The Na+-K+-Cl- cotransporters (NKCCs), which belong to the cation-Cl- cotransporter (CCC) family, are able to translocate across cell membranes. In this study, we have used the oocyte expression system to determine whether the K+-Cl- cotransporters (KCCs) can also transport and whether they play a role in pH regulation. Our results demonstrate that all of the CCCs examined (NKCC1, NKCC2, KCC1, KCC3, and KCC4) can promote translocation, presumably through binding of the ion at the K+ site. Moreover, kinetic studies for both NKCCs and KCCs suggest that is an excellent surrogate of Rb+ or K+ and that transport and cellular acidification resulting from CCC activity are relevant physiologically. In this study, we have also found that CCCs are strongly and differentially affected by changes in intracellular pH (independently of intracellular ). Indeed, NKCC2, KCC1, KCC2, and KCC3 are inhibited at intracellular pH <7.5, whereas KCC4 is activated. These results indicate that certain CCC isoforms may be specialized to operate in acidic environments. CCC-mediated transport could bear great physiological implication given the ubiquitous distribution of these carriers.
- ion affinity
- collecting duct
various members of the cation-Cl- cotransporter (CCC) family, namely, K+-dependent CCCs (K+-CCCs), have been shown to mediate transport through binding of the ion at the K+ site (1, 24, 30, 31, 51). The K+-CCCs, which are highly homologous to one another, include two types of carriers: the Na+-K+-Cl- cotransporters (NKCC1 and 2; Refs. 18, 41, 53) and the K+-Cl- cotransporters (KCC1–4; see Refs. 20, 27, 38, 42, 45). Evidence supporting a direct implication of the K+-CCCs in transport, however, is only available for NKCC1 and 2 (24, 30, 31, 51); for these two isoforms, interestingly, behaves as an almost perfect surrogate of K+ or Rb+.
Four of the six known K+-CCCs (NKCC1, KCC1, KCC3, KCC4) have wide tissue distribution, with expression occurring in nonpolarized as well as polarized cell types (25, 29, 35, 36, 53). The two other K+-CCCs (NKCC2 and KCC2) are tissue specific; i.e., NKCC2 is found exclusively in the kidney (28, 41) and KCC2 in the brain (42). As a group of carriers, hence, the K+-CCCs could potentially facilitate the transmembrane movement of and modulate intracellular concentration () in a wide variety of cell types.
Several physiological effects may result from changes in . By way of illustration, an increase in cellular influx of (but not of NH3) is typically accompanied by a decrease in intracellular pH (pHi) as dissociation into NH3 and H+ takes place (1, 30, 33). Similarly, an increase in may accentuate glutamine synthesis in certain cell types (52) and increase vectorial - movement across epithelial tissues (12). In the kidney, for example, such movement is essential to ensure NH3 transfer from the proximal tubule, where the gas is produced, to the collecting duct (CD), where H+ ions are secreted (12, 14, 54).
Different groups have shown that extracellular acidification affects the normal operation of NKCCs and of the KCC1 isoform (Isenring and Forbush, unpublished observations; see also Refs. 26, 34); for example, the activity of NKCC1 in duck erythrocytes decreases progressively as extracellular pH (pHo) is lowered from 7.2 to 6.0. The physiological relevance of these findings is unknown, in part because the effect of changes in pHi has not been concomitantly determined. Because members of the K+-CCC family share several functional and structural characteristics, changes in pHi could also influence the operation of KCC2, KCC3, and KCC4. In such a case, titration of residues within the K+-CCCs could correspond to a key mechanism by which NKCC- and KCC-mediated - transport is autoregulated.
In this work, we demonstrate for the first time that the KCCs (including KCC1, KCC3, and KCC4) are able to transport probably at the K+ site. We also provide evidence that several members within the K+-CCC family (both within the NKCC group and the KCC group) are regulated by changes in pHi. A preliminary report has been presented (17).
