Global discovery of high-NaCl-induced changes of protein phosphorylation

Abstract

High extracellular NaCl, such as in the renal medulla, can perturb and even kill cells, but cells mount protective responses that enable them to survive and function. Many high-NaCl-induced perturbations and protective responses are known, but the signaling pathways involved are less clear. Change in protein phosphorylation is a common mode of cell signaling, but there was no unbiased survey of protein phosphorylation in response to high NaCl. We used stable isotopic labeling of amino acids in cell culture coupled to mass spectrometry to identify changes in protein phosphorylation in human embryonic kidney (HEK 293) cells exposed to high NaCl. We reproducibly identify >8,000 unique phosphopeptides in 4 biological replicate samples with a 1% false discovery rate. High NaCl significantly changed phosphorylation of 253 proteins. Western analysis and targeted ion selection mass spectrometry confirm a representative sample of the phosphorylation events. We analyze the affected proteins by functional category to infer how altered protein phosphorylation might signal cellular responses to high NaCl, including alteration of cell cycle, cyto/nucleoskeletal organization, DNA double-strand breaks, transcription, proteostasis, metabolism of mRNA, and cell death.

interstitial NaCl can vary considerably between tissues and can be very high, particularly in the renal medulla; yet despite such continuous and, at times, variable stress, cells survive and function by means of a number of osmoprotective responses (6). Hypertonicity, such as that caused by high NaCl, shrinks cells, increases intracellular ionic strength, delays the cell cycle, rearranges the cytoskeleton, increases DNA breaks, and alters transcription and translation (6). Protective responses include increase in regulatory volume, increase in heat shock proteins (HSPs), and accumulation of organic osmolytes (6). The signaling involved in the damage and the protective responses is incompletely understood. Phosphorylation of proteins is a common form of cell signaling. Therefore, in the present studies, to further elucidate the signaling pathways involved, we used stable isotopic labeling of amino acids in cell culture (SILAC) (52) coupled to proteomic mass spectrometry [MS; multidimensional liquid chromatography (LC) coupled with tandem MS (LC/LC-MS/MS)] to profile high-NaCl-induced changes in the phosphoproteome of human embryonic kidney (HEK 293) cells. On the basis of these results and known effects of the changes in protein phosphorylation, we infer how changes in phosphorylation of particular proteins might signal damage to cells and cellular protective responses.

METHODS AND MATERIALS

Cell culture in SILAC.

HEK 293 cells (American Type Culture Collection, Manassas, VA) were cultured in 90% Eagle's minimal essential medium (Invitrogen, Carlsbad, CA) plus 10% fetal bovine serum in 5% CO₂-95% air at 37°C. Cells were used between passages 38 and 50. Cells were transferred to SILAC medium (Invitrogen, Carlsbad, CA), either [¹³C₆,¹⁵N₄-arginine,¹³C₆-lysine]DMEM (“heavy”) or [¹²C₆,¹⁴N₄-arginine,¹²C₆-lysine]DMEM (“light”), at 300 mosmol/kg for six generations. Incorporation of labeled amino acids was measured by LC-tandem MS (LC-MS/MS) analysis of tryptic peptides from heavy amino acid-equilibrated cells. After six generations, 97.5% of cellular proteins had incorporated the heavy amino acids. Osmolality of the heavy medium bathing 80% confluent equilibrated cells was increased to 500 mosmol/kg (NaCl added), while the light medium was exchanged for an identical medium at 300 mosmol/kg. After 1 h, the cells were washed once with PBS of the same osmolality (37°C), harvested in PBS of the same osmolality, and pelleted at 3,000 g for 4 min at 4°C. The experiment was performed three times in this manner. To ensure against bias due to isotope equilibration, a fourth experiment was performed with the same isotopes, but with the labeling reversed, so that medium bathing the light-equilibrated cells was increased to 500 mosmol/kg, while heavy-equilibrated cells remained at 300 mosmol/kg medium.

Sample preparation for MS.

The pelleted cells were lysed in 8 M urea, 50 mM Tris·HCl, and 75 mM NaCl with added protease inhibitor (Roche Diagnostics, Indianapolis, IN) and phosphatase inhibitors I and II (Sigma, St. Louis, MO). Proteins were digested with sequencing-grade trypsin (Promega, Madison, WI) at a ratio of 1:20 at 37°C for 16 h; then the samples were desalted with HLB cartridges (Waters, Milford, MA). The peptide mixture was fractionated by strong cation-exchange chromatography; then phosphopeptides were enriched using immobilized metal ion affinity chromatography columns (Pierce, Rockford, IL) before MS.

MS.

LC-MS/MS was carried out with an Eksigent nanoflow LC system connected to a mass spectrometer (LTQ-Orbitrap XL, Thermo Scientific, Waltham, MA). Data-dependent acquisition mode was enabled, and each survey MS scan was followed by six MS2 scans with dynamic exclusion of 20 s. For targeted ion selection (TIS), peptides were analyzed with an LTQ-Orbitrap Velos (Thermo Scientific), preselecting 50 phosphopeptide targets by specification of their predicted sizes for MS1. The identity of the MS1 peak areas of those phosphopeptides was confirmed by MS2 and quantified as described below.

Protein identification.

MS raw data were searched with the SEQUEST (20) and InsPectT (61) algorithms. Peptides were identified by search against a target-decoy human protein database (downloaded from the National Center for Biotechnology Information website ftp://ftp.ncbi.nih.gov/refseq/H_sapiens/H_sapiens/protein/) with 1% false discovery rate (FDR). Phosphorylation site was assigned using the scoring algorithm phosphate localization score (54). Phosphorylation sites were confirmed using the National Heart, Lung, and Blood Institute in-house programs ProMatch (63) and PhosphoPIC (28).

Phosphopeptide quantification.

Relative phosphopeptide abundance was calculated from the MS1 peak area using in-house software (QUIL) (66) with FDR of 1%. Comparisons were either heavy or light in the SILAC experiments or either 300 or 500 mosmol/kg in the TIS experiments. The weighted means of MS1 peak areas of phosphopeptides were used to calculate relative abundance ratio: heavy-to-light ratio (H/L) = [∑_{i = 0}ⁿ(H_i/L_i∗√L_i∗H_i)]/[∑_{i = 0}ⁿ(√L_i∗H_i)], where L_i is peak area integrated from ion intensity for light peak and H_i is peak area integrated from ion intensity for heavy peak. R and Perl scripts were used for data analysis. Students' t-test was used to determine the significance of changes in abundance of phosphopeptides. We used a threshold of P ≤ 0.05 (t-test) and further selected for log₂(500/300 mosmol/kg) less than or equal to −1 or ≥1 to identify important changes. Phosphorylation changes were in satisfactory agreement between the four replicates (Pearson's correlation coefficient = 0.63–0.80, with arginine-to-proline conversion taken into account).

iTRAQ.

For independent quantification of protein abundance, HEK 293 cells were grown in non-SILAC medium at 300 mosmol/kg. Osmolality of the medium bathing 80% confluent cells was increased to 500 mosmol/kg (NaCl added) for 1 h; then proteins were extracted and trypsinized, as described above, and the peptides were labeled differentially with isobaric tags (115 and 117 for 300 and 500 mosmol/kg, respectively) using an iTRAQ 8plex kit (AB Sciex, Framingham, MA). The iTRAQ-labeled peptides were combined and quantified with an Orbitrap Velos mass spectrometer. Results from three biological replicates were analyzed for differences between 300 and 500 mosmol/kg using Proteome Discoverer (Thermo Scientific). Changes were deemed significant if a protein's abundance ratio (500/300) deviated >10% from unity and P < 0.05 in a one-sample t-test. Of the ∼2,500 proteins identified, only 4 met these criteria.

