A brain-specific SGK1 splice isoform regulates expression of ASIC1 in

Maria F. Arteaga, Tatjana Coric, Christoph Straub, and Cecilia M. Canessa*

Department of Cellular and Molecular Physiology, Yale University, 333 Cedar Street, New Haven, CT 06520

Communicated by Steven C. Hebert, Yale University School of Medicine, New Haven, CT, January 30, 2008 (received for review October 23, 2007) Neurodegenerative diseases and noxious stimuli to the brain two additional full-length human and mouse SGK1 cDNAs. These enhance transcription of serum- and -induced cDNAs differ from the canonical one at the 5Ј end. The spans kinase-1 (SGK1). Here, we report that the SGK1 gene encodes a Ϸ118 kb, with the additional exons located far upstream of the brain-specific additional isoform, SGK1.1, which exhibits distinct initially identified 5Ј end. These exons are differentially spliced to regulation, properties, and functional effects. SGK1.1 decreases give rise to three different transcripts designated here as SGK1, expression of the acid-sensing -1 (ASIC1); thereby, SGK1.1, and SGK1.2. Fig. 1A shows a schematic representation of SGK1.1 may limit neuronal injury associated to activation of ASIC1 the exon–intron organization of the mouse SGK1 gene. The N in . Given that neurons express at least two splice iso- termini of the three splice isoforms are encoded by different exons, forms, SGK1 and SGK1.1, driven by distinct promoters, any changes whereas they share an identical catalytic domain and C-terminal in SGK1 transcript level must be examined to define the isoform hydrophobic motif (exons 6–16). induced by each stimulus or neurological disorder. The region upstream of exon 1 contains signature sequences consistent with a TATA box Ϸ1 kb and 0.7 kb from the initiation alternative promoter ͉ Proton-activated channel ͉ site. A different promoter in intron 4 controls transcription of the serum- and glucocorticoid-induced kinase canonical SGK1 isoform. This is the only promoter that has been characterized experimentally (21); it contains a TATA box near the GK1 is a S/T kinase expressed in many mammalian start site of transcription and a glucocorticoid-responsive element Stissues. It was originally identified as a glucocorticoid (1) and (GRE), consistent with the observation that in- cell-volume-responsive gene (2). The most extensively studied crease mRNA abundance of the canonical isoform in most tissues. target of SGK1 is the epithelial sodium channel, ENaC, which plays No GRE was identified in the 10 kb upstream of exons 1 or 4. a crucial role in the regulation of body sodium (3). The phenotype Tissue distribution and relative abundance of SGK1.1 isoform of SGK1-null mice is a partial deficit in renal sodium reabsorption were examined by quantitative (q)RT-PCR of mouse tissues. whereby the mice are prone to volume depletion when exposed SGK1.1 mRNA was detected exclusively in brain; in all other tissues to a low-salt diet (4). There is substantial evidence that SGK1 examined expression was negligible (Fig. 1C Left). Similar experi- also works in the signaling pathways increasing cell survival and ments conducted with SGK1.2 primers showed very low expression in all tissues; therefore, we did not pursue further studies with this in vertebrates (5) and in the nematode Caenorhabditis isoform. In brain, expression of SGK1.1 transcript was Ϸ1/10 of the elegans (6). level of SGK1 (Fig. 1C Right). The functional role of SGK1 in the mammalian nervous system Transcriptional regulation of the SGK1.1 isoform was examined has been explored by numerous studies that have reported increases by qRT-PCR in differentiated neuronal mouse B1E-115 cells in SGK1 transcript induced by diverse stimuli and conditions. High exposed to dexamethasone or to depolarization (by increasing the levels of SGK1 mRNA have been observed in ischemia (7), injury ϩ concentration of K ), a method that simulates neuronal activity. (8), in various animal models of Parkinson’s disease (9, 10), Depolarization increased SGK1.1 transcript, whereas dexametha- amyotrophic lateral sclerosis (11), (12, 13), Hun- sone did not induce a significant change in mRNA levels of any tington’s disease (14), and in the dorsal horn of the spinal cord after isoform (Fig. 1D). induction of in the corresponding innervated periph- Distribution of SGK1.1 in brain structures was analyzed by in situ eral tissues (15). hybridization using a probe specific for the SGK1.1 splice isoform. Recent studies have started to probe the functional effects of Low magnification of a brain section shows staining of all regions elevated SGK1 expression in neurons of the central nervous system. of hippocampus, dentate gyrus, and cerebral