Regulation of an inwardly rectifying K + channel by nitric oxide in cultured human proximal tubule cells

【摘要】  We investigated the effects of nitric oxide (NO) on activity of the inwardly rectifying K + channel in cultured human proximal tubule cells, using the cell-attached mode of the patch-clamp technique. An inhibitor of NO synthases, N -nitro- L -arginine methyl ester ( L -NAME; 100 µM), reduced channel activity, which was restored by an NO donor, sodium nitroprusside (SNP; 10 µM) or 8-bromo-cGMP (8-BrcGMP; 100 µM). However, SNP failed to activate the channel in the presence of an inhibitor of soluble guanylate cyclase, 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one (10 µM). Similarly, the SNP effect was abolished by a protein kinase G (PKG)-specific inhibitor, KT-5823 (1 µM), but not by a protein kinase A-specific inhibitor, KT-5720 (500 nM). Another NO donor, S -nitroso- N -acetyl- D, L -penicillamine (10 µM), mimicked the SNP-induced channel activation. In contrast to the stimulatory effect of SNP at a low dose (10 µM), a higher dose of SNP (1 mM) reduced channel activity, which was not restored by 8-BrcGMP. Recordings of membrane potential with the slow whole cell configuration demonstrated that L -NAME (100 µM) and the high dose of SNP (1 mM) depolarized the cell by 10.1 ± 2.6 and 9.2 ± 1.0 mV, respectively, whereas the low dose of SNP (10 µM) hyperpolarized it by 7.1 ± 0.7 mV. These results suggested that the endogenous NO would contribute to the maintenance of basal activity of this K + channel and hence the potential formation via a cGMP/PKG-dependent mechanism, whereas a high dose of NO impaired channel activity independent of cGMP/PKG-mediated processes.

patch-clamp; human kidney; guanosine 3',5'-cyclic monophosphate; sodium nitroprusside; protein kinase G

【关键词】  Regulation inwardly rectifying cultured proximal

NITRIC OXIDE (NO) is a small and ubiquitous molecule in vivo, produced mainly by NO synthase (NOS), which is classified into the following three isoforms: neuronal NOS (nNOS), inducible NOS (iNOS), and endothelial NOS (eNOS) ( 2, 38 ). Many investigators so far reported that NO affected a variety of functions of the kidney, including renal hemodynamics, renin secretion, and tubular transport of solutes and water ( 15, 37 ). Furthermore, all of the three NOS isoforms were detected in the kidney ( 15, 37 ).

We have previously demonstrated that an inwardly rectifying K + channel with inward conductance of 40 pS is the most frequently observed K + channel in cultured human proximal tubule cells under the control condition ( 29 ). The activity of this K + channel was stimulated by protein phosphorylation mediated by protein kinase A (PKA; see Ref. 29 ) and protein kinase G (PKG; see Ref. 28 ). We further confirmed that atrial natriuretic peptide (ANP), which stimulates ANP receptor-coupled guanylate cyclase and hence cGMP synthesis, elevated channel activity through PKG-mediated phosphorylation ( 28 ). Such characteristics are similar to those of the K + channel identified in opossum kidney proximal tubule cells ( 17, 18 ).

NO is another well-known activator of cGMP synthesis, the action of which involves stimulation of cytosolic soluble guanylate cyclase ( 37, 38 ). Thus it is possible that NO would participate in the regulation of this K + channel. Indeed, several investigators reported that NO affected the activity of K + channels in the distal nephron segments, such as cortical collecting duct ( 21, 22, 27, 35 ) or thick ascending limb ( 9, 23 ). However, it still remains unknown whether NO would modulate the activity of K + channels in the proximal tubule cells.

Although NO has been demonstrated to affect transport of solutes and water in the proximal tubule, its mode of action is a matter of controversy ( 19 ). A study using microperfused rat proximal tubules revealed that an NOS inhibitor suppressed the absorption of Na + and HCO 3 -, whereas NO donors stimulated it, probably by affecting the apical Na + /H + exchange ( 33 ). Furthermore, NOS knockout mice exhibited a defect in the proximal tubular absorption of fluid and HCO 3 - ( 34 ). In contrast, other investigators demonstrated that NO inhibited the apical Na + /H + exchange ( 32 ) or the basolateral Na + -K + -ATPase activity ( 10, 20 ). Because the apical Na + entry and basolateral pump activity are closely related to activity of the basolateral K + channel ( 8, 30, 36 ), it is intriguing to examine whether NO modulates the K + channel activity. In this study, we attempted to clarify the involvement of NO action in modulating the K + channel activity in cultured human proximal tubule cells, using the patch-clamp technique.

