INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

WE HAVE PREVIOUSLY SHOWN that addition of 5 µM arachidonic acid (AA) inhibits ROMK1 channel activity. The effect of AA was specific, since other fatty acids failed to mimic the effect of AA. In addition to ROMK1, two additional ROMK channels have been cloned in the rat (and human) kidney, termed ROMK2 and ROMK3, which are also localized to the distal nephron and cortical collecting duct (CCD) (3, 15, 24, 32). The amino acid sequences of the ROMK channels are identical with the exception of the NH₂ terminus, which varies in both length and initial amino acid sequence, the result of alternate splicing of the 5' end of the precursor nucleotide (Fig. 1). ROMK2 lacks the 19-amino acid sequence of ROMK1, which contains two putative protein kinase C (PKC) phosphorylation sites at serine-4 and threonine-17. ROMK3 has a unique aa NH₂ terminus which lacks the serine-4 putative PKC phosphorylation site but maintains one potential PKC site at threonine-24. The distribution of the remaining putative PKC sites, as well as the three identified protein kinase A (PKA) phosphorylation sites are identical (31). Distribution of the ROMK channels in the kidney also varies (3). ROMK1 is found only in the medullary collecting duct and CCD, ROMK2 is found in all nephron segments from the medullary thick ascending limb (TAL) to the CCD, and ROMK3 is localized to the medullary and cortical TAL and distal convoluted tubule. The aims of this study were to determine the effect of AA on the two alternatively spliced variants of ROMK and to further elucidate the mechanism by which AA exerts its effect on channel activity.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Preparation of Xenopus oocytes. Oocytes were isolated as previously described (16). Viable oocytes were selected and microinjected with ROMK cRNA as follows. cRNA was made by in vitro transcription from ROMK1 cDNA, which was subcloned into the Xho I and Bgl II multicloning site of the pSPORT vector, flanked by SP6 and T7 RNA polymerase promoters (Life Technologies, Grand Island, NY). A micropipette with a tip of 10-20 µm was filled with paraffin oil and attached to a manipulator (Narishige). An aliquot of cRNA was heated to 65°C for 2 min and subsequently loaded by suction into the tip of the micropipette, and 5 ng cRNA/oocyte was injected. The oocytes were then incubated at 19°C, and patch-clamp experiments were performed on days 3-7 following injection.

Methods for mutagenesis. To mutate ROMK1 serine-4 to alanine-4 (R1S4A), competent CJ 236 cells were first transformed with ROMK1-pSport and infected with helper phages (M13K07). Single-stranded DNA containing Uracils was prepared and hybridized with a kinased mutagenic primer. The primer was then extended with T4 polymerase and ligated with T4 ligase to generate heteroduplex DNA, which was used to transform DH5 cells. Mutants were screened by sequencing. The truncated ROMK1 (R1N37D), in which residues 2-37 were deleted, was generated by a polymerase chain reaction reaction. ROMK1 was used as the template. Primer 1 (5' GGA ATT CGC CAC CAT GCA ACG GGC AAG GCT GG 3') contained an EcoR I site, a Kozak sequence, and the initiation codan ATG, followed by ROMK1 sequence. Primer 2 was located downstream of the unique Bgl II site. The amplified product was double digested with EcoR I and Bgl II and was used to replace the EcoR I-Bgl II fragment of ROMK1-pSport. The mutants were screened with polymerase chain reactions, and the authenticity of the constructs was confirmed by dideoxy sequencing, using sequenase version II (UBS).

Experimental media. The standard pipette solution and the bath solution used for cell-attached patches were 150 mM KCl, 1.0 mM MgCl₂, and 5.0 mM N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES), pH 7.4. Before inside-out patches were formed, the bath solution was switched to the standard Mg²⁺-free bath solution, which contained 150 mM KCl, 5.0 mM EDTA, and 5.0 mM HEPES, pH 7.4. To maintain the channel activity in inside-out patches, 200 µM MgATP and PKA catalytic subunit (50 nM/ml) were added to the bath. Stock solutions of 10 mM MgATP were made using standard bath solution, and pH was adjusted to 7.4. The catalytic subunit of PKA (39 picomolar units/µg protein) (Promega, Madison, WI) was diluted in standard bath solution. AA was purchased from LC Laboratories (Woburn, MA) and dissolved in ethanol. Calphostin C and chelerythrine chloride were purchased from LC Laboratories and dissolved in ethanol or dimethyl sulfoxide. We did not observe any effect of the solvents on channel activity at the final concentration used (~0.01%). The purified PKC (a mixture of -, -, and -isoforms) was obtained from Biomol (Plymouth Meeting, PA). All experiments were conducted at room temperature.

