Regulation of OSR1 and the sodium, potassium, two chloride cotransporter by convergent signals

Samarpita Senguptaa, Andrés Lorente-Rodrígueza, Svetlana Earnesta, Steve Stippeca, Xiaofeng Guob, David C. Trudgianb, Hamid Mirzaeib, and Melanie H. Cobba,1

Departments of aPharmacology and bBiochemistry, University of Texas Southwestern Medical Center, Dallas, TX 75390-9041

Contributed by Melanie H. Cobb, October 3, 2013 (sent for review October 6, 2012) The Ste20 family protein kinases oxidative stress-responsive 1 (ENaC) (23, 24). SGKs and the related Akt enzymes are acti- (OSR1) and the STE20/SPS1-related proline-, alanine-rich kinase vated by phosphorylation on multiple sites, most prominently a directly regulate the solute carrier 12 family of cation-chloride residue in the activation loop by the phosphoinositide-dependent cotransporters and thereby modulate a range of processes includ- protein kinase and on a second site in a C-terminal hydrophobic ing cell volume homeostasis, blood pressure, hearing, and kidney motif (25). The kinase that phosphorylates the hydrophobic function. OSR1 and STE20/SPS1-related proline-, alanine-rich kinase motif site under many circumstances is the mammalian target are activated by with no lysine [K] protein kinases that phosphory- of rapamycin complex 2 (mTORC2), which provides an addi- late the essential activation loop regulatory site on these kinases. tional phosphatidylinositol-3 kinase (PI3K)-dependent input to We found that inhibition of phosphoinositide 3-kinase (PI3K) these kinases (26–33). reduced OSR1 activation by osmotic stress. Inhibition of the PI3K In this study, we show that OSR1 is phosphorylated not only target pathway, the mammalian target of rapamycin complex 2 by WNKs but also on a C-terminal site, conserved in SPAK, by (mTORC2), by depletion of Sin1, one of its components, decreased mTORC2. These studies reveal a link between WNK-OSR1/SPAK activation of OSR1 by sorbitol and reduced activity of the OSR1 and the PI3K-mTORC2 cascade that suggests that OSR1 and substrate, the sodium, potassium, two chloride cotransporter, in SPAK integrate signals from osmosensing and survival pathways. HeLa cells. OSR1 activity was also reduced with a pharmacological inhibitor of mTOR. mTORC2 phosphorylated OSR1 on S339 in vitro, Results and mutation of this residue eliminated OSR1 phosphorylation by OSR1 Is Regulated by a PI3K-Dependent Mechanism. In response to mTORC2. Thus, we identify a previously unrecognized connection changes in tonicity, OSR1 is activated through phosphorylation of the PI3K pathway through mTORC2 to a Ste20 protein kinase and of its activation loop by WNKs. Previously, we failed to find ion homeostasis. activation of OSR1 by serum or epidermal growth factor (4). A recent report, however, indicates that it participates in insulin- phosphoregulation | Akt | ion transport | WNK1 regulated events in a PI3K-sensitive manner (34). To confirm the PI3K sensitivity and retest potential regulation by growth factors, he protein kinases oxidative stress-responsive 1 (OSR1) and we examined effects of the PI3K inhibitor wortmannin on OSR1 Tits homolog the STE20/SPS1-related proline-, alanine-rich activity in HeLa cells stimulated with sorbitol or serum. The kinase (SPAK or PASK) are the mammalian members of the sodium, potassium, two chloride cotransporters 1 and 2 (NKCC1 germ-cell kinase VI subgroup of the large Ste20 branch of the and NKCC2) are related ion cotransporters of the solute carrier mammalian kinome. OSR1 and SPAK directly regulate the sol- 12 family that are phosphorylated and activated by OSR1/SPAK ute carrier 12 family of cation-chloride cotransporters which (2, 15). The activity of immunoprecipitated OSR1 was measured modulate ion homeostasis throughout the body (1, 2). OSR1/ using a recombinant N-terminal fragment of NKCC2, residues – SPAK kinase domains lie close to their N-termini and they con- 1 175, as substrate (9). Sorbitol stimulated OSR1 activity, but A tain two additional conserved regions named “PF1” and “PF2” serum did not, consistent with earlier results (Fig. 1 ). Wortmannin [PASK and Fray (Drosophila homolog)] (3). PF1 is a C-terminal decreased serum-stimulated phosphorylation of Akt on its hydro- extension to the kinase domain and is required for enzyme ac- phobic motif site and also reduced sorbitol-stimulated OSR1 kinase tivity (4). PF2 binds the consensus motif [(R/K)FX(V/I)] (5) in substrates including ion cotransporters and in regulators. OSR1 Significance and SPAK are activated by with no lysine [K] (WNK) protein kinases, which phosphorylate the essential activation loop regu- With no lysine [K] (WNK) protein kinases are sensitive to latory site as well as a second site in the PF1 region with an un- changes in osmotic stress. Through the downstream protein defined function (6–9). kinases oxidative stress-responsive 1 (OSR1) and STE20/SPS1- The four WNK protein kinases are large enzymes notable for related proline-, alanine-rich kinase, WNKs regulate a family of the alternative placement of the essential ATP-binding lysine ion cotransporters and thereby modulate a range of processes residue in their catalytic domains, distinguishing them from other including cell volume homeostasis, blood pressure, hearing, members of the protein kinase superfamily (10, 11). Initial at- and kidney function. We found that a major phosphoinositide tention was focused on these enzymes because certain mutations 3-kinase target pathway, the mammalian target of rapamycin in two family members cause pseudohypoaldosteronism type II, a complex 2, also phosphorylates OSR1, coordinating with WNK1 heritable form of hypertension (12). WNKs are activated by to enhance OSR1 and ion cotransporter function. changes in tonicity. Cellular reconstitution studies and mouse genetics demonstrated the importance of WNK function in cell Author contributions: S. Sengupta, S.E., D.C.T., H.M., and M.H.C. designed research; – S. Sengupta, A.L.-R., S.E., S. Stippec, and X.G. performed research; X.G., D.C.T., and H.M. volume regulation and maintenance of blood pressure (13 19). contributed new reagents/analytic tools; S. Sengupta, A.L.-R., S.E., S. Stippec, D.C.T., and H.M. Control of cation-chloride cotransporters through OSR1 and analyzed data; and S. Sengupta, A.L.-R., and M.H.C. wrote the paper. SPAK is among the best-documented actions of WNKs in diverse The authors declare no conflict of interest. – tissues (5, 20 22). 1To whom correspondence should be addressed. E-mail: melanie.cobb@utsouthwestern. WNKs also regulate serum- and glucocorticoid-inducible pro- edu. tein kinases (SGKs) through a noncatalytic mechanism leading This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. to increased sodium influx through the epithelial sodium channel 1073/pnas.1318676110/-/DCSupplemental.