Interpretation.

Gene Ontology functional category enrichment was analyzed by Database for Annotation, Visualization, and Integrated Discovery (DAVID; http://david.abcc.ncifcrf.gov/gene2gene.jsp).

Western blotting.

HEK 293 cells at 70% confluence were maintained in 300 or 500 mosmol/kg (NaCl added) medium for 1 h. Whole cell protein was extracted using PhosphoSafe extraction reagent (EMD, La Jolla, CA). Protein (30 μg) was separated on 3–8% gradient Tris-acetate or 4–12% gradient Bis-Tris gel and transferred electrophoretically to nitrocellulose membranes. Membranes were incubated in Odyssey blocking buffer (LI-COR Biosciences, Lincoln, NE) for 1 h at room temperature. Then membranes were incubated with the following antibodies overnight at 4°C: rabbit anti-phosphorylated (T71) activating transcription factor 2 (ATF-2), rabbit anti-ATF-2, rabbit anti-phosphorylated (S82) HSP27, mouse anti-HSP27, rabbit anti-phosphorylated (S727) STAT1, and mouse anti-STAT1 (Cell Signaling Technology, Danvers, MA); rabbit anti-phosphorylated (S421/S423) histone deacetylase 1 (HDAC1; Millipore); mouse anti-HDAC1, rabbit anti-phosphorylated (S63) c-Jun, rabbit anti-phosphorylated (S73) c-Jun, and mouse anti-c-Jun (Cell Signaling Technology); and rabbit anti-phosphorylated (S232) Ras GTPase-activating protein-binding protein (G3BP), mouse anti-G3BP, rabbit anti-phosphorylated (S964) mediator of DNA damage checkpoint protein 1 (MDC1), and rabbit anti-MDC1 (Abcam, Cambridge, MA). After they were washed with 0.1% Tween 20 in PBS, the membranes were incubated with Alexa Fluor 680-conjugated goat anti-rabbit IgG (Invitrogen, Eugene, OR) or IRDye 800-conjugated goat anti-mouse IgG (Rockland Immunochemicals, Gilbertsville, PA). Blots were visualized and quantified using an Odyssey infrared imager (LI-COR Biosciences).

RESULTS AND DISCUSSION

Phosphoproteome profiling and quantification.

We performed the experiment four times: three times using cells equilibrated with the heavy SILAC label at 300 mosmol/kg and then increased to 500 mosmol/kg for 1 h by addition of NaCl and once with the amino acid labeling reversed. We analyzed the results with SEQUEST (20) and InsPecT (61) algorithms searched against human concatenated real and reversed protein databases with 1% FDR. Combining the results from SEQUEST and InsPecT, we reproducibly identify >8,000 unique phosphopeptides in four biological replicate samples with a 1% FDR (Fig. 1A). The unique phosphopeptides are contained in 2,862 different proteins (Fig. 1A). We assume that abundance of most proteins is not changed by 1 h of high NaCl, so the phosphopeptide changes represent altered phosphorylation at particular sites in proteins, rather than change in amount of the proteins. This assumption is confirmed by our finding that abundance of few proteins (independently measured by iTRAQ) changes significantly after 1 h of high NaCl (Fig. 1B) and none of the phosphopeptides whose abundance changes significantly are contained in proteins whose abundance changes (see Supplemental Table S1 in Supplemental Material for this article, available online at the Journal website). We quantified the effect of high NaCl on phosphorylation by determining the relative areas of the MS1 peaks for each phosphopeptide at 500 vs. 300 mosmol/kg. An example of the peptide containing p38/phosphorylated (Y182) MAPK14 is shown in Fig. 1, C and D. We calculated the means of those ratios, weighted for ion intensity. In all, we are able to quantify 7,362 unique phosphopeptides, containing 8,876 phosphorylation sites; 80% of the peptides are phosphorylated at a single site and 20% are phosphorylated at multiple sites. Abundance of 120 of the phosphopeptides increases significantly (P < 0.05) at 500 mosmol/kg, whereas abundance of 207 phosphopeptides decreases significantly (Fig. 1E; see Supplemental Table S1). The 324 differentially regulated phosphopeptides are contained in 253 different proteins (see Supplemental Table S1).

Fig. 1.

A: number of unique phosphopeptides identified by InsPecT and/or SEQUEST algorithms. B: effect of elevating NaCl for 1 h on abundance of proteins, as measured by iTRAQ. Abundance of only very few proteins changed significantly. C: phosphorylated (Y182) p38/MAPK14 peptide R.HTDDEMTGY*VATR.W(+2). Stable isotopic labeling of amino acids in cell culture (SILAC)-labeled isoforms were identified at 20.78 min. D: high NaCl increased abundance of R.HTDDEMTGY*VATR.W(+2). MS1 peaks are compared. E: phosphopeptides whose abundance was changed by high NaCl. High NaCl significantly increased abundance of 120 and decreased abundance of 207 of the 7,362 unique phosphopeptides that were quantified.

Comparison of quantification by SILAC with quantification by TIS and Western blotting.

We used TIS, a label-free method, and Western immunoblotting to further test changes in phosphorylation identified by SILAC. Figure 2 shows dynein, cytoplasmic 1, light intermediate chain 1 phosphorylated (S516) at high NaCl, as quantified using SILAC (Fig. 2A) and using TIS (Fig. 2B). In TIS, the MS1 peak area of a selected phosphopeptide quantifies its relative abundance. Using TIS, we were able to quantify the effect of high NaCl on 17 of the differentially changed phosphopeptides indicated by SILAC. The results from TIS correlate highly with those from SILAC (R² = 0.83; Fig. 3A), which supports the accuracy of both measurements. We also selected eight phosphorylation sites in phosphopeptides whose abundance changed in the SILAC experiment and for which phosphospecific antibodies were available to confirm the differential changes by Western blot analysis. Figure 3B shows that the relative phosphorylation abundance identified in SILAC correlates well (R² = 0.67) with that shown by Western immunoblotting. However, the magnitude of the changes, as measured by Western blotting, is less than that measured by MS. We previously observed that MS and Western blot analysis give largely equivalent results for changes in protein abundance (44). Our current results extend that conclusion to measurements of protein phosphorylation. The lack of even better agreement may be due to less accuracy of Western blotting (1).

Fig. 2.

High-NaCl-induced change in abundance of the phosphopeptide containing phosphorylated (S516) DYNC1LI1, as measured by SILAC (A) and targeted ion selection (TIS; B). Relative MS1 peak areas are in good agreement between the two methods.

Fig. 3.