MATERIALS AND METHODS
cDNA Construction and Vectors
The cDNAs used in this work (huNKCC1, rbNKCC2A, rbKCC1, rtKCC2, huKCC3A, and msKCC4) are the same as those described in previous studies (7, 15, 16, 37, 38, 42; hu, rb, rt, and ms are human, rabbit, rat, and mouse, respectively). The cDNA for huKCC3B was obtained by RT-PCR. Briefly, the entire coding region was amplified as three overlapping fragments from human kidney polyA-selected RNA using three pairs of primers1; here, the most 5′-primer was capped with the BamHI restriction site and the most 3′-primer with the XbaI site. Each fragment was then subcloned in the vector PCR2.1 (Invitrogen) using the BamHI-SphI (for the 5′-fragment), SphI-HindIII (middle fragment), and HindIII-XbaI (3′-fragment) sites. The resulting constructs were called yo1, yo2, and yo3, respectively, and the identity of their inserts was verified by automated sequencing.
Expression in Xenopus Laevis Oocytes
We used the cDNAs of huNKCC1, rbNKCC2A, rbKCC1, and rtKCC2 subcloned in the oocyte expression vector Pol1 (7, 15, 16) and that of huKCC3A and msKCC4 subcloned in the oocyte expression vector pGE-MHE (37, 38); rtKCC2 (originally in the vector pBF) was a gift from Dr. Eric Delpire (Vanderbilt University, Nashville, TN), and huKCC3A and msKCC4 in pGE-MHE were a gift from Dr. David B. Mount (Harvard University, Boston, MA). Both Pol1 and pGE-MHE contain (5′ to 3′) the T7 promoter, the X. laevis β-globin 5′-untranslated region, a multiple cloning site, the X. laevis β-globin 3′-untranslated region, a polyA tract, and a linearizing site. To obtain an expression construct containing huKCC3B, the yo1, yo2, and yo3 constructs described above were cut with BamHI-SphI, SphI-HindIII, and HindIII-XbaI, respectively, and their inserts were assembled into the BamHI-XbaI sites of Pol1.
cRNA was produced as previously described (7, 15, 16). Briefly, expression constructs were linearized with NheI and inserts were in vitro transcribed with T7 RNA polymerase using the mMESSAGE mMACHINE T7 kit (Ambion). Defolliculated stage V-VI oocytes were injected with 25 nl H2O or ≈5–25 ng cRNA diluted in 25 nl H2O. Functional expression was assessed 3–4 days after injection; oocytes were maintained at 18°C in Barth's medium (medium l; see Table 1)+125 μM furosemide.
All experiments were carried out at ≈22°C and, unless specified, pH of media was pH 7.2–7.4. When necessary, replacement of cations (Na+, Rb+) was done with N-methyl-d-glucamine and of anions (Cl-) with gluconate. For each experiment, H2O-injected oocytes were tested (along with CCC-expressing oocytes) to determine changes in flux due to endogenous cotransport activity. In our studies, furosemide (125–250 μM) was used to block KCC and bumetanide (250 μM) to block NKCC; at concentrations >125 μM, we previously observed that furosemide inhibits both types of K+-CCCs efficiently, but that bumetanide has an incomplete inhibitory effect (results not shown). Hence, apparent Ki values are somewhat lower than those reported by Mercado et al. (37) perhaps due to differences in flux protocols.
Unidirectional flux measurements were determined by using our routine “flux protocol” that includes the following four steps: 1) activation of CCCs by 1-h incubation in tracer-free hyposmolar low-Cl- medium (for KCCs) or tracer-free hyperosmolar medium (NKCCs); the composition of the activating media (medium a and medium b) is given in Table 1; 2) 45-min incubation in various influx media (see below) supplemented with ∼2 μCi/ml 86Rb+, 10 μM ouabain, and ±250 μM furosemide (KCC-expressing oocytes) or ±250 μM bumetanide (NKCC-expressing oocytes); 3) termination of influxes with washes in a basic medium (medium j; Table 1) containing 250 μM furosemide or bumetanide and 10 μM ouabain; and 4) solubilization of cells in 2% SDS and detection of 86Rb+ by liquid β-scintillation using the TopCount-NXT microplate counter (Packard).
Different measurements were obtained by varying step 2 of the flux protocol.