18826–18831 | PNAS | November 19, 2013 | vol. 110 | no. 47 www.pnas.org/cgi/doi/10.1073/pnas.1318676110 Downloaded by guest on September 28, 2021 A OSR1 Is Regulated by WNK1 and mTORC2. Because of the input of 5 *** mTORC2 to SGK (27), we tested the possibility that mTORC2 4 might be responsible for the PI3K-dependence of OSR1 acti- 3 vation by sorbitol. To evaluate the contribution of mTORC2, we inhibited its activity by depleting Sin1, a required component of 2 complex 2 (26). Knockdown of Sin1 reduced activation of OSR1 relative units 1 by sorbitol (Fig. 1B), supporting the conclusion that mTORC2 is OSR1 kinase activity, 0 a PI3K-dependent input to OSR1. In contrast to NKCC2 which Control Sorbitol FBS Wortmannin - + - + - + displays tissue-restricted expression in kidney, NKCC1 is expressed Kinase in most cell types including HeLa. To test the consequences of assay NKCC Sin1 depletion on cation-chloride cotransporter function, we OSR pAkt * A 120 B 5 ** 100 OSR 4 NKCC 80 3 60 relative units 2 40 OSR1 kinase activity, 1

Fold kinase activity 20 0 Sorbitol - - + + - - + + 0 siControl siSin1 Sorbitol - + + + Rapa KU Kinase assay

Sin1 BIOCHEMISTRY OSR1 ERK1/2 NKCC Akt pAkt B

C

** Rb uptake 86 WNK1 kinase activity, % WNK1 Fold bumetanide-sensitive Fold bumetanide-sensitive

Kinase assay Fig. 1. PI3K pathways influence OSR1 activation and function in response to sorbitol. (A) HeLa cells were pretreated with 50 nM wortmannin (Wort) and GST-OSR1KR then stimulated with 0.5 M sorbitol or 20% FBS to stimulate Akt. Endogenous WNK1 OSR1 was immunoprecipitated and its activity measured using recombinant Sin1- NKCC2 1–175. Four such experiments were quantified (Top). A representative autoradiogram is shown with immunoblots of OSR1 and Akt phosphorylated on S473. Phosphorylation of Akt at S473 measured as an indicator of mTORC2 Fig. 2. Inhibition of mTORC2 reduces OSR1 activity. (A) HeLa cells were activity and pAkt as a positive control of stimulation conditions. n = 4, one pretreated for 15 min with either 100 nM rapamycin (Rapa) or 300 nM KU- way ANOVA, P < 0.0001; ***P = 0.0003 (adjusted from Tukey’s test). Error 0063794 (KU) and then stimulated with 0.5 M sorbitol (Sorb). Endogenous bars show standard deviation. (B) Effect of depletion of endogenous mTORC2 OSR1 was immunoprecipitated and its autophosphorylation and phosphor- by knockdown of Sin1 on OSR1 autophosphorylation (white bars) and activity ylation of recombinant NKCC2 1–175 were measured by immune-complex toward substrate (gray bars) upon sorbitol stimulation. n = 5, **P < 0.05. Error kinase assay. The incorporated radioactivity was measured by scintillation bars show standard error. Immunoblots show efficiency of Sin1 knockdown. counting and normalized to the amount of OSR1 immunoprecipitated in ERK1/2 was used as the loading control. Expression of Sin1 was normalized to each case. Cells stimulated with sorbitol without any drugs and without ERK1/2 expression (Bottom). (C) HeLa cells were transfected with siRNA for any other treatment were used as controls. Four experiments with 10 rep- Sin1 and scrambled siRNA as control (siC). Bumetanide sensitive uptake of licates were used to calculate standard error (error bars). *P < 0.05. Immu- 86Rb was measured as an assay of NKCC1 activity. The fold change in NKCC1 noblots of Akt and pS473 Akt are shown (Lower). (B) After depletion of Sin1, activity was normalized to control and average of three separate experi- HeLa cells were stimulated with sorbitol in duplicate and endogenous WNK1 ments each performed in triplicate was plotted. Error bars are standard error. was immunoprecipitated and detected by blotting. Duplicate samples of **P < 0.05. WNK1 were used in immune complex kinase assays using OSR1 K46R as substrate. Incorporated radioactivity was detected by autoradiography and the Coomassie blue-stained gel of the substrate is immediately below. Sorbitol- stimulated activity is shown as percent of unstimulated activity (Upper); error activity, indicating an effect of PI3K on this WNK-dependent bars show standard deviation. Sin1 depletion was assessed by immunoblotting pathway. Unexpectedly, an increase in Akt phosphorylation was lysates (lanes in Lower). Sin1 is the upper band, indicated by the tick mark. also noted with sorbitol (Figs. 1A and 2A). n = 4. CTRL, control.

