In rabbit proximal tubules, a basolateral ATP- and taurine-sensitive K+ channel (KATP) was shown to be involved in the regulation of the basolateral K+ conductance as a function of the rate of apical Na+ entry. To establish the molecular identity of this channel, we used degenerated primers to look for cDNA transcripts for an inwardly rectifying K+ channel (Kir6.1 and Kir6.2) and sulfonylurea receptors (SUR1, SUR2A, and SUR2B) in a cDNA library obtained from rabbit proximal tubules. PCR products were found only for Kir6.1, SUR2A, and SUR2B. Expression of Kir6.1 in Xenopusoocytes generated an additional K+ current that was found to be sensitive to external barium and intracellular taurine and to changes in intracellular ATP concentrations. To study the specificity of the taurine sensitivity, intracellular taurine was tested on several members of the Kir family expressed in Xenopus oocytes. K+ currents induced by Kir1.1A, Kir2.1, Kir3.2, Kir4.1, or Kir5.1 were insensitive to taurine, but all tested combinations of Kir6.x with or without the SUR subunit were significantly inhibited by taurine. This study suggests that the taurine-sensitive KATP channel of rabbit proximal tubules is formed by a combination of Kir6.1 plus SUR2A and/or SUR2B.
- inwardly rectifying potassium channel
- sulfonylurea receptor
in the kidney, the greatest part of volume reabsorption is performed by the proximal convoluted tubule (PCT), which generates an intense transepithelial NaCl transport. As this transepithelial ionic flux may vary as a function of the availability of Na+-cotransported substrates in the lumen, regulatory mechanisms exist to minimize potentially dangerous changes in cell volume and ionic composition. It was previously shown that the basolateral K+ conductance is adjusted as a function of transepithelial Na+ transport (26) and that an ATP-sensitive K+ channel is responsible for this regulation (7, 17, 31, 43). Pump-dependent changes in local ATP concentration were suggested to constitute the likely mediator of this regulatory mechanism (8,38), but other mediators could also be involved. For example, the KATP channel of the rabbit PCT was recently demonstrated to be taurine sensitive (31). As intracellular taurine has been implicated in volume regulation as an exportable osmolyte, it is a very interesting candidate for the regulation of the coupling between the basolateral K+conductance and the transepithelial Na+ transport that is known to affect cell volume (27).
Taurine sensitivity of K+ channels is a relatively new finding that has been thus far observed for KATP channels in ventricular myocytes (16, 37), proximal tubules (31), Xenopus oocytes (10), and skeletal muscle (42). However, it had not been established whether taurine sensitivity is a specific property of KATPchannels or could be extended to other members of the inwardly rectifying K+ channel family.
Originally discovered in cardiac muscle (29), KATP channels have been subsequently identified in a variety of tissues. Since 1995 it has been recognized that KATP channels are heteromultimeric proteins formed from the combination of two types of proteins: pore-forming subunits (i.e., the inwardly rectifying K+ channel, Kir 6.1 or Kir 6.2) and sulfonylurea receptors (SUR1, SUR2A, or SUR2B) (1, 19,20, 23, 36). More recently, the combination of Kir1.1A and the cystic fibrosis transmembrane conductance regulator (CFTR, like SUR, a member of the ATP-binding cassette protein family) was also shown to produce a KATP channel activity (35). A preliminary report describes a ubiquitously expressed KATP channel (uKATP-1 or Kir6.1 cloned from rat pancreatic islets; see Ref. 22) that was localized at the basolateral membrane of rat proximal tubule (5). This subunit has been subsequently cloned in human lungs (21) and mouse brains (45). However, it is still unknown whether the Kir6.1 subunit, which develops a channel with a 70-pS conductance in Xenopus oocytes (22), is the one responsible for the 50-pS channel in rabbit PCT. Attempts to determine the nature of the SUR subunits in mice and rats show that SUR1 and SUR2A mRNAs are generally absent in the kidney (9, 19, 23). On the other hand, SUR2B mRNAs are expressed in kidney as well as in all other tissues examined (heart, eye, urinary bladder, skeletal muscle, brain, lung, liver, pancreas, spleen, stomach, small intestine, colon, uterus, ovary, and fat tissue) (23). In kidney, high staining was found with SUR2B antisera in distal nephron segments, whereas low, diffuse staining was found in proximal tubules (9).
KATP channels in native tissues yield a wide variety of characteristics (such as conductance and/or sensitivity to ATP or sulfonylurea), and it is not clear whether the two Kir6.x and three SUR subunits already identified could account for such a wide variety of native KATP channels. In the kidney in general and in PCT in particular, the molecular identity and precise composition of the KATP channel remain unsettled.
As a first step in the molecular identification of the different subunits that compose the basolateral PCT KATP channel, we searched for the presence of transcripts that correspond to the most likely composition (Kir6.x plus SURx). We then expressed the corresponding proteins in Xenopus oocytes to compare important functional characteristics with what is currently known from rabbit PCT. Finally, to define whether taurine sensitivity is a specific characteristic of the KATP channel, the effect of taurine was tested on several members of the Kir family after expression in oocytes.