cortex layers. A The current evidence points to a role of SGK1 in activity-dependent cerebellum section shows staining of Purkinje cells and granular facilitation of learning and memory formation (16, 17), consolida- layer (Fig. 1E). Comparison of the SGK1.1 isoform with the tion of long-term memory (18), facilitation of expression of long- distribution of SGK1 by in situ hybridization published online by the NEUROSCIENCE term potentiation in hippocampal neurons (19), and modulation of Allen Institute for Brain Science (www.brain-map.org) indicates synaptic plasticity in the dorsal horn of the spinal cord (15). overlap of expression of SGK1.1 and SGK1 in most areas of the The capacity of SGK1 to modulate expression of ion channels mouse central nervous system. and transporters at the plasma membrane of many cell types (20) also provides a means to alter membrane excitability in neurons. SGK1.1 Is the Most Abundant Protein Isoform in Brain. The relative This prompted us to further investigate SGK1 in the nervous abundance of SGK1 and SGK1.1 was determined by system. Here, we report a splice isoform, SGK1.1, exclusively expressed in the nervous system. We describe the distinct features, transcriptional regulation, and functional effects of SGK1.1 in Author contributions: C.M.C. designed research; M.F.A., T.C., C.S., and C.M.C. performed neurons. research; C.M.C. contributed new reagents/analytic tools; M.F.A. and C.M.C. analyzed data; and M.F.A. and C.M.C. wrote the paper. Results The authors declare no conflict of interest. SGK1 Splice Isoforms and Distribution in Mouse Brain. Hitherto, it was Freely available online through the PNAS open access option. thought that the SGK1 gene spanned Ϸ6 kb, a segment of genomic *To whom correspondence should be addressed. E-mail: [email protected]. DNA that contains the promoter and all of the exons of the This article contains supporting information online at www.pnas.org/cgi/content/full/ reference sequence transcript of SGK1. However, the GenBank 0800958105/DC1. database (National Center for Biotechnology Information) reports © 2008 by The National Academy of Sciences of the USA

www.pnas.org͞cgi͞doi͞10.1073͞pnas.0800958105 PNAS ͉ March 18, 2008 ͉ vol. 105 ͉ no. 11 ͉ 4459–4464 Downloaded by guest on September 27, 2021 Fig. 1. Schematic of the SGK1 gene, differential expression, and distribution of isoforms. (A) The SGK1 gene expands Ϸ100 kb. White and blank boxes indicate 5Ј and 3Ј untranslated regions and exons; line represents introns. Arrows and arrowheads indicate positions of qPCR primers and the in situ hybridization probe specific for the SGK1.1 isoform. The gene contains at least three isoforms with corresponding promoters located upstream of exon 1, intron 3, and intron 4 for transcription of SGK1.1, SGK1.2, and SGK1 (canonical isoform), respectively. The SGK1 promoter contains a GRE. (B) Amino acid sequence of the N termini of SGK1 isoforms; intron–exon boundaries are indicated with arrows. Polybasic motif (ϩ) with large hydrophobic residues (⌬) in the N terminus of SGK1.1 is shown above the protein. (C) qRT-PCR of mouse tissues normalized to the value of SGK1.1 in brain Ϯ SD. Comparison of SGK1 and SGK1.1 in mouse brain. (D) Expression of SGK1 and SGK1.1 transcripts examined by qRT-PCR of N1E-115 cells treated with dexamethasone or increasing external Kϩ concentration to 50 mM. Each bar is the mean of six experiments normalized to GAPDH Ϯ SD. (E) In situ hybridization of mouse brain and cerebellum with SGK1.1-specific antisense and sense probes. e1 antisense and e2 sense probe on brain and e3 antisense and e4 sense probe on cerebellum are shown.

quantitative Western blot analysis of mouse tissues. In previous homogenates of brain, heart, and lung, we identified SGK1 (49/45 work, we demonstrated that the abundance of SGK1 protein is kDa) and SGK1.1 (60 kDa) only in brain (Fig. 2A). lower than expected from the level of its own transcript. This The relative high protein abundance of SGK1.1 in brain when disparity is due to rapid degradation by the ubiquitin/proteasomal compared with its low mRNA level suggests that it might be more system (22). We used a transgenic mouse strain with insertion of a stable than the canonical SGK1. We confirmed by pulse–chase Ͼ bacterial artificial (BAC) containing the whole-mouse experiments that the half-life of SGK1.1 is longer (t1/2 180 min) SGK1 gene (Ϸ200 kb) modified by the addition of three HA than the one previously determined for SGK1 (t1/2 of 28 min) (Fig. epitopes at the C terminus of the coding region (23). In tissue 2B). The high stability of SGK1.1 protein is due to the absence of the proteasomal degradation signal in the N terminus of SGK1.