Patch-clamp technique. Patch-clamp electrodes were pulled from glass capillary tubes (Dagan, Minneapolis, MN), with an expected resistance of 4-6 M when filled with 140 mM KCl. The vitelline membrane was removed from oocytes with watchmaker's forceps by placing oocytes in a hypertonic solution. The pipette was placed against the plasma membrane of the oocyte, and a slight negative pressure was applied, resulting in a high resistance gigaohm seal between the micropipette and the membrane. Recordings were made using a List L/M-EPC7 patch-clamp amplifier (List Medical, Darmstadt, Germany), and single-channel currents were low-pass filtered at 1 kHz by an eight-pole Bessel filter (902LPF; Frequency Devices, Haverhill, MA). The recordings were digitized at a sampling rate of 44 kHz, using a modified digital data recorder (Instrutech, Great Neck, NY) and were stored on videotape (JVC HR-J400U). For analysis, data were acquired and stored on hard disk (Gateway 2000 4DX) at a sampling rate of 5 kHz and analyzed using the pCLAMP software system (version 6.0; Axon Instruments, Burlington, CA). Single-channel slope conductance was calculated by recording single-channel currents at a number of different holding potentials. We have selected the patches with only one or two levels of the channel current to calculate the channel open probability (P_o). Channel number (N) was determined by the maximum numbers of superpositions of current level observed over at least 5-min recording time. Because ROMK channels have a high P_o, the current level observed in 5-min recording time should be the number of the functional K⁺ channels in a given patch. P_o of individual single K⁺ channels was determined as follows where t_n is the fractional open time spent at each of the observed current levels.

We define channel activity as NP_o, a product of N and P_o. The NP_o was calculated from data samples of 60-s duration in the steady state. However, if the channel activity decreased close to zero, we shortened the recording time to 30 s to increase probability seeing a wash-out effect. The NP_o was calculated as follows We did not make efforts to determine whether a change in NP_o was the result of alteration of N or P_o. Data are means ± SE. Where appropriate, Student's t-test for paired data was used to assess significance of difference.

Protein kinase C assay. PKC activity was assessed using a commercially available kit from Life Technologies (Gaithersburg, MD), as described (20). The presence of a membrane-bound PKC was addressed by the measurement of PKC in an oocyte membrane preparation using a specific substrate peptide, N-acetylated myelin basic protein fragment 4-14 in the presence of [-³²P]ATP. Phosphorylation of this substrate was measured by ³²P incorporation, and PKC specificity was confirmed by using the PKC pseudosubstrate inhibitor peptide, PKC-(19-36), which acts as a potent inhibitor for this substrate. The specific substrate peptide, N-acetylated myelin basic protein fragment 4-14 (0.05 mmol/l), was combined with the membrane preparation of oocytes (0.075 to 75.0 mg/ml protein) in tris(hydroxymethyl)aminomethane buffer (20 mmol/l, pH 7.5) containing MgCl₂ (20 mmol/l), CaCl₂ (1 mmol/l), adenosine triphosphate (0.02 mmol/l), and [-³²P]ATP [0.625 µCi/assay tube; sp act, 3,000 Ci/mmol (DuPont-NEN)] in the presence of lipid reagent [0.01 mmol/l phorbol 12-myristate 13-acetate (PMA), 0.28 mg/ml phosphatidylserine, and Triton X-100 mixed micelles] or of pseudosubstrate peptide PKC-(19-36) and incubated for 5 min at 30°C. Samples were run in parallel with membranes prepared from aortic smooth muscle obtained from hypertensive rats as a positive control. Phosphorylation of the substrate peptide was assessed by scintillation spectrophotometry of aliquots of incubate that had been placed on phosphocellulose discs and washed with phosphoric acid and water. PKC specific activity was defined as the difference between activity in the presence of lipid reagent vs. that in the presence of inhibitory peptide and was expressed as picomoles of incorporated phosphate per minute per milligram of protein. Protein measurements were determined using the Bio-Rad protein assay with bovine serum albumin as a standard.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Biophysical and kinetic characteristics of ROMK isoforms. We confirmed the previous finding that AA inhibits the activity of the ROMK1 channel with a K_i of 2.5 µM (16). When we examined the effect of AA, PKA was used to prevent channel rundown. We have previously shown that AA had no effect on the activity of exogenous PKA (16). We had extended our study to examine the effect of AA on other ROMK homologs, ROMK2 and ROMK3. First, we investigated whether the structural differences, which existed at the NH₂ termini of ROMK1, ROMK2, and ROMK3, resulted in differences in their respective single-channel kinetics. Figure 1A is a representative single-channel tracing of ROMK1, ROMK2, and ROMK3 in an inside-out patch at a membrane potential of 50 mV. By inspection of Fig. 1A, it is apparent that ROMK1, ROMK2, and ROMK3 are all high P_o channels, like those described for the renal low-conductance K⁺ channels (6, 26, 29, 30). In addition, ROMK1, ROMK2, and ROMK3 each displayed a modest inward rectification, characteristic of the native low-conductance K⁺ channel (Fig. 1B) (29). The current-voltage relationship (I-V) yielded an inward slope conductance of 32 ± 2, 30 ± 2, and 28 ± 1 pS, respectively.

