Ion motive ATPases involved in transcellular ion transport by renal epithelial cells include two molecular families: P-ATPases (Na,K-ATPase and H,K-ATPase) and V-ATPases. Na,K-ATPase is found in the plasma membrane of every vertebrate cell and exchanges three intracellular Na + for two intracellular K + for each hydrolyzed ATP molecule. Na,K-ATPase is highly expressed in the kidney tubule where it is located in the basolateral membrane and energizes Na + reabsorption. Na,K-ATPAse is highly regulated by hormones and local factors via modulation of its rate of synthesis, degradation and post translational events including regulatory phosphorylation and intracellular trafficking. Both gastric and non-gastric H,K-ATPases are expressed in the distal part of the kidney tubule where they reabsorb K + and secrete H + . Vacuaolar H-ATPases is found in intracellular compartments of every renal epithelial cell. In intercalated cells of connecting tubules and collecting ducts, it is localized in plasma membrane where it plays a key role in acid–base transport.
ATPase; active ion transport; renal epithelial cell; sodium; potassium; proton
Active transport of solutes across membranes against their concentration or electrochemical gradients requires energy. For ion-motive ATPases (F-type, V-type, and P-type), this process is an exchange between energy contained in the electrochemical gradient and chemical energy provided by ATP hydrolysis. F-type ATPases or ATP synthases are responsible for the generation of ATP using energy of the proton gradient created by the respiratory chain in mitochondria or photosynthetic complexes in chloroplasts. V-type ATPases acidify vesicles by transporting protons from the cytoplasm to the lumen of intracellular organelles (endosomes, lysosomes, vacuoles, …). V-ATPases are also present in the plasma membrane of some epithelial cells. Despite sharing a common general architecture and a large number of subunits, F- and V-ATPases usually work in opposite directions. P-type ATPases (also called E1-, E2-ATPases) form a third group of ion-motive ATPases that perform unidirectional or exchange transport of monovalent (H + , Na + , K + ) or divalent (Ca 2+ , Cu 2+ , Mg 2+ , …) cations.
While some members of the P-ATPase family are probably active as a single polypeptide, the functional unit of others consists of several subunits. The major subunit (α- or catalytic subunit for multimeric P-ATPases) consists of a series of hairpins formed by pairs of transmembrane segments linked by short extracellular loops. Two large intracellular loops make the connection between the first, second, and third hairpins. The largest cytoplasmic loop contains the ATP-binding domain and the phosphorylation site. Transient phosphorylation of this aspartate residue occurring during the transport cycle is a hallmark of P-ATPases ( Figure 3.1 ).
Structure and Function of Ca 2+ -ATPases (SERCA and PMCA)
SERCA is found in intracellular organelles related to the endoplasmic reticulum, such as the sarcoplasmic reticulum of cardiac and skeletal muscle cells. Three differently expressed genes have been identified: SERCA1 in fast-twitch skeletal muscle; SERCA2 in slow-twitch skeletal muscle, heart, and smooth muscle; and SERCA3 in blood, endothelial, and epithelial tissue. SERCA mediates uptake of Ca 2+ from the cytoplasm into the sarcoplasmic reticulum following calcium release from intracellular stores. It therefore acts as a terminator signal in excitation–contraction coupling processes in muscle, and plays a key role in excitation–secretion coupling in neurons and other secretory cells. Extensive structure–function studies have been performed with SERCA and a high resolution (2.6 Å) structure of this protein was obtained.
PMCA is expressed at the plasma membrane, where it extrudes Ca 2+ out of the cell. Four PMCA genes are known, with multiple splicing variants for each gene, resulting in the existence of about 20 isoforms. PMCA1 and PMCA4 are ubiquitous, while PMCA2 and PMCA3 are restricted to neurons, brain, muscle, and kidney. PMCA plays an important role in tubular reabsorption of calcium. The long C-terminal intracellular domain maintains PMCA in an inactive state by interacting with the catalytic site. A rise of cytosolic Ca 2+ concentration increases Ca 2+ –calmodulin-binding, allowing calmodulin to interact with the PMCA C-terminal domain. This releases PMCA autoinhibition, activating the pump. Conversely, calcium extrusion decreases cytosolic Ca 2+ concentration, and consequently its association with calmodulin. Calmodulin release from PMCA increases PMCA auto-inhibition. Regulatory inhibition of SERCA is mediated by the small associated protein phospholamban, which plays a role equivalent to that of the PMCA C-terminal domain. Phosphorylation of phospholamban by protein kinase C releases SERCA inhibition.
Structure of Na,K-ATPase and H,K-ATPase
Na,K- and H,K-ATPases are heteromeric proteins consisting of an α-subunit and a smaller glycosylated β-subunit. Na,K-ATPase hydrolytic activity, cation transport activity, and ouabain-binding properties were demonstrated by co-expression of α- and β-subunits in several expression systems (mammalian cells, Xenopus oocytes, baculovirus-infected insect cells, and yeast). Na- (or H-) and K-activated ATPase and cation transport activities (i.e., uphill cation transport driven by ATP hydrolysis) characteristic of Na,K- or H,K-ATPases have been demonstrated only in the presence of both α- and β-subunits. Expression of the α-subunit alone in insect cells resulted in Mg 2 -dependent ATPase activity that was not specifically activated by Na + and K + . The exact stoichiometry of the minimal functional unit is still a matter of debate. However, Na,K-ATPase activity is associated with solublized α–β protomers, and cross-linking experiments did not show evidence for a close interaction between α-subunits. A third subunit, the γ-subunit, can be associated with the α–β complex (see below).
Na,K- and H,K-ATPase α-subunit peptides range in length from about 1000 to 1040 amino acids. Their primary structure is characterized by a first group of four transmembrane segments, followed by a large cytoplasmic loop, and a second group of six transmembrane segments ( Figure 3.1 ). Crystal structures of pig and shark Na,K-ATPase at 3.5 and 2.4 Å resolutions, respectively, confirmed that the α-subunit has three cytoplasmic domains and 10 transmembrane helices, designated M1 to M10. Two-thirds of its mass is contained in the large cytoplasmic domain, while one-third spans the lipid bilayer. Of the total mass, only a small part is extracellular. Sequence homology between Na,K- and H,K-ATPases is high enough to safely predict H,K-ATPase structure, at least for the general architecture of this molecule, and for large domains where homology is highest.
