Inwardly rectifying K+ current (I Kir) in freshly isolated bovine retinal pigment epithelial (RPE) cells was studied in the whole cell recording configuration of the patch-clamp technique. When cells were dialyzed with pipette solution containing no ATP, I Kir ran down completely in <10 min [half time (t 1/2) = 1.9 min]. In contrast, dialysis with 2 mM ATP sustainedI Kir for 10 min or more. Rundown was also prevented with 4 mM GTP or ADP. When 0.5 mM ATP was used,I Kir ran down by ∼71%. Mg2+ was a critical cofactor because rundown occurred when the pipette solution contained 4 mM ATP but no Mg2+(t 1/2 = 1.8 min).I Kir also ran down when the pipette solution contained 4 mM Mg2+ + 4 mM 5′-adenylylimidodiphosphate (t 1/2 = 2.7 min) or 4 mM adenosine 5′-O-(3-thiotriphosphate) (t 1/2 = 1.9 min), nonhydrolyzable and poorly hydrolyzable ATP analogs, respectively. We conclude that the sustained activity ofI Kirin bovine RPE requires intracellular MgATP and that the underlying mechanism may involve ATP hydrolysis.
- patch clamp
- potassium channels
- adenosine triphosphate
positioned between the distal retina and the choroidal blood supply, the retinal pigment epithelium (RPE) carries out a wide range of functions that are critical to the health and integrity of the adjacent photoreceptor cells. One of these functions is control of the composition and volume of the extracellular fluid surrounding the photoreceptor outer segments through the transport of fluid, ions, and metabolites. K+ channels in the RPE apical and basolateral membranes play a central role in RPE physiology. They not only participate directly in net K+ transport (25) but they also influence the active transport of (14) and Cl− (2, 9) by virtue of their impact on resting membrane potentials (22, 26). Patch-clamp studies on freshly isolated amphibian (31) and human (18) RPE cells have shown that the predominant conductance in the physiological voltage range is an inwardly rectifying K+ conductance. The inward rectification of this K+conductance is relatively weak, such that it supports substantial outward K+ current at potentials positive to the K+ equilibrium potential (E K). Pharmacological studies on the intact toad RPE-choroid preparation indicate that this channel is a major component of the K+conductance in the apical membrane (15), where it can support Na+-K+pump activity by recycling K+.
In other cell types, a wide variety of inwardly rectifying K+ channels have been identified (27), some of which are regulated by intracellular ATP levels. KATP channels, for example, have a relatively high open probability when cell ATP levels are low but become inactivated when the intracellular ATP levels are high (1). Other inwardly rectifying K+channels, such as the cloned channel Kir2.1 (ROMK1; Ref. 23), exhibit rundown in the absence of cytoplasmic ATP as a result of the net dephosphorylation of the channel. In the present study, we show that inwardly rectifying K+conductance in freshly isolated bovine RPE cells requires relatively high concentrations of intracellular ATP for sustained activity and, furthermore, that this requirement may be specific for hydrolyzable nucleotide triphosphate analogs. A preliminary account of this work has been presented previously in abstract form (17).
Solutions. The standard bath solution consisted of (in mM) 135 NaCl, 5 KCl, 10 HEPES, 10 glucose, 1.8 CaCl2, and 1.0 MgCl2 and was titrated to pH 7.4 with NaOH. In experiments where the concentrations of K+, Cs+, or Ba2+ were varied, NaCl was replaced by an equimolar amount of the appropriate Cl− salt. Quinidine was added to Ringer solution as a concentrated DMSO solution; the final concentration of DMSO was <0.1%. Glybenclamide was added directly to the Ringer solution. In some experiments, we added 100 μM Gd3+ to the Ringer solutions to block a noisy, nonselective cation current that was activated in some cells by membrane hyperpolarization. Gd3+ had no obvious effects on either the kinetics or voltage dependence of the inwardly rectifying K+ currents (I Kir). The osmolality of all external solutions was 288 ± 5 mosmol/kgH2O.
The standard pipette solution used in these experiments consisted of (in mM) 30 KCl, 83 potassium gluconate, 10 HEPES, 5.5 EGTA-KOH, 0.5 CaCl2 (free Ca2+ concentration ≈ 20 nM), and 2.0 or 4 MgCl2, titrated to pH 7.2 with KOH. Addition of ATP (K+salt), ADP, adenosine 5′-O-(3-thiotriphosphate) (ATPγS), 5′-adenylylimidodiphosphate (AMP-PNP; Li+ salt), and GTP (Na+ or Tris salt) caused acidification of the pipette solution, requiring further titration to pH 7.2. Mg2+-free internal solution was made by eliminating MgCl2 and substituting 5.5 mM EDTA for EGTA. The concentrations of MgATP, free Ca2+, and free Mg2+ were calculated using a Basic program Calcium (4). In some early experiments, we used an internal solution containing 100 mM potassium gluconate, which in some cells led to the development of a swelling-activated, outwardly rectifying Cl− current (3). Orthovanadate stock solution, prepared by adjusting the pH of a 100 mM NaVO4 solution to pH 10, was stored at 4°C and boiled for 15 min before an aliquot was added to the internal solution. The pH of the diluted orthovanadate solution was then adjusted to pH 7.2 with HCl. All internal solutions had an osmolality of 244 ± 5 mosmol/kgH2O except for the 100 mM potassium gluconate solution, which had an osmolality of 278 ± 5 mosmol/kgH2O.
