Staurosporine and NEM mainly impair WNK-SPAK/OSR1 mediated phosphorylation of KCC2 and NKCC1.

The pivotal role of KCC2 and NKCC1 in development and maintenance of fast inhibitory neurotransmission and their implication in severe human diseases arouse interest in posttranscriptional regulatory mechanisms such as (de)phosphorylation. Staurosporine (broad kinase inhibitor) and N-ethylmalemide (NEM) that modulate kinase and phosphatase activities enhance KCC2 and decrease NKCC1 activity. Here, we investigated the regulatory mechanism for this reciprocal regulation by mass spectrometry and immunoblot analyses using phospho-specific antibodies. Our analyses revealed that application of staurosporine or NEM dephosphorylates Thr1007 of KCC2, and Thr203, Thr207 and Thr212 of NKCC1. Dephosphorylation of Thr1007 of KCC2, and Thr207 and Thr212 of NKCC1 were previously demonstrated to activate KCC2 and to inactivate NKCC1. In addition, application of the two agents resulted in dephosphorylation of the T-loop and S-loop phosphorylation sites Thr233 and Ser373 of SPAK, a critical kinase in the WNK-SPAK/OSR1 signaling module mediating phosphorylation of KCC2 and NKCC1. Taken together, these results suggest that reciprocal regulation of KCC2 and NKCC1 via staurosporine and NEM is based on WNK-SPAK/OSR1 signaling. The key regulatory phospho-site Ser940 of KCC2 is not critically involved in the enhanced activation of KCC2 upon staurosporine and NEM treatment, as both agents have opposite effects on its phosphorylation status. Finally, NEM acts in a tissue-specific manner on Ser940, as shown by comparative analysis in HEK293 cells and immature cultured hippocampal neurons. In summary, our analyses identified phospho-sites that are responsive to staurosporine or NEM application. This provides important information towards a better understanding of the cooperative interactions of different phospho-sites.

In neurons, KCC2 and NKCC1 are involved in the development and maintenance of inhibitory neurotransmission. For instance, KCC2 generates a low [Cl -] i required for fast inhibitory neurotransmission. Binding of γ-aminobutyric acid (GABA) or glycine to their receptors, which are ligand-gated Clchannels, leads to a Clinflux and therefore hyperpolarization [8,12]. By contrast, NKCC1 is more prevalent in immature inhibitory neurons. There, its action results in a high [Cl -] i causing GABA or glycine to elicit a depolarizing action that opens L-type voltage-gated Ca 2+ channels necessary for proper synapse formation [13][14][15][16][17][18].
Initially, both staurosporine and NEM were thought to act through a similar mechanism [47], but recent findings revealed that they act differentially on specific KCC2 phospho-sites [40]. Furthermore, staurosporine and NEM mediated effects involve both KCC2 phosphorylation and dephosphorylation [40,48]. To gain insight into their mode of action, we analyzed the impact of these compounds on phosphorylation of specific KCC2 and NKCC1 phosphosites using large-scale phosphoproteomics studies and phospho-site specific antibodies in stably transfected HEK293 cells and immature primary cultures of hippocampal neurons.

Cell culturing of HEK293 cells
For K + -Clcotransporter activity measurements, stably transfected rat KCC2b HEK293 cells (HEK rnKCC2b ) [55] were plated in a 0.1 mg/ml poly-L-lysine coated black-well 96 well culture dish (Greiner Bio-One, Frickenhausen, Germany) at a concentration of 100,000 cells/well. The remaining cells were plated on a 0.1 mg/ml poly-L-lysine-coated class coverslip and~18 h later analyzed by immunocytochemistry.

Primary cultures of rat hippocampal neuron
All animal procedures were carried out in accordance with the European Union Directive of 22 September (2010/63/EU). The protocol was approved by the INSERM Local committee (Number 0287.01, delivered by Ministère de l'Education et de la Recherche). Hippocampi from 18-day-old rat embryos were dissected and dissociated using trypsin (0.05%) and plated in 60-mm culture dishes at a density of 100,000 cells cm −2 in minimal essential medium (MEM) supplemented with 10% NU serum (BD Biosciences, Le Pont de Claix, France), 10 mM glucose, 1 mM sodium pyruvate, 2 mM glutamine, and 100 U ml −1 penicillin-streptomycin as previously described [78]. On day 7 of culture incubation, half of the medium was changed to MEM with 2% B27 supplement (Invitrogen). 24 h prior to plating, dishes were coated with poly-ethylenimine (5 μg/ml).
To determine the dose-response profile, increasing concentration of staurosporine (8-80 μM) and NEM (25-3000 μM) were applied to the preincubation buffer 15 min prior flux measurement. This was done for three technical replicates and at least five independent measurements were performed.