METHODS

Cell culture. Renal proximal tubule epithelial cells (RPTECs) isolated from the normal kidney of a 31-yr-old woman (strain 5899, lot 8F0690) were purchased from Clonetics (Walkersville, MD). It is guaranteed 90% of the cells are positive for -GTP, a marker protein specific to the proximal tubule ( 11 ). These cells were provided as cryopreserved secondary cultures and maintained in the renal epithelial cell growth medium (Clonetics) in a humidified atmosphere of 5% CO 2 -95% air at 37°C. In the experiments, the cells were dispersed from 70 to 80% confluence at passages 3-6 with trypsin/EDTA, resuspended in the growth medium, and seeded on collagen-coated coverslips (Iwaki Glass, Tokyo, Japan) in 24 multiwells at a density of 2 x 10 4 cells/well. After 3-7 h of incubation, the coverslips were transferred to an open bath heating chamber (Warner, Hamden, CT).

Solutions. The control bath solution contained (in mM) 140 NaCl, 5 KCl, 2 CaCl 2, 1 MgCl 2, 5 glucose, and 10 HEPES. In some experiments, NaCl was replaced with KCl to yield a high-K + (145 mM) bath solution. The pipette solution contained (in mM) 145 KCl, 1 MgCl 2, 1 EGTA, and 1 HEPES. All of these solutions were titrated to pH 7.3 with 5 N NaOH or KOH.

Test substances. Sodium nitroprusside (SNP), N -nitro- L -arginine methyl ester ( L -NAME), human ANP (hANP), and 8-bromo-cGMP (8-BrcGMP) were purchased from Sigma (St. Louis, MO). S -nitroso- N -acetyl- D, L -penicillamine (SNAP) and 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one (ODQ) were from Cayman (Ann Arbor, MI). A membrane-permeant PKG-specific inhibitor, KT-5823, and a PKA-specific inhibitor, KT-5720, were purchased from Calbiochem (La Jolla, CA). SNAP, ODQ, KT-5823, and KT-5720 were dissolved in DMSO as stock solutions, whereas others were dissolved in water. These stock solutions were diluted with the appropriate amount of the bath solution and added to the bath by hand pipetting. The final concentration of DMSO in the bath ranged from 0.004 to 0.054%, which did not affect channel activity.

Patch-clamp technique. Single channel currents were recorded with cell-attached and inside-out patches applied to the surface membrane of single RPTECs. The pipette holding potential ( V p ) was set at 0 mV for cell-attached patches and +50 mV for inside-out patches, unless otherwise stated. All patch experiments were performed at 33°C, which was adjusted by a heater platform connected with a controller (TC-324B; Warner). This temperature setting was very suitable for both the gigaseal formation and cell viability. Patch pipettes were fabricated from glass capillaries (Warner), with the resistance ranging from 2 to 4 M when filled with the pipette solution. Current signals were recorded with a patch-clamp amplifier Axopatch 200B (Axon, Foster City, CA) and stored on a DAT recorder (RD-120TE; TEAC, Tokyo, Japan). The recorded signals were then low-pass filtered (3611 Multifunction Filter; NF electronic instruments, Tokyo, Japan) at 500 Hz and digitized at a rate of 2.5 kHz through an interface (Digidata 1200A; Axon). The acquired data were analyzed with acquisition/analysis software (pCLAMP6; Axon) on an IBM-compatible personal computer. Current traces of downward deflections represented inward currents. Channel activity was determined by NP o, which was calculated from amplitude histograms as

where N is the maximum number of channels observed in the patch, P o is the open probability, n is the number of channels observed at the same time, and t n is the probability that n channels are simultaneously open. The control values of NP o varied among patches, ranging from 0.15 to 6.12, with an average of 1.57 ± 0.10 ( n = 95). Because the values of control NP o include the wide variation, we calculated normalized channel activity ( NP o,e / NP o,c ) for the convenient evaluation of effects of the substances. NP o,c and NP o,e are the channel activities under control and experimental conditions, respectively. Routinely, we determined NP o,c from a 20-s sampling period just before adding the substance when the steady state lasted for at least 60 s. NP o,e was determined from a 20-s sampling period extracted from the steady state for at least 20-30 s made by the experimental substance. If the 20-s sampling impaired the precise estimation of NP o because of the baseline drift, a couple of 10-s sampling periods were taken, and the averaged NP o was adopted. In some experiments, membrane potential was recorded, using the slow whole cell configuration. To make this configuration, the very tip of the patch pipette was filled with the pipette solution, and the rest of the pipette was backfilled with that containing 100 µg/ml Nystatin (Sigma). Nystatin was dissolved in DMSO as a stock solution of 50 mg/ml and diluted with the pipette solution.

