INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

CONTROL OF K⁺ permeation through the cell membrane is a ubiquitous phenomenon (22). Multiple K⁺ channels are expressed in cells, with distinct biophysical, physiological, and pharmacological properties (2, 11, 22). These channels play a critical role in excitable cells by determining action potential firing or muscle contraction. In epithelial cells, selective expression in distinct membrane domains and regulation of K⁺ channels lead to control of resting membrane potential, cell volume, and transepithelial K⁺ transport, as well as K⁺ recycling (8). Recently, several renal K⁺ channel subunits have been characterized at the molecular level (3, 4, 12, 26).

Two main families of pore-forming K⁺ channel subunits sharing common structural motifs have been identified (5, 13, 21). Voltage-gated K⁺ channels have six transmembrane segments (TMS) and a so-called P domain, which forms part of the conduction pore. Among them, Shaker-like channels have been widely studied. Inward rectifying K⁺ channels have only two TMS separated by a P domain. The ROMK channels belong to this category (1, 10, 12). Importantly, these channels are likely to form multimers to yield a functional K⁺ pore (2). Recently, a new structural family of K⁺ channels has been discovered in mammals. These channel subunits have four TMS and two P domains (6, 16, 17). TWIK-1 (16) is the first identified member in this family. It exhibits weak inward rectifying properties when expressed in Xenopus oocytes. Its unitary conductance is 30-40 pS, and it can be blocked by quinidine. Ba²⁺ can also block its activity, while it is relatively insensitive to tetraethylammonium ion. Interestingly, TWIK-1 (16) is downregulated by internal acidification and upregulated by protein kinase C (PKC) (16, 17). TWIK-1 mRNA has been shown to be expressed mainly in brain, although other tissues were positive, such as lung, skeletal muscle, and kidney (16, 17). TWIK-1 proteins self-associate to form functional covalent dimers (17, 18).

Several K⁺ conductances have been reported in the different tubular epithelia lining the renal nephron, with distinct biophysical and pharmacological properties (8, 30). Attempts to correlate these conductances to the growing number of cloned K⁺ channels require comparison of their functional properties and their precise cellular specificity of expression (cell type and apical vs. basolateral membrane).

The aim of this study was to determine precisely the cellular expression of TWIK-1 within kidney tubular cells and its membrane domain of expression compared with other K⁺ channels that are known to be expressed in kidney cells such as ROMKs.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Animals. Adult male Sprague-Dawley rats were fed a normal diet (control rats), a low-K⁺ diet (UAR, Epinay sur Orge, France) containing 120 mg K⁺/kg for 10 days, or a low-Na⁺ diet (UAR) containing 90 mg Na⁺/kg for 8 days.

Antibodies. Affinity-purified rabbit polyclonal antibodies directed against the COOH-terminal region of TWIK-1 (amino acids 264-336) fused to glutathione S-transferase (GST) (17, 18) were used. Anti-TWIK-1 antibodies were raised against a GST fusion protein containing the COOH terminus of TWIK-1 (amino acids 264-336). Female New Zealand White rabbits were immunized with 300 µg of purified fusion protein in the presence of complete Freund's adjuvant and boosted 1 mo later with 150 µg of the immunogen in the presence of incomplete Freund's adjuvant. Rabbits were bled 15 days after the boost. The antibodies were affinity purified by using His-Tag fusion proteins containing the same domains of TWIK-1 as the GST fusion proteins used for the immunization. Briefly, the crude antisera were incubated for 4 h at 4°C with 100-200 µg of purified His-Tag fusion proteins previously transferred to Hybond C-extra nitrocellulose membranes (Amersham). After three washes in PBS [10 mM phosphate buffer (pH 7.2) and 0.15 M NaCl] and 0.1% Tween 20, the anti-TWIK-1 antibodies were recovered by a 1-min elution of each strip with 0.1 M glycine and 0.5% BSA (pH 2.8). After the elution the purified antibodies were rapidly brought to pH 7.6 with 1 M Tris (pH 8.0) and 0.5% BSA.

