G. Scheiner-Bobis, Institut für Biochemie und Endokrinologie, Fachbereich Veterinärmedizin, Justus-Liebig-Universität Giessen, Frankfurter Strasse 100, D-35392 Giessen, Germany Fax: +49 641 9938179 Tel: +49 641 9938180 E-mail: email@example.com
Using SK-N-AS human neuroblastoma cells, which co-express the α1 and α3 isoforms of the sodium pump α subunit, we selectively silenced either the α1 or α3 subunit by means of transfection with small interfering RNA, and investigated cell survival and the cellular response to ouabain. We found that both of the α subunits are essential for cell survival, indicating that substitution of one subunit for the other is not sufficient. In the presence of both α subunits, ouabain causes sustained activation of extracellular signal-regulated kinases 1 and 2 (Erk1/2). This activation is not affected when the α1 subunit is silenced. However, when α3 expression is silenced, ouabain-induced activation of Erk1/2 does not occur, even at a high concentration of ouabain (1 μm). Thus, ouabain-induced Erk1/2 activation is mediated in SK-N-AS cells by α3 only, and α1 does not participate in this event. This is a clear demonstration of selective involvement of a specific sodium pump α subunit isoform in ouabain-induced signaling.
The sodium pump (Na+,K+-ATPase; EC 184.108.40.206) maintains the Na+ gradient across plasma membranes of animal cells . By hydrolyzing ATP, the enzyme transports three Na+ ions out of the cell in exchange for two K+ ions that are brought into the cytosol. This activity can be interrupted by a group of substances that are referred to as cardiotonic steroids (CTS), a name linked to their clinical use for the treatment of heart failure .
In recent years, numerous publications have established that CTS not only inhibit the sodium pump but also induce signaling cascades that may be associated with cell growth and proliferation as well as with apoptotic cell death, depending upon the cell type or CTS investigated. CTS-induced signaling does not depend on sodium pump inhibition [3–8], as inactive sodium pump mutants can still transmit signals when CTS are added to the cell culture [9,10].
The sodium pump of animal cells is an oligomeric enzyme consisting of α and β subunits . In some tissues, a regulatory γ subunit is associated with the α and β subunits . All three subunits have been co-crystallized in several conformational states of the enzyme [12–14]. The α subunit, which is referred to as the catalytic subunit, has ten transmembrane domains, hydrolyzes ATP, transports the cations and is the pharmacological receptor for CTS. The β subunit is a highly glycosylated protein with a single transmembrane span, and appears to function as a molecular chaperone for correct folding of the α subunit and its transportation to the plasma membrane . The γ subunit (also termed FXYD2) is a member of the FXYD family of proteins that includes phospholemman (FXYD1) and corticosteroid hormone-induced factor (FXYD4). These very hydrophobic proteins are characterized by a single transmembrane span (except FXYD3, mammary tumor marker Mat-8, which has two transmembrane spans) and an FYXD motif near the transmembrane domain, in the extracellular N-terminal part of the protein . Each of the subunits exists in various isoforms . There are four α subunits (α1–α4), three β subunits (β1–β3) and two splice variants of the γ subunit (FXYD2a and FXYD2b). Various investigations have shown that at least five of the seven FXYD proteins interact with the α and β subunits of Na+,K+-ATPase and regulate functions of the enzyme . In several cases, multiple isoforms of the α subunit are found in the same cell type, thus raising questions about the physiological significance of such co-existence.
Given that CTS/sodium pump interactions resemble typical hormone/receptor-mediated events in many respects [7,17], it may be assumed that co-existing subunits are involved in other signaling events in addition to the ion pump function [6–8]; however, there is no direct evidence to support this notion so far. We investigated this question by using the human neuroblastoma cell line SK-N-AS. These neuroblastoma cells, which co-express the α1 and α3 subunits, were shown to interact with the CTS ouabain. The results provide evidence for distinctive roles in signal mediation for the two subunit isoforms of the sodium pump.