Sengupta et al. PNAS | November 19, 2013 | vol. 110 | no. 47 | 18827 Downloaded by guest on September 28, 2021 measured the activity of endogenous NKCC1 in HeLa cells. A OSR1 wt OSR1 KR Disruption of mTORC2 by Sin1 knockdown caused a substantial IP: Sin1WNK1 N - Sin1WNK1 N - decrease in endogenous NKCC1 activity, measured as rubidium Kinase uptake sensitive to the loop diuretic bumetanide, an NKCC in- assay hibitor (Fig. 1C). Taken together, these data suggest that mTORC2 regulates OSR1 activity and function. Coomassie To determine whether OSR1 is regulated by other mTOR com- 6 plexes, we treated sorbitol-stimulated cells with rapamycin which OSR wt inhibits mTORC1 better than mTORC2 and with KU0063794 (KU) 4 OSRKR

which inhibits both mTORC1 and mTORC2. OSR1 was immu- OSR1 2 noprecipitated from cells and assayed with NKCC. We observed that treatment with KU0063794 for 15 min reduced OSR1 ac- phosphorylation 0 IP Sin1 IP Wnk1 IP Pr tivity (Fig. 2A), although the effect of rapamycin was not statis- IP: Sin1 WNK1 Preimmune tically significant, suggesting that the regulation of OSR1 is primarily by mTORC2. Both compounds inhibited phosphor- B ylation of Akt, although rapamycin was less effective. mTORC2 100 activity toward both OSR1 and Akt is evident in sorbitol-treated 80 cells. Whether mTORC2 activity increases under these conditions 60 or if its access to substrates is enhanced is not known. The fact 40 that brief exposure to KU0063794 was sufficient to impair OSR1 20 activation implies that this is a rapid regulatory mechanism. For

% OSR1 Kinase Activity % OSR1 0 comparison, we examined the effects of the depletion of WNK1 Beads WNK1 Sin1 WNK1+Sin1 without or with KU0063794 to block mTORC2. Depletion of OSR plus WNK1 inhibited OSR1 activity strongly (Fig. S1), consistent with Beads + previous results from multiple groups indicating that phosphor- WNK1 + + + ylation of the activation loop by WNK1 is essential for activation IP Sin1 + + + + of OSR1 (6, 7, 9). Kinase assay WNK1 Is Not Regulated by mTORC2. Another mechanism by which OSR1 mTORC2 might regulate OSR1 is through direct effect on WNK1 WNK1 itself. Several mTOR sites have been reported on WNK1 (35), Sin although all lie in the C-terminal region not near the kinase do- input IP Sin1 main. To test this hypothesis, we depleted HeLa cells of Sin1 to disrupt mTORC2, immunoprecipitated endogenous WNK1 from Sin1 these cells, and tested its activity using an in vitro kinase assay with mTOR kinase-dead OSR1K46R as a substrate. No difference in WNK1 p70S6K activity in cells depleted of Sin1 was detected compared with Fig. 3. WNK1 and mTORC2 phosphorylate OSR1 and increase its activity. (A) WNK1 from control cells (Fig. 2B), consistent with previous Phosphorylation of OSR1 wild type (wt) and the inactive K46R (KR) mutant results showing no effects of wortmannin on WNK1 activity (36). by Sin1 and WNK1 immunoprecipitates was measured in vitro. Immuno- Furthermore, mTORC2 and WNK1 do not coimmunoprecipitate precipitates with nonimmune serum (N) and no added enzyme (−) were used (Fig. S2A). as controls. n = 3. The Coomassie blue-stained gel shows the relative amounts of OSR1 proteins present. Incorporated radiolabel was quantified mTORC2 Directly Phosphorylates OSR1 on S339. To determine if using scintillation counting and normalized to the amount of OSR1 in each reaction. Results are plotted (Lower). (B) Wild type OSR1 was incubated for OSR1 and SPAK are mTORC2 substrates, we tested whether – mTORC2 could directly phosphorylate OSR1. We used anti- 30 min with recombinant WNK1 1 490 alone, with Sin1 immunoprecipitate alone (IP Sin1), with beads alone or with both enzyme preparations. OSR1 bodies against Sin1 to immunoprecipitate mTORC2 from HeLa B Bottom kinase activity was then measured with NKCC (Top); error bars show stan- cells (Fig. 3 , ) and then measured phosphorylation of dard deviation. A representative assay is shown (Middle) with immunoblots wild type OSR1 and the inactive mutant OSR1K46R in immune- indicating the amounts of enzymes present. Bottom shows that mTOR is complex kinase assays. Like immunoprecipitated WNK1, mTORC2 present in Sin1 immunoprecipitates (IP sin1), but the mTOR substrate p70 S6 phosphorylated both forms of OSR1 in vitro (Fig. 3A). Re- kinase is not. combinant WNK1 kinase domain (which has higher activity than the endogenous protein) caused a marked stimulation of recombinant OSR1 kinase activity in vitro; a smaller increase We found that T183 and S325 were phosphorylated by WNK1, as was detected with immunoprecipitated mTORC2, and the com- reported previously (6). A single site, S339, also within the PF1 bination of both caused a greater increase in OSR1 kinase activity region, was phosphorylated by mTORC2 (Fig. 4A). This region is than either alone (Fig. 3B). nearly identical in SPAK and the comparable residue was iden- In vitro phosphorylation by WNK1 on an additional site, S325, tified as a phosphorylation site previously and associated with in the OSR1 PF1 segment was reported by Alessi and coworkers increased kinase activity (37). We mutated S339 to alanine and (6). To rule out an involvement of this residue as well as the confirmed that phosphorylation of this OSR1 mutant by mTORC2 adjacent serine S324 in phosphorylation by mTORC2, we again was lost. The OSR1 truncation mutant, residues 1–323, produced immunoprecipitated mTORC2 and tested phosphorylation of in bacteria can be activated by WNK1, but is unstable (9). Without OSR1 S324A and SS324/325AA (SASA) mutants. Both mutants S339, this fragment is not phosphorylated by mTORC2 (Fig. 4B). were phosphorylated as well as wild type OSR1, indicating that neither of these sites is phosphorylated by mTORC2 (Fig. S2B). Discussion Mass spectrometry (MS) was used to determine the site phos- Changes in tonicity regulate the WNK-OSR1/SPAK pathway to phorylated by mTORC2. OSR1 was phosphorylated by recombinant control ion cotransporters for volume and ion homeostasis. We WNK1 kinase domain alone or in combination with immuno- find that mTORC2 also contributes to enhanced OSR1 activity. precipitated mTORC2 or a control serum immunoprecipitate. Inhibiting mTORC2 does not inhibit WNK1 activity, indicating