SGK1.1 Resides at the Plasma Membrane by Binding to PtdIns(4,5)P2. Cellular localization of SGK1.1 was examined by immunofluores- cence of CHO cells cotransfected with SGK1.1-V5 and PH-GFP, a fusion of GFP and the pleckstrin-homology domain of phospho- lipase C␦, which selectively binds PtdIns(4,5)P2.Fig.3A shows colocalization of the two proteins at the plasma membrane. Hy- drolysis of PtdIns(4,5)P2 by activation of phospholipase C produced rapid and almost complete translocation of both PH-GFP and SGK1.1 to the cytosol (Fig. 3B). These results suggest that SGK1.1 binds PtdIns(4,5)P2. To define this interaction more specifically, we cotransfected SGK1.1-RFP (fusion of DsRed-monomeric fluores- cent protein to the carboxyl terminus of SGK1.1), Lyn11-FRB (a membrane anchored domain of mTOR that binds rapamycin, FRB), and CF-Inp (a fusion of cyan fluorescent protein, the domain of FK506 that binds rapamycin (FKBP), and the yeast inositol polyphosphate 5-phosphatase that cleaves the phosphate at the 5 position of PtdIns(4,5)P2) (24). Upon addition of rapamycin, the protein domains FKBP and FRB dimerize, resulting in selective depletion of PtdIns(4,5)P2 without the production of Fig. 2. Expression of SGK1.1 in mouse tissues and calculation of protein 2ϩ diacylglycerol (DAG), inositol 1,4,5-triphosphate (IP3), or Ca half-life. (A) Relative abundance of SGK1 and SGK1.1 proteins in tissues of an release. Fig. 3C shows that rapamycin induces translocation of SGK1-BAC transgenic mouse (TG) and wild-type littermates (WT). All isoforms CF-Inp to the plasma membrane and SGK1.1-RFP to the of SGK1 were first immunoprecipitated (IP) with a polyclonal antibody di- rected to the C-terminal 3ϫ HA tag of the protein, followed by immunoblot- cytosol/nucleus, demonstrating that release of SGK1.1 from the ting (WB) with anti-HA monoclonal Ab. Asterisks indicate the expected mo- plasma membrane parallels depletion of PtdIns(4,5)P2 and is not lecular mass of SGK1.1 (*) and SGK1 (**). (B) Half-life of SGK1.1 and SGK1 the consequence of other signaling pathways activated by PLC or 2ϩ calculated by pulse–chase with [35S]methionine. The graph represents the increase in intracellular Ca . mean of densitometric values from three independent experiments Ϯ SD. The N terminus of SGK1.1 contains a cluster of positively

4460 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0800958105 Arteaga et al. Downloaded by guest on September 27, 2021 Fig. 3. Subcellular localization of SGK1.1. (A) CHO cells cotransfected with SGK1.1 and PH-GFP show colocalization of the two proteins at the plasma membrane. A fraction of SGK1.1 is also seen in the nucleus. (B) Activation PH-PLC with m-3M3FBS translocates SGK1.1 and PH-GFP to the cytosol. (C) Cells transfected with Ds-RedSGK1.1, CF-Ins, and LRD. Upon addition of rapa- mycin (5 ␮M), CF-Ins moves from the cytosol to the plasma membrane, whereas Ds-RedSGK1.1 moves from the plasma membrane to the cytosol. Images in A and B were obtained with a Zeiss IE-24 LSM Meta laser confocal microscope and in C with a two-photon microscope IE-25LSM510NLO.

charged residues intercalated with bulky hydrophobic residues (Fig. 1B) reminiscent of protein motifs that bind plasma membrane phospholipids (24). of three consecutive positive charges Fig. 4. A polybasic motif in the N terminus of SGK1.1 tethers the protein to (K21K22R23) for neutral residues eliminated SGK1.1 from the plasma membrane and relocated the protein to the cytosol (Fig. the plasma membrane. (A) Removal of three positive residues from the polybasic motif in the N terminus of SGK1.1 changes its localization from the 4A). In contrast, substitution of the hydrophobic residues F19F20 or plasma membrane to the cytosol. (B) Mutations of large hydrophobic residues W27 for alanine targeted the protein mainly to the nucleus (Fig. 4B). in the polybasic motif, F19F20,orW27 change distribution of SGK1.1 from This indicates that the cluster of positively charged residues together plasma membrane to predominantly nuclei. with hydrophobic residues in the N terminus of SGK1.1 encode a motif that binds PtdIns(4,5)P2. Elimination of the cationic or hydrophobic element markedly diminishes the affinity of the motif SGK1.1 subcellular localization on ASIC1 activity, we injected for PtdIns(4,5)P2. Because enrichment in positively charged resi- oocytes with SGK1.1 mutants that displace the protein to the dues can also serve as a nuclear localization signal, SGK1.1 migrates cytosol and nucleus. Fig. 5C shows no effect of SGK1.1RRK and to the cytosol and is partially sequestered in the nucleus upon SGK1.1FF, whereas a small but statistically significant decrease, hydrolysis of PtdIns(4,5)P2 either by activation of phospholipase C 30% and 20%, was observed with the constitutively active double or inositol phosphatases. mutants SGK1.1RRK/S515D and SGK1.1FF/S515D. This indicates that Furthermore, preferential binding to PtdIns(4,5)P2 over membrane localization of SGK1.1 is important for the modulation PtdIn(3,4,5)P3 was demonstrated by depletion of cellular of ASIC1, although