Expression of α subunit isoforms of Na+,K+-ATPase in SK-N-AS cells before and after transfection with Stealth™ RNAi to silence α1 and α3
SK-N-AS is a cancerous neuroblastoma cell line that, like other neuronal cells [18,19], expresses α1 and α3 subunit isoforms (Fig. 1). As shown in Fig. 1A, in control (untreated) SK-N-7AS cells, α1 and α3 subunit-specific cDNA bands are present in equivalent quantities (left lanes). The same result is seen when the cells are transfected with control RNAi: expression of both α1-specific and α3-specific cDNA is like that of the control cells (Fig. 1A, middle lanes). However, transfection of the SK-N-AS cells with either α1-specific RNAi (Fig. 1A, upper panel) or α3-specific RNAi (Fig. 1A, lower panel) leads to a reduction in the corresponding mRNA/cDNA (Fig. 1A, right lanes). In both cases, the reduction in expression of either α1- or α3-specific mRNA is significant (Fig. 1B). mRNA/cDNA for the α2 subunit was not detected. For normalization of data, glyceraldehyde 3-phosphate dehydrogenase-specific mRNA/cDNA from every probe was amplified in parallel experiments (Fig. 1A). Its expression was not affected.
Survival of SK-N-AS cells before and after transfection with Stealth™ RNAi to silence α1 and α3
Experiments with knockout mice demonstrated that either the α1 or α2 isoforms are essential for survival of the animals . Nevertheless, it is not known whether survival of cells that co-express multiple α subunits depends on the simultaneous presence of the various subunits or whether one subunit isoform can substitute for the other. To investigate this question, SK-N-AS cells were treated with specific RNAi to silence the expression of either the α1 or α3 subunit of the sodium pump, and cells were cultured for several days. Untreated cells served as a control. The MTT assay was used to determine the number of living cells under each condition. Figure 2 shows that, for the first 48 h, growth is the same for all cell types. However, cells lacking either α1 or α3 do not multiply further thereafter, and after 8 days (192 h), the number of living cells was reduced by more than 25% compared to the original number of cells, and by about 70% when compared to the number of living cells expressing both α1 and α3 subunits. It should be noted, however, that the cells expressing both subunits reached confluence after the 4th day of incubation and therefore did not multiply any further.
Ouabain-induced signaling in SK-N-AS cells
Ouabain and other CTS induce signaling cascades in a variety of cells and also in the neuroblastoma cell line SH-SY5Y [7,21,22]. One of the first events seen upon exposure of various cell types to CTS is activation of extracellular signal-regulated kinases 1 and 2 (Erk1/2) [3,21,23–26]. Therefore, we focused our attention on possible stimulation (phosphorylation) of these kinases in SK-N-AS cells. As in other cell types, low concentrations of ouabain trigger significant activation of Erk1/2 within 30 min in SK-N-AS cells (Fig. 3), thus raising the question of whether ouabain-induced activation of these kinases is mediated through the α1 or α3 subunit or through both subunit isoforms of the sodium pump.
Ouabain-induced activation of Erk1/2 after silencing either α1 or α3 subunits
Cells expressing only the ubiquitous α1 subunit of the sodium pump respond to CTS by induction of a variety of signaling cascades. However, in the present study, cells lacking α1 clearly show concentration-dependent ouabain-induced activation of Erk1/2 (Fig. 4). In contrast, cells lacking α3 did not respond to ouabain, even at the very high concentration of 1 μm (Fig. 4).
Previous experiments with knockout mice have demonstrated that the presence of either α1 or α2 subunits is critical for animal survival . However, as single cells can simultaneously express various α subunits, we addressed the question of whether survival of cells would be affected by loss of only one α subunit isoform, and, if so, which one is more essential for survival. Using the same experimental set-up, we also determined whether α1 and α3 subunits respond to ouabain by induction of different signaling events or whether signaling cascades are isoform-independent.