MATERIALS AND METHODS
A suspension of cortical renal tubules from New Zealand White rabbits was prepared as described previously (24). The tubule fragments were separated in a Percoll gradient using the method of Vinay et al. (44). The F4 layer, 98% of which represents proximal tubules, was then removed from the Percoll gradient and washed three times by centrifugation. Total RNA was extracted from the PCT suspension using TRIzol reagent (Life Technologies) according to the manufacturer's instructions. Double-stranded cDNA was obtained using the Smart PCR cDNA synthesis kit (Clontech, Palo Alto, CA), which presents the advantage of a significant enrichment in full-length cDNAs.
Cloning of rabbit PCT Kir6.x.
Two degenerated PCR primers corresponding to the first 28 nucleotides (the 5′ end) and the last 22 nucleotides (the 3′ end) of human, mouse, and rat Kir6.1 (respectively, 5′-GATGYTGGCCAGRAAGAGYATCACCCGGA-3′ and 5′-TCYTGCYSTCATGATTCYGATG-3′) were used to amplify a 1,286-bp product from PCT cDNA corresponding to the full-length rabbit renal Kir6.1. The PCR product was then ligated into the pGEMT-Easy vector (Promega, Madison, WI).
Two primers directly designed from the 5′ and 3′ ends (respectively, 5′-GGCATGCTGTCCCGAAAGGGG-3′ and 5′-TCAGGACAAGGAATCTGGAGAGATGC-3′) of the cloned rabbit heart Kir6.2 (GenBank accession no. AF006262) were used in our attempt to isolate the renal Kir6.2 from the PCT cDNA.
Cloning of rabbit PCT SURx.
Four degenerated PCR primers corresponding to the 5′ end (5′-CATGCCYYTGGCCTTCTGCGG-3′ or 5′-CATGCCYYTGGCCTTCTGCGGYASC-3′) and the 3′ end (5′-TCAYTTGTCYGCRCGGAC-3′ or 5′-TCAYTTGTCYGCRCGGACRAAGG-3′) of human, rat, and hamster SUR1 were used to isolate the rabbit renal SUR1 from the PCT cDNA.
Two degenerated PCR primers corresponding to the 5′ and 3′ ends (respectively, 5′-ATGAGCCTTTCHTTYTGTGG-3′ and 5′-TCACATGTCTGCRCGRAC-3′) of rat, mouse, and human SUR2B were used to amplify a 4.8-kb product. It could correspond to the full-length rabbit renal SUR2A or SUR2B, taking into account that the 5′ ends of SUR2A and SUR2B are identical and that the selected sequence of the PCR primer corresponding to the 3′ end of SUR2B is also present in the 3′ untranslated region of SUR2A. In fact, the nucleotide sequences of SUR2A and SUR2B cloned in other species are identical apart from an insertion in the 3′ end of the coding region for SUR2A, which generates a C-terminus different from SUR2B. The 4.8-kb PCR product was purified with GFX PCR DNA and the Gel Band purification kit (Amersham Pharmacia Biotech Life Science, Cleveland, OH) and ligated into pGEMT-Easy vector. The ligated product was used to transform competent X1–2 Blue E. colibacteria (Stratagene, La Jolla, CA), which were plated onto petri dishes. Plasmids of five different colonies were purified. To determine which colony corresponded to SUR2A or SUR2B, a new pair of PCR primers was designed to frame the 3′ insert region of SUR2A and therefore amplify products of different sizes from either SUR2A or SUR2B. It was expected that the 5′ and 3′ PCR primer pair (respectively, 5′-CCG TGG TCA CCA TAG CTC-3′ and 5′-TCA CAT GTC TGC ACG AAC-3′) would amplify a 336-bp product for SUR2A and a 162-bp product for SUR2B.
The nucleotide sequence of the isolated clones was determined by a commercial sequencing service (BioS&T, Lachine, PQ) and confirmed twice. For expression in oocytes, 10 μg of pGEMT-Easy vector that contained the full-length cDNA encoding Kir6.1, SUR2A, or SUR2B were linearized with NdeI (for Kir6.1) or SpeI (for SUR2A or SUR2B) and transcribed in vitro using the mMESSAGE mMACHINE T7 Transcription Kit (Ambion, Austin, TX). Each transcript (12–30 ng) was injected into the oocytes.
pSPORT (10 μg) containing cDNA that encodes rat Kir1.1A (or ROMK1, GenBank accession no. AF081365; a gift of Dr. S. J. White), was linearized with NotI and transcribed in vitro using the mMESSAGE mMACHINE T7 transcription kit. RNA (5 ng) was injected into the oocytes.
Human Kir 2.1 cDNA (1 ng; hIRK; given by Dr. R. Sauvé) and rGFP (1 ng; in pMT21 vectors) were injected intranuclearly.
Human Kir3.2 (or pBlue-HGIRK2, GenBank accession no. U52153) and human Kir4.1 (or pBlue-HGIRK15–2, GenBank accession no. HSU52155, gifts of Dr. H. Van Tol) were linearized with BamHI. Each clone was transcribed with the mMESSAGE mMACHINE T7 transcription kit, and each transcript (16 ng) was injected into the oocytes.
Rat Kir5.1 (or BIR9, GenBank accession no. AF249676; provided by Dr. Larry Salkoff) was linearized with MluI and transcribed in vitro with the mMESSAGE mMACHINE SP6 transcription kit (Ambion). RNA (16 ng) was injected into the oocytes.
