Kidney International, Vol. 61 (2002), pp. 967–974

ION CHANNELS – MEMBRANE TRANSPORT – INTEGRATIVE PHYSIOLOGY

The �-Na�,K�-ATPase subunit assembles selectively with�1/�1-Na�,K�-ATPase but not with the colonic H�,K�-ATPase

JUAN CODINA, JIAN LI, YAN HONG, and THOMAS D. DUBOSE, JR.

Department of Internal Medicine and Kidney Institute, University of Kansas School of Medicine, Kansas City, Kansas, USA

The �-Na�,K�-ATPase subunit assembles selectively with unit assembly. These functions include a decrease in Na�

�1/�1-Na�,K�-ATPase but not with the colonic H�,K�-ATPase. and K� affinity [4], induction of cation channel activityBackground. The ubiquitous Na�-pump (Na�,K�-ATPase) [5] and an increase the apparent affinity of the enzymeassembles as a heterodimer of composition �/� in some neph-

for adenosine 5�-triphosphate (ATP) [6].ron segments, while in other segments it may exist as a hetero-The �1, as a member of the X�,K�-ATPase superfamilytrimer of composition �/�/�. The �-subunit has been reported

to increase the affinity of the Na�-pump for adenosine 5�-triphos- of transport proteins, displays 75% sequence similarity atphate (ATP), and decrease affinity for both Na� and K�. The the amino acid level with cHK. Moreover, both �-subunits�-subunit of the colonic H�,K�-ATPase (cHK) shares 75% share and assemble stably with the same �1 [7, 8]. Whilesequence similarity with �1-Na�,K�-ATPase (�1) and assembles

cHK protein is expressed in low abundance in the renalwith �1-Na�,K�-ATPase (�1) in distal colon and renal medulla.medulla [9, 10] this transporter is up-regulated selectivelyDifferences in pharmacological properties have been ascribed

to when heterologously expressed function has been compared in this region by chronic hypokalemia [10, 11]. Severalto function in vitro. The purpose of this study was to determine groups have demonstrated that cHK, when expressed het-if cHK might associate with the �-subunit of the Na�,K�-ATPase

erogeneously, is insensitive to Sch-28080, a relatively spe-(�) as a possible explanation for these variations in function.cific inhibitor of the gastric H�,K�-ATPase (gHK), andMethods. An antibody specific for the � was used in coimmuno-

precipitation experiments to determine if the � assembles stably is partially sensitive to ouabain, to which gHK is resistantin vitro with cHK and �1 in rat renal medulla or distal colon. [12–14]. Discrepancies exist, however, when functional

Results. Our results demonstrate that the �-subunit assem- properties defined by studies performed in vitro are com-bles specifically with the Na�-pump, but not with cHK. Further-

pared with studies performed in heterologous expressionmore, the �-subunit assembly was specific for rat kidney andsystems. Chronic dietary K�-depletion was shown to in-was not observed in distal colon.

Conclusion. Since the �-subunit did not assemble with the crease the fraction of bicarbonate absorption (JCO2) sensi-

cHK/�1, complex, �-subunit assembly cannot explain those tive to Sch-28080 in rat isolated perfused collecting ductvariations in ex vivo and in vitro pharmacologic properties segments [15, 16]. This increase in JCO2

has been assumedascribed to cHK.to be the result of up-regulation of gHK [17, 18]. Neverthe-less, neither Northern blot or immunoblot analysis re-vealed changes in gHK mRNA or protein abundance inIt has been widely accepted that to be functional, therat renal medulla during chronic hypokalemia [9, 10]. InNa�,K�-ATPase requires assembly of � and � subunits.contrast, several groups have reported a selective in-The catalytic �-subunit is responsible for the secretioncrease in abundance of both cHK mRNA and proteinof Na� in exchange for K� and contains the ouabain-during hypokalemia that was localized specifically to thebinding site, while the �-subunit is responsible for migra-medullary collecting tubule [11]. However, our labora-tion and proper assembly of the �-subunit into the

plasma membrane [1, 2]. A third subunit (�-subunit) has tory [16] and others [19] have shown that a significantbeen identified in association with the Na�,K�-ATPase component of the JCO2

in IMCD, which is up-regulatedin certain regions of the nephron [3, 4]. by chronic hypokalemia, is sensitive to ouabain. These

Several functional properties were conferred on �-sub- data have been interpreted to suggest the induction ofa non-gastric H�,K�-ATPase with unique pharmacologicsensitivities to Sch-28080 and ouabain [15, 16, 19], orKey words: kidney, subunit composition, colonic H�,K�-ATPase,

Na�,K�-ATPase, immunoprecipitation, distal colon, sodium pump. induction of a unique isoform or activity (type III) forwhich there is, at present, no identification at the molecu-Received for publication May 30, 2001lar level [20–22].and in revised form August 13, 2001

