The cystic fibrosis mutation G551D alters the non-Michaelis-Menten behavior of the cystic fibrosis transmembrane conductance regulator (CFTR) channel and abolishes the inhibitory Genistein binding site.

Loss of cystic fibrosis transmembrane conductance regulator (CFTR) channel activity explains most of the manifestations of the cystic fibrosis (CF) disease. To understand the consequences of CF mutations on CFTR channel activity, we compared the pharmacological properties of wild-type (wt) and G551D-CFTR. Dose-dependent relationships of wt-CFTR activated by genistein follows a non-Michaelis-Menten behavior consistent with the presence of two binding sites. With phosphorylated CFTR, a high affinity site for genistein is the activator (K(s) approximately 3 microm), whereas a second site of low affinity (K(i) approximately 75 microm) is the inhibitor. With non-phosphorylated CFTR, K(s) was increased (K(s) approximately 12 microm), but K(i) was not affected (K(i) approximately 70 microm). In G551D-CFTR cells, channel activity was recovered by co-application of forskolin and genistein in a dose-dependent manner. A further stimulation of G551D-CFTR channel activity was measured at concentrations from 30 microm to 1 mm. The dose response is described by a classical Michaelis-Menten kinetics with only a single apparent site (K(m) approximately 11 microm). Our results suggest glycine 551 in NBD1 as an important location within the low affinity inhibitory site for genistein and offers new evidence for pharmacological alteration caused by an NBD1 mutation of CFTR. This study also reveals how a mutation of an ion channel converts a non-Michaelis-Menten behavior (two binding sites) into a classical Michaelis-Menten model (one binding site).

The protein product of the cystic fibrosis (CF) 1 gene is a cAMP-regulated chloride channel, the cystic fibrosis transmembrane conductance regulator (CFTR), which mediates Cl Ϫ transepithelial transport in epithelia (1)(2)(3). CF, one of the most common lethal autosomal recessive genetic diseases, is caused by mutations of the CF gene and generates defective Cl Ϫ transport across the affected epithelium (1,3). CFTR mutations (www.genet.sickkids.on.ca/cftr) can be assigned to one of five classes of mutations (4), leading to a protein in which chloride channel function is altered (classes III and IV) or lacking (classes I, II, and V) at the apical membrane. The glycine-to-aspartic acid missense mutation at codon 551 (G551D) is a class III mutation located within the nucleotide binding domain 1 (NBD1) (5). Glycine 551 lies within a sequence in NBD1, DX(G/ A)GQ, which shows remarkable conservation in ATP binding cassette transporters. G551D is one of the five most frequent CF mutations with a frequency of 2-5% depending of the population of origin but is associated with a severe CF phenotype (5). Class III mutations disrupt activation and regulation of CFTR at the plasma membrane. The G551D mutated protein is fully glycosylated, correctly located at the apical membrane (i.e. normal biosynthesis, trafficking, and processing), and normally phosphorylated at the R domain by cAMP-dependent protein kinase (6). However, G551D proteins have a decreased nucleotide binding (7) and a reduced ATPase activity at NBD1 (8,9).
Understanding the molecular consequences of a mutation on CFTR function and activity and its impact on the severity of the disease must be associated with the development of agents able to modulate the activity of CFTR. Our knowledge of the pharmacological modulation of non-mutated wild-type CFTR (wt-CFTR) chloride channel activity is now increasing, and a number of molecules have been described as activators (reviewed in Ref. 10). However, the pharmacological consequences of disease-causing mutations of CFTR have not been systematically characterized and compared with non-mutated CFTR. In this work, we compared the effect of genistein on both activation and inhibition of wt-CFTR and G551D-CFTR chloride channels. Analyzing dose-response relationships pointed to a novel type of alteration due to the disease-causing mutation G551D. A non-Michaelis-Menten behavior describes the effect of genistein on wt-CFTR and shows the presence of one inhibitory site and one activatory site. On the contrary, mutation of glycine 551 in NBD1 alters this kinetic behavior, which is transformed into a classical Michaelis-Menten mechanism indicating a single apparent site for genistein on CFTR.