Four different clones were given to us by Dr. Seino: rat Kir6.1 (linearized with HindIII), mouse Kir6.2 (linearized withXbaI), rat SUR1 (linearized with XbaI), and human SUR2A (linearized with ClaI). Each linearized clone was transcribed in vitro with the mMESSAGE mMACHINE T7 transcription kit. Oocytes were injected with 12.5 ng of each transcribed RNA.
Oocyte Preparation and Injection
Stage V and VI oocytes were removed from Xenopus laevisfrogs anesthetized with 3-aminobenzoic acid ethyl ester and were placed into Barth's solution [(in mM) 88 NaCl, 3 KCl, 0.82 MgSO4, 0.41 CaCl2, 0.33 Ca(NO3)2, and 5 HEPES; pH 7.60]. The follicular layer was removed by incubation in a Ca2+-free Barth's solution containing 1.6 mg/ml collagenase (Sigma, St. Louis, MO) as previously described (11). Defoliculated oocytes were stored at 18°C in Barth's solution supplemented with 5% horse serum, 2.5 mM sodium pyruvate, 100 U/ml penicillin, and 0.1 mg/ml streptomycin for 48 h. Healthy oocytes were then selected and injected either into the cytoplasm with 46 nl of water (±mRNA) or into the nucleus with 4.6 nl of water containing recombinant expression vector coding for green fluorescent protein with or without recombinant pMT21 vector containing the cDNA insert. The oocytes were incubated in antibiotic Barth's solution for 3–8 days before the experiments.
Two-Microelectrode Voltage-Clamp Technique
Oocyte currents were measured with the two-microelectrode voltage-clamp technique using a commercial amplifier (oocyte clamp model OC-725, Warner Instrument, Hamden, CT) as described previously (12). Current-voltage curves were obtained from 11 pulses (500-ms duration) separated by 600-ms periods at the resting potential of −40 mV. The voltage range covered was from −165 to +85 mV. Three different bath solutions were used: 1) a modified Barth's solution (M-Barth) that contained (in mM) 40 NaCl, 3 KCl, 50 n-methyl-d-glucamine-Cl (NMDG-Cl), 0.82 MgCl2, 0.74 CaCl2, and 5 HEPES/Tris, at pH 7.60; 2) a K+-enriched Barth's solution (K-Barth) that contained (in mM) 40 NaCl, 20 KCl, 33 NMDG-Cl, 0.82 MgCl2, 0.74 CaCl2, and 5 HEPES/Tris, at pH 7.60; and 3) a high-K+ Barth's solution (HK-Barth) that contained (in mM) 93 KCl, 0.82 MgCl2, 0.74 CaCl2, and 5 HEPES/Tris, at pH 7.60. Current and voltage traces were analyzed by averaging the signals in a window of 25 ms positioned after the decay of all capacitative transients.
Statistical analysis was performed using paired or unpairedt-tests as appropriate. Average values are given as means ±SE, and n represents the number of oocytes that were obtained from at least three different animals.
Cloning of Rabbit PCT Kir6.1
As there has been no reported sequence for a Kir6.x of renal origin, we first looked for the presence of transcripts corresponding to the known Kir6.x subunits, i.e., Kir6.1 or Kir6.2, in rabbit proximal tubule cDNA. Degenerated primers designed from sequences of Kir6.1 cloned in human lung, mouse brain, and rat pancreatic cells were used to amplify a 1,286-bp product (Fig.1 A). This clone, which corresponds to the full-length renal rabbit Kir6.1 (GenBank accession no. AF417509), was totally sequenced; it was predicted to code for a 424-amino-acid polypeptide where alignment through GenBank scored 95–97% identity with rat, mouse, and human Kir6.1 (Fig.1 B) and 62% identity with the rabbit Kir6.2 previously cloned from heart. As shown in Fig. 1 A, a second PCR product of lower molecular weight was also found to be amplified by the same primers but was not further studied. When primers designed from the rabbit cardiac Kir6.2 were used, no signal could be detected by PCR from our rabbit PCT cDNA (Fig. 1 A).
Cloning of Rabbit PCT SUR2A and SUR2B
A similar homology strategy was used to isolate the SUR subunits from our PCT cDNA. No product could be detected with two different pairs of degenerate PCR primers corresponding to human, rat, or hamster SUR1 (Fig. 2 A). On the other hand, two degenerated PCR primers corresponding to the 5′ and 3′ ends of mouse, rat, and human SUR2B were effective in amplifying a PCR product (Fig. 2 B) that corresponded to the size expected (about 4.8 kb) for the full-length renal rabbit SUR2A or SUR2B. A new pair of PCR primers designed to frame the 3′ end insert region of SUR2A was used to amplify (from plasmids isolated from colonies of bacteria transformed with the previously obtained 4.8-kb product) new PCR products of different size from either SUR2A and SUR2B. Among the five different plasmids isolated from five colonies of bacteria, one (no. 2 clone) yielded a 336-bp product (the expected size of a SUR2A fragment) and the four others (nos. 1, 3, 4, and 5 clones) yielded a 162-bp product (the expected size of a SUR2B fragment; Fig. 2 C; for details, see materials and methods). The no. 2 clone was predicted to code for a full-length 1,549-amino acid protein that was 100% identical to the rabbit cardiac SUR2A (GenBank accession no.AF417510), and the no. 1 clone was predicted to correspond to its splice variant SUR2B (GenBank accession no. AF417511). As expected, sequences of SUR2A and SUR2B only differ in the 3′ end, which is represented in Fig. 2 D. Alignment through GenBank has shown that the rabbit SUR2B from PCT was 96–98% identical to the rat, mouse, and human SUR2B, including the rat kidney SUR2B homologue (41).
