LETTERTO THE EDITOR T h e M e c h a n i s m o f KAa-p C h a n n e l Inhibition by ATP Dear Sir, Ribalet et al. (1989) have presented interesting data on the regulation of KAav channel activity by phosphorylation, and an intriguing hypothesis to explain KAyP channel inhibition by ATP. They propose that the opening and closing of the KAvp channel can be explained by a one-site phosphorylation model, outlined in Fig. 1 A. According to the model of Ribalet et al. (1989), the opening of the KAXpchannel at low [ATP]i occurs as a result of channel phosphorylation by a cAMP-dependent protein kinase. At high [ATP]~ this kinase is inhibited by an ATP-dependent protein kinase inhibitor (PKI) so that background dephosphorylation leads to inhibition of channel activity. Any subsequent opening of the channel requires the channel to be rephosphorylated and consequently requires the presence of some high-energy phosphate. The experiment shown in Fig. 1 B appears to refute this hypothesis. The experiment was performed using a modified oil-gate bath (Qin and Noma, 1988; Lederer and Nichols, 1989), which allows rapid and complete change of the solution bathing the intracellular surface of an inside-out patch of membrane. In each of the three panels, KATP channel activity is initially high after channel isolation into an ATP-free solution. According to the hypothesis being tested, channel activity should have disappeared due to dephosphorylation with no substrate available for rephosphorylation. We have observed such behavior on isolating m e m b r a n e patches into ATP-free solution in a bath that has never had ATP in it. In such a case, it is impossible for a small residual amount of ATP to be available in the solution. In the continued absence of ATP, switching to a solution containing the nonhydrolyzable ATP analogue AMP-PNP (left panel), the low energy phosphate AMP (middle panel), or ATP itself (right panel) all caused rapid channel inhibition. We presume that each nucleotide acted in a similar manner to inhibit channel activity. According to the hypothesis of Ribalet et al. (1989), channel activity would be decreased in each case because of reduced phosphorylation of the KAaVchannel with continued background dephosphorylation. The critical observation with respect to the hypothesis being tested is the rapid reactivation of channel activity on switching t o a nucleotide-free solution. In each case a complete restoration of channel activity is seen when the patch is returned to the nucleotide-free solution. According to the hypothesis of Ribalet et al. (1989), reopening of the KAVpchannel should not occur in at least the first two of these experiments, because on exposure to a nucleotide-free solution (after having been exposed to AMP or AMP-PNP) the catalytic subunit will have AMP or AMP-PNP bound, and no high energy phosphate is available to reprime the catalytic subunit and rephosphorylate the channel. Fig. 1 C shows that the extent and time course of recovery from repeated addition or removal of AMP-PNP remain constant for at least 1 min after removal from an ATP-containing solution. Thus, the results shown in Fig. 1, B and C, and, for instance, Fig. 8 of Dunne et al. (1988) are J. GEN.PHYSIOL.© The Rockefeller University Press • 0022-1295/91/05/1095/4 $2.00 Volume 97 May 1991 1095-1098

1095

FIGURE 1. (A) Model of KArP c h a n n e l regulation according to the hypothesis of Ribalet et al. (1989). In the absence of ATP ADP c® phosphorylation, channels are closed. W h e n phosphorylated, channels are activated (shifted to a gating m o d e with high Pl o p e n probability). D e p h o s p h o r ylation is responsible for chanINHIBITIBN nel closure at low [ATP] ("rundown") a n d at high [ATP] 50 mM AMP 1 mM ATP B 2 mM AMP-PNP ("inhibition"). (B ) Reversible c h a n n e l closure on exposure to , r I I nucleotides. Records of Kmp , ,7 ~l' c h a n n e l current from isolated inside-out m e m b r a n e patches from rat ventricular myocytes. AMP-PNP AMP-PNP AMP-PNP ATP C I Channels are completely closed on exposure to 2 mM adenyllimido diphosphate, "AMPPNP" (left), 50 mM AMP (m/ddie), or 1 mM ATP (right). In each case, channels r e o p e n fully on removal of the nucleotide. In each case the calibration bars represent 2 s a n d 10 pA. T h e e x p e r i m e n t s were performed at room temperature. T h e patch electrode contained 4 m M K + a n d the bath contained 140 m M K +. T h e memb r a n e potential was 0 inV. (C) Recovery of channel activity on RUN-UP INHII3ITInN removal of nucleotides can be observed for many seconds alATP ADP C® NucLeo±ide ter removal fi'om ATP-containing solution. Records of KA]P channel current from insidel .. f Nucteoi;ide out m e m b r a n e patches from rat PI . ( .( ventricular myocytes. Channels RUNDDVN ACTIVATIDN were isolated into nucleotidefree solution. O n exposure to 1 mM AMP-PNP (indicated by bar above record) channels were rapidly inhibited. C h a n n e l activity rose back to maximal on removal of AMP-PNP. T h e patch was exposed to AMP-PNP two more times before being exposed to 0.5 mM ATP. T h e calibration bars r e p r e s e n t 10 s and 30 pA. T h e conditions were as in B above. (D) Alternative model of KATe channel regulation. Channels require phosphorylation to be activated. Channels are closed by the binding of ATP (or o t h e r nucleotide) a n d on removal of ATP channels o p e n as nucleotide unbinds. This model explains why a gradual run-down of c h a n n e l activity occurs at very low [ATP] (because there is no high energy p h o s p h a t e to support phosphorylation), but also explains how rapid, reversible blockade by ATP a n d o t h e r nucleotides can be obtained.

