SIK2 not only caused disruption of SIK2⅐PP2A complex but also resulted in elevation of the Bcl-2 level in the ER membrane (data not shown). This may represent an alternative mechanism through which SIK2 regulates CaMKI-associated signaling; when the Bcl-2 level is elevated in the ER, Ca 2ϩ may be released and consequently activates CaMKI (i.e. elevation of CaMKI/Thr 177 phosphorylation level) (41). Furthermore, CaMKI-mediated phosphorylation of SIK2/ Thr 484 is known to down-regulate SIK2 through protein degradation (8). Indeed, the SIK2 level was reduced by overexpressing constitutively active CaMKI (Fig. 4A). These findings are thus in line with the negative feedback regulatory mechanism between these two kinases. In this scenario, in addition to sta-bilizing and activating PP2A, formation of the SIK2⅐PP2A complex and the consequently elevated phosphatase activity may also block CaMKI-mediated phosphorylation and degradation of SIK2. Together, these results highlight a tightly linked regulatory loop composed of SIK2⅐PP2A, CaMKI, and PME-1 that may function antagonistically in fine-tuning cell proliferation and stress response.
The N-terminal sequence of PME-1 is an unstructured, flexible region that may have hitherto unknown regulatory functions. Our present results showed that PME-1/Ser 15 is phosphorylated by CaMKI and dephosphorylated by PP2A, suggesting that the phosphorylation/dephosphorylation of PME-1/Ser 15 may have important regulatory functions. Our FIGURE 5. CaMKI is the upstream kinase for PME-1/Ser 15 , whose activity can be modulated by PP2A. A, HEK293T cells were transfected with either FLAG-PME-1-WT or FLAG-PME-1/S15A mutant, and cell extracts were analyzed for phosphorylated PME-1/Ser 15 level by Western blot. HEK293T cells were co-transfected with FLAG-PME-1 and pSuper-shLuc/shSIK2 (B) or FLAG-SIK2-WT/KD (C) for 2 days. Cell lysates were subjected to immunoblotting analysis for phosphorylated level of PME-1/Ser 15 . D, 48 h after transfected with FLAG-PME-1, HEK293T cells were treated with or without 100 nM OA for 2 h. E, HEK293T cells were co-transfected with pSuper-shLuc or pSuper-shPP2Ac and FLAG-PME-1 for 48 h. F, HEK293T cells were co-transfected with HA-PP2Ac-WT or HA-PP2Ac/ H59Q and FLAG-PME-1 for 48 h. Cell lysates were subjected to Western blot analysis with the indicated antibodies. G, for phosphatase preparation, the SIK2⅐PP2A complex was purified from HEK293T cells transfected with control vector, FLAG-SIK2-WT, or FLAG-SIK2-KD. For substrate preparation, FLAG-PME-1-expressing HEK293T cells were pretreated with 100 nM OA for 2 h to enrich the phosphorylated PME-1/Ser 15 level, and the purity of FLAG-PME-1 was confirmed by Coomassie Blue (CB) staining (right). The phosphatase assay mixtures containing the PME-1 substrate in 30 l of assay buffer were incubated at 30°C for 30 min. The reaction was terminated by SDS sample buffer and subjected to Western blot analysis (left). Quantitative data of B-F are presented as means Ϯ S.D. of three independent experiments (*, p Ͻ 0.05; **, p Ͻ 0.005; ***, p Ͻ 0.001; N.S., not significant). SIK2⅐PP2A Regulates CaMKI and PME-1 JULY 25, 2014 • VOLUME 289 • NUMBER 30 results also demonstrated that the phosphorylated levels of PME-1/Ser 15 and CaMKI/Thr 177 are inversely correlated with the phosphatase activity of SIK2⅐PP2A complex, further implying that the demethylase activity of phosphorylated PME-1/ Ser 15 may be higher than that of its unphosphorylated state. Furthermore, our findings on the SIK2⅐PP2A complex and their relationship with PME-1 may provide important insights into several human diseases and pathophysiological conditions. Given that PP2A is a major Tau phosphatase, its activity may also contribute to the development of Alzheimer disease. In addition, PP2A is also known to function as a tumor suppressor (42,43), whereas PME-1 is a repressor of PP2A that enhances oncogenic signaling (32). Aza-␤-lactam, an inhibitor of PME-1, is a known powerful drug for inhibiting cancer cell growth (44). When the catalytic center is bound by aza-␤-lactam, PME-1 can no longer inhibit PP2A, whose activity may thus be preserved for attenuating the oncogenic signaling pathways. Therefore, based on the observation that SIK2-depleted cancer cells are susceptible to apoptosis (7), we propose that inhibitor of PME-1 and down-regulation of SIK2, either pharmacologically or genetically, may be used simultaneously for inhibiting cancer cell growth. Our results thus have important implications in cancer research and treatment. Moreover, in addition to PME-1, a number of endogenous and viral inhibitors of PP2A have been studied (16,45). Whether the SIK2⅐PP2A complex could protect the inhibition of PP2A by one or more of these inhibitors remains unclear.