Accepted for publication October 1, 2001 Although it is generally accepted that in order to func-tion, cHK must assemble as a heterodimer of composition 2002 by the International Society of Nephrology

967

Codina et al: c-subunit of the Na�,K�-ATPase968

cHK/�, it is not clear if, in some cases, the �/� dimer might The particulate matter was removed by centrifugationfor two minutes at 200 � g. To the supernatant, 1 mol/Lassociate with a �-subunit. Assembly of an ��� hetero-

trimer conceivably could confer unique pharmacological MgSO4 was added to a final concentration of 10 mmol/LMgSO4. The sample was placed on ice for 15 minutesproperties which might explain such in vitro/ex vivo varia-

tions in functionality. Indeed, such alteration in functional and was shaken intermittently. The aggregated materialwas removed by centrifugation for 12 minutes at 2500 � gproperties has been observed for the Na�,K�-ATPase

[4–6]. at 4�C and apical enriched membranes from the superna-tant were collected by centrifugation at 27,000 � g forThe present study was designed to determine whether,

under physiological conditions, cHK associates in renal 20 minutes at 4�C. The pellet was resuspended in 5 mLbuffer B containing 10 mmol/L MgSO4, and homoge-medulla and/or in distal colon with the �-subunit to provide

a possible explanation for such differences in function. nized with pestle B of a Dounce homogenizer. Afterremoval of the aggregated material at 3100 � g for 12minutes at 4�C, membranes were collected by centrifuga-

METHODStion at 27,000 � g for 20 minutes at 4�C. The final mem-

Experimental animals branes were resuspended in buffer A and stored in ali-quots at �70�C. The protein concentration was measuredMale Sprague-Dawley rats weighing between 150 and

180 g had free access to drinking water and food. Chronic according to the Lowry method [25].hypokalemia was induced by feeding the animals a diet

Anti-� antibodydeficient in K� (cat # 960189; ICN, Cleveland, OH, USA)for 14 days [23]. Rats were sacrificed and their organs A synthetic peptide (KHRQVNEDEL) that extends

from amino acid 57 to 66 of the published sequence of �(distal colon, kidney cortex and medulla) were dissectedas quickly as possible, and membranes were prepared [6] was synthesized (Genosys Biotechnology, Inc.). The

peptide was cross-linked to KLH. Two rabbits were in-as described below.jected and bled according to standard protocols. The �

Preparation of plasma membranes is a member of a family of small membrane proteins that,upon expression in Xenopus laevis oocytes, induce slowlyTo prepare plasma membranes [8, 24], the rat distal

colon (1 g tissue) was homogenized in 10 mL buffer A activating voltage-dependent currents. Other membersof this family include CHIF [27], mat-8 [28] and phospho-[10 mmol/L Tris-HCl, pH 8.0, 1 mmol/L ethylenedi-

aminetetraacetic acid (EDTA)-Tris, 1 mmol/L phenyl- lemman [29]. Our immunizing peptide was only 10%similar to rat CHIF and mat-8 and, 30% similar to phos-methylsulfonyl fluoride (PMSF), 3 mmol/L benzamidine,

and 1 g/mL soybean trypsin inhibitor] containing 27% pholemman.The sensitivity and specificity of the antibodies againstsucrose (wt/vol). Homogenization was performed using

a Kinematica polytron, Model PT 10/35, with the poly- � was defined by immunoblot analysis of � synthesizedin vitro using rabbit reticulocyte lysate [30]. The recombi-tron set at 5. Tissues were homogenized for 10 seconds

at 4�C, the samples were placed on ice for 15 seconds, nant � displays mobility similar to hemoglobin on SDS-PAGE, which makes resolution of the protein difficult,and then homogenized again. This was followed by ho-

mogenization in a Dounce homogenizer using pestle A or impossible. This problem was avoided by subcloning,in frame, the NcoI/NcoI fragment (987 bp) of the amino(4 to 5 strokes). Nuclei were removed by centrifugation

at 2000 � g for four minutes at 4�C. The supernatant terminal of gHK to the NcoI site of the initial methionineof � in the plasmid pAGA#2. The open reading framewas applied to the top of 45% (wt/wt) sucrose in buffer

A and was centrifuged at 200,000 � g for 45 minutes at of the construct (�NK[gHK]) encodes a protein with anexpected mobility of approximately 40 kD. The resulting4�C. The membranes in the interphase 27/45% sucrose

were diluted in buffer A and collected by centrifugation protein can be resolved clearly from hemoglobin.at 25,000 � g, resuspended in buffer A, and stored in

