CFTR (cystic fibrosis transmembrane conductance regulator) is an epithelial Cl channel inhibited with high affinity and selectivity by the thiazolidinone compound CFTRinh-172. In the present study, we provide evidence that CFTRinh-172 acts directly on the CFTR. We introduced mutations in amino acid residues of the sixth transmembrane helix of the CFTR protein, a domain that has an important role in the formation of the channel pore. Basic and hydrophilic amino acids at positions 334–352 were replaced with alanine residues and the sensitivity to CFTRinh-172 was assessed using functional assays. We found that an arginine-to-alanine change at position 347 reduced the inhibitory potency of CFTRinh-172 by 20–30-fold. Mutagenesis of Arg347 to other amino acids also decreased the inhibitory potency, with aspartate producing near total loss of CFTRinh-172 activity. The results of the present study provide evidence that CFTRinh-172 interacts directly with CFTR, and that Arg347 is important for the interaction.

INTRODUCTION

The CFTR (cystic fibrosis transmembrane conductance regulator) protein is a Cl channel that has been studied extensively because of its involvement in the genetic disease cystic fibrosis and its pathogenic role in secretory diarrhoeas [1,2]. CFTR is composed of 12 TMDs (transmembrane domains), two NBDs (nucleotide-binding domains) and a cytosolic R region that contains multiple sites for cAMP-dependent phosphorylation [3,4]. Transport of anions through the transmembrane helices is controlled by the NBDs. It is believed that these structures interact with two molecules of ATP to form a dimer and that binding/hydrolysis of ATP molecules control CFTR channel opening [5].

In contrast with cation channels, which are characterized by a wide panel of specific and potent blockers, including natural toxins, anion channels are sensitive to a limited number of non-selective inhibitors that have relatively low potencies. With regards to CFTR, previously known inhibitors, including glibenclamide, niflumic acid and diphenylamine-2-carboxylic acid, have low potency and poor selectivity [1,6]. High-throughput screening of large chemical libraries has allowed the discovery of more selective CFTR inhibitors [2]. So far, inhibitors with two chemical scaffolds have been identified: the thiazolidinone CFTRinh-172 [7] and the glycine hydrazide GlyH-101 [8]. The former compound reduces CFTR activity at nanomolar concentrations in a voltage-independent manner by increasing the time spent in the closed state [9]. The latter is an open-channel blocker, acting from the extracellular side to occlude the channel pore. Other recently discovered CFTR inhibitors include α-aminoazaheterocycle-methylglyoxal adducts [10] and a peptide toxin [11].

The mechanism and site of action of CFTRinh-172 are unknown. Experiments indicating that CFTRinh-172 is not effective on other ion channels and transporters suggest that it may be a direct inhibitor of the CFTR protein [7]. Based on its lack of voltage-dependence (despite its negative charge in solution) and the effect on CFTR kinetics (closed-state stabilization), we postulated that CFTRinh-172 is not an open-channel blocker but an inhibitor of channel gating [9]. Since CFTR gating is controlled by NBDs, we hypothesized that CFTRinh-172 binds to a site located within these protein domains. However, NBD mutants such as G551D, G1349D or P574H, do not show an altered sensitivity to CFTRinh-172 [9,12]. In contrast, the potency of CFTR activators, whose effect involves direct interaction with the NBDs, is greatly altered by mutations such as G551D which resides in NBD1 [1315]. Therefore we considered the possibility that CFTRinh-172 binds to other regions of the CFTR protein. In the present study, we have performed an alanine scanning of the sixth TMD, a CFTR structure that is important in the formation of the CFTR pore and in anion transport [1619]. We found that mutation of Arg347 markedly reduces CFTR sensitivity to CFTRinh-172.

EXPERIMENTAL

Cell culture

FRT (Fisher rat thyroid) epithelial cells were cultured on plastic in Coon's modified F12 medium supplemented with 5% (v/v) FCS (foetal calf serum), 2 mM L-glutamine, 100 units/ml penicillin and 100 μg/ml streptomycin. COS-7 cells were instead cultured in DMEM (Dulbecco's modifed Eagle's medium)/F12 (1:1) medium with 10% (v/v) FCS, 2 mM L-glutamine, 100 units/ml penicillin and 100 μg/ml streptomycin. HEK (human embryonic kidney)-293 cells were grown in DMEM with Glutamax containing 10% (v/v) FCS, 100 units/ml penicillin and 100 μg/ml streptomycin.

CFTR mutagenesis

Mutations were introduced into a pcDNA3.1 plasmid carrying the wild-type CFTR coding sequence using the QuikChange® XL site-directed mutagenesis kit (Stratagene) following the manufacturer's protocol. The CFTR coding sequence was fully sequenced to confirm the mutagenesis reaction and to exclude the presence of undesired mutations.

CFTR microfluorimetric functional assay

A CFTR functional assay was performed on transiently transfected COS-7 cells using the halide-sensitive YFP (yellow fluorescent protein) YFP-H148Q/I152L [20]. Accordingly, COS-7 cells were seeded in 96-well microplates (25000 cells/well) in 100 μl of antibiotic-free culture medium (DMEM/F12). After 6 h, cells were co-transfected with pcDNA3.1 plasmids carrying the coding sequence for CFTR (wild-type or mutant) and for the halide-sensitive YFP. The transfection reagent was Lipofectamine™ 2000 (Invitrogen). For each well, 0.2 μg of plasmid DNA and 0.5 μl of Lipofectamine™ 2000 were first pre-mixed in 50 μl of OPTI-MEM (Invitrogen) to generate transfection complexes, and this was added to the cells. After 24 h, the complexes were removed by replacement with fresh culture medium. The CFTR functional assay was performed after a further 24 h. For this purpose, the cells were washed two times with PBS and incubated for 20–30 min with 60 μl of PBS containing 20 μM forskolin with and without CFTRinh-172 at various concentrations. After incubation, cells were transferred on to an Olympus IX 50 fluorescence microscope equipped with a 20× objective, optical filters for detection of EYFP (enhanced YFP) fluorescence (Chroma; excitation: HQ500/20X, 500±10 nm; emission: HQ535/30M, 535±15 nm; dichroic: 515 nm) and a photomultiplier tube (Hamamatsu). For each well, cell fluorescence was continuously measured before and after addition of 165 μl of a modified PBS containing 137 mM NaI instead of NaCl (final NaI concentration in the well: 100 mM). Where necessary, the added solution contained CFTRinh-172 at the same concentration present in the well. The output from the photomultiplier tube was digitized using a PowerLab 2/25 acquisition system (ADInstruments) and stored on a Macintosh computer. After background subtraction, cell fluorescence recordings were normalized for the initial average value measured before the addition of I. The signal decay caused by YFP fluorescence quenching was fitted with a double exponential function to derive the maximal slope that corresponds to the initial influx of I into the cells [15,21]. Maximal slopes were converted into rates of variation of the intracellular I concentration (in mM/s) using the equation:

 
formula

where KI is the affinity constant of YFP for I [20], and F/F0 is the ratio of the cell fluorescence at a given time versus the initial fluorescence. Initial fluorescence values were similar among cells transfected with different CFTR constructs. This finding allowed us to conclude that differences in I influx were due to intrinsic differences in the halide transport ability of CFTR mutants and not to changes in cell density or transfection efficiency.

