Proteomic analysis of high NaCl-induced changes in abundance of nuclear proteins

Abstract

Mammalian cells are normally stressed by high interstitial NaCl in the renal medulla and by lesser elevation of NaCl in several other tissues. High NaCl damages proteins and DNA and can kill cells. Known protective responses include nuclear translocation of the transcription factor NFAT5 and other proteins. In order better to understand the extent and significance of changes in nuclear protein abundance, we extracted nuclear and cytoplasmic proteins separately from HEK293 cells and measured by LC-MS/MS (iTRAQ) changes of abundance of proteins in the extracts in response to high NaCl at three time points: 1 h, 8 h, and adapted for two passages. We confidently identified a total of 3,190 proteins; 163 proteins changed significantly at least at one time point in the nucleus. We discerned the biological significance of the changes by Gene Ontology and protein network analysis. Proteins that change in the nucleus include ones involved in protein folding and localization, microtubule-based process, regulation of cell death, cytoskeleton organization, DNA metabolic process, RNA processing, and cell cycle. Among striking changes in the nucleus, we found a decrease of all six 14-3-3 isoforms; dynamic changes of “cytoskeletal” proteins, suggestive of nucleoskeletal reorganization; rapid decrease of tubulins; and dynamic changes of heat shock proteins. Identification of these changes of nuclear protein abundance enhances our understanding of high NaCl-induced cellular stress, and provides leads to previously unknown damages and protective responses.

cells respond in a complex manner to the hypertonicity that is produced by high NaCl. Hypertonicity perturbs cells, including cell cycle arrest, apoptosis, DNA damage, oxidative stress, inhibition of transcription and translation, and mitochondrial depolarization (6). Nevertheless, cells survive and function in hypertonic environments. NFAT5 (22, 24) is a transcription factor that contributes to the protective response. In a normotonic environment it is distributed between cytoplasm and nucleus. However, when activated by hypertonicity, NFAT5 translocates from the cytoplasm to nucleus and regulates downstream genes, including transporters and enzymes involved in accumulation of protective organic osmolytes and heat shock proteins (HSPs). The present studies were aimed at determining what additional proteins change in the nucleus because of hypertonicity, what their roles are in the response to hypertonicity, and what signaling pathways are involved.

Quantitative proteomics has emerged as an effective tool to reveal global changes in protein abundance (3). High-throughput analysis and highly sensitive mass spectrometers allow us to identify and quantify thousands of proteins from a cellular sample. Quantization of proteins in subcellular fractions can determine changes in abundance of proteins in individual cell compartments and can provide information about protein trafficking between compartments (4). Here we used iTRAQ-based quantitative subcellular proteomics to provide an insight into the time course of changes of protein abundance in cytoplasm and nucleus when cells experience high NaCl. Finding that high NaCl changes abundance of 163 proteins in nucleus at least at one time point, we consider the biological significance of those changes.

MATERIALS AND METHODS

Cell culture and treatment.

Low-passage (44–45) HEK293 cells were grown at 300 mosmol/kg in Eagle's minimal essential medium (American Type Culture Collection, Manassas, VA) with 10% fetal bovine serum and 4 mM glutamine added at 37°C incubator in 5% CO₂. Medium bathing the HEK293 cells was changed to elevate it from 300 to 500 mosmol/kg (NaCl added) for 1 h, 8 h, or for 2 passages (“adapted”). The corresponding three controls were handled the same but kept at 300 mosmol/kg.

Preparation of cytoplasmic and nuclear extracts for mass spectrometry.

The HEK293 cells were scraped off 10 cm dishes. Cytoplasmic and nuclear proteins were separated with NE-PER (Pierce, Rockford, IL) according to the supplier's manual. HALT protease and phosphatase inhibitor (Pierce) was added to the reagents immediately before use. The proteins were precipitated by a mixture of methanol-chloroform-water, as previously described, to remove salts and detergents (1). Protein pellets were resuspended in 40 μl of 3 M urea containing 1 M triethyl ammonium bicarbonate. An aliquot of 90 μg of protein was reduced with 2.5 mM Tris(2-carboxyethyl)phosphine and then alkylated with 12.5 mM iodoacetamide. The solution containing 90 μg of protein was diluted with water to reduce urea concentration to 1 M and digested with 5 μg trypsin gold (Promega, Madison, WI) overnight at 37°C. The peptides were labeled differentially with one of the isobaric tags (113–121) from an iTRAQ 8-plex kit (AB Sciex, Framingham, MA). Six cytoplasmic and six nuclear samples (three time points and their controls) were combined.

Western blot.

