Skeletal muscle Na,K-ATPase alpha and beta subunit protein levels respond to hypokalemic challenge with isoform and muscle type specificity.

During potassium deprivation, skeletal muscle loses K+ to buffer the fall in extracellular K+. Decreased active K+ uptake via the sodium pump, Na,K-ATPase, contributes to the adjustment. Skeletal muscle expresses α1, α2, β1, and β2 isoforms of the Na,K-ATPase αβ heterodimer. This study was directed at testing the hypothesis that K+ loss from muscle during K+ deprivation is a function of decreased expression of specific isoforms expressed in a muscle type-specific pattern. Isoform abundance was measured in soleus, red and white gastrocnemius, extensor digitorum longus, and diaphragm by immunoblot. α2 expression was uniform across control muscles, whereas α1 and β1 were twice as high in oxidative (soleus and diaphragm) as in fast glycolytic (white gastrocnemius) muscles, and β2 expression was reciprocal: highest in white gastrocnemius and barely detectable in soleus and diaphragm. Following 10 days of potassium deprivation plasma K+ fell from 4.0 to 2.3 mM, and there were distinct responses in glycolytic versus oxidative muscles. In glycolytic white gastrocnemius α2 and β2 fell 94 and 70%, respectively; in mixed red gastrocnemius and extensor digitorum longus both fell 60%, and β1 fell 25%. In oxidative soleus and diaphragm α2 fell 55 and 30%, respectively, with only minor changes in β1. Although decreases in α2 and β2 expression are much greater in glycolytic than oxidative muscles during K+ deprivation, both types of muscle lose tissue K+ to the same extent, a 20% decrease, suggesting that multiple mechanisms are in place to regulate the release of skeletal muscle cell K+.

Mammals closely control extracellular fluid (ECF) 1 and intracellular fluid (ICF) potassium within a very narrow range since the ratio of the ICF potassium (K ϩ ) to ECF K ϩ is the primary determinant of resting membrane potential. This is especially important in excitable tissues such as the heart, where disturbances in resting membrane potential can compromise cardiac contractility and become life threatening (1). ECF K ϩ is maintained by the interplay of two key organ systems: the kidneys control K ϩ excretion by actively reabsorbing or secreting K ϩ (1), while skeletal muscle, which contains the largest intracellular pool of K ϩ in the body (75% of total intracellular K ϩ ), adjusts ECF K ϩ by regulating K ϩ transport between the ICF and ECF compartments (2,3).
During dietary K ϩ deprivation, K ϩ output can exceed input, leading to a fall in plasma potassium, termed hypokalemia (1). In response, skeletal muscle selectively loses ICF K ϩ into the ECF. This loss buffers the fall in ECF K ϩ , minimizing the hyperpolarization of cell membranes due to the increased ratio of ICF K ϩ to ECF K ϩ . There is a concomitant decrease in the number of active sodium pumps (Na,K-ATPase) in skeletal muscle plasma membranes, postulated to be the primary event leading to the transfer of ICF K ϩ into the ECF (4).
Our previous studies provided evidence for isoform-specific regulation in response to hypokalemia: ␣2 protein and mRNA, but not ␣1 or ␤1, decreased in hindlimb skeletal muscle of hypokalemic rats (11). Since skeletal muscle is a heterogeneous mix of skeletal muscle tissue types that vary greatly in both metabolic and contractile characteristics (12), we aimed to investigate whether the response to hypokalemia was both isoform-and muscle type-specific.

EXPERIMENTAL PROCEDURES
Animals and Diets-To establish the relative levels of Na,K-ATPase isoform expression in a panel of distinct skeletal muscles, male Sprague-Dawley rats (300 g) were anesthetized with 0.2 ml of sodium pentobarbital/100 g of body weight. Muscles were removed, frozen in liquid nitrogen, and stored at Ϫ80°C. To provoke hypokalemia, Sprague-Dawley rats, approximately 8 weeks of age, were placed on a K ϩ -deficient diet (Harlan Teklad, TD 88239, Madison, WI) for 10 days and paired to a control group of rats fed a comparable diet with K ϩ restored (Harlan Teklad, TD 88238).