Effect of high NaCl on phosphopeptide abundance as determined by TIS vs. SILAC (A) and Western blot with phosphospecific antibodies vs. SILAC (B). Proteins include those whose phosphorylation changed significantly and those whose phosphorylation did not change significantly. Scales are log₂(500/300 mosmol/kg). High correlation coefficients between the paired methods support the accuracy of both. Gene symbols-(phosphorylation sites) are as follows: in A, CCNL2(S330) (1), CTTN(T364) (2), GIGYF2(S26) (3), AKT151(S212) (4), TBC1D4(S570) (5), CTNNB1(S552) (6), NEDD4L(S327) (7), GRLF1(S1179) (8), RABEP1 (9), GAPVD1(S929) (10), DTNA(S605) (11), MDC1(S376) (12), MYO9B(S1290) (13), TRIM28(S473) (14), EPS15(S482) (15), DYNC1LI1(S51) (16), MAVS(S222) (17), KIAA1432(S1017) (18), ZFYVE16(S939) (19); in B, G3BP1(S232) (1), MDC1(S964) (2), HDAC1(S421/S423) (3), JUN(S73) (4), STAT1(S727) (5), ATF2(T71) (6), HSPB1/HSP27(S82) (7), JUN(S63) (8).

Role of phosphorylation in signaling known effects of high NaCl.

High NaCl is known to affect cell cycle, cyto/nucleoskeletal organization, DNA double-strand breaks (DSBs), transcription, proteostasis, metabolism of mRNA, and cell death. We used the DAVID bioinformatic tool (http://david.abcc.ncifcrf.gov/) as a basis for identification of proteins whose phosphorylation is altered by high NaCl and whose functions are related to the known effects of high NaCl. This analysis provides insight into the roles of those proteins in response to high NaCl. Where other references are not cited, information about function of specific proteins and of phosphorylation sites within them is from PhosphoSitePlus (30).

High-NaCl-induced cell cycle delay.

Acute elevation of NaCl produces rapid arrest at all phases of the cell cycle (47). We find that high NaCl alters phosphorylation in 20 proteins involved in regulating the cell cycle (Table 1).

Table 1.

Regulation of cell cycle

Official Gene Symbol	Change, log2(fold)	Phosphorylated Amino Acid Site	Name
EGFR	3.46	T693	Epidermal growth factor receptor
TP53BP2	2.90	S433	Tumor protein p53-binding protein 2
MAPK14	2.87	Y182	MAP kinase p38α
ATM	2.67	S1981	Ataxia telangiectasia-mutated
SIPA1	2.13	S53	Signal-induced proliferation-associated 1
HCFC1	1.57	S598	Host cell factor C1 (VP16-accessory protein)
	2.42	S666
MDC1	1.42	S1605	Mediator of DNA-damage checkpoint 1
	1.55	T1239
ANLN	1.36	S792	Anillin, actin-binding protein
TTK	1.21	S281	TTK protein kinase
DYNC1LI1	1.16	S510, S516	dynein, cytoplasmic 1, light intermediate chain 1
	1.32	T515
	3.25	T513, S516
RBL1	−1.11	S1037, S1041	Retinoblastoma-like 1 (p107)
PML	−1.69	S518, S527	Promyelocytic leukemia
JUN	−1.86	S243	jun proto-oncogene
	2.07	S63
	2.13	T62
TPR	−1.87	S1185	Translocated promoter region, nuclear basket protein
RB1	−2.16	S249, T252	Retinoblastoma 1
TP53BP1	−2.09	S557	Tumor protein p53-binding protein 1
FOXC1	−2.23	S235, S241	Forkhead-related transcription factor 3
CDK13	−2.33	S1054, T1058	Cyclin-dependent kinase 13
DLG1	−2.61	S575	Synapse-associated protein 97
RBL2	−2.69	S413, T417	Retinoblastoma-like 2 (p130)

Previously, we found that high-NaCl-induced increase in phosphorylation and activation of p38/MAPK14 kinase is responsible for rapid initiation of delay of the G₂/M phase (16). We confirm here that high NaCl increases phosphorylated (Y182) MAPK14 (Table 1). We also previously found that high-NaCl-induced activation of p53/tumor protein 53 (TP53) is responsible for delay of the G₁/S phase (15). We do not identify changes in phosphorylation of p53 itself in the present experiments, but we do observe altered phosphorylation of related proteins, TP53-binding protein (TB53BP) 2 and TP53BP1 (Table 1). TP53BP2 and TP53BP1 can affect the cell cycle, but their role in high-NaCl-induced cell cycle delay remains to be confirmed.

Retinoblastoma (RB) tumor suppressor proteins (RB1, RBL1, and RBL2) regulate entry into the S phase of the cell cycle after the G₁ phase (68). Early in the G₁ phase, RB proteins are hypophosphorylated, but they become hyperphosphorylated late in the G₁ phase and maintain hyperphosphorylation through the remainder of the cell cycle. Hypophosphorylated RB binds to E2F transcription factors, which prevents transcription of E2F target genes that drive the cell cycle. Hyperphosphorylation of RB disrupts its inhibitory binding to E2Fs. Accordingly, overexpression of RBL1 and RBL2, in which phosphorylation sites have been mutated so that they cannot be phosphorylated, arrests the cell cycle at the G₁ phase (22). Phosphorylation inactivates the growth-inhibitory functions of RB. Increased phosphorylation of RB1 at S249 and T252 has been reported to reduce delay of the G₁/S phase caused by RB1 (8). Therefore, decreased phosphorylation, such as occurs when NaCl is elevated (Table 1), should increase the delay. We conclude that high-NaCl-induced decrease in phosphorylation of RBL1, RBL2, and RB1 (Table 1) may contribute to high-NaCl-induced G₁/S phase delay. Furthermore, rapid high-NaCl/p38-dependent initiation of G₂/M phase delay is followed by inhibition of cyclin-dependent kinase 1 (CDK1)/Cdc2 kinase (16). Since CDK1/Cdc2 phosphorylates RB1 at S249 and T252 (43), its inhibition could contribute to termination of G₁/S phase delay.

High-NaCl-induced cyto/nucleoskeletal reorganization.

Hypertonicity induces reorganization of the cytoskeleton, including submembranous F-actin assembly, which reinforces the cell structure to withstand the physical challenge of cell shrinkage (14, 29). Hypertonicity also causes nucleoskeletal reorganization involving changes in abundance in the nucleus of several “cytoskeletal” proteins (44). At 1 h after we increased NaCl, changes occurred in phosphorylation of 27 proteins involved in regulating cyto/nucleoskeletal organization (Table 2). The proteins include those involved with GTPase signaling (SIPA1, DOCK7, CDC42EP3, SEPT7, ARHGEF18, and RALGAPA1), actin cyto/nucleoskeleton (ANLN, ZYX, ABLIM1, CTTN, KIF13B, and FARP1), microtubules (MAP2, MAP4, MAP7, MAP1B, CLIP1, and CLASP1), and cell adhesion (CTNNB1, CTNND1, and PKP2) and include a protein kinase (LIMK1), a myosin [myosin heavy chain 9 (MYH9)], a HSP (HSPB1), and adapter proteins (SHKBP1 and DLG1).

Table 2.