To confirm that cytoplasmic and nuclear proteins were accurately separated we ran them on a NuPAGE 4–12% Bis-Tris gels (Invitrogen) and electrophoretically transferred them to nitrocellulose membranes (0.45 μm pore size, Invitrogen). The membranes were blocked for 1 h in Odyssey blocking buffer (LI-COR Biosciences, Lincoln, NE) to minimize nonspecific binding and then were incubated with mouse anti-BRG1 (Santa Cruz Biotechnologies, Santa Cruz, CA) or rabbit anti-GAPDH or rabbit anti-CREB (Cell Signaling Technology, Danvers, MA) antibody overnight at 4°C. After washing with 0.1% Tween in PBS three times, the membranes were incubated with Alexa Fluor 680 goat anti-mouse IgG or Alexa Fluor 780 or 800 goat anti-rabbit IgG (Molecular Probes, Grand Island, NY) for 1 h in the dark at room temperature. After washing with 0.1% Tween in PBS, the membranes were visualized on an Odyssey Infrared Imager (LI-COR Biosciences). We performed Western analysis of other proteins using antibodies against Ku80/XRCC5 (Cell Signaling), Calnexin/CANX (Enzo), 14-3-3 θ/YWHAQ (Santa Cruz), HSP 90β/HSP90AB1 (Santa Cruz), and Annexin II/ANXA2 (Santa Cruz).

Analysis of high NaCl-induced alternate splicing of NFAT5.

HEK293 cells were incubated at 300 or 500 mosmol/kg for 24 h. Total RNA was extracted with RNeasy mini kit (Qiagen, Valencia, CA) and reverse transcribed to cDNA using High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA) according to the supplier's manual. We used 25 ng cDNA for real-time PCR with the QuantiFast SYBR Green PCR Kit (Qiagen). The relative amounts of mRNAs were determined based on the threshold cycle for PCR product (Ct). Primers (sequences in Table 1) match those previously published (15). The primers produced single amplicons whose size was confirmed by gel electrophoresis.

Table 1.

NCBI Designations2	Designations in Ref. 15	Forward Primer Sequence (5′–3′)	Reverse primer sequence (5′–3′)
E7/E8	total NFAT5	AGTGGACATTGAAGGCACTAC	TTGGAACCAGCAATTCCTATTC
E3/E5	E3/E6	AGCTGTTGTTGCTGCTGATGC*	CATAGCCTTGCTGTCGGTGAC
E4/E5	E5/E6	GATTTGCCTCTGAAGCAGGG	CATAGCCTTGCTGTCGGTGAC

^*This primer actually overlaps exon 3 and 5.

Strong cation exchange chromatography.

The peptides were prefractionated via strong cation exchange chromatography on a PolySULFOETHYL A column with a dimension of 2.1 × 200 mm and packing size of 5 μm (PolyLC, Columbia, MD) on an Agilent 1200 HPLC (Agilent Technologies, Palo Alto, CA). A 60 min gradient (0–10% B for 5 min, 10–40% B for 35 min, 40–50% B for 5 min, 100% B held for 5 min, and 0% B held for 10 min) was used to separate the peptides (buffer A: 10 mM KH₂PO₄/25%ACN, pH 2.67; buffer B: 10 mM KH₂PO₄/500 mM KCl/25% ACN, pH 2.67). Flow rate was 200 μl/min. Fractions were collected at 1 min interval on a 96-well microtiter plate for a total of 60 fractions. The chromatographic peaks were monitored using the built-in UV detector (214 nm), and the number of fractions was reduced to 25 by combining fractions.

LC-MS/MS.

Each of the 25 peptide fractions was separated on a BetaBasic C18 PicoFrit column (75 μm ID, 10 cm length, tip 15 μm; New Objective, MA) on an Eksigent nanoLC Ultra coupled with an LTQ-Orbitrap Velos mass spectrometer (Thermo Scientific). The trapping column is Zorbax 300SB-C18 (5 μm, 5 × 0.3 mm, Agilent). Peptides were separated using a linear gradient from 5%-35% buffer B in 40 min (buffer A: 0.1% FA in water; buffer B: 0.1% FA in ACN). MS2 spectra were acquired by data-dependent mode. An Orbitrap full MS scan (resolution: 3 × 10⁴, mass range 300–2,000 Da) was followed by six MS2 scans (resolution: 7.5 × 10³, mass range 105–2,000 Da), which were performed via HCD on the top six most abundant ions. Isolation window for ion selection was 3 Da. Normalized collision energy was set at 45%. Dynamic exclusion time was 20 s (± 7 ppm relative to the precursor ion m/z). The entire experiment was performed three times.

Database search and statistical analysis.

