Ubiquitous [Na+]i/[K+]i-Sensitive Transcriptome in Mammalian Cells: Evidence for Ca2+ i-Independent Excitation-Transcription Coupling

Stimulus-dependent elevation of intracellular Ca2+ ([Ca2+]i) affects the expression of numerous genes – a phenomenon known as excitation-transcription coupling. Recently, we found that increases in [Na+]i trigger c-Fos expression via a novel Ca2+ i-independent pathway. In the present study, we identified ubiquitous and tissue-specific [Na+]i/[K+]i-sensitive transcriptomes by comparative analysis of differentially expressed genes in vascular smooth muscle cells from rat aorta (RVSMC), the human adenocarcinoma cell line HeLa, and human umbilical vein endothelial cells (HUVEC). To augment [Na+]i and reduce [K+]i, cells were treated for 3 hrs with the Na+,K+-ATPase inhibitor ouabain or placed for the same time in the K+-free medium. Employing Affymetrix-based technology, we detected changes in expression levels of 684, 737 and 1839 transcripts in HeLa, HUVEC and RVSMC, respectively, that were highly correlated between two treatments (p<0.0001; R2>0.62). Among these Na+ i/K+ i-sensitive genes, 80 transcripts were common for all three types of cells. To establish if changes in gene expression are dependent on increases in [Ca2+]i, we performed identical experiments in Ca2+-free media supplemented with extracellular and intracellular Ca2+ chelators. Surprisingly, this procedure elevated rather than decreased the number of ubiquitous and cell-type specific Na+ i/K+ i-sensitive genes. Among the ubiquitous Na+ i/K+ i-sensitive genes whose expression was regulated independently of the presence of Ca2+ chelators by more than 3-fold, we discovered several transcription factors (Fos, Jun, Hes1, Nfkbia), interleukin-6, protein phosphatase 1 regulatory subunit, dual specificity phosphatase (Dusp8), prostaglandin-endoperoxide synthase 2, cyclin L1, whereas expression of metallopeptidase Adamts1, adrenomedulin, Dups1, Dusp10 and Dusp16 was detected exclusively in Ca2+-depleted cells. Overall, our findings indicate that Ca2+ i-independent mechanisms of excitation-transcription coupling are involved in transcriptomic alterations triggered by elevation of the [Na+]i/[K+]i ratio. There results likely have profound implications for normal and pathological regulation of mammalian cells, including sustained excitation of neuronal cells, intensive exercise and ischemia-triggered disorders.


Introduction
Gene expression is regulated by diverse stimuli to achieve tissuespecific functional responses via coordinate synthesis of the cell's macromolecular components [1]. Electrochemical gradients of monovalent cations across the plasma membrane (high intracellular potassium, [K + ] i vs low intracellular sodium, [Na + ] i ) are created by the Na + ,K + -pump and determine a large variety of physiologically important processes. These processes include maintenance of resting and action electrical membrane potentials, regulation of cell volume, secondary transport of mono-and divalent ions (such as chloride, calcium and phosphate), and accumulation of nutrients (glucose, amino acids, nucleotides) and other relevant molecules [2]. More recent studies demonstrated that side-by-side with the above-listed ''classic'' Na + i ,K + i -depen-dent cellular processes, sustained elevation of the [Na + ] i /[K + ] i ratio in vascular smooth muscle cells, cardiomyocytes, hepatocytes, renal epithelial and neuronal cells causes differential expression of c-Fos and other immediate response genes (IRG), as well as cell type-specific late response genes, such as tumour growth factor-b, the a1and b1-subunits of Na + ,K + -ATPase, myosin light chain, skeletal muscle actin, atrial natriuretic factor and mortalin (for review see [3][4][5]). According to the generally accepted paradigm Na + i /K + isensitive mechanism of excitation-transcription coupling is driven by changes in intracellular [Ca 2+ ] and activation of several Ca 2+sensitive pathways -a phenomenon termed excitation-transcription coupling [6][7][8]. Indeed, it is well-documented that elevation of the [Na + ] i /[K + ] i ratio typically leads to increases in [Ca 2+ ] i via activation of the Na + /Ca 2+ exchanger [9] and/or voltage-gated Ca 2+ channels [10]. It has also been shown that promoters of numerous genes including c-Fos contain serum response element (SRE) and Ca 2+ +cAMP response element (CRE) activated by [Ca 2+ ] increments in the cytoplasm and nucleus, respectively [11].
In contrast to the aforementioned mechanistic view, we found that in vascular smooth muscle cells from the rat aorta (RVSMC) and the human adenocarcinoma cell line (HeLa) the ouabaininduced changes in the c-Fos expression were preserved in the presence of Ca 2+ channel blockers and extra-and intracellular Ca 2+ chelators [12,13]. These results made us conclude that along with canonical Ca 2+ i -mediated signaling, sustained elevation of the [Na + ] i /[K + ] i ratio affects gene transcription via unknown Ca 2+ iindependent mechanism(s) [4]. In the present study, we deployed Affymetrix technology to characterize the relative impact of Ca 2+ imediated and -independent signaling on changes in gene expression triggered by sustained elevation of the [Na + ] i /[K + ] i ratio. To accomplish this goal, we compared transcriptomes in 3 different cell types treated with 2 distinct Na + ,K + -ATPase inhibitors in the absence and presence of Ca 2+ chelators. Our results strongly indicate that, in mammalian cells, Ca 2+ i -independent pathways contribute to ubiquitous and cell type-specific transcriptomic alterations triggered by elevation of the [Na + ] i / [K + ] i ratio.

Results
Effect of ouabain and K + -free medium on the [Na + ] i /[K + ] i ratio and cell viability Figure 1 illustrates that 3-hr inhibition of the Na + ,K + -ATPase by ouabain in HeLa, human umbilical vein endothelial cells (HUVEC) and RVSMC increased Na + i content from ,50 to 400-600 nmol/mg protein and decreased K + i from 600-800 to ,100 nmol/mg protein. In RVSMC the action of K + -free medium on the [Na + ] i /[K + ] i ratio was similar, whereas in HUVEC and HeLa cells the gain of Na + i and the loss of K + i triggered by K + -free medium were higher compared to ouabain by ,20% by ,50%, respectively. Importantly, in all type of cells exposed to K + -free medium, the intracellular content of monovalent cations was not significantly affected by the addition of ouabain (Fig. 1). The control experiments demonstrated that 4 hr treatment with ouabain or K + -free medium as well as the addition of 50 mM EGTA and 10 mM BATPA-AM in Ca 2+ -free medium did not impact HeLa, HUVEC and RVSMC survival, as determined by lactate dehydrogenase (LDH) release, caspase-3 activity and chromatin cleavage assay ( Table 1).