expression of a constitutively active mutant can PtdIn(3,4,5)P3 by treatment of cells with inhibitors of phosphati- partially compensate for localization. dylinositol 3-kinase (PI3K), LY294002, and wortmannin. These N1B-115 cells, which express endogenous ASIC1 transcript as NEUROSCIENCE were compared with cells treated with and IGF, both demonstrated by RT-PCR (Fig. 6A), were transfected with SGK1.1- activators of PI3K. Confocal analysis of SGK1.1 distribution did not GFP. Fluorescence distributed over the plasma membrane of the reveal a significant change in the subcellular localization (data not soma and neurites as expected for colocalization with PdtIns(4,5)P2 shown). (Fig. 6B). When the cells were examined for the presence of proton-activated currents, we observed transient inward currents Functional Effects of SGK1.1. The canonical SGK1 isoform increases characteristic of ASIC1 (Fig. 6C). The average ASIC current in cells activity of ENaC in the kidney and also modulates other channels transfected with GFP alone was 418.18 Ϯ 78.62 pA, whereas, in cells and transporters in various tissues (20). We first examined whether transfected with SGK1.1S515D-GFP, the average current was SGK1.1 regulates ENaC in Xenopus oocytes. Whole-cell amiloride- 66.73 Ϯ 13.03 pA, an 84% reduction from the control (Fig. 6D). sensitive currents in oocytes coinjected with or without SGK1.1 Transfection with the inactive form, SGK1.1S515A-GFP, did not were similar (Fig. 5A). By contrast, activation of the closely related change the magnitude of proton-activated currents. The effect on neuronal specific channel ASIC1 was decreased in the presence of ASIC1 was specific because measurements of endogenous voltage- SGK1.1 (Fig. 5B). The decrease depended on kinase activity activated sodium currents in the same neurons did not differ in cells because the constitutively active form (SGK1.1S515D) (25) induced expressing GFP or SGK1.1S515D-GFP (Fig. 6E). the greatest decrease, whereas a mutant that cannot be activated As the rates of activation (6.5 Ϯ 2.9 sϪ1) and desensitization Ϫ1 (SGK1.1S515A) did not show an effect. To investigate dependence of (0.77 Ϯ 0.13 s ) of ASIC1 were not affected by SGK1.1, it is

Arteaga et al. PNAS ͉ March 18, 2008 ͉ vol. 105 ͉ no. 11 ͉ 4461 Downloaded by guest on September 27, 2021 consistent with the absence of a SGK1.1’s consensus phosphory- lation motif (RXRXXS/T) in both the N and C termini of the ASIC1. Thus, the effect on surface expression must be indirect, as is the case for the regulation of ENaC expression by the canonical SGK1. In the latter instance, SGK1 phosphorylates the ubiquitin ligase Nedd4–2, decreasing its ability to bind to the C terminus of the channel subunits, thereby changing the rate of ENaC endocy- tosis (reviewed in ref. 26). It is unlikely, however, that ASIC1 follows such a mechanism because the C terminus of this channel does not have the proline-rich motif for binding of Nedd4–2. At this stage, we can only speculate that SGK1.1 modifies the traffic of ASIC1 in or out of the membrane by targeting an as-yet-unidentified substrate. Localization of SGK1.1 to the plasma membrane was also required because disruption of the PtdIns(4,5)P2-binding motif abrogated the effect on ASIC1. In a similar manner, removal of the N terminus of the canonical SGK1 eliminates regulation of ENaC (27). Therefore, this result confirms and extends our previous conclusion that correct subcellular localization is key for achieving functional specificity of the SGK1 isoforms (23). Regulation by compartmentation is partly overcome by rendering the cytosolic SGK1.1 mutants constitutively active because, owing to their dis- tribution over the whole cell, they have access to substrates that cannot be reached under normal conditions. Because SGK1.1 is released from the plasma membrane toward the cytosol and nucleus when the level of PtdIns(4,5)P2 is transiently diminished, it is possible that SGK1.1 may also play a role in transcriptional regulation of . The presence of several promoters in the SGK1 gene allows induction by a large and diverse set of conditions in a tissue- specific manner. For instance, glucocorticoids are among the main stimuli promoting transcription of the canonical SGK1 in Fig. 5. Effect of SGK1.1 on ENaC and ASIC1 currents in oocytes. (A) Oocytes peripheral tissues, whereas the SGK1.1 isoform is not regulated expressing ENaC alone or with SGK1.1 cRNA. Data represent the amiloride- by glucocorticoids in brain. By contrast, in neurons, we observed sensitive current normalized to control oocytes (ENaC alone) measured with the TEVC at a holding membrane potential of Ϫ60 mV. (B) Oocytes expressing increased expression of SGK1.1 by depolarization of the plasma membrane. It is known that activity-dependent neuronal mem- ASIC1 alone or with wild-type SGK1.1, constitutively active (SGK1.1S515D), or nonactivatable (SGK1.1S515A) cRNAs. (C) ASIC and mutants of SGK1.1 that brane depolarization induces transcription of genes, thereby change subcellular localization (RRK and FF) and same mutants