Immunoblotsaliquots at �70�C. The final protein concentration wasmeasured using the method of Lowry [25]. Proteins were dissolved in Laemmli sample buffer [31],

separated on a 10 or 15% sodium dodecyl sulfate-poly-Preparation of apical enriched membranes acrylamide gel electrophoresis (SDS-PAGE) and trans-

ferred overnight to a nitrocellulose membrane [10]. Non-Membranes from distal colon, kidney cortex and me-dulla were prepared as described by Aronson [26]. Distal specific binding was blocked for one hour at room

temperature with 10 mmol/L Na-phosphate, pH 7.5,colon (1 g in 10 mL buffer B, consisting of 300 mmol/Lmanitol, 1 mmol/L Tris-HCl pH 7.2, 1 mmol/L PMSF, 150 mmol/L NaCl and 0.05% Tween-20 containing 1%

non fat dry milk. The primary antibody was diluted in3 mmol/L benzamidine and 1 g/mL soybean trypsininhibitor) was homogenized with a polytron, followed the same buffer but in the absence of non-fat dry milk.

Membranes were incubated for two hours at room tem-by 4 to 5 strokes with pestle A of a Dounce homogenizer.

Codina et al: c-subunit of the Na�,K�-ATPase 969

perature, washed three times for 10 minutes each andincubated with a donkey anti-rabbit IgG (cat# NA934;Amersham, Piscataway, NJ, USA) for one hour at roomtemperature. After extensive washing the bands weredetected using an enhanced chemiluminescence (ECL)system (cat. #54-61-00; Kirkegaard and Perry Labora-tories, Gaithersburg, MD, USA) following the manufac-turer’s instructions.

Assembly of � with cHK/�1 complex

�-Subunit assembly to the cHK/�1 complex was de-tected by coimmunoprecipitation of the cHK/�1/� com-plex using an anti-cHK antibody [8, 10]. Immune serum(150 L) was incubated overnight at 4�C with 5 mg (1 mg/ Fig. 1. mRNA for � is expressed in kidney only. mRNA (20 g) iso-

lated form rat kidney, distal colon, stomach and brain where hybridizedmL) of rat renal medulla plasma membranes preparedas described in the Methods section using a probe specific for �-subunit.from rats made chronically hypokalemic, as described

above. The membranes were rinsed twice with 1 mL ofbuffer C (10 mmol/L Tris-HCl, pH 8.0, 150 mmol/LNaCl, 1 mmol/L PMSF, 3 mmol/L benzamidine and 1 g/ reaction (PCR), the complete open reading frame of theL soybean trypsin inhibitor). Membrane proteins were � from a rat kidney cDNA library in �gt11 (cat# RL3002b;extracted with 500 L buffer C containing 1% CHAPS Clontech, Palo Alto, CA, USA). The PCR product wasfor one hour at 4�C. The insoluble material was removed subcloned in the PCR2.1 vector (cat# K-2030-01; Invitro-by centrifugation and the cHK/anti-cHK antibody com- gen, Carlsbad, CA, USA). The insert was sequenced andplex was precipitated by addition of 40 L protein A/G the fragment EcoRI/EcoRI was used as a probe to deter-Plus Agarose (cat# sc-2003; Santa Cruz Biotechnology, mine the presence or absence of � mRNA in differentSanta Cruz, CA, USA) for two hours at 4�C [12]. The tissues, as described previously by our laboratory [23, 32].resin was washed extensively and the bound protein was Loading control, quality of the RNA and efficiency ofdeglycosylated with PNGase F (25,000 U/mL; cat# 704L; transfer to nitrocellulose membranes were monitored byNew England Biolabs, Beverly, MA, USA) for one hour

visualization of transferred RNA under ultraviolet light.at 37�C or with a mixture of PNGase (25,000 U/mL),endo-�-N-acetylgalactosaminidase (0.03 U/mL; cat# E2391; Sigma, Saint Louis, MO, USA), neuramidase (0.07 RESULTSU/mL; cat# N7885; Sigma) and �-l-fucosidase (0.05 Tissue distribution of �-subunit mRNAU/mL; cat# F7753; Sigma) for three hours at 37�C in the

Total RNA (20 g) purified from rat kidney, distalpresence of Na-phosphate, pH 6.8 (10 mmol/L), PMSFcolon, stomach and brain were separated on an agarose(1 mmol/L), benzamidine (3 mmol/L), soybean trypsingel, transferred to a nylon membrane as described pre-inhibitor (1 g/mL) and CHAPS 1%, as described pre-viously by our laboratory [23] and probed with the fullviously by our laboratory [8]. The bound protein waslength open reading frame of the �. Figure 1 demon-extracted with Laemmli buffer [31], and separated in astrates that m-RNA for the �-subunit is expressed in15% SDS-PAGE. The presence of a �1-subunit in thethe kidney only. Based on these observations, the nextimmunoprecipitated protein was detected with a mono-experiments were performed using kidney membranes.clonal anti-�1 antibody (cat# 05-382; Upstate Biotechnol-

ogy, Lake Placid, NY, USA) as described previously byCharacterization of the anti-� antibody usingour laboratory [8]. The presence of � was detected usingrecombinant and native proteinsour antibody against � as described above.