Six different α-subunit genes have been identified in mammals: α1-4 isoforms of the Na,K-ATPase α-subunit; gastric H,K-ATPase α-subunit (αHKg); and colonic H,K-ATPase α-subunit (αHKc). Related isoforms have been identified in birds and amphibians. Na,K-ATPase sequences from Caenorhabditis elegans or Drosophila melanogaster do not show close similarity with any mammalian isoform. This suggests that the divergence between Na,K- and H,K-ATPase α-subunits precedes the divergence between mammals, amphibians and birds, and has occurred early in vertebrate evolution.
All Na,K-ATPase isoforms primarily maintain Na + and K + gradients across the cell membrane. The large inward electrochemical gradient for Na + is in turn used by numerous secondary active transport systems for various “housekeeping” functions: maintenance of intracellular pH via Na–H exchangers; extrusion of calcium via Na–Ca exchanger; control of cell volume via Na–K–2Cl symport and other coupled transport systems; and import of amino acids, nucleotides, and other nutrients or osmolytes through various Na + -coupled co-transport systems. The outward electrochemical gradient for K + is responsible for the intracellular negative membrane potential, because K + flows out of the cell through K + selective channels that are active in most cells. In addition to these general functions, Na + and K + gradients across cell membranes are essential for specialized functions, such as the generation and propagation of action potentials in excitable cells, neurotransmitter uptake, and transcellular transport of solutes and water by epithelial cells.
The α1 isoform is the most ubiquitous and abundant α isoform, and is responsible for the maintenance of whole-cell Na + and K + gradients necessary for housekeeping functions. Because of its abundance (it is the only α isoform present in many epithelial cells, including renal cells) it provides the driving force for solute and water transepithelial transport. The α2 isoform is found in skeletal and heart muscle, and in the nervous system (neurons and glial cells). The α3 isoform is essentially neuronal, but is also found in blood cells and macrophages. The α4 isoform is mostly expressed in testes, and plays a critical role in sperm motility.
αHKg is abundantly expressed in parietal cells of the gastric gland, where it plays a central role in proton secretion. Under resting conditions, it is mainly located in an intracellular tubulo–vesicular network that fuses with the apical membrane of parietal cells in response to stimuli, allowing the H,K-pump to secrete protons into the gastric gland lumen in exchange for potassium. αHKc is mainly expressed in the (rat) distal colon, but also in the kidney, uterus, and, to a lesser extent, in the heart.
As stated above, the β-subunit is an essential constituent of functional Na,K-ATPase and H,K-ATPase. Five different genes encoding similar proteins are known in mammalian genomes: β1; β2; β3; βHK; and βm (“m” emphasizes its predominant expression in skeletal muscle). β1, β2, and β3 are clearly Na,K-ATPase β-subunit isoforms, while βHK is co-expressed with gastric H,K-ATPase. Although usually described as ubiquitous, β1 appears to be absent, or at best is only a minor component, in several tissues such as liver and red blood cells. The β2 isoform was initially identified in glial cells, but is also present in other cell types, including neurons, blood cells, and epithelial cells. The β3 isoform, initially identified in nervous systems, is also widely distributed, being most abundant in testes, liver, and lungs and less so in skeletal muscle and kidney. Despite sharing sufficient sequence similarity to be classified in the same family, βm does not associate with any known mammalian α-subunit.
Structure and α–β Interaction
β-subunit peptides range in length from 288 to 315 amino acids, and show a lower degree of homology (about 30%–40% identity between isoforms) than α isoforms. Crystal structures of Na,K-ATPase obtained at resolutions of 3.5 Å and 2.4 Å have lent detailed insight into β-subunit structure. These studies, together with experimental evidence, show that the β-subunit is a type II membrane protein consisting of a single transmembrane segment, a ~35-amino acid N-terminal domain, and a large extracellular domain containing two to seven glycosylation sites, depending on the isoform, and six cysteine residues that form three disulfide bridges ( Figure 3.1 ).
Experimental modeling and analysis of Na,K-ATPase crystal structures has revealed complex interactions between α- and β-subunits. The transmembrane helix of the β-subunit forms several hydrogen bonds and numerous contacts with M7 and M10 transmembrane helices of the α-subunit, primarily via clusters of aromatic residues. At the extracellular side of the β-subunit, a stretch of amino acids adjacent to its transmembrane domain interacts with the α M7 / M8 extracellular loop that contains a consensus sequence SYGQ. Further downstream, Lys250 of the β-subunit forms a salt bridge with Glu899, located in the α M7 / M8 extracellular loop.
Except for gastric αHK and βHK, which are most abundantly expressed in a single cell type (parietal cells of gastric glands), there is no obvious common pattern of distribution between α isoforms and β isoforms that would define preferential physiological associations. Indeed, some cells even express as many as three α isoforms and at least two β isoforms. Unless formation of specific complexes is favored or repressed by unknown mechanisms, numerous combinations are possible, as suggested by studies using artificial expression systems. β1 is abundantly expressed in tissues in which α1 predominates, such as the kidney, strongly suggesting that α1β1 represents the predominant isozyme in these tissues. The nature of the β-subunit associated with αHKc is also a matter of debate, since all β isoforms are able to associate with αHKc, depending on the expression system used.
The functional interaction between α- and β-subunits has been studied and reviewed in detail. By acting as a molecular chaperone, the β-subunit plays a critical role in the maturation of the α-subunit. Indeed, the α-subunit reaches a mature and functional conformation, ready to be translocated from the endoplasmic reticulum to the plasma membrane, only when associated with a β-subunit.
The β-subunit contributes to intrinsic transport properties of the whole enzyme in several expression systems by influencing its apparent K + and Na + affinities. Biochemical analysis and crystallization of Na,K-ATPase has lent some mechanistic insight as to how this is achieved. By interacting with the M7 transmembrane domain of the α-subunit, Tyr40 and Tyr44 of the β-subunit transmembrane helix help confer intrinsic transport properties of the Na,K-ATPase enzyme, as suggested by a mutagenic study. Unwinding of M7 via hydrogen bonding of Tyr44 with Gly855 appears to be of central importance to K + binding. The role of the β-subunit ectodomain in modulating cation transport is further illustrated by the complex interactions between this domain and the α M7 / M8 extracellular loop.
