Kidney International, Vol. 60 (2001), pp. 427–430

Ion pump sorting in polarized renal epithelial cells

MICHAEL J. CAPLAN

Department of Cellular and Molecular Physiology, Yale University School of Medicine, New Haven, Connecticut, USA

Ion pump sorting in polarized renal epithelial cells. The plasma sorting process appears to occur as the newly synthesizedmembranes of renal epithelial cells are divided into distinct polypeptides pass through the trans-most cisterna of theapical and basolateral domains, which contain different inven- Golgi complex [3]. During their brief sojourn in thistories of ion transport proteins. Without this polarity vectorial

compartment, referred to as the trans Golgi networkion and fluid transport would not be possible. Little is known(TGN), apical and basolateral proteins appear to be sep-of the signals and mechanisms that renal epithelial cells use

to establish and maintain polarized distributions of their ion arated from one another. They become concentrated intransport proteins. Analysis of ion pump sorting reveals that discrete regions of the TGN membrane from which car-multiple complex signals participate in determining and regu- rier vesicles will bud to carry these proteins directly tolating these proteins’ subcellular localizations.

their sites of functional residence at the apical or basolat-eral cell surface. In order for this sorting process to occurin the TGN, proteins destined for polarized plasmalem-From a physiologic perspective, the salient feature ofmal distributions must contain within their structuresthe epithelial cells that line the nephron is their capacitysorting information that serves to specify their appro-to catalyze vectorial solute and fluid transport againstpriate destination. Furthermore, the cells must possesssteep concentration gradients. This capacity is, in turn,machinery that is capable of interpreting these signalscritically dependent on one of the renal epithelium’sand acting upon the pathways that they specify [4].most dramatic cell biologic characteristics. The plasma

Little is known of the signals and cellular machinerymembranes of the renal tubular epithelial cells involvedthat collaborate to achieve the polarized distributions ofin active transport are structurally and functionally po-ion transport proteins. Because the proteins that mediatelarized. They are divided into two biochemically distinction transport tend to be large and span the membranedomains that are endowed with distinct subsets of ionmultiple times, it has been difficult to isolate and identifytransport proteins [1]. These apical and basolateral do-narrowly defined sequence determinants that might func-mains are separated from one another by tight junctions,tion to transmit sorting information. Our efforts to ad-which serve both to regulate paracellular permeabilitydress this problem have taken advantage of families ofand to prevent the intermixing of their unique proteinion transport proteins whose members all share a highand lipid constituents. It is precisely this anisotropic dis-degree of sequence homology but are sorted to distincttribution of transport proteins among these defined plas-locations in polarized epithelial cells.malemmal surfaces that permits unidirectional and con-

centrative transport.While the importance of renal epithelial polarity is ION PUMPS IN POLARIZED EPITHELIA

clear, the mechanisms through which it is establishedThe Na,K-ATPase, or sodium pump, is perhaps theand maintained are far less obvious. Newly synthesized

best known member of the P-type family of ion trans-membrane proteins are cotranslationally inserted into theporting ATPases [5]. In the kidney, the Na,K-ATPasemembrane of the rough endoplasmic reticulum (RER).resides at the basolateral surfaces of tubule epithelialFrom there, they follow the secretory pathway throughcells, where it serves to generate the gradients that ener-the Golgi complex and onward to the plasma membranegize almost every renal transport process [6]. The sodium[2]. In polarized epithelial cells, proteins destined forpump’s closest cousin within the P-type family is thethe apical membrane must be segregated from thosegastric H,K-ATPase, which is responsible for the regu-earmarked for residence at the basolateral surface. Thislated acidification of the lumen of the stomach [7]. Thegastric H,K-ATPase is also expressed in the kidney,where it may function in potassium reabsorption and inKey words: Na,K-ATPase, H,K-ATPase, cell polarity, sorting, protein

transport, solute and fluid transport, tight junction. the maintenance of acid-base balance [8]. Both the Na,K-and H,K-ATPases are composed of an � subunit that 2001 by the International Society of Nephrology