Recombinant � synthesized in rabbit reticulocyte ly-Northern blots for � sate was used to define the specificity and sensitivity of

our anti-� antibody, as described in the Methods section.Two oligonucletides designed after the rat � sequenceWe subcloned, in frame, the NcoI/NcoI fragment (987published by Minor et al [5] were synthesized by Genosysbp) of the amino terminal of gHK to the NcoI site ofBiotechnology. The sense oligonucleotide (5�-ATGACthe starting methionine of the � in the plasmid pAGA#2.AGAGCTGTCAGC TAACCATG-3�) extends betweenThe open reading frame of the construct (�NK[gHK])bases �1 and �25 and, the antisense oligonucleotideencodes a protein with an expected mobility near 40 kD.(5�-TCACAGCTCATCTTC ATTGAC-3�) extends fromThe results of a representative experiment are displayedbases �181 to �201 of the �A sequence [6]. These oligo-

nucleotides were used to amplify, by polymerase chain in Figure 2A. �NK[gHK] and �1 synthesized in rabbit

Codina et al: c-subunit of the Na�,K�-ATPase970

Fig. 2. Specificity of anti-� antibody. (A) Re-combinant [�HK�1]�NK (7 ng) and recombi-nant �1 (8 ng) were synthesized in vitro usingrabbit reticulocyte lysate. Proteins were re-solved on a 10% SDS-PAGE, transferred toa nitrocellulose membrane, and blotted withthe anti-� antibody (1:1000). (B) Renal plasmaand apical enriched membranes from cortexand medulla (50 g) were resolved on a 15%SDS-PAGE, transferred to a nitrocellulosemembranes and immunoblotted with anti-�antibody (1:750).

reticulocyte lysate were dissolved in Laemmli sample chronic hypokalemia [9, 10]. Although it is logical toassume that most of the �-protein detected in the apicalbuffer [31] and the immunoblot was performed as de-

scribed in the Methods section. The results demonstrate enriched membranes represented contamination frommedullary thick ascending limb basolateral membranes,that the immune serum recognized �NK[gHK], as ex-

pected. The signal was blocked upon preincubation of it is possible that a small fraction of the �-subunit de-tected in the apical enriched membranes could assemblethe immune serum with the immunizing peptide (data

not shown). However, the antibody did not recognize with the cHK/�1 complex. This possibility was exploredby co-immunoprecipitation of cHK/�1/� complex withthe recombinant �1, used as a negative control. By addi-

tion of [35S]-methionine during the synthesis of the re- an antibody specific for cHK [8].combinant proteins, we detected between 2 and 5 ng of

Is the �-subunit an integral component of cHK/�1�NK[gHK].complex in renal medulla?We reasoned that if the � is an integral component of

cHK/�1 complex it also must be expressed in an enriched Based on the observation that our anti-� antibodyrecognizes recombinant and native proteins, we sought tofraction of apical membranes prepared from renal me-

dulla. Based on this hypothesis, we studied the distribu- determine if the � is an integral component of cHK/�1

complex in the renal medulla. Apical enriched mem-tion of � in apical enriched and plasma membranes pre-pared from renal cortex and renal medulla. Protein (50 g) branes obtained from chronically hypokalemic rats were

purified as described in the Methods section. Membraneswas resolved on a SDS-PAGE (15%) and the immuno-blots were performed with the anti-rat-� antibody. The 10 mg (1 mg/mL) were incubated with preimmune serum

(150 L) or anti-cHK antibody (150 L) overnight at 4�Cresults of a representative experiment are displayed inFigure 2B. The antibody detected a doublet (�A and �B) in buffer C (see Methods). The membranes were washed

extensively and immunoprecipitated as described. Figurein the 8 to 14 kD molecular weight range in all membranepreparations from rat kidney. The upper band (�A) was 3 (left panel) demonstrates that the � was not detected

by our antibody (1:250) in samples immunoprecipitatedmore abundant than the lower band except with apicalenriched membranes of renal medullary origin, where with the anti-cHK antibody. However, Figure 3 (right

panel) demonstrates, as expected, that �1 was immuno-the lower band (�B) was more abundant. Both the upperand the lower bands were blocked when the immune precipitated by our anti-cHK antibody [8]. It is generally

accepted that the �-subunit is not a glycosylated proteinserum was preincubated for one hour at 4�C with immu-nizing peptide (data not shown). [5, 33]. To rule out the unlikely possibility that in the

renal medulla � is a heavily glycosylated protein (withThe �-subunit was expressed in an apical enrichedmembrane fraction prepared from renal medulla, where O-glycosylation), and thus not accessible to our antibody,

the immunoprecipitation experiment was repeated withcHK protein is also expressed and up-regulated during