Effects of ouabain and K + -free medium on gene expression profile
Together with elevation of the [Na + ] i /[K + ] i ratio, ouabain and K + -free medium may affect cells independently of suppression of Na + ,K + -ATPase-mediated ion fluxes. Thus, recent studies have revealed that ouabain triggered interaction of the Na + ,K + -ATPase a-subunit with the membrane-associated nonreceptor tyrosine kinase Src, activation of Ras/Raf/ERK1,2, phosphatidyl inositol 3-kinase (PI(3)K), PI(3)K-dependent protein kinase B, phospholipase C, [Ca 2+ ] i oscillations and augmented production of the reactive oxygen species (for review, see [14,15]). On the other hand, the transfer of highly K + -permeable cells to K + -free medium results in transient membrane hyperpolarization, affecting the activity of diverse voltage-sensitive membrane-bound proteins [16,17]. Considering this, we compared the actions of ouabain and K + -free medium on gene expression profiles in HUVEC, RVSMC and HeLa cells with a final goal of identifying ubiquitous and cell type-specific Na + i ,K + i -sensitive transcriptomes.
The data obtained in 4 independent experiments were normalized and then analyzed by principal component analysis (PCA) [18]. Each point on the PCA represents the gene expression profile of an individual sample. Samples that are near each other in the resulting 3-dimensional plot have a similar transcriptome while those that are further apart have dissimilar transcriptional profiles. This approach identified treatments with ouabain and K + -free medium as major sources of variability within datasets ( Fig. 2A). Figure 2B disclosed that the total numbers of differentiallyexpressed transcripts in HeLa, HUVEC and RVSMC treated for 3 hr with ouabain were 819, 886 and 3199, whereas inhibition of the Na + ,K + -ATPase in K + -free medium altered the expression of 3078, 2858 and 3677 transcripts, respectively. In all types of cells, the number of up-and down regulated genes affected by these stimuli was about the same, whereas maximal fold of activation and suppression was 65 and 24, respectively ( Table 2). The decreased number of differentially expressed transcripts detected in ouabain-treated HUVEC and HeLa cells, compared to RVSMC, can be explained by retarded kinetics of elevation of [Na + ] i in ouabain-treated human cells compared to ouabaintreated RVSMC (Fig. 3A) and cells subjected to Na + ,K + -ATPase inhibition in K + -free medium (Fig. 3B). This observation is consistent with the slow kinetics of ouabain interaction with the human house-keeping a1 Na + ,K + -ATPase isoform demonstrated in early investigations [19].
Further analysis determined that the expression of 684, 737 and 1839 transcripts in HeLa, HUVEC and RVSMC, respectively, was affected by both stimuli (Fig. 2B). Importantly, we observed highly significant (p,0.0003) and positive (R 2 .0.62) correlations between levels of differentially expressed transcripts identified in the presence of ouabain and K + -free medium (Fig. 4). Because the gain of Na + i and loss of K + i in cells treated with ouabain and K +free medium are similar (Fig. 1), the results strongly suggest that the changes in gene expression evoked by both stimuli occur in response to elevation of the [Na + ] i /[K + ] i ratio rather than due to Na + i ,K + i -independent events. Considering this, we classified Na + i ,K + i -sensitive transcriptomes as sets of genes whose expression was impacted by both ouabain and K + -free medium.
To confirm the ability of microarrays to resolve the differences in expression levels, we selected several Na + i /K + i -sensitive genes (Egr1, Ptgs2 and Ppp1r15a) for additional validation by quantitative reverse transcription polymerase chain reaction (qRT-PCR) analysis. These experiments revealed highly significant correlations between the RT-PCR results and the validated microarray data (R 2 = 0.94, p,0.0000002; Fig. 5).

Ubiquitous Na
Among the Na + i ,K + i -sensitive genes detected in HeLa, HUVEC and RVSMC, we identified 80 common genes, i.e. genes whose differential expression was increased or decreased by both ouabain and K + -free medium by at least 1.2-fold (p,0.05) in all 3 cell types (Fig. 6A). We noted that the list of ubiquitous Na + i ,K + i -sensitive transcriptome was enriched with genes involved in the regulation of transcription/translation (49%), cell cycle, adhesion and migration (24%), and inflammatory and immune responses (6%) ( Table 3, Fig. 6B).
Ubiquitous Na + i /K + i -sensitive genes, whose expression was increased by more than 3-fold, included the transcriptional regulator of C2H2-type zinc-finger protein Egr-1, members of the superfamily of b-zip transcriptional factors possessing basic DNA-binding domain and leucine-zipper dimerization motif and forming heterodimeric activating protein-1 (AP-1) (Fos, FosB, Jun, JunB, Atf3), transcription factor of the steroid-thyroid hormone-retinoid receptor superfamily Nr4a2 and the basic helix-loop-helix transcription regulator Hes1. Nfkbiz and Nfkbia are transcriptional regulators of genes encoding intermediates of inflammation whereas interleukin 6 (Il6) is a potent controller of the acute inflammatory response phase. Prostaglandin-endoperoxide synthase 2 (Ptgs2) also known as cyclooxygenase-2 is a key enzyme in the biosynthesis of prostaglandins implicated in inflammatory responses and mitogenesis. Na + i /K + i -sensitive regulators of cell  proliferation, differentiation and death whose expression is sharply increased are represented by dual specificity protein phosphatase Dusp8, inhibitor of the receptor-transduced mitogen-activated protein kinase signaling pathway Spry4, the protein phosphatase 1 regulatory subunit Ppp1r15a, the cytokine inducible kinase Plk3, a member of the epidermal growth factor (EGF) family amphiregulin (Areg), and heparin-binding EGF-like growth factor Hbegf. Among ubiquitous Na + i ,K + i -sensitive genes from other functional categories, we noted augmented expression of oxidative stress mediator Txnip, the low-density lipoprotein receptor Ldlr, the regulator of cholesterol synthesis Insig1 and 2 carriers involved in intracellular phosphate handling, i.e. the Ca 2+ -dependent mitochondrial solute carrier Slc25a25 mediating ATP-Mg/P i exchange and the sodium-phosphate symporter Slc20a1 (Table 3). In contrast to substantial number of ubiquitous Na + i /K + i -sensitive genes that were strongly up-regulated, we identified only one transcript -the epsilon subunit of GABA-activated Cl 2 channel (Gabre) -whose expression was decreased by more than 3-fold (Table 3).