Expression of Rabbit Kir6.1 in Xenopus Oocytes
To characterize the properties of the cloned rabbit Kir6.1 and to compare these properties with the known characteristics of the native PCT KATP channel, we first expressed Kir6.1 RNA in the absence of an exogenous SUR subunit into Xenopus laevisoocytes. We compared whole-cell currents of control oocytes with Kir6.1-injected oocytes (at 5–8 days after injection). Control oocytes developed a current of 166 ± 11 nA at +85 mV and −71 ± 13 nA at −110 mV in M-Barth (n = 7; Fig.3). Conductance was higher in Kir6.1-injected oocytes: 441 ± 39 nA at +85mV and −192 ± 23 nA at −110 mV in M-Barth (n = 23). When oocytes expressed Kir6.1 RNA, the reversal potential measured in control conditions (M-Barth) was slightly depolarized (−31 ± 1 mV) compared with control oocytes (−40 ± 5 mV). In Kir6.1-expressing oocytes, as extracellular K+ was increased (from 3 mM in M-Barth to 20 mM in K-Barth and 93 mM in HK-Barth), the reversal potentials shifted to more-positive values (−31 ± 1 mV in M-Barth, −21 ± 2 mV in K-Barth, and −3 ± 0.5 mV in HK-Barth; see Fig. 3). These values indicate a partial selectivity of the whole-cell current for K+ (the partial conductance of potassium, GK/Gtot= 0.2 and 0.5 when calculated from the reversal potential, respectively, in M-Barth vs. K-Barth and K-Barth vs. HK-Barth). This estimation is probably an underevaluation of true selectivity considering that changes in local intracellular K+ concentration will follow changes in external K+ concentration and affect the selectivity calculation. In addition, inward currents activated by external addition of K+ (90 mM K+ increase from M-Barth to HK-Barth) were significantly higher in Kir6.1-expressing oocytes (285 ± 20 nA at −110 mV; n = 23;P < 0.001; see Fig. 3) than in control oocytes (101 ± 9 nA at −110 mV; n = 7; see Fig. 3).
To confirm the presence of a K+ current, the inhibitory effect of barium was tested. In the presence of 5 mM external barium, the currents of Kir6.1-expressing oocytes decreased to a level close to control currents (Fig. 4 A). External barium inhibited the currents in both directions (barium-sensitive currents of 154 ± 18 nA and −108 ± 17 nA at +85 and −110 mV; n = 7). The reversal potential of the Ba2+-sensitive current was found to be −22 ± 4 mV in K-Barth, which is consistent with an intracellular K+concentration of ∼48 mM.
As Kir6.1 is a KATP channel, we characterized its ATP sensitivity. A 10-min treatment with 3 mM sodium azide, which has been shown to decrease the oocyte ATP level from 2.3 to 1.7 mM on average (14), had no effect on control oocytes (see Fig.4 B, inset). In contrast, in Kir6.1-expressing oocytes, sodium azide increased the currents (from 560 ± 120 nA to 930 ± 230 nA at +85 mV and from −180 ± 78 nA to −340 ± 82 nA at −110 mV), and a significant sodium azide current was measured (375 ± 140 nA; P < 0.05, and 160 ± 35 nA; P < 0.025, respectively, at +85 and −110 mV; n = 4; see Fig. 4 B). The reversal potential of the current induced by the sodium azide treatment was found to be −22 ± 1 mV in K-Barth.
Finally, as the KATP channel of the PCT was shown to be taurine sensitive, we tested the sensitivity to taurine of the Kir6.1 clone. Taurine sensitivity was tested by comparing the effect of injecting taurine versus mannitol in two sets of Kir6.1-expressing oocytes. More precisely, the current-voltage curves of a given number of oocytes were measured just before they were injected with either mannitol or taurine (final intracellular concentration reached ∼40 mM). After a period of 1–2 h to allow for recovery toward the resting oocyte resistance, the total current-voltage curve was determined for each oocyte a second time. Although mannitol injection produced an increase in the current measured at both +85 and −110 mV (150 ± 49 nA at +85 mV; P < 0.025, and 59 ± 66 nA at −110 mV; not significant; n = 8), taurine produced a significant decrease at the same potentials (135 ± 36 nA; P < 0.025, and 150 ± 43 nA;P < 0.010, respectively, measured at +85 mV and at −110 mV after taurine injection; n = 8). This is illustrated for the current measured at +85 mV in Fig.5 A. Taurine-sensitive currents were determined by subtraction with each Kir6.1-expressing oocyte serving as its own control (Fig. 5 B). The reversal potential of this current was −23 ± 2 mV.
