Trinitrophenyl-ATP blocks colonic Cl- channels in planar phospholipid bilayers. Evidence for two nucleotide binding sites.

Outwardly rectifying 30-50-pS Cl- channels mediate cell volume regulation and transepithelial transport. Several recent reports indicate that rectifying Cl- channels are blocked after addition of ATP to the extracellular bath (Alton, E. W. F. W., S. D. Manning, P. J. Schlatter, D. M. Geddes, and A. J. Williams. 1991. Journal of Physiology. 443:137-159; Paulmichl, M., Y. Li, K. Wickman, M. Ackerman, E. Peralta, and D. Clapham. 1992. Nature. 356:238-241). Therefore, we decided to conduct a more detailed study of the ATP binding site using a higher affinity probe. We tested the ATP derivative, 2',3',O-(2,4,6-trinitrocyclohexadienylidene) adenosine 5'-triphosphate (TNP-ATP), which has a high affinity for certain nucleotide binding sites. Here we report that TNP-ATP blocked colonic Cl- channels when added to either bath and that blockade was consistent with the closed-open-blocked kinetic model. The TNP-ATP concentration required for a 50% decrease in open probability was 0.27 microM from the extracellular (cis) side and 20 microM from the cytoplasmic (trans) side. Comparison of the off rate constants revealed that TNP-ATP remained bound 28 times longer when added to the extracellular side compared with the cytoplasmic side. We performed competition studies to determine if TNP-ATP binds to the same sites as ATP. Addition of ATP to the same bath containing TNP-ATP reduced channel amplitude and increased the time the channel spent in the open and fast-blocked states (i.e., burst duration). This is the result expected if TNP-ATP and ATP compete for block, presumably by binding to common sites. In contrast, addition of ATP to the bath opposite to the side containing TNP-ATP reduced amplitude but did not alter burst duration. This is the result expected if opposite-sided TNP-ATP and ATP bind to different sites. In summary, we have identified an ATP derivative that has a nearly 10-fold higher affinity for reconstituted rectifying colonic Cl- channels than any previously reported blocker (Singh, A. K., G. B. Afink, C. J. Venglarik, R. Wang, and R. J. Bridges. 1991. American Journal of Physiology. 260 [Cell Physiology. 30]:C51-C63). Thus, TNP-ATP should be useful in future studies of ion channel nucleotide binding sites and possibly in preliminary steps of ion channel protein purification. In addition, we have obtained good evidence that there are at least two nucleotide binding sites located on opposite sides of the colonic Cl- channel and that occupancy of either site produces a blocked state.

Regulation of CFTR Cl- channel gating by ADP and ATP analogues.

The cystic fibrosis gene product (CFTR) is a chloride channel which, once phosphorylated, is regulated by nucleotide phosphates (Anderson, M. P., and M. J. Welsh. 1992. Science. 257:1701-1704; Venglarik, C. J., B. D. Schultz, R. A. Frizzell, and R. J. Bridges. 1994. Journal of General Physiology. 104:123-146). Nucleotide triphosphates initiate channel activity, while nucleotide diphosphates and nonhydrolyzable ATP analogues do not. To further characterize the role of these compounds on CFTR channel activity we examined their effects on chloride channel currents in excised inside-out membrane patches from CFTR transfected mouse L cells. ADP competitively inhibited ATP-dependent CFTR channel gating with a Ki of 16 +/- 9 microM. AMP neither initiated CFTR channel gating nor inhibited ATP-dependent CFTR channel gating. Similarly, ATP analogues with substitutions in the phosphate chain, including AMPCPP, AMPPCP, AMPPNP, and ATP gamma S failed to support CFTR channel activity when present at the cytoplasmic face of the membrane and none of these analogues, when present at three to 10-fold excess of ATP, detectably altered ATP-dependent CFTR channel gating. These data suggest that none of these ATP analogues interact with the ATP regulatory site of CFTR which we previously characterized and, therefore, no inference regarding a requirement for ATP hydrolysis in CFTR channel gating can be made from their failure to support channel activity. Furthermore, the data indicate that this nucleotide regulatory site is exquisitely sensitive to alterations in the phosphate chain of the nucleotide; only a nonsubstituted nucleotide di- or triphosphate interacts with this regulatory site. Alternative recording conditions, such as the presence of kinase and a reduction in temperature to 25 degrees C, result in a previously uncharacterized kinetic state of CFTR which may exhibit distinctly different nucleotide dependencies.