The cell isolation medium contained (in mM) 135N-methyl-d-glucamine (NMDG) chloride, 5 KCl, 10 HEPES, 3 EDTA-KOH, 10 glucose, and 3 cysteine, as well as 0.2 mg/ml papain (type III), and was titrated to pH 7.4 with NMDG free base.
All chemicals were of reagent grade and were obtained from Sigma Chemical (St. Louis, MO) except for GdCl3, which was obtained from Aldrich Chemical (Milwaukee, WI).
Cell isolation. Adult bovine eyes were obtained from a local abattoir. Eyes were enucleated within 30 min of death and placed on ice. Cells were isolated by enzymatic dispersion as described previously for human RPE (20). Briefly, 5-mm-square pieces of RPE-choroid were dissected from the inferior or superior pigmented region of bovine eyecups and incubated for 5–10 min in the cell isolation medium at room temperature. The tissue was then incubated in normal Ringer solution containing 0.1% BSA for 3 min and finally in normal Ringer solution for another 10 min. This series of incubations was repeated three or four times before dissociating cells by vortexing the tissue in 2–3 ml of Ringer solution. Isolated cells were stored in normal Ringer solution at 4°C for up to 24 h before use.
Electrophysiological methods. Isolated RPE cells were transferred to a continuously perfused Lucite recording chamber (31). Cells selected for recording had a bright appearance under phase-contrast microscopy. All experiments were conducted at room temperature (23–25°C).
Membrane currents were recorded using the conventional whole cell recording technique (12) or the amphotericin B perforated-patch method (30). Pipettes were pulled from 7052 glass tubing (Garner Glass, Claremont, CA) with the use of a multistage programmable puller (Sutter Instruments, San Rafael, CA) and coated with Sylgard (Dow Corning, Midland, MI). Before use, pipette tips were fire polished to resistances in the range 1–5 MΩ. For perforated-patch recording, patch pipettes were front filled with a 200- to 300-μm column of internal solution and then backfilled with internal solution containing amphotericin B. The amphotericin B solution was made just before use by adding 10 μl of a freshly made stock solution (1.5 mg amphotericin B/50 μl DMSO) to 2.5 ml of internal solution to give a final concentration of 120 μg/ml. Series resistance commonly decreased to 25 MΩ or less within 5–15 min after gigaohm seal formation on the basolateral membrane.
Currents were recorded with Axopatch 1D or Axopatch 200 amplifiers (Axon Instruments, Foster City, CA) with the built-in low-pass filter set to 1 kHz unless noted otherwise. Recordings were referenced to a Ag-AgCl electrode separated from the bath by a short column of 150 mM KCl set in 4% agar. Command potentials were generated by software control (pCLAMP, Axon Instruments, Foster City, CA). Signals were digitized on line every 1.2–2.5 ms and stored on the hard drive on a 486 microcomputer for subsequent analysis. Series resistance (R s) and membrane capacitance were calculated from uncompensated capacitative transients (low-pass filter 5 kHz) as described previously (31).R s averaged 13 ± 6 (SD) MΩ (n = 91) for whole cell recordings and 13 ± 11 (n = 6) for perforated-patch recordings and was commonly compensated 50% by amplifier circuitry. Built-in circuits also compensated for pipette and membrane capacitances. The pipette tip potential (−10 mV) was measured against a flowing saturated KCl bridge and was used to correct the apparent membrane potential as described previously (16). Liquid junction potentials at interface between the agar bridge and bath were <3 mV and were ignored.
Virtually every cell exhibited inwardly rectifying and delayed rectifier K+ currents; in addition, ∼24% of the cells also expressed M-type K+ currents (32). The delayed, outwardly rectifying current was routinely inactivated by holding the membrane potential at −10 mV between voltage pulses and ramps. Cells containing M-type K+currents were excluded. In most experiments, we quantified theI Kir by measuring the component of whole cell current that was blocked by 20 mM external Cs+, a near-saturating concentration (see results).
Dose-response relationships. The blocker sensitivity of theI Kir was determined by measuring currents at selected voltages in the presence of various concentrations of blocking ions and fitting the results to the first-order equation Equation 1 whereI K is the component of whole cell current blocked by a particular concentration of blocker ([B]),I max is the component of current blocked by a saturating concentration of blocker, and K D is the concentration at which the block is half maximal. Data for each cell were fitted using curve-fitting algorithms in SigmaPlot (SPSS, Chicago, IL) to obtain the parametersI max andK D.