Expression of Kir6.1 and SUR2A or SUR2B in Xenopus Oocytes
To compare the properties of the native PCT KATPchannel with those of cloned rabbit subunits, we expressed Kir6.1 RNA combined with SUR2A or SUR2B RNA. Expression in oocytes of Kir6.1 combined with SUR2A or SUR2B induced similar K+ currents (see Fig. 6 A: for SUR2A in K-Barth, the current measured 650 ± 110 nA at +85 mV and −220 ± 33 nA at −110 mV with a reversal potential of −30 ± 4 mV; n = 4; and for SUR2B in K-Barth, it was 1,470 ± 515 nA at +85 mV and −250 ± 75 nA at −110 mV with a reversal potential of −28 ± 8 mV; n = 4).
As the SUR subunits are responsible for the glibenclamide sensitivity, we tested the effect of increasing the concentration of glibenclamide (from 1 to 100 μM) on oocytes expressing Kir6.1/SUR2A. In four oocytes injected with a combination of Kir6.1/SUR2A, with respect to four control oocytes of the same batches, a specific current of 460 ± 120 nA at +85 mV could be measured. Application of 1-, 10-, 50-, and 100-μM glibenclamide produced a rapid inhibition (within 1 min) by 8, 32, 49, and 57%, respectively (P < 0.05 in all cases; see Fig. 6 B). Currents of Kir6.1/SUR2B-expressing oocytes were also inhibited by a 100-μM concentration of glibenclamide (74 ± 9% inhibition at +85 mV; n = 3). In contrast, glibenclamide had no measurable effect on oocytes expressing Kir6.1 alone (1 ± 3% inhibition at +85 mV using 100 μM glibenclamide; n = 3).
Taurine Sensitivity of Several Kir Family Members
Effect of taurine on ROMK1 channels.
The ATP-dependent rat Kir1.1A (ROMK1) RNA expressed in oocytes developed large currents (−13,000 ± 3,700 nA at −110 mV;n = 10) compared with that measured in control oocytes (−283 ± 67 nA at −110 mV in K-Barth; n = 3). A weak inward rectification was observed. Taurine sensitivity of the ROMK1 clone was then tested as previously described comparing the effect of intracellular mannitol versus taurine. Mannitol injection produced a weak decrease of the whole-cell currents at both +60 mV and −110 mV, which turned out to be statistically nonsignificant (−36 ± 19% and −20 ± 31%, respectively, at +60 mV and −110 mV; nonsignificant; n = 5). Similarly, taurine injection did not significantly modify whole-cell currents at either +60 mV or at −110 mV (−28 ± 24% and −12 ± 26%, respectively, at +60 mV and −110 mV; nonsignificant; n= 5; Table 1).
Effect of taurine on Kir2.1 channels.
Kir2.1 (IRK1) cDNA expression developed large inwardly rectifying currents (−17,000 nA at −80 mV; n = 7) in the presence of 20 mM K+ in the bath solution. Clearly, taurine injection failed to produce an inhibitory effect at all tested potentials (65 ± 53% and 12 ± 33% increases, respectively, at +60 mV and −80 mV; n = 7; Table 1).
Effect of taurine on Kir3.2 channels.
Human Kir3.2 RNA expression produced a modest but significant increase in whole-cell currents compared with control oocytes (−333 ± 63 nA and −77 ± 21 nA, respectively, for Kir3.2-expressing oocytes and control oocytes at −110 mV in K-Barth; n = 3). Unlike Kir2.1, which is a strongly inwardly rectifying K+channel, Kir3.2-expressing oocytes developed currents with an apparent outward rectification in the presence of 20 mM external K+. Either mannitol or taurine produced a slight nonsignificant decrease of membrane currents (−35 ± 23% and −16 ± 43%, respectively, at +85 mV and −110 mV after mannitol injection; and −10 ± 23% and −4 ± 53%, respectively, at +85 mV and −110 mV after taurine injection; n = 8; see Table 1).
Effect of taurine on Kir4.1 channels.
Expression of the human Kir4.1 ATP-dependent K+ channel in oocytes developed an apparent outwardly rectifying K+current in the presence of 20 mM external K+ (411 ± 34 nA and −214 ± 17 nA, respectively, at +85 mV and −110 mV in K-Barth; n = 13). In the same batch of oocytes, control currents of 130 ± 15 nA at +85 mV and −102 ± 16 nA at −110 mV were measured (n = 9). Although mannitol injection produced a weak decrease in membrane currents (−29 ± 12% and −35 ± 16%, respectively, at +85 mV and −110 mV), membrane currents measured before and after taurine injection were similar (+15 ± 26% at +85 mV and −22 ± 18% at −110 mV after taurine injection; see Table 1).
Effect of taurine on Kir5.1 channels.
Kir5.1 expression produced a two-to-threefold increase in whole-cell currents compared with control oocytes (−278 ± 61 nA and −115 ± 4 nA, respectively, for Kir5.1-expressing oocytes and control oocytes at −110 mV in K-Barth; n = 10). In rat Kir 5.1 RNA-expressing oocytes, membrane currents measured in K-Barth before mannitol injection were not significantly different from those measured after mannitol injection (−22 ± 30% at +85 mV and +5 ± 52% at −110 mV; not significant). The same observation was made with taurine injection (−25 ± 19% at +85 mV and −17 ± 43% at −110 mV; not significant; see Table 1).
Effect of taurine on different combinations of Kir6.x and SUR subunits.