Statistical analyses
The majority of data populations illustrated in Figs 1, 3 and 4 showed non-normal distributions (verified using Shapiro-Wilk normality test at 0.05 significance level). Therefore, a nonparametric test (Wilcoxon-Mann-Whitney rank sum test) was employed for the comparison between different groups of data. The resulting p-values were adjusted with the Bonferroni correction for multiple testing. Note that only p-values < 0.01 were considered to reduce the chances of false positives (type I errors). The sample size n always refers to the number of biological replicates (independent preparations). The activity of each independent preparation was determined as mean over three technical replicates. The average scatter within technical replicates was 3 times smaller than the scatter across biological replicates for a given treatment.
For dose-response analyses, we used the nls function from the stats package in R (version 3.5.1) to model the data points with the Hill-Langmuir equation, where y is the Tl + uptake of HEK rnKCC2b cells, y 0 is the Tl + independent-baseline activity (y 0 = 100%), y sat is the maximum activity (equivalent to V max in reaction rate equation) relative to baseline, x is the concentration of the agonist, n is the Hill coefficient, and ED 50 is the agonist concentration that produces 50% of the saturation response.

Treatment of stably transfected HEK rnKCC2b cells and hippocampal neurons
For mass spectrometry and immunoblot analyses, stably transfected HEK rnKCC2b cells [55] were grown in 75 cm 2 cell-culture flask. Cells were washed with 5 ml flux hypotonic preincubation buffer (100 mM N-methyl-D-glucamine-chloride, 5 mM HEPES, 5 mM KCl, 2 mM CaCl 2 , 0.8 mM MgSO 4 , 5 mM glucose, pH 7.4; 160 mmol/kg±2,04 mmol/kg) and then treated with or without 8 μM staurosporine or 1 mM NEM. Cells were centrifuged at 500 rpm for 3 min and the resulting pellets used for immunoblot analyses or mass spectrometry analyses. For immunoblot analyses of immature (9 days in vitro, (DIV)) rat hippocampal neurons, half of the media from culture dishes containing neurons (2.5 ml from each dish) was collected to prepare samples including 16 μM staurosporine (from 10 mM stock solution in DMSO), 1 mM NEM (by direct dissolving of the NEM powder) or DMSO (vehicle, same volume as for staurosporine), respectively. The samples were distributed dropwise into the culture dishes that gave final concentration of 8 μM staurosporine or 0.5 mM NEM and cultures were incubated 15 min at 37˚C (5% CO 2 ). After incubation, neurons were rinsed twice with Hanks' Balanced Salt solution (HBSS, ThermoFisher Scientific) precooled to 4˚C. The rinsage solution was replaced with ice-cold HBSS containing a cocktail of protease and phosphatase inhibitors (ThermoFisher Scientific) and neurons rapidly scraped and centrifuged at 4˚C (10,000 g, 3 min). Pellets were frozen in liquid nitrogen and kept for future analyses at -80˚C.

Multi-Enzyme Digestion Filter Aided Sample preparation (MED FAS)
For mass spectrometry analyses, cell pellets of untreated, staurosporine or NEM treated HEK rnKCC2b cells were lysed in 2% SDS in 0.1 M Tris-HCl, pH 8.0, containing 0.1 M DTT as described previously [81]. Total protein was determined using WF-assay in micro-titer plate format [82]. Aliquots of the cell lysate containing 1-2 mg were processed in Amicon Ultra 15 Ultracel 30k (Merck Millipore, Darmstadt) devices as described previously [83] with several modifications using the MED FASP method [84]. Briefly, SDS and other low molecular weight material were removed by centrifugation at 4,000 x g in 5 consecutive washes with UA buffer containing 8 M urea (ultrapure, Merck, Darmstadt) in 0.1 M Tris pH 8. Following two washes with 5 ml of 0.1M Tris-HCl, pH 8 (DB buffer), 20 μg of LysC (Wako, Neuss) in 0.5 mL of DB was added to the filter. Samples were digested overnight at 37˚C and peptides were collected by centrifugation. Next, the material retained in the filter was cleaved with 10 μg trypsin in 0.5 ml DB at 37˚C for 4 h and the peptides were eluted as previously. To increase the yield of peptides, filters were washed twice with 0.5 mL DB. Concentration of peptides was determined by WF-assay [82].