Iodide Efflux Experiments-CFTR chloride channel activity was assayed by measuring the rate of iodide ( 125 I) efflux versus time from cells as described previously (11,12). All experiments were performed at 37°C. Cells were cultured in 24-well plates to perform parallel experiments and comparison analysis. The efflux buffer contained 137 mM NaCl, 5.36 mM KCl, 0.8 mM MgCl 2 , 5.5 mM glucose, and 10 mM HEPES, pH 7.4. Cells are loaded in efflux buffer containing 1 M KI (1 Ci of Na 125 I/ml, PerkinElmer Life Sciences) for 30 min at 37°C. The loss of intracellular 125 I was determined by removing the medium with efflux buffer every 1 min for up to 11 min. The first four aliquots were used to * This work was supported by Vaincre La Mucoviscidose. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ To whom correspondence should be addressed. establish a stable baseline in efflux buffer alone. A medium containing the appropriate drug was used for the remaining aliquots. Residual radioactivity was extracted with 0.1 N NaOH and determined using a ␥ counter (Cobra II, Packard-Bell). The fraction of initial intracellular 125 I lost during each time point was determined, and time-dependent rates of 125 I efflux were calculated from ln ( 125 I t1 / 125 I t2 )/(t 1 Ϫ t 2 ) where 125 I t is the intracellular 125 I at time t, and t 1 and t 2 are successive time points (13). Curves were constructed by plotting the rate of 125 I versus time. The relative rate R corresponds to R peak /R basal . All comparisons were based on maximal values for the time-dependent rates excluding the points used to establish the baseline (12).
Patch Clamp Experiments-Whole cell recordings were performed as described elsewhere (12). Briefly, currents were recorded with a List EPC-7 patch clamp amplifier. I-V relationships were built by clamping the membrane potential to Ϫ40 mV and by pulses from Ϫ100 mV to ϩ100 mV by 20-mV increments. Media generating a chloride gradient of an external concentration of 151 mM and internal concentration of 28 mM were used. The pipette solution contained 113 mM L-aspartic acid, 113 mM CsOH, 27 mM CsCl, 1 mM NaCl, 1 mM EGTA, 10 mM TES, 285 mosM (pH 7.2). MgATP (3 mM) was added just before patch clamp experiments. The external solution consists of 145 mM NaCl, 4 mM CsCl, 1 mM CaCl 2 , 5 mM glucose, 10 mM TES, 340 mosM (pH 7.4). Results were analyzed with the pCLAMP6 package software (pCLAMP, Axon Instruments).
Statistics-Results are expressed as means Ϯ S.E. of n observations. To compare sets of data, we used either an analysis of variance or Student's t test. Differences were considered statistically significant when p Ͻ 0.05. All statistical tests were performed using GraphPad Prism version 3.0 for Windows (GraphPad Software, San Diego, California, CA).

Non-Michaelis-Menten Behavior of wt-CFTR in Presence of
Genistein-We used CHO cells as a model to study the pharmacology of wt-CFTR chloride channel activity. Because phosphorylation is an important mode of regulation of CFTR (reviewed in Ref. 14) and because cAMP-dependent phosphorylation is a prerequisite for genistein to activate CFTR (15,16), we examined the effect of genistein on partially and non-phosphorylated CFTR channels. In the presence of increasing concentrations of forskolin (1 nM-10 M), the wt-CFTR channel activity determined from iodide efflux experiments and whole cell patch clamp was stimulated with an EC 50 Ϸ 1 M (not shown). In this study, partial phosphorylation of the CFTR channel was achieved in the presence of 500 nM forskolin (FSK). Concentration-dependent activation relationships were then generated using genistein (GST, chemical structure shown in Because genistein activates CFTR at low concentrations but inhibits CFTR at high concentrations, the half-maximal concentrations for activation (K s ) and inhibition (K i ) cannot be determined by a classical analysis from the fit of the data. This phenomenon is reminiscent of an enzymatic process in which an inhibition by excess of substrate occurs, a process known as non-Michaelis-Menten behavior. In such a reaction, an enzyme and its substrate give an intermediate complex and the product. When a second molecule of substrate interacts with this complex, the new complex formed is inactive or partially inactive. From the study of Randak et al. (17), we learned that genistein binds to the second nucleotide binding domain (NBD2) of CFTR. We thus hypothesized that the inhibition of CFTR with high concentrations of genistein could be due to the interaction of a second molecule of genistein with the complex formed by CFTR and genistein, as illustrated below in Reaction 1, In such a non-Michaelis-Menten reaction, for low concentrations of GST, the K s of the reaction is given by the Lineweaver-Burk representation (as shown in Equation 1), The relative rate of efflux, designated as R, is considered equivalent to the velocity V of the reaction. As shown in Fig. 2, the intersection on the x axis of the dotted line (which represents the fit of first-order regression to the data) corresponds to Ϫ1/K s in the presence of forskolin ( Fig. 2A) and in its absence (Fig. 2B). From six separate experiments, we found K s ϭ 3 Ϯ 0.8 M for GST ϩ FSK and K s ϭ 12 Ϯ 1.5 M in the absence of forskolin. Both values fit well with the graphs presented in Fig.  1A and are significantly different (p Ͻ 0.001). The substrate inhibition constant (K i ) represents the parameter for inhibition at high concentrations of genistein and is given by a modification of the Lineweaver-Burk representation (as shown in Equa- The intersection on the x axis of the dotted line indicates ϪK i , the reverse concentration of genistein that produced half-maximal inhibition. Fig. 3 shows examples for such an analysis. We found K i values of 75 Ϯ 3.2 M and 70 Ϯ 4.1 M from six separate experiments with GST ϩ FSK and six others with GST alone, respectively. In contrast to the K s values, these values for K i are not significantly different. Finally, dosedependent relationships presented in Fig. 1A were analyzed using the new parameters K s and K i fitted with the following equation (Equation 3) that describes a non-Michaelis-Menten kinetics (18), VЈ is the rate of iodide efflux, S is the concentration of genistein, and V is the new non-Michaelis-Menten velocity. Fig. 4 presents the curves obtained with Equation 3 (to be compared with Fig. 1A). This demonstrates that the pharmacological behavior of CFTR activated by genistein fully follows a non-Michaelis-Menten function.

