Renal Physiology

Potassium secretion and the regulation of distal nephron K channels

Lawrence G. Palmer


K-selective channels in the luminal membranes of distal nephron segments form a key pathway for the secretion of K ions into the urine. This process is important to the control of K balance, particularly under conditions of normal or high K intake. This brief review will cover three issues:1) the identification of apical K channels, 2) the role of these channels in the maintenance of K homeostasis, and3) the role of aldosterone in this regulatory process. The large amount of literature on renal K transport has been elegantly summarized in a recent review in this journal [G. Giebisch. Am. J. Physiol.274 (Renal Physiol. 43): F817–F833, 1998]. Here I will focus on a few prominent unsolved problems.

  • ROMK channels
  • potassium adaptation
  • mineralocorticoids
  • kaliuretic hormones

the kidney is able to adjust the excretion of potassium in the urine to match dietary intake. This may be accomplished in part by the regulation of the activity of apical K channels in the luminal membrane of the collecting tubule. These channels have been identified at the molecular level as ROMK, a member of the inward-rectifier K channel family. The pathways involved in their regulation during changes in K intake remain unidentified.


Physiological identification. Three types of K-selective channels have been identified in the apical membrane of distal nephron cells, especially from the cortical collecting tubule (CCT) and the thick ascending limb (TAL), using the patch-clamp technique. The most prominent type in the rat CCT is a low-conductance (40 pS) channel with the properties of high, voltage-independent open probability (P o) and moderate inward rectification (9, 54). We refer to these as SK channels. Very similar channels were observed in the rabbit and rat TAL (51, 53). This segment also has a second inward-rectifying apical K channel type with a higher single-channel conductance (70 pS) (51). High-conductance (100–200 pS) K channels, activated by intracellular Ca and membrane depolarization, are present in the CCT (8, 17). These so-called BK channels are thought to mediate at most a small fraction of K secretion in the mammalian kidney, for several reasons. First, they are found at relatively low densities relative to the SK channels (9, 54). Second, they are preferentially expressed in intercalated cells, which are thought not to participate in an important way to K secretion (29). Third, they have a very lowP o at the apical membrane resting potential, at least under normal conditions (8). However, these channels may represent important routes for K secretion in other epithelia such as the amphibian collecting tubule (43) and the mammalian distal colon (5).

Molecular identification. ROMK channels were isolated from a rat kidney cDNA library using expression cloning (14). These channels have been designated as Kir 1.1, indicating that they are the prototypes for the inward rectifier family of K channels, as distinct from the voltage-gated or Kv family. There are at least three isoforms of ROMK in the rat (4, 60) and at least five in the human (41), apparently arising from different splicing of the same gene. ROMK proteins form the SK channels in the distal nephron. This identification is based in part on the similarities of biophysical properties (31) and regulatory properties (52) between the SK channels in the CCT and TAL and the ROMK channels expressed in oocytes. ROMK mRNA is expressed in both the TAL and the CCT, as well as the intervening distal convoluted tubule and connecting tubule segments (4,11). Antibodies raised against several different ROMK partial sequences recognize epitopes on the luminal membranes of these same nephron segments (18, 26, 57).

RACKT1 is a second candidate K channel gene, which was isolated from a rabbit CCT-cell cDNA library (46). Antibodies raised against the RACKT1 sequence also recognize apical membrane epitopes in the rat CCT (47). RACKT1 cDNA induces K-selective channels when expressed in CHO cells (45) but the biophysical properties of these channels do not match those observed in native cells. The single-channel conductance (80 pS) is higher than that of the SK channels of the CCT and TAL. They are also activated by intracellular Ca2+, which distinguishes them from both the SK and 70-pS channels. One report challenged the identity of RACKT1 as a bona fide mammalian gene product (40). This issue has not been fully resolved (44).

TWIK-1 (KCNK1) is another candidate for a renal secretory K channel. This protein, which is similar to two inward-rectifier channel subunits arranged in tandem, is widely expressed in many different tissues including the kidney (6, 19, 28). It forms K-selective channels when expressed in Xenopus oocytes (19), although, like RACKT1, the characteristics of these channels do not precisely match those of native renal cells. In the rabbit kidney, expression of mRNA for this channel was mainly in the cortical TAL and the collecting duct (28). In the rat, immunolocalization indicated expression of protein predominantly in the proximal tubule brush border membrane, the TAL and intercalated cells of the collecting duct (6). The role of intercalated cells in general, and TWIK channels in particular, in K secretion of the CCT remains to be established.