Statistics. Data are expressed as means ± SE from 4-18 patches. Student's t -test or ANOVA in conjunction with Bonferroni t -test was used for statistical comparisons. A P value <0.05 was considered to be significant.

RESULTS

Figure 1 A shows a serial current recording of an inwardly rectifying K + channel in RPTECs, which was obtained using a cell-attached patch under the condition with high-K + bath solution at V p shown on the left of each trace. It is apparent that the amplitude of inward current is larger than that of outward current. Data from four patches were pooled and plotted to draw a current-voltage curve ( Fig. 1 B ). The inward slope conductance, which was measured between -90 and -30 mV of - V p, was 42.8 ± 1.1 pS, whereas the outward slope conductance between +30 and +90 mV was 8.5 ± 2.7 pS. These measurements almost fit with the previous results obtained from inside-out patches under the symmetrical high-K + condition ( 28, 29 ). When the bath was replaced with the control solution, the current-voltage curve shifted to the right with the change in reversal potential from +10 to +75 mV ( Fig. 1 B ).

Fig. 1. Current-voltage relationships of the inwardly rectifying K + channel in renal proximal tubule epithelial cells (RPTECs). A : representative current traces recorded with a cell-attached patch at different holding potentials ( V p ). The bath solution used was a high-K + (145 mM) solution, in which K + was substituted for Na +. Dotted line, closed channel level; short thick horizontal bar to the left of a trace, open channel level. B : current-voltage curves were obtained from 4 cell-attached patches, using the high-K + bath solution ( ) and 5 cell-attached patches, using the control bath solution ( ).

We first examined the effect of an NOS inhibitor on the activity of this K + channel in cell-attached patches. Figure 2 A is a representative current trace in response to a nonselective NOS inhibitor, L -NAME. Addition of L -NAME (100 µM) suppressed channel activity ( Fig. 2 A ), which was reversed by an NO donor, SNP, at 10 µM ( Fig. 2 B ). Because NO is known to stimulate soluble guanylate cyclase and to elevate the cytosolic cGMP level, the effect of a membrane-permeant cGMP analog, 8-BrcGMP, was also examined in the presence of L -NAME. As shown in Fig. 2 C, the channel activity suppressed by L -NAME (100 µM) was restored by the subsequent addition of 8-BrcGMP at 100 µM. Summarized data are shown in Fig. 2 D. L -NAME significantly reduced channel activity to 37.8 ± 8.1% of the control, which was restored to the levels near the initial control values by SNP or 8-BrcGMP. These results strongly suggested that the inhibitory effect of L -NAME would primarily be induced by the depletion of intracellular NO. It was also suggested that the depletion of NO might result in decreased cGMP production.

Fig. 2. Representative current traces showing reversible suppression of the channel by N -nitro- L -arginine methyl ester ( L -NAME; A ) and reactivation of the suppressed channel by either a nitric oxide (NO) donor, sodium nitroprusside (SNP; B ), or a membrane-permeant analog of cGMP, 8-bromo-cGMP (8-BrcGMP; C ). These traces were obtained from separate cell-attached patches. L -NAME and 8-BrcGMP were used at 100 µM. The dose of SNP was 10 µM. NP o values calculated from these current traces were 0.77-0.88 for the control, 0-0.12 for L -NAME, 0.91 for L -NAME + SNP, and 0.78 for L -NAME + 8-BrcGMP. D : summary of the effects of L -NAME alone ( n = 18), L -NAME + SNP ( n = 5), and L -NAME + 8-BrcGMP ( n = 7). NP o,e and NP o,c, channel activities under experimental and control conditions, respectively; N, maximum no. of channels observed in the patch; P o, open probability. **Significantly different ( P < 0.01) compared with respective initial control levels.

A specific inhibitor of soluble guanylate cyclase, ODQ (10 µM), was also tested. As shown in Fig. 3 A, ODQ suppressed channel activity, which could not be reversed by 10 µM SNP. In contrast, 8-BrcGMP (100 µM) reactivated the channel, even when ODQ was present in the bath ( Fig. 3 B ). Data on the effects of ODQ are summarized in Fig. 3 C. Channel activity was reduced significantly to 23.3 ± 3.5% of the control by ODQ. Because this suppressive effect was reversed by 8-BrcGMP, but not by SNP, it is likely that NO stimulates soluble guanylate cyclase and elevates intracellular cGMP, which leads channel activation.