ATP alters current fluctuations of cystic fibrosis transmembrane conductance regulator: evidence for a three-state activation mechanism.

The cystic fibrosis gene product cystic fibrosis transmembrane conductance regulator (CFTR) is a low conductance, cAMP-regulated Cl- channel. Removal of cytosolic ATP causes a cessation of cAMP-dependent kinase-phosphorylated CFTR channel activity that resumes upon ATP addition. (Anderson, M. P., H. A. Berger, D. R. Rich, R. J. Gregory, A. E. Smith, and M. J. Welsh. 1991. Cell. 67:775-784). The aim of this study was to quantify possible effects of ATP on CFTR gating. We analyzed multichannel records since only 1 of 64 patches contained a single channel. ATP increased the channel open probability (Po) as a simple Michaelis-Menten function of concentration; the effect was half maximal at 24 microM, reached a maximum of 0.44, and had a Hill coefficient of 1.13. Since the maximum Po was not 1, the simplest description of the effect of ATP on CFTR gating is the noncooperative three-state mechanism of del Castillo and Katz (1957. Proceedings of the Royal Society of London. B. 146:369-381). We analyzed current fluctuations to quantify possible changes in CFTR gating. The power density spectra appeared to contain a single Lorentzian in the range of 0.096-31 Hz. Analysis of the corner frequency (fc) of this Lorentzian revealed that ATP increased 2 pi fc as a Michaelis-Menten function with a Hill coefficient of 1.08, and it provided estimates of the ATP dissociation constant (44 tau open (154 ms), and the ATP-sensitive tau close [(185 ms) (44 microM/[ATP] + 1)]. These results suggest that the binding reaction is rapid compared to the opening and closing rates. Assuming that there is a single set of closed-to-open transitions, it is possible to verify the outcome of fluctuation analysis by comparing fluctuation-derived estimates of Po with measures of Po from current records. The two values were nearly identical. Thus, noise analysis provides a quantitative description of the effect of ATP on CFTR opening. The noncooperative three-state model should serve as a basis to understand possible alterations in CFTR gating resulting from regulators or point mutations.

Cholinergic regulation of Na absorption by turtle colon: role of basolateral K conductance.

The mechanism underlying the muscarinic inhibition of colonic Na absorption is unknown. In this study the effects of carbachol on active Na transport and basolateral K conductance were compared in the isolated turtle colon. Carbachol produced a biphasic response in both Na transport and basolateral K conductance. The response consisted of a transient activation followed by a sustained inhibition and was blocked by atropine. Submucosal cholinergic neurons were implicated in the regulation of colonic transport by employing depolarizing agents to release endogenous acetylcholine. Depolarizing agents produced a carbachol-like response that was atropine-sensitive. Finally, experiments with the Ca ionophores, A23187 and ionomycin, suggested that the muscarinic response may be mediated, at least in part, by changes in cellular Ca. These experiments provide evidence that cholinergic neurons are present in the turtle colon submucosa, muscarinic agonists cause a change in basolateral K conductance that may be an important event in the regulation of colonic Na absorption, and a Ca second messenger system may be involved in mediating the response.

Development of chloride channel modulators.

Chloride channels are ubiquitously distributed, biophysically varied and functionally diverse. Despite the known contribution of chloride channels to the physiology of various cell types and the pathology of several diseases, high affinity ligands are not available to study these channels. Here we report the iterative and integrated use of ion channel kinetic analysis and computational chemical methods in the development of high affinity blockers of the outwardly rectifying chloride channel (ORCC). Kinetic analysis, with emphasis on estimation of the block time constant as determined from critical closed time plots, was used to guide the synthesis of new disulfonic stilbene derivatives. Computational chemical methods were used to deduce the important features of the disulfonic stilbene molecule necessary for potent blockade of ORCC and ultimately led to the discovery of the calixarenes. Para-sulfonated calixarenes were found to be potent blockers of ORCC with subnanomolar inhibition constants and exceptionally long block times.