Basal 86Rb+ transport by heterologous CCCs. Here, two types of influx media (medium c5 mM and medium d5 mM; Table 1) were used. Medium c5 mM is a basic solution that contains 5 mM Rb+, whereas medium d5 mM contains 5 mM instead of 5 mM Rb+. These experiments were devised to determine whether KCCs are able to transport in the absence of K+ or Rb+ in the external medium.
Dependence of 86Rb+ transport on Rb+ concentration and . Sixteen types of influx media were employed; eight of these (medium c0.1-20 mM) are free and differ in Rb+ concentration ([Rb+]; from 0.1 to 20 mM), whereas the eight others (medium d0.1-20 mM) are Rb+ free and differ in  (also from 0.1 to 20 mM). These experiments were conducted to determine apparent affinities of various CCCs for both Rb+ and .
Effect of changes in pHi or pHo on CCC-mediated 86Rb+ influx. For these measurements, we used two types of media titrated at different pH with HCl or NaOH; one medium (media e) contains 60 mM Na+ acetate (medium epH x.x) and the other (media f) 60 mM Na+ gluconate (medium fpH x.x).2 Here, an intermediate 10-min equilibration step was added between steps 1 and 2 of the flux protocol using medium e or medium f without the tracer or the inhibitors.2 In some of the studies, Rb+ in medium e was replaced by (medium gpH x.x). These measurements were obtained to determine the pH sensitivity of various CCC isoforms and the topology of the pH effect (intra- vs. extracellular).
Effect of glycerol vs. sucrose on CCC-mediated 86Rb+ influx. These experiments were conducted as above, except that 60 mM Na+ acetate in medium epH x.x was replaced by 60 mM glycerol+30 mM Na+ gluconate (medium hpH x.x), and 30 mM Na+ gluconate in medium FpH x.x was replaced by 60 mM sucrose (medium ipH x.x). Here, the effect of a permeable osmole on CCC activity is compared with that of an impermeable osmole, independently of changes in pHi or pHo.
These studies were carried out in oocytes using conventional and pH-sensitive microelectrodes. Conventional electrodes were filled with a 1 M KCl solution; their resistances were 4–6 MΩ. The pH-sensitive electrodes were silanized in 0.5% dichlorophenylsilane (dissolved in ultrapure acetone) and baked overnight. Subsequently, the tips of the electrodes were filled with a pH-sensitive liquid ion exchanger (WPI, Sarasota, FL), and the remainder were backfilled with a calibration solution titrated at a pH of 7.4 (medium kpH 7.4; see Table 1). After each experiment, the pH-sensitive electrodes were calibrated in medium kpH 7.4 as well as in two other buffers (medium kpH 6.4 and medium kpH 6.9); for over 30 electrodes used, the slope of the voltage-pH curve averaged 47 ± 4 (SD) mV/pH unit.
In these studies, noninjected vs. CCC-injected oocytes were incubated first in medium a or medium b for at least 1 h. They were then transferred to an experiment chamber and impaled with a conventional electrode connected to a voltage-clamp amplifier (model OC-725B, Warner Instrument, Hamden, CT) and with a pH-sensitive electrode connected to a high-impedance electrometer (model FD-234, WPI). In each experiment, the starting bathing solution was medium d0 mM. After a 1- to 2-min stabilization period, this solution was replaced by medium d5 mM or medium d20 mM to which oocytes were exposed for an additional 1.5–3.0 min before being returned to medium d0 mM. Signals were digitized at 10 Hz to obtain pHi and recorded with a data-acquisition system (pClamp8 and clampfit8, Axon Instrument, Union City, CA). acidification rates (dpHi/dt) were measured by performing a linear fit to the pHi data once a steady-state acidification rate was observed (usually between t = 20 s and t = ≈80 s after the switch to the -containing solution).
Calculations and Statistics
Transport rates are expressed per hour as total counts in the sample (cpm) × nonradioactive Rb+ or (in some experiments) (nmol/μl) × normalization factor (in some experiments)/counts in the influx medium (cpm/μl). In each study, transport rates among 2–12 oocytes (usually, from 4 to 7 oocytes) were averaged; for studies in which the ion dependence of the 86Rb+ influx was assessed, these averaged rates were also normalized to the value measured at the highest ion concentration. Flux values (absolute or normalized) from 1 to 11 experiments (usually, from 3 to 4 experiments) were subsequently reaveraged to obtain the means ± SE. Affinity constant (Km) values were obtained by nonlinear least squares analysis using the Michaelis-Menten equation (1-binding-site model). For pHi measurements, dpH/dt values were obtained at 5 or 20 mM , and values among 4–11 oocytes were averaged. When appropriate, differences between groups of variables were analyzed by Student's two-tailed t-tests, and the null hypothesis was rejected for P values <0.05.