Perturbation of the non-Michaelis-Menten Behavior of CFTR by the G551D Mutation-
The glycine-to-aspartic acid missense mutation at codon 551 (G551D) is a class III mutation located within the NBD1 domain (5). We used CHO cells stably expressing G551D-CFTR to analyze the effect of genistein. We first determined the relative rate R of 125 I efflux in experiments in which genistein was added in the absence of forskolin. The addition of up to 100 M genistein failed to activate chloride transport in G551D-CFTR-expressing cells. We also noticed that up to 10 M forskolin failed to stimulate G551D-CFTR channels (not shown). However, when we applied simultaneously 10 M forskolin (to phosphorylate CFTR) and 30 M genistein, the average efflux peak rate was dramatically increased to 2.8 Ϯ 0.11 (n ϭ 35) as compared with control (i.e. 10 M FSK:1.40 Ϯ 0.17, n ϭ 12, p Ͻ 0.001 or 30 M GST:1.17 Ϯ 0.13, n ϭ 12, p Ͻ 0.001).
To compare the effect of genistein on G551D-CFTR with our study with wt-CFTR cells, we generated dose-response rela-tionships for genistein in the presence of 10 M forskolin. Example traces from one experiment are shown in Fig. 5A. Surprisingly, at 50 and 400 M genistein, the CFTR-mediated efflux was further increased rather than decreased with efflux rates of 3 and 4.1, respectively. Complete dose-response relationships for activation of G551D-CFTR using concentrations of genistein from 0.1 M to 1000 M were generated from six separate experiments. Results are presented in Fig. 5B. Two  . In B, the Michaelis-Menten equation was used to fit the data. The kinetic parameters were determined assuming V equivalent to the relative rate R and V max corresponding to R max with R max being the maximal rate of the reaction that occurs when [GST] Ͼ Ͼ K m . K m is the substrate concentration that gives half-maximal velocity. Each data point represents the mean Ϯ S.E. from six separate experiments. The corresponding Lineweaver-Burk representation is presented in the inset (see "Results" for equations). ϭ 4 and K m ϭ 11 Ϯ 1.5 M (n ϭ 6). These results obtained from the analysis of six separate experiments with the mutated protein are in sharp contrast with those obtained for wt-CFTR (compare Fig. 5B with Fig. 1A).
Finally, whole cell patch clamp experiments were performed using cells first exposed to 10 M forskolin and then to genistein (30 M). As expected from our iodide efflux experiments, a large current, which was time-and voltage-independent, was stimulated only in cells treated with FSK ϩ GST (Fig. 6, A and  B). The currents reversed at Ϫ40 mV, as shown in Fig. 6D, a value close to the theoretical Cl Ϫ equilibrium potential imposed by our conditions (E Ce Ϫ ϭ Ϫ42 mV, see "Experimental Procedures"), confirming the chloride nature of the current (11,12). The genistein-activated conductance had a current density of 16.42 Ϯ 3.77 pA/picofarads (n ϭ 15) and was statistically different from control conditions (1.92 Ϯ 0.23, pA/picofarads, n ϭ 17, p Ͻ 0.001) when measured at ϩ40 mV (Fig. 6D). Several inhibitors were also used to further characterize G551D-CFTR chloride channel activity. Fig. 6, C-E, shows that in the pres-ence of 100 M glibenclamide, the activation of G551D-CFTR currents is abolished. A time course of the activation of CFTR current at ϩ40mV is illustrated in Fig. 6E. Interestingly, the outwardly rectifying chloride channel blocker calixarene (100 nM, see Ref. 10) has no effect on the currents, whereas a further addition of glibenclamide fully blocked the current. A summary of the effect of four different inhibitors on the rate of iodide efflux stimulated by FSK ϩ GST is also presented in Fig. 6F, showing the inhibition of G551D-CFTR channel activity by 100 M glibenclamide and 250 M DPC but not by 200 M DIDS nor by 100 nM calixarene. DISCUSSION We described here the mechanism of action and kinetic parameters of wt-CFTR chloride channel activated by genistein and the profound alteration resulting from the G551D mutation. This study reveals two important results: (i) a non-Michaelis-Menten behavior describes the pharmacological effect of genistein on wt-CFTR, suggesting the presence of one activatory site and one inhibitory site; (ii) the non-Michaelis-Menten behavior of wt-CFTR is absent and transformed into a classical Michaelis-Menten behavior by the CF mutation G551D. This demonstrates the presence of a single genistein binding site that is activatory and the loss of the inhibitory site. Thus, glycine at position 551 in the NBD1 domain of CFTR is an important location within the inhibitory binding site for genistein.