Comparison of -nitro versus -amino 4,4'-substituents of disulfonic stilbenes as chloride channel blockers.

We showed previously that the disulfonic stilbene DNDS (4,4'-dinitrostilben-2,2'-disulfonic acid) was a potent blocker of outwardly rectifying chloride channels (ORCC). The studies reported here were designed to quantify the relationship between electron withdrawal by the 4,4'-substituents and blocker potency. Specifically we compared the blocking effects and molecular properties of the symmetrically substituted 4,4'-diaminostilben-2, 2'-disulfonic acid (DADS) and the hemi-substituted 4-amino, 4'-nitrostilben-2,2'-disulfonic acid (ANDS) with those of DNDS. Blockade was studied using outwardly rectifying colonic chloride channels incorporated into planar lipid bilayers. DADS was 430-fold and ANDS 44-fold less potent than DNDS as blockers of ORCC. Amplitude distribution analysis revealed that all three disulfonic stilbenes act as open channel blockers. Furthermore, this kinetic analysis indicated that the lower potency of DADS and ANDS was due to an increase in off rate. These results support the conclusion that the 4,4'-substituents make an important contribution to blockade by stabilizing the channel-blocker complex. Isopotential electron contour maps illustrated the dramatic shift in charge at the 4,4'-poles of the disulfonic stilbene molecule from electronegative in DNDS to electropositive in DADS as well as the bipolar contour of ANDS. Thus, the greater potency of DNDS results from the symmetric electronegative regions at the 4,4'-poles of the molecule. We hypothesize that the channel protein has two corresponding electropositive areas at the blocker binding site.

A simple assay for agonist-regulated Cl and K conductances in salt-secreting epithelial cells.

We developed a convenient flux assay that permits simultaneous measurement of Cl and K conductance pathways in Cl-secreting epithelial cells. Monolayers of the colonic tumor cell line T84 were preloaded with 125I and 86Rb, and isotope effluxes were monitored by a sample-replace procedure. The adenosine 3',5'-cyclic monophosphate (cAMP)-mediated agonists forskolin and prostaglandin E2 increased I efflux with little effect on Rb efflux, whereas the Ca-mediated agonists ionomycin, A23187, and carbachol increased both I and Rb effluxes. Simultaneous determinations of I and Cl or Rb and K effluxes indicated that I and Rb provide good measures of the effluxes of Cl and K, respectively. Forskolin- and ionomycin-stimulated I effluxes were inhibited by the Cl-channel blockers diphenylamine-2-dicarboxylate (DPC), 5-nitro-2-(3-phenylpropyl-amino)benzoic acid (NPPB), and 2-[cyclopentyl-6,7-dichloro-2,3-dihydro-2-methyl-1-oxo-1H- inden-5-yl)oxy]acetic acid (IAA-94) and by high external K. The Rb efflux evoked by ionomycin was inhibited by the K-channel blockers Ba and charybdotoxin. These findings suggest that I and Rb effluxes provide qualitative estimates of agonist-stimulated Cl and K conductance pathways. Thus this method can provide a simple and relatively inexpensive screening assay for Cl and K conductances in cultured cells to assess the effects of agonist, blockers, or genetic manipulations.

Improved oxygenation promotes CFTR maturation and trafficking in MDCK monolayers.

Culturing airway epithelial cells with most of the apical media removed (air-liquid interface) has been shown to enhance cystic fibrosis transmembrane conductance regulator (CFTR)-mediated Cl(-) secretory current. Thus we hypothesized that cellular oxygenation may modulate CFTR expression. We tested this notion using type I Madin-Darby canine kidney cells that endogenously express low levels of CFTR. Growing monolayers of these cells for 4 to 5 days with an air-liquid interface caused a 50-fold increase in forskolin-stimulated Cl(-) current, compared with conventional (submerged) controls. Assaying for possible changes in CFTR by immunoprecipitation and immunocytochemical localization revealed that CFTR appeared as an immature 140-kDa form intracellularly in conventional cultures. In contrast, monolayers grown with an air-liquid interface possessed more CFTR protein, accompanied by increases toward the mature 170-kDa form and apical membrane staining. Culturing submerged monolayers with 95% O(2) produced similar improvements in Cl(-) current and CFTR protein as air-liquid interface culture, while increasing PO(2) from 2.5% to 20% in air-liquid interface cultures yielded graded enhancements. Together, our data indicate that improved cellular oxygenation can increase endogenous CFTR maturation and/or trafficking.