View larger version (26K): [in this window] [in a new window]   Fig. 1.   A: single-channel recordings of ROMK isoforms in 150 mM symmetrical KCl solution at a membrane potential of 50 mV. C, closed state. B: current-voltage curves of ROMK1 (), ROMK2 (), and ROMK3 () for K+ permeation in symmetrical KCl solutions (150 mM). Data are expressed as means ± SE. Bottom: amino acid differences in NH2 termini of ROMK homologs. * Putative PKC phosphorylation sites.

Histograms were likewise constructed to determine the open and closed durations of ROMK1, -2, and -3, as indicated in Fig. 2. Single-channel inside-out recordings revealed two open states and one closed state. Thus the kinetic behavior of the ROMK channels revealed a transition between a longer and a shorter open state (25 ± 1 and 0.92 ± 0.24 ms) and a short closed state (0.85 ± .07 ms). These data are consistent with those described in the apical K⁺ channel of the rabbit TAL, with two open states with time constants of 33 ± 1 (₁) and 1.5 ± 0.10 ms (₂) and one closed state with a time constant of 1.0 ± 0.20 ms (30). Complete single-channel kinetic data for ROMK1, -2, and -3 are summarized in Table 1.

View larger version (45K): [in this window] [in a new window]   Fig. 2.   Representative histograms of open (left) and closed duration (right) of ROMK1, ROMK2, and ROMK3. Mean individual open (o) and closed (c) times are summarized in Table 1. Arrows indicate each open state.

Effect of AA on ROMK isoforms. Having established that the three cloned ROMK channels, ROMK1, ROMK2 and ROMK3, were indistinguishable based on kinetic criteria, we next extended our study to determine whether the structural differences that exist at the NH₂ termini had any impact on inhibition of channel activity by AA. We previously found that the addition of 5 µM AA completely blocked ROMK1 channel activity. In contrast, 5-10 µM AA reduced the activity of ROMK2 and ROMK3 by only 15 ± 8% inhibition and 13 ± 9%, respectively (Fig. 3), suggesting that the NH₂ terminus is crucial for the inhibitory effect of AA, since this is the only observed difference among these channels.

View larger version (18K): [in this window] [in a new window]   Fig. 3.   Effect of 5-10 µM arachidonic acid (AA) on ROMK channel activity: ROMK1 (, n = 27), ROMK2 (, n = 6), and ROMK3 (, n = 6) in channel recordings obtained in inside-out patches. Addition of AA and washout of bath occurred at times indicated by arrows. Data were normalized using equation (NPo experimental/NPo control) × 100%, where NPo is product of channel number and open probability. * P < 0.05, results are significantly different from the control group in the absence of AA, as determined by paired Student's t-test.