Preparation of Tissue Homogenates-Skeletal muscle was weighed, minced into small pieces, and homogenized at an approximate 1:20 (w:v) ratio for 45-60 s using a Polytron homogenizer (Brinkmann Instruments) in a buffer containing 5% sorbitol, 25 mM histidine-imi-* This work was supported by National Science Foundation Grant IBN 9S13958 and National Institutes of Health Grant DK 34316. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Immunoprecipitation of ␣␤ Complexes-Immunoprecipitation of nondenatured ␣␤ heterodimers was accomplished using techniques reported previously (23). Briefly, skeletal muscle homogenate (300 g of protein at 3-10 g/l) was incubated for 1 h at 4°C in a lysis buffer, at an approximate 1:3 (v:v) ratio, containing 50 mM Tris-buffered saline at pH 8.0, 1% Nonidet P-40, 1% bovine serum albumin, and proteolytic enzyme inhibitors: 0.2 units/ml aprotinin, and 1.0 mg/ml phenylmethylsulfonyl fluoride. The samples were centrifuged at 250 ϫ g for 10 min, and the supernatant lysate was pretreated with mouse serum-agarose beads and nonspecific goat IgG-agarose beads at an approximate 1:5 bead to supernatant (v:v) ratio, at 4°C for 1 h with gentle rocking. Samples were centrifuged at 2,500 ϫ g for 30 -45 s, the supernatant incubated with IEC 1/48 (24) obtained from A. Quaroni (Cornell University, Ithaca), at a approximate 1:5 antibody to supernatant (v:v) ratio, at 4°C for 4 h with gentle rocking. This monoclonal antibody recognizes undenatured ␤, thus it can be used to immunoprecipitate ␣␤ heterodimers. Goat anti-mouse IgG-agarose beads at 1:5 beads were added to supernatant and the samples incubated at 4°C for 1 h with gentle rocking to isolate ␣␤ complexes. After washing, the precipitate was heated to 90°C for 5 min to dissociate antibody-antigen complexes and ␣ and ␤ subunits. Finally, the samples were resolved by SDS-PAGE, subjected to immunoblot analysis, probed with ␣ and ␤ isoformspecific antibodies, and visualized by the CDP-Star chemiluminescence detection system (Tropix, Bedford, MA) for ␣ subunits or 125 Iprotein A for ␤ subunits.
Plasma and Cellular K ϩ Concentrations-Plasma K ϩ and muscle K ϩ were measured by flame photometry. Specific muscles and cardiac ventricles were quickly dissected, flash frozen in liquid nitrogen, and stored until assayed. They were later thawed, blotted lightly to remove adherent ECF or blood, homogenized in approximately 1:50 (w:v) 0.3 M trichloroacetic acid for 1-2 min using a Tissuemizer (Tekmar, Cincinnati, OH), and centrifuged at 2,000 rpm for 20 min to remove cell debris. K ϩ and Na ϩ were measured using a Radiometer FLM 3 flame photometer (Radiometer, Copenhagen, Denmark), with lithium as internal standard (25).
N-Glycanase Treatment-Sugars were removed from ␤ subunits with peptide N-glycosidase F (N-glycanase, Genzyme Corp., Cambridge, MA). Crude skeletal muscle homogenate (100 g/sample) was centrifuged at 48,000 rpm for 60 min. The pelleted crude membranes were incubated with N-glycanase as described in Azuma et al. (22) to yield a final protein/enzyme ratio of 2.5 units of N-glycanase/100 g of protein.
Quantitation-Autoradiograms were scanned and quantitated using the Bio-Rad GS670 Imaging Densitometer and software. All data are expressed as means Ϯ S.E. Significance was assessed by a two-tailed Student's t test for unpaired samples, and differences were regarded as significant at p Ͻ 0.05.
Materials-Chemicals were reagent grade, spectroquality, or electrophoresis purity reagents. SDS-PAGE reagents were from Bio-Rad. Leupeptin, phenylmethylsulfonyl fluoride, 4-aminobenzamidine dichloride, SDS-PAGE molecular weight standards, and antibodies coupled to agarose beads were purchased from Sigma. Nitrocellulose transfer membrane was from Micron Separations. Immobilon-P transfer membrane was obtained from Millipore. 125 I-Protein A was from ICN. Rabbit anti-mouse secondary antibody was from Calbiochem. Anti-mouse and anti-rabbit alkaline phosphatases were from Tropix. Anti-calsequestrin antibody (IgG2b) was from Affinity Bioreagents.