In oocytes, each of the wild-type CCCs (rbNKCC2A, rbKCC1, huKCC3A, and msKCC4) induces 86Rb+ influx (Fig. 1A, filled bars) well above that of the endogenous NKCC (27-, 4-, 11-, and 6-fold, respectively); after correction for specific activity, 86Rb+ influx is 8.3, 1.3, 3.4, and 1.7 nmol · oocyte-1 · h-1, respectively, compared with 0.3 nmol · oocyte-1 · h-1 for controls. In these experiments, flux assays were performed after incubation of oocytes in hyposmolar medium to activate KCCs (17, 20, 35, 38) or hyperosmolar medium to activate NKCC2 (15, 16, 25). The assays were also performed in the presence of 10 μM ouabain to block the Na+ pump. Over 90% of the 86Rb+ influx measured under such conditions was bumetanide sensitive (results not shown). Even if all CCC-expressing oocytes are found to have significantly higher transport rates than controls3, it is clear from Fig. 1A that KCCs exhibit lower rates compared with NKCC2. These differences in activity could be due to differences in cell-surface expression or maximal transport capacity; alternatively, the procedure used to activate KCCs may have been suboptimal compared with NKCCs.
To determine whether KCCs can transport (as shown previously for NKCCs; see Refs. 13, 23, 24, 30, 31, 50, 51), the aforementioned measurements were repeated using a similar flux medium, except that Rb+ was replaced by (Fig. 1A, open bars). In fact, this medium still contains 3 μM Rb+ from the addition of 86Rb+; at this [Rb+], however, neither NKCCs nor KCCs can support measurable ion transport rates because the Km for Rb+ (and K+) is in the millimolar range (see below). As illustrated in Fig. 1A for the three KCCs studied here (and for rbNKCC2A), 86Rb+ influx in the -containing medium is not significantly different from that measured in the Rb+-containing medium. These results suggest that all K+-CCCs are able to support ion transport by using as a surrogate of Rb+ (or of K+).
More direct proof that is actually transported across the membrane, and does not merely interfere with the normal operation of KCCs (and NKCCs) or with the binding of 86Rb+ at the K+ site, can be obtained through pHi measurements. Indeed, an increase in influx due to an increase in CCC activity at the oocyte surface should lead to acidification of the cytosol as accumulates in the cell and dissociates into NH3+H+. Importantly, the concentration of free NH3 does not increase in the cytosol of noninjected oocytes after NH4Cl loading (4, 5, 11). Figure 2 shows typical time courses of pHi changes obtained at two different  (medium d5 mM or medium d20 mM) for noninjected oocytes (Fig. 2A) and for rbNKCC2A-, rbKCC1-, huKCC3A-, and msKCC4-expressing oocytes (Fig. 2, B–E), and Table 2 presents averaged dpHi/dt values derived from several of these experiments.
As demonstrated previously (4, 11), the cytosol of noninjected oocytes becomes weakly acidic at 5 mM (Table 2). In our studies, changes in dpHi/dt values are statistically significant at 20 mM but are still relatively small (Table 2, Fig. 2A). It is noteworthy that dpHi/dt for noninjected oocytes are not influenced by the type of media used (medium a vs. medium b) to activate CCCs (Table 2).
For oocytes expressing rbNKCC2A or huNKCC1, dpHi/dt are much larger compared with noninjected oocytes (Fig. 2B and/or Table 2). These results confirm that NKCCs are able to transport (see Refs. 13, 23, 24, 30, 31, 50, 51), most likely by using the K+ site. Interestingly, the rates are similar in amplitude at 5 vs. 20 mM , suggesting that the () values for NKCCs examined in this study are <5 mM.