Fig. 3. Representative current traces showing that the 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one (ODQ)-induced suppression of the channel was not reversed by SNP ( A ), whereas it was reversed by 8-BrcGMP ( B ). Each current trace was recorded with separate cell-attached patches. ODQ and SNP were used at 10 µM. The dose of 8-BrcGMP was 100 µM. NP o values calculated from these current traces were 0.83-0.86 for the control, 0.08-0.14 for ODQ, 0 for ODQ + SNP, and 0.88 for ODQ + 8-BrcGMP. C : summary of the effects of ODQ alone ( n = 14), ODQ + SNP ( n = 7), and ODQ + 8-BrcGMP ( n = 7). **Significantly different ( P < 0.01) compared with respective initial control levels.

In the next step, we examined whether the cGMP-dependent action of NO had some relation to PKG-mediated phosphorylation processes. Figure 4 A is a representative current trace showing that a PKG-specific inhibitor, KT-5823 (1 µM), suppressed channel activity and that 10 µM SNP could not reactivate the channel. Figure 4 B shows that 100 µM 8-BrcGMP also failed to reactivate the channel in the presence of KT-5823 (1 µM). As reported previously, PKA-mediated phosphorylation also upregulates this K + channel ( 29 ). In fact, a PKA-specific inhibitor, KT-5720 (500 nM), suppressed channel activity, as shown in Fig. 4 C. However, 10 µM SNP reactivated the channel in the presence of 500 nM KT-5720 ( Fig. 4 C ). This is in good agreement with the previous observation that 8-BrcGMP restored the suppressed channel activity by KT-5720 ( 28 ). Summarized data are shown in Fig. 4 D. KT-5823 significantly reduced channel activity to 11.7 ± 6.9% of the control. The suppressed level of channel activity lasted after the addition of SNP or 8-BrcGMP. In contrast, the channel activity suppressed by KT-5720 to 29.6 ± 12.1% of the control recovered after the addition of SNP. These results indicated that the PKG-mediated, but not PKA-mediated, phosphorylation is indispensable for the NO-induced channel activation.

Fig. 4. Representative current traces showing that KT-5823 suppressed channel activity, which could not be reactivated by either SNP ( A ) or 8-BrcGMP ( B ) and that the channel suppressed by KT-5720 was reactivated by SNP ( C ). Each current trace was recorded with separate cell-attached patches. KT-5823, SNP, 8-BrcGMP, and KT-5720 were used at 1, 10, and 100 µM and 500 nM, respectively. NP o values calculated from these current traces were 0.88-1.83 for the control, 0 for KT-5823, 0 for KT-5823 + SNP/8-BrcGMP, 0.89 for K-T5720, and 1.71 for KT-5720 + SNP. D : summary of the effects of KT-5823 alone ( n = 14), KT-5823 + SNP ( n = 7), KT-5823 + 8-BrcGMP ( n = 7), KT-5720 alone ( n = 6), and KT-5720 + SNP ( n = 6). **Significantly different ( P < 0.01) compared with respective initial control levels.

In another series of experiments, we further examined the stimulatory effects of NO donors on channel activity. As shown in Fig. 5 A, SNP added at 10 µM to the bath caused a reversible activation of the channel. This effect of 10 µM SNP was abolished by the subsequent addition of 1 µM KT-5823 ( Fig. 5 B ). Channel activation with another NO donor, SNAP, used at 10 µM was also abolished by 1 µM KT-5823 ( Fig. 5 C ). Data are summarized in Fig. 5 D. Both SNP and SNAP alone caused about twofold increases in channel activity compared with the respective controls, which was almost completely abolished by KT-5823. These results again confirmed that the NO-induced activation of the channel is dependent on the PKG-mediated phosphorylation.

Fig. 5. Representative current traces showing that SNP reversibly activated the channel ( A ) and that KT-5823 abolished the channel activation induced by SNP ( B ) or S -nitroso- N -acetyl- D, L -penicillamine (SNAP; C ). Each current trace was recorded with separate cell-attached patches. NO donors were used at 10 µM. The dose of KT-5823 was 1 µM. NP o values calculated from these current traces were 0.19-0.84 for the control, 0.86-1.67 for SNP, 0.21 for SNP + KT-5823, 1.82 for SNAP, and 0.83 for SNAP + KT-5823. D : summary of the effects of SNP alone ( n = 10), SNP + KT-5823 ( n = 5), SNAP alone ( n = 8), and SNAP + KT-5823 ( n = 5). * P < 0.05 and ** P < 0.01, significantly different compared with respective initial control levels.