427

Caplan: Ion pump sorting in renal epithelial cells428

Fig. 1. Schematic representation of ion pumpdistribution in a gastric parietal cell. The Na,K-ATPase (thin black bars) is limited to the baso-lateral cell surface. In contrast, the H,K-ATPase(thin gray bar) is localized at the apical plasmamembrane, as well as in the membranes ofthe tubulovesicular elements (TVE, circles).Secretagogue stimulation leads to fusion ofthe TVEs with the apical plasma membrane.

spans the membrane ten times [9] and a glycosylated � ion pumps composed of complementary portions of thesubunit that crosses the membrane once in a type II � subunit polypeptides [13–15]. By expressing these chi-orientation [10]. The � subunits of these pumps share meric pump constructs in polarized epithelial cells inapproximately 65% protein sequence identity, whereas culture and determining their subcellular distributions,the � subunits are approximately 40% identical. Consis- we were able to characterize portions of these proteinstent with their high degree of structural homology, these that determine their sorting properties. We found thatpumps share a common reaction mechanism involving both pump subunit proteins embody signals that deter-sequential transitions through a series of distinct confor- mine and regulate the distribution of the holoenzymemational states [11]. Despite their close structural and complex.functional relationship, these pumps manifest markedly Analysis of the distributions of numerous chimeras,distinct cell biologic properties. expressed by transfection in the proximal tubule-like

In contrast to the sodium pump’s basolateral distribu- LLC-PK1 line of pig kidney epithelial cells, reveals thattion, the H,K-ATPase functions at the apical surfaces of the fourth transmembrane spanning sequence (TM4) ofgastric and renal epithelial cells (Fig. 1). Furthermore, the H,K-ATPase manifests dominant sorting informa-while the Na,K-ATPase is a fairly stable resident of the tion [15]. A chimera composed of Na,K-ATPase �-sub-basolateral surface, the gastric H,K-ATPase undergoes unit sequence in which only the TM4 is replaced by itsa process of regulated insertion into and removal from H,K-ATPase counterpart behaves as an apical mem-the plasmalemma. In the resting gastric parietal cell, the brane protein. This result is somewhat surprising, sinceH,K-ATPase occupies the membranes of an intracellular most sorting signals identified to date are present in thestorage compartment known as the tubulovesicular ele- cytoplasmic domains of membrane proteins, where theyment (TVE). Secretagogue stimulation results in fusion are available to interact with the wide variety of cytosolicof the TVEs with the apical plasma membrane, resulting proteins that participate in membrane trafficking [3]. Wein the insertion of the H,K pump. Cessation of gastric wondered, therefore, whether the TM4 signal mightsecretion is initiated through the endocytic retrieval of function through interactions with lipid molecules in thethe H,K pump and its return to the TVE [7]. A similar plane of the membrane.regulated exocytosis and retrieval mechanism may also

Recently, a large body of evidence has accumulatedgovern H,K-ATPase activity in the kidney [12].

demonstrating that some membrane proteins may besorted by virtue of their capacity to partition into glyco-

ROLE OF THE � SUBUNIT IN sphingolipid (GSL)-rich rafts that self-assemble in thePOLARIZED SORTING plane of the Golgi membrane [16]. These rafts appear

to nucleate the budding of transport vesicles, which carryTo identify the molecular determinants responsiblefor these distinct behaviors, we have generated chimeric their cargo of membrane proteins to the apical plasma