For KCC-expressing oocytes, substantial decreases in pHi are also seen. In the presence of 20 mM , e.g., dpHi/dt for KCC1, huKCC3A, and KCC4 (Fig. 2, C–E, Table 2) are more than twice those for noninjected oocytes (Fig. 2A, Table 2); noticeably for KCC4, rates are also comparable at 5 vs. 20 mM . Hence, similar to NKCCs, KCCs can mediate transport at appreciable rates. Based on these results, () values for different CCCs are probably as follows: rbKCC1< huNKCC3A < msKCC4 ≤ rbNKCC2A.
In principle, the rate of entry into cells is equal to measured dpHi/dt × intracellular buffering power (33). Thus the interpretation of measurements shown in Fig. 2 could be affected by systematic changes in initial pHi values. Table 2 shows that for several conditions tested, however, initial pHi values are not statistically different from, or are relatively close to, one another. Based on previous studies, in addition, the intracellular buffering power of X. laevis oocytes was shown to be similar at pHi values ranging between 6.85 and 7.45 (10) and was estimated at 23.8 mM/pH unit when the average pHi was 7.69 (48). Taken together, these results indicate that differences in dpHi/dt values reported here are unlikely to be accounted for (to an appreciable extent) by pH-dependent changes in buffering power.
Rb+ Influx as a Function of [Rb+] and 
The principle of competitive inhibition (e.g., influx rates measured at fixed, or radioactive [86Rb+] but various, or cold [85Rb+]) is often used to determine apparent ion affinities of a transport protein (fluxes are then corrected for specific activity). Km(Rb+) values shown in Fig. 3 were obtained using this principle. To determine affinities of K+-CCCs for and facilitate comparisons between kinetic parameters, 86Rb+ was also employed as the fixed substrate when  was varied, and 86Rb+ influx as a measure of transport. Here, correction for specific activity is also made assuming a simple model in which analogous substrates ( and Rb+ have the same ionic radius once hydrated) (30, 32) compete for the same translocation site. Previous studies for NKCCs and the results shown in Fig. 2 suggest these assumptions are correct.
As illustrated in Fig. 3, the dependence of 86Rb+ influx on [Rb+] for NKCC2 and for the KCCs obeys a single-ligand-binding-site model (Fig. 3, A–D). Interestingly, Km(Rb+) values derived from these measurements are similar among KCCs, varying between 12 and 17 mM, and are close to those reported by other groups (37, 45). The dependence of 86Rb+ influx on  for each CCC examined in this work is also best fit with a one-binding-site model (Fig. 3, E–H); noticeably, the influx-concentration relationships translate into () values that are very close to those for Km(Rb+) and that are in agreement with the order estimated from acidification rate measurements. Taken together, the results presented here suggest that is a very good surrogate of K+ or Rb+ for all carriers belonging to the K+-CCC family.
pHi and CCC Activity
Changes in pH can affect the operation of NKCCs and of KCC1; hence, they could affect that of the other KCCs, which are also K+ dependent and similar to one another. The mechanisms of the interaction between pH and CCC-mediated transport are ill defined. If they involve titration of intracellular residues, the implication could be that K+-CCC-mediated transport is regulated by pHi and that Km values derived from the measurements shown above represent a possibly inaccurate estimate of Km values derived from more direct measurements, e.g., true dissociation constants (Kd). The following experiments were devised to examine the effect of changes in pH on K+-CCCs and to determine the sideness of the effect. Results are shown in Fig. 4 and discussed below.
When oocytes are bathed in a medium containing 60 mM acetate, a change in pHo will lead to a proportional change in pHi (slope ∼0.64) so that pHi will be relatively close to (but smaller than) pHo (8, 9, 48).2 In such a medium, interestingly, the activity of all K+-CCCs examined is shown to vary as a function of pH (see Fig.4, A–E). For example, huKCC3B (results not shown), rbKCC1, and huKCC3A exhibit low activity at pH <7.0 or >7.5, whereas rtKCC2 and rbNKCC2A exhibit low activity only at pH <7.5, and msKCC4 only at pH >7.5. These studies show for the first time that the K+-CCC isoforms are all sensitive to changes in pH and that they are differentially affected by such changes.