The role of the β-subunit in cell–cell adhesion will be discussed in the section “New Physiological Functions of Na,K-ATPase.”
FXYD proteins are a third component of Na,K-ATPase. There are seven isoforms in mammals, ranging from 61 to 95 amino acids in length, except for FXYD5, which consists of 178 amino acids due to an N-terminal extension. Most FXYD proteins are small type 1 membrane proteins containing an extracellular N-terminus. This family of proteins is so named since all members contain a FXYD (Phe-X-Tyr-Asp) sequence located immediately downstream of the transmembrane segment. All members also contain two conserved glycine residues in the transmembrane domain, as well as a serine residue located further downstream. As with α and β isoforms, the tissue distribution of FXYD proteins is isoform-specific. FXYD1 (phospholemman) is predominately expressed in heart and skeletal muscle and, to a lesser extent, in brain, FXYD2 (γ-subunit) in kidney, FXYD3 (MAT-8) in stomach and colon, FXYD4 (CHIF) in kidney and distal colon, FXYD5 (RIC) in kidney, intestine and lung, and FXYD6 and FXYD7 in brain. In the kidney FXYD2 is mostly expressed in proximal tubule and thick ascending limb of Henle, while FXYD4 is exclusively found in the collecting duct where its expression increases from cortical to medullary portions.
Function and Interaction with α- and β-Subunits
FXYD proteins were first thought to be regulators of ion channels or even to act as ion channels themselves. It has now been demonstrated that FXYD proteins interact with Na,K-ATPase. Contrary to the β-subunit, most FXYD proteins do not associate with H,K-ATPase and do not appear to act as a molecular chaperone. Rather, they regulate Na,K-ATPase functional properties. As recently reviewed, analysis of FXYD-deficient mice and in vitro modulation of Na,K-ATPase activity by different FXYD proteins has shown that FXYD1–4, 6, and 7 all decrease Na,K-ATPase apparent affinity for Na + and/or K + , with the exception of FXYD4 which has been shown to decrease apparent affinity for K + but increase that for Na + . FXYD5 does not appear to influence Na,K-ATPase affinity for either Na + or K + , but enhances maximal transport activity. By modulating Na,K-ATPase activity, FXYD proteins play important physiological roles, each isoform playing a different role depending on its tissue-specific distribution. For instance, FXYD1 influences myocardial contractility by modulating Na,K-ATPase activity and calcium handling, as demonstrated in FXYD1-deficient mice that display depressed cardiac contractile function and increased cardiac mass. FXYD6, the only FXYD isoform expressed in the inner ear, may play an auditory and vestibular role by contributing to endolymph ionic compostion, whose production depends on Na,K-ATPase activity. Reflecting altered expression levels in various types of cancer, FXYD3 may be implicated in the control of cell differentiation, proliferation, and/or apoptosis through regulation of Na,K-ATPase activity. Finally, the renal tubule segment-specific distribution of FXYD2 and FXYD4 may explain, at least in part, the higher apparent Na + affinity of collecting duct Na,K-ATPase as compared to that of more proximal nephron segments.
Crystallization of the Na,K-ATPase holoenzyme has deciphered FXYD structure, and revealed how it interacts with α- and β-subunits, providing some insight into how it modulates Na,K-ATPase activity. This notably involves interaction between the transmembrane domain of FXYD proteins, particularly Gly34, with the α M9 transmembrane domain. Hydrogen bonds between Cys31 and αGlu960 may additionally play a structural role important for FXYD functional regulation. The FXYD motif helps confer β conformational structure. This is partly achieved via Phe12, which anchors this segment to the β-subunit, and Tyr14, which forms a cluster of aromatic residues with Tyr69 of the β-subunit and Trp987 of the α M9 / M10 extracellular loop.
Properties of Na,K-ATPase and H,K-ATPase
Under physiological conditions, Na,K-ATPase exchanges three intracellular Na + for two extracellular K + at the expense of one ATP during each cycle, generating an outward current of one net charge per cycle. The outward current generated by the Na,K-pump tends to hyperpolarize the cell membrane to a few millivolts under steady-state conditions. The majority of the 50 to 80 mV resting-membrane potential is due to flow of K through K channels, which relies on Na,K-pump activity, since high intracellular K + concentration is maintained by the Na,K-pump. The apparent affinities for Na + and K + are dependent on experimental conditions, but under physiological conditions intracellular Na + activates the Na,K-pump with a K½ of 10 to 20 mM and a Hill coefficient between 2 and 3. Extracellular K + has a K½ of about 1 mM, with a Hill coefficient between 1 and 2.
Despite extensive studies, the difference in transport properties between Na,K-ATPase α isoforms has not been entirely resolved. Initial studies performed in native tissues indicated that α2/α3 isoforms have a higher affinity for Na + . More recent studies comparing isoforms in the same artificial expression system revealed that α2 has a slightly lower (about 20 mM), and α3 a much lower (30–70 mM), affinity for Na. However, other factors, such as the type of associated β-subunit and FXYD protein, also modulate Na,K-ATPase transport properties (see above).
The mechanism of cation translocation by Na,K-ATPase can be summarized as follows ( Figure 3.2 ). Na,K-ATPase exists under two main conformations, E1 and E2, and transport activity is performed via a cycle in which the protein is transiently phosphorylated, and alternately adopts E1 or E2 conformations. These two conformations differ in their apparent affinities for Na + and K + . The E1 conformation has high affinity sites for Na + exposed at the intracellular side of the membrane, while the E2 conformation has high affinity sites for K + exposed at the extracellular side of the membrane. Three Na + and two K + ions are alternately bound to the enzyme and then “occluded,” that is, tightly bound inside the protein. This model is compatible with the reported structure of the major conformations of SERCA, and the E2-P conformation of Na,K-ATPase.
Non-electrogenic transport is achieved by gastric H,K-ATPase, indicating that a symmetrical number of H + and K + ions are translocated across the membrane during each cycle. Transport stoichiometry depends on pH. Under conditions of high or near neutral pH, the stoichiometry is 2H,2K + -1ATP, and shifts to 1H,1K + -1ATP under physiological conditions, i.e., conditions of very low extracellular pH. The transport properties of colonic H,K-ATPase and its close relatives (human ATP1AL1 and toad Bufo marinus bladder H,K-pump) are not yet completely defined. Artificial systems have demonstrated an inward transport of K + , and an outward transport of H + . However, data obtained in heterologous expression systems have suggested that these enzymes may exchange Na + for K + .