Cell type-specific Na
Because the number of cell-type specific Na + i ,K + i -sensitive transcripts is very large (Fig. 2B, Table 2), we restricted their functional characterization to genes whose expression was altered by more than 4-fold. The relevant analysis led us to several conclusions. First, ubiquitous genes comprised up to ,50% of Na + i ,K + i -sensitive genes whose expression was increased in HeLa, HUVEC and RVSMC by more than 4-fold (Tables 4,5,6). Second, strongly up-regulated Na + i ,K + i -sensitive genes are abundant with the transcripts that were detected in 2 cell types. These transcripts (underlined and bold) accounted for 37%, 18% and 16% of the total numbers of genes that manifested up-and down-regulation in Figure 2. Comparative analysis of the actions of Na + ,K + -ATPase inhibition by ouabain and K + -free medium on HeLa, HUVEC and RVSMC transcriptomes. A. Principal component analysis of the transcriptomes of HeLa, HUVEC and RVSMC. Cells were incubated for 3 hr in control Ca 2+ -containing media and processed for oligonucleotide microarray analysis as indicated in the Methods section. Ouabain was added at a final concentration of 3 mM (HeLa and HUVEC) or 3 mM (RVSMC). All experiments are repeated 4 times. Ellipsoids highlight portioning of samples based on the type of treatment. The principal components in 3-dimensional graphs (PC#1, PC#2 and PC#3) represent the variability of gene expression level within datasets. The total percentage of PCA mapping variability is shown on top. B. The total number of genes whose expression is altered by ouabain and K + -free medium by more than 1.2-fold with p,0.05 is indicated; numbers of genes affected by both stimuli appear in bold. doi:10.1371/journal.pone.0038032.g002 HeLa, HUVEC and RVSMC, respectively. Third, similar to ubiquitously regulated genes (Table 3), the list of cell type-specific Na + i ,K + i -sensitive genes that were strongly up-and downregulated was enriched with the transcriptional regulators and regulators of immune responses and inflammation. Examples included nuclear receptor subfamily 4 group A Nr4a1 and Nr4a3 in HeLa and RVSMC, early growth response 2 Egr2 in HeLa and RVSMC, immediate early response 2 Ier2 in RVSMC, interleukin Ouabain was added at a final concentration of 3 mM (HeLa and HUVEC) or 3 mM (RVSMC). Intracellular Na + content in the absence of Na + ,K + -ATPase inhibitors was taken as 100%. Means obtained in experiments performed in triplicate are shown. doi:10.1371/journal.pone.0038032.g003 Table 2. Total numbers of differentially expressed transcripts in HeLa, HUVEC and RVSMC in 3-hr of Na + ,K + -ATPase inhibition in control (Ca 2+ containing) medium.