Mascot database search was carried out via Proteome Discoverer (Thermo Scientific) against the current Sprot human database, specifying trypsin with two enzymatic missed cleavages allowed. Precursor mass tolerance was 50 ppm, while that for fragment mass was 0.05 Da. Carbamidomethyl cysteine was indicated as a static modification while oxidation of methionine, and iTRAQ8plex labeling of lysine, tyrosine, and peptide NH₂ terminus were indicated as dynamic modifications. Peptides were filtered by Mascot significance threshold (<0.05) and false discovery rate (<1%, obtained by decoy database search). Peptides shared between proteins were excluded, so each protein is identified solely by unique peptides. Median iTRAQ reporter ion intensity was used to normalize all the reporter ion intensities in the same channel. Natural log ratios of peptides between treatment (500 mosmol/kg) and control (300 mosmol/kg), weighted by ion intensity, were calculated for each replicate using the iTRAQ reporter ion intensities. The average of the weighted log ratio for each protein was pooled across the three biological replicates; proteins with fewer than three peptide identifications were excluded from the analysis. The within-protein variance was obtained from the average of the weighted peptides across replicates while between-protein variance was calculated at the protein-to-protein level. The total variance for each protein is the combination of the within-protein and between-protein variance. A Gaussian distribution and a two-tailed t-test were used to transform the weighted ratios to ζ_ij values (Eq. 2). The transformed ζ values were normalized using the inverse of the standard normal cumulative distribution. We then used an optimized FDR approach to obtain q values <0.05 on a protein-to-protein basis. Supplementary Table S1 contains the data for all of the proteins that were identified.¹

For protein i under condition j,

$${\text{stat}}_{ij}=\frac{\text{average}{\left(\log W{R}_{ij,l}\right)}_{ij}}{\sqrt{\text{variance}{\left(\log W{R}_{ij,l}\right)}_{ij}/\left(n_{ij}-1\right)}},1,2,\dots ,n_{ij}and{n}_{ij}\ge 3,$$ (Eq. 1)

n_ij is the total number of peptides matched to the protein i across the three replicates of an experiment.

$${\zeta }_{ij}={\Phi }^{-1}\left(\text{Probability}\left(\text{t}<{\text{stat}}_{\text{ij}}\right)\right)$$ (Eq. 2)

RESULTS

Extraction of nuclear and cytoplasmic proteins.

Osmolality bathing HEK293 cells grown at 300 mosmol/kg was raised to 500 mosmol /kg by adding NaCl for 1 h, 8 h, or two passages (adapted). Corresponding controls were handled the same way, except kept at 300 mosmol/kg. Cells from all six conditions were fractionated to provide cytoplasmic (CE) and nuclear (NE) protein extracts. The accuracy and consistency of the subcellular fractionation is demonstrated by results with a cytoplasmic (GAPDH) and two nuclear (BRG1 and CREB) markers. GAPDH is consistently enriched in cytoplasmic extracts, while BRG1 and CREB are present only in the nuclear extracts (Fig. 1A).

Fig. 1.

A: Western blot of cytoplasmic (GAPDH) and nuclear (CREB, BRG1) marker proteins, showing that nuclear and cytoplasmic proteins were separately extracted. B: comparison of natural log of reporter ion abundances between biological replicate 2 and 3 of cytoplasmic samples from cells adapted to high NaCl. The high correlation is representative of the good agreement between biological replicates. C: distribution of high NaCl-induced changes in abundance of cytoplasmic and nuclear proteins at 1 h, 8 h, and in adapted cells. D: numbers of cytoplasmic proteins identified to increase, decrease, or remain the same after NaCl was increased. E: numbers of nuclear proteins identified to increase, decrease, or remain the same after NaCl was increased. F: Venn diagram of proteins whose abundance changed significantly in the cytoplasm and/or nucleus. G: Venn diagram of proteins whose abundance changed significantly in the nucleus in response to high NaCl at 1 h, 8 h and/or in adapted cells.

Protein identifications.

Three biological replicates were analyzed. Peptides produced by trypsin digestion of nuclear and cytoplasmic extracts at each time point (six samples) from each biological replicate were labeled by iTRAQ and combined for analysis by LC-MS/MS. The results correlate well between the three biological replicates. Pearson correlation coefficients between any two replicates range from 0.52 to 0.80 in 36 comparisons of three replicates of six samples in nucleus and cytoplasm. A representative comparison is shown in Fig. 1B. We identified 3,496 proteins from CE and 3,058 from NE. Results were further filtered to include only proteins identified by more than two unique peptides. After the additional filtering, we identify 2,335 proteins from CE and 2,074 from NE. All together, we identified 3,190 unique proteins (Supplementary Table S1). Figure 1C demonstrates that the observed changes in protein abundance in cytoplasm and nucleus at all times are normally distributed and are centered close to zero with standard deviations (SD) ranging from 0.12 to 0.23. Cells typically express more than twice the number of proteins that we detected. We presumably failed to detect many of the less abundant proteins, which limits the scope of our conclusions. For example, NFAT5, known to translocate to the nucleus when NaCl is elevated, was not detected, presumably because, like most transcription factors, its abundance is low.