with addition establishing structural and functional changes characteristic of of the activation residue S515D. Data are peak currents induced by changing neuronal plasticity (28). extracellular pH from 7.4 to 6.0. N ϭ number of independent experiments; It is relevant to notice that all microarray and qPCR studies thus n ϭ number of oocytes; error bars, SEMs; *, P Յ 0.01. far conducted to examine changes in SGK1 mRNA expression by various conditions and diseases have used probes that recognize cDNA regions common to all SGK1 isoforms; therefore, those unlikely that these processes are targets of SGK1.1 modulation but studies did not identify the isoform induced by the specific stimulus. rather reduction in number of channels at the cell surface. Transient The distinction is important owing to the functional differences of transfection of SGK1.1 in a stable CHO cell clone expressing the isoforms. Moreover, the relative low basal level of SGK1.1 ASIC1-FLAG showed a decrease in the level of channels expressed mRNA indicates that even a small change in total mRNA, if due at the plasma membrane examined by biotinylation of surface exclusively to increases in SGK1.1, represents a major change in proteins (Fig. 6F), whereas transfection with the SGK1.1FF mutant expression of the latter isoform. At the protein level, the transcrip- did not induced a significant change of ASIC1 surface expression. tional effect is further amplified because of the much higher We also examined whether SGK1.1S515D phosphorylates ASIC1. stability of SGK1.1 compared with SGK1. An ASIC-FLAG clone was transfected with SGK1.1S515D and It has become apparent that a recurrent theme among the SGK 32 analyzed for incorporation of [P] into immunoprecipitated proteins is tethering to membranes of organelles by means of ASIC1. These experiments did not show a significant change when specific motifs localized in the N terminus of these proteins. The compared with cells not transfect with SGK1.1S515D, suggesting that canonical SGK1 contains an amphipathic ␣-helix that targets the phosphorylation of the channel is not the mechanism whereby protein to the cytosolic surface of the endoplasmic reticulum (22, channel expression at the cell surface is attenuated [supporting 29); SGK3 contains a PX domain that binds to PtdIns(3)P, thereby information (SI) Fig. 8]. targeting the protein to early endosomes (30); SGK1.1 uses a cluster of cationic and large hydrophobic residues to bind PtdIns(4,5)P2 Discussion and tether the protein to the plasma membrane. Such motifs have We report a brain-specific SGK1.1 isoform that down-regulates the been described and shown to be necessary for targeting small activity of the neuronal ASIC1 channel, at least in part, by decreas- guanosine triphosphatases (GTPases) from the Ras, Rho, Arf, and ing its expression at the cell surface. Activation of SGK1.1 by Rab subfamilies to the plasma membrane (24). Because phosphorylation of the C-terminal hydrophobic motif (S515) en- PtdIns(4,5)P2 are the most abundant phosphoinositides in the hanced the effect on ASIC1, indicating that catalytic activity is plasma membrane, in resting conditions, SGK1.1 resides in this important to induce the signaling pathway that diminishes expres- compartment unless levels decrease transiently upon activation of sion of ASIC1 at the plasma membrane. Although an active kinase PLC or inositol phosphatases. Under this condition, SGK1.1 moves is necessary, SGK1.1 does not phosphorylate the channel protein, to the cytosol and accumulates in the nucleus using the same

4462 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0800958105 Arteaga et al. Downloaded by guest on September 27, 2021 Fig. 6. Functional effects of SGK1.1 on neuronal cells. (A) Products of RT-PCR from N1B-115 cells with (ϩ) and (Ϫ) serum amplified with specific primers from mouse ASIC1, -2, -3, and VR1 receptor. (B) Confocal image of N1B-115 cells transfected with SGK1.1-GFP and maintained in serum-free medium for 2 days. (C) Representative examples of endogenous ASIC1 whole-cell currents in differentiated N1B-115 cells elicited by changing external pH from 7.4 to 5.0 in cells transfected with GFP or SGK1.1S422D-GFP. (D) Average proton-activated currents of N1B-115 cells transfected with GFP or SGK1.1S515D-GFP, n ϭ 13, P ϭ 0.00018. ϩ (E) Endogenous TTX-sensitive voltage-activated Na currents in neurons transfected with GFP or SGK1.1S515D, n ϭ 12. (F) CHO cells stably expressing ASIC1-FLAG transiently transfected with SGK1.1 or SGK1.1FF. Representative gel of surface biotinylated proteins probed with anti-FLAG monoclonal and densitometric analysis of three independent experiments normalized to controls.