Effects of Ca 2+ -free medium and intracellular Ca 2+ chelation on gene expression
To trigger Ca 2+ -depletion, we incubated cells in Ca 2+ -free medium containing extra-and intracellular Ca 2+ chelators (50 mM EGTA and 10 mM BAPTA-AM, respectively). Figure 7A displays that in HeLa cells this procedure almost completely abolished sharp elevation of [Ca 2+ ] i evoked by activation of purinergic receptors by ATP. We did not reveal any significant elevation of [Ca 2+ ] i in Ca 2+ -depleted, ATP-treated HUVEC and RVSMC (Fig. 7B).
Similarly to the results obtained in Ca 2+ -containing media, PCA found that treatments with ouabain and K + -free medium produced dramatic changes in gene expression in Ca 2+ -depleted HUVEC, RVSMC and HeLa cells (Fig. 8A). Figures 8B and Table 7 show that in all types of Ca 2+ -depleted cells, the numbers of differentially expressed genes were increased in comparison to cells incubated in Ca 2+ -containing media (Fig. 3B, Table 2). As in Ca 2+ -containing media, we ascertained significant positive correlations between the differential expression of transcripts triggered by ouabain and K + -free medium in Ca 2+ -depleted RVSMC (R 2 = 0.98; p,0.03), HUVEC (R 2 = 0.96; p,3610 27 ) and HeLa cells (R 2 = 0.93; p,2610 220 ) (Fig. 9). Importantly, the number of Na + i ,K + i -sensitive transcripts, i.e. transcripts whose expression is affected by both ouabain and K + -free medium, was also increased in the presence of extra-and intracellular Ca 2+ chelators by ,3 fold in HeLa and HUVEC and by 2-fold in RVSMC (Fig. 10A, compare to Fig. 6A).
Further analysis disclosed that (i) less than 25% of ubiquitous Na + i ,K + i -sensitive genes were detected in the Ca 2+ -containing . Correlation analysis of transcripts whose expression is altered by ouabain and K + -free medium in HeLa, HUVEC and RVSMC by more than by 1.2-fold with p,0.05. Cells were incubated during 3 hr; ouabain was added at a final concentration of 3 mM (HeLa and HUVEC) or 3 mM (RVSMC). Incubation medium contains 1.8 mM CaCl 2 . The total number of transcripts subjected to analysis is shown in Figure 2B. Transcript expression in control cells was taken as 1.00. doi:10.1371/journal.pone.0038032.g004 media only, (ii) the expression of ,15% of ubiquitous Na + i ,K + isensitive genes was independent on the presence of extra-and intracellular Ca 2+ chelators, and (iii) more than 60% of ubiquitous and cell type-specific Na + i ,K + i -sensitive genes were found exclusively in Ca 2+ -depleted cells (Fig. 11). In HeLa cells, among the Na + i ,K + i -sensitive genes whose expression was changed by ouabain or K + -free medium by more than 4-fold, we did not observe any genes whose differential expression was abolished by Ca 2+ depletion (Table 4). In HUVEC and RVSMC this procedure eliminated differential expression of ,3 and 20% of transcripts, respectively (Tables 5,6).
To further verify the efficacy of Ca 2+ depletion procedure employed in our study, we compared intracellular concentration of BAPTA, Ca 2+ i and expression of 3 selected genes in HeLa cells incubated in the presence of 10 or 100 mM BAPTA-AM. In control Ca 2+ -containing medium, 3 hr incubation of HeLa cells with ouabain elevated [Ca 2+ ] i by ,35% and increased content of EGR1, PTGS2 and PPP1R15A RNAs by ,56-, 6-and 9-fold, respectively ( Table 8). Addition of 10 mM BAPTA-AM in Ca 2+free medium containing 50 mM EGTA completely abolished an increment of [Ca 2+ ] i triggered by ouabain and increased the content of EGR1, PTGS2 and PPP1R15A RNAs by ,6-, 2-and 1.3-fold, respectively. Elevation of BATPA-AM concentration in the incubation medium up to 100 mM increased intracellular [BAPTA] from 127 to 735 mM and augmented expression of EGR1, PTGS2 and PPP1R15A by ,9-, 2,5-and 2.2-fold, respectively. This action of Ca 2+ -depletion procedure is probably caused by elevation of the passive permeability of the plasma membrane for Na + and [Na + ] i /[K + ] i ratio detected in EGTA-and BAPTA-AM-treated RVSMC [20]. Importantly, elevation of intracellular [BAPTA] up to 735 mM did not abolished increments of gene expression evoked by 3 hr incubation with ouabain (Table 8). Table 9 lists the ubiquitous Na + i ,K + i -sensitive genes, which were detected in Ca 2+ -depleted RVSMC, HUVEC and HeLa cells. Similarly to Ca 2+ -containing conditions, ubiquitous Na + i ,K + isensitive transcriptomes in the Ca 2+ -depleted cells were abundant with regulators of transcription/translation, cell cycle control and inflammatory/immune responses (Fig. 10B). Importantly, the actions of elevated [Na + ] i /[K + ] i ratio on the differential expression of several genes shown in Table 8 in italics, such as nuclear receptor Nr4a1, anti-proliferative gene Btg2, angiogenic inducer Cyr61, adrenomedulin Adm and regulator of G-protein signaling Rgs2, were abolished in the presence of Ca 2+ . We also noted that in Ca 2+ -depleted cells 23 genes or 13% of ubiquitous Na + i ,K + isensitive transcriptome were involved in protein folding and ubiquitination. This is in contrast to 2 genes in the same functional category that were detected in the presence of Ca 2+ (Fig. 6B). These results are consistent with numerous observations that Ca 2+depletion causes endoplasmic reticulum stress and activates unfolded protein response (for review, see [21]).
Based on this work we selected 8 ubiquitous Na + i ,K + i -sensitive genes whose expression was increased by more than 5-fold in Ca 2+ -containing media, and whose augmented expression was preserved in the presence of extra-and intracellular Ca 2+ chelators. As seen in Figure 12, Ca 2+ -depletion strongly attenuated (but did not completely abolish) Na + i /K + i -dependent changes in gene expression for FosB, Il6, Pppr1r15a and Ptgs2 in all cell types. In contrast, Ca 2+ -depletion slightly decreased, did not affect, or increased expression of Fos, Zfp36, Jun and Dusp8 in a cell typespecific manner. These results further illustrate that (i) [Ca 2+ ] i elevation is not obligatory for changes in gene expression, and (ii) Ca 2+ i -dependent signaling may have complex effects on ubiquitous and cell type specific Na + i ,K + i -sensitive transcriptomes.

Discussion
In the present study we utilized genome-wide Affymetrix arrays to identify [Na + ] i /[K + ] i -sensitive transcriptomes in 3 diverse cell lines. Based on the results presented here we can conclude that elevation of the [Na + ] i /[K + ] i -ratio causes robust changes in gene expression in both ubiquitous and cell type-dependent manners. The number of regulated transcripts was high and represented more than 2% of the total genome. Surprisingly, Ca 2+ -depletion increased rather than decreased number of the [Na + Table 3. Ubiquitous Na + i /K + i -sensitive genes whose expression was up-and down-regulated by more than 1.

