Department of Physiology, McGill University, Montreal, Quebec, Canada











































































1. Introduction

The cystic fibrosis transmembrane conductance regulator (CFTR) is a membrane glycoprotein which consists of 1480 amino acids and is mutated in the autosomal recessive disease cystic fibrosis (CF). CF is characterized by abnormal ion transport across exocrine epithelia. Defects in CFTR processing or function lead to viscous secretions in the pancreatic ducts, airways, and intestine, however many aspects of its pathogenesis remain poorly understood.

CFTR has two membrane domains (TMD1 and TMD2), each of which contains six transmembrane segments that may form alpha helices. Nucleotide binding domains (called NBD1 and NBD2, respectively) follow each TMD in the linear sequence and place CFTR in a large superfamily of “ATP binding cassette” (ABC) transport proteins that are prevalent in organisms from bacteria to humans. Various members of the protein superfamily transport a vast array of substrates that include peptides (eg STE6,TAP1/TAP2), amino acids and sugars (eg HisP and MalK, respectively), lipophilic drugs and xenobiotics (eg. MDR and MRP1, respectively) and phospholipids (eg. MDR2 (11)). CFTR is the only known ion channel in the superfamily and the only member of the family that has a regulatory (R) domain.


Localization of the small GTPase rab5 (green) and CFTR (red). Baby hamster kidney cells were co-transfected with cDNA encoding Rab5 fused to green fluorescent protein and CFTR that contains a biotinylation target sequence. CFTR was specificially biotinylated using recombinant BirA from E. coli and visualized by binding fluorescent streptavidin.

2. CFTR is an anion channel

When the cftr gene was functionally expressed in mammalian and insect cells, a cAMP-stimulated conductance appears that is mediated by ohmic Cl- channels having relatively low (7 – 10 pS) conductance. These and other single channel properties (slow gating, flickering at negative membrane potentials, weak inhibition by external disulfonic stilbenes) are reminiscent of endogenous Cl- channels that were characterized previously in epithelia, therefore it was proposed that CFTR is itself a non-rectifying, low-conductance Cl channel. Mutations in predicted transmembrane segments altered the selectivity of macroscopic CFTR currents, providing further evidence that CFTR forms a pore in the membrane. The channel activity of CFTR was formally demonstrated when it was purified to homogeneity and reconstituted into planar bilayers.

Except for mouse NBD1 (aa 389-673), which has been solved, the structure of intact CFTR has not yet been solved, nevertheless, amino acids that line its pore have been inferred by comparing ion permeation in wild-type and mutant CFTR channels (i.e. their anion permeability ratios, conductances and blocker sensitivities). The sixth transmembrane segment (TM6) is likely to line the pore because mutations there affect channel selectivity, conductance , multi-ion pore behaviour, and blocker sensitivity, however many mutagenesis/patch clamp results are difficult to interpret and need confirmation by other (eg biochemical) methods. PCl/Pcation of 8-14 were reported for channels expressed in human colonic T84 and intestinal CaCo2 cell lines. Among the halides, permeability ratios (PX/PCl) generally follow an (inverse) lyotropic series with large, weakly hydrated ions having the highest values. One exception is iodide (I-), which gives inconsistent results even within laboratories. PI/PCl was near unity when determined at the single channel level, but is lower (~0.4) in whole cell recordings and other macroscopic measurements. Indeed, I- currents were not detected in one study of endogenous CFTR channels in cultured thyroid cells, and iodide block of Cl- currents through CFTR has also been reported.

3. Phosphorylation of the R domain regulates CFTR channel gating:

Phosphorylation of the R domain stimulates CFTR channel activity and may also influence its interactions with other proteins. Regulation by PKA and PKC are well established although other serine-threonine and tyrosine kinases may also have a role. Sites that become conspicuously phosphorylated under in vivo conditions are RRNS660I, KRKNS700I, RKVS795L and RRLS813 Q. Phosphorylation of other sites (i.e., RKFS712I, RIS753V and RRQS768V) has been detected in vitro, presumably because kinase activity is higher (or phosphatase activity lower) in vitro than in intact cells. The role of consensus sequences that are not detectably phosphorylated in vivo are rather mysterious probably have important roles since they can support ~ 50% channel activity and are highly conserved between species despite sequence divergence within the R domain. Some PKA sites (S737, S768) down-regulate channel activity when phosphorylated. Protein kinase C (PKC) is much less effective in activating CFTR channels compared to PKA (<10%) but does enhance CFTR responsiveness to PKA by increasing the rate and magnitude of PKA activation and helping to maintain CFTR in a PKA-responsive state. Several PKC consensus sequences are essential for CFTR to be competent for gating, and at least one of these is actually phosphorylated by PKC. There is also evidence for inhibitory PKC sites on the R domain.