View larger version (54K): [in this window] [in a new window]   Fig. 1.   Western blots of TWIK-1 and aquaporin-2 (AQP-2) in kidney homogenates. A: crude rat kidney homogenates were blotted in presence of affinity-purified anti-TWIK-1 antibody, in absence () or presence (+) of immunizing fusion protein, and with an anti-actin antibody as an internal control. B: signal observed with unpurified anti-AQP-2 antibody, alone () or in presence of immunizing peptide (+); actin was used as an internal control.

Detection of TWIK-1 in COS cells. The TWIK-1 sequence was excised from the pEXO-TWIK-1 plasmid (16) and subcloned into the pIRES-cd8 vector to obtain pIREScd8-TWIK-1. The pIRES-cd8 vector was obtained by replacing the neo gene in pIRESneo (Clontech) by the CD8 gene. COS cells were seeded at a density of 3 × 10⁴ cells/35-mm dish 24 h before transfection. Cells were transiently transfected by the classical DEAE-dextran method with 1 µg of pIREScd8-TWIK-1 plasmid per dish. After 1 day, cells were dissociated and plated on polylysine-coated coverslips in a 24-well cluster. Transfected cells were visualized 48 h after transfection by application of anti-CD8 antibody-coated beads (Dynabeads, Dynal). TWIK-1 immunodetection (Fig. 2) was then performed as previously described (18), except cells were permeabilized by addition of 0.1% Triton X-100 in the blocking solution (PBS supplemented with 2% BSA and 5% normal goat serum). For control, immunodetection was performed on cells transfected in the same manner but by using affinity-purified anti-TWIK-1 antibodies preincubated over 30 min with the GST-TWIK-1 fusion protein that has been used for rabbit immunization (25 µg GST-TWIK-1/ml detection solution). For Western blot, proteins from COS cells and synaptic membranes from adult rat brain were prepared and analyzed in the absence of reducing agents, as previously described (17, 18).

Immunolocalization of TWIK or AQP-2 on kidney sections. Kidneys from male Sprague-Dawley rats were obtained after in vivo perfusion (at 80 mmHg) of the aorta with 2% paraformaldehyde, then 4-6 h of immersion in the same fixative at 4°C. The tissue was then cryoprotected by immersion in 30% sucrose in phosphate buffer (120 mM Na₂HPO₄-NaH₂PO₄, pH 7.2) overnight at 4°C, frozen in liquid nitrogen, and stored at 80°C. Cryostat sections (10 µm) were deposited on Superfrost slides, dried in air, and immersed in phosphate buffer.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Specificity of TWIK-1 antibody. Antibody specificity was tested on transiently transfected COS cells. Cells were transfected with a polycistronic vector to coexpress TWIK-1 and the CD8 protein from the same transcript. Expressing cells were visualized by application of anti-CD8 antibody-coated beads (Fig. 2, A and B). As expected, a positive signal was obtained by incubating TWIK-1-expressing cells with anti-TWIK-1 antibodies. TWIK-1 channels were not detected from mock-transfected cells (not shown) or from TWIK-1-expressing cells when anti-TWIK-1 antibodies were preincubated with the immunizing fusion protein before the detection step (Fig. 2B). Figure 2C shows that TWIK-1 channels were specifically detected from TWIK-1-expressing COS cells (lane 2) and from rat brain synaptic membranes (lane 3). The M_r of the rat brain TWIK-1 (~80,000) is identical to the previously reported M_r of the mouse brain TWIK-1 (17). It probably corresponds to a disulfide-bridged dimeric form of the protein, as shown for the human channel (18). The M_r of TWIK-1 is lower (70,000-75,000) in transiently transfected COS cells than in brain (Fig. 2C, lane 2). This could be due to differences in posttranslational modifications of the protein. A similar lower M_r has been observed in another transient expression system (18).