Characterization of ubiquitous Na
We observed that sustained Na + ,K + -ATPase inhibition by ouabain led to the differential expression of 886 transcripts in primary cultured human endothelial cells, 819 transcripts in human epitheloid carcinoma cell line and 3191 transcripts in smooth muscle cells obtained from the rat aorta (Fig. 2). Differentially regulated genes represent ,3 and ,10% of genes in human and rat genome, respectively. Keeping in mind that ouabain may affect signaling pathways independently of inhibition of Na + ,K + -ATPase [14,15], we tested K + -free medium as an alternative approach to elevation of the [Na + ] i /[K + ] i ratio. Three hour incubation of cells in K + -free medium resulted in gain of Na + i and loss of K + i that were quantitatively similar to those triggered by ouabain (Fig. 1). We found highly significant positive correlations between the levels of transcripts impacted by both stimuli (Fig. 4). These data strongly indicate that differential gene expression is evoked by elevation of the [Na + ] i /[K + ] i ratio rather than by Na + i ,K + i -independent mechanisms. Among 684, 737 and 1839 Na + i ,K + i -sensitive transcripts detected in HeLa, HUVEC and RVSMC, we discerned 80 genes whose expression was increased up to 60-fold or decreased up to ,3-fold independently of the origin of cultured cells (Table 3). It should be noted that because of the retarded kinetics of elevation of the [Na + ] i /[K + ] i ratio in human cells (Fig. 3), the number of Na + i ,K + i -sensitive genes in HeLa and HUVEC as well as the number of ubiquitous Na + i ,K + i -sensitive genes is probably underestimated.
Functional analysis demonstrated that almost the half of the ubiquitous Na + i ,K + i -sensitive genes belonged to transcription and translation regulators (Fig. 6). Important examples include Egr1, Fos, Fosb, Atf3, Jun, Ddit3, Junb, Cyr61. Although functional characterization is somewhat artificial -because genes are usually multifunctional and fall into several categories -the relative content of transcriptional regulators identified in the ubiquitous Na + i ,K + i -sensitive transcriptome was ,7-fold higher than in total human genome [22]. These findings strongly suggest that heightened expression of ubiquitous Na + i ,K + i -sensitive regulators of transcription underlie ubiquitous and cell-type specific transcriptomic modifications triggered by elevation of the [Na + ] i / [K + ] i . This conclusion is supported by analysis of the top gene expression network revealed using the IPA Knowledge Base software. Indeed, Figure 13 illustrates the central role of ubiquitous Na + i /K + i -sensitive regulators of gene expression such as Fos, FosB, Jun, JunB, Atf3, Cyr61 in the triggering of diverse cellular signals. Figure 14 shows that cellular development, gene expression, cell death, immunological and inflammatory responses are among top altered biological functions controlled by ubiquitous Na + i /K + isensitive transcriptome and mapped in this software with p,10 25 . Importantly, cellular responses triggered by differential expression of Na + i ,K + i -sensitive genes might be cell type-specific. Thus, for example 24 hr incubation with ouabain resulted in death of renal epithelial cells [23] but rescued RVSMC from apoptosis triggered by serum deprivation [24].  [7,[25][26][27]. Unexpectedly, we found that in the presence of extra-and intracellular Ca 2+ chelators (EGTA and BAPTA, respectively) the number of cell type-specific and ubiquitous Na + i ,K + i -sensitive genes was increased rather than decreased. Importantly, differential expression of more than 50% of ubiquitous and cell type-specific Na + i ,K + i -sensitive genes was detected in Ca 2+ -depleted cells only (Fig. 11). Furthermore, the expression of several genes such as Fos, Zfp36, Jun and Dusp8 was not affected or even activated in the presence of extra-and intracellular Ca 2+ chelators (Fig. 12).
Recently, Akita and Okada reported that activation of volumesensitive anion channels in astrocytes by bradykinin was suppressed by addition of BAPTA-AM at concentration higher HeLa, HUVEC and RVSMC were treated with ouabain or K+-free medium for 3 hr. Genes whose expression is not affected by Na + ,K + -ATPase inhibition at least in one type of Ca 2+ -depleted cells are shown in italics. Genes whose expression is not affected by Na + ,K + -ATPase inhibition in all 3 types of cells are shown in underlined italics. than 100 mM [28]. Thus, it may be proposed that at concentration 10 mM used in our study BAPTA-AM was unable completely abolish elevation of [Ca 2+ ] i triggered by sustained inhibition of the Na + /K + -ATPase. However, data listed below did not support this assumption. First, 30 min incubation in Ca 2+ -free medium containing 10 mM BAPTA-AM and 50 mM EGTA sharply suppressed or completely abolished an increment of [Ca 2+ ] i triggered by activation of purinergic receptors (Fig. 7). Second, in contrast to overwhelming number of electrically excitable cells, 2 hr exposure of RVSMC to ouabain did not affect [Ca 2+ ] i but Table 4. HeLa cells: the list of genes whose expression was changed by more than 4-fold in 3 hr of Na + ,K + -ATPase inhibition in control (Ca 2+ containing) medium. Genes whose differential expression is limited to HeLa cells are shown in bold. Genes whose differential expression were also detected in HUVEC or RVSMC are shown in underlined bold. All transcripts listed in this table were also differentially expressed in Ca 2+ -depleted cells. doi:10.1371/journal.pone.0038032.t004 Table 5. HUVEC: the list of genes whose expression was changed in by more than 4-fold in 3 hr of Na + ,K + -ATPase inhibition in control (Ca 2+ containing) medium. Genes whose differential expression is limited to HUVEC are shown in bold. Genes whose expression were also detected in HeLa or RVSMC are shown in underlined bold. Genes whose expression was not affected by Na + ,K + -ATPase inhibition in Ca 2+ -depleted cells are shown in italics. doi:10.1371/journal.pone.0038032.t005 Table 6. RVSMC: the list of genes whose expression was changed by more than 4-fold in 3 hr of Na + ,K + -ATPase inhibition in control (Ca 2+ containing) medium. Genes whose differential expression is limited to RVSMC are shown in bold. Genes whose expression were also detected in HeLa or HUVEC are shown in underlined bold. Genes whose expression was not affected by Na + ,K + -ATPase inhibition in Ca 2+ -depleted cells are shown in italics. doi:10.1371/journal.pone.0038032.t006 sharply increased expression of Fos and Jun superfamily IRG [12].
Here, we demonstrated that addition of 50 mM EGTA and 10 mM BAPTA-AM completely abolished a modest increment of [Ca 2+ ] i triggered by 3 hr incubation of HeLa cells with ouabain whereas augmented expression of EGR1, PTGS2 and PPP1R15A was preserved (Table 8). Third, in the presence of 10 mM BAPTA-AM, intracellular concentration BAPTA in HeLa cells reached a value of ,130 mM (Table 8). Based on the volume of intracellular water (,2 ml/mg of protein), the intracellular BAPTA content on these cells (,260 pmol/mg protein) was sufficient to bind the total pool of intracellular exchangeable calcium (,200 pmol/mg protein) [13]. Fourth, elevation of intracellular [BAPTA] up to ,700 mM did not abolished increments of gene expression evoked by ouabain (Table 8). It is important to note that long-term exposure of cells to Ca 2+ chelators at higher concentrations causes diverse side-effects including elevated permeability of the plasma membrane for Na + [20]. Moreover, because of the high affinity to other di-and trivalent cations [29], Ca 2+ chelators can affect transcriptome via irreversible conformational transition and inactivation of transcriptional adaptor Zn 2+ -binding domain [30]. Viewed collectively, these data strongly indicate that besides canonical Ca 2+ -mediated signalling, gain of [Na + ] i and/or loss of [K + ] i influence gene expression via Ca 2+ i -independent mechanism.