Caplan: Ion pump sorting in renal epithelial cells 429

membrane. It is thought that this mechanism accounts motif. Rather, it seems to play a role in creating a confor-for the apical sorting of glycosylphosphotidyl inositol mational determinant that is recognized by the cell’s sort-(GPI)-linked proteins observed in many epithelial cell ing machinery. Chimeras composed of the Na,K-ATPasetypes [17]. It has also been shown that several apical � subunit in which one or the other sequence domainstransmembrane proteins associate with GSL rafts and, that flank TM4 are replaced by their H,K-ATPase counter-furthermore, that this association is required for them to parts behave as basolateral polypeptides. Simultaneousachieve their apical distribution [18]. GSL-rich rafts and substitution of both flanking domains with H,K-ATPasetheir associated proteins resist solubilization in 1% Tri- sequence, however, results in the creation of an apicallyton X-100 at 4�C [19]. To test whether the H,K-ATPase directed ion pump [15]. The apical localization of thisTM4 signal operates through a similar mechanism, we chimera occurs despite the fact that its TM4 is derivedanalyzed the detergent solubility properties of our apical from the Na,K-ATPase. Thus, the apical message con-chimeras. We found that the chimeras were fully deter- tained in the H,K-ATPase TM4 can be recapitulated bygent soluble under these conditions, as was the endoge- the discontinuous sequences that border it, but only ifnous Na,K-ATPase [15]. Thus, lipid partitioning is un- both of those sequences are present. We interpret theselikely to account for the sorting information intrinsic to surprising observations to mean that apical sorting ofthe H,K-ATPase TM4. the H,K-ATPase and the chimeric pumps is not the prod-

The mechanism through which the H,K-ATPase TM4 uct of a signal constituted of a single linear stretch ofbrings about apical localization is even more difficult to amino acids. Rather, the apical sorting determinant ap-fathom when considered in light of the sorting behavior of pears to arise from a conformational conversation thatthe nongastric H,K-ATPase. The nongastric H,K-ATPase occurs between TM4 and its neighboring sequences inis a P-type ion pump expressed in kidney and in colon [8]. the protein’s structure.Like its close relatives Na,K and gastric H,K-ATPase, thenongastric H,K-ATPase is a heterodimer composed of

ROLE OF THE � SUBUNIT IN IONan � and a � subunit. While the molecular identity ofPUMP REGULATIONthe � subunit has yet to be fully established, the sequence

While the signal responsible for apical versus basolat-of the � subunit is �60% identical to those of both theeral pump targeting does not appear to be reducible toNa,K- and gastric H,K-ATPase � subunits [20]. The non-a short motif, the opposite is true of the signal that specifiesgastric H,K-ATPase plays a role in potassium reabsorptionthe targeting of the gastric H,K-ATPase to a regulatedfrom the lumen of the colon and the renal collecting tu-storage compartment. Whereas the Na,K-ATPase ap-bule, especially during systemic potassium depletion. Thispears to reside constitutively at the basolateral surfacespump may also play a role in acid-base regulation, althoughof renal epithelial cells, the presence of the gastric H,K-the extent to which it is involved in active proton trans-ATPase at the parietal cell apical surface is subject toport remains open to discussion. Recent in vitro expres-tight physiologic control. Hormonal stimulation of gas-sion studies demonstrate that the nongastric H,K-ATPasetric acid secretion leads to fusion of the TVE membranesmay preferentially employ sodium ions rather than pro-with the apical plasmalemma, resulting in the cell surfacetons as the counter ions for potassium uptake [21–23].delivery of a large cohort of H,K-ATPase [7]. Acid secre-Since it appears to function in lumenal potassium in-tion is terminated through the endocytic internalizationflux, it is not surprising that physiologic data indicate thatof the H,K-ATPase, which returns the pump to the TVEthe nongastric H,K-ATPase resides in the apical plasmacompartment.membrane [24]. Recent expression studies further sup-

A great deal has been learned about the mechanismsport this localization, demonstrating that the human or-that mediate membrane protein endocytosis. Many pro-thologue of this pump accumulates at the apical surfaceteins subject to internalization contain within the sequencesin transfected Madin-Darby canine kidney (MDCK) re-of their cytoplasmic tails a short tyrosine-containing mo-nal epithelial cells [25]. Based on the chimera studiestif that is necessary and sufficient to ensure their suscepti-implicating the gastric H,K-ATPase TM4 in this pump’sbility to endocytosis [26]. The cytoplasmic tail of the gastricapical sorting, one might expect that the TM4 of theH,K-ATPase � subunit includes a tyrosine-containing se-nongastric H,K-ATPase should resemble that of its gas-quence that is quite similar to well-characterized endocy-tric cousin. In marked contrast, however, the nongastrictosis signals [13]. No such motif is present in the cytosolicH,K-ATPase TM4 is almost identical to that of the baso-tail of the Na,K-ATPase � subunit. We speculated thatlateral Na,K-ATPase. Therefore, it appears that otherthe H,K-ATPase � subunit sequence might be responsi-determinants must be responsible for the apical sortingble for the regulated internalization of the H,K pumpof the nongastric H,K-ATPase.and thus for the cessation of gastric acid secretion.While the nature of these other determinants has yet