Stable transfections

FRT cells expressing YFP-H148Q/I152L were plated on to 60 mm Petri dishes and transfected with the CFTR plasmids using Lipofectamine™ 2000. Cell clones co-expressing both CFTR and YFP proteins were generated by continuous selection with 0.5 mg/ml Hygromycin B and 0.75 mg/ml G418 (Geneticin). Positive clones were identified by performing the CFTR functional assay in a microplate reader, as previously described [21].

CFTR protein analysis: transient transfection and immunoprecipitation/Western blot analysis

All cells were grown on 60- or 100-mm-diameter dishes. Subconfluent cells (60%) were transfected by lipofection using Lipofectamine™ and Plus reagent (Invitrogen) according to the manufacturer's protocol. Confluent monolayers were harvested and used 24 h post-transfection for immunoprecipitation.

Cells grown on 100-mm-diameter dishes were harvested 24 h after transfection in 1× PBS, pelleted at 2000 g at 4 °C for 10 min, and then resuspended in 600 μl of cold lysis buffer [20 mM Hepes (pH 7), 150 mM NaCl, 1 mM EDTA and 1% Igepal; all from Sigma] supplemented with complete protease inhibitors (Roche Applied Sciences) for 30 min at 4 °C. Cell lysates were spun at 20000 g for 15 min at 4 °C to pellet insoluble material. Supernatants were pre-cleared three times with Pansorbin cells (Calbiochem) and 150 μl of 5× RIPA [250 mM Tris/HCl (pH 7.5), 750 mM NaCl, 5% (v/v) Triton X-100, 5% sodium deoxycholate and 0.5% SDS] were added before the addition of 0.8 μg of mAb (monoclonal antibody) 24-1 (R&D Systems) overnight at 4 °C on a rotating wheel. A total of 25 μl of Pansorbin cells were added to the lysates and the incubation continued for a further 1 h at 4 °C. These were then spun at 12000 g for 1 min at 4 °C and the pellets were washed three more times with 1× RIPA. Pellets were resuspended in ESB buffer [125 mM Tris/HCl (pH 6.8), 5% (w/v) SDS, 25% (w/v) sucrose and 5% (v/v) 2-mercaptoethanol], heated at 37 °C for 15 min and spun at 12000 g for 5 min at 4 °C. All of the immunoprecipitated material was separated on SDS/PAGE (7.5% gel) and electroblotted from the gels on to PVDF membrane (GE Healthcare). CFTR was immunodetected using mAb MM13-4 (1:1000) followed by HRP (horseradish peroxidase)-conjugated anti-mouse antibody (1:50000), and visualized by chemiluminescence with West Dura kit (Pierce) according to the manufacturer's protocol. Direct recording of the chemiluminescence was performed using the CCD (charge-coupled-device) camera of the GeneGnome analyser, and quantification was achieved using the GeneTools software (Syngene BioImaging Systems, Synoptics, sold by Ozyme).

Immunoprecipitation/PKA (protein kinase A) assay

Transiently transfected HEK-293 cells grown on 60-mm-diameter dishes were harvested 24 h post-transfection in 1×PBS, pelleted at 2000 g at 4 °C for 10 min, and then resuspended in 200 μl of cold lysis buffer supplemented with complete protease inhibitors (Roche Applied Sciences) for 30 min at 4 °C. After preclearing with Pansorbin cells, 50 μl of 5×RIPA was added and the CFTR protein was immunoprecipitated as described above with 0.4 μg of mAb 24-1. For the PKA assay, the immunoprecipitated proteins were phosphorylated in vitro with 5 units of the catalytic subunit of PKA (Promega) and 10 μCi of [γ-33P]ATP (GE Healthcare), and then were separated by SDS/PAGE (5% gel), dried and autoradiographed. Radioactivity was quantitated by radioanalytic scanning (using a Molecular Dynamics PhosphoImager).

Transepithelial Cl current measurements

FRT cells expressing wild-type or mutant CFTR were seeded at a high density on Snapwell permeable supports (Corning Costar; 500000 cells/insert). The culture medium was replaced every 48 h on both apical and basolateral sides. At 4–7 days from plating, the Snapwell supports carrying the FRT monolayers were mounted in a home-made Ussing chamber-like system. The apical chamber was filled with a low-Cl-containing solution [65 mM NaCl, 65 mM sodium gluconate, 2.7 mM KCl, 1.5 mM KH2PO4, 2 mM CaCl2, 0.5 mM MgCl2, 10 mM glucose and 10 mM Na-Hepes (pH 7.4)]. The basolateral chamber was instead filled with 130 mM NaCl, 2.7 mM KCl, 1.5 mM KH2PO4, 1 mM CaCl2, 0.5 mM MgCl2, 10 mM glucose and 10 mM Na-Hepes (pH 7.4). The basolateral membrane was permeabilized with 250 μg/ml amphotericin B for 30 min before starting the recording. Experiments were performed at 37 °C with a continuous bubbling with air on both sides.

Hemichambers were connected to a DVC-1000 voltage clamp (World Precision Instruments) via Ag/AgCl electrodes and 1 M KCl agar bridges. The transepithelial electrical potential difference was clamped at zero to measure CFTR Cl currents across the apical membrane.