Effect of AA on amino terminal variant of ROMK1. We examined the effect of AA on a ROMK1 variant, R1ND37, in which the initial 2-37 amino acids of the NH₂ terminus were deleted. Figure 4A is a recording made in an inside-out patch showing the single-channel current of R1ND37. It is apparent that the mutant, R1ND37, has similar channel kinetics as the wild type (data not shown). Further analysis showed that R1ND37 had the same single-channel conductance and P_o as that of the wild type (data not shown). Addition of 5 µM AA to inside-out patches of R1ND37 failed to inhibit channel activity, confirming the significance of the NH₂ terminal region of ROMK1 on AA sensitivity (Fig. 4B). Comparison of the NH₂ termini of the ROMK isoforms revealed that ROMK2 and ROMK3 both lack the serine-4 putative PKC phosphorylation site, suggesting this serine-4 phosphorylation site may play a role in the AA-mediated inhibition of ROMK1. To test this hypothesis, the effect of AA was examined on a mutant ROMK1, R1S4A, in which the serine-4 residue was mutated to alanine, using site-directed mutagenesis. Figure 4A shows a typical channel recording of R1S4A in a cell-attached patch. Single-channel analysis showed that R1S4A had the same channel kinetics, P_o, and conductance as ROMK1 (data not shown). However, compared with ROMK1, the elimination of this PKC consensus sequence abolished the effect of AA to the same extent as that observed in ROMK2, ROMK3, and R1ND37 (Table 2), suggesting the importance of serine-4 residue in determination of the AA effect.

View larger version (21K): [in this window] [in a new window]   Fig. 4.   A: single-channel recordings of ROMK mutants, R1ND37 and R1S4A, in 150 mM symmetrical KCl solution at a membrane potential of 60 mV. C, closed state. B: effect of 5 µM AA on a ROMK1 (, n = 6); a ROMK1 variant, R1ND37 (, n = 4), in which the 37-amino acid NH2 terminus was deleted; and on a mutant ROMK1, R1S4A (, n = 6), in which the serine-4 residue was mutated to alanine-4. Channel recordings were obtained in inside-out patches, and addition of AA and washout of bath occurred at times indicated by arrows.

Effect of PKC on ROMK1. Because serine-4 residue is a putative PKC phosphorylation site and AA has been shown to activate PKC (14, 19, 22, 23), it is possible that the effect of AA is the result of stimulation of a membrane-bound PKC which phosphorylates serine-4 residue. It has also been suggested that a pool of inactive but primed PKC, which has been translocated to the membrane, exists and that this pool may be responsive to stimuli, such as fatty acids, resulting in activation of PKC without the necessity of translocation (2). Thus we first examined the effect of PKC on channel activity in inside out patches. Figure 5A is a representative channel recording showing the effect of PKC on the wild type of ROMK1. To activate PKC, we added 200 µM MgATP and 1 µM PMA. Neither MgATP nor PMA had a significant effect on channel activity (data not shown), suggesting that there is no membrane-bound and PMA-sensitive PKC. However, addition of 1 nM exogenous PKC completely inhibited ROMK1 within 100-120 s (Fig. 5A). Longer latent time of the PKC effects than that observed in the rat CCD may result from a different experimental temperature, since the PKC experiments were performed at 37°C in the native CCD, whereas the present experiments were carried out at room temperature. Figure 5B shows that addition of 1 nM PKC reduced NP_o from 2.05 ± 0.58 under control conditions to 0.2 ± 0.2 (n = 4) and that the effect of PKC is reversible. In contrast, 1 nM PKC failed to inhibit activity of the mutant, R1S4A (Fig. 5C), and the NP_o in the presence of PKC was 2.2 ± 0.6 (n = 3) and was not different from the control value (NP_o = 2.3 ± 0.6). However, it was observed that increasing PKC concentration to 4 nM caused a moderate 30% inhibition of R1S4A (data not shown). Therefore, our data indicate that ROMK1 was inhibited by PKC as observed in the native tissue (28) and that serine-4 residue is important for the PKC effect.