To test this hypothesis, the tyrosine residue presentto be determined, further analysis of Na,K/gastric H,K-in the cytoplasmic tail of the gastric H,K-ATPase � sub-ATPase chimeras suggests that the TM4 sorting deter-

minant does not function as an autonomous amino acid unit was changed to alanine and the resultant mutated

Caplan: Ion pump sorting in renal epithelial cells430

Golgi network to the plasma membrane in polarized cells. Seminprotein (H-Y20A) was expressed in transgenic mice [27].Cell Dev Biol 9:503–509, 1998

Animals expressing the H-Y20A protein exhibited a ge- 4. Matter K: Epithelial polarity: Sorting out the sorters. Curr Biol10:R39–R42, 2000netically dominant gastric acid hypersecretion associated

5. Pedersen PL, Carafoli E: Ion motive ATPases. I. Ubiquity, prop-with the constitutive presence of the H,K-ATPase at theerties and significance to cell function. Trends Biol Sci 12:146–150,

apical cell surface. Their phenotype also included the 19876. Caplan MJ: Ion pumps in epithelial cells: Sorting, stabilizationdevelopment of gastric erosions, ranging in severity from

and polarity. Am J Physiol 272:G1304–G1313, 1997mild gastritis to severe ulceration. Analysis of renal po-7. Hersey SJ, Sachs G: Gastric acid secretion. Physiol Rev 75:155–

tassium clearance revealed that the H-Y20A–expressing 189, 19958. Codina J, Wall SM, DuBose TD, Jr: Contrasting functional andanimals demonstrated markedly reduced potassium ex-

regulatory profiles of the renal H�,K�-ATPases. Semin Nephrolcretion, consistent with increased renal H,K-ATPase ac-19:399–404, 1999

tivity [12]. Therefore, these data demonstrate that the 9. Shull GE, Lingrel J: Molecular cloning of the rat stomach H,K-ATPase. J Biol Chem 261:16788–16791, 1986gastric H,K-ATPase � subunit’s tyrosine-containing mo-

10. Shull GE: cDNA Cloning of the �-subunit of the rat gastric H,K-tif constitutes a biologic cell signal that is responsibleATPase. J Biol Chem 265:12123–12126, 1990

for the endocytic regulation of two extremely important 11. Jorgensen PL: Mechanism of the Na,K pump: Protein structureand conformations of the pure Na,K-ATPase. Biochim Biophysion transport processes.Acta 694:27–68, 1982

12. Wang T, Courtois-Coutry N, Giebisch G, Caplan MJ: A tyro-sine-based signal regulates H,K-ATPase-mediated potassium reab-CONCLUSIONsorption in the kidney. Am J Physiol 275:F818–F826, 1998

13. Gottardi CJ, Caplan MJ: An ion transporting ATPase encodesClearly, a number of molecular signals appear to bemultiple apical localization signals. J Cell Biol 121:283–293, 1993involved in establishing the polarized distributions of ion

14. Muth TR, Gottardi CJ, Roush DL, Caplan MJ: A dominantpump proteins and in regulating their functional tenure basolateral sorting signal is encoded in the �-subunit of the Na,K-