Time course of changes of protein abundance.

We observed both increases and decreases of protein abundance in the cytoplasm (Fig. 1D) and nucleus (Fig. 1E). More proteins changed in the cytoplasm after the cells adapted to high NaCl than at earlier times (Fig. 1D), but fewer changed in the nucleus after adaptation (Fig. 1E). We cannot exclude the possibility that protein degradation produced fragments that were present in extracts prior to trypsin digestion and therefore do not represent intact proteins. Most proteins are too large to diffuse into the nucleus from the cytoplasm through nuclear pores. However, it is possible that such fragments could diffuse into the nucleus because of their small size.

High NaCl-induced changes of abundance of proteins in nucleus and cytoplasm 1 h after NaCl is elevated.

Figure 1F shows the number of proteins whose abundance changed significantly in the cytoplasm or nucleus after 1 h in response to high NaCl; 228 unique proteins changed significantly in one compartment or the other, 109 proteins in cytoplasm, and 163 proteins in nucleus (Supplementary Table S2). Seventeen proteins change in the same direction in both the cytoplasm and nucleus; 73 proteins changed significantly in the nucleus but not in the cytoplasm; 25 changed significantly only in the cytoplasm (Supplementary Table S1). The abundance of a protein in the nucleus could change either because it moves into or out of the nucleus from the cytoplasm or because the total cellular abundance of the protein has changed without redistribution. Translocation is a more likely contributor at 1 h for the latter 73 proteins because we expect that nuclear translocation causes a protein to change relatively less in the cytoplasm than in the nucleus because the cytoplasmic volume is so much greater. After NaCl is elevated only 44 out of 109 proteins whose abundance changes significantly in the nucleus also change significantly in the cytoplasm (Fig. 1F). It is interesting that nuclear abundance of more proteins changed 1 h after NaCl was elevated than at later times and that abundance of only 18 nuclear proteins changed at all of the times, pointing to the dynamic nature of the stress response (Fig. 1G).

Gene Ontology analysis of proteins whose abundance changes in the nucleus in response to high NaCl.

We performed Gene Ontology analysis (20) of biological processes overrepresented among the 163 proteins whose abundance changed in the nucleus in response to high NaCl at one time or another. Many of the proteins are involved in protein folding and localization, microtubule-based process, regulation of cell death, cytoskeleton organization, DNA metabolic process, RNA processing, and cell cycle (Fig. 2A). We also performed GO analysis of the molecular function of proteins whose abundance changed in the nucleus in response to high NaCl. The identified proteins are involved in binding to unfolded proteins, cytoskeletal proteins, nucleotides or RNA, are involved in helicase activity, or are structural constituents of cytoskeleton (Fig. 2B). Many of the biological processes were previously known to be affected by high NaCl (6), and the dynamic changes of nuclear proteins helps to understand the basis for the effects.

Fig. 2.

Gene Ontology analysis of the 163 proteins whose abundance was changed significantly in the nucleus by high NaCl, grouped by biological process (A) and molecular function (B). Data were analyzed with the Gene Ontology tool (DAVID). The P values displayed for each time point indicate significance of enrichment for that process.

High NaCl induced changes in the nucleus of “cytoskeletal” proteins (Fig. 3A). At first it may seem surprising that so many proteins whose abundance in the nucleus is affected by high NaCl are designated as cytoskeletal (Fig. 2A). However, it is now recognized that there is a nucleoskeleton that contains many cytoskeletal proteins (10). Most of the nucleoskeletal proteins that change significantly increase at 1 h after NaCl is elevated, at 8 h all are lower than at 1 h, but in adapted cells two distinct clusters are apparent (Fig. 3A). In one cluster the nucleoskeletal proteins are higher in adapted cells at 500 than at 300 mosmol/kg. Those proteins include myosins (MYH9 and MYH11), filamins (FNLA and FNLC), vimentin (VIM), and desmoplakin (DSP). Proteins in the other cluster are less elevated in adapted cells than at 1 h. Those proteins include actin (ACTB), actinin (ACTN4), spectrins (SPTAN1, SBTBN1, and SPTBN4), tropomyosins (TPM1, TPM3, and TPM4), and radixin (RDX).

Fig. 3.

Natural log of ratios of high NaCl-induced (500/300 mosmol/kg) changes in abundance of proteins in the nucleus or cytoplasm. A: nucleoskeletal proteins. B: nuclear tubulins. C: nuclear and cytoplasmic heat shock proteins. D: 14-3-3 isoforms. E: 14-3-3 protein interaction network generated with IPA (Ingenuity Systems, http://www.ingenuity.com). F: RNA processing proteins. G: alternate splicing of NFAT5 induced by high NaCl. H: proteins related to DNA metabolic processes. Official gene names are given.