Time course experiments. After the membrane patch beneath the recording pipette was ruptured, the steady-state current-voltage (I-V) relationship of the cell was measured every 15 s by ramping the membrane voltage from +40 to −160 mV (in some cells, to −110 mV) at a rate of 50 mV/s, with the voltage held at −10 mV between ramps. This voltage-clamp protocol inactivated the delayed rectifier K+ current and elicited currents through inwardly rectifying K+ channels as well as Cl− and Na+-selective channels. The time course of I Kirand residual current (Na+ + Cl− currents) was constructed from this I-V data by sampling currents at specific voltages. ForI Kir, we first determined the reversal potential of the residual current by measuring the zero-current potential of the I-Vcurve obtained in the presence of 20 mM external Cs+, a concentration that blocks nearly all ofI Kir. For the residual current, we sampled the current at the reversal potential of the Cs+-sensitive current, which generally was within a few millivolts ofE K (−83 mV).
Statistics. Values are means ± SE, except where noted. Statistical significance was determined using Student’s two-tailed t-test.P ≤ 0.05 was considered significant.
Membrane parameters. Freshly isolated bovine RPE cells had a “figure-eight” shape with distinct apical and basolateral domains, as reported previously for frog (16) and human (20) RPE. Cell size varied considerably, with the diameter of the basolateral hemisphere ranging from 15 to 30 μm. The apical hemisphere was generally two to four times smaller and had numerous processes projecting from its surface. To ensure effective space clamp, we selectively recorded from the smallest cells in the recording chamber; these cells had an average membrane capacitance of 82 ± 26 (SD) pF (n = 118). The zero-current potential (V 0) of bovine RPE cells, measured in the zero-current clamp mode immediately after rupturing the membrane patch, averaged −64 ± 11 (SD) mV (n = 118). This value compares favorably with the membrane potentials measured at 35°C in the intact bovine RPE-choroid preparation with intracellular microelectrodes (22).
Properties of IKir.
To characterize the properties ofI Kir, we carried out a series of experiments on bovine RPE cells dialyzed with 4 mM ATP; under this condition,I Kir was sustained for 2 h or more. Figure 1 shows representative recordings of whole cell currents in a cell that had been dialyzed with 4 mM ATP (3.36 mM MgATP, 0.64 mM free ATP) for ∼5 min. Currents were essentially time independent over the voltage range examined, but their amplitude was substantially larger at hyperpolarized voltages (Fig. 1 A,top). The steady-stateI-V relationship (Fig.1 B, control) shows that the whole cell current reversed at −75 mV (V 0) and exhibited mild inward rectification, with the slope conductance decreasing gradually with membrane depolarization. Superfusing the cell with 20 mM Cs+, an effective blocker of I Kirin amphibian (31) and human (18) RPE cells, substantially reduced the amplitude of the whole cell currents (Fig.1 A,middle), depolarizedV 0 to −31 mV (Fig. 1 B), and markedly reduced inward and outward slope conductances (Fig.1 B). The Cs+-sensitive current, obtained by taking the difference between the family of currents recorded in the absence and presence of Cs+, is shown in Fig. 1 A,bottom. The current activated rapidly in response to hyperpolarizing voltage steps and did not exhibit inactivation. The I-V relationship of the difference current (Fig. 1 C) shows that the underlying conductance was mildly voltage dependent and that a substantial portion was activated in the physiological voltage range, which lies between −70 and −50 mV (22). The reversal of the difference current at −82 mV near the predicted K+ equilibrium potential (−83 mV) confirms that the Cs+-sensitive, inwardly rectifying current is carried by K+.
Similar results were obtained in 10 other cells dialyzed with 4 mM ATP for 5–25 min. Extracellular Cs+ (10–50 mM) depolarizedV 0 from a mean value of −69.0 ± 1.3 mV to −46.1 ± 2.5 mV (P < 10−5), decreased the inward slope conductance measured between −120 and −110 mV from 10.47 ± 1.26 to 2.35 ± 0.30 nS (P < 10−5), and decreased the outward slope conductance measured between −60 and −50 mV from 5.95 ± 1.21 to 2.28 ± 0.37 nS (P = 0.001). These results indicate that the inwardly rectifying K+conductance accounts for ∼78% of the inward slope conductance and 62% of the outward slope conductance of the bovine RPE cell membrane.