When either the α1 or α3 subunits are silenced, cells proliferate over a period of 2 days in a manner similar to the control, in which neither of the two subunits had been silenced (Fig. 2). After that, the number of living cells starts declining, until, at day 8, the numbers of cells that lack either α1 or α3 are only about 30% of the number of control cells that express both subunit isoforms. The data in Fig. 2 indicate that cells lacking α1 show an earlier loss of viability than those lacking α3; however, the impact of this is not known. Nevertheless, the results clearly show that both α1 and α3 subunits are essential for survival, and that the α1 and α3 isoforms have distinct roles in SK-N-AS cells and loss of one cannot be compensated by the other. Based on these and previous findings, it is possible to speculate that these subunits are similarly essential in other cell types as well.
What niche of cell biological functions do α1 or α3 subunits occupy that makes them essential for survival? Are their functions identical, or do they differ in some respects? In the investigation presented here, the latter seems to be the case; cells lacking the α1 subunit respond to ouabain by activation of Erk1/2, indicating that the α3 subunit can transmit CTS-induced signaling (Fig. 4A,C). However, silencing the α3 subunit abolishes the ability of the cells to respond to CTS (Fig. 4B), indicating that the presence of α3 is essential for this signaling pathway in SK-N-AS cells. Activation of Erk1/2 was not observed in these cells even at the high concentration of 1 μm ouabain (Fig. 4B), indicating that inhibition of the sodium pump, which occurs at this concentration, is not a requirement for CTS-induced signaling, as shown previously [9,10]. Figure 4B,D additionally shows that, after silencing the α3 subunit, the Erk1/2 activity is already reduced compared to the activity seen in control cells, indicating that, even in the absence of externally added ouabain, part of the Erk1/2 basic activity is contributed by the α3 subunit of the sodium pump.
Here we demonstrate for the first time that α1 and α3 have distinct functions in cell physiology, but this leads to a fundamental question: why does α1 not mediate CTS-induced activation of Erk1/2 in SK-N-AS cells when it has been shown to be involved in CTS-induced signaling in various other cells? This question can be addressed by taking into consideration the mechanisms by which CTS generate signals and the structural differences between the α1 and α3 sodium pump subunits. CTS-induced signaling can be explained by two different models. In the first model, it is assumed that cell signaling induced by CTS is due to inhibition of the sodium pump and a local increase in intracellular [Na+] followed by a subsequent increase in [Ca2+] in the small space between the plasmalemma and endoplasmic/sarcoplasmic reticulum. This space, referred to as the plasmerosome, contains sodium pump isoforms α2 and α3 but not α1 in smooth muscle cells and astrocytes [6,8,27]. In an alternative model, the sodium pump is considered to be a member of a caveolae-defined environment of proteins that are capable of communicating with each other. This entity is referred to as the signalosome . This model proposes that it is not inhibition of the sodium pump but rather conformational changes of the CTS/Na+, K+-ATPase complex that trigger the signaling cascade. This is supported by the fact that signaling cascades are activated when CTS interact with non-pumping sodium pump mutants . Although the models differ in their basic assumption, they have in common the requirement that the sodium pump be targeted to a defined environment.
The results presented here show that the α3 isoform can generate CTS-induced activation of Erk1/2, indicating its localization in an environment different to that of the α1 isoform. The basis for these different environments is may be due to differences in the structure of the two proteins. The primary structures of the human α1 and α3 isoforms are 87% identical and display a similarity of 94%. Nevertheless, at the level of tertiary structure, they display significant differences. In a recently published comprehensive work, comparison of the tertiary structures of α1/α2 and α1/α3 subunits revealed that surface-exposed areas of the α2 or α3 isoforms were very different from the corresponding areas of the α1 isoform . These areas are found mainly within the N-domain but also within the A-domain of the proteins. The membrane-spanning segments of the three isoforms are rather conserved . We assume that the clusters of isoform-specific differences in the surface-exposed regions might be important for isoform-specific interactions with other proteins. These specific interactions, which may result in either distinctive targeting of the isoforms to different areas of the plasma membrane or specific interactions with signaling molecules (or both), could be the reason for the differences found for ouabain-induced signaling through the α1 or α3 isoforms. The recent demonstration that the α1 isoform is recruited to the plasma membrane via interaction of adaptor protein 1 with Tyr255 of this isoform supports this hypothesis, and demonstrates that even small differences in surface-exposed areas may have a big impact in targeting of the proteins . Isoforms α2 or α3 lack this tyrosine residue, and we assume that their targeting to specific areas of the plasma membrane must be defined by other parameters.