As shown in Fig. 5, the cloned rabbit proximal Kir 6.1 expressed in oocytes developed taurine-sensitive K+ channels. To determine whether taurine sensitivity is a specific property of KATP channels, the effects of taurine were tested on different combinations of Kir and SUR subunits (Kir6.1, Kir6.2, SUR1, and SUR2A) cloned in different species.
In oocytes expressing rat Kir6.1 plus rat SUR1 (whole-cell current of −1,300 ± 300 nA at −110 mV in K-Barth), mannitol injection induced a slight increase in membrane currents (+14 ± 24% at +60 mV and +51 ± 42% at −110 mV, not significant; n= 8), whereas taurine significantly inhibited membrane currents at all tested membrane potentials (−46 ± 16% at +60 mV;P < 0.005, and −39 ± 16% at −110 mV;P < 0.05; n = 10; see Table 1).
Rat Kir6.1 was also expressed with human SUR2A (whole-cell current of −518 ± 68 nA at −110 mV in K-Barth; n = 14), and the effects of both mannitol and taurine were tested. It was observed that taurine significantly reduced the membrane currents at all tested potentials (−37 ± 14% at +85 mV, P< 0.05, and −40 ± 15% at −110 mV, P < 0.05;n = 6; see Table 1). In comparison, mannitol had no significant effect at all tested potentials (−31 ± 26% at +85 mV and +14 ± 67% at −110 mV).
The effect of taurine was also tested on the Kir6.2 subunit coexpressed with SUR1 (whole-cell current of −570 ± 83 nA at −110 mV in K-Barth; n = 24) and SUR2A (whole-cell current of 830 ± 174 nA at −110 mV in K-Barth; n = 24). It was observed that membrane currents in taurine-injected Kir6.2/SUR1-expressing oocytes were smaller (−30 ± 14% at +85 mV, P < 0.05, and −38 ± 13% at −110 mV,P < 0.01; n = 24; see Table 1) than membrane currents in mannitol-injected Kir6.2/SUR1-expressing oocytes. In Kir6.2/SUR2A-expressing oocytes, taurine significantly reduced the membrane currents (−24 ± 11% at +85 mV, P < 0.05, and −58 ± 18% at −110 mV; P < 0.005; Table 1). Mannitol had no significant effect (−16 ± 26% at +85 mV and +3 ± 40% at −110 mV, not significant; n= 24).
Taurine-sensitive inward currents expressed as a percentage of the measured inward current before taurine injection are systematically correlated with the known ATP sensitivity on all tested Kir members in Table 1. If we pool all the experiments with the different combinations involving Kir6.1, taurine produced an inhibition of 40% of the inward current measured at −110 mV, whereas the effect of taurine on Kir6.2 averaged 47%. The two effects are not significantly different. This inhibition is an underestimation, because injection of mannitol produces an average increase of 27%.
The goal of this study was to determine the molecular identity of the KATP channel of the rabbit PCT and to define some of the regulatory properties of the cloned channels. A homology cloning strategy was used to isolate from our PCT cDNA the Kir6.1, SUR2A, and SUR2B subunits. Expressed in Xenopus oocytes, Kir6.1 by itself forms a functional K+ channel, and the induced K+ current was shown to be sensitive to extracellular K+, external barium, cytosolic ATP levels, and intracellular taurine. Expression of Kir6.1 combined with SUR2A or SUR2B generates a glibenclamide-sensitive K+ current. These properties, including the specific inhibitory effect of taurine, are also found for the native PCT KATP channel.
Molecular Identity of the Rabbit PCT KATP Channel
A wide variety of KATP channels have been directly observed in different epithelial and nonepithelial cells. The conductances range from 9 to 300 pS and the sensitivities to ATP and glibenclamide vary considerably: the IC50 ranges from 10 μM to 1 mM for ATP and from 3 nM to 250 μM for glibenclamide (30, 33, 38). At the molecular level, reconstitution of ATP-sensitive K+ channels with cloned Kir6.x and SURx subunits suggests that Kir6.2/SUR1 could represent the pancreatic β-cell KATP channel, Kir6.2/SUR2A or Kir6.2/SUR2B could be responsible, respectively, for the cardiac and smooth muscle KATP channels, and Kir6.1/SUR2B could account for the vascular smooth muscle KATP type (for review, see Ref.39). However, other combinations could be required to explain the functional characteristics of all KATP channels previously identified. For example, it was recently shown (13) that Kir6.1 and Kir6.2 can coassemble to produce functional channels. In addition, other subunits could be implicated, as it was shown that a combination of Kir1.1 (ROMK1) with CFTR could produce a bona fide KATP channel (35).
In the case of PCT and other renal segments, the composition of the observed KATP channels is not clear. In a preliminary report (5), the ubiquitously expressed Kir6.1 subunit was localized at the basolateral membrane of rat proximal tubule, but it is unclear whether the Kir6.1 subunit, which develops a channel with a 70-pS conductance in Xenopus oocytes (22), could be responsible for the 50-pS activity found in rabbit PCT. In addition, it seems that SUR1 and SUR2A mRNAs were absent in mouse and rat kidney (19, 23), although SUR2B mRNA was shown to be abundantly expressed in rat proximal tubule, cortical thick ascending limb (CTAL), and outer medullary collecting duct (OMCD) (41). Interestingly, the combination of mouse Kir6.1/SUR2B has been described as the vascular smooth muscle KATP type, with functional characteristics quite different from those observed in PCT cells (45). In summary, the molecular identity and the precise composition of the KATP channel in rabbit PCT is in need of clarification.