TiO 2 -based enrichment of phosphopeptides
Phosphopeptides were enriched using TiO 2 -beads [85] with several modifications [83]. Briefly, 25 mg of 'Titansphere TiO 2 10 μm' (GL Sciences, Inc., Japan) were suspended in 50 μl of 3% (m/v) dihydroxybenzoic acid in 80% (v/v) CH 3 CN, 0.1% CF 3 COOH and diluted 1:4 with water before use. Ten microliters of this slurry (1 mg beads) were added and samples incubated under continuous agitation for 20 min. The mass ratio of the beads and peptides was 3:1. Then, the titanium beads were sedimented by centrifugation at 5,000 x g for 1 min and the supernatants were collected and mixed with another portion of the beads and incubated as above. The bead-pellets were resuspended in 150 μl of 30% (v/v) CH 3 CN containing 3% (v/v) CF 3 COOH and transferred to a 200 μl pipet tip plugged with one layer of glass microfiber filter GFA (Whatman). The beads were washed three times with 30% (v/v) CH 3 CN, 3% CF 3 COOH (v/v) solution and three times with 80% CH 3 CN (v/v), 0.3% CF 3 COOH (v/v) solution. Finally, the peptides were eluted from the beads with 100 μL of 40% CH 3 CN (v/v) containing 15% NH 4 OH (m/v) and were vacuum-concentrated to *4 μl.

Mass-spectrometric analysis
Analysis of peptide mixtures was conducted using a QExactive HF-X mass spectrometer (Thermo-Fisher Scientific, Palo Alto). Aliquots containing <1 μg of total peptide were chromatographed on a 50 cm column with 75 μm inner diameter packed C 18 material. Peptide separation occurred at 300 nl/min for 95 min using an acetonitrile gradient of 5-30%. The temperature of the column oven was 55˚C. The mass spectrometer operated in data-dependent mode with survey scans acquired at a resolution of 60,000. Up to the top 15 most abundant isotope patterns with charge � +2 from the survey scan (300-1650 m/z) were selected with an isolation window of 1.4 m/z and fragmented by HCD with normalized collision energies of 25. The maximum ion injection times for the survey scan and the MS/MS scans were 20 and 28 ms, respectively. The ion target value for MS1 and MS2 scan modes was set to 3×10 6 and 10 5 , respectively. The dynamic exclusion was 30 s. The MS spectra were searched using MaxQuant software. A maximum of two missed cleavages was allowed. The maximum false peptide and protein discovery rate was specified as 0.01. The whole mass spectrometry analyses were performed in three technical and two biological replicas per treatment.
For phospho-antibody immunoprecipitation, KCC isoforms were immunoprecipitated from indicated cell extracts. 2 mg of the indicated clarified cell extract was mixed with 15 μg of the indicated phospho-specific KCC antibody conjugated to 15 μl of protein-G-Sepharose, in the added presence of 20 μg of the dephosphorylated form of the phosphopeptide antigen, and incubated 2 hours at 4˚C with gentle shaking. Immunoprecipitates were washed three times with 1 ml of lysis buffer containing 0.15 M NaCl and twice with 1 ml of buffer A. Bound proteins were eluted with 1x LDS sample buffer.