Evidence supporting interactions between SIK2 and PP2A includes the following: 1) the lack of interaction between the catalytically inactive H59Q mutant of the C subunit, which is defective in holoenzyme formation (29), and SIK2, and 2) the association of SIK2 with recombinant PR55-containing PP2Aa and PP2Ac. These data clearly suggest that the PP2A holoenzyme structure is important for SIK2⅐PP2A complex formation. However, the actual physical interaction between SIK2 and PP2A remains to be elucidated by structural studies. The outcome of these future investigations may shed light on how to manipulate PP2A activity in SIK2⅐PP2A complex. Intriguingly, regulatory parallelism between SIK2⅐PP2A and immunoglobulinbinding protein 1 (IGBP1/␣4)⅐PP2A complexes was noticed. The IGBP1⅐PP2A complex preserves PP2A activity and functions in microtubule dynamics and mTORC1 signaling (46,47). The Opitz syndrome protein midline 1 (MID1) serves as a microtubule targeting subunit as well as a negative regulator of IGBP1⅐PP2Ac. MID1 has been shown to function as an E3 ligase targeting the PP2Ac for ubiquitin-mediated degradation (48, 49). Depletion of MID1 or expression of Opitz syndrome-derived mutated MID1 was shown to compromise association with microtubules and/or transport along microtubules. In an analogous manner, CaMKI is a negative regulator of phosphorylating SIK2/Thr 484 and targeting it for proteasomal degradation. Thus, activation of CaMKI by abnormal elevation of calcium may trigger the disruption of the SIK2⅐PP2A complex, exposing PP2A to its endogenous inhibitor, such as PME-1. The effects of decreased SIK2⅐PP2A complex may thus be linked to the dysregulation of microtubule vesicle transport and have great impacts on the functions of SIK2 in autophagy and ER-as-sociated protein degradation, as exemplified by our previous publications (9,10). The physiological functions of SIK2⅐PP2A complex are important questions to be addressed in future research.
SIK1, a closely related kinase of SIK2, exhibits a seemingly distinct role in the regulation of the CaMKI and PME-1 pathway. In response to elevated intracellular calcium, CaMKI is activated to phosphorylate SIK1 on Thr 322 , which in turn is responsible for the phosphorylation and activation of PME-1. Consequently, PME-1 dissociates from PP2A and exerts activation of the Na ϩ /K ϩ -ATPase (34). Contrary to this regulatory FIGURE 7. Depletion of SIK2 results in activation of CaMKI and phosphorylation of HDAC5 and up-regulation of MEF2-dependent transcription. A, HEK293T cells co-transfected with pSuper-shSIK2 or -shLuc and FLAG-HDAC5 were lysed and subjected to Western blot analysis with the indicated antibodies. B, HEK293T cells ectopically expressing FLAG-HDAC5 along with pSuper-shSIK2 or -shLuc were fixed and immunostained with anti-FLAG antibody to detect the subcellular distribution of HDAC5. GFP, shRNA knockdown cells. Scale bar, 10 m. C, HEK293T cells were co-transfected with pSuper-shSIK2 or -shLuc and FLAG-HDAC5 as well as HA-14-3-3⑀ for 48 h. Cell lysates were immunoprecipitated (IP) by anti-FLAG beads and then subjected to immunoblotting analysis with the indicated antibodies. D, 24 h after transduction with lentiviral shSIK2 or shScramble, HEK293T cells were co-transfected with 3ϫMEF2-Luc reporter (encoding firefly luciferase) and the internal control, pTK-Renilla-Luc, along with control vector, FLAG-MEF2C, or FLAG-HDAC5 for 24 h. Cells were then harvested and assayed following the manufacturer's instructions for the Dual-Luciferase reporter assay system (Sigma). All values were calculated as firefly luciferase activity normalized with Renilla luciferase activity. E, HEK293T cells were co-transfected with control vector, FLAG-HDAC5, and FLAG-SIK2-WT, -KD, or -S587A mutant, along with 3ϫMEF2-Luc reporter (encoding firefly luciferase) and pTK-Renilla-Luc for 48 h. Cell lysates were assayed with the Dual-Luciferase reporter assay system (Sigma). All values were calculated as firefly luciferase activity normalized with Renilla luciferase activity. The bar graph shows mean Ϯ S.D. (error bars) of three experimental replicates (*, p Ͻ 0.05; ***, p Ͻ 0.001). F, HEK293T cells were co-transfected with either control vector, FLAG-SIK2-WT, or FLAG-SIK2-KD and HA-PP2Ac-WT or HA-PP2Ac/H59Q mutant for 48 h. Then cell lysates were subjected to Western blot analysis for phosphorylated HDAC5/ Ser 259 . The ratio of quantitative results of A and C is presented as means Ϯ S.D. (error bars) of three independent experiments (*, p Ͻ 0.05). Quantitative data of D are means Ϯ S.D. of three independent experiments (*, p Ͻ 0.05; ***, p Ͻ 0.001).

SIK2⅐PP2A Regulates CaMKI and PME-1
mechanism, our results showed that CaMKI directs phosphorylation of PME-1/Ser 15 and that the SIK2 level is inversely correlated with the phosphorylation states of CaMKI/Thr 177 and PME-1/Ser 15 . Together, these findings emphasize distinct roles of these related kinases that closely depend on cellular contexts.