18828 | www.pnas.org/cgi/doi/10.1073/pnas.1318676110 Sengupta et al. Downloaded by guest on September 28, 2021 A IP Sin1 + + - PF1 and PF2 regions. The PF1 segment is essential for their activity WNK1 -+ + (4). Interaction of an N- or C-terminal extension, such as PF1, with its protein kinase catalytic core is a common stabilizing and regu- OSR1 K46R latory event in the protein kinase family (38, 39). Protein kinase C phosphorylation and Akt, for example, are regulated by phosphorylation on residues C-terminal to their catalytic cores; the best understood of these sites are often referred to as hydrophobic motif and turn motif sites. From structural studies of these kinases, hydrophobic motif phosphorylation promotes activation by docking a C-terminal ex- tension onto a regulatory site on the catalytic core (40). Phos- phorylation of the turn motif enhances stability of the enzyme by protecting it from degradation directly or indirectly and perhaps by providing additional ligands for interaction between the C-terminal extension and the kinase core (28, 41). OSR1 is phosphorylated on the activation loop and an addi- B IP: Preimmune Sin1 tional site in the PF1 region C-terminal to the core kinase domain by WNK1. Phosphorylation of S325 or its mutation to glutamate wt SA 323 wt SA 323 modestly increased OSR1 activity (6). However, most dramatic Kinase activity increase comes from phosphorylation of the activation loop assay by WNK1. Additional phosphorylations have been reported to further increase activity including phosphorylation of S339 (37). Coomassie Additionally, OSR1 S339 and the comparable site in SPAK (as well as several additional sites in the PF1 region) have been identified by MS in several proteomic studies (www.phosphositeplus.com) (35, 42). Here we identify a major S339 kinase as mTORC2, and we confirm that mutation of S339 to alanine reduces OSR1 activity. Phosphorylation by mTORC2 on the OSR1 PF1 segment may

facilitate the interaction of PF1 with the kinase core to enhance the BIOCHEMISTRY active state of the enzyme. The structural impact of dual PF1 phosphorylation by WNK1 and mTORC2 may resemble the im- pact of Akt hydrophobic and turn motif phosphorylation by mTORC2, which remains to be determined. Calcium binding protein 39 (CAB39; originally known as MO25) together with the Ste20-related adaptor STRAD, an OSR1/SPAK pseudo kinase homolog, activate the protein kinase C WNK1 PI3K LKB1 (43, 44). OSR1 and SPAK have also been shown to bind and be activated by CAB39. CAB39 is thought to bind the region Akt of OSR1 and SPAK that contains the mTORC2 phosphorylation site and increase their activity independently of WNK1. Modi- fi OSR1 mTORC2 cation of this site may recruit CAB39 or other regulators to OSR1 and SPAK. It might also be involved in stabilizing the protein. Another Ste20 kinase was recently shown to auto- NKCC phosphorylate on a comparable C-terminal site (45). The phos- phorylation itself was not thought to directly change activity, but Fig. 4. mTORC2 phosphorylates OSR1 on S339. (A) Recombinant OSR1K46R was proposed to be involved in docking CAB39 (44). Thus, the was phosphorylated using Sin1 immunoprecipitates (IP Sin1) with or without PF1 is a regulatory nexus in these kinases. recombinant WNK1-194-483 (Inset) and samples were analyzed using MS. The turn motif and hydrophobic motif sites of Akt are phos- Representative spectrum and sequence of OSR1 shown identifying the residue phosphorylated specifically by mTORC2, S339. (B) Recombinant GST-tagged phorylated by mTORC2 with distinct kinetics. The turn motif is OSR1 wild type (wt), S339A (SA), and 1-323 (323) were phosphorylated with phosphorylated cotranslationally, whereas the hydrophobic motif Sin1 immunoprecipitates (IP) as the kinase. Incorporated radioactivity was is phosphorylated in response to insulin and growth factors (28, measured and plotted relative to the intensity of the substrate band; error 46). The fact that KU0063794 rapidly interfered with activation bars show standard deviation. (C) Model suggesting interconnections of the of OSR1 suggests that phosphorylation of S339 by mTORC2 is WNK-OSR1-NKCC pathway with the PI3K-mTORC2-Akt pathway. not cotranslational but instead occurs acutely in response to the osmotic stimulus, and may underlie regulation of OSR1 by in- sulin (34). A second not mutually exclusive possibility is that this that mTORC2 regulates OSR1 activation independently of the site is phosphorylated cotranslationally to stabilize OSR1 before effects of WNK1. The intersection of the WNK-OSR1/SPAK association with partners, yet turns over rapidly as a consequence pathway with the mTORC2 pathway suggests that mTORC2, which of its localization with NKCC. In addition to binding OSR1, the is active in a normal growing cell with ample nutrients, is also im- OSR1-activated cotransporter NKCC binds the phosphatase portant to muster ubiquitous responses to osmotic change mediated protein phosphatase 1, which contributes to rapid changes in by WNK protein kinases (Fig. 4C). OSR1/SPAK appear to be co- NKCC activity (47, 48). This phosphatase may also dephosphory- incidence detectors, requiring a real-time, regulated input from late OSR1 and SPAK when complexed to NKCCs. Active OSR1 mTORC2 to enable optimum WNK-dependent homeostasis. Our and SPAK associated with NKCC substrate may require rephos- findings add two members of the Ste20 branch of the kinome to the phorylation in situ to retain activity. list of mTORC2 targets apparently regulated in this manner. We found that OSR1 was less sensitive to mTOR inhibition In contrast to the p21-activated kinase-like Ste20 kinases, in with longer times of drug treatment. The reduced capacity of OSR1 and SPAK the kinase domain lies near the N terminus of the inhibitors to block OSR1 phosphorylation with longer times of proteins and OSR1 and SPAK contain the additional C-terminal treatment most likely accounts for the fact that this site was not