ATPase. Am J Physiol 274:C688–C696, 1998at the plasma membrane. It remains to be determined15. Dunbar LA, Aronson PL, Caplan MJ: A transmembrane segmenthow these signals are interpreted and with what molecu- determines the steady-state localization of an ion transporting

lar machinery they interact. Future studies need to focus adenosine triphosphatase. J Cell Biol 148:769–778, 200016. Brown DA, London E: Functions of lipid rafts in biological mem-on identifying the protein partners that associate with

branes. Annu Rev Cell Dev Biol 14:111–113, 1998these signals to transduce their messages. Defining the 17. Brown D, Rose JK: Sorting of GPI-anchored proteins to glyco-proteins that can bind to the relevant domains of the lipid-enriched membrane subdomains during transport to the api-

cal cell surface. Cell 68:533–544, 1992ion pump subunit polypeptides should provide valuable18. Scheiffele P, Roth MG, Simons K: Interaction of influenza virus

clues as to the nature of the molecular mechanisms that haemagglutinin with sphingolipid-cholesterol membrane domainsvia its transmembrane domain. EMBO J 16:5501–5508, 1997epithelial cells use to establish and modify the polarized

19. Arreaza G, Melkonian KA, Bernt-LaFevre M, Brown DA:distributions of these structurally complex and physio-Triton X-100 membrane complexes from cultures kidney epithelial

logically critical transport proteins. cells contain the Src family protein tyrosine kinase p62yes. J BiolChem 269:19123–19127, 1994

20. Crowson MS, Shull GE: Isolation and characterization of aACKNOWLEDGMENTScDNA encoding the putative distal colon H�,K(�)-ATPase: Simi-larity of deduced amino acid sequence to gastric H�,K(�)-ATPaseThe work from the laboratory discussed here is supported by grantsand Na�,K(�)-ATPase and mRNA expression in distal colon,from the National Institutes of Health (GM-42136 and DK-17433).kidney, and uterus. J Biol Chem 267:13740–13748, 1992The author is grateful to the many members of my lab group, past

21. Grishin AV, Caplan MJ: ATPAL1, a member of the non-gastricand present, who have carried out the studies described in this review.H,K-ATPase family, functions as a sodium pump. J Biol Chem273:27772–27778, 1998Reprint requests to Michael J. Caplan, M.D., Department of Cellular

22. Cougnon M, Bouyer P, Planelles G, Jaisser F: Does the colonicand Molecular Physiology, Yale University School of Medicine, 333H,K-ATPase also act as an Na,K-ATPase? Proc Natl Acad SciCedar Street, New Haven, Connecticut 06510, USAUSA 95:6516–6520, 1998E-mail: michael.caplan@yale.edu

23. Codina J, Pressley TA, DuBose TD Jr: The colonic H�,K�-ATPase functions as a Na�-dependent K�(NH4�)-ATPase in api-cal membranes from rat distal colon. J Biol Chem 274:19693–19698,

APPENDIX 199924. Doucet A: H,K-ATPase in the kidney: Localization and functionAbbreviations used in this article are: GPI, glycosylphotidyl inositol;

in the nephron. Exp Nephrol 5:271–276, 1997GSL glycosphingolipid; H-Y20A, mutated protein from H,K-ATPase25. Reinhardt J, Grishin AV, Oberleithner H, Caplan MJ: Differ-�-subunit changed to alanine; RER, rough endoplasmic reticulum;

ential localization of the human non-gastric H�,K�-ATPaseTGN, trans Golgi network; TME4, fourth transmembrane spanningATP1AL1 in polarized renal epithelial cells. Am J Physiol 279:sequence; TVE, tubulovesicular element. F417–F425, 2000

26. Thomas DC, Roth MG: The basolateral targeting signal in thecytoplasmic domain of glycoprotein G from vesicular stomatitisREFERENCESvirus resembles a variety of intracellular targeting motifs related

1. Caplan MJ: Membrane polarity in epithelial cells: Protein sorting by primary sequence but having diverse targeting activities. J Bioland the establishment of polarized domains. Am J Physiol 272: Chem 269:15732–15739, 1994F425–F429, 1997 27. Courtois-Coutry N, Roush DL, Rajendran V, et al: A tyrosine-

2. Caplan MJ: Biogenesis and sorting of plasma membrane proteins. based signal targets H,K-ATPase to a regulated compartment andCurr Top Membr 39:37–86, 1991 is required for the cessation of gastric acid secretion. Cell 90:501–

510, 19973. Ikonen E, Simons K: Protein and lipid sorting from the trans-