Unsolved problems. Do ROMK channels account for K secretion by the distal nephron? Although ROMK is almost certainly an important pathway for K secretion, mutations of this protein leading to loss of function can produce Bartter's syndrome, whose phenotype includes urinary K loss and hypokalemia (42). A likely cause is inhibition of salt transport by the TAL. This would increase both flow rate and Na delivery to the distal nephron, both of which will enhance K secretion (12). This implies, however, that there is a route for K secretion other than ROMK. What this route might be is unclear. It is possible that other K channels (e.g., BK, RACKT1, TWIK-1) which do not appear to be abundant in the CCT under normal conditions become more prevalent when ROMK is absent or dysfunctional. Furthermore, little is known about whether K channels other than ROMK or other K transporters are expressed in the connecting tubule and/or the late distal convoluted tubule. Since these segments represent important sites for K secretion, this information would be very useful to have. It will require development of methods to prepare these tubules for patch-clamp analysis of their apical membranes. A related unsolved issue is the molecular identification of the 70-pS channel in the TAL.


Response to changes in K intake. It has been known for a long time that rats fed a diet rich in K adapt by upregulating their machinery for secreting K (23). One component of this adaptation involves an increase in the density of SK channels in the CCT. The number of channels per patch increased by a factor of 3–4 when the animals ate a high-K diet for 1–2 wk (30, 54). The densities of conducting apical membrane Na channels, and the activity of the Na/K pump also increased (30). More recent measurements indicated that this stimulation of channel activity can occur over a much shorter time course. SK channel density increased by more than threefold after 48 h and by more than twofold after just 6 h (essentially one meal) on the high-K diet (10). This time course is similar to that of the increase in K excretion observed when rats are fed a high-K diet (35). This suggests that changes in SK channel activity play a role in the day-to-day regulation of K balance by helping to match renal K excretion with K intake.

There is no evidence that these responses involve a concomitant increase in the amount of channel mRNA or protein in the cell. ROMK mRNA abundance was not significantly enhanced in animals adapted to a high-K diet when assessed using Northern blot analysis of whole kidney cortex (49) or in isolated CCTs using a semiquantitative in situ hybridization assay (11). Similarly, the amount of ROMK protein in rat kidney cortex, estimated from immunoblots using antibodies raised against the COOH terminus, was unchanged in K-adapted animals (P. A. Mennitt et al., unpublished observations). This has not been confirmed at the single-tubule level. The most likely interpretation is that the increase in density of conducting channels involves an activation or translocation of existing ROMK proteins.

Developmental regulation. Newborn animals generally need to retain K for growth. In rabbits, the ability of the CCT to secrete K does not appear until the third week after birth (37). Patch-clamp studies on the rabbit CCT showed that the abundance of SK channels was well matched with the onset of K secretion (38). The channels were undetected in week 1 after birth, and increased in density gradually fromweeks 2 to5. This developmental regulation of channel function likely involves changes in the abundance of channel protein. Single-tubule RT-PCR measurements indicated that ROMK mRNA was absent in 1-wk-old and scarce in 2-wk-old rabbits relative to older animals (3). Immunocytochemical analysis of sections of newborn rat kidney indicated the absence of ROMK protein in the cortex. Clear apical staining of the CCTs was observed in 3-wk-old animals (61), consistent with the time course of the appearance of conducting channels.

Unsolved problems. How could quiescent channels be activated during K adaptation? The answer is not known, but there are a number of regulatory mechanisms known to affect SK/ROMK channels (see Refs. 12 and 52 for more thorough discussion). In inside-out patches, the channel activity can be increased by phosphorylation by protein kinase A (PKA) and decreased by dephosphorylation (24). Furthermore, expression of ROMK mutants in which potential PKA phosphorylation sites are abolished reduces channel activity (58). Phosphorylation of the channel protein is an attractive possibility, since it could well survive the process of tubule isolation. The channels can also be regulated by cell pH (48), ATP (25), phosphatidylinositol 4,5-bisphosphate (16), and arachidonic acid (22). It is not known whether any of these pathways might play a role in K adaptation or the maintenance of K homeostasis.


Is aldosterone a kaliuretic hormone?Many studies have shown that elevation of plasma aldosterone levels can stimulate renal K excretion. Primary or secondary hyperaldosteronism often leads to hypokalemia from excess K excretion, whereas aldosterone deficiency is usually associated with reduced K excretion and hyperkalemia (21). Experimental animals infused with mineralocorticoids also excrete more K (15, 34, 59).