In contrast to the stimulatory effect of SNP at 10 µM, a higher dose (1 mM) of SNP (high SNP) caused inhibition of channel activity ( Fig. 6 A ). Addition of 100 µM 8-BrcGMP to the bath could not relieve the channel suppression induced by high SNP ( Fig. 6 B ). Data on the effects of high SNP are summarized in Fig. 6 C. Channel activity was reduced significantly to 32.5 ± 5.2% of the control by high SNP. The reduced channel activity continued after addition of 8-BrcGMP, suggesting that high SNP impaired the functions of PKG and/or the channel protein itself. Although current traces are not shown, SNP had no significant effect on channel activity in inside-out patches at 10 µM ( NP o,e / NP o,c = 0.99 ± 0.04, n = 8) and 1 mM ( NP o,e / NP o,c = 0.92 ± 0.14, n = 8).

Fig. 6. Suppressive effect of a high dose of SNP. A : representative current trace showing reversible suppression of the channel by high SNP. B : ineffectiveness of 8-BrcGMP to reactivate the channel in the presence of high SNP. Each current trace was recorded with separate cell-attached patches. SNP and 8-BrcGMP were used at 1 mM and 100 µM, respectively. NP o values calculated from these current traces were 0.89-1.81 for the control, 0.13-0.23 for high SNP, and 0.13 for high SNP + 8-BrcGMP. C : summary of the effects of high SNP alone ( n = 12) and high SNP + 8-BrcGMP ( n = 8). **Significantly different ( P < 0.01) compared with respective initial control levels.

We previously reported that hANP increased the activity of this K + channel through cGMP/PKG-mediated processes ( 28 ). Although ANP is believed to act on its target cells through activation of particulate guanylate cyclase, a few reports suggest that ANP stimulated NO release from human proximal tubular cells ( 5, 24 ). Therefore, it is possible that the effects of hANP we previously observed were mediated by the NO pathway. To address this issue, the effect of hANP was reevaluated in the presence of L -NAME or ODQ. Channel activity, which was reduced significantly ( P < 0.01) by 100 µM L -NAME ( NP o,e / NP o,c = 0.24 ± 0.09, n = 6) or by 10 µM ODQ ( NP o,e / NP o,c = 0.30 ± 0.09, n = 7), was recovered to its initial control level ( NP o,e / NP o,c = 1.15 ± 0.17 or 0.87 ± 0.18) by the following addition of hANP at 20 nM (data not shown). These results indicate that the action of hANP is independent of NO-mediated processes.

Finally, we examined effects of NO reagents on the membrane potential, using the slow whole cell configuration. As shown in Fig. 7 A, L -NAME added to the bath at 100 µM depolarized the cell membrane. In 11 experiments, the membrane potential of -64.4 ± 1.7 mV in control conditions was depolarized to -54.3 ± 2.7 mV (depolarization by 10.1 ± 2.6 mV, P < 0.01). In contrast, SNP of 10 µM (low SNP) caused hyperpolarization ( Fig. 7 B ). The value of membrane potential in control conditions was -59.4 ± 1.9 mV ( n = 9), which was hyperpolarized by low SNP to -66.5 ± 1.8 mV (hyperpolarization by 7.1 ± 0.7 mV, P < 0.05). However, high SNP caused depolarization ( Fig. 7 C ). The value ( n = 10) was changed from -68.0 ± 2.3 to -58.8 ± 2.5 mV (depolarization by 9.2 ± 1.0 mV, P < 0.05). In all cases, the membrane potential became almost zero when the control bath solution was replaced with the high-K + solution ( Fig. 7, A-C ), indicating that the potential formation would largely be dependent on K + conductance.

Fig. 7. Representative recordings of membrane potential ( V m ) showing depolarization by 100 µM L -NAME ( A ), hyperpolarization by 10 µM SNP (low SNP; B ), and depolarization by 1 mM SNP (high SNP; C ). Each trace was obtained from separate patches of slow whole cell configuration. Short bar on the top of each trace, high-K + (145 mM) bath solution.

DISCUSSION

The present study has revealed that NO modulates the activity of the inwardly rectifying K + channel in cultured human proximal tubule cells. Although many investigators so far reported that NO affected renal tubular functions ( 15, 37 ), including the K + channel activity in the distal nephron segments ( 9, 21 - 23, 27, 35 ), the mode of NO action on K + channels in proximal tubule cells remained to be elucidated.

Because an NOS inhibitor suppressed the activity of the inwardly rectifying K + channel in RPTECs, it seems likely that the endogenous production of NO in these cells would play a role in the regulation of this channel. The suppressive effect of the NOS inhibitor was reversed by the subsequent addition of an NO donor or 8-BrcGMP. An inhibitor of soluble guanylate cyclase also suppressed channel activity, which was reversed by 8-BrcGMP but not by the NO donor. Furthermore, the stimulatory effects of NO donors were abolished by a PKG-specific inhibitor, but not by a PKA inhibitor. Thus it can be concluded that NO stimulated channel activity through the activation of soluble guanylate cyclase, consequent elevation of intracellular cGMP, and PKG-mediated phosphorylation. Previously, the direct effect of internal SNP on K + channel activity was demonstrated in rat kidney cortical collecting duct ( 12 ). However, we confirmed with inside-out patches that cytoplasmic SNP did not alter the K + channel activity in RPTECs.