View larger version (24K): [in this window] [in a new window]   Fig. 5.   A: channel recording made in an inside-out patch showing effect of 1 nM protein kinase C (PKC) on ROMK1 with Mg2+-free solution in the bath. MgATP (200 µM) and 1 µM phorbol 12-myristate 13-acetate (PMA) were used to activate PKC, and neither MgATP nor PMA had an effect on channel activity (data not shown). Channel closed level is indicated by C, and the cell membrane potential was 40 mV. Top: whole time course of experiment. Bottom: four parts of the trace indicated by 1-3 are extended to show the channel activity at fast time resolution. B and C: summary from 4 experiments showing effects of PKC on NPo of ROMK1 (B) and mutant R1S4A (C). , Channel activity (NPo) obtained from each experiment under control conditions and after application of PKC; solid symbols are means ± SE of the experimental values.

Although we did not observe the effect of PMA, it is possible that AA may stimulate atypical isoforms of PKC, which are not sensitive to PMA. We tested this possibility by pretreatment of cells with 50 nM calphostin C, specific inhibitor of PKC, for 10 min and then followed by the addition of 5 µM AA. Preincubation of the channel with calphostin C had no effect on the ability of AA to inhibit channel activity, since channel activity was similarly inhibited by AA (Fig. 6). We also tested the effect of AA in the presence of chelerythrine, a specific inhibitor of PKC, that is nonspecific for the different PKC isoforms. Unlike calphostin C, chelerythrine inhibits PKC by interacting with the highly conserved catalytic domain in a noncompetitive manner with respect to ATP and in a competitive manner with respect to substrate binding (9). Like calphostin C, chelerythrine had no effect on AA mediated inhibition of ROMK1 (Fig. 6), indicating that addition of exogenous AA did not result in the activation of a PKC isozyme. The notion that oocytes have no membrane-bound PKC is supported by experiments in which no membrane-bound PKC activity in oocytes was found compared with positive control in the presence of lipid reagents (Table 3). Compared with the control (aortic smooth muscle cell membranes from hypertensive rats), there was no detectable PKC activity (0 ± 0 vs. 212 ± 20 pmol ³²P incorporated · min¹ · mg protein¹). In addition, we could still not detect the PKC activity in the presence of 5 µM AA.

View larger version (25K): [in this window] [in a new window]   Fig. 6.   Time course showing effect of 5 µM AA on ROMK1 channel activity alone (, n = 6) in the presence of the PKC inhibitors, 50 nM calphostin C (, n = 4) and 10 µM chelerythrine (, n = 4). Channel recordings were obtained in inside-out patches, and addition of AA and washout of bath occurred at times indicated by arrows.

Because inhibition of PKC failed to abolish the effect of AA, we examined whether AA activates a membrane-bound protein kinase, accordingly phosphorylating the serine-4 residue. Staurosporine at 5 µM has been used as a nonspecific protein kinase inhibitor (1). Therefore, we studied the effect of AA in the presence of 5 µM staurosporine to determine whether inhibition of the protein kinase can abolish the effect of AA (Fig. 7). It is apparent that in the presence of staurosporine, addition of 10 µM AA has no significant effect on channel activity, since channel activity was 85 ± 12% of the control value (n = 5).

View larger version (16K): [in this window] [in a new window]   Fig. 7.   Time course showing effect of 10 µM AA on ROMK1 in the presence of 5 µM staurosporine. Channel activity was recorded in inside-out patches with Mg2+-free bath solution.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

ROMK1 has been proposed as a key component of the low-conductance K⁺ channel of the CCD, an observation based on the similarities between the biophysical properties, intracellular regulators, and the exclusive expression of the channel to the collecting duct (10, 15, 17, 25, 28). In addition, the intracellular signaling pathways that regulate the secretory K⁺ channel of the CCD have been investigated. It was found that AA (K_i = 2.6 µM) could inhibit the native low-conductance K⁺ channel activity in inside-out patches. The effect of AA was not mediated by the AA metabolites, since inhibition of the three metabolic pathways of AA had no impact on the effect of AA (27). Likewise, we have demonstrated that addition of AA (K_i = 2.5 µM) to inside-out patches resulted in reversible inhibition of ROMK1, an effect which was preserved in the presence of inhibitors of the enzymes which metabolize AA (16). These results indicate that, like the native low-conductance K⁺ channel, AA is able to directly inhibit ROMK1 channel activity.