View larger version (117K): [in this window] [in a new window]   Fig. 2.   Immunodetection of TWIK-1 in transfected COS cells and in rat brain. A and B: TWIK-1 was coexpressed in transfected COS cells together with CD8 protein. Transfected COS cells were visualized with anti-CD8-coated beads (arrowheads). Then cells were fixed, permeabilized, and incubated successively with affinity-purified anti-TWIK-1 antibodies in absence (A) or presence (B) of immunogenic glutathione S-transferase-TWIK-1 fusion protein and with FITC-conjugated goat anti-rabbit IgG. Immunocomplexes (stained green) were visualized by fluorescence microscopy at ×400 magnification. C: Western blot analysis of TWIK-1 protein in mock-transfected COS cells (lane 1), TWIK-1-expressing COS cells (lane 2), and synaptic membranes of adult rat brain (lane 3).

Immunolocalization of TWIK-1. Immunofluorescence was apparent in several cell types of the nephron, as illustrated in Fig. 3, which shows the results obtained in the superficial cortex, the deep cortex, the medulla, and the papilla of paraformaldehyde-fixed kidneys. In the cortex (Fig. 3A) the apical membrane of the proximal convoluted tubule showed a strong signal that was clearly limited to the brush-border membrane (Fig. 3E). The expression of TWIK-1 decreased along the length of the proximal tubule and appeared much higher in its initial portion (proximal convoluted tubule) than in its terminal portion, i.e., the pars recta, located in the deep cortex, as shown in Fig. 3B (cf. Fig. 3A). However, the brush-border membranes were clearly positive in these paraformaldehyde-fixed kidneys, whereas the signal was more diffuse after methanol fixation (see Fig. 5). The distal tubules and early collecting ducts (Fig. 3A) were also positive, with cytoplasmic as well as apical staining, whereas glomeruli were negative. In the collecting duct, some cells were positive (Fig. 3F), with a cellular pattern of expression that varies from cell to cell: in some cases, TWIK-1 appears intracellular; in other cases, it is clearly restricted to the apical membrane (Fig. 3G). The reason for this variable expression in the collecting duct is unclear, and the cytoplasmic signal may represent a storage compartment of TWIK-1. To see whether this pattern of expression could vary with the K⁺ status of the animal, immunofluorescence studies were performed in rats fed a low-K⁺ diet (to enhance renal K⁺ reabsorption) and in rats fed a low-Na⁺ diet, a condition that is known to enhance plasma aldosterone concentration and thus K⁺ secretion in the collecting duct. Neither of these manipulations altered the apparent level of expression or the cellular pattern of expression of TWIK-1 in the collecting duct (not shown). In the outer medulla (Fig. 3C), TWIK-1 was found in the medullary collecting duct and in the thick ascending limbs of the loop of Henle. Identification of cortical thick ascending limbs was difficult on these sections. TWIK-1 immunofluorescence was also present in the papillary collecting duct (Fig. 3D), where the signal was diffuse throughout the cell; the loops of Henle (thin limbs) were negative.

View larger version (136K): [in this window] [in a new window]   Fig. 3.   Immunofluorescence of TWIK-1 in kidney sections. Confocal microscopy was used to show expression of TWIK-1 at protein level in renal superficial cortex (A), deep cortex (B), medulla (C), and papilla (D). Immunofluorescence was restricted to brush-border (apical) membrane of initial parts of proximal tubules (pt in A, and E) and to cells of distal tubule and early collecting duct (dct), whereas glomerulus (g) was negative. Collecting duct cells (cd) remain positive along its entire length, including its papillary portion (D). Signal observed in proximal tubule was strongly reduced in its terminal part, i.e., pars recta (pr in B). In cortical collecting duct, immunofluorescence was intracellular or apical (F and G) in a minority of cells. No clear signal was visible over cells of thick ascending limb of loop of Henle (hl in C) or thin limbs of loop of Henle (hl in D). Magnification ×170 in A-D, ×240 in E and F, and ×360 in G.