Although we did not investigate which NOS isoform(s) would actually be involved, it has been reported that iNOS mRNA was expressed in proximal tubule cells ( 1, 19, 25 ). In contrast to nNOS and eNOS, the activity of iNOS is independent of intracellular Ca 2+, and its mRNA expression is enhanced by lipopolysaccharide or cytokines ( 19, 37, 38 ). Interestingly, several investigators reported that ANP affected NO production probably by modulating expression of iNOS mRNA in human proximal tubule cells ( 5, 24 ) or mouse macrophages ( 13, 14 ). However, the present study showed that hANP activated the channel even when the NOS inhibitor or ODQ was present in the bath. Therefore, it is likely that the majority of the stimulatory effect of hANP on channel activity would be independent of the NO pathway under our experimental conditions, even though these observations may not necessarily negate the possible interaction between hANP and NO.

Using the slow whole cell technique, we obtained values of -76 to -52 mV for membrane potential in control conditions of RPTECs. The membrane potential became almost zero when the bath was replaced with the high-K + solution. This observation indicates that the cell membrane of RPTECs is highly K + conductive. It is generally accepted that the K + conductance of the cell membrane largely depends on the K + channel activity. Thus the changes in K + channel activity could alter the membrane potential. Indeed, the NOS inhibitor and the high SNP, which reduced channel activity, depolarized the membrane, whereas the low SNP, which stimulated channel activity, hyperpolarized it. However, the relative changes in membrane potential induced by these substances were much smaller than those in single channel activity. One of the plausible explanations for this discrepancy is that other types of K + channels, which are not affected by NO, are involved in formation of the membrane potential. It has recently been reported that another inwardly rectifying K + channel, Kir7.1, was expressed in the human proximal tubule ( 7 ). Although Kir7.1 is hardly detectable at the single channel level because of its small conductance ( 50 fS; see Ref. 16 ) and the effect of NO on its activity is still unknown, the K + channels with such small conductance may interfere with the changes in the steady-state membrane potential. Further experiments will be necessary to clarify this issue. Anyway, even though the small-conductance K + channels, such as Kir7.1, might be important in maintaining the potential formation, it can be concluded that the membrane potential in RPTECs is modulated, at least in part, by the 40-pS K + channel activity, as observed in this study. It is also conceivable that regulation of membrane potential by several types of K + channels might be beneficial for the fine tuning of cellular functions.

In previous reports, some investigators demonstrated the stimulatory effect of NO on the proximal tubular fluid and Na + transport ( 3, 33, 34 ), whereas others showed the inhibitory effect ( 10, 20, 32 ). The diversity of NO action on the proximal tubular cell might partly be caused by the biphasic action of NO ( 19, 22 ). Our experiments using NO donors have also provided evidence that the effect of NO on channel activity is dose dependently biphasic. Namely, a relatively low dose of NO stimulates the channel activity, whereas a high dose of NO suppresses it. With regard to the NO-induced suppression, some reaction sites after cGMP production would be impaired, since application of 8-BrcGMP could not restore the channel activity. It has been demonstrated that NO reacts not only with soluble guanylate cyclase but also with superoxide (O 2 - ), the latter forming a potent oxidant, peroxynitrite (OONO - ), which induces oxidation or thiol nitrosylation of various proteins ( 4 ). Although O 2 - is scavenged by superoxide dismutase (SOD), an excess amount of NO competes with this enzyme for O 2 - and enhances the formation of OONO -, impairing protein functions ( 4 ). Furthermore, SNP was reported to inhibit SOD ( 26 ). Thus it is likely that the suppressive effect of SNP on channel activity at a high dose involves accelerated formation of OONO -.

The endogenous NO would not be so high as to induce excessive OONO - under our experimental conditions, since the low dose of NO donors stimulated the activity of the inwardly rectifying K + channel in RPTECs. The relatively low concentration of endogenous NO would contribute to the maintenance of basal activity of this K + channel. According to the report by Chatterjee et al. ( 5 ), nitrite production in the primary culture of human proximal tubule cells was <5 nmol·mg protein -1 ·24 h -1 under the basal condition, whereas it exhibited a 10-fold increase after treatment with cytokines. They also demonstrated that the plasma level of nitrite/nitrate significantly increased in rats subjected to bilateral renal ischemia ( 6 ). Therefore, immune-inflammatory responses or some tissue damage might be involved in the NO-mediated modulation of K + channel activity. In relation to this issue, it has been reported that the K + channel blockers, such as tetraethylammonium and glibenclamide, reduced hypoxic injury of the proximal tubule ( 31 ). Although the high concentration of NO would have some toxic effects, as mentioned above ( 4 ), it is conceivable that the high NO might also protect the cells from hypoxic injury by inhibiting K + channel activity in proximal tubule cells. Further studies are required to clarify this hypothesis.