Regulation of cytoskeletal and nucleoskeletal organization

Official Gene Symbol	Change, log2(fold)	Phosphorylated Amino Acid Site	Name
SIPA1	2.13	S53	Signal-induced proliferation-associated protein 1
DOCK7	1.94	S30	Dedicator of cytokinesis 7
HSPB1	1.92	S82	Heat shock protein β1
MYH9	1.69	S1943	Myosin-9
ANLN	1.36	S792	Anillin, actin-binding protein
CDC42EP3	1.18	S89	Cdc42 effector protein 3
ZYX	1.16	S344	Zyxin
MAP7	−1.21	S209	Microtubule-associated protein 7
CTNNB1	−1.22	S552	Catenin (cadherin-associated protein), β1, 88 kDa
SEPT7	−1.25	T426	Cell division cycle 10 isoform 2
ABLIM1	−1.27	S655	Actin-binding LIM protein
FARP1	−1.33	S427	FERM, RhoGEF and pleckstrin domain-containing
ARHGEF18	−1.33	S1103	Rho-specific guanine nucleotide exchange factor
EPB41L3	−1.59	S460, T469	Erythrocyte membrane protein band 4.1-like 3
CTTN	−1.59	T401	Cortactin isoform a
MAP2	−1.69	T306	Microtubule-associated protein 2 isoform 5
LIMK1	−1.72	S310	LIM domain kinase 1
MAP1B	−1.77	S1779, S1782, T1788	Microtubule-associated protein 1B
	1.22	S1785
RALGAPA1	−1.94	S775	GTPase-activating Rap/RanGAP domain-like 1
CLIP1	−2.19	S204	Restin isoform a
CLASP1	−2.21	S1071	CLIP-associating protein 1 isoform 2
	−1.35	S797
SHKBP1	−2.37	S587	SH3-domain kinase-binding protein 1 isoform a
DLG1	−2.61	S575	Discs, large homolog 1 isoform 2
PKP2	−2.62	S329	Plakophilin 2 isoform 2a
	1.42	S151
CTNND1	−2.98	S167	Catenin, δ1 isoform 3ABC
KIF13B	−3.53	S1644	Kinesin family member 13B
MAP4	−4.86	S1073	Microtubule-associated protein 4 isoform 1
	−4.01	S636

Rho family small G proteins are pivotal regulators of actin organization (29). They are highly sensitive to cell volume changes. However, the events between cell volume decrease and Rho protein activation remain enigmatic, at least in part due to the daunting number of upstream regulators of these proteins. Our observation of high-NaCl-induced changes in phosphorylation of proteins involved in GTPase signaling (Table 2) provides clues for identification of additional upstream regulators.

High NaCl decreases phosphorylation of β-catenin/CTNNB1 at S552, catenin D1/CTNND1 at S167, and cortactin/CTTN at T401 (Table 2). β-Catenin links members of the cadherin family of transmembrane cell-cell adhesion receptors to the actin cytoskeleton. AKT (21) and PKA (62) were previously shown to increase phosphorylation of β-catenin/CTNNB1 at S552. AKT (74) and PKA (23) are activated by high NaCl, but their activation should increase phosphorylation of β-catenin/CTNNB1 at S552, not cause the decrease that we found. We speculate that some other (unidentified) kinase or phosphatase may be involved in the decreased phosphorylation of β-catenin/CTNNB1 at S552.

We previously found that high NaCl decreases the abundance of 10 different microtubule proteins in the nucleus within 1 h (44). We now find decreased phosphorylation of six microtubule-associated proteins (MAP2, MAP4, MAP7, MAP1B, CLIP1, and CLASP1) at that time, suggesting that these microtubule-associated proteins might regulate the decrease in nuclear tubulins. High NaCl decreases phosphorylation of MAP7 at S209 (Table 2). The decreased phosphorylation could result from high-NaCl-induced inhibition of CDK1/Cdc2 (16), which phosphorylates MAP7 at S209 (3). These findings reinforce the idea (44) that decrease in nuclear tubulins is associated with rapid high-NaCl-induced cell cycle delay and suggest a previously unknown function of microtubule-associated proteins.

High NaCl increases phosphorylation of MYH9 at S1943 (Table 2), a site whose phosphorylation is known to be involved in cytoskeletal organization, since reduction of phosphorylation of MYH9 at S1943 is associated with its redistribution during ionizing radiation-induced senescence of human mesenchymal stem cells (65). Casein kinase 2 catalyzes phosphorylation of MYH9 at S1943 (65). We do not know, however, whether hypertonicity increases casein kinase 2 activity, which requires further investigation.

High-NaCl-induced DNA damage.

High NaCl increases DNA DSBs in cell culture and in Caenorhabditis elegans (18), renal inner medullas (17), and marine invertebrates (19). The DSBs are not repaired while NaCl remains high but are rapidly repaired when NaCl is lowered. Importantly, high NaCl also decreases efficiency of repair of DNA damage caused by UV radiation, which ordinarily is rapid (17). Some DNA damage response proteins are known to be inhibited by high NaCl. While NaCl remains high, meiotic recombination 11 homolog 1 [Mre11 (MRE11A)] exonuclease is mainly present in the cytoplasm, rather than the nucleus, and histone H2AX (H2AFX) is not phosphorylated, as it normally would be in response to DNA damage (17). If NaCl is subsequently reduced, Mre11 returns to the nucleus and H2AX becomes phosphorylated, accompanying the DNA repair. The changes that we now find in phosphorylation of other DNA damage response proteins (Table 3) can add to our understanding of how high NaCl inhibits DNA repair.

Table 3.

Response to DNA damage stimulus

Official Gene Symbol	Change, log2(fold)	Phosphorylated Amino Acid Site	Name
ATM	2.67	S1981	Ataxia telangiectasia mutated protein isoform 1
BAZ1B	1.87	S947	Tyrosine-protein kinase BAZ1B
MDC1	1.42	S1605	Mediator of DNA damage checkpoint 1
	1.55	T1239
SFPQ	1.11	S273	Splicing factor proline/glutamine rich (polypyrimidine tract-binding protein-associated)
LIG1	−1.53	S76	DNA ligase I
PML	−1.69	S518, S527	Promyelocytic leukemia protein isoform 1
TP53BP1	−2.09	S557	Tumor protein p53-binding protein 1 isoform 2
PNKP	−2.26	T118, T122	Polynucleotide kinase 3¢-phosphatase
CHAF1B	−2.51	T433	Chromatin assembly factor 1 subunit B

Polynucleotide kinase 3′-phosphatase (PNKP) is involved in DSB repair by nonhomologous end-joining. High NaCl reduces phosphorylation of PNKP at T118 and T122 (Table 3). This is opposite to the response in yeast to ionizing radiation, which is a nearly threefold increase in phosphorylation at those sites (2). The possibility that reduced phosphorylation at those sites contributes to inhibition of DNA repair by high NaCl suggests a direction for further investigation.

High NaCl increases phosphorylation of BAZ1B [Williams syndrome transcription factor (WSTF)] at S947 (Table 3). Repair of DSBs induced by genotoxic agents (e.g., ionizing radiation) is initiated by assembly on chromatin of foci that contain H2AX phosphorylated on S139 (“γH2AX”). In the absence of genotoxic stress, BAZ1B phosphorylates histone H2AX at Y142 (70), which prevents formation of γH2AX. This maintains a “standby” mode in which DNA is not repaired unnecessarily. When DSBs are induced by genotoxic agents, WSTF dissociates and is replaced by eyes absent (EYA1/EYA3) phosphatases, which dephosphorylate phosphorylated (Y142) H2AX, facilitating formation of γH2AX (12). In contrast to other genotoxic agents, high-NaCl-induced DSBs do not increase γH2AX (17). The explanation may be that high-NaCl-induced phosphorylation of BAZ1B at S947 maintains its activity and/or that high NaCl inhibits EYA1/EYA2 activity. Both of these possibilities warrant further investigation.

High-NaCl-induced changes in transcription.