polybasic cluster, which also serves as a nuclear localization signal. TAA (SGK1.1 sense), and GGTTTGGCGTGAGGGTTGGAGGAC (antisense for both In nuclei, SGK1.1 may have specific targets implicated in transcrip- amplicons). PCRs were prepared with iQ SYBR green SuperMix (Bio-Rad) with 3 ␮ tion regulation that need to be identified. mM MgCl2, and 0.5 l of cDNA template. All samples were run in triplicate on The existence of multiple SGK1 promoters and transcripts allows iCycler (Bio-Rad). qPCR conditions were 95°C for 2 min, followed by 40 cycles of SGK1 expression to be regulated at multiple levels including 95°C for 15 s and 68°C for 45 s. Melting curve analysis was added after the final transcription, stability, subcellular localization, and translatability. PCR cycle. Control reactions were run with no cDNA template or with non- reverse-transcribed RNA. Starting mRNA quantities were calculated from the The fact that SGK1 expression is regulated at several levels is standard curves generated by using serial dilutions of plasmid DNA containing relevant to SGK1’s multiple functions as a modulator of neuronal SGK1.1 or SGK1 and respective qPCR primers. Calculated mRNA expression levels plasticity and a mediator of neuronal survival. were normalized to the expression levels of GAPDH in the same cDNA sample. qPCR for GAPDH was performed as described above for SGK1.1, by using the Methods following primers: GATGGTGAAGGTCGGTGTGAACGGAT (sense), and CCTTG- Cell Culture, Transfection, and Treatments. CHO cells and N1E-115 cells seeded on GAGGCCATGTAGGCCATGA (antisense). coverslips or six-well Petri dishes were transfected with Lipofectamine-2000 (Invitrogen). One day after transfection, the medium of N1E-115 cells was Metabolic Labeling and Pulse–Chase Experiments. Transfected cells were washed changed to serum-free to induce cell differentiation. Experiments were carried with methionine- and cysteine-free medium, followed by incubation with 150 out 2–3 days after induction of differentiation. For RNA extraction, N1E-115 cells ␮Ci/ml of Express Cell Labeling Mix (PerkinElmer Life Sciences) for 20 min. Cells were kept 3 days on serum-free medium, followed by treatment with 1 ␮M ϩ were chased with medium containing a 10-fold molar excess of both methionine/ dexamethasone for3hormedium modified to contain 50 mM K maintaining NEUROSCIENCE cysteine and 0.1 mg/ml cycloheximide for the indicated time periods. Cells were 290 mOsm for 90 min. lysed and prepared for with anti-SGK1 antibody as de- scribed (22). Gels were exposed to x-ray film and analyzed by densitometry using Cloning of SGK1 Spliced Isoforms and Mutagenesis. Mouse brain total RNA was Bio-Rad G800 and QuantityOne software. isolated with TRIzol (Invitrogen). Single-strand cDNA was synthesized with Su- perScript III reverse transcriptase and primed with oligo(dT). Forward (GGAA- Cell lines constitutively expressing ASIC1-FLAG GATGGTAAACAAAGACATGAATGG and CTCGGTCCGCAGCTATGGGCGAGATG) Membrane Protein Biotinylation. and reverse (GAGGAAGGAATCCACAGGAGGTGC) primers were used to amplify were biotinylated with Sulfo-NHS-SS-Biotin (Pierce). Protein concentration was the coding region of spliced SGK1 isoforms by using high-fidelity Taq polymerase measured with the BCA kit (Pierce), and equal amounts of total protein were (Roche). Products were ligated to pCDNA3.1/V5-His TOPO vector (Invitrogen). processed. Biotinylated proteins were recovered with Streptavidin-agarose Mutations were inserted with a QuikChange mutagenesis kit (Stratagene). All beads (Pierce). The amount of added beads was adjusted to ensure complete final constructs were sequenced. recovery of biotinylated proteins from lysates. Biotinylated proteins were eluted from the beads by heating to 90°C in SDS/PAGE sample buffer. qRT-PCR. Total RNA from mouse tissues (four control and four dexamethasone- treated mice (0.5-mg i.p. injection for 18 h) was treated with RNase-free DNase Immunoblot Analysis. Samples were separated by electrophoresis in 10% SDS/ (New England BioLabs). A similar procedure was used for N1E-115 cells. Reverse PAGE and transferred to Immobilon-P membrane (Millipore). After blocking with transcription of 3 ␮g of total RNA was performed by using SuperScript RTIII 5% dry milk, the membranes were probed with primary antibody, anti-FLAG-HRP (Invitrogen) and oligo(dT) primer. The primers used for qPCR were CGTCAAAGC- (Sigma). Signals were developed with ECLϩ (Amersham), and blots were exposed CGAGGCTGCTCGAAGC (SGK1 sense), GAAGGCGGATCGGGATACAGATGCAG- to BioMax MR Film (Eastman Kodak).

Arteaga et al. PNAS ͉ March 18, 2008 ͉ vol. 105 ͉ no. 11 ͉ 4463 Downloaded by guest on September 27, 2021 Immnunofluorescence Microscopy. Cells were fixed with 4% paraformaldehyde, external pH to 6.0, and peak currents were computed for analysis. The ENaC permeabilized with 1% Triton X-100 in PBS, and blocked with 1% goat serum. V5 current was measured by adding 20 ␮M amiloride to the bath. monoclonal Ab was incubated for 1–2 h, washed with PBS five times, and secondary anti-mouse IgG goat secondary Ab conjugated with Alexa Fluor 594 Patch-Clamp of N1E-115 Cells. Fluorescent cells were selected under the micro- was added for 1 h. Cells were examined with a Zeiss IE-24 LSM Meta laser confocal scope for whole-cell patch-clamp measurements with an Axopatch 200B ampli- microscope or a two-photon microscope IE-25LSM510NLO. fier (Axon Instruments). Neither cell capacity nor series resistance (Ϸ8M⍀) was compensated. For voltage-clamp experiments, the membrane potential was held In Situ Hybridization. In situ hybridization was performed on 30-␮m-thick free- at Ϫ60 mV. ASIC was activated by changing the bath solution from pH 7.4 to pH floating sections of paraformaldehyde-fixed mouse brain. The antisense probe 5.0 by using a modified mechanical switching perfusion system (SF-77B, Perfusion was specific for SGK1.1, ϩ354 to Ϫ456 of the coding sequence (GenBank Fast-Step; Warner Instrument) (34). Voltage-dependent Na channels were acti- BC070401). Riboprobes were labeled with DIG-dUTP, DIG-RNA-labeling kit SP6/T7 vated by depolarizing voltage steps with 20 mV increasing amplitude starting (Roche). Sections were treated with proteinase K (Roche; 15 ␮g/ml in TBS with 2 from holding potential to ϩ80 mV. To minimize voltage-errors, cell capacity and series-resistance compensation were used at 60%. Leakage currents were sub- mM CaCl2) for 20 min at 37°C, washed in cold TBS [50 mM Tris (pH 7.5) and 150 mM NaCl], followed by a 10-min incubation in 0.5% acetic anhydride in 0.1 M Tris (pH tracted online by a P/6 program with hyperpolarizing pulses from the holding Ϫ 8.0), washed in TBS, and incubated in hybridization buffer (50% formamide, 10% potential of 80 mV. For selective analysis of voltage-dependent Na currents, dextrane sulfate, 4ϫ SSC, 2.5ϫ Denhardt’s solution, 0.25 mg/ml ss DNA, 0.6 mg of inward peak values after depolarization were determined before and after a ␮ yeast tRNA, and 0.025% SDS) at 55°C for 1 h. Denatured riboprobes (final 1-min perfusion with 2 M TTX, and differences were plotted against voltage. concentration 0.5 ␮g/ml) added to hybridization buffer were incubated at 55°C The bath solution contained 150 mM NaCl, 5 mM KCl, 2 mM MgCl2, 2 mM CaCl2, for 16 h. Posthybridization washes: three for 10 min in 2ϫ SSC at 23°C, three for and 10 mM glucose, and the pH was adjusted to 7.45 (Hepes 10) or 5 (Mes 10). The 20 min in 50% formamide/1ϫ SSC at 55°C, one for 10 min in 1ϫ SSC, and one for pipette solution was made of 140 mM KCl, 5 mM NaCl, 2 mM MgCl2, 5 mM EGTA, 2 mM MgATP, and 10 mM Hepes, and the pH was 7.35. Membrane currents were 10 min in TBS, incubated in blocking solution (1% DIG blocking reagent, Roche, low-pass filtered at 1 kHz, digitized with 16 bits, and stored on a computer for 10% sheep serum, and 2% mouse serum) for1hat23°C. Anti-DIG-AP (Roche) was analysis. AISC currents were determined as the peak value of change in current applied overnight at 4°C, diluted 1:4,000 in blocking solution. Detection: NBT/ after switching to pH 5. BCIP substrate (Roche) with 240 ␮g/ml levamisole in 100 mM Tris (pH 9.5), 100 mM NaCl, and 50 mM MgCl2 in the dark for color development. Sections were mounted in 0.1 M phosphate buffer (pH 7.0), 7.5% gelatin, and 50% glycerol. Data Analysis. For every experiment, equal numbers of cells were measured for experimental and control conditions (n ϭ 8–15 oocytes per experiment). Results were normalized to the mean value of the control and computed as the mean Ϯ Xenopus laevis Oocyte Injection and Two-Electrode Voltage Clamp (TEVC). Stage SEM of control and experimental conditions. Experiments were repeated with V and VI Xenopus laevis oocytes were injected with 5–50 ng of cRNA and 7–12 independent batches of oocytes (N). Statistical significance was proved by incubated 24 h for ENaC and 72 h for ASIC1 at 16°C, supplemented with amiloride Student’s t test for overall experiments (n). for oocytes injected with ENaC. cRNA from rASIC1, rENaC (␣, ␤, and ␥), and mSGK1.1 were transcribed with T7 mMessagemMachine (Ambion). Channel and ACKNOWLEDGMENTS. We thank Drs. T. Inoue and T. Meyer for the constructs SGK1.1 cRNAs were injected in a 1:2 ratio in a volume of 50 nanoliters per oocyte. LDR and CF-Inp, and P. De Camilli for PH-GFP plasmid. This work was supported TEVC experiments were performed as in ref. 31. The composition of bath solution by National Institutes of Health Grant DK054062.06A1 and American Heart was 150 mM NaCl, 1 mM CaCl2, 1 mM KCl, 10 mM Hepes (pH 7.5). The holding Association Grant 0555777 (to C.M.C.). T.C. is the recipient of an American membrane potential was Ϫ60 mV. ASIC1 was activated by a rapid change of the Physiological Society training grant.

1. Webster MK, Goya L, Ge Y, Maiyar AC, Firestone GL (1993) Characterization of , a 16. Tsai KJ, Chen SK, Ma YL, Hsu WL, Lee EH (2002) sgk, a primary glucocorticoid-induced novel member of the serine/threonine protein kinase gene family which is transcrip- gene, facilitates memory consolidation of spatial learning in rats. Proc Natl Acad Sci tionally induced by glucocorticoids and serum. Mol Cell Biol 13:2031–2040. USA 99:3990–3995. 2. Waldegger S, Barth P, Raber G, Lang F (1997) Cloning and characterization of a putative 17. Chao CC, Ma YM, Lee EH (2007) Protein kinase CK2 impairs spatial memory formation human serine/threonine protein kinase transcriptionally modified during anisotonic through differential cross talk with PI-3 kinase signaling: activation of Akt and inac- and isotonic alterations of cell volume. Proc Natl Acad Sci USA 94:4440–4445. tivation of SGK1. J Neurosci 27:6243–6248. 