In summary, NO affects the activity of an inwardly rectifying K + channel in cultured human proximal tubule cells. The production of NO in these cells would be relatively low under the control condition and involved in the maintenance of basal channel activity through PKG-mediated phosphorylation. In contrast, an excess amount of NO suppresses channel activity by impairing some protein functions, which are independent of the cGMP/PKG pathway.

ACKNOWLEDGMENTS

We thank Dr. Y. Sohma (Osaka Medical College, Osaka, Japan) for valuable discussion on preparing this manuscript.

【参考文献】
  Ahn KY, Mohaupt MG, Madsen KM, and Kone BC. In situ hybridization localization of mRNA encoding inducible nitric oxide synthase in rat kidney. Am J Physiol Renal Fluid Electrolyte Physiol 267: F748-F757, 1994.

Alderton WK, Cooper CE, and Knowles RG. Nitric oxide synthases: structure, function and inhibition. Biochem J 357: 593-615, 2001.

Amorena C and Castro AF. Control of proximal tubule acidification by the endothelium of the peritubular capillaries. Am J Physiol Regul Integr Comp Physiol 272: R691-R694, 1997.

Beckman JS and Koppenol WH. Nitric oxide, superoxide, and peroxynitrite: the good, the bad, and ugly. Am J Physiol Cell Physiol 271: C1424-C1437, 1996.

Chatterjee PK, Hawksworth GM, and McLay JS. Cytokine-stimulated nitric oxide production in the human renal proximal tubule and its modulation by natriuretic peptides: a novel immunomodulatory mechanism? Exp Nephrol 7: 438-448, 1999.

Chatterjee PK, Patel NS, Sivarajah A, Kvale EO, Dugo L, Cuzzocrea S, Brown PA, Stewart KN, Mota-Filipe H, Britti D, Yaqoob MM, and Thiemermann C. GW274150, a potent and highly selective inhibitor of iNOS, reduces experimental renal ischemia/reperfusion injury. Kidney Int 63: 853-865, 2003.

Derst C, Hirsch JR, Preisig-Muller R, Wischmeyer E, Karschin A, Doring F, Thomzig A, Veh RW, Schlatter E, Kummer W, and Daut J. Cellular localization of the potassium channel Kir7.1 in guinea pig and human kidney. Kidney Int 59: 2197-2205, 2001.

Giebisch G. Renal potassium channels: function, regulation, and structure. Kidney Int 60: 436-445, 2001.

Gu RM, Wei Y, Jiang HL, Lin DH, Sterling H, Bloom P, Balazy M, and Wang WH. K depletion enhances the extracellular Ca 2+ -induced inhibition of the apical K channels in the mTAL of rat kidney. J Gen Physiol 119: 33-44, 2002.

Guzman NJ, Fang MZ, Tang SS, Ingelfinger JR, and Garg LC. Autocrine inhibition of Na + /K + -ATPase by nitric oxide in mouse proximal tubule epithelial cells. J Clin Invest 95: 2083-2088, 1995.

Hanigan MH and Frierson HF Jr. Immunohistochemical detection of gamma-glutamyl transpeptidase in normal human tissue. J Histochem Cytochem 44: 1101-1108, 1996.

Hirsch JR, Cermak R, Forssmann WG, Kleta R, Kruhoffer M, Kuhn M, Schafer JA, Sun D, and Schlatter E. Effects of sodium nitroprusside in the rat cortical collecting duct are independent of the NO pathway. Kidney Int 51: 473-476, 1997.

Kiemer K and Vollmar AM. Autocrine regulation of inducible nitric-oxide synthase in macrophages by atrial natriuretic peptide. J Biol Chem 273: 13444-13451, 1998.

Kone BC. Molecular biology of natriuretic peptides and nitric oxide synthases. Cardiovasc Res 51: 429-441, 2001.

Kone BC and Baylis C. Biosynthesis and homeostatic roles of nitric oxide in the normal kidney. Am J Physiol Renal Physiol 272: F561-F578, 1997.

Krapivinsky G, Medina I, Eng L, Krapivinsky L, Yang Y, and Clapham DE. A novel inward rectifier K + channel with unique pore properties. Neuron 20: 995-1005, 1998.