To assess blocker sensitivity, we estimated the apparent dissociation constants (K D) of the Cs+- and Ba2+-induced blocks of the inwardly rectifying K+ conductance by measuring I Kirin the presence of various blocker concentrations. Figure2 A shows the dose-response relationship of the Cs+-induced block ofI Kir measured at −35 and −110 mV. Symbols represent the mean values of normalized K+ current measured in the presence of various Cs+concentrations (n = 5), and the smooth curves are the best fits of the data to Eq.1 (see methods). The Cs+-induced block was mildly voltage dependent and had a meanK D of 2.01 ± 0.32 mM at −35 mV and aK D of 0.73 ± 0.08 mM at −110 mV (P = 0.005). At none of the concentrations tested did Cs+ exhibit a time-dependent effect on I Kir(not shown). These results indicate that 20 mM Cs+ blocks >90% ofI Kir, thus validating the use of this blocker to isolateI Kir from other currents.
Figure 2 B shows that external Ba2+ also blockedI Kir in a weakly voltage-dependent manner; the apparentK D of the Ba2+-induced block averaged 219 ± 68 μM at −35 mV and 85 ± 22 μM at −110 mV (P = 0.02;n = 6). In contrast to Cs+, the Ba2+-induced block was time dependent at voltages negative toE K (not shown). Quinidine, a blocker of several types of K+ channel (7), blocked ∼50% of the inwardly rectifying K+conductance at a concentration of 20 μM (n = 2), and at a concentration of 100 μM the block was essentially complete (n = 4; not shown). Glybenclamide, a potent inhibitor of ATP-sensitive K+ channels (1), had no significant effect on K+ currents in bovine RPE cells at a concentration of 100 μM (n = 5).
Dependence on intracellular ATP. To investigate the possible dependence ofI Kir on intracellular ATP, we dialyzed cells with ATP-free pipette solution and monitored the time course of changes in current at specific voltages (see methods). The results of a representative experiment are shown in Fig.3. Figure3 A plots the time course of currents measured at −30 and −82 mV, the reversal potentials of the Cs+-insensitive (residual) and Cs+-sensitive components of the whole cell current, respectively (see Fig.3 B). At −30 mV, the sum of all currents excludingI Kir was zero, and hence the current at this voltage represents outward K+ movement through the inwardly rectifying K+ conductance. Likewise, at −82 mV (≈E K), there was no current through the inwardly rectifying K+ conductance, and hence the current at this voltage represents the residual current, consisting largely of Cl− and Na+ currents. As shown in Fig.3 A,I Kir ran down over a period of 8 min, while the residual current remained essentially unchanged. A decrease inI Kir is also evident from the data depicted in Fig.3 B, which comparesI-V curves measured after 30 s of whole cell recording (a) and again after 8 min of dialysis (b). Dialysis with the ATP-free solution led to a large decrease in inwardly rectifying current, resulting in a shallowI-V relationship. The difference between the initial and steady-stateI-V curves (Fig.3 C, a− b) exhibits mild inward rectification and a reversal potential of −82 mV, confirming that the component of whole cell current that ran down was mainlyI Kir. This conclusion was corroborated by the finding that 20 mM Cs+ had no effect on theI-V relationship at 9 min, when the current at −30 mV had reached zero (Fig.3 A). Similar results were obtained in nine other cells dialyzed with ATP-free pipette solution, with the half time forI Kir rundown averaging 1.9 ± 0.3 min (Fig. 3 Dand Table 1). The time course of rundown was approximately five times slower than the rate expected for simple diffusion of MgATP from the cytoplasm to the pipette,1suggesting that cellular ATP may be compartmentalized.
As already mentioned,I Kir was sustained when bovine RPE cells were recorded with pipette solution containing 4 mM ATP. Figure 4 shows the results of dialyzing a cell with the pipette solution containing 2 mM ATP (1.90 mM MgATP, 0.10 mM free ATP). During the first minute of dialysis, I Kir(measured at −36 mV) increased slightly (Fig.4 A), possibly due to an increase in intracellular K+ concentration. Thereafter, I Kirwas stable until the cell was superfused with 20 mM Cs+ (Fig.4 A, open box), which produced a reversible inhibition. Figure 4,B andC, shows that theI-V relationship changed little after 6 min of dialysis and that the cell retained a large inwardly rectifying K+ conductance. The results obtained in this and nine other cells dialyzed with 2 mM ATP are summarized in Fig. 4 D, which shows the time course of normalizedI Kir.