High NaCl decreases transcription in general (11) but increases transcription of osmoprotective genes (6, 11). In the present study we find changes in phosphorylation of 44 proteins involved with regulation of transcription (Table 4). Those changes may help us understand why high NaCl specifically increases transcription of the osmoprotective genes. The proteins include transcription factors (FOXK2, YBX1, ZNF446, NFX1, ZNF687, JUN/c-Jun, ATF2, ZMYM2, RFX7, AFF4, CREB5, and STAT1), transcription coactivators/corepressors (MED24, DMAP1, PAWR, DAXX, LRRFIP1, ATF7IP, TLE3, MTA1, HMGA2, HMGA1, SLTM, FOXC1, DNTTIP2, DPF2, PML, HCFC1, GATAD2B, and GTF2F1), regulators of transcription from RNA polymerase III promoter (MAF1 and ARID1A), regulators of transcription from RNA polymerase II promoter (YEATS2, BRWD1, SAFB, ZNF462, and ZNF768), transcription elongation factors (SUPT5H, RDBP, and TCEB3), histone modifiers (RNF20 and SUV39H1), and RNA polymerase II (POLR2A).

Table 4.

Regulation of transcription

Official Gene Symbol	Change, log2(fold)	Phosphorylated Amino Acid Site	Name
TCEB3	2.91	S542	Elongin A
CREB5	2.69	T59, T61	cAMP responsive element-binding protein 5 isoform α
ATF7	2.64	T51, T53	Activating transcription factor 7 isoform 3
JUN	2.13	T62	jun oncogene
	2.07	S63
	−1.86	S243
ATF7IP	2.11	S113, T118	Activating transcription factor 7-interacting protein
PAWR	1.95	T230	PRKC, apoptosis, WT1, regulator
ZMYM2	1.77	S305	Zinc finger protein 198
ATF2	1.76	T69, T71	Activating transcription factor 2
HCFC1	1.57	S598	Host cell factor 1
	2.42	S666
STAT1	1.24	S727	Signal transducer and activator of transcription 1
SUV39H1	−1.01	S391	Suppressor of variegation 3–9 homolog 1
YBX1	−1.02	S165	Nuclease-sensitive element-binding protein 1
BRWD1	−1.06	S1475	Bromodomain and WD repeat domain-containing 1
DMAP1	−1.13	T445	DNA methyltransferase 1-associated protein 1
LRRFIP1	−1.19	S88	Leucine-rich repeat (in FLII)- interacting protein 1 isoform 2
ARID1A	−1.20	S696, S702	AT-rich interactive domain 1A isoform a
DNTTIP2	−1.22	S117	Deoxynucleotidyltransferase, terminal, interacting protein 2
MED24	−1.26	S860	Mediator complex subunit 24 isoform 2
	2.55	S849
SUPT5H	−1.32	S666	Suppressor of Ty 5 homolog isoform a
RDBP	−1.34	S353	RD RNA-binding protein
	2.25	S115
	2.80	S51
	3.10	S49
HMGA1	−1.37	S91, S92	High-mobility group AT-hook 1 isoform b
DPF2	−1.39	S142	Zinc finger protein ubi-d4
FOXK2	−1.40	S398	Forkhead box K2
AFF4	−1.53	S1043	ALL1 fused gene from 5q31
DAXX	−1.63	S683	Death domain-associated protein isoform b
RNF20	−1.65	S138	Ring finger protein 20
PML	−1.69	S518, S527	Promyelocytic leukemia protein isoform 1
HMGA2	−1.69	S101	High-mobility group AT-hook 2 isoform a
TLE3	−1.74	T334	Transducin-like enhancer protein 3 isoform a
SAFB	−1.86	S604	Scaffold attachment factor B
YEATS2	−1.86	S519	YEATS domain-containing 2
ZNF462	−1.92	S688	Zinc finger protein 462
NFX1	−1.96	S50	Nuclear transcription factor, X-box-binding 1 isoform 2
SLTM	−2.00	S1002	Modulator of estrogen-induced transcription isoform a
RFX7	−2.16	S1178	Regulatory factor X domain-containing 2
FOXC1	−2.23	S235, S241	Forkhead box C1
ZNF768	−2.36	S90, S97	Zinc finger protein 768
MAF1	−2.44	S75	MAF1 protein
ZNF446	−2.45	S137	Zinc finger protein 446
POLR2A	−2.46	S1878, T1884	DNA-directed RNA polymerase II polypeptide A
GTF2F1	−2.60	T389	General transcription factor IIF subunit 1
ZNF687	−2.74	S253	Zinc finger protein 687
GATAD2B	−2.88	S333	GATA zinc finger domain-containing 2B
	1.01	S486
MTA1	−3.21	T578	Metastasis-associated protein
	−1.70	S576

High NaCl activates the transcription factor nuclear factor of activated T cells (NFAT5), which increases transcription of osmoprotective target genes. Promoter regions of those target genes contain not only DNA elements specific for binding NFAT5, but also nearby activator protein 1 (AP-1) sites that bind the AP-1 proteins FOS/c-Fos and JUN/c-Jun (32). FOS/c-Fos and JUN/c-Jun are activated by high NaCl and contribute to increased transcription of NFAT5 target genes (32). Also, inhibition of MAPK14/p38, which is activated by high NaCl, reduces high-NaCl-dependent activation of a transcriptional reporter that contains both the NFAT5-specific DNA element and an AP-1 site, but only if the AP-1 site is intact (32). MAPK14/p38 increases c-Jun abundance by activating the c-jun promoter (46). In the present study we find additional evidence for pathways that activate AP-1 (Fig. 4). High NaCl increases phosphorylation of JUN/c-Jun at T62 and S63 and reduces phosphorylation at S243 (Table 4). We are unaware of any references to the effect on its activity of phosphorylation of JUN/c-Jun at T62. However, increased phosphorylation at S63 (26) and decreased phosphorylation at S243 (5) promote activity of JUN/c-Jun, and identification of the kinases and phosphatases that affect phosphorylation at those sites points to the regulatory pathways that are involved. Considering that MAPK8/JNK activates JUN/c-Jun by phosphorylation at sites including JUN/c-Jun phosphorylated at S63 (26) and high NaCl increases activity of MAPK8/JNK (73), we suggest that the MAPK8/JNK-JUN/c-Jun pathway contributes to high-NaCl-induced activation of NFAT5. Furthermore, reduced phosphorylation of JUN/c-Jun at S243 increases DNA-binding activity of JUN/c-Jun, and, conversely, elevated phosphorylation at that site decreases JUN/c-Jun activity (5). Also, mutation of JUN/c-Jun phosphorylated at S243 to phenylalanine, which cannot be phosphorylated, greatly increases the transactivating ability of JUN/c-Jun. In resting cells, JUN/c-Jun is in a latent, phosphorylated form that is activated by dephosphorylation at S243. GSK3B/GSK-3β phosphorylates JUN/c-Jun at S243 in resting cells, which maintains JUN/c-Jun in an inactive state. High NaCl inhibits GSK3B/GSK-3β by increasing phosphorylation of GSK3B/GSK-3β at S9 (76). Decreased GSK-3β activity, by reducing phosphorylation of JUN/c-Jun at S243, increases JUN/c-Jun activity. Further upstream, the high-NaCl-induced phosphorylation of GSK-3β at S9 depends on PKA, phosphatidylinositol 3-kinase, and AKT (76), which are themselves activated by high NaCl and contribute to activation of NFAT5 (23, 31, 57). PKC may also contribute to high-NaCl-induced activation of JUN/c-Jun. Application of the phorbol ester 12-O-tetradecanoylphorbol-13-acetate, which activates PKC, causes dephosphorylation of c-Jun-S243 (5). High NaCl is reported to activate PKC in most, but not all, studies (reviewed in Ref. 75), and PKCμ contributes to the hypertonicity-induced increase in HSP70 abundance (45). Further studies are necessary to determine whether PKC contributes to high-NaCl activation of JUN/c-Jun and, if so, what isoforms of PKC are involved. High NaCl also increases phosphorylation of two other AP-1 proteins, ATF2 phosphorylated at T69/T71 and ATF7 phosphorylated at T51/T53 (Table 4). ATF2 and ATF7 (25) are highly homologous AP-1 transcription factors that form homodimers or heterodimers with other AP-1 factors, such as c-Fos and c-Jun. ATF2 is activated by phosphorylation at T69 and T71 and ATF7 by phosphorylation at T51 and T53, as we observe in response to high NaCl. MAPK8/JNK and/or MAPK14/p38 phosphorylate those sites (13). However, we do not know what role, if any, ATF2 and ATF7 have in high-NaCl-induced activation of AP-1 sites flanking NFAT5 target genes. Figure 4, which is based on the above-described observations, presents a model of the signaling pathways by which high NaCl activates JUN/c-Jun and other AP-1 factors, including ATF2 and ATF7.