Tubulins are cytoskeletal proteins that also occur in the nucleus. High NaCl changes the abundance of α- and β-tubulins in the nucleus (Fig. 3B). Nuclear tubulins decrease within 1 h after NaCl is raised, probably associated with the high NaCl-induced disruption of the cell cycle (23, 28). When the cells have become adapted to high NaCl and the cell cycle is restored, nuclear tubulins return closer to the level before NaCl was elevated, despite NaCl remaining high.

High NaCl-induced changes of HSPs.

HSPA1A, HSPA1L, and HSPA2, which are HSP70s, increase in the nucleus, but not in the cytoplasm, within 1 h after NaCl is elevated but do not remain significantly elevated in adapted cells (Fig. 3C). The increase at 1 h in the nucleus, coupled with a decrease in the cytoplasm, suggests nuclear translocation. Constitutively expressed (“cognate”) HSPs HSPA9, HSPA8, HSPA5, HSP90B1, and HSPD1 mostly remain elevated in cytoplasm and/or nucleus at all times after NaCl is increased (Fig. 3C). HSP90AB1 and HSP90AA1 decreased significantly in the nucleus at 1 h but did not change at later times.

High NaCl-induced changes in the nucleus of 14-3-3 proteins.

Six 14-3-3 isoforms decrease significantly in the nucleus within 1 h after NaCl is elevated, remain low at 8 h, and then return to the 300 mosmol/kg control level after the cells adapt to high NaCl (Fig. 3D). The six 14-3-3 isoforms that change significantly are YWHAQ, YWHAB, YWHAG, YWHAE, YWHAZ, and YWHAH. 14-3-3 proteins heterodimerize creating multiple protein-protein interactions among the isoforms (32). The network diagram in Fig. 3E depicts interactions of these 14-3-3 proteins with each other and with other proteins.

High NaCl-induced changes in the nucleus of proteins that process RNA.

Nuclear abundance of a total of 25 proteins involved in RNA processing changes significantly at one time or another after NaCl is increased (Fig. 2A). Changes of individual proteins in the nucleus follow different temporal patterns (Fig. 3F). The proteins that change significantly in the nucleus at one time or another after NaCl is elevated include heterogeneous nuclear ribonucleoproteins (hnRNPs)(SYNCRIP, HNRNPH1, HNRNPF, HNRNPC, HNRNPA2B1, HNRPDL, HNRNPA3, HNRNPD, HNRNPA1, HNRNPM, and PTBP1), which complex with heterogeneous nuclear RNA. The proteins also include RNA helicases (DDX17, DHX9, and DDX5). DDX5 is elevated at 1 and 8 h, but DHX9 and DDX17 are increased only in adapted cells. DDX5 and DDX17 are of interest because of their role in regulation of promigratory transcription by NFAT5 (15), the transcription factor whose role in protection of cells against hypertonicity had originally lead to its identification (22, 24). In the context of cell migration DDX5 and DDX17 cause alternate splicing that increases the inclusion of exon 5² of NFAT5. Since exon 5² contains a premature translation termination codon, its inclusion lowers NFAT5 mRNA via the nonsense-mediated mRNA decay pathway, resulting in less NFAT5 protein. The previous study (15) does not investigate the role of DDX5 and DDX17 in splicing of NFAT5 in response to hypertonicity. Therefore, we measured the effect of high NaCl on abundance of the alternately spliced variants of NFAT5 (Fig. 3G). As previously observed, high NaCl increases the abundance of NFAT5 mRNA that includes all splice variants (Fig. 3G, “Total”). In contrast, high NaCl reduces abundance of the alternately spliced NFAT5 mRNA variant that includes exon 5² (Fig. 3G, PCR primers in exon 5 and exon 6), but not the variant that does not necessarily contain exon 5 (Fig. 3G, primers in exon 3 and exon 6). These results are consistent with DDX5 and DDX17 promoting alternative splicing of NFAT5 in the context of hypertonicity, as well as cell migration.

High NaCl-induced changes in nuclear proteins involved in DNA metabolic process.

High NaCl affects nuclear abundance of 17 proteins involved in “DNA metabolic process” (Fig. 2A). Figure 3H depicts the time course of high NaCl-induced changes in the abundance of those proteins. PKRDC (DNA-PKca), XRCC5 (Ku80), and XRCC6 (Ku70) all increase in the nucleus within 1 h. They are the components of the DNA-dependent protein kinase (DNA-PK) complex that is involved in sensing and repair of DNA double strand breaks. PPIA (a peptidyl-prolyl cis-trans isomerase), DNMT1 (a DNA methyltransferase), and MCM5/6/7 (mini-chromosome maintenance proteins) all decrease within 1 h after NaCl rises but rise closer to their original level at later times. MDC1 (mediator of DNA damage checkpoint 1), NONO (an RNA- and DNA-binding protein) and SFPQ (a splicing factor that heterodimerizes with NONO) all decrease, but only after cells become adapted to high NaCl. In contrast, APEX1 (a DNA repair enzyme) and TOPA2 (a topoisomerase) both rise in adapted cells.