Typical autoradiograms of rat skeletal muscle homogenates probed with isoform-specific antibodies against ␣1, ␣2, ␤1, and ␤2 are shown in Fig. 1. Whereas ␣1 and ␣2 are expressed in all skeletal muscles, ␤1 was very low in WG and ␤2 barely detected in soleus and diaphragm. The relative levels of expression of Na,K-ATPase subunits are summarized in Fig. 2. The ␣1, ␣2, and ␤1 subunit abundance values were normalized to abundance in soleus muscle, defined as 1. ␤2 was normalized to abundance in RG. ␣1 is 2-4-fold higher in the oxidative rich soleus and diaphragm than in glycolytic WG, EDL, and RG. ␣2 has much less variability than ␣1, is lowest in WG, and twice as high in diaphragm. ␤1 parallels the pattern of ␣1 expression: higher in both fast and slow oxidative rich fibers (soleus, diaphragm, and RG), and barely detected in WG. ␤2 expression, reciprocal to that of ␤1, is highest in muscles containing fast glycolytic fibers (WG, EDL, RG) and is below detection in FIG. 1. Detection of Na,K-ATPase ␣ and ␤ isoform abundance in skeletal muscles. Shown are representative autoradiograms of skeletal muscle homogenates probed with isoform-specific antibodies. Homogenates from each sample (100 g for ␣ and 50 g for ␤ detection) were resolved by SDS-PAGE, blotted, and incubated with isoformspecific antibodies as described under "Experimental Procedures." Ten g of brain homogenate was assayed in parallel as a positive control. ␤ subunit abundance was measured following removal of sugar residues with N-glycanase, described under "Experimental Procedures." The antibody-antigen complexes were detected with 125 I-protein A and autoradiography. gastroc, gastrocnemius.
oxidative dominated muscle (soleus and diaphragm). Consistent with previous observations that no ␣3 mRNA is expressed in mature rat skeletal muscle (6, 10), ␣3 protein was not detected in any individual skeletal muscle. These result suggest that ␣2␤2 is the predominant heterodimer in fast glycolytic muscle (WG) and that both ␣1␤1 and ␣2␤1 heterodimers are expressed in muscles rich in oxidative fibers (soleus and diaphragm). Whether ␣2␤1 and ␣2␤2 are limited to oxidative fibers and fast glycolytic fibers, respectively, cannot be resolved in mixed muscle homogenates.
Characterization of ␣␤ Heterodimers in Soleus-The detection of ␣1, ␣2, and ␤1 but not ␤2 in soleus suggested that both ␣ subunits form heterodimers with ␤1. To test this hypothesis an antibody to ␤1 was used to immunoprecipitate ␣␤ heterodimers. Immunoblot analysis of the immunoprecipitates was used to determine whether both ␣1 and ␣2 were precipitated with the anti-␤1 antiserum; samples of brain and kidney homogenate were assayed directly by immunoblot in parallel as positive controls. Typical autoradiograms of these immunoblots are shown in Fig. 3. As shown in panel A, the IEC 1/48 monoclonal antibody immunoprecipitated ␤1 from both brain and kidney, but did not immunoprecipitate ␤2, demonstrating that the antibody is ␤1-specific. As shown in panel B, the antibody coimmunoprecipitated ␣2 along with ␤1 from soleus muscle, evidence that ␤1 forms heterodimers with ␣2. ␣1 was also detected in these immunoprecipitated samples from soleus (not shown), evidence for ␣1␤1. A corollary to the findings that both ␣1 and ␣2 form heterodimers with ␤1 and that ␤2 is not detected in this muscle is that ␤1 abundance may provide a measure of total sodium pump ␣␤ heterodimer pool size in this muscle.