Thus, based on our results and those of others discussed above, we suggest that the differences seen in the CTS-induced signaling through the α1 or α3 isoforms are associated with differences in the surface-exposed regions of the two proteins. These might lead to specific targeting of each isoform to different micro-environments of the plasma membrane or to isoform-specific interactions with other proteins of the micro-environments. Future work should help to verify this assumption.
SK-N-AS cells (American Type Culture Collection, Manassas, VA) were cultured in Dulbecco’s modified Eagle’s medium (PromoCell, Heidelberg, Germany) supplemented with 10% v/v fetal bovine serum (PromoCell) and 100 IU·mL-1 Pen/Strep (PromoCell). The culture was maintained in a humidified incubator at 37 °C in 5% CO2. The medium was replaced twice per week. Cells were harvested by incubating with trypsin (0.25%; PromoCell) for 2 min at 37 °C.
Preparation of cell lysates
Cell lysates were prepared as described previously .
SDS/PAGE and western blotting of isolated proteins
A total of 10–50 μg of protein was separated by SDS/PAGE using 10% acrylamide and 0.3%N,N′-methylene-bis-acrylamide gels. Biotinylated molecular weight markers (Cell Signaling Technology, Frankfurt am Main, Germany) were run in parallel. After SDS/PAGE, proteins were electroblotted onto nitrocellulose membranes (Schleicher & Schuell, Dassel, Germany) at 500 mA for 30–40 min. Detection of proteins was performed as described by the manufacturers of the antibodies (Cell Signaling Technology, Santa Cruz Inc. or Dianova, Hamburg, Germany) in combination with an enhanced chemiluminescence (ECL) kit (GE HealthCare, Munich, Germany). Chemiluminescence was visualized and quantified using a molecular imager ChemiDoc XRS system (Bio-Rad, Munich, Germany).
Detection of α1-, α2- and α3-specific mRNA in SK-N-AS cells
Cells were grown to 75% confluence before isolation of total mRNA using the RNeasy mini kit (Qiagen, Hilden, Germany). RNase-free DNase I (Qiagen) was used to eliminate potential contamination by DNA. The concentration and purity of total mRNA were determined by measuring the absorbance at 260 and 280 nm.
The OneStep RT-PCR kit (Qiagen) was used for reverse transcription and PCR amplification of DNA. In a total volume of 50 μL, 20–40 ng of mRNA, 1 μL (20 pmol·mL−1) of each primer, 10 μL of 5× buffer (12.5 mmol·L−1 MgCl2, 20 mmol·L−1 Tris/HCl, 100 mmol·L−1 KCl, 10 mmol·L−1 dNTPs), 2 μL of a mixture of Omniscript and Sensiscript reverse transcriptases and HotStar Taq DNA polymerase were incubated in a MasterCycler gradient (Eppendorf, Hamburg, Germany) at 50 °C for 30 min for the reverse transcription reaction. Then the mixture was heated at 95 °C for 15 min, followed by 40–45 cycles of denaturation at 94 °C for 1 min, annealing at 58 °C for 1 min and extension at 72 °C for 1 min, with a final extension at 72 °C for 10 min. For specific amplification of α1, the forward and reverse primers were 5′-GTTGGGGCTCCGATGTGTTGGGGT-3′ and 5′-CTGGCTGGAGGCTGTCATCTTCTTCAT-3′, respectively; for specific amplification of α2, the forward and reverse primers were 5′-CTGGCTGGAGGCTGTCATCTTCTTCAT-3′ and 5′-GGCTCTTGGGGGCTGTCTTCTCGCT-3′, respectively; for specific amplification of α3, the forward and reverse primers were 5′-CTGGCTTGAGGCTGTCATCTTCTTCAT-3′ and 5′-ATCGGTTGTCGTTGGGGTCCTCGGT-3′ respectively. The corresponding fragment sizes are 560 bp (α1), 557 bp (α2) and 560 bp (α3). To control PCR efficiency and the quality of the cDNA, primers 5′-TGGGGAAGGTGAAGGTCGGAGTCAA-3′ and 5′-TAAGCAGTTGGTGGTGCAGGAGGCA-3′ were used to co-amplify a specific fragment of 469 bp coding for the housekeeping gene glyceraldehyde 3-phosphate dehydrogenase.