A group of natural compounds known as “cardiac steroids,” so named because they contain a steroid nucleus attached to a lactone ring and are used for treatment of congestive heart failure, are potent Na,K-ATPase inhibitors. In addition, one or several endogenously related compounds may also act as hormonal agents that participate in regulating Na,K-ATPase activity.
Cardiac steroid interaction with Na,K-ATPase, particularly ouabain, has been extensively studied. Differences of ouabain affinity between Na,K-ATPase isoforms, together with mutagenesis studies and recently obtained crystal structures of ouabain bound to Na,K-ATPase, all show that ouabain binds to a deep cavity formed by the transmembrane helices M1 , M2 , M4 , M5 , and M6 at the proximity of the K + -binding site. The slow kinetics of ouabain binding may be associated with partial unwinding of the M4 helix.
Large differences in ouabain sensitivity occur among animal species. The α1 isoform is ouabain-resistant in rat, mouse, and toad Bufo marinus , but is sensitive in human, rabbit, sheep, and Xenopus . The α2 and α3 isoforms are more sensitive than the α1 isoform in “resistant” species. However, in humans, little difference of equilibrium binding is found among α1, α2, and α3 isoforms, except for a slightly higher K I for α2. However, the α2 isoform also exhibits a faster ouabain association and dissociation rate constant than α1 and α3. The “resistant” phenotype of some species is linked to charged amino acids in the first extracellular loop between M1 and M2 . The presence of endogenous circulating inhibitors of Na,K-ATPase (“endo-ouabain”) is well-demonstrated. However, their precise chemical nature needs to be clarified, their controlled synthesis and release better understood, and their specific effects more precisely described, before the hypothesis of controlled Na,K-ATPase activity by such circulating hormones can be considered as fully established.
Gastric H,K-ATPase is insensitive to ouabain, as demonstrated by an absence of detectable effects at millimolar concentrations of ouabain. On the other hand, non-gastric H,K-ATPases, expressed in Xenopus oocytes, show some sensitivity to ouabain and exhibit inhibitory constants (K I ) between 10 and 100 mM. Two types of gastric H,K-pump inhibitors are known. SCH-28080 is a reversible inhibitor that competes with extracellular K + , while substituted benzimidazole compounds irreversibly inhibit gastric H,K-ATPase by forming a covalent (disulfide) bond between the sulfonamide form of the compound (produced in very acid pH) and the thiol group of a cysteine residue exposed at the cell surface.
Gene inactivation studies have shown that Na,K-ATPase is essential for life. Absence of Na,K-ATPase α1 expression is embryonically lethal, even though α1 −/− embryos develop beyond the blastocyst stage. In contrast, α2 −/− animals are born alive but die soon after birth, most likely as the result of dysfunctional neuronal circuits involved in controlled respiration. α1 or α2 gene inactivation in heterozygous animal tissues leads mostly to disturbances of cardiac phenotype, supporting an important role for the α2 isoform in intracellular calcium homeostasis. Only two human genetic diseases are known to be related to mutations of Na,K-ATPase α isoforms. A familial form of hemiplegic migraine has been associated with mutations in the α2 gene that abolish or greatly reduce enzyme activity (OMIM #602481). Rapid onset dystonia parkinsonism has been associated with an inactivating mutation of the α3 gene (OMIM #128235). In both cases, the mode of inheritance is dominant, suggesting that the disease is due to a haploinsufficiency of α2 and α3 isoforms, respectively. A dominant form of familial hypomagnesemia has been associated with a mutation in the FXYD2 gene.
Regulation of Na,K-ATPase
Obviously, the activity of an enzyme of such paramount importance for so many cellular and organ functions must be tightly regulated. Physiological regulation of Na,K-ATPase activity is complex and occurs at several levels. First, the activity of Na,K-ATPase depends on its three substrates, Na + , K + , and ATP. Second, for short-term regulation, the activity of Na,K-ATPase present at the cell surface can be regulated by post-translational modifications, such as phosphorylation. Third, Na,K-ATPase density at the cell surface is controlled by its rate of synthesis, at both transcriptional and post-transcriptional levels, by its degradation, and by its distribution between the cell surface and intracellular pools. Finally, there is some evidence that protein–protein interactions further regulate Na,K-ATPase activity and abundance.
Interaction of intracellular Mg-ATP with Na,K-ATPase is complex and involves high- and low-affinity sites. Since intracellular ATP concentration is usually largely above K ½ values, ATP is not considered a physiological limiting factor for Na,K-ATPase activity. However, under pathological conditions, such as ischemia/hypoxia, ATP may become rate limiting. Similarly, since normal concentrations of extracellular K + (3.5–4.5 mM) are above K ½ values, physiological variations are not expected to influence Na,K-ATPase activity, although decreased activity could occur under conditions of severe hypokalemia. In contrast, intracellular Na + concentration is below, or close to, K ½ values. Considering the steep concentration–activity relationship, with a Hill coefficient between 2 and 3, a low concentration of intracellular Na + implies that Na,K-ATPase operates far from its maximal rate under physiological conditions. It follows that small variations in intracellular Na + are immediately followed by parallel variations in Na,K-ATPase activity, in order to maintain a constant intracellular Na + concentration.
Intracellular Na + and extracellular K + concentrations can also have a long-term effect on Na,K-pump density. The influence of extracellular K + on α2 expression was demonstrated in skeletal muscle. Under conditions of hypokalemia, α2 is downregulated, allowing release of K + from large intracellular pools in muscle and fine regulation of K + homeostasis in small extracellular pools. Similarly, intracellular Na + modulates Na,K-ATPase expression (see “Na,K-ATPase and the Kidney”).