Sengupta et al. PNAS | November 19, 2013 | vol. 110 | no. 47 | 18829 Downloaded by guest on September 28, 2021 attributed to mTOR previously (35). It seems likely that one or siRNA. HeLa cells were detached from 10-cm dishes with trypsin and imme- more other kinases can also phosphorylate the PF1 site on OSR1 diately transfected with 20 nM Sin1 dsRNA oligonucleotides (hSin1), or 2.5 nM following prolonged mTOR inhibition. Enzymes other than small interfering WNK1 dsRNA oligonucleotides [sense, cagacagugcaguauu- mTORC2 phosphorylate the Akt hydrophobic motif under cer- cacTT; antisense, gugaauacugcacugucugTT; (Ambion) using Lipofectamine tain circumstances; these include Akt itself, the DNA-dependent RNAiMax (Invitrogen)] according to manufacturer’s protocols. protein kinase (DNA-PK), and the Tank binding kinase (30, 49– μ 51). DNA-PK and casein kinase II are predicted to phosphory- Kinase Assays. For immune complex kinase assays, 20 L immunoprecipitate μ – γ 32 late OSR1 S339 (http://scansite.mit.edu). Akt, like the Ste20 on beads were incubated with 50 M ATP (5,000 13,000 dpm/pmol [ - P]), kinase noted above, can autophosphorylate its own hydrophobic 10 mM MgCl2, and 20 mM Hepes pH 7.4 for 30 min at room temperature. Proteins were resolved as above. Gels were dried and exposed to film for motif site; this mechanism is enhanced by mutation of the pleck- autoradiography. Incorporation of radioactivity was quantified by scintilla- strin homology domain or prolonged inhibition of mTORC1 (30, tion counting of the bands excised from the gel. To obtain a 0.1-mol/mol 49, 52). Autophosphorylation is an intriguing possibility because incorporation of phosphate into substrate, 2 μM OSR1 K46R were used as a OSR1 can exist as an activation loop-swapped dimer (53). substrate with 500 μM ATP and the reactions were incubated for 4 h at 30 °C. WNKs are interconnected with PI3K-dependent pathways. Akt phosphorylates WNK1 on a threonine residue, T60 (T58 in NKCC Assay. Knockdown proceeded 48 h following transfection and then cells

rat), preceding the kinase domain and the comparable residue in were washed 2× in 140 mM NaCl, 5 mM KCl, 1 mM MgCl2, 1 mM CaCl2,10 WNK4 (24, 36, 54, 55). Phosphorylation of this site has little mM Hepes pH 7.4, 10 mM glucose, 10 mM Na pyruvate, and 0.1% BSA. After effect on WNK1 kinase activity. However, mutation of T60 in- 30 min, cells were incubated for 5 min with 107 cpm/mL 86Rb and 0.5 mM hibits SGK activation by WNKs, suggesting that phosphorylation ouabain with or without 10 μM bumetanide. After washing 2× in cold 100