In addition to AA, PKC has been found to play an important role in the regulation of the native low-conductance K⁺ channel (28). In the present study, we show that ROMK1 is also inhibited by PKC. The ROMK1 channel has three putative PKC phosphorylation sites (10, 31); however, the present study strongly suggests that the serine-4 residue plays a critical role for the effect of PKC, since mutation of the serine-4 residue significantly attenuates the response of the channel to PKC. The notion that the serine-4 residue is important for the PKC-induced effects is also supported by the observation of McNicholas et al. (17) that inhibition of PKC activity abolished the effect of raising intracellular calcium on ROMK channel activity and that the calcium effect on channel activity is significantly diminished in the ROMK mutant R1S4A. Interestingly, we observed that increasing PKC concentration to 4 nM can still lead to a moderate inhibition of channel activity by 30% (data not shown), suggesting that phosphorylation of serine-4 residue may be important for facilitating phosphorylation of other PKC sites. We need further experiments to confirm this hypothesis.

In the present study, we show that the serine-4 residue in the NH₂ terminus plays a key role in the determination of the AA effect. First, AA completely blocks ROMK1 channel activity but not that of ROMK2 and ROMK3, which lack the serine-4 residue. Second, the truncated ROMK1 (R1ND37) is not sensitive to AA. Finally, single mutation of the serine-4 site to alanine results in the loss of sensitivity to AA. Therefore, the results indicate that the phosphorylation of the serine-4 residue is involved in mediating the effect of AA. However, three lines of evidence indicate that the effect of AA on ROMK1 channel activity was not mediated by stimulation of a membrane-bound PKC. First, preincubation of ROMK1 with the PKC inhibitors, calphostin C and chelerythrine, had no effect on the ability of AA to inhibit ROMK1 channel activity. Second, addition of PMA failed to mimic the effect of AA. Finally, measurement of membrane-bound PKC activity, using a specific substrate peptide, revealed the absence of PKC activity.

The observation that staurosporine, a nonspecific protein kinase inhibitor, abolished the AA effect suggests that the phosphorylation of serine-4 site might be induced by a membrane bound protein kinase. Although serine-4 residue is located among a consensus sequence for the PKC recognition as a substrate, the sequences do not absolutely predict the kinase responsible for phosphorylation (12). Interestingly, it was reported that protein kinase activity was closely associated with a reconstituted Ca²⁺-activated K⁺ channel (4). Although we were unable to identify the protein kinase responsible for the phosphorylation of serine-4 site, our experimental results strongly indicated the importance of serine-4 site for the AA effects. We have attempted to examine the effect of AA in the absence of MgATP or in the presence of nonhydrolyzable ATP analogs to determine whether blocking channel phosphorylation can abolish the effect of AA. However, channel activity was progressively decreased in the MgATP-free solutions. In addition, it is unlikely that the effect of AA on the native K⁺ channel is mediated by stimulating the same kinase observed in oocytes. However, it is conceivable that AA may block the native K⁺ channel in the rat CCD by phosphorylation of serine-4 site. Thus the main finding of this study is that phosphorylation of serine-4 site is involved in the AA effect. The kinase responsible for the phosphorylation in the oocytes and in the native tissue is still unknown and needs further experiments to identify the enzyme responsible for the effect.

AA and the metabolites of AA have been shown to have a variety of effects on a number of intracellular signaling targets, including PKC (7, 8, 14, 19, 22, 23). AA and oleic acid were recently demonstrated to activate a unique kinase that phosphorylates the large subunit of myosin phosphatase (11). In addition, it has been shown that the AA-mediated effect on channel activity of several different channels was due to activation of PKC (5, 13, 21). Although we have established that the serine-4 residue plays an important role for determination of the AA effect, it is still not known whether the other two putative PKC sites are also involved in mediating the effect of AA. We need further experiments to address this issue.

In conclusion, the ROMK1 channel preserves the properties to be sensitive to AA and to PKC, and the serine-4 residue in the NH₂ terminus plays a key role for the effect of AA and PKC.