Synthesis of a zwitterionic analogue of CFTRinh-172: thiaxo N-O {5-(1-oxido-4-pyridinyl)methylene-2-thioxo-3-[3-(trifluoromethyl)phenyl]-4-thiazolidinone}

A mixture of 2-thioxo-3-(3-trifluoromethyl phenyl)-4-thiazolidinone (55 mg, 0.2 mmol; synthesized as described previously [22]), 4-pyridinecarboxaldehyde-1-oxide (25 mg, 0.2 mmol) and sodium acetate (10 mg) in glacial acetic acid (0.5 ml) was refluxed for 8 h. The solvent was evaporated, and the residue was crystallized from ethanol and further purified by normal phase flash chromatography to yield 22 mg of a yellow powder (yield 29%); melting point: 209–210 °C (decomposed); MS (ES+) (m/z): [M+H]+ calculated for C16H9F3N2O2S2, 382.39, found 383.01.

Patch-clamp recordings

The whole-cell configuration of the patch-clamp technique was used to measure the effect of CFTRinh-172 on CFTR Cl currents. Experiments were carried out on FRT cells with stable expression of wild-type or R347A CFTR. Borosilicate glass pipettes were pulled on a vertical two steps puller to a final resistance of 1–3 MΩ measured in the working solution. Currents were recorded with an EPC-7 patch-clamp amplifier (List Medical) using PULSE software (Heka) and sampled at 2 kHz with an Instrutech ITC-16 AD/DA interface. Data were filtered at 1 kHz and analysis was done with IgorPro (Wavemetrics). The bath solution contained 150 mM NaCl, 1 mM CaCl2, 1 mM MgCl2, 10 mM glucose, 10 mM mannitol and 10 mM Na-Hepes (pH 7.4). The pipette (intracellular) solution contained 120 mM CsCl, 10 mM TEA-Cl (triethanolamine chloride), 0.5 mM EGTA, 1 mM MgCl2, 40 mM mannitol, 10 mM Cs-Hepes and 1 mM MgATP (pH 7.3). To build current-to-voltage relationships, pulses to membrane voltages from −100 to 100 mV were applied in 20 mV steps from a holding potential of 0 mV. CFTRinh-172 was applied by extracellular perfusion. All experiments were performed at room temperature (22–24 °C).

Data analysis and statistics

Dose–response relationships from each experiment were fitted with the Hill equation using Igor software (WaveMetrics) to calculate Ki and the Hill coefficient. We did not test CFTRinh-172 at concentrations higher than 50 μM due to limited solubility of this compound at high micromolar concentrations. Data are reported as representative traces and as means±S.E.M. To determine the significance of differences between groups of data, we used ANOVA by means of Statview software. Differences were considered statistically significant when P<0.01.

RESULTS

Anion transport and sensitivity to CFTRinh-172 in a series of CFTR point mutants was measured with the functional assay based on the halide-sensitive YFP. COS-7 cells were transiently co-transfected with the plasmids coding for the fluorescent protein and the wild-type or mutant CFTR. Anion transport was determined from the rate of fluorescence decrease upon extracellular addition of a modified PBS containing I instead of Cl. Cells expressing wild-type CFTR and stimulated with the cAMP-elevating agent forskolin showed a fast fluorescence decrease, as expected, due to the large CFTR-mediated influx of I which replaces intracellular Cl (Figure 1A). The rate of I influx was 45-fold greater than that in mock-transfected cells (Figure 1B). Mutants of the sixth TMD showed different rates of anion transport. T338A and R347A were similar to wild-type CFTR. R334A and S341A showed reduced anion transport, although this was significantly greater than cells transfected with the fluorescent protein alone (Figure 1B). More marked was the effect of R352A, which showed undetectable levels of CFTR activity. This finding may depend on the observed ability of mutations to Arg352 to strongly decrease CFTR single-channel conductance [23]. The absence of function did not allow us to measure the potency of CFTRinh-172 for the R352A mutant.

Mutants of the sixth TMD

Figure 1
Mutants of the sixth TMD

(A) Representative traces showing normalized cell fluorescence recordings and quenching upon I addition in COS-7 cells transfected with wild-type (top panel) or R347A (bottom panel) CFTR. Experiments were performed after stimulation of CFTR with 20 μM forskolin in the presence or absence of 10 μM CFTRinh-172. (B) Rate of I transport measured from experiments as in (A) in COS-7 cells transfected with the wild-type or mutant CFTR. Cells were stimulated with forskolin in the absence of CFTRinh-172. Wild-type CFTR and all mutants except R352A showed an I influx significantly higher (P<0.01) than mock-transfected cells. (C and D) Dose–response relationships determined for CFTRinh-172 on wild-type and mutant CFTR. Values are the means±S.E.M. of four to ten experiments. WT, wild-type.

Figure 1
Mutants of the sixth TMD

(A) Representative traces showing normalized cell fluorescence recordings and quenching upon I addition in COS-7 cells transfected with wild-type (top panel) or R347A (bottom panel) CFTR. Experiments were performed after stimulation of CFTR with 20 μM forskolin in the presence or absence of 10 μM CFTRinh-172. (B) Rate of I transport measured from experiments as in (A) in COS-7 cells transfected with the wild-type or mutant CFTR. Cells were stimulated with forskolin in the absence of CFTRinh-172. Wild-type CFTR and all mutants except R352A showed an I influx significantly higher (P<0.01) than mock-transfected cells. (C and D) Dose–response relationships determined for CFTRinh-172 on wild-type and mutant CFTR. Values are the means±S.E.M. of four to ten experiments. WT, wild-type.

Anion transport was measured in the presence of CFTRinh-172. As shown by representative traces and dose–response relationships (Figures 1A, 1C and 1D), most of the CFTR mutants showed sensitivity to CFTRinh-172 comparable with that of the wild-type protein with Ki approx. 1–3 μM (Table 1). The exception was R347A. This mutant showed a dramatic shift in sensitivity with a Ki of 44.98±4.71 μM (Figures 1A and 1D, and Table 1). We also introduced a mutation at position 349 (an alanine residue replaced by a serine residue). This amino acid change made the sequence of the sixth TMD identical with that of murine CFTR. The A349S mutant had anion transport ability (Figure 1B) and sensitivity to CFTRinh-172 (Ki=1.23±0.41 μM, Table 1) similar to those of the wild-type channel. This result is consistent with the strong CFTRinh-172 inhibition of mouse CFTR [24].