Codina et al: c-subunit of the Na�,K�-ATPase 971

Fig. 3. The � is not immunoprecipitated byanti-cHK antibody from apical enriched mem-branes prepared from LK rats. (Left panel)Renal medulla apical enriched membranes (50g) were used as a control. In the middle lane,renal medulla apical enriched membranes (10mg) were incubated with anti-cHK antibody(125 L) overnight at 4�C, as described inthe Methods section. The immunoprecipitatedproteins were deglycosylated and resolved ona 15% SDS-PAGE and hybridized against theanti-�-Na�,K�-antibody (dilution 1:250). Inthe right lane, preimmune serum was used inthe immunoprecipitation experiment. (Rightpanel) Left lane, �1 was immunoprecipitatedby the anti-cHK antibody. Right lane, �1 wasnot immunoprecipitated when incubated withpreimmune serum.

the anti-cHK antibody. Prior to the solubilization of the the immune serum was preincubated with immunizingpeptide (second lane). In the same experiment the preim-sample with Laemmli sample buffer we treated the im-

munoprecipitated samples with a mixture of PNGase F, mune serum did not immunoprecipitated �1 (third lane).Figure 4B demonstrates that the anti-� antibody immu-endo-�-N-acetylgalactosaminidase, neuraminidase, �-l-

fucosidase and O-glycosydase as described previously by noprecipitates �1 as well as �1 (lane 1). As expected,immunoprecipitation of this subunit was blocked byour laboratory, to “fully” deglycosylate the �1 assembled

with cHK in rat distal colon [34]. As expected, the results the immunizing peptide (lane 2). Preimmune serum didnot immunoprecipitate �1 (lane 3).demonstrate that the �-subunit was not immunoprecipi-

tated by anti-cHK antibody (data not shown).Sensitivity of anti-� and anti-�1 antibodiesIn the experiments described above we demonstrated

that the �-subunit does not assemble with cHK. To rule Our immunoprecipitation experiments demonstratedthat anti-cHK antibody could immunoprecipitate �1 butout the possibility that CHAPS (1%) disrupted the sta-

bility of a possible ��� complex during incubation, the �-subunit was not detected. A potential explanation forthese data is that anti-� antibody, as used in these studies,following experiment was performed. Rat renal medulla

plasma membranes (3.0 mg) were diluted to 1 mg/mg in was not sufficiently sensitive. To rule out this possibilitywe performed the experiment displayed in Figure 5. �1the presence of buffer C. Membranes were divided into

three groups. The first group was incubated overnight and �NK[gHK] proteins were synthesized using rabbitreticulocyte lysate [30]. Each aliquot was subdivided inat 4�C with 125 L anti-� antibody. The second group

was incubated with 125 L of immune serum previously two identical groups. To one, [35S]-methionine (1.5 Ci/mmol) was added to quantify protein concentration [30].preabsorbed with immunizing peptide (500 mol/L) and

the third group was incubated with 125 L of preimmune The second aliquot was used to synthesize protein in theabsence of [35S]-methionine. The synthesized proteinsserum. The membranes were washed with buffer C, and

complex bound to anti-� antibody was extracted with were resolved on a 10% SDS-PAGE, transferred to anitrocellulose membrane and the presence of �- or �-sub-CHAPS (1%). The insoluble material was removed by

centrifugation and, that bound to anti-� antibody was units was detected by immunoblot analysis (dilution1:250 for anti-�, and 1:1000 for anti-�1). These conditionsprecipitated with agarose A/G plus, as described in the

Methods section. The bound proteins were deglycosy- were designed to simulate our immunoprecipitation ex-periments.lated with PNGase F and extracted with Laemmli sample

buffer [31]. Proteins were resolved on a 10% SDS- Figure 5, demonstrates that using recombinant pro-teins (which are not glycosylated proteins and fully ac-PAGE, and the immunoblot was performed using a

monoclonal antibody against �1 (cat# 05-382; Upstate cessible to the antibody), the sensitivity of the anti-�antibody is equal to or higher than the sensitivity of theBiotechnology, Upstate, NY, USA) or a monoclonal an-

tibody against �1 (cat# 05-369; Upstate Biotechnology). anti-�1 antibody. The native �1 associated in kidney withthe cHK is an N-glycosylated protein [8]. Although itThe results of a representative experiment are displayed

in Figure 4. Figure 4A demonstrates that the anti-� anti- appears that our assay deglycosylates the majority of the�1 associated with cHK, the possibility that a fraction ofbody immunoprecipitated �1 (first lane from the left), as

expected, and the immunoprecipitation was blocked if protein not fully deglycosylated persists, and may not be