Na,K-ATPase α-subunits can be phosphorylated by several protein kinases. PKA phosphorylates a serine residue (S 943 ) located in a short intracellular loop that links the M8 and M9 segments, conserved in all Na,K-ATPase and H,K-ATPase isoforms. However, the functional relevance of this PKA site has been questioned, because it should be poorly accessible under native conformational states, and because its mutation did not alter the effect of PKA on Na,K-ATPase activity in renal epithelial cells. Activation of PKC also results in Na,K-ATPase α-subunit phosphorylation. A conserved, non-conventional PKC phosphorylation site, S 16 , located in the intracellular N-terminal domain of the α1-subunit, as well as two species-specific PKC sites (S 23 in rat α1, and T 15 in Bufo marinus α1) have been identified. α2- and α3-subunits are poor substrates for PKC or are not phosphorylated at all. PKC phosphorylation of the α-subunit may regulate both apparent Na + affinity and cell surface expression of the enzyme. More recently, tyrosine phosphorylation of α1 at Y 10 in response to insulin, and also at Y 260 in response to ouabain, has been described. Similarly, tyrosine phosphorylation of α2 was reported, although the phosphorylation site has not been mapped. Tyrosine phosphorylation of Na,K-ATPase is functionally associated with both an increase of apparent Na + affinity and an increase of the enzyme’s cell surface expression. Finally, phosphorylation of α1 by ERK has been shown in vitro and in response to insulin in skeletal muscle. This ERK-dependent phosphorylation is associated with stimulation of Na,K-ATPase activity via increased cell surface expression.
Synthesis and Degradation
Since αβ-subunit stoichiometry is inflexible, controlled synthesis of both subunits is expected to participate in regulating Na,K-ATPase activity. Hormonal control by glucocorticoids, mineralocorticoids, and thyroid hormones has been demonstrated. Thyroid hormones have been shown to regulate the synthesis of both subunits, but the mechanism of this regulation is complex and differs between organs. For instance, glucocorticoids stimulate the transcription of mostly β-subunit mRNA in lungs, even though expression of both α and β protein is increased, indicative of complex transcriptional and post-transcriptional control. Regulation of Na,K-ATPase by mineralocorticoids will be discussed later in the section “Na,K-ATPase in the Kidney.”
In addition to the rate of synthesis of its subunits, Na,K-ATPase abundance is dependent on its degradation rate. Recent experimental evidence obtained by pulse-chase experiments in lung alveolar and renal (E. Feraille, unpublished results) epithelial cells indicates that the half-life of plasma membrane Na,K-ATPase is approximately 4 hours, while that of newly synthesized α-subunits is approximately 6 hours. Therefore, in contrast to current belief, Na,K-ATPase half-life is relatively short in epithelial cells. The majority of Na,K-ATPase is degraded by lysosomes, and ubiquitination of its α-subunit may participate in this process. Regulation of Na,K-ATPase degradation in response to variations of transepithelial sodium transport will be discussed later in the section “Na,K-ATPase in the Kidney.”
An increasing amount of experimental evidence indicates that plasma membrane Na,K-ATPase expression is controlled via regulated membrane trafficking. Regulated Na,K-ATPase endocytosis was first demonstrated in response to dopamine by atypical PKC-ζ in rat proximal tubule cells. This will be discussed later in the section “Na,K-ATPase in the Kidney.” The role of PKC-ζ in Na,K-ATPase endocytosis has been confirmed in alveolar lung cells in response to hypoxia. Downregulation of plasma membrane Na,K-ATPase has also been demonstrated in response to ouabain in renal epithelial cells (LLCPK1 cells), and to AMPK activation in response to CO 2 in lung alveolar cells. Moreover, direct interaction of the Na-K-ATPase α-subunit with arrestins and spinophilin modulates its endocytotic rate in COS-7 cells. It remains to be determined whether internalized Na,K-ATPase units are degraded or recycled back to the plasma membrane.
Rapid recruitment of Na,K-ATPase to the plasma membrane has first been demonstrated in response to insulin in skeletal muscle and following an increase of intracellular sodium concentration in renal collecting ducts. This latter event will be discussed in the section “Na,K-ATPase in the Kidney.” Increased cell surface expression of Na,K-ATPase was subsequently demonstrated in response to aldosterone in renal collecting duct cells, and to increased cAMP levels in both renal and lung epithelial cells. This process is sensitive to brefeldin A, suggesting that a latent pool of Na,K-ATPase units is recruited from the trans-Golgi network. In addition, activation of PKC-β in proximal tubule cells and PKC-δ–εδ in lung alveolar cells increases Na,K-ATPase plasma membrane expression. Exocytosis of Na,K-ATPase is dependent on RhoA, kinesin, and myosin-Va in lung alveolar cells. The amount of active Na,K-ATPase units at the cell surface is therefore highly regulated via both endocytotic and exocytotic processes.
Interaction with the Cytoskeleton
Na-K-ATPase interacts both directly and indirectly with the membrane cytoskeleton in several ways. Na,K-ATPase binds directly to ankyrin, which itself links Na,K-ATPase to spectrin (fodrin) and consequently to the basolateral cytoskeleton. Ankyrin-spectrin interaction is thought to participate in specific targeting of Na-K-ATPase to the basolateral membrane of epithelial cells. Na-K-ATPase activity may also be modulated via its association with ankyrin.
Na,K-ATPase and actin filaments display a complex relationship that is not fully understood. Actin filaments may directly modulate Na,K-ATPase activity or indirectly control Na,K-ATPase plasma membrane expression via adducins which control actin polymerization. Mutant α-adducin associated with hypertension in rats and humans was shown to decrease constitutive Na,K-ATPase endocytosis, and thereby increase the number of active Na,K-ATPase units at the cell surface. Direct activation of Na,K-ATPase activity by adducin was also demonstrated. In addition, Na,K-ATPase directly binds to cofilin, which modulates actin filament polymerization. Na,K-ATPase basolateral targeting and modulation of its activity appears additionally to depend on its interaction with multiple cytoskeletal elements.
New Physiological Functions of Na,k-ATPase
In addition to its well-known function, i.e., Na + and K + membrane translocation, Na,K-ATPase modulates a variety of cellular processes involved in cell growth, differentiation, and intercellular adhesion.
Cell Signaling by Na,K-ATPase
Since the pioneering work of Askari et al. which demonstrated that ouabain induces early response genes involved in cardiac hypertrophy in a Ca 2+ -dependent manner, a large body of experimental evidence collectively indicates that Na,K-ATPase behaves as a signaling platform in both mesenchymal and epithelial cells. Activation of Src and Epidermal Growth Factor (EGF) receptor tyrosine kinases, and downstream activation of MAP kinases (ERK) by ouabain at concentrations that do not alter intracellular Na + concentration was demonstrated in cardiac myocytes and in LLC-PK1 renal epithelial cells. Ouabain induces dissociation and activation of Src from Na,K-ATPase, and subsequent transactivation of EGF receptor. This ouabain-induced signaling depends on a pool of non-functional Na,K-ATPase specifically located in caveola. In addition to ERK activation, ouabain induces calcium oscillations via close association of phospholipase C and IP3 receptors in cell signaling microdomains. However, these ouabain-dependent signaling events remain to be demonstrated in native tissues, and their physiological relevance is not yet established.