Table 1
Properties of wild-type and mutant CFTR

Sensitivity to CFTRinh-172 and I transport ability for wild-type and mutant CFTR. Results were obtained with the halide-sensitive YFP assay on transfected COS-7 cells. The Table shows Ki and Hill coefficients (calculated by fitting CFTRinh-172 dose–response relationships), maximal I influx values in the absence of inhibitor and the number of experiments (n). For all fits, the correlation coefficient R2 was higher or equal to 0.97. **Indicate that the R347A Ki was significantly higher (P<0.01) compared with wild-type CFTR. For Arg347 mutants other than R347A, sensitivity to CFTRinh-172 was so low that the Ki could not be determined precisely (see Experimental section). I influx values for all CFTR constructs were significantly higher than mock-transfected cells (P<0.01).

CFTR form CFTRinh-172 Ki (μM) Hill coefficient I influx (mM/s) n 
Wild-type 1.32±0.25 1.03±0.07 0.1336±0.0107 10 
S341A 0.57±0.17 1.21±0.37 0.0297±0.0064 
T338A 3.20±0.86 1.13±0.20 0.1260±0.0225 
R347A 44.98±4.71** 0.91±0.04 0.1288±0.0154 
R334A 2.39±0.74 0.93±017 0.0313±0.062 
A349S 1.23±0.41 1.11±0.25 0.1500±0.011 
R347D >50 Not determined 0.1160±0.0136 
R347D/D924R >50 Not determined 0.1008±0.0504 
R347C >50 Not determined 0.1437±0.0123 
Mock   0.003±0.001 10 
CFTR form CFTRinh-172 Ki (μM) Hill coefficient I influx (mM/s) n 
Wild-type 1.32±0.25 1.03±0.07 0.1336±0.0107 10 
S341A 0.57±0.17 1.21±0.37 0.0297±0.0064 
T338A 3.20±0.86 1.13±0.20 0.1260±0.0225 
R347A 44.98±4.71** 0.91±0.04 0.1288±0.0154 
R334A 2.39±0.74 0.93±017 0.0313±0.062 
A349S 1.23±0.41 1.11±0.25 0.1500±0.011 
R347D >50 Not determined 0.1160±0.0136 
R347D/D924R >50 Not determined 0.1008±0.0504 
R347C >50 Not determined 0.1437±0.0123 
Mock   0.003±0.001 10 

We tested the effect of replacement of Arg347 with other amino acids. As found for R347A, the mutants R347C and R347D also showed a normal rate of anion transport but altered sensitivity to CFTRinh-172 (Figures 2A and 2B). Remarkably, an arginine-to-aspartate residue change produced a near total loss of CFTRinh-172 activity (Figure 2B).

Mutagenesis of Arg347 and Asp924 residues

Figure 2
Mutagenesis of Arg347 and Asp924 residues

(A) Rate of I transport measured in COS-7 cells transfected with the indicated constructs. In contrast with all other constructs, D924A and D924R did not show a significant I transport compared with mock-transfected cells. (B) CFTRinh-172 dose–response relationships for wild-type, R347D and R347D/D924R CFTR. Values are the means±S.E.M. of four to ten experiments. (C) Analysis of CFTR protein maturation by immunoprecipitation/radioactive labelling (top panel) or by immunoprecipitation/Western blot analysis (middle panel). The position of mature fully glycosylated (band C) and immature core glycosylated (band B) CFTR is indicated. The bottom panel shows ratios of band C/(band C+band B) intensities (means±S.E.M.) measured for the two types of experiments. Open bars indicate immunoprecipitation/radioactive labelling and solid bars indicate immunoprecipitation/Western blot analysis (n=3 for both conditions). ** indicate a significant difference relative to wild-type CFTR (P<0.01). WT, wild-type.

Figure 2
Mutagenesis of Arg347 and Asp924 residues

(A) Rate of I transport measured in COS-7 cells transfected with the indicated constructs. In contrast with all other constructs, D924A and D924R did not show a significant I transport compared with mock-transfected cells. (B) CFTRinh-172 dose–response relationships for wild-type, R347D and R347D/D924R CFTR. Values are the means±S.E.M. of four to ten experiments. (C) Analysis of CFTR protein maturation by immunoprecipitation/radioactive labelling (top panel) or by immunoprecipitation/Western blot analysis (middle panel). The position of mature fully glycosylated (band C) and immature core glycosylated (band B) CFTR is indicated. The bottom panel shows ratios of band C/(band C+band B) intensities (means±S.E.M.) measured for the two types of experiments. Open bars indicate immunoprecipitation/radioactive labelling and solid bars indicate immunoprecipitation/Western blot analysis (n=3 for both conditions). ** indicate a significant difference relative to wild-type CFTR (P<0.01). WT, wild-type.

Because Arg347 is believed to be involved in the formation of a salt bridge with Asp924 [25], we also tested the CFTR mutants D924A and D924R. Unfortunately, these mutants showed no detectable CFTR activity (Figure 2A). To further investigate the importance of the salt bridge, we generated a double mutant, R347D/D924R, in which the positions of the charged amino acids are inverted. Interestingly, in contrast with D924R, the double mutant was able to transport anions but showed a low sensitivity to CFTRinh-172, with an estimated Ki greater than 50 μM (Figures 2A and 2B), similar to that of the single R347D mutant.

We evaluated the maturation of the D924R mutant by immunoprecipitation followed by Western blot analysis or in vitro phosphorylation (Figure 2C). Wild-type protein and the R347D mutant showed a normal pattern of electrophoretic mobility with a prevalent abundance of the band C which represents the fully glycosylated mature form of CFTR. As expected, the trafficking mutant F508del was associated with the presence of only band B, the core-glycosylated immature form of CFTR (results not shown). In contrast with R347D, the D924R mutation produced a partial defect in maturation, with increased band B intensity compared with band C. Interestingly, this defect seemed to be corrected in the double mutant R347D/D924R (Figure 2C).

We next determined the sensitivity of the Arg347 mutants to CFTRinh-172 by direct measurement of electrogenic Cl transport (Figures 3A–3E). FRT cells were stably transfected with wild-type, R334A, R347A and R347D CFTR, and transepithelial Cl currents were measured. The R347A and R347D mutants showed significantly reduced sensitivity to CFTRinh-172 compared with the wild-type CFTR, thus confirming the results obtained with the fluorescence assay (Figures 3A, 3B and 3D). The calculated Ki for wild-type CFTR, R347A and R347D was 0.85±0.13 μM (n=8), 17.35±3.90 μM (n=13) and 53.10±4.74 μM (n=6) respectively (Figure 3E). The values for the two mutants were significantly higher than that of wild-type CFTR (P<0.01). For comparison, the sensitivity of the R334A mutant was not significantly altered (Ki=0.50±0.18 μM, n=6; Figures 3C and 3E), in agreement with the fluorescence assay results. Interestingly, the R347D mutant, although insensitive to CFTRinh-172, was fully inhibited by the open-channel blocker GlyH-101 (Figure 3D). Dose-responses carried out on wild-type CFTR and R347D cells showed identical sensitivity to GlyH-101 with a Ki of 5.02±0.90 μM (n=6) and 4.99±0.66 μM (n=6) respectively (Figures 4A–4C).