The RT-PCR products were analyzed by electrophoresis in a 1.7% agarose gel. The correct identity of α1- or α3-specific amplified sequences was further verified by digestion with Bpi, Eco47I, BspTI (all MBI Fermentas, St Leon-Rot, Germany) and subsequent agarose gel electrophoresis.
Silencing a1 or a3 mRNA biosynthesis by siRNA transfection
Stealth™ RNAi (Invitrogen, Karlsruhe, Germany) for silencing α1 and α3 isoforms of Na+,K+-ATPase was transfected into SK-N-AS cells using Lipofectamine 2000 according to the manufacturer’s protocol. In brief, 2 × 105 SK-N-AS cells were placed in each well of a six-well culture vessel in Dulbecco’s modified Eagle’s medium without antibiotics. Cells were 30–50% confluent at the time of transfection. Before transfection, Stealth™ RNAi and Lipofectamine 2000 were diluted with Opti MEM I Reduced Serum Medium (Invitrogen) and incubated for 5 min at room temperature. Then, diluted Stealth™ RNAi and diluted Lipofectamine 2000 were combined, mixed gently and added to the cells after 20 min. The final concentration of the Stealth™ RNAi was 100 nm. Control cells were treated with Lipofectamine 2000 only. In parallel, cells were transfected with Stealth™ RNAi Negative Control High or Medium GC (negative control for either α1 or α3 Na+,K+-ATPase Stealth™ RNAi). After 72 h of incubation at 37 °C in a CO2 incubator, transfection efficiencies of 81 ± 3% in the SK-N-AS cells were estimated using a Block-iT™ transfection kit (Invitrogen) according the manufacturer’s protocol. Total RNA was isolated from SK-N-AS cells, and the extracted RNA was subjected to RT-PCR to amplify α1-, α2- and α3-specific DNA fragments as described above.
The sequences for the Stealth™ RNAi were: 5′-GGGUGUGGUGCUAUCAGCCGUUGUA-3′ and 5′-UACAACGGCUGAUAGCACCACACCC-3′ (α1 subunit); 5′-ACGACAACCGAUACCUGCUGGUGAU-3′ and ‘5-AUCACCAGCAGGUAUCGGUUGUCGU-3′ (α3 subunit).
Seventy-two hours after transfection with Stealth™ RNAi, cells were exposed to various ouabain concentrations for 3 h. These cells were used to produce lysates that were subsequently used in western blot experiments.
In a different set of experiments, the survival of cells transfected with Stealth™ RNAi was followed over a longer period of time using the MTT assay described below.
Cells were treated with Stealth™ RNAi (2 × 104 cells per well/24-well culture vessel) under the conditions described above. After replacing the transfection medium by complete growth medium, incubation was continued for various times (1–8 days). Afterwards, the medium was aspirated and replaced by 300 μL of fresh medium containing 0.5 mg·mL−1 MTT. The cells were incubated for an additional 4 h. The medium containing MTT was removed by inverting the plate, and the resulting formazan crystals were solubilized by adding 200 μL dimethyl sulfoxide to each well. After 10 min of vigorous vortexing, the absorbance in each well was read in a microplate reader at 540 nm.
Data were analyzed by one-way anova and by applying Dunnett’s comparison for evaluation of all data with respect to control values. Significance was accepted at P < 0.05.
L.K. was supported by a stipend granted through the German Academic Exchange Service (Deutscher Akademischer Austausch Dienst).