We further tested the dependence ofI Kir on intracellular ATP by first depleting cells of ATP by metabolic inhibition and then dialyzing them with ATP. We found that the exposure of bovine RPE cells to 250 ng/ml oligomycin plus 10 mM deoxyglucose (substituted for glucose) caused nearly complete rundown of the inwardly rectifying K+ conductance (not shown). In six cells recorded in the amphotericin B perforated-patch configuration, the combined metabolic inhibitors reduced the inwardly rectifying K+chord conductance (g Kir) at −110 mV from an average of 4.54 ± 0.74 to 0.34 ± 0.13 nS (P = 0.007). Figure5 shows the results of an experiment in which a cell preexposed to oligomycin and deoxyglucose for ∼30 min was recorded in the whole cell configuration with a pipette solution containing 4 mM ATP (3.36 mM MgATP, 0.64 mM free ATP). Figure5 A plots the time course of currents measured at −38 and −83 mV, which representI Kir and the residual current, respectively. At the first measurement after rupture of the membrane patch,I Kir was near zero, but it increased rapidly until it reached a steady level after ∼3 min of dialysis. Inspection ofI-V relationships corresponding to these time points (Fig. 5, B andC) reveals that dialysis led to a dramatic increase in both inward and outward conductances and that these conductance increases were completely blocked by 20 mM extracellular Cs+. The data indicate that the inwardly rectifying K+ conductance was small or absent before dialysis with ATP began and that the change in whole cell current can be attributed solely to an increase in inwardly rectifying K+ conductance. Similar results were obtained in two other cells that were first metabolically inhibited and then dialyzed with 4 mM ATP. For these three cells,I Kir rose with an average half time of ∼1.8 min (Fig.5 D), and the steady-stateg Kir at −110 mV averaged 6.69 ± 0.49 nS. These results provide strong support for the idea that the magnitude ofI Kir is determined by intracellular ATP concentration.
To specify the ATP concentration range over whichI Kir is modulated, we measuredg Kir at −110 mV in cells dialyzed with various amounts of ATP for 10–15 min, when diffusional exchange between the pipette and cytoplasm was likely complete. The results of these experiments are summarized in Fig. 6, which depicts average conductances for cells dialyzed with 0 ATP, 0.5 mM ATP (0.46 mM MgATP, 0.04 mM free ATP), 2 mM ATP (1.90 mM MgATP, 0.10 mM free ATP), and 4 mM ATP (3.36 mM MgATP, 0.64 mM free ATP). Compared with the conductance present in cells dialyzed with 4 mM ATP, cells dialyzed with 2 mM, 0.5 mM, and 0 ATP had conductances that were 28% (P = 0.09), 71% (P = 0.0004), and 95% smaller (P < 10−7), respectively. Together, these results suggest that the concentration of intracellular ATP necessary to support a half-maximalg Kir probably lies in the range 0.5–2 mM. It is worthwhile noting that the local concentration of ATP in the vicinity of the K+ channels or the regulatory protein that modulates them may differ from the concentration of ATP in the pipette because of diffusion limitations and endogenous ATP production and consumption. Nevertheless, the data are consistent with the idea that relatively high ATP concentrations, in the millimolar range, are required to sustain maximal activity ofI Kir.
We also tested whether other nucleoside triphosphates could sustainI Kir. As shown in Fig. 7, internal dialysis with 4 mM GTP (3.4 mM MgGTP, 0.6 mM free GTP) was effective in preventing the rundown. During the initial 8 min of dialysis, there was a transient increase in I Kir, but comparison of initial and steady-stateI-V relationships showed little net change (Fig. 7 B,a andb, and Fig.7 C, a− b). Exposure of the cell to 20 mM Cs+ after 8 min of dialysis blocked a major fraction of whole cell current, demonstrating the presence of a large inwardly rectifying K+ conductance (Fig. 7,B andC). Although results were somewhat variable, there was no evidence ofI Kir rundown in cells dialyzed with 4 mM GTP (Fig.7 D). After 10–20 min of dialysis, g Kir at −110 mV averaged 5.47 ± 1.09 nS (n = 10), which is not significantly different from the value obtained in cells dialyzed with 4 mM ATP (P = 0.14). These results suggest that the mechanism responsible for sustainingI Kir can utilize GTP as well as ATP.
Effect of ATP analogs. If inwardly rectifying K+ channel activity were regulated by nonhydrolytic ATP binding, thenI Kir should be sustained by the nonhydrolyzable ATP analog, AMP-PNP. Figure8 shows, however, that dialysis with 4 mM AMP-PNP plus 4 mM Mg2+ led toI Kir rundown. This effect cannot be attributed to the small amount of Li+ that was liberated from AMP-PNP, becauseI Kir was stable in two cells dialyzed with 4 mM ATP plus 4 mM Li+ (not shown). In six cells dialyzed with 4 mM AMP-PNP,I Kir declined to near zero with an average half time of 2.7 ± 0.6 min (Fig.8 D and Table 1); after 5–10 min of dialysis, the magnitude ofg Kir at −110 mV averaged 0.13 ± 0.07 nS.