of T60 is a permissive event that exerts positive feedback on mM MgCl2 and 10 mM Hepes pH 7.4 buffered with Tris, cells were lysed in some WNK1 functions (24, 36). Phosphorylation of T60 is re- 2% SDS. 86Rb uptake was measured in 100 μL lysates by liquid scintillation quired for maximal activation of ENaC by WNK1, consistent counting. Lysate protein concentration was determined using Pierce MicroBCA with the effect on SGK1 activation (36); regulation of the renal Protein Assay Kit. Uptake was an average of triplicates. NKCC activity outer medullary potassium channel by WNK1 also depends on (bumetanide-sensitive uptake) was taken as the difference with and without phosphorylation of T60 (56). WNK1 activity is essential for bumetanide. Effects of treatments are expressed relative to the control. phospholipase C beta PLCβ function downstream of Gq-coupled receptors, which may also involve PI3K (57). The PI3K pathway Statistical Analysis. Comparison between two groups was made using the two- has recently been implicated in WNK-OSR1/SPAK-NCC path- tailed unpaired t test. One-way ANOVA followed by Tukey’s multiple compar- way in animal studies (58). With the finding that activation of ison test were performed as indicated. P values are indicated in figure legends. OSR1/SPAK by WNK1 is enhanced by mTORC2, we have demonstrated the mechanism underlying another WNK1-regu- Proteomics and MS. Protein bands from polyacrylamide gels in SDS were lated cascade that requires a coincident signal from PI3K. We excised and digested overnight with trypsin (Promega) after reduction and alkylation with DTT and iodoacetamide (Sigma-Aldrich). The resulting sam- conclude that cell homeostasis requires the multilevel integration ples were analyzed by tandem MS using a Q Exactive mass-spectrometer of WNK osmosensing and PI3K survival pathways. (Thermo Electron) coupled to an Ultimate 3000 RSLC-Nano liquid chroma- μ Experimental Procedures tography system (Dionex). Peptides were loaded onto a 180 m inner di- ameter, 10 cm self-packed column containing 3 μm C18 resin (Dr. Maisch), Materials. Human Sin1 siRNA oligonucleotides were from Ambion (hSin1, and eluted with a gradient of 1–41% buffer B in 40 min. Buffer A contained sense, gauuagaacgacuccgaaaTT, antisense, uuucggagucguucuaaucTT). Rapa- 0.1% formic acid in water and buffer B, 0.1% formic acid in acetonitrile. The mycin and wortmannin were from LC Laboratories, KU0063794 from Tocris + + Q Exactive instrument acquired up to 10 high-energy collision-induced biosciences, and ouabain (Na /K inhibitor) and bumetanide (NKCC inhibitor) from Sigma. Antibodies were from the following: Sin1 (NovusB biological), dissociation fragment spectra for each full spectrum acquired. mTOR (Santa Cruz Biotechnology, immunoblotting) and Cell Signaling (Santa Raw MS data were converted to peak list format using ProteoWizard Cruz Biotechnology, immunoprecipitation), and Akt and pS473 Akt (Cell msconvert (version 3.0.3535) (61). The resulting files were analyzed using the Signaling). Anti-WNK1 (Q256), anti-OSR1 (U5567) and plasmids encoding Central Proteomics Facilities Pipeline (CPFP) (62). Peptide identification was WNKs and NKCC2 were as described (59). 86Rb and [γ-32P]ATP were from performed using the X!Tandem (63) and Open MS search algorithm (64) Perkin-Elmer. Recombinant proteins used for kinase assays were purified search engines against the UniProtKB human whole proteome sequence from Escherichia coli strain BL21 using standard protocols. database (release 2012_02) (65), with reversed decoy sequences appended (66). Fragment and precursor tolerances of 20 ppm and 0.5 Da were speci- Cell Culture and Transfection. HeLa cells were grown in DMEM supplemented fied and two missed cleavages were allowed. Carbamidomethylation of with 10% (vol/vol) FBS (Sigma) and 2 mM glutamine. Cells were harvested in 50 cysteine was specified as a fixed modification. Oxidation of methionine and mM Hepes, pH 7.7, 150 mM NaCl, 1.5 mM MgCl2, 1 mM EGTA, 10% (vol/vol) phosphorylation (Ser/Thr/Tyr) were specified as variable modifications. Val- β glycerol, 100 mM NaF, 0.2 mM NaVO4,50mM -glycerophosphate, 0.1% Triton idation and combination of results was performed within CPFP using Trans- X-100, 0.1 μM phenylmethyl sulfonyl fluoride, 10 mg/L each N-α-p-tosyl-L-lysine Proteomic Pipeline tools (67). Identifications were filtered to <1% false chloromethyl ester, N-α-p-tosyl-L-arginine methyl ester, N-α-p-tosyl-L-lysine discovery rate. Initial assessment of localization ambiguity for phos- chloromethyl ketone, leupeptin, and pepstatin A as described (60). phorylation assigned by the search engines was performed using the modification localization score tool within CPFP, based on the post trans- Immunoprecipitation and Immunoblotting. Proteins were immunoprecipitated lational modification score method (68, 69). overnight at 4 °C from 750 μg of soluble lysate protein with 5 μL each an- tibody. Antibodies were collected with 30 μL of a 50% protein A-Sepharose (GE Healthcare) slurry following a 2-h incubation at 4 °C. Beads were washed ACKNOWLEDGMENTS. We thank members of the M.H.C. laboratory for suggestions about this work and Dionne Ware for administrative assistance. 3× in lysis solution and proteins were eluted using 5X electrophoresis sample This work was supported by National Institutes of Health (NIH) Grant R01 buffer. Proteins in cell lysates or immunoprecipitates were resolved by SDS GM53032 and Robert A. Welch Foundation Grant I1243 (to M.H.C.). H.M. and polyacrylamide gel electrophoresis and transferred to nitrocellulose. D.C.T. were supported by Cancer Prevention and Research Institute of Texas Immunoblots were developed using a Li-COR Odyssey infrared imaging Grants RP120613 and R1121 (to H.M.), and A.L.-R. was supported by NIH system. Training Grant 2T32 CA124334-06.

1. Gagnon KB, Delpire E (2013) Physiology of SLC12 transporters: Lessons from inherited 3. Leiserson WM, Harkins EW, Keshishian H (2000) Fray, a Drosophila serine/threonine human genetic mutations and genetically engineered mouse knockouts. Am J Physiol kinase homologous to mammalian PASK, is required for axonal ensheathment. Cell Physiol 304(8):C693–C714. Neuron 28(3):793–806. 2. Hebert SC, Mount DB, Gamba G (2004) Molecular physiology of cation-coupled Cl- 4. Chen W, Yazicioglu M, Cobb MH (2004) Characterization of OSR1, a member of the cotransport: The SLC12 family. Pflugers Arch 447(5):580–593. mammalian Ste20p/germinal center kinase subfamily. JBiolChem279(12):11129–11136.