Dose-response analyses of staurosporine and NEM in stably transfected HEK rnKCC2b cells
Staurosporine and NEM generally activate KCCs [40,47,86,87]. To closer characterize their mode of action, we used stably transfected HEK rnKCC2b cells, as high amount of the cotransporter is advantageous for subsequent biochemical analyses. Immunoreactivity of HEK rnKCC2b cells against KCC2 was detected in all cells with clear labeling at the cell membrane and cytosol ( Fig 1A). This is in agreement with previous cell surface expression analyses that detected 11.8 ± 1.4% of total KCC2 in stably transfected rnKCC2b at the cell surface [79].
Next, we determined the dose-response relationships of staurosporine and NEM on rnKCC2b transport activity by treating HEK rnKCC2b cells with different concentrations of staurosporine (5-80 μM) or NEM (25-3,000 μM). Fig 2 represents the dose-response curve for both agents. The dose-response curve for staurosporine (Fig 2A) was approximately a rectangular hyperbola (n � 1, not significantly different from unity). This reflects Michaelis-Menten kinetics and suggests absence of cooperative effects. In contrast, the dose response curve for NEM ( Fig 2B) had a pronounced sigmoidal shape (n � 5), which reflects cooperative binding kinetics. The effective dose ED 50 (representing the potency) for staurosporine of stably transfected HEK rnKCC2b was 12.8 ± 4.9 μM and the maximal efficacy (E max ) was 205 ± 40%. The ED 50 value for NEM was 0.5 ± 1.3 mM and E max was 105 ± 6%. For further analyses, we used a concentration of 8 μM staurosporine and 1 mM NEM, if not indicated otherwise as these concentrations significantly increase KCC2b transport activity.

Identification of CCC phosphorylation sites in stably transfected HEK rnKCC2b cells by mass spectrometry analyses upon treatment with staurosporine or NEM
Phosphoproteomics by mass spectrometry has the advantage of providing an unbiased survey of phospho-sites. Therefore, we here used for the first time this technique to gain insight on  Table 1. We first mapped KCC2b phosphosites in untreated HEK rnKCC2b cells. Twelve phospho-sites were identified: Ser 25 , Ser 26 , Ser 31 , Thr 32 , Thr 34 in the cytoplasmic N-terminus and Thr 906 , Ser 937 , Ser 940 , Thr 1007 , Ser 1022 , Ser 1025 , and Ser 1026 in the C-terminus (Table 1, untreated). These sites include all phospho-sites already present in PhosphositePlus (Table 2) [91]. Recently phosphoproteomic data deposited in Phosida and Phosphosite plus revealed that rat KCC2b tissue only harbors seven phosphosites (Table 2). Thus, we here report five phospho-sites (Ser 25 , Thr 32 , Thr 906 , Ser 937 and Thr 1007 ) that were so far only reported for mouse but not rat KCC2 tissue (Table 2). These sites most likely reflect different expression of kinases and phosphatases in different tissues (HEK293 vs. rat brain tissue) [69] or increased detection rate in stably transfected HEK rnKCC2b cells.
As KCC1, KCC4, and NKCC1 are endogenously expressed in HEK293 cells and as mass spectrometric analysis provides data on most proteins in a given sample, we also investigated phosphorylation sites of the following proteins: hsKCC1 (Thr 893 , analogous to rnKCC2b Thr 1007 ), hsKCC4 (Thr 926 and Thr 980 , analogous to rnKCC2b Thr 906 and Thr 1007 ) and hsNKCC1 (Thr 212 , Thr 217 , Ser 242 , Thr 266 , Thr 268 , Ser 940 , Tyr 956 , and Ser 957 ). These phosphosites were already previously deposited in PhosphositePlus and Phosida (S1-S3 Tables). Overall, we detected only a low proportion of all so far deposited phospho-sites for these three cotransporters. This might reflect low expression levels in HEK293 cells.
Next, we investigated the phosphorylation pattern of rnKCC2b upon staurosporine and NEM treatment. Phosphorylation at Ser 26 , Thr 32 , Thr 34 , Thr 906 , Ser 940 , Thr 1007 , Ser 1022 , Ser 1025 , Ser 1026 of rnKCC2b were still present upon treatment with staurosporine or NEM, whereas  (Table 1). Additionally, no phosphorylation was detected for Ser 31 after treatment with staurosporine.
Regarding endogenously expressed CCCs, the phospho-site Thr 983 of hsKCC1 (analogous to rnKCC2b Thr 1007 ), Thr 926 of hsKCC4 (analogous to rnKCC2b Thr 906 ) and Thr 212/217 and Thr 266/268 of hsNKCC1 could not be detected upon application of either of the two reagents. Additionally, the phospho-sites Thr 980 of hsKCC4 (analogous to rnKCC2b Thr 906 ) and Ser 242 of hsNKCC1 were not detected anymore after staurosporine treatment. The phospho-sites Thr 266 , Ser 940 and Tyr 956 /Ser 957 of hsNKCC1 were still present upon treatment with staurosporine and NEM.
Several kinases were described to directly phosphorylate KCC2 and NKCC1. This includes kinases of the WNK-SPAK/OSR1 and PKC mediated phosphorylation pathways. To gain further insight into the regulatory phosphorylation mechanism, we explored their phosphorylation pattern as well. We detected several phosphorylation sites in hsWNK1, hsWNK2, and hsSPAK [69,93,94] (Table 3). Upon all, we observed phosphorylation of the activating T-loop residue Ser 382 of hsWNK1 and the S-loop phosphorylation site of Ser 372/373 of hsSPAK that is phosphorylated by WNK1 [63,69,93]. Upon treatment with staurosporine or NEM, Ser 382 of hsWNK1 and Ser 372/373 of hsSPAK were not detected anymore (Table 4). We were not able to detect phosphorylation of Thr 233 that is located in the T-loop kinase domain of hsSPAK. Normally, this site is directly phosphorylated by WNK1 and WNK4 to activate SPAK [60, 63, 69, Table 1. Phospho-sites of stably transfected HEK rnKCC2b cells. Stably transfected HEK rnKCC2b were treated with or without 1 mM NEM or 8 μM staurosporine before they were analyzed by mass spectrometry. The protein accession numbers are: hsNKCC1 (P55011), hsKCC1 (Q9UP95), rnKCC2b (Q9H2X9-2), hsKCC4 (Q9Y666).
To summarize, according to mass spectrometry based phosphoproteome analyses, staurosporine and NEM reduce the number of detected phospho-sites of stably expressed rnKCC2b and endogenously expressed hsKCC1, hsKCC4, hsNKCC1, hsWNK1 and hsSPAK. Phosphorylation of some sites (rnKCC2b: Ser 31 , hsNKCC1:Ser 242 , hsKCC4: Thr 980 ) was absent only after staurosporine treatment. Yet, these results can only be used as an indication since the absence of phosphorylation sites can reflect detection problems caused by low phosphorylation rates.