Kubokawa M, Mori Y, and Kubota T. Modulation of inwardly rectifying ATP-regulated K + channel by phosphorylation process in opossum kidney cells. Jpn J Physiol 47: 111-119, 1997.

Kubokawa M, Nakaya S, Yoshioka Y, Nakamura K, Sato F, Mori Y, and Kubota T. Activation of inwardly rectifying K + channel in OK proximal tubule cells involves cGMP-dependent phosphorylation process. Jpn J Physiol 48: 467-476, 1998.

Liang M and Knox FG. Production and functional roles of nitric oxide in the proximal tubule. Am J Physiol Regul Integr Comp Physiol 278: R1117-R1124, 2000.

Linas SL and Repine JE. Endothelial cells regulate proximal tubule epithelial cell sodium transport. Kidney Int 55: 1251-1258, 1999.

Lu M, Giebisch G, and Wang W. Nitric oxide links the apical Na + transport to the basolateral K + conductance in the rat cortical collecting duct. J Gen Physiol 110: 717-726, 1997.

Lu M and Wang WH. Reaction of nitric oxide with superoxide inhibits basolateral K + channels in the rat CCD. Am J Physiol Cell Physiol 275: C309-C316, 1998.

Lu M, Wang X, and Wang W. Nitric oxide increases the activity of the apical 70-pS K + channel in TAL of rat kidney. Am J Physiol Renal Physiol 274: F946-F950, 1998.

McLay JS, Chatterjee PK, Jardine AG, and Hawksworth GM. Atrial natriuretic factor modulates nitric oxide production: an ANF-C receptor-mediated effect. J Hypertens 13: 625-630, 1995.

McLay JS, Chatterjee P, Nicolson AG, Jardine AG, McKay NG, Ralston SH, Grabowski P, Haites NE, MacLeod AM, and Hawksworth GM. Nitric oxide production by human proximal tubular cells: a novel immunomodulatory mechanism? Kidney Int 46: 1043-1049, 1994.

Misra HP. Inhibition of superoxide dismutase by nitroprusside and electron spin resonance observations on the formation of a superoxide-mediated nitroprusside nitroxyl free radical. J Biol Chem 259: 12678-12684, 1984.

Muto S, Asano Y, Wang W, Seldin D, and Giebisch G. Activity of the basolateral K + channels is coupled to the Na + -K + -ATPase in the cortical collecting duct. Am J Physiol Renal Physiol 285: F945-F954, 2003.

Nakamura K, Hirano J, Itazawa S, and Kubokawa M. Protein kinase G activates inwardly rectifying K + channel in cultured human proximal tubule cells. Am J Physiol Renal Physiol 283: F784-F791, 2002.

Nakamura K, Hirano J, and Kubokawa M. An ATP-regulated and pH-sensitive inwardly rectifying K + channel in cultured human proximal tubule cells. Jpn J Physiol 51: 523-530, 2001.

Palmer LG. Renal ion channels. In: Handbook of Physiology: Renal Physiology. Bethesda, MD: Am Physiol Soc, 1992, vol. I, chapt. 16, p.715-738.

Reeves WB and Shah SV. Activation of potassium channels contributes to hypoxic injury in proximal tubules. J Clin Invest 94: 2289-2294, 1994.

Roczniak A and Burns KD. Nitric oxide stimulates guanylate cyclase and regulates sodium transport in rabbit proximal tubule. Am J Physiol Renal Fluid Electrolyte Physiol 270: F106-F115, 1996.

Wang T. Nitric oxide regulates HCO 3 - and Na + transport by a cGMP-mediated mechanism in the kidney proximal tubule. Am J Physiol Renal Physiol 272: F242-F248, 1997.

Wang T. Role of iNOS and eNOS in modulating proximal tubule transport and acid-base balance. Am J Physiol Renal Physiol 282: F658-F662, 2002.

Wei Y and Wang W. Angiotensin II stimulates basolateral K channels in rat cortical collecting ducts. Am J Physiol Renal Physiol 284: F175-F181, 2003.

Welling PA. Cross-talk and the role of K ATP channels in the proximal tubule. Kidney Int 48: 1017-1023, 1995.

Wilcox CS. L -Arginine-nitric oxide pathway. In: The Kidney: Physiology and Pathophysiology. Philadelphia, PA: Williams & Wilkins, 2000, vol. I, chapt. 32, p. 849-871.

Wu G and Morris SM Jr. Arginine metabolism: nitric oxide and beyond. Biochem J 336: 1-17, 1998.

作者单位：Department of Physiology II, Iwate Medical University School of Medicine, Morioka, 020-8505 Japan