When the influx medium contains gluconate instead of acetate, changes in pHo are not accompanied by changes in pHi (6).2 In such a medium (Fig. 4, F–J), the effect of changes in pH on CCCs is much smaller (except for rbKCC1, which is still inhibited at higher pHo). These results indicate that, to a large extent, intracellular domains within K+-CCCs mediate the pH effect described in our study. They also suggest that cellular accumulation of can lead to a change in K+-CCC activity as the ion dissociates into H+ and NH3. Accordingly, the Km values reported here could be underestimated (rbNKCC2A, rbKCC1, rtKCC2, huKCC3A) or overestimated (msKCC4).
The formula [Vn at pHi+o x.x] - [Vn at pHo x.x - Vn at pHo 7.2], where Vn = 86Rb+ influx normalized to the highest flux values and x.x = pH values, can be used to estimate CCC activity resulting from changes in pHi alone. By using this formula, the effect of changing pHo from 7.2 to another value is subtracted, assuming that the relationship between pHo and CCC activity is independent of that between pHi and CCC activity, and assuming that estimated pHi is close to measured pHi. Results of these calculations (Fig. 4, K–O) show that for four of the K+-CCCs examined, the shape of the curve is similar to those in corresponding top panels (Fig. 4, A–E); for rbKCC1, the curve is only changed at higher pHs.
Because cell membranes are more permeable to acetate than to gluconate, medium e could have reduced effective osmolality compared with medium f, leading to overestimated 86Rb+ fluxes. Hence, values reported in Fig. 4, K–O, could also be overestimated. However, because osmolalities of medium e or f do not change at different pH values, the look of the curves in Fig. 4, I–L, are probably not determined by differences in effective osmolality. Similar experiments wherein the permeant anion is replaced by glycerol, and a fraction of the impermeant anion by sucrose (Table 1, media h and i), confirm that this is the case. In such experiments, the effect of changes in pHo (6.0 vs. 7.5) on two different CCCs was found to be the same in either medium. Indeed, 86Rb+ influx rates were all between 0.5 and 0.7 ± 0.1 nmol · oocyte-1 · h-1 (n = 7–8/condition) and were not statistically different from one another.
Effect of pHi on Estimated Rb+ vs. Transport
Although both Rb+ and share the same ionic radii once hydrated (as mentioned), and although estimated () and Km(Rb+) are quite similar for any given CCCs, the pH effect reported in this work indicates that the relationship between () and () may differ from that between Km(Rb+) and Kd(Rb+). Accordingly, changes in pHi could have altered the preference for one substrate vs. the other. To test this possibility, we measured CCC activity in an Na+ acetate influx solution (pH 6.5) containing either 5 mM Rb+ (medium epH 6.5) or 5 mM (medium gpH 6.5). The results of these studies are illustrated in Fig. 1B. They show that for any given CCCs, 86Rb+ influx is the same whether measurements are obtained in medium e or f. Hence, changes in pHi do not appear to alter substrate specificity at the K+ site (at least when these substrates are used at a concentration of 5 mM).
In this study, we used an expression system to determine whether KCCs, similar to NKCCs (24, 30, 31, 50, 51), are directly involved in transport. Based on various studies, we were able to conclude 1) that the K+ site of K+-CCCs can interact with and 2) that these carriers can promote translocation. The first conclusion is supported by the findings that KCC-mediated 86Rb+ influx decreases in the presence of and that estimated () is similar to Km(Rb+) for any given CCC. The second conclusion is supported by the finding that the cytosol of KCC-expressing oocytes becomes acidic at a rate that is significantly faster than that of noninjected oocytes after incubations in -containing media (Fig. 2).
In mammals, the physiological importance of NKCC-mediated transport has been documented previously (2, 12, 19, 50, 51). Results presented here provide evidence that transport by KCCs may also be relevant physiologically. Indeed, estimated () values for these carriers are in the millimolar range, only ∼0.5–1 order of magnitude lower than those reported for NKCCs (see Fig. 3 and Refs. 26, 31, 51). It is noteworthy that these values are also close to  found in various tissues (13, 23, 44, 51). Hence, K+-CCCs may be important accessory pathways for transport in a number of cell types.