Quantitative analyses of phospho-sites of rnKCC2b and hsNKCC1 upon staurosporine and NEM treatment in HEK rnKCC2b cells
The experimental setup of our mass spectrometry-based analysis precluded quantification of changes at individual phospho-sites. We therefore applied in a next step phospho-site-specific antibody, as they were previously shown to quantitatively monitor changes in KCC2, NKCC1 and SPAK phosphorylation [23, 48,59,66,70,102]. Currently, a limited number of this class  (Fig 3). These data corroborate the phosphoproteome analyses which revealed phosphorylation of Ser 940 , Thr 906 and Thr 1007 in KCC2b and Thr 212 /Thr 217 in NKCC1 as well (Table 1). Next, we observed the impact of staurosporine and NEM on these phospho-sites (Fig 3). Both agents decreased the phosphorylation status of the WNK/SPAK sites Thr 906 (p-value for NEM or staurosporine: p = 0.0026) and Thr 1007 (p-value for NEM or staurosporine: p = 0.0026) of rnKCC2 and Thr 203/207/212 of hsNKCC1 (p-value for NEM and staurosporine:  [72][73][74][75]103]. The reduced phosphorylation of Thr 212 in hsNKCC1 agrees with our phosphoproteome analyses as no phosphorylation of Thr 212 /Thr 217 in NKCC1 was observed (Table 1). Previous analyses showed that SPAK directly phosphorylates Thr 1007 of rnKCC2 [48] and Thr 203/207/212 of hsNKCC1 [70]. Treatment of HEK293 cells with NEM resulted in a decrease of phosphorylation of Thr 233 (p = 0,0026), that is located in the T-loop kinase domain, and the S-loop phosphorylation site Ser 373 of hsSPAK (p = 0.0026) (Fig 3). Both are targets of WNKs [48,69,93]. As no data were available for the staurosporine mediated effect, we additionally analyzed its impact on these phospho-sites in HEK293 cells. Staurosporine also reduced the phosphorylation of Thr 233 (p = 0.0026) and Ser 373 (p = 0.0026) in hsSPAK (Fig 3). Thus, both agents reduced phosphorylation levels of these SPAK phospho-sites. These data conform well to our phosphoproteomic analyses, as no phosphorylated Ser 372/373 of hsSPAK was detected after treatment with either of the two agents.
Furthermore, staurosporine reduced phosphorylation of Ser 940 (p = 0.0026) in HEK rnKCC2b , whereas NEM increased phosphorylation of Ser 940 (p = 0.046) significantly (Fig 3). Since, Ser 940 is directly phosphorylated by PKC [77,104], we here analyzed the impact of both agents on the T-loop phosphorylation site Thr 505 of PKC-δ. Autophosphorylation of this site is most probably essential for kinase activity [105][106][107]. Staurosporine significantly decreases Thr 505 phosphorylation (p = 0.0026, Fig 3), whereas NEM slightly, but not significantly, increases Thr 505 phosphorylation (p = 0.064). The different impact of both agents on the phosphorylation of Thr 505 of PKC correlates well with their impact on Ser 940 phosphorylation of KCC2b.
We also determined whether staurosporine or NEM altered the total protein amount of rnKCC2b, hsNKCC1 or hsSPAK (Fig 3). Whereas NEM resulted in increased KCC2 amount (p = 0.0026), no obvious change was detected upon staurosporine treatment. NKCC1 and SPAK levels were not changed significantly upon treatment with either agent.
The number in brackets indicates in how many technical replica a given phospho-site was detected (max. 3). Each bracket provides the results of one biological experiment. https://doi.org/10.1371/journal.pone.0232967.t004