Physiological and pathophysiological implications
The Na + i /K + i -dependent regulation of gene expression may have numerous physiological and pathological implications (Fig. 15). Here, we discuss a few of tissue-specific examples.
Excitation of neuronal cells. Almost 30 years ago, several research teams reported that expression of Erg1 and other Na + i ,K + i -sensitive IGR listed in Table 2 is strongly increased in neuronal cell subjected to excitation by diverse stimuli including neurotransmitters, depolarization and light [31][32][33]. The mechanisms of this phenomenon, widely employed for the identification of excited brain areas, remain largely unknown. It has been shown that activation of N-methyl-D-aspartate (NMDA) receptor contributes to elevation of Egr1 mRNA content in excited neuronal cells. This link is so prominent that monitoring Egr1 expression has been proposed as a biological assay for NMDA receptor activity. Given a key role of NMDA receptor activation in regulating synaptic strength, Egr1 accumulation has also been connected with learning and memory (for review, see [34]). Indeed, experiments performed on Egr1 deficient mice showed that their inability to form long-term memory in a variety of behavioural tasks [35]. It is generally accepted that Egr1 expression in neuronal cells is triggered by [Ca 2+ ] i elevation [25]. However, ion currents through NMDA channels are mainly mediated by monovalent cations (P Na ,P K &P Ca ), and short periods of synaptic activity in apical dendrites and dendritic spines produce increases in [Na + ] i , from ,10 to 30 and 100 mM, respectively [36]. Recently, sharp elevation of [Na + ] i in response to local application of glutamate was demonstrated in neocortical neurons loaded with sodiumsensitive nanoprobe [37]. Here, we report that elevation of the [Na + ] i /[K + ] i ratio in Ca 2+ -depleted HUVEC and HeLa cells augmented Egr1 expression by ,10-fold, which is comparable to increments detected in Ca 2+ -containing medium (Tables 4,5). In contrast, in RVSMC, Egr1 up regulation was exclusively detected in the presence of Ca 2+ (Table 6). Therefore, the relative impact of Ca 2+ i -mediated and Ca 2+ i -independent mechanisms on Egr1 expression triggered by elevation of the [Na + ] i /[K + ] i ratio in neuronal cells and their roles in memory formation and storage should be further examined in light of the data presented here.  Total numbers of genes whose expression is altered by ouabain and K + -free medium by more than 1.2-fold with p,0.05 are indicated; numbers of genes affected by both stimuli appears in bold. doi:10.1371/journal.pone.0038032.g008 Figure 9. Correlation analysis of transcripts whose expression is altered by ouabain and K + -free medium in Ca 2+ -depleted HeLa, HUVEC and RVSMC more than by 1.2-fold with p,0.05. Cells were incubated during 3 hr in Ca 2+ -free medium containing 50 mM EGTA and 10 mM BAPTA-AM. Ouabain was added at a final concentration of 3 mM (HeLa and HUVEC) or 3 mM (RVSMC). The total number of transcripts subjected to analysis is shown in Figure 8B. Transcript expression in control cells was taken as 1.00. doi:10.1371/journal.pone.0038032.g009 Intensive exercise. Numerous studies demonstrated that the plasma concentration of IL6 increases up to 100-fold during muscular exercise. This increase is followed by the expression of the IL1 receptor agonist (Il1ra) and the anti-inflammatory cytokine IL10. Importantly, contracting skeletal muscle rather than the immune cells is the only source of the IL6 in circulation in response to exercise. The mechanism of this phenomenon, which plays a key role in the energy supply via elevation of glucose  uptake in hepatocytes and lipolysis in adipose tissue, remains poorly understood (for comprehensive review, see [38]). It is known, however, that in both humans and experimental animals, intensive exercise increases [Na + ] i in skeletal muscles by 3-4-fold and decreases [K + ] i by 15-25% via activation of voltage-gated K + and Na + channels and partial inactivation of Na + ,K + -ATPase [39,40]. Here, we demonstrated that Il6 is among the ubiquitous genes whose expression is strongly increased by elevation of the [Na + ] i /[K + ] i ratio (Table 3). Therefore, our data allow us hypothesize that elevation of plasma level of IL6 during intensive exercise is caused by elevation of the [Na + ] i /[K + ] i ratio in skeletal muscle that, in turn, increases expression of the [Na + ] i /[K + ] isensitive genes, including Il6. Ischemia. To date, transcriptomic alterations in response to oxygen deprivation, such as modest hypoxia in solid tumours or in the adipose tissue of obese patients, have been largely ascribed to the overexpression of hypoxia-inducible factor 1 (HIF-1) known to be the molecular pO 2 sensor that affects gene expression via its binding to cis-acting hypoxic-response elements (HRE) [41]. It should be noted, however, that the overwhelming number of genes, such as Egr1, Atf3, Ptgs2, Il6, Ppp1r5, Hes1, Nfkbiz, Txnip, Adamts1, Egr 3, Cxcl2, Hsp70, whose expression is strongly increased in cells subjected to ischemia/reperfusion both in vivo and in vitro [42][43][44][45][46][47][48][49][50][51][52], lack HRE. Significantly, we detected the above-listed genes among ubiquitous (Egr1, Atf3, Ptgs2, Il6, Ppp1r5, Hes1, Nfkbiz, Txnip, Tables 3 and 9) or cell-type specific (Adamts1, Egr 3, Cxcl2, Hsp70,  [53], it can be hypothesized that inversion of the [Na + ] i /[K + ] i ratio contributes to transcriptomic changes triggered by ischemia, with pathways that are complementary to those previously associated with activation of HIF-1. This conclusion is consistent with IPA Knowledge Base data showing abnormalities of gene expression revealed in ischemic liver, heart and kidney among disorders linked with ubiquitous Na + i /K + i -sensitive transcriptome (Fig. 16). In conclusion, we report here that elevation of the [Na + ] i /[K + ] i ratio affects expression of hundreds of genes via the activation of Ca 2+ -mediated and most importantly Ca 2+ -independent signaling pathways. The proposed hypothetical hierarchy of these pathways Figure 11. Pie-chart showing the action of Ca 2+ depletion on Na + i ,K + i -sensitive transcriptomes. Ca 2+ -depletion was triggered by omission of extracellular Ca 2+ and addition of 50 mM EGTA and 10 mM BAPTA-AM. The numbers of ubiquitous Na + i ,K + i -sensitive genes and Na + i ,K + i -sensitive genes detected in HUVEC, HeLa and RVCSM are shown in italics. doi:10.1371/journal.pone.0038032.g011 Table 8. Intracellular concentration of BAPTA, Ca 2+ and expression of EGR1, PTGS2 and PPP1R15A in HeLa cells.