18830 | www.pnas.org/cgi/doi/10.1073/pnas.1318676110 Sengupta et al. Downloaded by guest on September 28, 2021 5. Piechotta K, Lu J, Delpire E (2002) Cation chloride cotransporters interact with the 38. Knighton DR, et al. (1991) Crystal structure of the catalytic subunit of cyclic adenosine stress-related kinases Ste20-related proline-alanine-rich kinase (SPAK) and oxidative monophosphate-dependent protein kinase. Science 253(5018):407–414. stress response 1 (OSR1). J Biol Chem 277(52):50812–50819. 39. Zhang J, Zhang F, Ebert D, Cobb MH, Goldsmith EJ (1995) Activity of the MAP kinase 6. Vitari AC, Deak M, Morrice NA, Alessi DR (2005) The WNK1 and WNK4 protein kinases ERK2 is controlled by a flexible surface loop. Structure 3(3):299–307. that are mutated in Gordon’s hypertension syndrome phosphorylate and activate 40. Frödin M, et al. (2002) A phosphoserine/threonine-binding pocket in AGC kinases and SPAK and OSR1 protein kinases. Biochem J 391(Pt 1):17–24. PDK1 mediates activation by hydrophobic motif phosphorylation. EMBO J 21(20): 7. Moriguchi T, et al. (2005) WNK1 regulates phosphorylation of cation-chloride-coupled 5396–5407. cotransporters via the STE20-related kinases, SPAK and OSR1. J Biol Chem 280(52): 41. Newton AC (2003) Regulation of the ABC kinases by phosphorylation: Protein kinase 42685–42693. C as a paradigm. Biochem J 370(Pt 2):361–371. 8. Gagnon KB, England R, Delpire E (2006) Volume sensitivity of cation-chloride co- 42. Hornbeck PV, et al. (2012) PhosphoSitePlus: A comprehensive resource for in- transporters is modulated by the interaction of two kinases: SPAK and WNK4. Am J vestigating the structure and function of experimentally determined post-trans- Physiol Cell Physiol 290:C134–C142. lational modifications in man and mouse. Nucleic Acids Res 40(Database issue): 9. Anselmo AN, et al. (2006) WNK1 and OSR1 regulate the Na+,K+, 2Cl- cotransporter in D261–D270. HeLa cells. Proc Natl Acad Sci USA 103(29):10883–10888. 43. Boudeau J, et al. (2003) MO25alpha/beta interact with STRADalpha/beta enhancing 10. Xu B, et al. (2000) WNK1, a novel mammalian serine/threonine protein kinase lacking their ability to bind, activate and localize LKB1 in the cytoplasm. EMBO J 22(19): – the catalytic lysine in subdomain II. J Biol Chem 275(22):16795–16801. 5102 5114. 11. Veríssimo F, Jordan P (2001) WNK kinases, a novel protein kinase subfamily in multi- 44. Filippi BM, et al. (2011) MO25 is a master regulator of SPAK/OSR1 and MST3/MST4/ – cellular organisms. Oncogene 20(39):5562–5569. YSK1 protein kinases. EMBO J 30(9):1730 1741. 12. Wilson FH, et al. (2001) Human hypertension caused by mutations in WNK kinases. 45. Fuller SJ, et al. (2012) A novel non-canonical mechanism of regulation of MST3 – Science 293(5532):1107–1112. (mammalian Sterile20-related kinase 3). Biochem J 442(3):595 610. 13. Xie J, Craig L, Cobb MH, Huang C-L (2006) Role of WNKs in the pathogenesis of 46. Oh WJ, et al. (2010) mTORC2 can associate with ribosomes to promote cotranslational – Gordon’s Syndrome. Pediatr Nephrol 21:1231–1236. phosphorylation and stability of nascent Akt polypeptide. EMBO J 29(23):3939 3951. 14. Zambrowicz BP, et al. (2003) Wnk1 kinase deficiency lowers blood pressure in mice: A 47. Darman RB, Flemmer A, Forbush B (2001) Modulation of ion transport by direct gene-trap screen to identify potential targets for therapeutic intervention. Proc Natl targeting of protein phosphatase type 1 to the Na-K-Cl cotransporter. J Biol Chem – Acad Sci USA 100(24):14109–14114. 276(37):34359 34362. 15. McCormick JA, Ellison DH (2011) The WNKs: Atypical protein kinases with pleiotropic 48. Gagnon KB, Delpire E (2010) Multiple pathways for protein phosphatase 1 (PP1) actions. Physiol Rev 91(1):177–219. regulation of Na-K-2Cl cotransporter (NKCC1) function: The N-terminal tail of the Na- 16. Lalioti MD, et al. (2006) Wnk4 controls blood pressure and potassium homeostasis via K-2Cl cotransporter serves as a regulatory scaffold for Ste20-related proline/alanine- – regulation of mass and activity of the distal convoluted tubule. Nat Genet 38(10): rich kinase (SPAK) AND PP1. J Biol Chem 285(19):14115 14121. 49. Warfel NA, Niederst M, Newton AC (2011) Disruption of the interface between the 1124–1132. pleckstrin homology (PH) and kinase domains of Akt protein is sufficient for hydro- 17. Kahle KT, Wilson FH, Lifton RP (2005) Regulation of diverse ion transport pathways by phobic motif site phosphorylation in the absence of mTORC2. J Biol Chem 286(45): WNK4 kinase: A novel molecular switch. Trends Endocrinol Metab 16(3):98–103. 39122–39129. 18. Liu Z, et al. (2011) Downregulation of NCC and NKCC2 cotransporters by kidney- 50. Feng J, Park J, Cron P, Hess D, Hemmings BA (2004) Identification of a PKB/Akt hy- specific WNK1 revealed by gene disruption and transgenic mouse models. Hum Mol drophobic motif Ser-473 kinase as DNA-dependent protein kinase. J Biol Chem BIOCHEMISTRY Genet 20(5):855–866. 279(39):41189–41196. 19. Xie J, et al. (2009) Endothelial-specific expression of WNK1 kinase is essential for 51. Ou YH, et al. (2011) TBK1 directly engages Akt/PKB survival signaling to support angiogenesis and heart development in mice. Am J Pathol 175(3):1315–1327. oncogenic transformation. Mol Cell 41(4):458–470. 20. Dowd BF, Forbush B (2003) PASK (proline-alanine-rich STE20-related kinase), a regu- 52. Rodrik-Outmezguine VS, et al. (2011) mTOR kinase inhibition causes feedback- latory kinase of the Na-K-Cl cotransporter (NKCC1). J Biol Chem 278(30):27347–27353. dependent biphasic regulation of AKT signaling. Cancer Discov 1(3):248–259. 