View larger version (78K): [in this window] [in a new window]   Fig. 4.   Control immunofluorescence. Kidney sections from outer cortex (A), inner cortex (B), medulla (C), and papilla (D) incubated with preimmune serum show nonspecific signal in some cells, mainly in cortex (glomerulus) and papillary collecting duct cells. Signal is clearly reduced compared with that observed with TWIK-1 antibody (Fig. 3). TWIK-1 immunofluorescence is displaced in presence of immunizing fusion protein in outer cortex (E), inner cortex (F), and papilla (G). Magnification ×100.

View larger version (116K): [in this window] [in a new window]   Fig. 5.   Immunolocalization of TWIK-1 and Tamm-Horsfall protein (THP) in kidney. Methanol-fixed cryostat cortex (A-C) and medulla (D-F) sections were incubated with anti-TWIK-1 (B and E, stained red) or anti-THP (A and D, stained green) antibody (to identify thick ascending limbs of loop of Henle). C and F: double immunofluorescence showing colocalization of TWIK-1 and THP (stained yellow) in cortical and medullary parts, respectively, of thick ascending limb of loop of Henle. Magnification ×160.

View larger version (109K): [in this window] [in a new window]   Fig. 6.   Immunolocalization of TWIK-1 and AQP-2 in kidney cortex. Paraformaldehyde-fixed kidney sections were incubated with anti-AQP-2 antibody alone (A) or in presence of immunizing peptide (B); preimmune serum gave low background signal (C). D: coimmunolocalization of TWIK-1 (stained red) and AQP-2 (stained green) showing distinct cellular expression. Such distinct cellular expression is also illustrated by incubating serial sections with anti-AQP-2 antibody (E), a marker of principal cells of cortical collecting duct, or anti-TWIK antibody (F). Magnification ×100 for A-C and ×250 for D-F. Arrows, intercalated cells.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Several K⁺ conductances have been reported in kidney cells over the last few years (for review see Refs. 8, 9, and 30). Detailed information is now available on their biophysical and pharmacological properties in distinct cell types. They are involved in the establishment of a K⁺ concentration gradient (which can be used to generate membrane potential of tubular cells and Na⁺-coupled transport), in regulation of cell volume, in recycling of K⁺ across apical and basolateral membranes, and in K⁺ secretion in the collecting duct (8, 9, 30). Low-conductance (30-pS) K⁺ channels have been identified in the apical membrane of collecting duct principal cells (7, 8, 28, 31); these channels share some properties with TWIK-1. The low-conductance apical K⁺ channel of the cortical collecting duct and TWIK-1 are weakly inward rectifying, with a low conductance (30-35 pS) and a low sensitivity to tetraethylammonium ion, and are downregulated by internal acidification (7, 28, 30, 31). Their sensitivity to PKC activation appears to differ, since PKC upregulates TWIK-1 activity in oocytes (16, 17) but reduces apical K⁺ channel activity in the collecting duct (8, 29); protein kinase A activates apical K⁺ conductance (29), whereas TWIK-1 seems to be insensitive to cAMP (16). Whether these differences are real or depend on the cellular context of expression (oocytes injected with TWIK-1 mRNA vs. native cells of the collecting duct) remains unknown. Finally, apparent divergences may be due to more complex phenomena; the notion of possible heteromultimerization of K⁺ channels (compared with homomultimerization) has been proposed and may suggest a new diversity of function among cloned K⁺ channels (2).