These results suggest that simple nonhydrolytic ATP binding is not sufficient to sustain the activity ofI Kir. An alternative explanation, however, is that AMP-PNP does not have the appropriate three-dimensional structure to interact with the nucleotide-binding site of the channel or regulatory protein. To examine this possibility, we tested whether an excess of AMP-PNP could prevent ATP from sustainingI Kir. In eight cells, dialysis with 3.5 mM AMP-PNP plus 0.5 mM ATP led to a partial rundown of I Kir, with the steady-stateg Kir at −110 mV averaging 3.5 ± 0.72 nS. This value is somewhat greater but not significantly different from the average steady-state conductance measured in cells dialyzed with 0.5 mM ATP alone (P = 0.21), indicating that AMP-PNP does not bind to the ATP regulatory site or else it does so with a much lower affinity than does ATP. Hence the results with AMP-PNP do not allow us to rule out the possibility that the mechanism responsible for the sustained activity ofI Kir involves nonhydrolytic binding of ATP.
The poorly hydrolyzable ATP analog ATPγS can also substitute for ATP in some mechanisms involving nonhydrolytic binding. In addition, it can serve as a substrate for a number of protein kinases, whereas most ATPases do not accept it (8). Figure 9shows, however, that dialysis of a bovine RPE cell with 4 mM ATPγS plus 4 mM Mg2+ did not prevent rundown of I Kir. The change in current at −84 mV (Fig.9 A) indicates that the conductance for an ion other than K+ also decreased, but the fact that the rundown current reversed nearE K (−75.8 ± 2.8 mV, n = 4) leads to the conclusion that the main effect of ATPγS was a decrease in K+ conductance. In four cells dialyzed with 4 mM ATPγS,I Kir rapidly declined with an average half time of 1.8 ± 0.2 min (Fig.9 D), and the steady-stateg Kir measured at −110 mV averaged 0.26 ± 0.16 nS. In contrast to AMP-PNP, dialysis with ATPγS together with ATP causedI Kir to run down to a greater extent than did ATP alone. In eight cells dialyzed with 3.5 mM ATPγS plus 0.5 mM ATP, the steady-stateg Kir at −110 mV averaged 0.30 ± 0.13 nS, which is significantly smaller than the average steady-state conductance measured in cells dialyzed with 0.5 mM ATP (P = 0.03). These results suggest that ATPγS can bind to the same regulatory site as does ATP but is incapable of maintainingI Kir activity. Hence we can exclude nonhydrolytic ATP binding as the mechanism underlying regulation of the channel.
The failure of ATPγS to prevent or slowI Kir rundown also suggests that ATP dependence for sustained activity ofI Kir does not involve a phosphorylation mechanism. To test this possibility further, we carried out a series of experiments in which cells were dialyzed with an ATP-free pipette solution containing various protein phosphatase inhibitors (Table 1). The inclusion of calyculin, a specific inhibitor of protein phosphatases 1 and 2A, in the pipette solution did not significantly affect the rate of rundown at concentrations of either 10 nM or 1 μM. Likewise, rundown was not affected by the nonspecific protein phosphatase inhibitor orthovanadate. Rundown was somewhat faster in cells dialyzed with pipette solution containing the serine-threonine protein phosphatase inhibitor okadaic acid, but this could have arisen from a smaller average cell volume for this experimental group. Although we cannot exclude the possibility that the inwardly rectifying K+ channel is regulated by a protein kinase at some level, the data suggest that the rundown ofI Kir observed in cells dialyzed with ATP-free solution is not the result of net dephosphorylation by a protein phosphatase. Rather, it seems likely that some other ATP-dependent mechanism is required to maintain sustained channel activity.
ATP metabolites. It is possible that the inwardly rectifying K+conductance is regulated by ATP breakdown products rather than by ATP itself. To test this possibility, we dialyzed cells with ADP or AMP. A representative example of an experiment involving dialysis with 4 mM ADP plus 4 mM Mg2+ is shown in Fig. 10. ADP did not sustain theI Kir at its initial level, but the rundown that did occur was relatively slow and incomplete. I Kiralso exhibited modest rundown in three other cells dialyzed with ADP, but in two cells no significant rundown occurred. For all cells, the steady-state g Kirat −110 mV averaged 4.54 ± 1.45 nS, which is significantly lower than that obtained with 4 mM ATP (P = 0.03, one-tailt-test). It is possible that the partial preservation of theI Kir by ADP resulted from the phosphorylation of ADP to ATP by endogenous adenylate kinases. In contrast, cells dialyzed with 4 mM AMP consistently exhibited a complete rundown ofI Kir, with an average half time of 1.9 ± 1.6 min (n = 4). These results provide additional support for the hypothesis that the sustained activity ofI Kir requires ATP.
For a number of enzymatic processes that involve ATP hydrolysis, Mg2+ is a requisite cofactor. To test whether MgATP is required to maintainI Kir, we dialyzed four cells with a pipette solution containing 4 mM K2ATP and 0 Mg2+. In each cell, the current measured at the reversal potential of the residual (Cs+-insensitive) current rapidly declined to near zero within 5 min (half time = 1.8 ± 0.2 min), with little change in the current measured atE K, indicating rundown of I Kir(not shown). This conclusion is supported by the findings that the reversal potential of the rundown current was close toE K (−75.6 ± 6.8 vs. −83 mV) and that 20 mM external Cs+ had no significant effect on the whole cell current after rundown was complete. These results indicate that MgATP, rather than free ATP, is the molecule required for sustained activation of theI Kir.