CFTR Cl current inhibition by CFTRinh-172

Figure 3
CFTR Cl current inhibition by CFTRinh-172

(AD) Representative traces showing recordings of transepithelial Cl currents measured in FRT cells with stable expression of wild-type (WT), R347A, R334A and R347D-CFTR. Cells were first stimulated with 20 μM forskolin to activate CFTR and then tested with increasing concentrations of CFTRinh-172. In (D), 25 μM GlyH-101 was added at the end of experiment to block the remaining CFTR activity. (E) The CFTRinh-172 dose–response relationships obtained from experiments (AD). Values are means±S.E.M. of six to thirteen experiments.

Figure 3
CFTR Cl current inhibition by CFTRinh-172

(AD) Representative traces showing recordings of transepithelial Cl currents measured in FRT cells with stable expression of wild-type (WT), R347A, R334A and R347D-CFTR. Cells were first stimulated with 20 μM forskolin to activate CFTR and then tested with increasing concentrations of CFTRinh-172. In (D), 25 μM GlyH-101 was added at the end of experiment to block the remaining CFTR activity. (E) The CFTRinh-172 dose–response relationships obtained from experiments (AD). Values are means±S.E.M. of six to thirteen experiments.

CFTR Cl current inhibition by GlyH-101

Figure 4
CFTR Cl current inhibition by GlyH-101

(A and B) Representative traces showing transepithelial Cl currents measured in FRT cells with stable expression of wild-type (WT) and R347D-CFTR. Cells were first stimulated with 20 μM forskolin to activate CFTR and then were tested with increasing concentrations of GlyH-101. (C) GlyH-101 dose-responses. Values are the means±S.E.M. of six experiments.

Figure 4
CFTR Cl current inhibition by GlyH-101

(A and B) Representative traces showing transepithelial Cl currents measured in FRT cells with stable expression of wild-type (WT) and R347D-CFTR. Cells were first stimulated with 20 μM forskolin to activate CFTR and then were tested with increasing concentrations of GlyH-101. (C) GlyH-101 dose-responses. Values are the means±S.E.M. of six experiments.

We also tested the sensitivity of wild-type and R347D to a zwitterionic, net neutral analogue of CFTRinh-172, thiazo N-O, in which the carboxyphenyl group is replaced by an oxido-4-pyridinyl group (Figure 5A). Although less potent than CFTRinh-172, the thiazo N-O compound caused a dose-dependent decrease of wild-type CFTR currents (Ki=31.42±7.30 μM, n=4), but very weak inhibition of the R347D mutant (Figures 5B and 5C).

CFTR Cl current inhibition by thiazo N-O

Figure 5
CFTR Cl current inhibition by thiazo N-O

(A) Chemical structures of thiazo N-O and CFTRinh-172. (B) Representative recordings showing effect of increasing concentrations of thiazo N-O on wild-type and R347D activity. (C) Thiazo N-O dose-response. Values are the means±S.E.M. of four experiments. WT, wild-type.

Figure 5
CFTR Cl current inhibition by thiazo N-O

(A) Chemical structures of thiazo N-O and CFTRinh-172. (B) Representative recordings showing effect of increasing concentrations of thiazo N-O on wild-type and R347D activity. (C) Thiazo N-O dose-response. Values are the means±S.E.M. of four experiments. WT, wild-type.

Whole-cell patch-clamp experiments were carried out to further investigate the mechanism of action of CFTRinh-172. Experiments on cells expressing wild-type CFTR (n=6) showed large cAMP-activated currents that were inhibited by more than 90% with CFTRinh-172 (5 μM; Figures 6A and 6B). The dependence of CFTR currents versus the applied membrane potential was linear and the extent of inhibition by CFTRinh-172 was not affected by membrane potential, as previously described [7]. In cells expressing the R347A mutant (n=6), cAMP stimulation elicited currents with a moderate outward rectification of the current–voltage relationship (Figures 6C and 6D). Despite the use of a CFTRinh-172 concentration an order of magnitude higher than the half-effective concentration for wild-type CFTR, the Cl currents of the R347A mutant were only partially inhibited (Figures 6C and 6D). Summary of results in Figure 6(E) shows that CFTRinh-172 at 10 μM produced an approx. 50% reduction of R347A currents whereas the inhibition of wild-type currents at 5 μM was nearly total. The graph also shows that CFTRinh-172 inhibition was not affected by the applied membrane potential.

Patch-clamp analysis of CFTR inhibition by CFTRinh-172

Figure 6
Patch-clamp analysis of CFTR inhibition by CFTRinh-172

(A and C) Superimposed membrane currents recorded from cells expressing wild-type and the R347A mutant at membrane potentials between −100 and +100 mV. Currents were recorded in the absence and in the presence of CFTRinh-172 (5 μM for wild-type and 10 μM for R347A). (B and D) Plots of current-to-voltage relationships for the experiments shown in (A) and (C) respectively. (E) Summary of block caused by CFTRinh-172 at all membrane potentials on wild-type (at 5 μM) and R347A (at 10 μM). Values are means±S.E.M. (n=6 for both wild-type and R347A).

Figure 6
Patch-clamp analysis of CFTR inhibition by CFTRinh-172

(A and C) Superimposed membrane currents recorded from cells expressing wild-type and the R347A mutant at membrane potentials between −100 and +100 mV. Currents were recorded in the absence and in the presence of CFTRinh-172 (5 μM for wild-type and 10 μM for R347A). (B and D) Plots of current-to-voltage relationships for the experiments shown in (A) and (C) respectively. (E) Summary of block caused by CFTRinh-172 at all membrane potentials on wild-type (at 5 μM) and R347A (at 10 μM). Values are means±S.E.M. (n=6 for both wild-type and R347A).