Evidence for the presence of Kir6.1 in PCT.
Our homology cloning strategy allowed us to isolate the renal rabbit Kir6.1 although no signal for Kir6.2 could be detected. Inagaki et al. (22) found that uKATP-1 or Kir6.1 transcripts were expressed in all rat tissues tested: in the heart, ovary, and adrenals at high levels and in the kidney at low levels. Subsequently, Anzai et al. (5) found that Kir6.1 was expressed along the nephron except for the ascending and descending thin limbs. It appears that, in accordance with the intensity of Na+ reabsorption, Kir6.1 expression was as follows: PCT, medullary thick ascending limb > proximal straight tubule, CTAL, cortical collecting duct > OMCD > inner medullary collecting duct. Kir6.1 has been localized in cytoplasm and at the basolateral membrane of proximal tubule, CTAL, and collecting ducts. On the contrary, Kir6.2 mRNA seems to be absent in the kidney (19). In fact, the Kir6.2 subunit has been typically associated with pancreatic β-cells (4, 19, 36), cardiac and skeletal muscle (20, 32,36), smooth muscle (23), or brain (19,36). All of these results are in good agreement with evidence of the presence of Kir6.1 (and absence of Kir6.2) in PCT that was found in the present study.
Evidence for the presence of SUR2A and SUR2B in PCT.
With the same homology cloning strategy, we isolated from our proximal tubule cDNAs the rabbit renal SUR2A and its splice variant SUR2B; no signal could be detected with the degenerated PCR primer of SUR1. In agreement with this result, Inagaki and colleagues (19) found that SUR1 mRNA is absent in the kidney. However, using an RT-PCR approach as in the present study, Isomoto et al. (23) found a mouse SUR2B but not SUR2A signal in the kidney. This was also reported by Beesley et al. (9), who found a PCR product for SUR2B but not for SUR2A or SUR1 in the mouse kidney. These authors also used an anti-SUR2B antisera that revealed specific staining of SUR2B in distal nephron segments of mouse kidney, and a diffuse, low-level staining was observed in proximal tubules. In contrast, the rat kidney homologue of SUR2B was recently isolated and shown to be abundantly expressed in rat proximal tubules (41). Beesley et al. (9) found that no protein could be detected with anti-SUR2A antiserum in Western blot studies, and no staining could be observed in the immunohistochemistry at dilution of antisera equivalent to that used with anti-SUR2B. Interestingly, at a lower dilution they found a low level of diffuse staining in proximal tubules, which suggests that a low level of SUR2A could be present in the proximal tubule. We conclude that SUR2B is probably the main SUR transcript in the proximal tubule, with a lower level for SUR2A, which could explain the low level of staining observed at the protein level of SUR2A.
As it has been shown that Kir6.1 and Kir6.2 could coassemble to produce a functional K+ channel (13), it is possible to imagine that two different SUR subunits (SUR2A and SUR2B) could combine with Kir6.1 to form the proximal KATP channel. Such a coassembly would need to be proven at the protein level but would help in explaining the observed diversity of native KATPchannels.
Reconstitution of KATP Channels in Xenopus Oocytes With Kir6.1, Kir6.1/SUR2A, and Kir6.1/SUR2B
Expression of rabbit renal Kir6.1 yields a functional K+ channel.
Functional expression of Kir6.1 in Xenopus oocytes has revealed an unexpected result: contrary to a previous suggestion (14), the expression of an exogenous SUR subunit is not required to generate a KATP channel. Whole-cell currents of Kir6.1-expressing oocytes are significantly higher than those measured in control oocytes. However, the induced currents are <1 μM. This low current could be explained by the absence of an expressed SUR subunit. Indeed, it has been suggested, for example, that the SUR1 subunit could increase membrane insertion of the Kir6.2 subunit (25); however, we found that expression of Kir6.1 combined with the SUR2A or SUR2B subunit developed similar, quite-low K+ currents (Fig. 6). The high mean ATP level of healthy oocytes could also explain the low activity of these KATPchannels. Indeed, lowering the mean ATP level (by a 10-min treatment with sodium azide) increased the K+ current.
As expected for a K+ current, barium addition returned the conductance to the level of noninjected oocytes, and the inward currents induced by increases in external K+ concentration were significantly higher in Kir6.1-expressing oocytes than in control oocytes. In Kir6.1-expressing oocytes but not in control oocytes, a slight decrease in cytosolic ATP concentration induced a significant current increase. Finally, as for the native KATP channel, K+ currents of Kir6.1-expressing oocytes can be inhibited by injection of intracellular taurine. These different lines of evidence point to a significant KATP activity in Kir6.1-injected oocytes. It remains to be shown whether the oocyte can provide an endogenous SUR subunit to form a complete KATPchannel or whether such a subunit is unnecessary for promoting a KATP channel activity. However, the presence of a sufficient level of endogenous SUR subunits to form a KATPchannel with the overexpressed Kir6.1 is doubtful, because the K+ current of Kir6.1-injected oocytes is insensitive to glibenclamide.
The present study shows that coexpression of Kir6.1 with SUR2A induced a glibenclamide-sensitive K+ current in oocytes. These results are at variance with previous work, which suggests that this combination failed to induce a significant channel activity (2). On the other hand, our results (45) with expression of Kir6.1 and SUR2B confirmed earlier results in HEK293T cells that this subunit combination was fully capable of generating a functional K+ channel.