An interesting issue regarding K+-CCC-mediated transport is “net direction of flux.” In theory, this direction should depend on the K+-CCCs involved, the intra- and extracellular concentration of transported ions (,,,), and on pHi and pHo, which will also influence and . Because the intracellular-to-extracellular concentration gradient of is much less than that of K+ in most cell types and because pHi is usually close to pHo, the direction of fluxes could differ from that of Rb+ fluxes, especially when is high or pHi is low. During our transport studies, e.g., the direction of 86Rb+ flux in KCC-expressing oocytes is inward at > 1 mM, pHo ≈7.3, and pHi <7.8.
Other cell types that may display inwardly directed transport via KCCs (and also via NKCC1) include the periportal hepatocytes. Indeed, the concentration of in portal blood is usually very high (up to 20-fold higher than that in systemic blood) because urea-derived products are absorbed in large quantities from the gut (44, 49). Hence, basolateral K+-CCCs in hepatocytes could contribute to uptake by the liver parenchyma, and via the Kreb-Henseleit-urea cycle, promote metabolic elimination of this product, which is toxic to the brain (40, 44) and several other tissues. Members within the CCC family would then collaborate with other systems (enzymes and carrier molecules) to maintain / in systemic blood (and in the cerebrospinal fluid) at very low levels.
Based on kinetic measurements reported in this work (and in other studies; see Refs. 13, 23, 50, 51), KCC- and NKCC1-mediated transport in various renal cells is probably inward (as in hepatocytes) because  is increased throughout the interstitium (23, 51), reaching 15 mM in certain regions (46). This concentration is close to, or even above, the estimated () values for various K+-CCCs (Fig. 3), indicating, in addition, that intracellular accumulation of across basolaterally disposed renal K+-CCCs may be physiologically important. For example, uptake in α-intercalated cells, which were recently shown to express NKCC1 and KCC4 (3, 22), would contribute to distal acidification by promoting secondary NH3 secretion. On the other hand, KCC-mediated uptake in the thick ascending limb of Henle's loop (TAL), a nephron segment that expresses several KCCs (39), would serve an undetermined role as it would in fact limit net reabsorption (at least when pHi is 7.0 and pHo 7.4; see below).
The presence of pathways in the CD may seem difficult to reconcile with an observation by Flessner et al. (14) that transcellular movement of across this nephron segment is low and with the fact that formed in the lumen (from buffering of secreted H+ by NH3) must not be reabsorbed for net H+ excretion to occur. However, the observation above does not rule out the possibility that only the apical membrane of the CD has low permeability; in such a case, appreciable back-leak from the lumen would still be prevented. The possibility entertained herein needs to be confirmed through additional microperfusion studies, as data obtained in heterologous systems may not compare with those obtained in more complex systems.
Previous studies have demonstrated that the function of NKCC1, NKCC2, and KCC1 decreases when pHo is <7.0 (Isenring and Forbush, unpublished observations; see also Refs. 26 and 34). The results presented in this study show that this effect probably results to a large extent from secondary changes in pHi and that all KCCs are affected by such changes. Interestingly, KCC4 was shown to be more active at low pHi, whereas all other K+-CCCs were less active. This finding suggests that certain isoforms may be specialized to operate in acidic environments, e.g., in α-intercalated cells.
The interdependence observed between pH and transport, for example, our finding that K+-CCCs are pH sensitive (Fig. 4) and that they can also affect pHi by regulating transport (Fig. 2), points to the possibility that these transporters are involved in regulation of intracellular [H+]; in this regard, changes in pHi could correspond to an important signaling intermediately involved in the autoregulation of K+-CCC-mediated transport. Here, remarkably, we have shown that low pHi led to the inhibition of NKCC2 (a carrier that promotes influx) and activation of KCC4 (a carrier that promotes efflux below a certain pHi). In certain cell types, those of the TAL for instance, pHi-mediated regulation of K+-CCCs may be important in preventing excessive cellular acidification resulting from the transepithelial movement of .