DISCUSSION

The thiazolidinone CFTRinh-172 is now widely used as a high-affinity and selective inhibitor of the CFTR Cl channel [2632]. However, its mechanism of action is unknown. It has been uncertain whether CFTRinh-172 is a direct inhibitor of the CFTR protein or a modulator of an interacting regulatory protein. In the present study, we have analysed the effect on CFTRinh-172 activity of mutations in the sixth TMD of the CFTR protein. We introduced mutations in residues Arg334, Thr338 and Ser341, which are believed to be located in the narrowest part of the CFTR pore, thus determining channel conductance and ion selectivity. We also introduced mutations in Arg347 and Arg352, which are reported to influence CFTR channel properties but are located more distant from the channel pore, probably in a cytosolic vestibule. We found, using three different techniques, that mutagenesis of Arg347 strongly reduced CFTRinh-172 potency. Although, the absolute values of CFTRinh-172 Ki were dependent on the type of experiment (with fluorescence assays giving the highest values), there was a consistent 20–30-fold loss of potency in the Arg347 mutants compared with the wild-type protein. In contrast, mutagenesis of Arg334, Thr338 and Ser341 did not significantly alter CFTRinh-172 affinity. This finding is consistent with previous patch-clamp results showing that CFTRinh-172 inhibition does not interact with the inner CFTR pore since its effect is not affected by transmembrane voltage. Furthermore, CFTRinh-172 does not change the mean open time of CFTR channel, a behaviour that would be expected for a blocker interacting with the pore.

Arg347 lies on the cytosolic part of the sixth TMD although it is not clear whether it is exposed to the solvent [17]. Arg347 was initially considered as an amino acid residue directly involved in the interaction of CFTR with permeating anions [33,34]. Indeed, mutagenesis of this residue caused a change in the anomalous mole fraction behaviour of CFTR, i.e. in the intrinsic properties of channel conductance and selectivity. Nevertheless, a subsequent study has instead pointed out that Arg347, together with Asp924, forms a salt bridge which is important in the determination of the CFTR channel structure [25]. We have investigated the importance of the R347–D924 salt bridge in the inhibitory activity of CFTRinh-172. We removed the negative charge at position 924 by replacing an aspartate residue with an alanine or arginine residue. Unfortunately, these mutations abolished anion transport thus impeding determination of CFTRinh-172 activity. Determination of CFTR protein electrophoretic mobility revealed that mutagenesis of Asp924 causes a partial impairment in maturation. Indeed, D924R causes a reduced amount of fully glycosylated protein (band C) that may indicate that the mutant protein is less stable. This defect may be responsible for the absence of ion transport function through a reduction of protein in the plasma membrane and/or defective gating and conductance. Interestingly, when we generated the double mutant R347D/D924R, in which the positions of positive and negative charges are inverted but the salt bridge is maintained [25], we found that anion transport and CFTR protein maturation were rescued compared with single Asp924 mutants. On the other hand, we found that the double mutant R347D/D924R did not behave as the wild-type CFTR in terms of CFTRinh-172 sensitivity but was more similar to single Arg347 mutants. Therefore we conclude that it is not the presence of the salt bridge but rather that of an arginine residue at position 347 that is required for CFTRinh-172 inhibitory activity.

To further investigate the role of Arg347 in the CFTRinh-172 mechanism of action, we tested an analogue, thiazo N-O, having the carboxyphenyl moiety replaced by a zwitterionic neutral moiety. We hypothesized that the negatively charged carboxyl group in CFTRinh-172 interacts with the positive charge of Arg347, such that the effects of Arg347 mutations could be explained by loss of electrostatic attraction (R347A) or generation of electrostatic repulsion (R347D). Accordingly, the activity of the neutral thiazo N-O compound had to be less affected by the R347D mutation. However, the potency of this compound on the R347D mutant was reduced, as found for CFTRinh-172, providing evidence against a direct electrostatic interaction between the CFTRinh-172 carboxyl group and Arg347.

The finding that a single amino acid change in the CFTR protein sequence alters CFTRinh-172 potency has two interesting implications. First, it represents the first indication that CFTRinh-172 is a direct inhibitor of CFTR, thus supporting its value as a tool of research to selectively inhibit CFTR function. Secondly, it indicates that CFTRinh-172 may be a useful probe in investigating the CFTR structure–function relationship. The results of the present study may imply that Arg347 is directly involved in binding to CFTRinh-172. Alternatively, giving the reported importance of Arg347 in CFTR channel structure, and our results with the thiazo N-O analogue, it is possible that the binding site lies elsewhere in the CFTR sequence and that replacement of Arg347 has an allosteric effect. For example, the R347D mutation has been shown to alter ATPase activity in NBDs [35], a finding that points to strong conformational coupling between TMDs and NBDs. However, it is important to point out that in the present study the R347D mutant, although being poorly inhibited by CFTRinh-172, showed an unaltered sensitivity to GlyH-101, an open-channel blocker acting from the extracellular side [8]. This finding indicates that Arg347 mutations do not cause a gross disruption of the CFTR pore structure.

In theory, a compound causing CFTR Cl current inhibition may act in two ways: by occluding the channel pore or by impairing the channel gating [1]. Occlusion of the pore, as is the case of GlyH-101 [8], is characterized by shortening of the mean open time and voltage-dependence of the block (if the blocking molecule is electrically charged). In contrast, CFTRinh-172 acts in a voltage-independent manner and by increasing the mean closed time [9]. This behaviour is different from that of classical open-channel blockers and more suggestive of a mechanism of action involving gating inhibition, probably by stabilizing the closed state. CFTR gating is controlled by events occurring at the NBDs, i.e. cycles of ATP binding/hydrolysis that are coupled with NBD dimerization. Such processes in turn induce changes in the conformation of TMDs leading to opening of the CFTR pore. In theory, CFTRinh-172 may act at one of the steps coupling NBDs to TMDs. The results of the present study, by showing a change in potency caused by Arg347 mutations, indicate that CFTRinh-172 interacts directly with CFTR. Future studies are needed to assess whether CFTRinh-172 binds to TMDs or whether the role of Arg347 is indirect, and involves a binding site located more distant in the CFTR protein, possibly in the NBDs.

We thank Telethon-Italy (GGP05103), CIPE (Comitato Interministeriale per la Programmazione Economica)/Regione Liguria 2007 (Drug Discovery and Delivery), the NIH (National Institutes of Health; P30 DK072517) and Cystic Fibrosis Foundation Therapeutics for support. This work was also supported by public grants from Institut National de la Santé et de la Recherche Médicale (INSERM) and Chancellerie des Universités de Paris, and the French Association ‘Vaincre la Mucoviscidose’.