4. CFTR is down-regulated by a membrane-associated protein phosphatase:

CFTR channels run down soon after membrane patches are excised from cAMP-stimulated cells. This spontaneous decline in open probability (Po) has been attributed to phosphatase activity in the excised patch because channels could be reactivated by exposure to MgATP + PKA. Rundown is not inhibited significantly by high concentrations of okadaic acid, and does not require Ca or calmodulin, indicating that it is not mediated by PP1, PP2A or PP2B. However, rundown is sensitive to the alkaline phosphatase inhibitors levamisole and bromotetramisole at high concentrations (> 200 mM) that inhibit all four types of serine/threonine phosphatases in biochemical studies. Reducing Mg2+ concentration slows rundown of CFTR channel activity in excised patches, consistent with the dependence of PP2C phosphatase activity on millimolar concentrations of Mg2+. In patches from T84 cell monolayers, deactivation of cAMP-stimulated currents was also insensitive to okadaic acid and calyculin A, suggesting most CFTR deactivation in epithelia is also mediated by PP2C. PP2C is the main phosphatase physically associated with CFTR but it is not yet known if this interaction is direct or mediated by other proteins.

5. Nucleotides are required for CFTR gating

ATP is required for CFTR channel gating, but CFTR activity has been recorded in nominally Mg-free solutions that should prevent hydrolysis. Moreover, the temperature dependence of CFTR gating in planar bilayers suggests that the open and closed states of CFTR have similar free energies and therefore ATP hydrolysis energy is not required for closing the channel and it has been proposed that the energy comes instead from the interaction of MgATP with CFTR. Different functions have been ascribed to NBD1 and NBD2 based on electrophysiological studies of mutant channels. Photoaffinity labeling studies with  8-azido-[a-32P]ATP indicate that nucleotides bind stably at NBD1 whereas rapid ATP turnover occurs at NBD2.

6. CFTR interacts physically and functionally with other proteins

CFTR may be associated in a regulatory complex with other proteins at the plasma membrane where it has been reported to affect other transport proteins such as the epithelial sodium channel (ENaC). The best characterized interaction is that involving the amino terminal tail (N-tail) of CFTR and syntaxin 1A. Syntaxin 1A is part of the vesicle fusion apparatus at neuronal synapses and is also expressed in airway epithelium. It binds to the N-tail of CFTR and inhibits its channel activity, apparently by disrupting interaction of the N-tail with the R domain of CFTR. The carboxy-terminal amino acids of CFTR (DTRL) form a motif that binds PDZ (Post synaptic density 95, Discs large, ZO-1) domains in scaffolding/adaptor proteins such as EBP50 (also called NHERF for sodium hydrogen exchange regulatory factor). Other PDZ proteins such as E3KARP, CAL (CFTR associated ligand), CAP70, and Shank2 also been proposed to bind to the C terminus of CFTR and regulate its maturation, oligomeric state, channel gating, or phosphorylation. Whether protein-protein interactions mediate CFTR’s effects on other transporters such as the epithelial sodium channel ENaC remains to be established.


Book Chapter (PDF file) :
"The Cystic Fibrosis Transmembrane Conductance Regulator (ABCC7)"



1. Introduction

It has been appreciated since the 1920s that extracellular ATP regulates the contractile activity of cardiac and smooth muscle cells.  ATP release by sensory neurons and its effect on vascular resistance were demonstrated later, and Burnstock and colleagues demonstrated the importance of extracellular ATP for nociception in non-adrenergic, non-cholinergic sensory nerves of the gut.  It is now widely accepted that extracellular ATP mediates autocrine and paracrine signaling throughout the body and plays diverse roles that include regulating epithelial secretion and signaling visceral pain.  Most cells release nucleotides and express multiple purinoceptor subtypes that signal through many pathways including activation of phospholipase C and mobilization of calcium and PKC, and activate a wide range of cell-specific functions.  Epithelial cells express receptors for nucleotides and adenosine and may be regulated “downstream” by luminal secretion of ATP into exocrine glands and renal tubules.