Effect of Low K ϩ Diet on Na,K-ATPase Isoform Expression in Specific Skeletal Muscles-In 8-week-old rats placed on a K ϩdeficient diet for 10 days, serum K ϩ fell from 4.0 Ϯ 0.14 mM to 2.3 Ϯ 0.08 mM (control animals, n ϭ 12; low K ϩ animals, n ϭ 11). K ϩ -deprived rats gained slightly less weight during the 10 days, attaining a final body weight of 267 Ϯ 0.9 g, which was 7% lower than controls (287 Ϯ 1.9 g). Typical autoradiograms of skeletal muscle homogenates from control and hypokalemia rats are shown in Fig. 4, and the relative changes in Na,K-ATPase ␣ and ␤ subunit isoforms are summarized in Fig. 5. There were no significant changes in ␣1 abundance. In contrast, ␣2 decreased significantly in all skeletal muscles and cardiac left ventricular muscle. The decline in ␣2 was most pronounced in fast glycolytic dominated muscles, falling to 0.06 of control in WG, followed by decreases to 0.39, 0.42, 0.44, and 0.71 of control in RG, EDL, soleus, and diaphragm, respectively. ␤1 decreased 25% or less in RG, EDL, and diaphragm (muscles containing oxidative fibers) to 0.74, 0.76, and 0.83 of control, respectively. However, ␤1 did not decrease significantly in soleus, which is puzzling given that no ␤2 was detected in this muscle, and ␣2 decreased by 56%. This result suggests a dissociation between pool sizes of ␣ and ␤ in this muscle. ␤2 decreased 60% or more in all muscles where it was detected, falling in RG, EDL, and WG to 0.27, 0.40, and 0.30 of control, respectively. These results demonstrate distinct muscle-specific responses to K ϩ deprivation: ␣2␤2 decreases more than 50% in fast glycolytic muscle fibers (WG, EDL, and RG), whereas ␣2 alone decreases in oxidative rich soleus, with only FIG. 3. Coimmunoprecipitation of ␣2␤1. Shown are representative autoradiograms of immunoprecipitated sodium pump subunits analyzed after SDS-PAGE and immunoblot with subunit-specific antibodies. Lane 1, sample immunoprecipitated with anti-␤1 IEC 1/48 antibody; lane 1Ј, supernatant from sample in lane 1 reimmunoprecipitated with IEC 1/48 antibody to establish that enough antibody was added to precipitate most of the ␤ during the first round; lane 2, immunoprecipitation with nonspecific mouse monoclonal antibody rather than IEC 1/48 antibody; lane 3, positive controls for immunoblot detection: direct SDS-PAGE and blot of brain membranes (10 g), kidney membranes (10 g), or soleus homogenate (50 g) without immunoprecipitation. Panel A demonstrates that IEC 1/48 recognizes and immunoprecipitates ␤1 but not ␤2 subunit from brain and kidney. Panel B demonstrates that ␣2 is immunoprecipitated along with ␤1 in soleus and brain samples.

FIG. 2. Relative levels of Na,K-
ATPase ␣ and ␤ isoform abundance in skeletal muscles. The relative abundance of ␣ and ␤ isoform proteins was assessed by quantitating immunoblots of skeletal muscle as described under "Experimental Procedures." Data are expressed in arbitrary units, normalized to a standard defined as 1. ␣1, ␣2, and ␤1 are normalized to abundance in soleus, and ␤2 is normalized to abundance in RG. ␣1 and ␣2 panels summarize results from five rats, and ␤1 and ␤2 panels summarize results from three rats, all expressed as mean Ϯ S.E. Sol, soleus; Dia, diaphragm. a minor decrease in ␣2␤1 in mixed oxidative diaphragm. Cardiac ventricle was assayed in parallel, and in agreement with previous findings (11) ␣2 but not ␣1 abundance decreased to 0.69 of control.
Since changes in ␣2 abundance were so pronounced, we considered the possibility that the NH 2 -terminal epitope recognized by the monoclonal McB2 had been cleaved. Blotting with an additional anti-␣2 polyclonal antibody, anti-HERED, which recognizes a sequence of amino acids (491-495) found on the early part of the large cytoplasmic loop, between transmembrane segments H4/H5, and unique to ␣2 isoforms (12,18), confirmed that the decrease in ␣2 signal was not due to cleavage of the epitope (Fig. 6). The pattern of change in ␣2 in response to K ϩ deprivation is identical with both antibodies.
Effect of Low K ϩ Diet on ICF K ϩ in Control versus Hypokalemic Rats-Based on the distinct muscle-specific changes in Na,K-ATPase expression in hypokalemia we predicted that intracellular K ϩ would fall more in WG, which lost greater than 70% of both ␣2 and ␤2, than in diaphragm, which lost only 30% of ␣2 and ␤1. To test this hypothesis, ICF K ϩ was measured in paired muscles from the same pool of rats used for subunit abundance analysis. Fig. 7 summarizes ICF K ϩ levels expressed as mol/g, wet weight, in skeletal muscles from control and K ϩ -deprived rats. In contrast to our prediction, by day 10 all skeletal muscles lost about 20% of ICF K ϩ . Additionally, ICF K ϩ measured in left ventricles from the same animals did not change. These results suggest a revised hypothesis, that muscles may have multiple mechanisms in place to regulate intracellular K ϩ loss besides a decrease in pump number, such as pump translocation to internal stores, as described previously for ␣2␤1 in soleus in response to insulin stimulation (30,31). DISCUSSION The existence of isoforms suggests the potential for differential isoform-specific function, tissue-specific expression, and regulation. Functional differences in enzymatic and transport properties between the ␣1 and ␣2 isoforms expressed in skeletal muscle are subtle if any (32). This study demonstrates a muscle type-specific pattern of Na,K-ATPase isoform expression and distinct muscle-specific patterns of regulation of Na,K-ATPase isoforms in response to K ϩ deprivation.