Codina et al: c-subunit of the Na�,K�-ATPase972

Fig. 4. The �1 (A) and �1 (B)-subunits areimmunoprecipitated by anti-� antibody. Lane1 (positive control): The �-subunit from renalmedulla apical enriched membranes (1 mg)was immunoprecipitated with the anti-� anti-body. The presence of �1 or �1 in the immu-noprecipitated sample was detected by immuno-blot analysis as described under “Methods.”Lane 2: The immunoprecipitation of the �1 or�1 by anti-� antibody was blocked by immuniz-ing peptide. Lane 3: The �1 or �1 was notimmunoprecipitated by preimmune serum.

colon and renal medulla, might serve as an explanationfor variations in ex vivo and in vitro pharmacologicalsensitivities. Our study suggests that the renal Na�-pumpmay be the only X�,K�-ATPase that associates with the�-subunit in these tissues.

As indicated previously, two in vitro techniques, bicar-bonate absorption (JCO2

) in isolated perfused medullarycollecting tubules, and pHi-recovery after an NH4Cl loadin renal medullary cells in culture, have been used to inves-tigate the mechanism of K�-absorption during chronic

Fig. 5. The sensitivities of anti-� and anti-�1 antibodies are similar. hypokalemia [15, 16]. Both studies demonstrated thatRabbit reticulocyte lysate synthesized [gHK]�NK- and �1 (0.2 pmol/

with chronic hypokalemia a unique pharmacological pro-lane) were immunoblotted against anti-� (dilution 1:250) and anti-�1

antibodies. file emerged. Namely, a significant component of the in-cremental up-regulation of JCO2

was sensitive to both Sch-28080 and ouabain, or combinations thereof. For example,Wall, Mehta and DuBose demonstrated that a ouabain-

accessible to our anti-�1 antibody. Nevertheless, the sensitive component could account for at least 20% ofstudy using recombinant proteins indicates clearly that JCO2

when ouabain was added to Sch-28080 [16]. Thisif the �-subunit did assemble with the cHK/�1 complex, profile for bicarbonate absorption in terminal inner med-it would be expected to display a stronger signal than �1 ullary collecting duct (IMCD) is compatible with featuresin the immunoprecipitation experiments. Since the data of the unique K�-ATPase activity reported by Doucet’sin Figure 3 show that the �-subunit was not detected as

group [20, 22] for which no molecular counterpart hasan integral component of cHK/�1 complex in the renal

been discovered. These functional properties not with-medulla, we conclude that � was not present.standing, chronic hypokalemia has been associated inseveral reports with an increase in cHK mRNA and

DISCUSSION protein abundance in the renal medulla [10, 23], wherechanges in the level of expression of gHK mRNA andOur results support the view that the � can serve asprotein were not observed [10, 23].an integral component of the �1/�1 complex in specific

Recently, Sangan et al suggested that �3 Na�,K�-regions of the renal medulla [4]. Nevertheless, the func-ATPase is the specific �-subunit for cHK [35]. Studiestional contribution of the �-subunit in vivo, which hasby our laboratory also have shown that both mRNA andbeen inferred from expression studies in heterologousprotein for �3 are present in the distal colon and in thesystems, remains unclear. Putative functions include: (a)renal medulla (unpublished observations). Nevertheless,regulation of the affinity of the Na�,K�-ATPase for Na�

the cHK/�1complex, but not a cHK/�3 complex, was de-and K� [4], (b) regulation of the affinity of the enzymetected by coimmunoprecipitation in renal medulla duringfor ATP [6] and, (c) induction of K�-channel activityboth control conditions and after chronic hypokalemia.[5]. The present study examined the hypothesis that as-

sembly of �-subunit with a related protein, cHK in distal Moreover, in immunoprecipitation experiments performed

Codina et al: c-subunit of the Na�,K�-ATPase 973

a small protein associated with the Na,K-ATPase. J Cell Biolby our laboratory, �-subunit association with the cHK/�121:579–586, 1993

complex was not demonstrated. Therefore, when taken 4. Arystarkhova E, Wetzel RK, Asinovski NK, Sweadner KJ: Thegamma subunit modulates Na(�) and K(�) affinity of the renaltogether, these results suggest that neither the �-subunitNa,K-ATPase. J Biol Chem 274:33183–33185, 1999per se nor variations in �-subunit assembly could serve

5. Minor NT, Sha Q, Nichols CG, Mercer RW: The gamma subunitas an explanation for reported differences in pharmaco- of the Na,K-ATPase induces cation channel activity. Proc Natl

Acad Sci USA 95:6521–6525, 1998logical sensitivities and properties ascribed to cHK by6. Therien AG, Karlish SJ, Blostein R: Expression and functionaldifferent laboratories [16, 19, 36].

role of the gamma subunit of the Na,K-ATPase in mammalianThe present study shows that �-Na�,K�-ATPase is cells. J Biol Chem 274:12252–12256, 1999