Incubation medium
Ouabain, mM  Table 9. Genes whose expression was changed in HeLa, HUVEC and RVSMC by more than 1.2-fold (p#0.05) in 3 hr of Na + ,K + -ATPase inhibition in Ca 2+ -free medium containing extra-and intracellular Ca 2+ chelators.   is depicted in Figure 15. Previously, we demonstrated that gain of Na + i rather than loss of K + i triggers augmented c-Fos expression in RVSMC [12]. We also reported that c-Fos expression triggered by [Na + ] i elevation occurs via Ca 2+ i -independent signaling mechanism that is not mediated by any known transcriptional elements in 59-promoter [13]. We firmly believe that the newly gained knowledge of the Na + i /K + i -sensitive transcripts (genes X 1,2,3… , Fig. 15) will allow researchers to take powerful bioinformatics approaches to identify previously-unfamiliar Ca 2+ i -independent mechanism of the excitation-transcription coupling as well as its involvement in cellular responses triggered by sustained elevation of the [Na + ] i /[K + ] i ratio.

Cell cultures
We studied three cell types in the experiments reported here. RVSMC cells were isolated from rat aortae according to the procedures outlined in the Guide for the Care and Use of Experimental Animals endorsed by the Canadian Institutes of Health Research and accepted by the Institutional Animal Protection Committee of the CRCHUM. These cells maintain a number of characteristics of primary cultured RVSMC including high expression of smooth muscle-specific a-actin, SM22 protein and myosin light chain kinase [54]. The human cervical adenocarcinoma cell line HeLa was purchased from the American Type Culture Collection (Rockville, MA, USA). The human umbilical vein endothelial cells (HUVEC) were purchased from Lonza (Walkersville, MD, USA) and passaged 4-12 times. RVSMC and HeLa were maintained in Dulbecco's Modified Eagle Medium (DMEM, Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum (FBS) and 100 U/ml penicillin and 100 mg/ml streptomycin. HUVEC were cultured in complete endothelial cell growth medium-2 (EGM-2 BulletKit, CC3162, Lonza). All cell cultures were maintained in a humidified atmosphere with 5% CO 2 /balance air at 37uC. To establish quiescence, cells were incubated for 24 hr in the media in which concentration of FBS was reduced to 0.2%. The possible impact of Na + ,K + -ATPase inhibition and Ca 2+ -depletion on cell viability and apoptosis was studied by lactate dehydrogenase Genes whose expression was not affected by ouabain or K + -free medium in the presence of Ca 2+ in at least one of cell types are shown in italics. Genes whose expression was not affected by ouabain or K + -free medium in the presence of Ca 2+ in all 3 types of cells shown in underlined italics. Gene functions are indicated in the left column as: t -regulators of transcription/translation, RNA processing and degradation; d -regulators of cell adhesion, migration, proliferation, differentiation and death; f -protein folding and ubiquitination; i -inflammation and immune response; o -other functional categories and genes withy unknown function. doi:10.1371/journal.pone.0038032.t009 (LDH) release and measurement of caspase-3 activity and chromatin cleavage, as described in details elsewhere [23,24,55].

Elevation of the [Na + ] i /[K + ] i ratio
Quiescent cells were washed with Ca 2+ -and K + -free DMEM (Sp-DMEM; Invitrogen, Carlsbad, CA) and incubated for 3 hr in either control medium containing 1.8 mM CaCl 2 and 5 mM KCl (Sp-DMEM+Ca,K) or in Ca 2+ -free medium (Sp-DMEM+K) containing 50 mM EGTA and 10 mM BAPTA-AM. To increase the [Na + ] i /[K + ] i ratio, the Na + ,K + -ATPase activity was inhibited by the addition of ouabain or the omission of extracellular K + (K +free medium, Sp-DMEM+Ca). Because the affinity of the ubiquitous a1-Na + ,K + -ATPase for ouabain and other cardiotonic steroids (CTS) in rodents is ,1000-fold lower than in other mammalian species [56], ouabain was added to the media with human and rat cells at the concentrations of 3 and 3,000 mM, respectively. Intracellular content of exchangeable K + and Na + Intracellular content of exchangeable K + and Na + was measured as the steady-state distribution of extra-and intracellular 86 Rb and 22 Na, respectively. To establish isotope equilibrium, cells growing in 12-well plates were preincubated for 3 hr in control or K + -free medium (Sp-DMEM+Ca) containing 0.5 mCi/ml 86 RbCl or 3 mCi/ml 22 NaCl and ouabain was added for the next 3 hr. To test the action of K + -free medium, cells were washed twice with ice-cold Sp-DMEM+Ca. Then, cells loaded with 22 Na were transferred to Sp-DMEM+Ca, containing 3 mCi/ml 22 NaCl, whereas cells loaded with 86 Rb were transferred to isotope-free Sp-DMEM+Ca. After 3 hr, the cells were transferred onto ice, washed 4 times with 2 ml of ice-cold medium W containing 100 mM MgCl 2 and 10 mM HEPES-tris buffer (pH 7.4). The washing medium was aspirated and cells were lysed with 1% SDS and 4 mM EDTA solution. Radioactivity of incubation media and cell lysates was quantified, and intracellular cation content was calculated as A/am, where A was the radioactivity of the samples (cpm), a was the specific radioactivity of 86 Rb (K + ) and 22 Na in the medium (cpm/nmol), and m was protein content. For more details, see [57].

Measurement of intracellular Ca 2+
Cells grown on glass cover slips were incubated for 30-40 min in medium containing 5 mM fura 2-AM, washed twice and kept for up to 30 min at room temperature before the experiments. Then, the cover slips treated as indicated in figure and table legends were mounted in a diagonal position in a 161 cm cuvette, and fluorescence was determined under permanent stirring at 37uC (l ex = 340 and 380 nm, slit 4 nm; l em = 510 nm, slit 12 nm), using a SPEX FluoroMax spectrofluorimeter (Edison, NJ). Free [Ca 2+ ] i was quantified as [Ca 2+ ] i = K d (R2R min )x(R max 2R) 21 , where K d is the dissociation constant of the Ca 2+ -fura 2 complex (224 nM at 37uC), and R = F 340 /F 380 is the ratio of fluorescence at l ex = 340 and 380 nm. To determine F max , the cells were treated with 0.5 mM ionomycin in the presence of 1 mM CaCl 2 . To determine F min , MnCl 2 was added at a final concentration of 2 mM.