Regarding muscle-specific expression, Hundal et al. (9) reported that pooled membranes from muscles enriched in slow twitch oxidative fibers expressed ␤1, not ␤2; that pooled membranes from muscles composed of fast twitch glycolytic fibers expressed ␤2, not ␤1 isoform; and that ␣1 and ␣2 were similar in the two types of muscle membranes. This distinction prompted our examination of ␣ and ␤ isoform expression in five distinct skeletal muscles. ␣1 and ␤1 were detected in all muscles and were at least twice as abundant in oxidative fiber rich soleus and diaphragm than in fast twitch glycolytic WG, whereas ␣2 expression was equivalent in all five muscles. In agreement with Hundal (9), ␤2 expression was limited to fast glycolytic muscle types. Additionally, we note that ␤2 expression appears directly proportional to the relative percentage of fast glycolytic fibers, that is, ␤2 is most abundant, relative to cellular protein, in WG, which is 84% fast glycolytic; less abundant in EDL, which is 56% fast glycolytic; and below detection in slow oxidative soleus, which has no fast glycolytic fibers. In contrast to the Hundal study, we detected significant levels of ␤1 in EDL, an almost pure fast muscle with 42% fast glycolytic- FIG. 4. Detection of Na,K-ATPase ␣ and ␤ isoform abundance in skeletal muscles from control and K ؉ -deprived rats. Shown are representative autoradiograms of skeletal muscle homogenates from control and hypokalemic rats probed with isoform-specific antibodies to ␣1, ␣2, ␤1, and ␤2. ␤ subunit abundance was measured following removal of sugar residues with N-glycanase, as described under "Experimental Procedures." Homogenate from each sample (100 g for ␣ and 50 g for ␤) was assayed as described in Fig. 1.   FIG. 5. Effect of K ؉ deprivation on Na,K-ATPase ␣ and ␤ isoform abundance in skeletal muscles. The relative abundance of ␣ and ␤ isoforms was assessed as described in the legend for Fig.  2. Values for K ϩ -deprived muscle samples are expressed relative to mean of control values, defined as 1. Results are expressed as mean Ϯ S.E. * ϭ p Ͻ 0.05, n ϭ 6 for both control and K ϩ -deprived groups. Sol, soleus; Dia, diaphragm. oxidative fibers. This may be a consequence of deglycosylation of EDL samples prior to analysis, allowing for enhanced ␤1 subunit detection in this study. We conclude that ␣1, ␤1, and ␣2 are expressed ubiquitously in skeletal muscle, ␣1 and ␤1 enriched in oxidative fibers and ␣2 about the same relative abundance across muscle, whereas ␤2 is expressed in only a subset of muscles containing fast twitch glycolytic fibers.
The expression of two ␣ and two ␤ isoforms suggests the potential for four different ␣␤ heterodimer combinations unless constraints prevent an ␣ from combining with a ␤ subunit. We know from both purification and immunoprecipitation studies that ␣1␤1 heterodimers are the active sodium pumps in kidney and epithelial cell lines (23). The immunoprecipitation of both ␣1␤1 and ␣2␤1 from soleus muscle in this study provides in vivo evidence that both ␣s complex with ␤1. The corollary to this finding is that if nearly all the ␤ is assembled as heterodimers, ␤1 subunit abundance will provide a relative measure of total sodium pump pool size in soleus, since it lacks ␤2. The work of Gloor et al. (33) demonstrates that glial ␤2 forms heterodimers with ␣2 but not ␣1 subunits. Whether ␤2 will be constrained to form heterodimers with only ␣2 in skeletal muscles remains to be established.