Fig. 4.

Pathways by which high-NaCl-induced changes in phosphorylation of signaling molecules activate activator protein 1 (AP-1), contributing to activation of the osmoprotective transcription factor nuclear factor of activated T cells (NFAT5). Green indicates increase and red indicates decrease in activity or phosphorylation of kinases (rectangles) or transcription factors (ovals).

Hypertonicity induced by sorbitol increases activity of STAT1 in COS-7 cells via MAPK14/p38-mediated phosphorylation of STAT1 at T701 (4). We do not find a significant increase in phosphorylation of STAT1 at T701 in HEK 293 cells in response to high-NaCl-induced hypertonicity. However, we do find substantially increased phosphorylation of STAT1 at S727 (Table 4). That may be pertinent, since phosphorylation of STAT1 at S727 increases transactivating activity of STAT1 (69). MAPK14/p38 could be responsible for phosphorylation of STAT1 at S727 in response to high NaCl, since MAPK14/p38 was previously found to phosphorylate STAT1 at S727 in response to interleukin-13 (71). However, we have reservations about that conclusion, since hypertonicity induced by sorbitol only slightly increases phosphorylation of STAT1 at S727 in COS-7 cells, and that phosphorylation is independent of p38 (4). Also, PKC-δ (PRKCD) (64) and phosphatidylinositol 3-kinase/AKT (50) are responsible for phosphorylation of STAT1 at S727 in response to interferon. Therefore, we used the p38 inhibitor SB-203580 to determine whether p38 contributes to high-NaCl-induced increase in phosphorylation of STAT1 at S727 in HEK 293 cells. SB-203580 inhibits high-NaCl-induced phosphorylation of STAT1 at S727 (Fig. 5), indicating that p38 does contribute to high-NaCl-induced phosphorylation of STAT1 at S727 in HEK 293 cells. Nevertheless, additional studies are required to identify the gene targets of the high-NaCl-induced increase in STAT1 activity.

Fig. 5.

p38/MAPK14 activity contributes to high-NaCl-induced increase in phosphorylation of STAT1 and HSP27/HSPB1. HEK 293 cells were preincubated for 30 min with the p38/MAPK14 inhibitor SB-203580 (8 μM) at 300 mosmol/kg; then the medium was changed for 1 h to an identical medium (control) or a medium still containing the inhibitor and with osmolality increased to 500 mosmol/kg by addition of NaCl. Immunoblots were prepared using anti-STAT1 and anti-phosphorylated (S727) STAT1 (A) or anti-HSP27/HSPB1 and anti-phosphorylated (S82) HSP27/HSPB1 antibodies (B). Top: representative Western blots. Bottom: means ± SE; n = 3. *P < 0.05 vs. control.

High NaCl decreases phosphorylation of general transcription factor II F (GTF2F1)/transcription factor II F (TFIIF)/RNA polymerase II-associating protein 74 (RAP74) at T389 (Table 4). GTF2F1/TFIIF/RAP74 is a general transcription factor that binds to RNA polymerase II, helps recruit it to the initiation complex, and promotes transcription elongation. GTF2F1/TFIIF/RAP74 has serine/threonine kinase activity that autophosphorylates it at S385 and T389 (56). The autophosphorylation downregulates RNA polymerase II activity. Therefore, the high-NaCl-induced decrease in phosphorylation of TFIIF at T389 could increase transcriptional activity (56). It is unclear, however, whether transcription of osmoprotective genes would be differentially regulated by the activity of GTF2F1/TFIIF/RAP74.

High-NaCl-induced perturbation of proteostasis.

Proteostasis refers to the biogenesis, folding, trafficking, and degradation of proteins. All phases of proteostasis apparently are affected by hypertonicity. Hypertonicity not only causes dramatic changes in translation in mammalian cells, but evidence has been emerging in C. elegans that high NaCl also induces rapid protein aggregation in vivo and that many of the genes that are essential for survival during hypertonic stress function to prevent accumulation of aggregated proteins (10). Whether hypertonicity induces comparable protein damage in mammalian cells, however, has been less clear.

Hypertonicity, including that produced by high NaCl, rapidly inhibits translation of most proteins, but not translation of osmoprotective proteins. The difference in response apparently depends on whether translation is dependent on the 5′ cap of the mRNA (which is true of most proteins) or is independent of the 5′ cap (which is true of osmoprotective proteins) (55). We find that high NaCl changes phosphorylation of five proteins involved in regulation of translation (Table 5): eukaryotic translation initiation factor (EIF) 4γ 2 (EIF4G2), EIF 4E-binding protein 1 (EIF4EBP1)/PHAS-1, EIF3B, HSPB1, and RPTOR/Raptor.

Table 5.

Perturbation of proteostasis

Official Gene Symbol	Change, log2(fold)	Phosphorylated Amino Acid Site	Name
EIF3B	−3.08	S119	Eukaryotic translation initiation factor 3, subunit 9η, 116 kDa
SQSTM1	−3.05	T269	Sequestosome 1
EIF4EBP1	−2.57	S65	Eukaryotic translation initiation factor 4E-binding protein 1
RPTOR	−1.66	S859, S863	Raptor
USP47	−1.65	S822	Ubiquitin-specific protease 47
RNF20	−1.65	S138	Ring finger protein 20
UBR1	−1.60	T21	Ubiquitin protein ligase E3 component n-recognin 1
USP42	−1.55	S856	Ubiquitin-specific protease 42
EIF4G2	−1.31	T508	Eukaryotic translation initiation factor 4 γ2 isoform 1
PSMA5	1.01	S56	Proteasome α5 subunit
PSMF1	1.02	S153	Proteasome inhibitor subunit 1
PSMD1	1.12	T273	Proteasome 26S non-ATPase subunit 1
HSPB1	1.92	S82	Heat shock protein β1
DOCK7	1.94	S30	Dedicator of cytokinesis 7
USP8	2.22	S389	Ubiquitin-specific peptidase 8

The large high-NaCl-induced decrease in phosphorylation of EIF4EBP1 at S65 (Table 5) (48) helps us understand what determines whether hypertonicity decreases or increases translation of particular proteins. Cap-dependent translation (60) is initiated by association of the cap-binding protein eIF4E with eIF4G, which then recruits the ribosomal 43S complex to the 5′ end of mRNA. EIF4EBP1 directly interacts with eIF4E, and that interaction inhibits formation of the eIF4E-eIF4G complex, which represses translation. High-NaCl-induced dephosphorylation of EIF4EBP1 (Table 5) increases its affinity for eIF4E, which inhibits general cap-dependent translation, but not the cap-independent translation of osmoprotective proteins. Interestingly, in C. elegans, the high-NaCl-induced general inhibition of translation may provide the signal for transcription of osmoprotective genes (40).