7. Kraut JA, Hiura J, Shin JM, et al: The Na(�)-K(�)-ATPase beta 1associated with �1/�1 Na�,K�-ATPase in medullary tis-subunit is associated with the HK alpha 2 protein in the rat kidney.sue. Pu et al, using immunolocalization techniques, didKidney Int 53:958–962, 1998

not identify the �-subunit in the medullary collecting 8. Codina J, Delmas Mata JT, DuBose TD Jr: The alpha-subunitof the colonic H�,K�-ATPase assembles with beta1-Na�,K�-duct per se [37]. Rather, the � subunit was immunoloca-ATPase in kidney and distal colon. J Biol Chem 273:7894–7899,lized to the basolateral membranes of cells of the thick1998

ascending limb of Henle’s loop. However, since cHK is 9. Kraut JA, Hiura J, Besancon M, et al: Effect of hypokalemia onthe abundance of HK alpha 1 and HK alpha 2 protein in the ratexpressed at very low levels, the possibility that cHKkidney. Am J Physiol 272:F744–F750, 1997was associated with the �-subunit could not be ruled out

10. Codina J, Delmas-Mata JT, DuBose TD Jr: Expression of HK-in Pu’s studies. cHK expression in renal medulla and alpha2 protein is increased selectively in renal medulla by chronic

hypokalemia. Am J Physiol 275:F433–F440, 1998its up-regulation by chronic hypokalemia is specifically11. Sangan P, Rajendran VM, Mann AS, et al: Regulation of colonicconfined to the medullary collecting duct [10]. Studies

H-K-ATPase in large intestine and kidney by dietary Na depletionperformed thus far have not revealed cHK (or gHK) and dietary K depletion. Am J Physiol 272:C685–C696, 1997

12. Codina J, Kone BC, Delmas Mata JT, DuBose TD Jr: FunctionalmRNA or protein in the thick ascending limb of Henle’sexpression of the colonic H�,K�-ATPase alpha- subunit. Pharma-loop where � was localized [37]. Therefore, prior to per-cologic properties and assembly with X�,K�-ATPase beta-sub-

forming the coimmunoprecipitation experiments in this units. J Biol Chem 271:29759–29763, 199613. Cougnon M, Planelles G, Crowson MS, et al: The rat distalstudy, Pu’s data could not be taken as evidence that �

colon P-ATPase alpha subunit encodes a ouabain-sensitive H�,K�-might not associate with cHK.ATPase. J Biol Chem 271:7277–7280, 1996

In summary, our study suggests that the colonic H�, 14. Asano S, Hoshina S, Nakaie Y, et al: Functional expression ofputative H�-K�-ATPase from guinea pig distal colon. Am J PhysiolK�-ATPase is a heterodimer of composition cHK/�1 in275:C669–C674, 1998both the renal medulla and distal colon. In contrast, �

15. Nakamura S, Wang Z, Galla JH, Soleimani M: K� depletionappears to be an integral but selective component of the increases HCO3

� reabsorption in OMCD by activation of colonicH(�)-K(�)-ATPase. Am J Physiol 274:F687–F692, 1998Na�,K�-ATPase, and not the colonic H�,K�-ATPase.

16. Wall SM, Mehta P, DuBose TD Jr: Dietary K� restriction upreg-ulates total and sch-28080-sensitive bicarbonate absorption in rat

ACKNOWLEDGMENT tIMCD. Am J Physiol 275:F543–F549, 199817. Wallmark B, Briving C, Fryklund J, et al: Inhibition of gastricThis work was supported in part by a National Institutes of Health

H�,K�-ATPase and acid secretion by SCH 28080, a substituted(Institute of Diabetes, Digestive and Kidney Diseases) research grant,pyridyl(1,2a)imidazole. J Biol Chem 262:2077–2084, 1987DK-30603, awarded to Dr. DuBose.

18. Mendlein J, Sachs G: Interaction of a K(�)-competitive inhibitor,a substituted imidazo[1,2a] pyridine, with the phospho- and de-phosphoenzyme forms of H�,K(�)-ATPase. J Biol Chem 265:5030–

APPENDIX 5036, 199019. Nakamura S, Amlal H, Galla JH, Soleimani M: Colonic H�-K�-Abbreviations user in this article are: cHK, �-subunit of the colonic

ATPase is induced and mediates increased HCO3� reabsorption inH�,K�-ATPase; gHK, �-subunit of the gastric H�,K�-ATPase; �1, �1- inner medullary collecting duct in potassium depletion. Kidney Intsubunit of the Na�,K�-ATPase; �3, �3-subunit of the Na�,K�-ATPase;