In our initial investigation into the effect of hypokalemia on Na,K-ATPase expression we reported decreased abundance of ␣2, but not ␣1 or ␤1, in pooled hindlimb skeletal muscle, heart, and brain. The changes in skeletal muscle ␣2 were by far the greatest: protein and mRNA levels decreased 80 and 40%, respectively, after 14 days of K ϩ deprivation (11). This study examined whether changes were restricted to a subset of muscle types found in the rat hindlimb. The results establish that ␤2 protein levels are also depressed more than 50% during 10 days of K ϩ deprivation in all muscles expressing this isoform (RG, EDL, and WG). The failure of our prior study to detect changes in ␤1 is likely a consequence of enhanced sensitivity of ␤1 subunit detection following deglycosylation of muscle samples prior to analysis. Overall, three distinct responses to K ϩ deprivation were observed in the panel of muscles: 1) in WG the response was the greatest, ␣2 and ␤2 decreased 94 and 70%, respectively; 2) in RG and EDL ␣2 and ␤2 decreased about 60%, and ␤1 decreased about 25%; and 3) in soleus and diaphragm ␣2 decreased 55 and 25%, respectively, with little if any accompanying change in ␤1 and no detection of ␤2. From this pattern we predicted that the WG would lose more intracellular K ϩ than soleus or diaphragm. However, intracellular K ϩ decreased equivalently in all five skeletal muscles after 10 days of K ϩ deprivation. In comparison, while we observed a 31% decrease in ␣2 abundance in K ϩ -deprived cardiac left ventricular muscle, there was no significant decrease in cardiac intracellular K ϩ , which can be attributed to the low percentage of ␣2 type pumps (25% or less) in the heart (6). Norgaard et al. (34) observed that after 4 weeks of K ϩ -deficient diet there was only a 13% fall in left ventricle ICF K ϩ and a concomitant 43% decrease in [ 3 H]ouabain binding (a measure of ␣2 abundance at the plasma membrane), the difference likely due to the longer deprivation period. Taken together, the results of this study demonstrate that the abundance of ␣2 and ␤2, and not the ubiquitous ␣1 isoform, is depressed in hypokalemia, and demonstrate small but significant decreases in ␤1 in a subset of muscles assayed, which is in contrast to our previous report of no change in ␤1 (11). These results, coupled with information on the relative ratios of ␣1 to ␣2 protein in various tissues, provide a molecular explanation for the significant and selective loss of K ϩ from skeletal muscle and only minor loss of K ϩ from cardiac muscle. We hypothesize that the presence and regulation of the ␣2 isoform in skeletal muscle provided an evolutionary advantage to complex organisms that needed to maintain transmembrane K ϩ gradients in the face of fluctuations in K ϩ availability (3).
We have hypothesized that expressing Na,K-ATPase subunits in a fiber type-specific pattern would provide an organism with the means to regulate fiber type involvement in the regulatory response to a K ϩ challenge (35). Specifically, one subset of muscles may be adapted to respond to hypokalemia by expressing isoforms that respond to insulin and rapidly move dietary K ϩ from ECF to ICF after a meal to avert cardiac complications from elevated extracellular K ϩ , while another subset of muscles may be adapted to respond to hypokalemia by expressing isoforms that decrease in response to chronic K ϩ deprivation leading to the loss of K ϩ from the ICF to the ECF. Parallel specialization would allow an animal both to chronically lose K ϩ during deprivation and rapidly clear a ECF K ϩ load when a meal (and insulin surge) ends the K ϩ deprivation. Recent evidence from Lavoie et al. (31) supports this hypothesis by demonstrating that insulin increases Na,K-ATPase transport activity only in oxidative fiber type skeletal muscles (for example, soleus and most RG) by inducing translocation of ␣2 and ␤1 from intracellular to plasma membranes. Additional support comes from the study of Hsu and Guidotti (36), demonstrating that this insulin-stimulated K ϩ uptake response is still present in a hypokalemic animal with depressed sodium pump expression. Arguing against specialization during K ϩ deprivation are the findings of this study. After 10 days of K ϩ deprivation, the oxidative fiber muscle types lose as much intracellular K ϩ as the fast glycolytic muscle even though there are much smaller decreases in ␣ and ␤ subunit pool sizes in oxidative fiber types. One hypothesis that reconciles these findings is that the molecular mechanisms mediating the responses may be muscle fiber type-specific. For example, sodium pump transport activity may be depressed in slow oxidative fiber muscle types by both decreasing abundance and translocation of the predominant ␣2␤1 heterodimers from plasma membrane to internal stores (where they will be poised to respond to insulin), while sodium pump transport activity in fast glycolytic muscle fiber types, where ␣2␤2 heterodimers predominate, may be depressed solely by decreasing pool sizes of this isoform. Such a scenario would allow skeletal muscle K ϩ stores to buffer the fall in extracellular K ϩ during K ϩ deprivation while retaining the ability to clear a potassium load from the ECF into the skeletal muscle during hyperkalemia.