Genome-wide RNAi screen and in vivo protein aggregation reporters identified degradation of damaged proteins as essential for the hypertonic stress response in C. elegans (10). Hypertonic stress causes loss of cellular water, cell shrinkage, elevated intracellular ionic strength, and macromolecular crowding (6). In vitro studies using simple mixtures of proteins, solutes, and artificial crowding agents have shown that high concentrations of inorganic ions, such as K⁺, Na⁺, and Cl⁻, destabilize protein secondary structure and disrupt enzyme activity and that macromolecular crowding promotes nonnative protein-protein interactions, which can lead to protein aggregate formation. There is little evidence for hypertonicity-induced protein damage similar to that in C. elegans in intact mammalian cells. However, we observe numerous high-NaCl-induced changes in phosphorylation in proteins involved in protein degradation, namely, proteins in the ubiquitin-conjugating system (SQSTM1, USP47, RNF20, UBR1, USP42, and USP8) and proteosome subunits and regulators (PSMA5, PSMF1, and PSMD1) (Table 5).

Phosphorylation of PSMD1 at T273 is increased by hypertonicity produced by sorbitol (41) and NaCl (Table 5). The increased phosphorylation results from hypertonicity-induced activation of MAPK14/p38 (41), and it contributes to an inhibition of proteosomal activity that results in an increase in ubiquitinated proteins (41). We do not know whether altered phosphorylation of the other proteins involved in ubiquitination and proteasomal activity (Table 5) affects their activity, but that seems a likely possibility, pointing to altered proteostasis. Along the same line, HSPB1/HSP27 is a ubiquitin-binding protein proposed to favor the degradation of ubiquitinated proteins (53). Stress-induced activation of MAPK14/p38 promotes phosphorylation of HSPB1/HSP27 at S82 in rabbit muscle (59) and HeLa cells (39). We used the p38 inhibitor SB-203580 to test whether MAPK14/p38 also contributes to high-NaCl-induced increase in phosphorylation of HSPB1/HSP27 at S82 in HEK 293 cells. SB-203580 inhibits high-NaCl-induced phosphorylation of HSPB1/HSP27 at S82 (Fig. 5), confirming that MAPK14/p38 does contribute to high-NaCl-induced phosphorylation of HSPB1/HSP27 at S82 in HEK 293 cells. Phosphomimetic mutation at sites including S82 promotes activity of HSPB1/HSP27 (37). We suggest that increased phosphorylation of HSPB1/HSP27 at S82, enhanced by activation of MAPK14/p38, contributes to changes in proteostasis induced by high NaCl.

High-NaCl-induced effects on metabolism of mRNA.

We were unaware of previous evidence that high NaCl affects mRNA stability or splicing other than increased stability of NFAT5 mRNA (7). However, we now find that high NaCl increases phosphorylation of several proteins involved in metabolism of RNA (Table 6). Known functions of those proteins include mRNA degradation (PARN, SMG9, DCP1A, and DCP1B), pre-mRNA splicing (SRRM2, RBMX, SNRNP70, RBM10, RBM25, SFRS16, KHDRBS1/SAM68, YBX1, SF3B2, SFPQ, ZCCHC8, and HNRNPM), and multiple effects on mRNA processing (PCBP2/hnRNP-E2). High NaCl increases phosphorylation of PCBP2/hnRNP-E2 at S189 (Table 6). Phosphorylation of PCBP2/hnRNP-E2 at S189 contributes to stabilization of PCBP2/hnRNP-E2 protein in murine myeloid cells (9). Phosphorylation at S189 is catalyzed by MAPK1/ERK1/2, which is, in turn, activated by ABL1/c-ABL. High NaCl is known to activate MAPK1/ERK1/2 (34) and ABL1/c-ABL (24), suggesting a pathway by which high NaCl may increase phosphorylation of PCBP2/hnRNP-E2. Other than PCBP2/hnRNP-E2, we do not have information on the effects of site-specific phosphorylation of the proteins involved in metabolism of mRNA. Nevertheless, the results suggest that there may be previously unreported effects of high NaCl on mRNA, including effects on splicing and stability.

Table 6.

Metabolism of mRNA

Official Gene Symbol	Change, log2(fold)	Phosphorylated Amino Acid Site	Name
SRRM2	−2.28	S1541, S1552	Splicing coactivator subunit SRm300
	1.35	S2272
	1.71	S2118
	2.08	T2289
PARN2	−2.08	S102	Poly(A)-specific ribonuclease (deadenylation nuclease) isoform 2
SMG9	−2.04	S32	Hypothetical protein LOC56006
RBMX	−1.99	S328	RNA-binding motif protein, X-linked
SNRNP70	−1.99	S410	U1 small nuclear ribonucleoprotein 70 kDa
HNRNPUL1	−1.83	S718	Heterogeneous nuclear ribonucleoprotein U-like 1 isoform a
RBM10	−1.66	S89	RNA-binding motif protein 10 isoform 1
DCP1A	−1.62	S525, T531	DCP1-decapping enzyme homolog A
RBM25	−1.61	S677	RNA-binding motif protein 25
SFRS16	−1.55	S547	Splicing factor, arginine/serine-rich 16
KHDRBS1	−1.04	S20	KH domain-containing, RNA-binding, signal transduction-associated 1
YBX1	−1.02	S165	Nuclease-sensitive element-binding protein 1
SF3B2	1.00	S307, S309	Splicing factor 3B subunit 2
SFPQ	1.11	S273	Splicing factor proline/glutamine-rich (polypyrimidine tract-binding protein-associated)
ZCCHC8	1.35	S658	Zinc finger CCHC domain-containing protein 8
PCBP2	1.62	S189	Poly(rC)-binding protein 2 isoform d
DCP1B	1.76	S275	Decapping enzyme Dcp1b
ADAR	3.18	T601	Adenosine deaminase, RNA-specific isoform a
HNRNPM	4.06	S579	Heterogeneous nuclear ribonucleoprotein M isoform b

High-NaCl-induced effects on cell death.

Acute elevation of NaCl beyond a threshold that depends on the type of cell induces cell death by apoptosis (6). In the present study we raised NaCl to a level that HEK 293 cells survive. Nevertheless, phosphorylation changed in a number of proteins involved in cell death (Table 4), including DPF2/requiem, DNM1L/DRP1, TP53BP2, PML, PAWR, DAXX, CTNNB1, SLTM, SQSTM1, BAG3, SHKBP1, HSPB1, ACIN1, BCL7C and RBM25. Also, JUN/c-Jun (Table 7) is involved in cell death. Some of the specific phosphorylations were previously observed to be proapoptotic, and others were observed to be antiapoptotic. In what follows we analyze the apparently pro- and antiapoptotic changes, since their balance presumably determines cell survival.