Role of Na,K-ATPase in Cell Adhesion
An increasing amount of experimental evidence indicates that the Na,K-ATPase β1-subunit plays a direct role in homotypic cell–cell adhesion. This cell–cell adhesion function is negatively correlated with the level of complexity of β glycosylation. Moreover, the transcription factor Snail, which plays a key role in epithelial-to-mesenchymal transition processes observed in poorly differentiated carcinoma cells, represses Na,K-ATPase β1-subunit transcription. In addition to a direct inhibition of cell motility via increased cell–cell adhesion, Na,K-ATPase may also modulate cell migration via its interaction with the PI3-kinase p85 regulatory subunit. This adhesive function of Na,K-ATPase may play important roles in organogenesis and carcinoma invasiveness.
Na,K-ATPase also modulates tight junction dynamics and permeability in cultured renal epithelial cells. Indeed, both ouabain and low extracellular K + , which inhibit Na,K-ATPase and increase intracellular Na + , prevent tight junction recovery in calcium switch experiments performed on MDCK cells. This effect is dependent on inhibition of RhoA GTPase activity by high intracellular Na + . In addition, ouabain at concentrations that do not alter transcellular K + transport increases tight junction permeability via activation of Src and ERK. The physiological relevance of these findings remains to be demonstrated.
Na,K-ATPase in the Kidney
Tubular epithelial cells are characterized by their functional polarity. Na,K-ATPase is exclusively located in the basolateral membrane, the infoldings of which are closely surrounded by mitochondria which provide a constant supply of ATP. The Na + gradient generated by Na,K-ATPase between intra- and extracellular compartments is mainly dissipated across the apical membrane. A net reabsorption of Na + results from this architectural organization, and the main role of Na,K-ATPase in the kidney is to energize Na + reabsorption. In humans, kidneys reabsorb over 500 g of sodium per day, and utilize over 2 kg of ATP to fuel Na,K-ATPase. Although renal Na,K-ATPase also energizes secondary active transport of other solutes, its main function is related to Na + transport.
Measurements in microdissected segments of the mammalian renal tubule indicate that Na,K-ATPase activity is high in the thick ascending limb of Henle’s loop (TAL) and the distal convoluted tubule (DCT), intermediate in the proximal tubule (PT) and the collecting duct (CD), and very low in the thin segments of Henle’s loop. This distribution profile of Na,K-ATPase activity is correlated with transtubular Na + reabsorption capacity in various nephron segments. In PT and CD, Na,K-ATPase activity declines from the kidney cortex toward the outer and inner medulla. This distribution profile is confirmed by immunocytochemistry on kidney sections and by quantification of the number of pumps by 3 H-ouabain binding or by Western blotting of α1- and β1-subunits. Immunohistochemistry on kidney sections indicate that Na,K-ATPase expression in the collecting duct is much higher in principal than in intercalated cells. In contrast, quantification of Na,K-ATPase mRNA along the rat nephron does not confirm this axial heterogenic distribution, at least for α- and β-subunits (A. Doucet, personnal communication). This suggests the presence of segment-specific control mechanisms for Na,K-ATPase translation and/or degradation.
Regulation of Na,K-ATPase in Proximal Tubule
Na,K-ATPase activity must be tightly controlled, since the PT reabsorbs the bulk of filtrated sodium (more than 60%). Na,K-ATPase is regulated by hormones, neurotransmitters, and para/autocrine factors acting via synergistic or antagonistic signaling pathways, and one should keep in mind that the final effect, i.e., stimulation of inhibition of Na,K-ATPase activity, is not the result of activation of a single signaling pathway, but rather of a highly complex integrated response. Modulation of Na,K-ATPase cell surface expression and its affinity for Na + are the most important mechanisms of regulation identified in PT.
Control of Na,K-ATPase by Insulin
Stimulation of Na,K-ATPase by insulin in the PT most likely participates in stimulation of Na + reabsorption in this nephron segment. Indeed, in isolated rat PT, insulin stimulates Na,K-ATPase transport activity in the presence of physiological concentrations of intracellular Na + , not by changing V max values, but by increasing apparent Na + affinity. Experiments performed in isolated rat PT and in cultured OK cells (a cellular model of PT) strongly suggest that stimulation of Na,K-ATPase activity relies on phosphorylation of the α-subunit at Y 10 . However, the causal relationship between tyrosine phosphorylation of Na,K-ATPase and increased Na + affinity remains to be directly demonstrated. Insulin also reduces the inhibitory effect of dopamine on Na,K-ATPase (see next section), and this may also participate in its overall stimulatory effect.
Control of Na,K-ATPase by Dopamine and Parathormone
Dopamine is produced from L-dopa by PT, and its synthesis is increased by high sodium intake, making it a putative local modulator of sodium and fluid handling. Dopamine decreases fluid and sodium reabsorption in vitro in microperfused rabbit PT. Accordingly, dopamine decreases Na,K-ATPase V max values in rat PT, as well as in OK cells. Although dopamine also increases Na,K-ATPase apparent affinity for Na + in the PT, the overall resulting effect in intact cells is Na,K-ATPase inhibition. This inhibition results from activation of both dopamine DA 1 – and DA 2 -like receptors, and is mediated by PKCζ.
Within minutes, dopamine induces Na,K-ATPase endocytosis from the plasma membrane to intracellular compartments. Decreased Na,K-ATPase expression at the basolateral membrane of rat PT is associated with a sequential increase of Na,K-ATPase abundance in clathrin-coated pits (1 minute), early endosomes (2.5 minutes), and late endosomes (5 minutes). Results obtained in OK cells strongly suggest that the inhibitory effect of dopamine is dependent on PKC-mediated phosphorylation of the Na,K-ATPase α-subunit on S 23 . Following endocytosis, the α-subunit is dephosphorylated in late endosome compartments ( Figure 3.3 ).