High values of plasma K can stimulate aldosterone secretion by the glomerulosa cells of the adrenal cortex (13). This leads naturally to the idea that high K intake leads to an increase in plasma K and hence to elevation of circulating aldosterone and finally to increased K excretion (56). However, the operation of this feedback loop in the maintenance of plasma K within normal limits has been questioned (32). According to this view, other homeostatic mechanisms will keep plasma K at levels below those required to stimulate aldosterone release under most circumstances, including a moderate increase in K intake.

Indeed, a number of studies have indicated that renal K transport can be regulated to balance K intake without a corresponding change in aldosterone. In sheep, substantial changes in renal K excretion were observed in response to an increase in K intake and these changes were not correlated with altered plasma levels of either K or aldosterone (33). At the tubular level, K secretion by the isolated perfused rabbit CCT was increased by high dietary K even in adrenalectomized animals (55). Apical membrane K conductance of the CCT was also observed to increase in response to K loading in adrenalectomized rabbits (27).

Does aldosterone regulate K channel activity? A separate but related question is whether secretory K channels are regulated by aldosterone. This need not be the case, since augmented K secretion in response to the hormone could be secondary to the activation of Na transport by the CCT. This would have two important effects. First, entry of Na into the cell through Na channels will depolarize the apical membrane and increase the electrical driving force for K efflux into the urine. Second, Na entering the cell will stimulate the Na/K pump in the basolateral membrane, increasing the influx of K into the cell from the interstitial space.

Again the results are conflicting. In the rabbit CCT, infusion of the mineralocorticoid deoxycorticosterone acetate for 48 h led to an increase in apical K conductance (36). The time course was significantly slower than that of the increase in amiloride-sensitive apical Na conductance. In contrast, in the rat CCT, analysis of the effects of aldosterone on K fluxes and transmembrane electrical potentials suggested that stimulation of these fluxes could be accounted for by changes in the electrical driving force for K (39).

At the single-channel level, neither elevation of endogenous aldosterone secretion with a low-Na diet nor infusion with exogenous aldosterone increased SK channel density in the rat CCT (30). The density of these channels was increased, as mentioned above, with a diet rich in K, and plasma aldosterone increased only modestly under these conditions. However, in primary cultures of rabbit CCT cells, the expression of SK channels was dependent on the presence of aldosterone in the growth medium (20).

At the molecular level, adrenalectomy was shown to decrease the expression of ROMK mRNA in the rat kidney cortex (49). It is not known which nephron segments are involved, or whether the effect could be reversed by aldosterone repletion. In addition, Beesley et al. (2) recently reported that the expression of mRNA for some variants of ROMK by the rat kidney was enhanced by aldosterone infusion. These observations need to be resolved with the lack of effect of aldosterone on the density of conducting K channels in the rat CCT (30). These various results are most easily understood in terms of separate effects of aldosterone and plasma K on the SK/ROMK channels. In some epithelia (e.g., rabbit CCT), either of these factors may be sufficient to enhance apical K conductance. In others (e.g., rat CCT) both may be required. In either case, aldosterone and high plasma K can increase K excretion in a synergistic manner. In this way, the regulation of apical Na and K channels can be at least partially separated. This makes physiological sense, since it is not always desirable to change Na absorption and K secretion simultaneously.

Unsolved problems. If elevation of plasma aldosterone does not account for activation of apical K channels under conditions of high K intake, then what does? One possibility is that the cells of the distal nephron respond directly to increases in plasma K. Alternatively, a circulating factor other than aldosterone could be released in response to high plasma K and act on the cells of the distal nephron. Both vasopressin (7) and glucagon (1) can increase distal nephron K secretion. However, these hormones probably act through the generation of cAMP, which activates the K channels through activation of PKA (see the accompanying review by W. Wang; Ref. 50). It is not clear whether such an effect in vivo would be retained as an increased channel density measured in vitro. It is also not known whether the circulating levels of either of these hormones would be chronically elevated by increased plasma K. A third hypothesis is that the body senses K intake directly, rather than circulating plasma K, through receptors in the gut or in the portal circulation (32). This idea would put the sensing mechanism where the changes in K concentration would be largest. The latter two scenarios suggest the existence of a specific kaliuretic hormone that could control renal K excretion through activation of SK/ROMK channels. Identification of such a factor could help to clarify this rather murky picture of the control of renal K channels and K secretion.


  • Address for reprint requests and other correspondence: L. G. Palmer, Dept. of Physiology and Biophysics, Weill Medical College of Cornell Univ., 1300 York Ave., New York, NY 10021 (E-mail:lgpalm{at}


View Abstract