21. Welling PA, Chang YP, Delpire E, Wade JB (2010) Multigene kinase network, kidney 53. Lee SJ, Cobb MH, Goldsmith EJ (2009) Crystal structure of domain-swapped STE20 transport, and salt in essential hypertension. Kidney Int 77(12):1063–1069. OSR1 kinase domain. Protein Sci 18(2):304–313. 22. Gagnon KB, Delpire E (2012) Molecular physiology of SPAK and OSR1: Two Ste20- 54. Vitari AC, et al. (2004) WNK1, the kinase mutated in an inherited high-blood-pressure related protein kinases regulating ion transport. Physiol Rev 92(4):1577–1617. syndrome, is a novel PKB (protein kinase B)/Akt substrate. Biochem J 378(Pt 1): 23. Xu BE, et al. (2005) WNK1 activates SGK1 to regulate the epithelial sodium channel. 257–268. Proc Natl Acad Sci USA 102(29):10315–10320. 55. Jiang ZY, et al. (2005) Identification of WNK1 as a substrate of Akt/protein kinase B 24. Heise CJ, et al. (2010) Serum and glucocorticoid-induced kinase (SGK) 1 and the epi- and a negative regulator of insulin-stimulated mitogenesis in 3T3-L1 cells. J Biol Chem thelial sodium channel are regulated by multiple with no lysine (WNK) family 280(22):21622–21628. members. J Biol Chem 285(33):25161–25167. 56. Cheng CJ, Huang CL (2011) Activation of PI3-kinase stimulates endocytosis of ROMK 25. Park J, et al. (1999) Serum and glucocorticoid-inducible kinase (SGK) is a target of the via Akt1/SGK1-dependent phosphorylation of WNK1. J Am Soc Nephrol 22(3): – PI 3-kinase-stimulated signaling pathway. EMBO J 18(11):3024 3033. 460–471. 26. Frias MA, et al. (2006) mSin1 is necessary for Akt/PKB phosphorylation, and its iso- 57. An SW, et al. (2011) WNK1 promotes PIP synthesis to coordinate growth factor and fi – 2 forms de ne three distinct mTORC2s. Curr Biol 16(18):1865 1870. GPCR-Gq signaling. Curr Biol 21(23):1979–1987. 27. García-Martínez JM, Alessi DR (2008) mTOR complex 2 (mTORC2) controls hydro- 58. Nishida H, et al. (2012) Phosphatidylinositol 3-kinase/Akt signaling pathway activates phobic motif phosphorylation and activation of serum- and glucocorticoid-induced the WNK-OSR1/SPAK-NCC phosphorylation cascade in hyperinsulinemic db/db mice. – protein kinase 1 (SGK1). Biochem J 416(3):375 385. Hypertension 60(4):981–990. 28. Facchinetti V, et al. (2008) The mammalian target of rapamycin complex 2 controls 59. Sengupta S, et al. (2012) Interactions with WNK (with no lysine) family members – folding and stability of Akt and protein kinase C. EMBO J 27(14):1932 1943. regulate oxidative stress response 1 and ion co-transporter activity. J Biol Chem 29. Pearce LR, Sommer EM, Sakamoto K, Wullschleger S, Alessi DR (2011) Protor-1 is re- 287(45):37868–37879. fi quired for ef cient mTORC2-mediated activation of SGK1 in the kidney. Biochem J 60. Barylko B, et al. (2002) Analysis of the catalytic domain of phosphatidylinositol 4-kinase – 436(1):169 179. type II. JBiolChem277(46):44366–44375. 30. Toker A, Newton AC (2000) Akt/protein kinase B is regulated by autophosphorylation 61. Kessner D, Chambers M, Burke R, Agus D, Mallick P (2008) ProteoWizard: Open source – at the hypothetical PDK-2 site. J Biol Chem 275(12):8271 8274. software for rapid proteomics tools development. Bioinformatics 24(21):2534–2536. 31. Lu M, et al. (2010) mTOR complex-2 activates ENaC by phosphorylating SGK1. JAm 62. Trudgian DC, et al. (2010) CPFP: A central proteomics facilities pipeline. Bioinformatics Soc Nephrol 21(5):811–818. 26(8):1131–1132. 32. Hakim S, Bertucci MC, Conduit SE, Vuong DL, Mitchell CA (2012) Inositol poly- 63. Craig R, Beavis RC (2004) TANDEM: Matching proteins with tandem mass spectra. phosphate phosphatases in human disease. Curr Top Microbiol Immunol 362:247–314. Bioinformatics 20(9):1466–1467. 33. Hong F, et al. (2008) mTOR-raptor binds and activates SGK1 to regulate p27 phos- 64. Geer LY, et al. (2004) Open mass spectrometry search algorithm. J Proteome Res 3(5): phorylation. Mol Cell 30(6):701–711. 958–964. 34. Sohara E, et al. (2011) Acute insulin stimulation induces phosphorylation of the Na-Cl 65. Magrane M, Consortium U (2011) UniProt Knowledgebase: A hub of integrated cotransporter in cultured distal mpkDCT cells and mouse kidney. PLoS ONE 6(8): protein data. Database (Oxford) 2011:bar009. e24277. 66. Elias JE, Gygi SP (2007) Target-decoy search strategy for increased confidence in large- 35. Yu Y, et al. (2011) Phosphoproteomic analysis identifies Grb10 as an mTORC1 sub- scale protein identifications by mass spectrometry. Nat Methods 4(3):207–214. strate that negatively regulates insulin signaling. Science 332(6035):1322–1326. 67. Keller A, et al. (2005) A uniform proteomics MS/MS analysis platform utilizing open 36. Xu BE, Stippec S, Lazrak A, Huang CL, Cobb MH (2005) WNK1 activates SGK1 by XML file formats. Mol Syst Biol, 1:2005.0017. a phosphatidylinositol 3-kinase-dependent and non-catalytic mechanism. J Biol Chem 68. Olsen JV, Mann M (2004) Improved peptide identification in proteomics by two 280(40):34218–34223. consecutive stages of mass spectrometric fragmentation. Proc Natl Acad Sci USA 37. Gagnon KB, Delpire E (2010) On the substrate recognition and negative regulation of 101(37):13417–13422. SPAK, a kinase modulating Na+-K+-2Cl- cotransport activity. Am J Physiol Cell Physiol 69. Olsen JV, et al. (2010) Quantitative phosphoproteomics reveals widespread full 299(3):C614–C620. phosphorylation site occupancy during mitosis. Sci Signal 3(104):ra3.

Sengupta et al. PNAS | November 19, 2013 | vol. 110 | no. 47 | 18831 Downloaded by guest on September 28, 2021