The CFTR (Cystic Fibrosis Transmembrane Conductance Regulator) Channel: Anion Permeation, and Regulation by Adenylyl Cyclase and ATP Hydrolysis

CFTR Cl- channel function and regulation were studied in guinea-pig ventricular myocytes, using either the whole-cell or excised inside-out patch configurations of the patch clamp technique. A typical myocyte cell membrane contained a total of -1400 CFTR channels; channel density was -0.11 Jlm-2. Using whole-cell CFTR Cl: conductance as an on-line assay of cAMP levels, we examined the interaction of the adenylyl cyclase (AC) modulators forskolin (Fsk), and the GTP-binding proteins Gs and Gi. In the presence of GTP, maximal activation of Gi using the muscarinic acetylcholine receptor agonist carbachol (CCh) reduced the efficacy with which the -adrenoceptor agonist isoproterenol (Iso) activated AC, with little effect on potency, whereas CCh decreased the potency but not the efficacy of Fsk, a direct activator of AC. Thus, Gi appeared to act like a competitive inhibitor of Fsk\u27s stimulatory action on AC, but like a noncompetitive inhibitor of Iso\u27s stimulatory action. Anion selectivity of CFTR channels was determined by shifts of the reversal potential (Erev) for CFTR channel current on changes in extracellular [Cl ] ([Cl-]o): Erev shifted roughly linearly with log [Cl-]o, with a slope of -57 ± 3 mY, close to the Nernst prediction of -61.5 mY, implying that CFTR channels are relatively anion selective. Relative permeability to several anions was determined by shifts of whole-cell bi-ionic reversal potential, both with normal channel gating, and with channels locked open using the ATP analog AMP-PNP. The sequence of relative permeabilities was unaffected by channel gating, and was: NOa- (1.74 ± 0.04) \u3e Br- (1.51 ± 0.06) \u3e 1- (1.30 ± 0.07) \u3e CI- (1.0) \u3e F- (0.25 ± 0.03) \u3e Aspartate- (0.07±O.01) » Isethionate (0.05), HEPES (0.03). Although 1- was more permeant than CI-, it reduced current both at positive and at negative potentials, indicating that 1- ions readily enter CFTR channels, but leave them more slowly than CIions. This sequence suggests that the anion binding site within the pore is a relatively weak site (Eisenman sequence II). The minimum pore diameter appears to be -7 A. The single channel conductance (y) in symmetric 160 mM CI-, determined with single channels isolated in excised patches and locked open with AMP-PNP, was 10.3 ± 0.4 pS; the permeability coefficient for CI- (PCl) was 2xl0-17 ± 9xl0-19 cm3s-1. Protein kinase A (PKA)-phosphorylated CFTR CI- channels require hydrolyzable nucleoside triphosphate to open and close normally. Individual CFTR channels were studied in excised membrane patches to examine how nucleoside triphosphate action at one or both of the nucleotide binding domains (NBDs) controls channel activity. At 1.2 mM free [Mg2+], channel open probability (Po) increased hyperbolically with [ATP] (KO.5 -35 JiM [ATP]), largely due to an increase in opening rate. That ATP hydrolysis governs channel opening is almost certain, since all ATPases require Mg2+, and Mg2+ ions were absolutely required for channel opening by ATP. Thus, at 2 mM [ATP], the mean opening rate at 20-22 °C was -0 at -0 [Mg2+], -0.03 s-1 at 5 JlM [Mg2+], and -0.22 s-1 at 1.2 mM [Mg2+]. This suggests that both ATP and Mg2+ must be bound before a channel can open. Free [Mg2+] levels also regulate closing. At 5 JlM [Mg2+], channels can stay open for tens of s, even after rapid (-1 s) ATP washout, but close promptly on raising [Mg2+]: the mean closing rate rises with [Mg2+], from -0.05 s-1 at 0 or 5 JiM [Mg2+], to -0.3 s-1 at 1.2 roM [Mg2+]. This confirms that hydrolysis of a second ATP prompts channel closure. Evidently, that second ATP stays tightly bound, stabilizing the open conformation, without Mg2+. The closing rate on Mg2+ readdition then reflects the probability that a Mg2+ ion will bind and catalyze hydrolysis of that ATP, leading to channel closure. The level of channel phosphorylation governs access to the NBDs and hence their ability to modulate channel activity. Partially phosphorylated channels can bind and hydrolyze ATP only at NBDl, resulting in exclusively brief channel openings. Fully phosphorylated channels can bind and hydrolyze ATP at both NBDI and NBD2: hydrolysis at NBDI opens channels, and binding of ATP at NBD2 stabilizes the channel in the open conformation, until that ATP is hydrolyzed, resulting in channel closure