Western analysis.

We measured by semiquantitative Western analysis the nuclear abundance of some of the proteins that changed significantly, at one time point or another, as measured by mass spectrometry when NaCl was elevated and against which we had antibodies available. The ratio of abundance at 500 vs. 300 mosmol/kg correlates well, comparing results from the two methods (Fig. 4). Incidentally, standard error/mean is 1.9 times greater for the Western analysis than for the mass spectrometry, indicating greater reproducibility of the mass spectrometry. We conclude that the mass spectrometry and Western analysis give largely equivalent results.

Fig. 4.

A: Western blot of 5 proteins showing their nuclear abundances at 3 different time points at 300 and 500 mosmol/kg (NaCl added). B: correlation (R²) of the changes in protein abundance measured by mass spectrometry vs. Western analysis is 0.77.

DISCUSSION

iTRAQ.

Our use of iTRAQ made it practical to track the time course of the global changes in abundance of nuclear and cytoplasmic proteins induced by high NaCl. However, iTRAQ is limited by precursor interference which causes systematic underestimation of observed differences (31). This is not what we observed in our data when we compared the protein abundance change by mass spectrometry with that by Western blot. Since the two methods produced consistent results, we do not find evidence for systematic suppression of the protein abundance change by iTRAQ.

Nucleoskeletal reorganization.

Hypertonicity causes major rearrangement of the cytoskeleton (19). Cell shrinkage increases cortical F-actin and modulates F-actin associated cytoskeletal proteins in the cortical region. In the present study we find that high NaCl also alters the abundance of nucleoskeletal proteins (Fig. 3A), which leads us to propose that hypertonicity induces nucleoskeletal reorganization as well as cytoskeletal reorganization. Osmotically induced changes in whole cell volume have been studied extensively, but we are unaware of any studies that specifically measure the volume of the nucleus, so we do not know whether hypertonicity shrinks the nucleus or whether there is regulatory nuclear volume increase like the regulatory cell volume increase that occurs. The large size of nuclear pores make osmosis between the nucleus and the rest of the cell unlikely, so it is difficult to guess what besides nucleoskeletal changes could modulate the volume of the nucleus. Nucleoskeletal proteins are known to not only determine the shape and mechanical properties of the nucleus, but to also interact and contribute to discrete nuclear functions. Thus, nuclear actin colocalizes with α-actinin, tropomyosin, spectrin, and filamin in HeLa cells (11). Actin interacts with the nuclear transcription apparatus, with chromatin-remodeling complexes, and with all three nuclear RNA polymerases. Nuclear actin contributes to regulation of gene transcription, to chromatin remodeling, to RNA processing, and to nuclear export . Nuclear actin-myosin interactions are involved in the transition of the initiation complex into the elongation complex, by triggering a structural change of the transcription apparatus or by generating force that supports RNA polymerase movement (17). The rapid increase of actin and its binding proteins in the nucleus 1 h after NaCl is raised suggests they may not only have mechanical functions but may be involved in the rapid changes of transcription and RNA processing that occur.

The abundance of nuclear tubulins decreases within 1 h after NaCl is elevated (Fig. 3B) and then increases toward or to the original level by the time the cells adapt. Tubulins are involved in mitosis, which is rapidly arrested after NaCl is elevated, and resumes as the cells become adapted. The abundance of nuclear tubulins follows the same time course. Nuclear tubulins are not only involved in mitosis, but in other processes, as well. For example, association between free β-tubulins and histone H3 in the nucleus prevents H3 from interacting with the lamin B receptor and heterochromatin protein 1, leading to reduced cellular proliferation (2). Nuclear tubulins are also involved in signaling by Notch (NOTCH1). Notch regulates cell cycle, differentiation, and apoptosis. Notch signaling is activated through the cleavage of Notch receptors, resulting in release and translocation into the nucleus of the Notch1 receptor intracellular domain (N1IC), the activated form of the Notch1 receptor. Nuclear α- and β-tubulins modulate Notch signaling by associating with N1IC (33).

HSPs.

HSPs are highly conserved proteins that are expressed constitutively and/or induced by stress. Constitutively expressed (“cognate”) HSPs participate in protein folding and assembly, elimination of misfolded proteins, and stabilization of newly synthesized protein in various intracellular compartments. Expression of inducible HSPs is evoked by various stresses, many of which denature proteins. The HSPs enhance cellular survival and aid cellular recovery by acting as molecular chaperones that limit or correct damage to proteins. Also, HSPs prevent apoptosis by inhibition of JNK activation, prevention of cytochrome c release from mitochondria, and disruption of apoptosomes. Hypertonicity damages proteins and detection and degradation of damaged proteins are essential mechanisms for survival under hypertonic conditions (7, 9). Hypertonicity was previously reported to increase expression of several HSPs (Ref. 6 and references therein).