2. Mechanisms and regulation of cellular ATP release

ATP release is not well understood in non-excitable cells and multiple mechanisms may coexist in the same cell.  ATP can permeate through high-conductance anion and connexin hemichannels, and channel-mediated efflux has been proposed in many cell types including Xenopus oocytes during large voltage pulses.  Other studies of non-excitable cells suggest that release occurs principally by vesicle exocytosis, or perhaps by a combination of mechanisms, eg exocytotic insertion of ATP channels.  We found that ATP release by airway and colonic epithelial cells is acutely sensitive to mechanical stimulation; simply tipping the culture dish when collecting samples or ejecting puffs of solution onto the cells is sufficient to elevate extracellular [ATP] by an order of magnitude.  Mechanically-induced ATP release from native rabbit urinary bladder occurs when it is stretched under physiological conditions, and this may play a role in stimulating micturition.  Recent studies of Xenopus oocytes indicate that the mechanical stimulation increases ATP efflux by 5,000-fold, and this can be blocked by maneuvers that inhibit vesicle trafficking and integrin binding.  The sites and mechanisms of ATP release and the signalling pathways activated by stretch and other mechanical stimuli remain poorly understood.  Once released, ATP is rapidly metabolized to other nucleoside tri-, di- and monophosphates and to adenosine by ecto-enzymes, and these metabolites may also have regulatory effects. 

3. Approaches for studying epithelial ATP release

 A commonly used, and highly sensitive method for detecting ATP is the luciferin-luciferase assay.  In this reaction, luciferase from the firefly Photinus pyralis catalyzes formation of luciferin-AMP, which reacts with molecular oxygen to form oxyluciferin-AMP in an excited state.  A single photon is emitted when the intermediate breaks down into oxyluciferin, CO2, and AMP.  Coenzyme A (CoA) prolongs the light emission but is not required for the reaction.  Luciferin-luciferase is remarkably sensitive, detecting less than 1 femtomolar (10-15 M) levels of ATP in solution. 

Extracellular luciferase can be used to detect released ATP without sampling the solution, however careful controls are needed to ensure that experimental interventions do not affect the bioluminescence reaction.  Bioluminescence is several orders of magnitude weaker than fluorescence, therefore imaging signals generated by the luciferase-luciferin reaction requires an ultra-sensitive (photon counting) camera.  However, if ATP release occurs by vesicle fusion it should be possible to label the vesicles with a targeted fluorescent protein and image single fusion events. 

4. Possible role of ATP release in disease

Stretching rabbit urinary bladder or ureter by imposing a transmural hydrostatic gradient in vitro stimulates ATP release to the basolateral side.  This ATP efflux, which had only minor effects on transepithelial transport in the bladder, was proposed to function in sensing bladder distension.  Extracellular ATP has recently been implicated in the disorder interstitial cystitis perhaps by increasing the sensitivity of subepithelial sensory neurons to other (normal physiological) stimuli.  Interestingly, the urothelium expresses transient receptor potential (TRP) channels, which belong to a large family of receptors that sense chemical, thermal and (perhaps mechanical) stimuli.  Whether epithelial vallinoid receptors are involved directly in nociception or form part of a cell stretch receptor-like complex in the epithelium is unknown.  Regardless, an involvement of ATP release in interstitial cystitis is suggested because stretch-induced ATP release is elevated in urinary bladder epithelial cells cultured from biopsies obtained from IC patients.  The fact that this release remains elevated after the cells have been passaged several times implies that excessive ATP release may be a fundamental defect that predisposes individuals to interstitial cystitis.  Elevated ATP release is also observed in cats with a disorder that closely resembles interstitial cystitis.  ATP release may have a role in other pathological conditions including those of the airways, where ATP regulates mucociliary clearance by stimulating ciliary beat frequency and mucus secretion.  


This page was last edited on 18 August, 2004