PLOS ONE
and Thr 203/207/212 in hsNKCC1. Additionally, both agents reduced phosphorylation of Thr 906 in rnKCC2b, which is phosphorylated by WNKs and a yet unknown kinase [48]. Staurosporine also reduced phosphorylation of the PKC site Ser 940 in rnKCC2b, whereas NEM increased its phosphorylation level. This correlated with the reduction of Thr 505 phosphorylation of PKC-δ upon staurosporine treatment and the impact of NEM to increase Thr 505 phosphorylation.
Quantitative analyses of phospho-sites of KCC2 and NKCC1 upon staurosporine or NEM treatment of rat immature hippocampal neurons As KCC2 is predominantly expressed in neurons [20], we analyzed for the first time the impact of staurosporine and NEM on the phosphorylation of specific phospho-sites of endogenously expressed KCC2 and NKCC1 using immature (9 DIV) primary rat hippocampal neurons (Fig  4). At this age cultured hippocampal neurons exhibit prominent level of Thr 906 and Thr 1007 KCC2 phosphorylation [66] that could be a subject of modulation by staurosporine and NEM.
To this end, we treated neurons with 8 μM staurosporine, 0.5 mM NEM or DMSO as a vehicle control for 15 min. NEM was reduced to 0.5 mM, since higher concentrations induced cell death. As described for the analyses in stably transfected HEK rnKCC2b cells, we used phosphospecific KCC2 and NKCC1 antibodies to quantify phosphorylation levels of each phospho-site relatively to the DMSO control. In untreated cultured immature hippocampal neurons, Ser 940 , Thr 906 and Thr 1007 in KCC2 and Thr 203, 207, 212 in NKCC1 were phosphorylated, similar to stably transfected HEK rnKCC2b cells (Fig 4). Most actions of staurosporine and NEM, as monitored by immunblots, were similar between HEK rnKCC2b cells and immature hippocampal neurons. Both agents resulted in decreased phosphorylation of Thr 1007 in KCC2 (p-value for NEM and staurosporine: p = 0.0026), Thr 203/207/212 in NKCC1 (p-value for NEM and staurosporine: p = 0.0026), and Ser 373 in SPAK (p-value for NEM: p = 0.0026 and staurosporine: p = 0.0026). Additionally, both agents reduced phosphorylation of the WNK-dependent phospho-site Thr 906 in KCC2 (p-value for NEM and staurosporine: p = 0.0026). Finally, NEM increased the total protein level of KCC2, as observed in HEK293 cells (p = 0.015) (Fig 4). We were not able to detect the phosphorylation of Thr 233 of SPAK using phospho-specific antibodies.
A marked difference, however, was observed for the PKC dependent phospho-site Ser 940 . Here, treatment with NEM reduced phosphorylation of Ser 940 (p = 0.0026) (Fig 4), contrary to the results obtained in stably transfected HEK rnKCC2b cells (Fig 3). Treatment with staurosporine also resulted in reduced phosphorylation of Ser 940 (p = 0.0026), which was similar to its action in HEK rnKCC2b cells. We could not detect Thr 505 phosphorylation of PKC-δ due to low expression rates.
To sum up, NEM affected the phosphorylation status of Ser 940 in immature hippocampal neurons (decrease) in the opposite way compared to HEK rnKCC2b cells (increase). All other effects of staurosporine and NEM were similar between immature hippocampal neurons and HEK293 cells, i.e. both reduced the phosphorylation status of Ser 940 , Thr 906 , and Thr 1007 in KCC2 and Thr 203/207/212 in hsNKCC1.