Dopamine-induced endocytosis of Na,K-ATPase is also associated with activation of phosphatidylinositol 3-kinase (PI3K), an enzyme critical for membrane trafficking. However, dopamine-induced activation of PI3K is not secondary to its phosphorylation. Rather, S 23 phosphorylation of the Na,K-ATPase α-subunit serves as an anchor signal for the sequential recruitment of 14-3-3 protein and PI3K to the membrane. Activation of PI3K in turn generates local production of phosphatidylinositol 3-phosphate, which allows binding of Na,K-ATPase with adaptor protein-2 (AP2), recruitment of clathrin, and endocytosis of Na,K-ATPase. Concomitantly, dopamine activates protein phosphatase 2A which in turn dephosphorylates dynamin 2, thus allowing Na,K-ATPase recruitment at the plasma membrane.
Parathormone (PTH) inhibits fluid and solute transport by PT, at least in part via inhibition of Na,K-ATPase activity, as first demonstrated in PT suspensions. This inhibitory effect relies partly on generation of arachidic acid metabolites via the cytochrome P450 pathway. Na,K-ATPase endocytosis in response to PTH has been observed both in vitro , in OK cells, and in vivo after infusion of PTH in normal rats, and may account for the inhibitory effect of this hormone. Studies performed in the OK cell model have shown that PTH-induced endocytosis of Na,K-ATPase is ERK-dependent, requires S 16 phosphorylation of the α-subunit and the scaffolding protein NHERF1.
Endocytotic removal of active Na,K-ATPase from the plasma membrane therefore constitutes a major regulatory mechanism of fluid and sodium reabsorption by the PT.
Angiotensin II Exerts a Biphasic Effect on Na,K-ATPase
Angiotensin II (ANG II) controls PT Na + reabsorption from both luminal and basolateral sides. High concentrations of ANG II can be found in the PT lumen with respect to plasma circulating concentrations, suggesting that the majority of luminal ANG II originates from local synthesis. Indeed, PT cells express angiotensinogen, renin, and angiotensin converting enzyme. Only a very small fraction of ANG II is excreted in the urine, since it is almost entirely reabsorbed and degraded by PT. ANG II exerts a biphasic effect. Low levels of ANG II (12 −12 to 10 −10 M) stimulate, while high levels of ANG II (10 −9 to 10 −7 M) inhibit, fluid and solute reabsorption in vivo in microperfused rat PT. ANG II also modulates Na,K-ATPase activity in a biphasic manner in isolated PT. At low concentrations, ANG II enhances Na,K-ATPase activity via an increase in its apparent Na + affinity. The mechanism by which high levels of ANG II inhibit Na,K-ATPase activity remains to be determined.
Regulation of Na,K-ATPase in Thick Ascending Limb of Henle’s Loop
The TAL reabsorbs close to 15% of the filtered Na + load and is impermeable to water. As a result, fluid delivered to the distal convoluted tubule is hypotonic, NaCl concentration levels being close to 50 mM. Na + enters the luminal side of the cell via a furosemide-sensitive Na-K-2Cl-co-transporter (BSC1 or NKCC2), and Cl − leaves the cell via Cl − -channels and K-Cl co-transporters. K + is recycled back to the lumen via inwardly rectifying and voltage-insensitive K + -channels expressed at the apical membrane. Conductive diffusion of Cl − and K + depolarizes the basolateral membrane and hyperpolarizes the apical membrane, respectively. The combination of both diffusion potentials generates positive transepithelial voltage which provides the driving force for paracellular cation reabsorption.
Micropuncture and in vitro microperfusion experiments show that cAMP analogs and hormones coupled to adenylyl cyclase activation enhance NaCl reabsorption in the TAL. Stimulation of Na + reabsorption by the cAMP signaling pathway is at least in part mediated by stimulation of Na,K-ATPase activity. This stimulatory effect is observed at V max values, and requires sufficient amounts of oxygen and metabolic substrates. Indeed, when metabolic supply is limiting, an increase of cellular cAMP content actually inhibits Na,K-ATPase activity via generation of arachidonic acid metabolites that in turn is dependent on the cytochrome P450 pathway. Na,K-ATPase stimulation by cAMP is correlated with increased phosphorylation levels of the Na,K-ATPase α-subunit. It remains to be determined whether this effect results from phosphorylation of the α-subunit by PKA or by another kinase.
The stimulatory effect of the cAMP/PKA signaling pathway is subject to negative modulation by numerous signaling pathways, including protein Gαi activation by prostaglandins, cGMP generation in response to nitric oxide and natriuretic peptides, and PKC stimulation in response to angiotensin II, bradykinin, and extracellular Ca 2+ via activation of the extracellular Ca 2+ receptor.
Control of Na,K-ATPase in Collecting Duct
Mammalian connecting tubules (CNT) and CD (cortical collecting duct: CCD, outer medullary collecting duct: OMCD, inner medullary collecting duct: IMCD) are the main sites for the fine-tuning of sodium reabsorption, crucial for the adjustment of daily urinary sodium excretion to dietary intake. In these renal tubule segments, apical sodium entry is primarily mediated by amiloride-sensitive sodium channels (ENaC), and accessorily mediated by a recently identified sodium-dependent chloride/bicarbonate exchanger, SLC4A8. ENaC was long thought to be the principal player for sodium transport, Na,K-ATPase activity being adapted secondarily to changes of intracellular Na + concentration, itself brought about by changes in ENaC activity. However, it is now clearly established that regulation of sodium transport results from the coordinated regulation of both ENaC and Na,K-ATPase.
Aldosterone Induces a Biphasic Stimulation of Na,K-ATPase
Na,K-ATPase activity (V max values) is decreased by ~70% in collecting ducts within 4–5 days of adrenalectomy, but is increased following administration of supraphysiological doses of mineralocorticoid. This latter effect appears after a 24-hour latency, and culminates after approximately 6 days. Administration of aldosterone to adrenalectomized animals or in vitro addition of aldosterone to renal tubules isolated from adrenalectomized animals also increases Na,K-ATPase activity (V max values) in the CD, but this effect is much more rapid than in adrenal-intact animals since it is observed after only an hour, and is maximal after only 2–3 hours. These data indicate that low levels of CCD Na,K-ATPase activity and plasma aldosterone concentration is associated with a fast stimulatory response.