Abbreviations

     
  • CFTR

    cystic fibrosis transmembrane conductance regulator

  •  
  • DMEM

    Dulbecco's modifed Eagle's medium

  •  
  • FCS

    foetal calf serum

  •  
  • FRT

    Fisher rat thyroid

  •  
  • HEK

    human embryonic kidney

  •  
  • mAb

    monoclonal antibody

  •  
  • NBD

    nucleotide-binding domain

  •  
  • PKA

    protein kinase A

  •  
  • thiaxo N-O

    5-(1-oxido-4-pyridinyl)methylene-2-thioxo-3-[3-(trifluoromethyl)phenyl]-4-thiazolidinone

  •  
  • TMD

    transmembrane domain

  •  
  • YFP

    yellow fluorescent protein

References

References
1
Hwang
T. C.
Sheppard
D. N.
Molecular pharmacology of the CFTR Cl− channel
Trends Pharmacol. Sci.
1999
, vol. 
20
 (pg. 
448
-
453
)
2
Verkman
A. S.
Lukacs
G. L.
Galietta
L. J. V.
CFTR chloride channel drug discovery: inhibitors as antidiarrheals and activators for therapy of cystic fibrosis
Curr. Pharm. Des.
2006
, vol. 
12
 (pg. 
2235
-
2247
)
3
Sheppard
D. N.
Welsh
M. J.
Structure and function of the CFTR chloride channel
Physiol. Rev.
1999
, vol. 
79
 (pg. 
S23
-
S45
)
4
Akabas
M. H.
Cystic fibrosis transmembrane conductance regulator. Structure and function of an epithelial chloride channel
J. Biol. Chem.
2000
, vol. 
275
 (pg. 
3729
-
3732
)
5
Vergani
P.
Lockless
S. W.
Nairn
A. C.
Gadsby
D. C.
CFTR channel opening by ATP-driven tight dimerization of its nucleotide-binding domains
Nature
2005
, vol. 
433
 (pg. 
876
-
880
)
6
Schultz
B. D.
Singh
A. K.
Devor
D. C.
Bridges
R. J.
Pharmacology of CFTR chloride channel activity
Physiol. Rev.
1999
, vol. 
79
 (pg. 
S109
-
S144
)
7
Ma
T.
Thiagarajah
J. R.
Yang
H.
Sonawane
N. D.
Folli
C.
Galietta
L. J. V.
Verkman
A. S.
Thiazolidinone CFTR inhibitor identified by high-throughput screening blocks cholera toxin-induced intestinal fluid secretion
J. Clin. Invest.
2002
, vol. 
110
 (pg. 
1651
-
1658
)
8
Muanprasat
C.
Sonawane
N. D.
Salinas
D.
Taddei
A.
Galietta
L. J. V.
Verkman
A. S.
Discovery of glycine hydrazide pore-occluding CFTR inhibitors: mechanism, structure-activity analysis, and in vivo efficacy
J. Gen. Physiol.
2004
, vol. 
124
 (pg. 
125
-
137
)
9
Taddei
A.
Folli
C.
Zegarra-Moran
O.
Fanen
P.
Verkman
A. S.
Galietta
L. J. V.
Altered channel gating mechanism for CFTR inhibition by a high-affinity thiazolidinone blocker
FEBS Lett.
2004
, vol. 
558
 (pg. 
52
-
56
)
10
Routaboul
C.
Norez
C.
Melin
P.
Molina
M. C.
Boucherle
B.
Bossard
F.
Noel
S.
Robert
R.
Gauthier
C.
Becq
F.
Decout
J. L.
Discovery of α-aminoazaheterocycle-methylglyoxal adducts as a new class of high-affinity inhibitors of cystic fibrosis transmembrane conductance regulator chloride channels
J. Pharmacol. Exp. Ther.
2007
, vol. 
322
 (pg. 
1023
-
1035
)
11
Fuller
M. D.
Thompson
C. H.
Zhang
Z. R.
Freeman
C. S.
Schay
E.
Szakacs
G.
Bakos
E.
Sarkadi
B.
McMaster
D.
French
R. J.
, et al. 
State-dependent inhibition of CFTR chloride channels by a novel peptide toxin
J. Biol. Chem.
2007
, vol. 
282
 (pg. 
37545
-
37555
)
12
Pedemonte
N.
Lukacs
G. L.
Du
K.
Caci
E.
Zegarra-Moran
O.
Galietta
L. J. V.
Verkman
A. S.
Small-molecule correctors of defective deltaF508-CFTR cellular processing identified by high-throughput screening
J. Clin. Invest.
2005
, vol. 
115
 (pg. 
2564
-
2571
)
13
Zegarra-Moran
O.
Romio
L.
Folli
C.
Caci
E.
Becq
F.
Vierfond
J. M.
Mettey
Y.
Cabrini
G.
Fanen
P.
Galietta
L. J. V.
Correction of G551D-CFTR transport defect in epithelial monolayers by genistein but not by CPX or MPB-07
Br. J. Pharmacol.
2002
, vol. 
137
 (pg. 
504
-
512
)
14
Pedemonte
N.
Sonawane
N. D.
Taddei
A.
Hu
J.
Zegarra-Moran
O.
Suen
Y. F.
Robins
L. I.
Dicus
C. W.
Willenbring
D.
Nantz
M. H.
, et al. 
Phenylglycine and sulfonamide correctors of defective delta F508 and G551D cystic fibrosis transmembrane conductance regulator chloride-channel gating
Mol. Pharmacol.
2005
, vol. 
67
 (pg. 
1797
-
1807
)
15
Pedemonte
N.
Boido
D.
Moran
O.
Giampieri
M.
Mazzei
M.
Ravazzolo
R.
Galietta
L. J. V.
Structure-activity relationship of 1,4-dihydropyridines as potentiators of the cystic fibrosis transmembrane conductance regulator chloride channel
Mol. Pharmacol.
2007
, vol. 
72
 (pg. 
197
-
207
)
16
Cheung
M.
Akabas
M. H.
Locating the anion-selectivity filter of the cystic fibrosis transmembrane conductance regulator (CFTR) chloride channel
J. Gen. Physiol.
1997
, vol. 
109
 (pg. 
289
-
299
)
17
Dawson
D. C.
Smith
S. S.
Mansoura
M. K.