The present study on freshly isolated bovine RPE cells demonstrates that the inwardly rectifying K+conductance, which is the predominant ion conductance at physiological voltages, requires intracellular ATP for sustained activity. The results suggest the possible involvement of a hydrolytic mechanism rather than simple nonhydrolytic ATP binding.
ATP dependence of IKir.
The sustained activity of the bovine RPEI Kir was found to depend on the concentration of intracellular ATP. Internal dialysis with an ATP-free pipette solution resulted in the rapid rundown (half time = 1.9 min) ofI Kir. Rundown was prevented, however, when the pipette solution contained 2 mM ATP. ATP prevented rundown in a dose-dependent manner: at a concentration of 0.5 mM ATP, partial rundown occurred, the inwardly rectifying K+ conductance declining ∼71%. ADP was partially effective in preventing the complete rundown ofI Kir, but this may have been due to the local generation of ATP from ADP by an adenylate kinase. The conductance appears to be regulated by ATP itself and not by its metabolites, because rundown occurred when AMP was used in place of ATP.
There are several possible mechanisms that could underlie this ATP regulation ofI Kir. These include simple ATP binding to a nucleotide-binding site, energy-consuming processes that require ATP hydrolysis, and phosphorylation. Our data suggest that hydrolysis of nucleotide triphosphates may be required for the activation of the inwardly rectifying K+ conductance. ATP and GTP were both effective in preventing rundown, but Mg2+, a cofactor required for ATP hydrolysis reactions, was also required for the sustained activity ofI Kir. The possibility that nonhydrolytic binding of ATP to a nucleotide-binding site regulates channel activity seems less likely, since high concentrations of the nonhydrolyzable ATP analogs AMP-PNP and ATPγS did not prevent rundown. It is possible, however, that these analogs do not have the appropriate three-dimensional structure to interact with the nucleotide-binding site of the channel (or regulatory protein). Consistent with this idea, an excess of AMP-PNP did not interfere with the regulatory effects of ATP. In contrast, similar experiments with ATPγS yielded results indicating competitive inhibition. Hence we conclude that the mechanism by which ATP regulates the inwardly rectifying K+ channel does not involve simple nonhydrolytic binding. ATPγS can also serve as a cofactor for protein kinase phosphorylation (5, 8), and its failure to support I Kirsuggests that phosphorylation of the channel alone is not sufficient to open the channel. This conclusion is supported by the finding that the inhibition of protein phosphatase activity by calyculin, okadaic acid, or orthovanadate did not slowI Kir rundown in cells dialyzed with 0 ATP. Thus we conclude that RPE inwardly rectifying K+ channel activity is sustained by an ATP-dependent process that may involve ATP hydrolysis.
It is conceivable that the rundown ofI Kir was an indirect result of inhibition of an ATPase, such as the plasma membrane or endoplasmic reticulum Ca2+pump. For example, inhibition of a Ca2+-ATPase could lead to an increase in intracellular Ca2+, which leads to the inhibition ofI Kir in both toad (19) and bovine (unpublished observations) RPE cells. There are two arguments, however, against this hypothesis. First, the pipette solutions used in present experiments contained 5.5 mM EGTA, which would be expected to buffer the free Ca2+ concentration to ∼10−8 M. Still, the Ca2+-buffering capacity may have been exceeded in the proximity of the plasma membrane if the rate of Ca2+ influx where high. Second, the dose dependency of the ATP regulation ofI Kir is inconsistent with the dependency of most ATPases on ATP. Our data suggest that theK D of the ATP-dependent regulation ofI Kir is >500 μM. This value is much higher than that obtained for Na+-K+-ATPase and Ca2+-ATPase [K D,ATP = 0.2–2 μM (34)]. Other results suggest thatI Kir rundown does not result from inhibition of the Na+-K+pump. In preliminary experiments on bovine RPE cells recorded in the amphotericin B perforated-patch configuration, we found that inhibition of the Na+-K+-ATPase with ouabain had no inhibitory effect on the inwardly rectifying K+ conductance (n = 5).
ATP-dependent regulation has been described in reports on various members of the inwardly rectifying K+ (Kir) channel family. For example, the functional activity of a cloned inward rectifier K+ channel, Kir2.1 (IRK1), requires ATP for both phosphorylation and a separate mechanism apparently involving ATP hydrolysis (10). On the other hand, ATP has a dual effect on native ATP-sensitive K+ channels (KATP). When present at low concentrations on the cytoplasmic face of excised membrane patches, it sustains channel activity through a phosphorylation mechanism (11, 35), but at high concentrations it inhibits activity through nonhydrolytic binding (6, 11, 28).