PLOS ONE
Posttranslational regulation via the WNK-SPAK/OSR1 dependent phosphorylation represents a potent mechanism to regulate transport activity of KCC2 and NKCC1 in a reciprocal way [60,[108][109][110] (Fig 5). Here, we show that staurosporine and NEM decrease phosphorylation of Thr 233 and Ser 373 in SPAK, of Thr 1007 in rnKCC2 and Thr 203 , Thr 207 and Thr 212 in hsNKCC1 in both HEK293 cells and immature cultured hippocampal neurons. Since SPAK directly impairs phosphorylation of Thr 1007 in rnKCC2 and Thr 203 , Thr 207 and Thr 212 in hsNKCC1 [38,48,59,60,63,[68][69][70][71][72][73][74][75][76], our data suggest that staurosporine and NEM directly affect the WNK-SPAK/OSR1 mediated phosphorylation of these residues in KCC2 and NKCC1. The data are in line with previous analyses showing that NEM reduces phosphorylation of Ser 373 in SPAK and Thr 1007 in KCC2 using HEK293 cells and immature cortical neurons [48]. Furthermore, application of staurosporine and NEM decreases phosphorylation of Thr 906 in rnKCC2. This site is directly phosphorylated by WNKs and a yet unknown kinase [48]. However, functional in-depth analyses such as mutagenic approaches are required to prove a causal relation between dephosphorylation of SPAK Thr 233 and Ser 373 and dephosphorylation of the specific KCC2 and NKCC1 phospho-sites upon staurosporine and NEM treatment.
Recent analyses demonstrated that dephosphorylation of Thr 906 and Thr 1007 increases KCC2 activity [61,[64][65][66], whereas dephosphorylation of Thr 203 , Thr 207 and Thr 212 decreases NKCC1 activity [45,51]. Furthermore, staurosporine and NEM results in activation of KCC2 (this study and [40,47,48,59,60,79,80], whereas they reduce NKCC1 activity [43][44][45][46]51]. This suggests, that staurosporine and NEM mediated dephosphorylation of these phosphosites result in a reciprocal regulation of KCC2 (activation) and NKCC1 (inactivation) activity most likely via the WNK/SPAK-dependent phosphorylation pathway. This is also in line with the observation in immature hippocampal neurons that KCC2 can rapidly be activated by staurosporine [42]. Another key regulatory KCC2 phospho-site is the PKC-mediated phosphorylation of Ser 940 . Phosphorylation of Ser 940 enhances KCC2 cell surface expression and increases ion transport activity, whereas mutation of serine to alanine (mimicking the dephosphorylated state) results in transport activity that is equal or decreased compared to wild-type KCC2 (KCC2 wt ) [65,77,111]. Our data demonstrate that staurosporine and NEM can differentially affect Ser 940 phosphorylation. Treatment of immature hippocampal neurons with either agent results in decreased phosphorylation of Ser 940 . However, treatment of HEK rnKCC2b cells with staurosporine decreases, whereas NEM increases phosphorylation of Ser 940 . This is in line with the different effect of these agents on transport activity of the phosphomutants Ser 31A/D , Thr 34A/D , Thr 999A and Thr 1008A/D [40].
Moreover, NEM has a cell-type specific impact on Ser 940 phosphorylation. In immature cortical neurons [48] and HEK293 cells, NEM increases Ser 940 phosphorylation, whereas it decreases Ser 940 phosphorylation in immature cultured hippocampal neurons (Fig 4). The mechanisms that cause this opposite effect of NEM on Ser 940 phosphorylation in different tissues is unclear. One possibility is that different NEM concentrations (HEK293 cells: 1 mM and 0.1 mM; immature hippocampal neurons: 0.5 mM; immature cortical neurons: 0.1 mM [48]) affect different regulatory pathways. Alternatively, tissues-specific sets of PKC isoforms and phosphatases result in different