ACTIVATIBN

I-.

i

i

15

\

J

-o0

LETTERS TO THE EDITOR

1097

n o t c o m p a t i b l e in any simple ~ way with the hypothesis o f Ribalet et al. (1989) a n d suggest that a different m o d e l may be n e e d e d to e x p l a i n the c h a n n e l behavior. Fig. 1 D illustrates a simplified m o d e l o f c h a n n e l r e g u l a t i o n that can account for the d a t a p r e s e n t e d by Ribalet et al. (1989) as well as the result shown in Fig. 1 B. In this m o d e l t h e r e a r e two kinds o f interactions o f A T P with the KA~ channel. T h e c h a n n e l c a n n o t o p e n without a necessary p h o s p h o r y l a t i o n step (the only step in the m o d e l o f Ribalet et al., 1989). This p h o s p h o r y l a t i o n - d e p h o s p h o r y l a t i o n transition constitutes w h a t has b e e n t e r m e d " r u n - d o w n " a n d " r u n - u p " by various investigators (Trube a n d Hescheler, 1984; Findlay a n d Dunne, 1986; Misler et al., 1986; Findlay, 1987; O h n o - S h o s a k u et al., 1987; Ribalet et al., 1989) a n d takes place slowly (seconds to minutes; T a k a n o et al., 1990). H i g h [ATP] (or o t h e r nucleotide) leads to inhibition o f c h a n n e l activity by an a d d i t i o n a l direct b i n d i n g o f o n e (or m o r e ) A T P molecules to the c h a n n e l ( L e d e r e r a n d Nichols, 1989; Qin et al., 1989). This second step does n o t r e q u i r e p h o s p h o r y l a t i o n , can occur in the a b s e n c e o f Mg ~÷, a n d occurs quickly (milliseconds to seconds). Thus, o t h e r nucleotides (AMP-PNP, ADP, AMP, a d e n o s i n e , NADP, NADPH), can all r a p i d l y a n d reversibly inhibit c h a n n e l activity by direct b i n d i n g at the same site(s). A l t h o u g h the e x p e r i m e n t shown in o u r Fig. 1 B was c a r r i e d out in KAXpchannels from h e a r t muscle, evidence consistent with the m o d e l shown in Fig. 1 D has b e e n p u b l i s h e d by various workers using b o t h insulin-secreting cells a n d h e a r t cells (Dunne a n d Peterson, 1986; Findlay, 1987, 1988; D u n n e et al., 1988; Q i n a n d N o m a , 1988; L e d e r e r a n d Nichols, 1989). Ribalet et al. (1989) have d e m o n s t r a t e d that b o t h p r o t e i n kinase a n d PKI can affect KATP c h a n n e l activity. T h e i r results a r e i m p o r t a n t in c o n s i d e r i n g the physiological r e g u l a t i o n o f the channel. However, results p r e s e n t e d h e r e a n d elsewhere seem to invalidate their hypothesis that d e p h o s p h o r y l a t i o n is r e s p o n s i b l e for the welld o c u m e n t e d n u c l e o t i d e inhibition o f the channel. This work was supported by grants from the NIH and the Maryland Affiliate of the American Heart Association. Original version received 6 April 1990 and accepted version received 1 August 1990.