Comparison Between Regulatory Properties of Cloned Rabbit KATP and Known Characteristics of Native KATPChannels in PCT
Expression of Kir6.1 in Xenopus oocytes has previously been shown to develop a K+ channel with a 70-pS conductance, whereas Kir6.1 combined with SUR2B produces a 29–33-pS K+ channel in HEK293 cells (13,45). It remains to be shown whether the expression of Kir6.1 with SUR2A and SUR2B could be responsible for a native 50-pS KATP channel in the rabbit proximal tubule. Comparison of the conductances of native KATP channels with those of Kir and SUR subunits expressed in heterologous cells (for review, see Ref.6) has already shown some discrepancy. However, focusing on the comparison of the specific properties of the proximal tubule KATP channel with the characteristics of the cloned Kir6.1/SUR2A and/or Kir6.1/SUR2B expressed in oocytes, our results suggest that the isolated clones are good candidates for a native KATP channel in PCT.
The best evidence that Kir6.1 could be involved in the PCT KATP is its taurine sensitivity. Indeed, we recently found that the native KATP channel of the proximal tubule was inhibited by intracellular taurine (31). Several physiological roles have been identified for intracellular taurine, including modulation of Na+, Ca2+, or Cl− transport (for review, see Ref. 18), but taurine sensitivity of the K+ channel is a relatively new finding. Direct inhibition by taurine of the KATP channel has been also shown in guinea pig and rabbit ventricular cardiomyocytes (16, 37). More recently, experiments (42) on skeletal muscle fibers have shown that taurine stabilizes the close state of the KATP channel. In addition, it has been found (42) that taurine could increase glibenclamide affinity, which suggests that taurine could interfere with the glibenclamide site on the SUR subunit or on the Kir subunit at a site allosterically coupled to the SUR subunit. We now demonstrate that the cloned renal rabbit Kir6.1 expressed in Xenopus oocytes develops K+ currents that are sensitive to intracellular taurine, which suggests that the taurine sensitivity of the KATPchannel does not require the presence of an expressed SUR subunit. Finally, we recently found (10) that the endogenous, ATP-sensitive K+ channel in Xenopus oocytes, which was activated when ATP levels were decreased below 1 mM, was significantly inhibited by intracellular taurine.
Until now, the only K+ channels reported to be taurine sensitive were KATP channels. Testing the effect of taurine on different K+ channels from the Kir family, we have demonstrated that taurine sensitivity is not a common characteristic of the Kir family but is a specific property of KATP channels.
Taurine sensitivity is particularly interesting in the context of the proximal tubule, which is subjected to frequent changes in cell volume in response to variations in transepithelial Na+ transport (27). If taurine can be shown to be an exportable osmolyte in PCT as for other tissues (15, 28, 34, 40), a PCT cell that is initially swollen by excessive apical Na+ uptake could react by releasing its intracellular taurine. This decrease in intracellular taurine concentration would activate basolateral KATP channels. In this way, loss of intracellular osmolytes and activation of a K+ current would act in concert to efficiently regulate PCT cell volume.
In rabbit PCT, Tsuchiya et al. (43) found that a high concentration of glibenclamide (IC50 of 250 μM) was needed to inhibit the basolateral KATP. The absence of SUR1 is consistent with this low sensitivity, as SUR1 is characterized by a glibenclamide affinity on the order of 1 nM (3, 19). Interestingly, SUR2A and SUR2B are low-affinity receptors that require 100–1,000-fold higher glibenclamide concentrations to inhibit the channel (20, 23). Identification of the presence of SUR2A and SUR2B in the present study is in qualitative agreement with the low sensitivity observed for the basolateral KATP of rabbit PCT. In addition, expression of Kir6.1 with SUR2A or SUR2B results in the development of K+ currents that are sensitive to micromolar concentrations of glibenclamide (such as 50% inhibition of the Kir6.1/SUR2A current with 50 μM glibenclamide).
In summary, cDNA for Kir6.1, SUR2A, and SUR2B were found in a rabbit PCT cDNA library, whereas Kir6.2 and SUR1 appeared to be absent. Expression of Kir6.1, Kir6.1/SUR2A, and Kir6.1/SUR2B inXenopus oocytes induces K+ currents that are sensitive to external K+, external barium, cytosolic ATP, and intracellular taurine; all of these properties were previously found for the native KATP channel. This suggests that the KATP channel of rabbit PCT could be formed from a combination of Kir6.1 with SUR2A and/or SUR2B. This suggestion must be further tested at the protein level.
The authors acknowledge S. J. White, R. Sauvé, H. Van Tol, L. Salkoff, and S. Seino who graciously provided the different Kir clones.
This work was supported by the Medical Research Council of Canada Grant MT-10900 to J.-Y. Lapointe and R. Laprade. E. Brochiero acknowledges a postdoctoral fellowship from the Ministère de l′Education du Québec.
Address for reprint requests and other correspondence: J.-Y. Lapointe, Département de Physique, P.O. Box 6128 Succursale “Centre-ville,” Montréal, Québec H3C 3J7, Canada (E-mail:).
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 © 2002 the American Physiological Society