Short-term (2–6 hour) aldosterone challenge increases both the activity and cell surface expression of Na,K-ATPase via recruitment of a latent pool of pumps, as shown both in CD isolated from adrenalectomized rats, and in cultured mpkCCD C14 cells, a model of CCD principal cells ( Figure 3.4 ). Although cell fractionation and cell-surface labeling studies suggest that a latent pool of Na,K-ATPase is intracellular, its exact localization has not yet been established. Short-term stimulation of Na,K-ATPase activity occurs independently of ENaC and apical sodium entry, but does depends on de novo transcription and translation. Experiments performed on Xenopus oocytes demonstrated that serum and glucocorticoid-regulated kinase-1 (SGK1), an early aldosterone-inducible gene, increased Na,K-ATPase activity and cell surface expression. This suggests that aldosterone-induced recruitment of latent Na,K-ATPase units may be mediated by SGK1. Early recruitment of latent Na,K-ATPase is followed by transcriptional stimulation and synthesis of Na,K-ATPase α1- and β1-subunits. In summary, after a latency period of 1 hour, aldosterone stimulates Na,K-ATPase activity in a biphasic manner. First via recruitment of an inactive Na,K-ATPase reservoir to the cell surface, and then by increased synthesis of Na,K-ATPase subunits.
Vasopressin Stimulates Na,K-ATPase
Vasopressin (AVP) is coupled to adenylyl cyclase via V 2 receptors, and stimulates the cAMP/PKA signaling pathway in CD principal cells. The major role of AVP in CD is to stimulate water reabsorption by increasing water permeability of the apical membrane of principal cells. However, in vitro microperfusion studies have shown that AVP also stimulates Na + reabsorption and K + secretion along the CD. This effect of AVP on urinary Na + excretion has recently been confirmed in humans.
Na,K-ATPase stimulation is a prerequisite for increased Na + reabsorption, but initial studies reported an inhibitory effect of AVP and cAMP analogs on Na,K-ATPase activity in isolated rat CCD. Results from our laboratory indicate that an arachidonic acid-dependent inhibitory pathway is induced by metabolic stress related to ex vivo experimental conditions. Indeed, in both well-oxygenated, isolated rat CCDs, and cultured mpkCCD Cl4 cells, cAMP analogs induced a two-fold stimulation of Na,K-ATPase activity. This stimulatory effect is rapid (5 min), and is associated with a proportional increase of Na,K-ATPase cell surface expression in the absence of a change in whole-cell Na,K-ATPase abundance in a PKA-dependent manner ( Figure 3.4 ). Identification of a cAMP-responsive Na,K-ATPase pool and its relation to an aldosterone controlled reservoir remains to be determined.
Na,K-ATPase Expression is Regulated by Sodium Availability
Acute (hour) increases of intracellular Na + in CD not only activate Na,K-ATPase activity via a substrate effect, but also rapidly increase V max values and the number of active Na,K-ATPase units present at the cell surface. This rapid stimulation occurs independently of protein synthesis, suggesting that pre-existing Na,K-ATPase units present in a latent pool are recruited to the cell surface (see above). The effect of intracellular Na + is mediated by cAMP-independent PKA activation that itself results from dissociation of a complex consisting of NF-κB, IκBα and the catalytic PKA subunit. This indicates that the pro-inflammatory NF-κB pathway may be part of a Na + sensing mechanism that mediates cross-talk between apical Na + entry and basolateral Na + exit ( Figure 3.4 ). Recent experimental evidence indicates that a sustained increase of apical Na + entry also increases Na,K-ATPase expression via inhibition of Na,K-ATPase endocytosis and degradation (E. Feraille, personal communication). Such modulation of Na,K-ATPase activity by Na + load may be involved, to some extent at least, in diuretic resistance. Accordingly, chronic administration of the loop diuretic furosemide stimulates CD Na,K-ATPase activity independently of variations of circulating aldosterone.
Induction of Na,K-ATPase is Associated with Sodium Retention in Nephrotic Syndrome and Liver Cirrhosis
Interstitial edema is a cardinal clinical manifestation in nephrotic syndrome. It is secondary to the accumulation of sodium in the extracellular compartment following imbalanced dietary sodium intake and urinary sodium output, and also results from alterations of fluid transfer across the capillary endothelial barrier. Mechanisms behind sodium retention have been extensively investigated using a rat model of puromycin aminonucleoside (PAN)-induced nephrotic syndrome. Both in vivo and in vitro studies demonstrated that the CCD is the main site of increased Na + reabsorption in PAN nephrotic rats. Na,K-ATPase hydrolytic and transport activities are increased two-fold in CCD of PAN nephrotic rats. Increased Na,K-ATPase stimulation, which culminates at day 6 following PAN administration, parallels decreased urinary sodium excretion and development of a positive sodium balance. Moreover, a linear inverse correlation between urinary sodium excretion and CD Na,K-ATPase activity is observed in three different experimental models of nephrotic syndrome. Stimulation of Na,K-ATPase activity is paralleled by increased abundance of α and β Na,K-ATPase mRNA and basolateral protein expression. Sodium retention and induction of Na,K-ATPase activity in CCD are independent of variations of circulating aldosterone.
Interstitial edema and ascites are frequently observed in liver cirrhosis. The mechanism governing sodium retention in liver cirrhosis is under debate; however, dysregulation of CD Na,K-ATPase activity may play a key role. Indeed, in the bile duct ligation model Na,K-ATPase activity is specifically increased in mouse CCD independently of variations of circulating glucocorticoids or aldosterone.
In summary, increased Na,K-ATPase activity in the CCD participates in sodium retention in both experimental nephrotic syndrome and liver cirrhosis, and may therefore play a key role in the pathogenesis of edematous diseases.
The previous paragraphs collectively outline the complexity of Na,K-ATPase regulation in the kidney. They demonstrate that: (1) hormonal triggering of intracellular signaling pathways can rapidly alter Na,K-ATPase V max values and/or modulate the pump’s affinity for Na + ; (2) these changes are true regulatory mechanisms that occur independently of changes in apical Na + entry; and (3) they are generally accompanied by a concomitant regulation of apical Na + entry. Thus, despite repetitive and rapid changes of Na + reabsorption, whole-body sodium balance and intracellular Na + homeostasis are maintained, at least in part, by short-term regulation of Na,K-ATPase activity. Finally, alterations of Na,K-ATPase activity are relevant to the pathophysiology of edematous diseases.