54:1233–1239, 1998�1, �1-subunit of the Na�,K�-ATPase; �3, �3-subunit of the Na�,K�- 20. Buffin-Meyer B, Younes-Ibrahim M, Barlet-Bas C, et al: KATPase; �NK[gHK], �-subunit of the Na�,K�-ATPase subcloned in depletion modifies the properties of Sch-28080-sensitive K-ATPaseframe with 987 nucleotides of the amino terminal of gHK; SDS, sodium in rat collecting duct. Am J Physiol 272:F124–F131, 1997dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis; PNGase, 21. Cheval L, Elalouf JM, Doucet A: Re-evaluation of the expres-glycosidase F. sion of the gastric H,K-ATPase alpha subunit along the rat neph-

ron. Pflugers Arch 433:539–541, 199722. Laroche-Joubert N, Marsy S, Doucet A: Cellular origin andREFERENCES

hormonal regulation of K(�)-ATPase activities sensitive to sch-1. Beggah A, Mathews P, Beguin P, Geering K: Degradation and 28080 in rat collecting duct. Am J Physiol (Renal Physiol) 279:

endoplasmic reticulum retention of unassembled alpha- and beta- F1053–F1059, 2000subunits of Na,K-ATPase correlate with interaction of BiP. J Biol 23. DuBose TD Jr, Codina J, Burges A, Pressley TA: RegulationChem 271:20895–20902, 1996 of H(�)-K(�)-ATPase expression in kidney. Am J Physiol 269:

2. Beggah AT, Beguin P, Bamberg K, et al: Beta-subunit assembly F500–F507, 1995is essential for the correct packing and the stable membrane inser- 24. Pohl SL, Krans HM, Kozyreff V, et al: The glucagon-sensitivetion of the H,K-ATPase alpha-subunit. J Biol Chem 274:8217–8223, adenyl cyclase system in plasma membranes of rat liver. VI. Evi-1999 dence for a role of membrane lipids. J Biol Chem 246:4447–4454,

3. Mercer RW, Biemesderfer D, Bliss DP Jr, et al: Molecular clon- 197125. Lowry OH, Rosebrough NJ, Farr AL, Randall RJ: Proteining and immunological characterization of the gamma polypeptide,

Codina et al: c-subunit of the Na�,K�-ATPase974

measurement with the Folin phenol reagent. J Biol Chem 193:265– 32. Codina J, Pressley TA, DuBose TD Jr: Effect of chronic hypoka-lemia on H(�)-K(�)-ATPase expression in rat colon. Am J Physiol275, 1951272:F22–F30, 199726. Aronson PS: Energy-dependent of phlorizin binding to isolated

33. Kuster B, Shainskaya A, Pu HX, et al: A new variant of therenal microvillus membranes. J Membr Biol 42:81–98, 1978gamma subunit of renal Na,K-ATPase. Identification by mass spec-27. Attali B, Latter H, Rachamim N, Garty H: A corticosteroid-trometry, antibody binding, and expression in cultured cells. J Biolinduced gene expressing an “IsK-like” K� channel activity in Xeno-Chem 275:18441–18446, 2000pus oocytes. Proc Natl Acad Sci USA 92:6092–6096, 1995

34. Codina J, Pressley TA, DuBose TD Jr: The colonic H�,K�-28. Morrison BW, Moorman JR, Kowdley GC, et al: Mat-8, a novelATPase Functions as a Na�-dependent K�(NH4

�)-ATPase in api-phospholemman-like protein expressed in human breast tumors,cal membranes from rat distal colon. J Biol Chem 274:19693–19698,induces a chloride conductance in Xenopus oocytes. J Biol Chem1999270:2176–2182, 1995

35. Sangan P, Thevananther S, Sangan S, et al: Colonic H-K-ATPase29. Moorman JR, Palmer CJ, John JED, et al: Phospholemman ex- alpha- and beta-subunits express ouabain-insensitive H-K-ATPase.pression induces a hyperpolarization-activated chloride current in Am J Physiol (Cell Physiol) 278:C182–C189, 2000Xenopus oocytes. J Biol Chem 267:14551–14554, 1992 36. Amlal H, Wang Z, Soleimani M: Functional upregulation of H�-

30. Sanford J, Codina J, Birnbaumer L: Gamma-subunits of G pro- ATPase by lethal acid stress in cultured inner medullary collectingteins, but not their alpha- or beta-subunits, are polyisoprenylated. duct cells. Am J Physiol 273:C1194–C1205, 1997Studies on post-translational modifications using in vitro translation 37. Pu HX, Cluzeaud F, Goldshlegger R, et al: Functional role andwith rabbit reticulocyte lysates. J Biol Chem 266:9570–9579, 1991 immunocytochemical localization of the {gamma}a and {gamma}b

31. Laemmli UK: Cleavage of structural proteins during the assembly forms of the Na,K-ATPase {gamma} subunit. J Biol Chem 276:20370–20378, 2001of the head of bacteriophage T4. Nature 227:680–685, 1970