3. Chen SY, et al. (1999) Epithelial sodium channel regulated by -induced 18. Von Hertzen LS, Giese K (2006) Memory reconsolidation engages only a subset of protein sgk. Proc Natl Acad Sci USA 96:2514–2595. immediate-early genes induced during consolidation. J Neurosci 25:1935–1942. 19. Ma YL, Tsai MC, Hsu WL, Lee EH (2006) SGK protein kinase facilitates the expression of 4. Wulff P, et al. (2002) Impaired renal Naϩ retention in the -knockout mouse. J Clin long-term potentiation in hippocampal neurons. Learn Mem 13:114–118. Invest 110:1263–1268. 20. Lang F, et al. (2006) (Patho)physiological significance of the serum- and glucocorticoid- 5. Brunet A, et al. (2001) Protein kinase SGK mediates survival signals by phosphorylating inducible kinase isoforms. Physiol Rev 86:1151–1178. the forkhead transcription factor FKHRL1 (FOXO3a). Mol Cell Biol 21:952–965. 21. Firestone GL, Giampaolo JR, O’Keeffe BA (2003) Stimulus-dependent regulation of 6. Hertweck M, Gobel C, Baumeister R (2004) C. elegans SGK-1 is the critical component serum and glucocorticoid inducible protein kinase (SGK) transcription, subcellular in the Akt/PKB kinase complex to control stress response and life span. Dev Cell localization and enzymatic activity. Cell Physiol Biochem 13:1–12. 6:577–588. 22. Arteaga MF, Wang L, Ravid T, Hochstrasser M, Canessa CM (2006) An amphipathic helix 7. Nishida Y, et al. (2004) Alteration of serum/glucocorticoid regulated kinase-1 (sgk-1) targets serum and glucocorticoid-induced kinase 1 to the endoplasmic reticulum- in rat hippocampus after transient global ischemia. Brain Res Mol associated ubiquitin-conjugation machinery. Proc Natl Acad Sci USA 103:11178–11183. Brain Res 123:121–125. 23. Arteaga MF, Alvarez de la Rosa D, Alvarez JA, Canessa CM (2007) Multiple translational 8. Imaizumi K, Tsuda M, Wanaka A, Tohyama M, Takagi T (1994) Differential expression isoforms give functional specificity to serum- and glucocorticoid-induced kinase 1. Mol of sgk mRNA, a member of the Ser/Thr protein kinase gene family, in rat brain after CNS Biol Cell 18:2072–2078. injury. Brain Res Mol Brain Res 26:189–196. 24. Heo WD, et al. (2006) PI(3,4,5)P3 and PI(4,5)P2 lipids target proteins with polybasic 9. Iwata S, Nomoto M, Morioka H, Miyata A (2004) Gene expression profiling in the clusters to the plasma membrane. Science 314:1458–1460. midbrain of striatal 6-hydroxydopamine-injected mice. Synapse 51:279–286. 25. Kobayashi T, Cohen P (1999) Activation of serum- and glucocorticoid-regulated protein 10. Stichel CC, et al. (2005) sgk1, a member of an RNA cluster associated with cell death in kinase by agonists that activate phosphatidylinositide 3-kinase is mediated by 3-phos- a model of Parkinson’s disease. Eur J Neurosci 21:301–316. phoinositide-dependent protein kinase-1 (PDK1) and PDK2. Biochem J 339:319–328. 11. Schoenebeck B, Bader V, Zhu XR, Schmitz B, Lubbert H (2005) Sgk1, a cell survival 26. Kamynina E, Staub O (2002) Concerted action of ENaC, Nedd4–2, and Sgk1 in trans- ϩ response in neurodegenerative diseases. Mol Cell Neurosci 30:249–264. epithelial Na transport. Am J Physiol Renal Physiol 283:F377–F387. 27. Na´ray-Fejes-To´th A, Helms MN, Stokes JB, Fejes-To´th G (2004) Regulation of sodium 12. Tudor M, Akbarian S, Chen RZ, Jaenisch R (2002) Transcriptional profiling of a mouse transport in mammalian collecting duct cells by aldosterone-induced kinase, SGK1: model for Rett syndrome reveals subtle transcriptional changes in the brain. Proc Natl Structure/function studies. Mol Cell Endocrinol 217:197–202. Acad Sci USA 99:15536–15541. 28. West AE, et al. (2001) Calcium regulation of neuronal gene expression. Proc Natl Acad 13. Nuber UA, et al. (2005) Up-regulation of glucocorticoid-regulated genes in a mouse Sci USA 98:11024–11031. model of Rett syndrome. Hum Mol Genet 14:2247–2256. 29. Bogusz AM, Brickley DR, Pew T, Conzen SD (2006) A novel N-terminal hydrophobic 14. Rangone H, et al. (2004) The serum- and glucocorticoid-induced kinase SGK inhibits motif mediates constitutive degradation of serum- and glucocorticoid-induced mutant -induced toxicity by phosphorylating serine 421 of huntingtin. Eur kinase-1 by the ubiquitin-proteasome pathway. FEBS J 273:2913–2928. J Neurosci 19:273–279. 30. Xu J, Liu D, Gill G, Songyang Z (2001) Regulation of -independent survival kinase 15. Ge´ranton SM, Morenilla-Palao C, Hunt SP (2007) A role for transcriptional repressor (CISK) by the Phox homology domain and phosphoinositid. J Cell Biol 154:699–705. methyl-CpG-binding protein 2 and plasticity-related gene serum- and glucocorticoid- 31. Coric T, Zhang P, Todorovic N, Canessa CM (2003) The extracellular domain determines inducible kinase 1 in the induction of inflammatory pain states. J Neurosci 27:6163– the kinetics of desensitization in acid-sensitive ion channel 1. J Biol Chem 278:45240– 6173. 45247.

4464 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0800958105 Arteaga et al. Downloaded by guest on September 27, 2021