Biochemical and biophysical identification of cystic fibrosis transmembrane conductance regulator chloride channels as components of endocytic clathrin-coated vesicles.

Cystic fibrosis results from mutations in the gene encoding the CFTR Cl- channel. Although CFTR occurs as an integral component of the plasma membrane, recent studies implicate CFTR in endocytic recycling and suggest that the protein may also exist in intracellular vesicular compartments. To test this, we analyzed CFTR in clathrin-coated vesicles (CCV) purified from cells constitutively expressing CFTR at high levels. CFTR immunoreactivity was detected in CCV by immunoblot and was identified as CFTR based on labeling of immunoprecipitates with protein kinase A and by tryptic phosphopeptide mapping. Fusion of uncoated CCV with planar lipid bilayers resulted in the incorporation of kinase- and ATP-activated Cl- channel activity (7.8 pS at 20 degrees C; 11.9 pS at 37 degrees C), with a linear current-voltage relation under symmetrical conditions. Thus, functional CFTR occurs in CCV. Moreover, CFTR interacts with the plasma membrane specific adaptor complex during endocytosis through clathrin-coated pits. Therefore, the abundance of CFTR in the plasma membrane may be regulated by exocytic insertion and endocytic recycling, and these processes may provide an augmentation to protein kinase A activation as a mechanism for regulating CFTR Cl channels in the plasma membrane.

Glibenclamide blockade of CFTR chloride channels.

The cystic fibrosis transmembrane conductance regulator (CFTR) is a protein kinase A- and ATP-regulated Cl- channel located in the apical membranes of epithelial cells. Previously Sheppard and Welsh (J. Gen. Physiol. 100: 573-591, 1992) showed that glibenclamide, a compound which binds to the sulfonylurea receptor and thus blocks nucleotide-dependent K+ channels, reduced CFTR whole cell current. The aim of this study was to identify the mechanism underlying this inhibition in cell-free membrane patches containing CFTR Cl- channels. Exposure to gliben-clamide caused a reversible reduction in current carried by CFTR which was paralleled by a decrease in channel open probability (Po). The decrease in Po was concentration dependent, and half-maximum inhibition (ki) occurred at 30 microM. Fluctuation analysis indicated a flickery-type block of open CFTR channels. Event duration analysis supported this notion by showing that the glibenclamide-induced decrease in Po was accompanied by interruptions of open bursts [i.e., an apparent reduction in the burst duration (Tburst)] with only a slight reduction in closed time (Tc). The plot of the corresponding open-to-closed (Tburst-1) and closed-to-open (Tc-1) rates as a function of glibenclamide concentration were consistent with a pseudo-first-order open-blocked mechanism and provided estimates of the on rate (kon = 1.17 microM-1S-1), the off rate (koff = 16 s-1), and the dissociation constant (Kd = 14 microM). The difference between the Ki (30 microM) and the Kd (14 microM) is the result expected for a closed-open-blocked model with an initial Po of 0.47. Since the initial Po was 0.50 +/- 0.02 (n = 12), we can conclude that glibenclamide blocks CFTR by a closed-open-blocked mechanism.

Tolbutamide causes open channel blockade of cystic fibrosis transmembrane conductance regulator Cl- channels.