CFTR: mechanism of anion conduction
Physiol. Rev.
1999
, vol. 
79
 (pg. 
S47
-
S75
)
18
McCarty
N. A.
Zhang
Z. R.
Identification of a region of strong discrimination in the pore of CFTR
Am. J. Physiol. Lung Cell. Mol. Physiol.
2001
, vol. 
281
 (pg. 
L852
-
L867
)
19
Ge
N.
Muise
C. N.
Gong
X.
Linsdell
P.
Direct comparison of the functional roles played by different transmembrane regions in the cystic fibrosis transmembrane conductance regulator chloride channel pore
J. Biol. Chem.
2004
, vol. 
279
 (pg. 
55283
-
55289
)
20
Galietta
L. J. V.
Haggie
P. M.
Verkman
A. S.
Green fluorescent protein-based halide indicators with improved chloride and iodide affinities
FEBS Lett.
2001
, vol. 
499
 (pg. 
220
-
224
)
21
Galietta
L. J. V.
Springsteel
M. F.
Eda
M.
Niedzinski
E. J.
By
K.
Haddadin
M. J.
Kurth
M. J.
Nantz
M. H.
Verkman
A. S.
Novel CFTR chloride channel activators identified by screening of combinatorial libraries based on flavone and benzoquinolizinium lead compounds
J. Biol. Chem.
2001
, vol. 
276
 (pg. 
19723
-
19728
)
22
Sonawane
N. D.
Muanprasat
C.
Nagatani
R.
Jr
Song
Y.
Verkman
A. S.
In vivo pharmacology and antidiarrheal efficacy of a thiazolidinone CFTR inhibitor in rodents
J. Pharm. Sci.
2005
, vol. 
94
 (pg. 
134
-
143
)
23
Aubin
C. N.
Linsdell
P.
Positive charges at the intracellular mouth of the pore regulate anion conduction in the CFTR chloride channel
J. Gen. Physiol.
2006
, vol. 
128
 (pg. 
535
-
545
)
24
Salinas
D. B.
Pedemonte
N.
Muanprasat
C.
Finkbeiner
W. F.
Nielson
D. W.
Verkman
A. S.
CFTR involvement in nasal potential differences in mice and pigs studied using a thiazolidinone CFTR inhibitor
Am. J. Physiol. Lung Cell. Mol. Physiol.
2004
, vol. 
287
 (pg. 
L936
-
L943
)
25
Cotton
J. F.
Welsh
M. J.
Cystic fibrosis-associated mutations at arginine 347 alter the pore architecture of CFTR. Evidence for disruption of a salt bridge
J. Biol. Chem.
1999
, vol. 
274
 (pg. 
5429
-
5435
)
26
Li
H.
Findlay
I. A.
Sheppard
D. N.
The relationship between cell proliferation, Cl− secretion, and renal cyst growth: a study using CFTR inhibitors
Kidney Int.
2004
, vol. 
66
 (pg. 
1926
-
1938
)
27
Bruscia
E. M.
Price
J. E.
Cheng
E. C.
Weiner
S.
Caputo
C.
Ferreira
E. C.
Egan
M. E.
Krause
D. S.
Assessment of cystic fibrosis transmembrane conductance regulator (CFTR) activity in CFTR-null mice after bone marrow transplantation
Proc. Natl. Acad. Sci. U.S.A.
2006
, vol. 
103
 (pg. 
2965
-
2970
)
28
Fang
X.
Song
Y.
Hirsch
J.
Galietta
L. J. V.
Pedemonte
N.
Zemans
R. L.
Dolganov
G.
Verkman
A. S.
Matthay
M. A.
Contribution of CFTR to apical-basolateral fluid transport in cultured human alveolar epithelial type II cells
Am. J. Physiol. Lung Cell. Mol. Physiol.
2006
, vol. 
290
 (pg. 
L242
-
L249
)
29
Wang
Y.
Soyombo
A. A.
Shcheynikov
N.
Zeng
W.
Dorwart
M.
Marino
C. R.
Thomas
P. J.
Muallem
S.
Slc26a6 regulates CFTR activity in vivo to determine pancreatic duct HCO3− secretion: relevance to cystic fibrosis
EMBO J.
2006
, vol. 
25
 (pg. 
5049
-
5057
)
30
Liu
X.
Luo
M.
Zhang
L.
Ding
W.
Yan
Z.
Engelhardt
J. F.
Bioelectric properties of chloride channels in human, pig, ferret, and mouse airway epithelia
Am. J. Respir. Cell. Mol. Biol.
2007
, vol. 
36
 (pg. 
313
-
323
)
31
Perez
A.
Issler
A. C.
Cotton
C. U.
Kelley
T. J.
Verkman
A. S.
Davis
P. B.
CFTR inhibition mimics the cystic fibrosis inflammatory profile
Am. J. Physiol. Lung Cell. Mol. Physiol.
2007
, vol. 
292
 (pg. 
L383
-
L395
)
32
Wolde
M.
Fellows
A.
Cheng
J.
Kivenson
A.
Coutermarsh
B.
Talebian
L.
Karlson
K.
Piserchio
A.
Mierke
D. F.
Stanton
B. A.
, et al. 
Targeting CAL as a negative regulator of ΔF508-CFTR cell-surface expression: an RNA interference and structure-based mutagenetic approach
J. Biol. Chem.
2007
, vol. 
282
 (pg. 
8099
-
8109
)
33
Tabcharani
J. A.
Rommens
J. M.
Hou
Y. X.
Chang
X. B.
Tsui
L. C.
Riordan
J. R.
Hanrahan
J. W.
Multi-ion pore behaviour in the CFTR chloride channel
Nature
1993
, vol. 
366
 (pg. 
79
-
82
)
34
Tabcharani
J. A.
Linsdell
P.
Hanrahan
J. W.
Halide permeation in wild-type and mutant cystic fibrosis transmembrane conductance regulator chloride channels
J. Gen. Physiol.
1997
, vol. 
110
 (pg. 
341
-
354
)
35
Kogan
I.
Ramjeesingh
M.
Huan
L. J.
Wang
Y.
Bear
C. E.
Perturbation of the pore of the cystic fibrosis transmembrane conductance regulator (CFTR) inhibits its ATPase activity
J. Biol. Chem.
2001
, vol. 
276
 (pg. 
11575
-
11581
)