In a recent report, Ishii et al. (21) presented molecular biochemical and immunocytochemical evidence that rat and rabbit RPE cells express Kir4.1 (KAB-2). This ATP-regulated K+ channel contains a Walker type-A domain, which has been proposed to serve as a hydrolytic ATP binding site that confers ATP dependence (33). This idea is somewhat controversial, however, because recent studies on Kir1 (ROMK1), another cloned inward rectifier K+ channel containing a Walker type-A domain (13), suggest that nonhydrolytic interactions between MgATP and the Walker A site are involved in channel inhibition (24). The expression pattern and ATP dependence of Kir4.1 suggest that this channel subtype may be the molecular basis for the RPE inwardly rectifying K+conductance, but several properties of the native RPE conductance differ from those of Kir4.1 expressed inXenopus oocytes; these include weaker inward rectification and an anomalous decrease in slope conductance when extracellular K+concentration is increased (18, 31). Clearly, additional experiments are needed to determine the relationship between the RPE inwardly rectifying K+ channel and Kir4.1, as well as the mechanism underlying its ATP-dependent regulation.
Physiological significance. There is a wide body of evidence indicating that the inwardly rectifying K+ conductance in the RPE apical membrane is an important control point for the transport of K+ and other ions between the subretinal space and the choroid. For instance, studies in intact bovine RPE have shown that the large apical membrane K+ conductance mediates the recycling of most of the K+ that enters the cell via the apical Na+-K+pump and that a decrease in its magnitude, produced by extracellular Ba2+, for example, stimulates net K+ absorption (25). In addition, the apical K+ conductance provides a crucial link in the communication between photoreceptors and the RPE. At light onset, the closure of cGMP-gated channels in the plasma membrane of photoreceptor outer segments produces a hyperpolarization, leading to a redistribution of K+from the extracellular space into the photoreceptor cytoplasm. This light-evoked decrease in subretinal K+ concentration hyperpolarizes the apical membrane of the RPE by virtue of its inwardly rectifying K+ conductance (15), and this triggers changes in several RPE transport mechanisms (9, 14).
The requirement of the inwardly rectifying K+ conductance for relatively high intracellular ATP levels has important implications with regard to RPE function. An increase in glycolytic metabolism, for example, may elevate the intracellular ATP concentration, causing an increase in apical K+ conductance; this conductance increase would lead to a greater degree of recycling of the K+ that enters the cell through the apical Na+-K+pump and a decrease in net K+absorption. Circumstances that would tend to decrease intracellular ATP levels, such as an increase in Na+-K+-ATPase activity secondary to an increase in Na+-coupled ion and nonelectrolyte transport, would have the opposite effect, decreasing the apical K+ conductance and increasing net K+ reabsorption (25). This decrease in the apical K+conductance would also diminish the responsiveness of the RPE to photoreceptor-induced changes in subretinal K+ concentration, which normally leads to, among other things, volume changes of the RPE cell and subretinal space (2). It is presently not known whether these aforementioned mechanisms produce changes in intracellular ATP concentration in the RPE under physiological conditions. Changes in the metabolic capacity of the RPE resulting from disease processes or aging could also lead to a decrease in the intracellular concentration of ATP, resulting in a diminishment of the apical K+ conductance, and abnormal transport function.
This work was supported by National Eye Institute Grant EY-O8850, Core Grant EY-O7703, a Retinitis Pigmentosa Research Center Grant, and a Research to Prevent Blindness Career Development Award to B. A. Hughes.
Address for reprint requests: B. A. Hughes, Dept. of Ophthalmology, Univ. of Michigan Medical School, 1000 Wall St., Ann Arbor, MI 48105.
Present Address of M. Takahira: Dept. of Ophthalmology, Kanazawa Univ. School of Medicine, Kanazawa, Ishikawa, Japan.
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. §1734 solely to indicate this fact.
↵1 The time course of dialysis between the patch pipette and the cell interior is a complex function of cell volume and geometry, access resistance between the pipette and the cell, and the molecular weight of the test molecule (29). To quantify the rate of diffusional equilibrium in bovine RPE cells, we carried out whole cell recordings with K+-free pipette solution (NMDG substitute) and measured the delayed rectifier current at +40 mV every 5 s. In eight cells, the outward K+ current declined exponentially, with an average half time of 10.2 ± 0.9 s. Assuming that the rate of current decline is solely a function of diffusion of K+ out of the cell and that the diffusion rate between cytoplasm and pipette is proportional to the inverse cube root of the molecular weight, then the half time for MgATP diffusion can be estimated to be ∼24 s. This value is about five times faster than the observed half time forI Kir rundown in cells dialyzed with ATP-free solution (112 s).
- Copyright © 1998 the American Physiological Society