Cystic fibrosis transmembrane conductance regulator (CFTR) is an epithelial Cl- channel that is regulated by protein kinase A and cytosolic nucleotides. Previously, Sheppard and Welsh reported that the sulfonylureas glibenclamide and tolbutamide reduced CFTR whole cell currents. The aim of this study was to quantify the effects of tolbutamide on CFTR gating in excised membrane patches containing multiple channels. We chose tolbutamide because weak (i.e., fast-type) open channel blockers introduce brief events into multichannel recordings that can be readily quantified by current fluctuation analysis. Inspection of current records revealed that the addition of tolbutamide reduced the apparent single-channel current amplitude and increased the open-channel noise, as expected for a fast-type open channel blocker. The apparent decrease in unitary current amplitude provides a measure of open probability within a burst (P0 Burst), and the resulting concentration-response relationship was described by a simple Michaelis-Menten inhibition function. The concentration of tolbutamide causing a 50% reduction of Po Burst (540 +/- 20 microM) was similar to the concentration producing a 50% inhibition of short-circuit current across T84 colonic epithelial cell monolayers (400 +/- 20 microM). Changes in CFTR gating were then quantified by analyzing current fluctuations. Tolbutamide caused a high-frequency Lorentzian (corner frequency, fc > 300 Hz) to appear in the power density spectrum. The fc of this Lorentzian component increased as a linear function of tolbutamide concentration, as expected for a pseudo-first-order open-blocked mechanism and yielded estimates of the on rate (koff = 2.8 +/- 0.3 microM-1 s-1), the off rate (kon = 1210 +/- 225 s-1), and the dissociation constant (KD = 430 +/- 80 microM). Based on these observations, we propose that there is a bimolecular interaction between tolbutamide and CFTR, causing open channel blockade.

Activation of DeltaF508 CFTR in an epithelial monolayer.

The DeltaF508 mutation leads to retention of cystic fibrosis transmembrane conductance regulator (CFTR) in the endoplasmic reticulum and rapid degradation by the proteasome and other proteolytic systems. In stably transfected LLC-PK1 (porcine kidney) epithelial cells, DeltaF508 CFTR conforms to this paradigm and is not present at the plasma membrane. When LLC-PK1 cells or human nasal polyp cells derived from a DeltaF508 homozygous patient are grown on plastic dishes and treated with an epithelial differentiating agent (DMSO, 2% for 4 days) or when LLC-PK1 cells are grown as polarized monolayers on permeable supports, plasma membrane DeltaF508 CFTR is significantly increased. Moreover, when confluent LLC-PK1 cells expressing DeltaF508 CFTR were treated with DMSO and mounted in an Ussing chamber, a further increase in cAMP-activated short-circuit current (i.e., approximately 7 microA/cm2; P < 0.00025 compared with untreated controls) was observed. No plasma membrane CFTR was detected after DMSO treatment in nonepithelial cells (mouse L cells) expressing DeltaF508 CFTR. The experiments describe a way to augment DeltaF508 CFTR maturation in epithelial cells that appears to act through a novel mechanism and allows insertion of functional DeltaF508 CFTR in the plasma membranes of transporting cell monolayers. The results raise the possibility that increased epithelial differentiation might increase the delivery of DeltaF508 CFTR from the endoplasmic reticulum to the Golgi, where the DeltaF508 protein is shielded from degradative pathways such as the proteasome and allowed to mature.

Colonic Cl channel blockade by three classes of compounds.

We compared the potency and inhibitory actions of three different classes of organic acids on a Cl channel derived from colonic enterocyte plasma membrane vesicles. Chloride channels were incorporated into planar lipid bilayer membranes to examine the effects of the anthranilic acids, diphenylamine 2-carboxylic acid (DPC) and 5-nitro-2-(3-phenylpropylamino)benzoic acid (NPPB), the indanyl alkanoic acids, 2-[(2-cyclopentyl-6,7-dichloro-2,3-dihydro-2-methyl-1-oxo-1H-inden -5-yl)oxy] acetic acid (IAA-94) and its stereoenantiomer IAA-95, and the disulfonic stilbene, 4,4'-dinitro-stilbene-2,2'-disulfonic acid (DNDS). Except for DNDS, each of the blockers was equipotent from both the outer membrane and the cytoplasmic side of the channel protein. The potency order from the outmembrane side was DNDS greater than IAA-94 = IAA-95 greater than NPPB much greater than DPC. In contrast, the potency order from the cytoplasmic side was IAA-94 = IAA-95 greater than NPPB greater than DNDS much greater than DPC. DPC and NPPB caused a concentration-dependent decrease in the single-channel conductance (fast block). DNDS, IAA-94, and IAA-95 caused a flickery-type block and a concentration-dependent decrease in open-channel probability. Kinetic analysis revealed that blockade could be explained by a linear closed-opened-blocked kinetic scheme. Similarities in the electrostatic potential maps of these open-channel blockers suggest they may bind to a single shared binding site within the channel protein.

Synthetic chloride channel restores glutathione secretion in cystic fibrosis airway epithelia.