One might argue that our assumption that the ATP falls quickly at the membrane is invalid. It could be argued that ATP remains bound to the membrane for a substantial time after removal from the bathing solution, and that ATP is still available to rephosphorylate the channel after addition and subsequent removal of other blocking nucleotides. Such an argument is difficult to counter directly, but indirect evidence seems to make this possibility very unlikely. If dephosphorylation and rephosphorylation are occurring over millisecond time scales, as would be necessary to account for the rate of channel closure on raising [ATP] (Fig. 1 B ), then to maintain channel activity over tens or hundreds of seconds after removal of ATP (as in Fig. 1 C) thousands of molecules of ATP would have to be accessible to each kinase catalytic subunit. Given the diffusion coefficient of ATP, Qin and Noma (1988) have modeled the diffusion profile of [ATP] at the surface of the membrane in an isolated patch experiment. A similar calculation predicts that the [ATP] at the membrane surface would fall to subpicomolar levels in ~ 2.5 s. The time course and extent of channel opening on removal of a blocking dose of AMP-PNP is constant, at least over 20 times this period of time after removal from ATP-containing solution (Fig. 1 C), so that one would have to argue that ATP is continuously available to the kinase long after the solution [ATP] has dropped to essentially zero. Given a Kmfor the kinase of 7.6 I~M ATP (Whitehouse et al., 1983), this possibility seems untenable.

1098

THE JOURNAL OF GENERAL PHYSIOLOGY • VOLUME 97" 1991 REFERENCES

Belles, B., J. Hescheler, and G. Trube. 1987. Changes of membrane currents in cardiac cells induced by long whole-cell recordings and tolbutamide. Pfli*gers Archly. 409:582-588. Dunne, M. J., and O. H. Petersen. 1986. Intracellular ADP activates K+ channels that are inhibited by ATP in an insulin-secreting cell line. FEBS Letters. 208:59-62. Dunne, M. J., West-Jordan, R.J. Abraham, R. H. T. Edwards, and O. H. Petersen. 1988. The gating of nucleotide-sensitive K + channels in insulin-secreting ceils can be modulated by changes in the ratio ATp4-/ADP 3- and by nonhydrolyzable derivatives of both ATP and ADP.Journal of Membrane Biology. 104:165-177. Findlay, I. 1987. ATP-sensitive K+ channels in rat ventricular myocytes are blocked and inactivated by internal divalent cations. Pfliigers Archiv. 410:313-320. Findlay, I. 1988. Effects of ADP upon the ATP-sensitive K+ channel in rat ventricular myocytes. Journal of Membrane Biology. 101:83-92. Findlay, I., and M. J. Dunne. 1986. ATP maintains ATP-inhibited K + channels in an operational state. Pfliigers Archiv. 407:238-240. Kakei, M., A. Noma, and T. Shibasaki. 1985. Properties of adenosine-triphosphate-regulated potassium channels in guinea-pig ventricular cells.Journal of Physiology. 363:441-462. Lederer, W. J., and C. G. Nichols. 1989. Nudeotide modulation of the activity of rat heart KA,, channels in membrane patches. Journal of Physiology. 419:193-211. Misler, D. S., L, C. Falke, K. Gillis, and M. L. McDaniel. 1986. A metabolite regulated potassium channel in rat pancreatic [3-cells. Proceedings of the National Academy of Sciences, USA. 83:7119-7123. Ohno-Shosaku, T., B. J. Zunkler, and G. Trube, 1987. Dual effects of ATP on K ÷ currents of mouse pancreatic [3-cells. Pfliigers Archiv. 408:133-138. Qin, D., and A. Noma. 1988. A new oil-gate concentration jump technique applied to inside-out patch-clamp recording. American Journal of Physiology. 255:H980-H984. Qin, D., M. Takano, and A. Noma. 1989. Kinetics of ATP-sensitive K÷ channel revealed with oil-gate concentration jump method. American Journal of Physiology. 257 :H 1624-H 1633. Ribalet, B., S. Ciani, and G. T. Ribalet. 1989. ATP mediates both activation and inhibition of K(ATP) channel activity via cAMP-dependent protein kinase in insulin-secreting cell lines.Journal of General Physiology. 94:693-717. Takano, M., D. Qin, and A. Noma. 1990. ATP-dependent decay and recovery of K+ channels in guinea-pig cardiac myocytes. American Journal of Physiology. 258:H45-H50. Trube, G., and .]. Hescheler. 1984. Inward-rectifying channels in isolated patches of the heart cell membrane: ATP-dependence and comparison with cell-attached patches. Pfliigers Archiv. 407:178184. Whitehouse, S., J. R. Feramisco, J. E. Casnellie, E. G. Krebs, and D. A. Walsh. 1983. Studies on the kinetic mechanism of the catalytic subunit of the cAMP-dependent protein kinase. Journal ~![ Biological Chemistry. 258:3693-3701. C. G. NICHOt,S

W.J. LEDERER Department of Physiology, University of Maryland, Baltimore, Maryland 21201