In the present study we find that several HSPs increase in nucleus and/or cytoplasm within 1 h, and that some, but not all, are still elevated after the cells adapt to high NaCl (Fig. 3C). The short time is not surprising since rapid response is a hallmark of the HSP response (25). HSP70s stabilize existing proteins against aggregation and mediate the folding of newly translated proteins. The fact that HSPA1A, HSPA1L, and HSPA2, which are HSP70s, increase significantly in the nucleus, but not in the cytoplasm, within 1 h after NaCl is elevated, but do not remain significantly elevated in adapted cells is consistent with our earlier conclusion that the rapid response of HSP70s protects cells early in the response to high NaCl, before protective organic osmolytes have had time to accumulate, but is no longer necessary after the osmolytes have accumulated (30). Also, increase of HSPA1L and HSPA1A at 1 h in the nucleus, coupled with a decrease in the cytoplasm, suggests nuclear translocation. Heat shock cognate proteins, HSPA9, HSPA8, HSPA5, HSP90B1, and HSPD1, mostly remain elevated in cytoplasm and/or nucleus at all times after NaCl is increased (Fig. 3C). High NaCl increases DNA breaks, and the DNA damage persists in adapted cells as long as NaCl remains high (13). The persistent elevation of the heat shock cognate proteins after cells have adapted to high NaCl suggests that stress to proteins also continues after cells adapt to high NaCl. High NaCl increases reactive oxygen species, resulting in carbonylation of proteins, and a high level of carbonylation exists in cells adapted to the normally high NaCl in the renal medulla in vivo (34).

HSP90A proteins are highly conserved molecular chaperones that have key roles in cell cycle control, signal transduction, protein folding, regulation of transcription, and protein degradation. HSP90AA1 is the inducible form, and HSP90AB1 is the constitutive form. HSP90s contribute to regulation of >100 client proteins, including the osmoprotective transcription factor NFAT5 (8). Both HSP90As decrease transiently in the nucleus within 1 h after NaCl is elevated (Fig. 3C).

14-3-3 proteins.

YWHAE maintains hypertonicity-induced autoactivation of MAP3K3 (MEKK3) by binding to it at S526, which protects this site from dephosphorylation by PP2A (1, 14) (Fig. 3, D and E). MAP3K3 in turn activates MAPK14 (p38) via MAP2K6 (MEK6)(14). Another 14-3-3 protein, YWHAG, activates PI3K (1), which contributes to high NaCl-induced increased phosphorylation of Akt (27). All of these proteins (MAP3K3, MAP2K6, MAPK14, PI3K, and Akt) contribute to high NaCl-induced activation of the osmoprotective transcription factor NFAT5 (21, 26, 27, 35). Although the 14-3-3 proteins are not known to directly activate NFAT5, tests of this possibility seem indicated.

14-3-3 proteins are highly conserved. They are involved in multiple cellular functions including protein trafficking, signal transduction, apoptosis, and cell cycle regulation. 14-3-3 proteins bind to phospho-serines and -threonines in diverse partners, including transcription factors, biosynthetic enzymes, cytoskeletal proteins, signaling molecules, apoptosis factors, and tumor suppressors. 14-3-3 proteins form homo- and heterodimers, and each monomer can bind independently to target proteins. 14-3-3 proteins can alter the stability and catalytic activity of their ligands; can physically occlude motifs to modify accessibility of ligands to kinases, phosphatases, and proteases; and can act as a scaffold to anchor target proteins close to one another. 14-3-3 proteins aid in protein trafficking from the nucleus via the client nuclear export signal, and a 14-3-3 protein can prevent its client from reentering the nucleus by masking the ligand nuclear localization signal (5, 29, 32).

Nuclear abundance of proteins that process RNA.

Nuclear abundance of proteins that process RNA is affected by high NaCl (Fig. 3F). Since the hnRNPs (“HNRNP” in their gene symbols) shuttle between the nucleus and the cytoplasm, translocation could contribute to early changes in the nucleus. High NaCl increases nuclear HNRNPA1 within 1 h, but the increase is only transient. Hypertonicity was previously known to cause HNRNPA1 to localize to cytoplasmic stress granules (18), and depletion of HNRNPA1 affects the recovery of cells from stress. High NaCl also increases nuclear HNRNPA2B1 in adapted cells. HNRNPA2B1 is a key regulator of mRNA stability (16). We do not know at this point what role high NaCl-induced changes in nuclear HNRNPA1 and HNRNPA2B1 may have.