Measurement if intracellular BAPTA
HeLa cells seeded in 12-well plates were incubated for 3 hr in control or Ca 2+ -free media containing 2 mCi/ml [ 14 C]-urea. Then, the cells were washed with 463 ml of ice-cold medium W and lysed for BAPTA measurement by addition of 200 ml of 10% trichloroacetic acid containing 1 mM CaCl 2 or for radioactivity measurement as described above. The content of BAPTA in cell Figure 12. Effect of Ca 2+ -depletion on expression of ubiquitous Na + i ,K + i -sensitive genes in HeLa, HUVEC and RVSMC. Cells were incubated during 3 hrs in control medium containing 1.8 mM CaCl 2 or in Ca 2+ -containing K + -free medium. To trigger Ca 2+ -depletion, CaCl 2 was omitted and 50 mM EGTA and 10 mM BAPTA-AM were added. Increments of gene expression triggered by Na + ,K + -ATPase inhibition in Ca 2+ -containing medium were taken as 100%. For absolute values of gene expression, see Tables 3 and 8. doi:10.1371/journal.pone.0038032.g012 lysates was estimated by the increment of absorbance at 255 nm triggered by addition of 10 mM EGTA (DA 255 ). Intracellular concentration of BAPTA (mmol/L) was calculated as DA 255 / DA 255St V i , where DA 255St is the Ca 2+ -dependent increment of A 255 in the presence of 0.001 mmol BAPTA (internal standard) and V i is the volume of intracellular water (ml/mg protein). V i was calculated as the volume of [ 14 C]-urea available space V i = V o A i /A o m, where A i and A o are the radioactivity of [ 14 C]-urea in the cell lysate and incubation medium, respectively (dpm), m is protein content in the cell lysate (mg), and V o is the volume of incubation medium (L) used for A o determination.

RNA isolation
Total RNA was extracted from cells grown in 6-well plates using TRIzolH reagent (Invitrogen, Carlsbad, CA and purified with the RNeasyH MinElute cleanup kit (Qiagen, Valencia, CA) following the manufacturers' protocols. Only the RNA samples that had more than 7.0 RNA integrity number (RIN) and no detectable genomic DNA contamination were used for the subsequent gene array analyses. RNA quality was assessed by 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA). Microarray experiments were performed with GeneChipH Human Gene 1.0 ST array (which detects 28,869 gene products) and GeneChipH Rat Gene 1.0 ST array (detects 27 342 gene products). On both arrays, each gene was represented by approximately 26 probes along the entire length of the transcript (Affymetrix, Santa Clara, CA). 100 ng of total RNA for each sample was processed with AmbionH WT Expression Kit (Invitrogen). This kit uses a reverse transcription priming method that specifically primes non-ribosomal RNA, including both poly(A) and non-poly(A) mRNA, and generates sense-strand cDNA as the final product. 5.5 mg of the singlestranded cDNA was fragmented and labeled using the Affymetrix GeneChipH WT Terminal Labeling Kit and 2.0 mg of the resulting cDNA was hybridized on the chip.

GeneChip expression analysis
The whole hybridization procedure was conducted with the Affymetrix GeneChipH system according to the protocol recommended by the manufacturer. The hybridization results were evaluated with Affymetrix GeneChipH Command Console Software (AGCC). Quality of the chips was determined using Affymetrix Expression Console. Data analysis was performed within Partek Genomics Suite (Partek, St. Louis, Missouri). The data were initially normalized by Robust Multichip Average (RMA) algorithm, which uses background adjustment, quantile normalization and summarization. Then, normalized data were Figure 13. The signaling network possessing the highest score of 48 for association with differential expression of ubiquitous Na + i ,K + i -sensitive genes. The p-value and geometric fold change for each gene listed in Table 3 were overlaid onto a global molecular network developed from information within the IPA Knowledge Base. Genes are represented as nodes of various shapes to represent the functional category of gene product as shown in the top corner, and the biological relationship between two nodes is represented as a line. The up-regulated Na + i ,K + isensitive genes are shown in pink. Note that network's Score of 48 is equal to -log(Fisher's Exact test result). It means that there is a 1 in 10 48 chance of getting a network from Ingenuity Knowledge Base containing at least the same number of eligible molecules by chance when randomly picking molecules derived form 80 ubiquitous Na + i ,K + i -sensitive genes listed in Table 3. doi:10.1371/journal.pone.0038032.g013 analyzed by principal component analysis (PCA) [18] to identify patterns in the dataset and high-light similarities and differences among the samples. Major sources of variability identified within the dataset by PCA were used as grouping variabilities for analysis of variance (ANOVA) with n = 4 for each group of samples. The ensuing data were filtered to identify transcripts with statistically significant variation of expression among the groups that are modulated by at least 20%, with multiple testing correction by the false discovery rate (FDR). The calculated p-value and geometric fold change for each probe set identifier were imported into Ingenuity Pathway Analysis (IPA, Ingenuity Systems, http://www. ingenuity.com) to ascertain networks, biological functions and their pathophysiological implications. Functional information on regulated gene was also obtained using PubMed and cited publications.

Real-time quantitative RT-PCR
To validate data obtained by genome-wide Affymetric gene arrays, we estimated changes in the gene expression for several selected transcripts by qRT-PCR, which was performed using Express SYBR GreenER qPCR Supermix kit (Invitrogen, Carlsbad, CA, USA) according to the manufacture's instructions. The reaction was carried out with a 7900 HT Fast RT -PCR system (Applied Biosystems, Foster City, CA, USA). Primers for Egr1, Ptgs2 and Ppp1r15a were designed using Primer3Plus online software from consensus sequences provided by Affymetrix for Figure 14. Top significantly altered biological functions associated with differential expression of ubiquitous Na + i ,K + i -sensitive genes. The p-value and geometric fold change for each gene listed in Table 3 were imported into Ingenuity Pathway Analysis. The significance criteria with a threshold of p = 0.05 (or 1.3 when expressed as -log(p-value) is shown by line. doi:10.1371/journal.pone.0038032.g014 Figure 16. Disorders significantly associated with differential expression of ubiquitous Na + i ,K + i -sensitive genes. The p-value and geometric fold change for each genes listed in Table 3 were imported into Ingenuity Pathway Analysis. The criteria with a threshold for significance of p = 0.05 (or 1.3 when expressed as -log(p-value) is shown by line. doi:10.1371/journal.pone.0038032.g016