phosphorylation patterns. Indeed, the PKC family consists of 10-12 isoforms grouped into three classes [98][99][100][101] that vary in their expression profile [112] and regulation of their activity through several regulatory proteins, co-factors and second messenger cascades [100,101,113]. This offers the opportunity to differentially regulate KCC2 function in distinct neuronal populations through PKC. In HEK rnKCC2b cells, we showed that staurosporine reduced and NEM slightly increased phosphorylation of the T-loop phosphosite Thr 505 of PKC-δ. This correlated with decreased KCC2 Ser 940 phosphorylation upon staurosporine treatment and increased phosphorylation of this residue upon NEM treatment. This suggests, that both agents directly act on PKC-δ mediating the phosphorylation of Ser 940 in HEK rnKCC2b cells. However, more functional in-depth analyses are required to elucidate the causal link between dephosphorylation of PKC-δ Thr 505 and dephosphorylation of KCC2 Ser 940 upon staurosporine treatment. Since we were unable to detect phosphorylation of PKCδ Thr 505 in immature hippocampal neurons, we suggest that other PKC isoforms are involved in the direct phosphorylation of KCC2 Ser 940 in immature hippocampal neurons.
Our data furthermore reveal that staurosporine (HEK rnKCC2b cells, immature hippocampal neurons) and NEM treatment (immature hippocampal neurons) decrease Ser 940 phosphorylation resulting in an equal or diminished transport activity compared to KCC2 wt [65,77,111]. We therefore conclude that Ser 940 is not the key regulatory phospho-site mediating the staurosporine and NEM-based stimulation effect on KCC2. This is in line with recent published analyses showing that NEM still enhances the transport activity of Ser 940A (mimicking dephosphorylated state), indicating that other phospho-sites are important in NEM-dependent stimulation [40,48].
NEM but not staurosporine increased total KCC2 amount in HEK293 cells and immature cultured hippocampal neurons. In immature cortical neurons, Deep and coworkers [48] detected the same trend of enhanced total KCC2 abundance (albeit not significant), resulting in increased cell surface expression [48]. This suggests that in contrast to staurosporine, NEM increases KCC2 expression and trafficking that could result in a higher KCC2 activity. The different impact of NEM on total KCC2 abundance in HEK293 in the study of Deep and coworkers (no increase) and our analyses (increase) could result from different NEM concentration used in the experiments (0.1 mM vs. 1 mM).
Via mass spectrometry analysis, we identified several new phosphorylation sites whose function awaits further investigation. These sites are the N-terminal KCC2 phospho-site Thr 32 , and the C-terminal NKCC1 phospho-sites Ser 242 , Thr 266 , Th 268 , Ser 940 , Tyr 956 and Ser 957 . Future studies should investigate their regulatory impact on KCC2 and NKCC1 activity.

Conclusions
In conclusion, our data identify molecular mechanisms involved in staurosporine and NEM mediated changes in transport activity of KCC2 and NKCC1, which are a defining feature of CCCs [114]. The observation of cell-type specific action of these agents is in line with different reversal potentials in mature neuronal populations [115] and calls for comprehensive neuron subtype-specific phospho analysis. The recently reported structural data of CCCs [116][117][118][119] finally lay the foundation to analyze jointly the physiological role of phosphorylation and underlying structural changes to obtain an integrative and mechanistic view of the action of phosphorylation.
Supporting information S1