Table 7.

Cell death

Official Gene Symbol	Change, log2(fold)	Phosphorylated Amino Acid Site	Name
BCL7C	−4.23	S126	B-cell CLL/lymphoma 7C
SQSTM1	−3.05	T269	Sequestosome 1
ACIN1	−2.43	S240	Apoptotic chromatin condensation inducer 1
SHKBP1	−2.37	S587	SH3-domain kinase-binding protein 1 isoform a
SLTM	−2.00	S1002	Modulator of estrogen-induced transcription isoform a
PML	−1.69	S518, S527	Promyelocytic leukemia protein isoform 1
DAXX	−1.63	S683	Death domain-associated protein isoform b
RBM25	−1.61	S677	RNA-binding motif protein 25
DPF2	−1.39	S142	Zinc finger protein ubi-d4
BAG3	−1.29	S284, S289	BCL2-associated athanogene 3
CTNNB1	−1.22	S552	Catenin (cadherin-associated protein), β1, 88 kDa
HSPB1	1.92	S82	Heat shock protein-β1
PAWR	1.95	T230	PRKC, apoptosis, WT1, regulator
DNM1L	2.05	S616	Dynamin 1-like protein isoform 1
HTT	2.47	S1876	Huntingtin
TP53BP2	2.90	S562	Tumor protein p53-binding protein, 2 isoform 2

High NaCl depolarizes mitochondria and promotes their fission, but the threshold level of NaCl at which that occurs and whether apoptosis occurs at that level of NaCl differ between cell types (6). In the present study we find that high NaCl increases phosphorylation of DNM1L/DRP1 at S616 (Table 7). DNM1L/DRP1 contributes to fission of mitochondria during apoptosis (42). Fission of mitochondria following depolarization depends on DNM1L/DRP1 (33). Phosphorylation of DNM1L/DRP1 at S616, which we find is induced by high NaCl, promotes mitochondrial fission (36). Thus a level of NaCl less than that which causes apoptosis in HEK 293 cells still increases phosphorylation of DNM1L/DRP1 at S616 that is associated with apoptosis.

High NaCl decreases phosphorylation of PML at S518/S527 (Table 7) and JUN/c-Jun at S243 (Table 4). Activation of JUN/c-Jun by PML promotes UV-induced apoptosis (58), and phosphorylation of PML at S518 in response to hypoxia promotes KLHL20-mediated PML destruction (72). If the decreased phosphorylation of JUN/c-Jun and PML that is caused by high NaCl has effects opposite to the previously observed increases, the net result would be antiapoptotic. We do not know whether PML protein is reduced by high NaCl, but, if it is, that would help explain how HEK 293 cells survive high NaCl up to the level in the present study. High NaCl inhibits CDK1 (16), and CDK1 phosphorylates PML at S518 (72). Thus inhibition of CDK1 by high NaCl explains how high NaCl reduces phosphorylation of PML at S518. With regard to JUN/c-Jun, phosphorylation of JUN/c-Jun at S243 primes JUN/c-Jun for phosphorylation of JUN/c-Jun at T239 by GSK3B, and phosphorylation of JUN/c-Jun at S243 creates a high-affinity binding site for FBXW7 E3 ligase. FBXW7 targets JUN/c-Jun for polyubiquitination and proteasomal degradation (67). High NaCl inhibits GSK3B (76). Thus high-NaCl-induced decrease in phosphorylation of JUN/c-Jun at S243 and inhibition of GSK3B could increase JUN/c-Jun protein expression by preventing its FBXW7-mediated degradation. Considering these possibilities, it would be of interest to know the effect of high NaCl on protein expression of JUN/c-Jun and PML.

High NaCl decreases phosphorylation of CTNNB1/-β-catenin at S552 (Table 7). Since phosphorylation of CTNNB1/-β-catenin at S552 increases its transcriptional coregulatory activity (49), decreased phosphorylation could decrease the activity. Decreased activity of CTNNB1/β-catenin could reduce DDK1, which is a transcriptional target of factors regulated by CTNNB1/β-catenin (49). DDK1 is proapoptotic (27). Thus high-NaCl-induced decrease in phosphorylation of CTNNB1/β-catenin at S552 could enhance cell survival by decreasing DDK1. In this respect, it would be of interest to know the effect of high NaCl on protein expression of DDK1.

High NaCl increases phosphorylation of HSPB1/Hsp27 at S82 (Table 7). Hypertonicity-induced phosphorylation of Hsp27 at S82 was previously reported to be catalyzed by activation of MAPK14/p38 in rat brain slices (51) and human epidermal keratinocytes (35). We confirm that MAPK14/p38 is responsible for the high-NaCl-induced phosphorylation of HSPB1/Hsp27 at S82 in HEK 293 cells, finding that inhibition of MAPK14/p38 by SB-203580 prevents the phosphorylation (Fig. 5). Considering that phosphorylation of HSPB1/Hsp27 at S82 protects cells from apoptosis (38), we suggest that that the phosphorylation contributes to survival of HEK 293 cells when osmolality is increased to 500 mosmol/kg by addition of NaCl. In addition to its prosurvival effect, HSPB1/Hsp27 is also involved in cell cycle progression, RNA metabolism, proteostasis, and cytoskeletal organization (38), all of which are affected by high NaCl (see above), so its phosphorylation at S82 could be involved as well in perturbation of those functions by high NaCl.

Perspective.

By combining stable isotope labeling of proteins, multidimensional LC, and high-resolution MS, we performed quantitative global analysis of high-NaCl-induced changes in phosphorylation of proteins. To our knowledge, our work represents the first global analysis of high-NaCl-induced changes in phosphorylation of particular amino acids in proteins. We quantified 7,362 unique phosphopeptides. High salt significantly changed phosphorylation in 324 peptides contained in 253 different proteins. We further analyzed how the specific changes in phosphorylation might be involved in known effects of high NaCl, including altered cell cycle, cyto/nucleoskeletal organization, DNA DSBs, transcription, proteostasis, metabolism of mRNA, and cell death.

Almost all the phosphorylation sites that we found were identified in previous phosphoproteomic screens. Despite so many previous identifications, relatively little has been reported that indicates the kinases and phosphatases involved and the functional consequences of the changes in phosphorylation. From the limited information that is available, we have extrapolated pathways by which high NaCl may alter the specific phosphorylations and the possible functional consequences. A limitation is that most of the studies from which we extrapolate were conducted with cells other than the HEK 293 cells that we used, so we cannot be sure whether certain of the effects depend on cell type. Also, only a few of the previous studies investigated hypertonicity. Despite these limitations, our analysis suggests previously unappreciated possibilities and generates hypotheses for future investigation.

GRANTS

This work was funded by the Division of Intramural Research, National Heart, Lung, and Blood Institute.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

R.W., J.D.F., M.G., and M.B.B. are responsible for conception and design of the research; R.W., Y.I., N.D., K.R., and G.W. performed the experiments; R.W., J.D.F., Y.I., G.W., and M.B.B. analyzed the data; R.W., J.D.F., and M.B.B. interpreted the results of the experiments; R.W. and M.B.B. prepared the figures; R.W. and M.B.B. drafted the manuscript; R.W., J.D.F., Y.I., G.W., M.G., and M.B.B. approved the final version of the manuscript; J.D.F. and M.B.B. edited and revised the manuscript.