To gain some insight into the involvement of TWIK-1 in K⁺ handling along the nephron, its expression at the protein level in the rat kidney was characterized. Results show that TWIK-1 renal expression is in the proximal tubule, the thick ascending limb of the loop of Henle, and collecting duct intercalated cells. This pattern of expression along the nephron is close to that observed with another class of low-conductance K⁺ channels, the ATP-sensitive ROMKs, which belong to the inward rectifying K⁺ channel family, characterized by two membrane-spanning segments (10, 12). ROMK mRNAs have been shown (10, 15) to be expressed essentially in the distal half of the nephron, with distinct expression according to the isoform. In particular, ROMK-2 and ROMK-3 mRNAs were found in the ascending limb of the loop of Henle and in the distal tubule and collecting duct, whereas ROMK-1 mRNA was absent in the loop of Henle and present in the distal nephron (1). In the cortical collecting duct, ROMKs are restricted to the apical membrane domain of principal cells (32), at variance with TWIK-1, which is in intercalated cells. Thus it appears that these two K⁺ channels have complementary patterns of expression in the cortical collecting duct.

Expression of apical K⁺ conductances along the nephron is interesting to discuss in view of genetic disorders such as those observed in Bartter's syndrome. This syndrome was initially attributed to a defect in the Na-K-2Cl cotransporter (NKCC2) in the loop of Henle (23), leading to impaired NaCl reabsorption in this epithelium, responsible for Na⁺ wasting, secondary hyperaldosteronism and hypokalemic alkalosis. More recently (14, 24), mutations of ROMK have been identified in some cases of Bartter's syndrome. The loss of ROMK function results in the inability to recycle K⁺ from the cells of the ascending limb of the loop of Henle, leading to severe impairment of the activity of the Na-K-2Cl cotransporter (all mutations identified are in the core peptide shared by all known ROMK isoforms; consequently, activity of all isoforms is expected to be affected by these mutations). Because ROMK isoforms are also expressed more distally along the nephron, i.e., in the distal tubule and collecting duct (which are major sites for net renal K⁺ secretion), it was expected that ROMK mutations would also impair distal K⁺ secretion and prevent the hypokalemia secondary to hyperaldosteronism. Of interest, hypokalemia, although less severe than in NKCC2 mutations, was also present, despite expected ROMK-dependent impairment in distal K⁺ secretion (24). This suggests that the ROMK K⁺ conductance plays a major role in the loop of Henle, not in the collecting tubule. We propose that TWIK-1 activity may compensate for the loss of function of ROMK in the collecting duct. Such functional compensation may not exist (or is not sufficient) in the loop of Henle, despite the presence of TWIK.

Two other nephron segments express TWIK-1, but not ROMKs: the proximal tubule and the inner medullary collecting duct. In the brush-border membrane of the proximal tubule, TWIK-1 may participate to maintain the negative potential of tubule cells or to regulate cellular volume (8), together with another K⁺ channel subunit, minK (also expressed in the brush-border membrane of the proximal tubule) (25). In the inner medullary (papillary) collecting duct, a basolateral Shaker-like K⁺ channel has been reported (27); together with TWIK-1, it may play a role in the final adjustments of K⁺ secretion in the urine.

In a recent report (20) a cDNA named KCNK1 has been cloned from human kidney, with complete identity to TWIK-1. A partial rabbit clone was also amplified, and rabbit-specific primers were used to probe the expression of KCNK1 along the nephron. Positive signals were obtained in the cortical part of the thick ascending limb of the loop of Henle and in the cortical and outer medullary collecting duct. The proximal tubule, the medullary thick ascending limb of the loop of Henle, and papillary collecting ducts were referred to as negative. If this cDNA is identical to TWIK-1, its expression at the mRNA level along the rabbit nephron is clearly distinct from our findings at the protein level in the rat. However, because no sequence information has been provided for the amplified rabbit sequences and because data originate from a single experiment using nonquantitative RT-PCR, it is difficult to interpret these findings in terms of specific K⁺ channel expression.

In conclusion, we have shown that TWIK-1, a new K⁺ channel with four transmembrane domains, is selectively expressed in the brush-border membrane of the proximal convoluted tubule, in the thick ascending limb of the loop of Henle, and in collecting duct intercalated cells, with intracellular and apical localization. TWIK-1 may play a significant role in K⁺ secretion, thus participating in the final adjustments of K⁺ handling in the kidney.