Because several K+-CCCs exhibit similar and Rb+ affinities [as shown in Fig. 3, estimated () and Km(Rb+) for rbNKCC2A are 1.7 and 3.7 mM], we can predict () values for other K+-CCCs not included in the present analysis (rbNKCC2F and B) but for which Km(Rb+) values are already known (15, 21, 43). Based on previous measurements, hence, () values for these “F” and “B” variants should be ∼8 and ∼3 mM, respectively; recent studies by us have shown that this was actually the case (results not shown). Because the NKCC2s are distributed differentially along the TAL (28, 41), i.e., F is in the inner medulla, A in the outer medulla, and B in the cortex, the affinity of these carriers for should increase progressively along the TAL, leading to optimized transport throughout the nephron segment. In the cortex, efficient transport may be important for NH3 recycling by the proximal tubule and NH3 secretion along the proximal CD.
The Km(Rb+) values reported here for KCC1, KCC3, and KCC4 were found to be relatively similar to one another (12, 17, and 12 mM, respectively) and similar to those reported for low-affinity NKCC2 splice variants (15, 21, 43). These results are interesting with respect to structure-function relationships. Indeed, we have shown that NKCC2 splice variants with the sequence I, L, T at positions 13, 16, and 18 of the second transmembrane α-helix (I13, L16, and T18) had higher cation affinities than splice variants with the sequence L13, I16, M18 (see Ref. 14, Fig. 3, Table 3). For all of the KCCs, remarkably, the sequence is also L13, I16, M18, similar to that of the low-affinity NKCC2 splice variants. Thus, for both the NKCCs and KCCs, residues at positions 13, 16, and 18 of the second transmembrane domain may play an important role in ion transport by modulating ion affinity within the translocation pocket.
In conclusion, we have shown that the KCCs, similar to the NKCCs, are able to transport probably at the K+ site and that they are sensitive to changes in pHi. In certain cell types, the dependence of KCC activity on pHi may be important for regulation of K+-CCC-mediated transport. In other cell types, e.g., the CD, KCC-mediated transport may play an important role in distal renal acidification.
The authors thank Luc Caron and Valerie Montminy for technical assistance.
This work was supported by grants from the Kidney Foundation of Canada and the Canadian Institute of Health and Research (MT-15405). P. Isenring is a Canadian Institutes of Health Research Clinician Scientist II.
↵1 The oligonucleotides used in the PCR reactions were 1) GGGATCCCAGAAAGAGCAAAGTATTATCTAAC and CTTGATGAAAGGGTGCGG (5′-fragment), where bold characters indicate the BamHI site; 2) TGCTCCTCCACACTTCCC and CACCAGCTGGCAGAATCC (middle fragment); and 3) GGTCTCACTATTGTGGGC and GTCTAGAGGCACTTCCATGGAGGACGTAGGCC (3′-fragment), where bold characters indicate the XbaI site.
↵2 When oocytes are bathed in an influx medium that contains 60 mM Na+ acetate, changes in pHo are accompanied by predictable changes in pHi because H+ acetate diffuses freely across plasma membranes (8, 9, 48). Under such circumstances, steady-state pHi values are observed after a 20-min incubation. When, on the other hand, the influx medium contains 60 mM Na+ gluconate, changes in pHo are not accompanied by appreciable changes in pHi because the oocyte's membrane has low permeability to H+ or Cl- (6). In the experiments described here, incubation in Na+ acetate or Na+ gluconate was limited to 55 min: 10 min in tracer-free medium followed by 45 min in tracer-supplemented medium. Under such circumstances, pHi values are probably stable for most of the period during which unidirectional fluxes are monitored. Longer incubations were avoided to limit potential damage to the oocytes in the lower pH solutions.
↵3 In these studies (Fig. 1), flux measurements for H2O controls were obtained after preincubation in hyposmolor medium only. When H2O controls are preincubated in hyperosmolar medium, a 30–40% increase in influx rates is generally observed (unpublished observations; see also Ref. 16). Hence, the difference shown here between NKCC-injected oocytes, which were preincubated in hyperosmolar medium, and H2O-injected oocytes is probably overestimated. However, because absolute flux rates in H2O controls are very small, subtracted flux rates (CCC-injected minus H2O-injected) are still significantly lower in the KCC group.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2003 the American Physiological Society