Cystic fibrosis (CF), an inherited disease characterized by defective epithelial Cl- transport, damages lungs via chronic inflammation and oxidative stress. Glutathione, a major antioxidant in the epithelial lung lining fluid, is decreased in the apical fluid of CF airway epithelia due to reduced glutathione efflux (Gao L, Kim KJ, Yankaskas JR, and Forman HJ. Am J Physiol Lung Cell Mol Physiol 277: L113-L118, 1999). The present study examined the question of whether restoration of chloride transport would also restore glutathione secretion. We found that a Cl- channel-forming peptide (N-K4-M2GlyR) and a K+ channel activator (chlorzoxazone) increased Cl- secretion, measured as bumetanide-sensitive short-circuit current, and glutathione efflux, measured by high-performance liquid chromatography, in a human CF airway epithelial cell line (CFT1). Addition of the peptide alone increased glutathione secretion (181 +/- 8% of the control value), whereas chlorzoxazone alone did not significantly affect glutathione efflux; however, chlorzoxazone potentiated the effect of the peptide on glutathione release (359 +/- 16% of the control value). These studies demonstrate that glutathione efflux is associated with apical chloride secretion, not with the CF transmembrane conductance regulator per se, and the defect of glutathione efflux in CF can be overcome pharmacologically.

Chlorzoxazone or 1-EBIO increases Na(+) absorption across cystic fibrosis airway epithelial cells.

Previous studies demonstrated that chlorzoxazone or 1-ethyl-2-benzimidazolinone (1-EBIO) enhances transepithelial Cl(-) secretion by increasing basolateral K(+) conductance (G(K)) (Singh AK, Devor DC, Gerlach AC, Gondor M, Pilewski JM, and Bridges RJ. J Pharmacol Exp Ther 292: 778-787, 2000). Hence these compounds may be useful to treat cystic fibrosis (CF) airway disease. The goal of the present study was to determine whether chlorzoxazone or 1-EBIO altered ion transport across Delta F508-CF transmembrane conductance regulator homozygous CFT1 airway cells. CFT1 monolayers exhibited a basal short-circuit current that was abolished by apical amiloride (inhibition constant 320 nM) as expected for Na(+) absorption. The addition of chlorzoxazone (400 microM) or 1-EBIO (2 mM) increased the amiloride-sensitive I(sc) approximately 2.5-fold. This overlapping specificity may preclude use of these compounds as CF therapeutics. Assaying for changes in the basolateral G(K) with a K(+) gradient plus the pore-forming antibiotic amphotericin B revealed that chlorzoxazone or 1-EBIO evoked an approximately 10-fold increase in clotrimazole-sensitive G(K). In contrast, chlorzoxazone did not alter epithelial Na(+) channel-mediated currents across basolateral-permeabilized monolayers or in Xenopus oocytes. These data further suggest that alterations in basolateral G(K) alone can modulate epithelial Na(+) transport.

Relationship between airway ion transport and a mild pulmonary disease mutation in CFTR.

Patients with cystic fibrosis (CF) display defects in airway ion transport, but the influence of airway transport phenotype on improved prognosis is not known. We studied airway bioelectric properties in five CF patients with the rare A455E mutation that is associated with mild pulmonary disease. We also evaluated five patients possessing premature truncation mutations (G542X and R553X) for which an association with mild pulmonary disease has not been as well established. We found no evidence in vivo that a mild lung disease mutation in the CF transmembrane regulator gene (CFTR) led to correction or partial correction of: (1) unstimulated Cl- secretion; (2) beta-agonist-activated Cl- secretion; (3) basal sodium reabsorption; or (4) amiloride-sensitive airway sodium transport. Early phase therapeutic trials in CF, including human gene transfer trials, rely heavily on improvements in airway potential difference to identify promising interventions and an improved prognosis. Based on our findings in a naturally occurring group of CF patients with an improved pulmonary prognosis (A455E), one can argue that marked clinical benefit might be possible without any improvement whatsoever in airway bioelectric phenotype. Moreover, if genetic modifiers exist that influence the severity of a particular CFTR mutation (e.g., A455E), these may be independent of human airway Cl-secretion in vivo, since we detected minimal Cl--secretory responses in patients with A455E.