High NaCl also affects nuclear abundance of RNA helicases. High NaCl increases nuclear DDX5 within 1 h, but the increase is transient. High NaCl causes later increases of nuclear DHX9 and DDX17. DDX5 and DDX17 are of interest because of their role in regulation of promigratory transcription by NFAT5 (15), the transcription factor whose role in protection of cells against hypertonicity had originally lead to its identification (22, 24). In the context of cell migration DDX5 and DDX17 affect NFAT5 in two ways (15). They act as transcriptional coactivators of NFAT5, required for activating NFAT5 target genes, and, at the splicing level, they increase the inclusion of exon 5² of NFAT5. Since exon 5² contains a premature translation termination codon, its inclusion lowers NFAT5 mRNA via the nonsense-mediated mRNA decay pathway, resulting in less NFAT5 protein. This dual regulation enhances transcriptional activity of NFAT5 despite lowering NFAT5 protein abundance. The previous study (15) did not investigate the role of DDX5 and DDX17 in activation of NFAT5 by hypertonicity, but we find that high NaCl increases overall abundance of NFAT5 splice variants but reduces abundance of NFAT5 splice variants that include exon 52 (Fig. 3G), consistent with activity of DDX5 and DDX17. At this point we do not know how the dynamic changes in their nuclear abundance (Fig. 3F) affect their activity.

Some of the other RNA processing proteins whose nuclear abundance is affected by high NaCl are involved in splicing (RBMX, SNRNP200, SF1, and SF3A1) or in nucleologenesis and nucleolar function (NOLC1, HEATR1). Some are multifunctional RNA-binding proteins involved in varied aspects of RNA metabolism, such as transcription, alternative pre-mRNA splicing, mRNA localization, mRNA degradation, and DNA repair (KHSRP, QKI, NONO, and SFPQ). Finally, ADAR encodes the enzyme responsible for RNA editing by site-specific deamination of adenosines. It seems likely that the high NaCl-induced changes in nuclear abundance of these RNA processing proteins must have effects on RNA function in addition to those described here, we are unaware that this has been directly studied to any great extent and suggest that further investigation is warranted.

Proteins involved in DNA metabolic process.

High NaCl increases DNA double strand breaks (DSBs)(13) (Fig. 3H). The DNA-PK complex ordinarily localizes to aberrant DSBs and rapidly repairs them. The components of the DNA-PK complex, PKRDC (DNA-PKca), XRCC5 (Ku80), and XRCC6 (Ku70), all increase in the nucleus within 1 h after NaCl is elevated (Fig. 3H). Despite this increase of DNA-PK in the nucleus, high NaCl inhibits repair of DNA breaks, whatever their cause, and increased DSBs are not repaired as long as NaCl remains elevated (13). Some of the factors are known that inhibit DNA repair while NaCl is high (12), but our understanding is incomplete. Nuclear abundance of other DNA damage response proteins (MDC1, PARP1, APEX1, and LIG3) also changes at one time or another after NaCl is elevated. However, those changes do not provide a ready clue to the continued inhibition of DNA repair. DNMT1, whose abundance is reduced for up to 8 h after NaCl is elevated, is a DNA (cytosine-5-)-methyltransferase. That raises the possibility that high NaCl may affect methylation of DNA. MCM5, MCM6, and MCM7 are mini-chromosome maintenance proteins that are essential for the initiation of eukaryotic genome replication. They all decrease within 1 h after NaCl is increased. Their decrease presumably contributes to the rapid inhibition of mitosis that occurs when NaCl is elevated.

Perspective.

The dynamic global changes in nuclear protein abundance that we find enhance our understanding of the functional changes caused by high NaCl and they provide leads to other consequences that were previously unknown. Although high NaCl was previously known to elevate HSPs, their dynamic changes in subcellular compartments at three different time points adds to that understanding. The dynamic, coordinate change of six 14-3-3 isoforms in the nucleus suggests that their role in the response to high NaCl should be investigated further. Also, the dynamic changes of nucleoskeleton proteins suggests that hypertonicity induces nucleoskeletal, as well as cytoskeletal, remodeling with both functional and structural consequences.

GRANTS

This research was supported by the IRP of the NHLBI.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: J.L., J.D.F., and M.B.B. conception and design of research; J.L., Y.I., and G.W. performed experiments; J.L., D.Y., T.S., and G.W. analyzed data; J.L., J.D.F., D.Y., and M.B.B. interpreted results of experiments; J.L., D.Y., and M.B.B. prepared figures; J.L., J.D.F., and M.B.B. drafted manuscript; J.L., J.D.F., D.Y., T.S., Y.I., G.W., M.G., and M.B.B. approved final version of manuscript; J.D.F., M.G., and M.B.B. edited and revised manuscript.