The concentration-response curve for genistein on wt-CFTR is bell-shaped with an upper plateau that is not straight but tends to flex downwards at high concentrations. A bell-shaped response is referred to as non-Michaelis-Menten behavior, a common mechanism observed for a broad range of enzymes including protein-tyrosine phosphatase (19), 4-␣-carbinolamine dehydratase (20), N-acylphosphatidylethanolamine synthase (21), and receptor activity such as the cAMP accumulation in brown adipocytes (22) associated with ␣1/␤-receptor interaction. Originally, the observation of non-Michaelis-Menten kinetics was considered either as an experimental artifact or as the result of an overstimulation leading to desensitization of the receptor or enzyme. However, in 1994, Rovati and Nicosia (23) demonstrated in a theoretical analysis that such a mechanism could be explained in terms of the resulting interaction of stimulatory and inhibitory components involving two binding sites, one being the activator while a second one is the inhibitor.
Our results support a model with two binding sites for genistein on wt-CFTR but only one binding site on G551D-CFTR. Evidence for two mechanisms of genistein interactions was obtained previously in a study on wt-CFTR channels by Wang et al. (24) and by Lansdell et al. (25), who observed an additional inhibitory effect with high concentrations of genistein. This was interpreted as the presence of a second, low affinity site that should reduce the opening rate of CFTR. In cardiac myocytes, genistein activates CFTR at 20 M but inhibits its activity at 100 M (26). Remarkably, the maximal stimulation of CFTR activity obtained by these authors was 35 M genistein, in close agreement with our study. Moreover, in rat epididymal epithelium, genistein stimulated CFTR current with an EC 50 of 10 M in the absence of forskolin (27), a value similar to that found here. In a recent study, Randak et al. (17) demonstrated that genistein interacts with a fusion protein (comprising a maltose-binding protein and NBD2 of CFTR) by inhibiting ATPase hydrolysis at NBD2. Other studies have demonstrated that ATPase hydrolysis at NBD2 closes the channel (4, 28). Thus, genistein most likely activates wt-CFTR at least in part through the inhibition of ATPase hydrolysis at NBD2 via an interaction with the high affinity site for genis- tein, which may be located in the vicinity of the ATP binding site at NBD2 of CFTR.
We found that K s , the kinetic parameter for activation of wt-CFTR by genistein, depends on the phosphorylation status of CFTR. Indeed, with partially phosphorylated CFTR (i.e. with low concentrations of forskolin present), K s was 4-fold smaller (K s Ϸ 3 M) as compared with non-phosphorylation conditions (i.e. in the absence of forskolin) with K s Ϸ 12 M. Phosphorylation of CFTR may facilitate the accessibility of the high affinity binding site to genistein through domain interactions or conformation changes. On the contrary, at high concentrations of genistein, the inhibition of wt-CFTR appears to be independent of the phosphorylation status of CFTR since K i was unchanged (K i Ϸ 70 -75 M with or without phosphorylation of CFTR). Thus, we hypothesize that phosphorylation of the R domain interacts only with the high affinity binding site for genistein (i.e. at NBD2) to increase the accessibility of the site. Indeed, one major difference concerning the phosphorylation process is that with G551D-CFTR, much higher concentrations of forskolin are required to achieve the stimulation by genistein of the CFTR channel activity. The G551D-CFTR protein is a class III mutation that produced a fully glycosylated protein correctly located at the apical membrane (i.e. normal biosynthesis, trafficking, and processing) and normally phosphorylated at the R domain by cAMP-dependent protein kinase (6). However, G551D proteins have decreased nucleotide binding (7) and a reduced ATPase activity at NBD1 (8,9). The mutation G551D also disrupts activation and regulation of CFTR at the plasma membrane since phosphorylation alone is not able to achieve efficient stimulation of chloride transport in G551D cells (12,29,30).
Our study of the activity of G551D-CFTR mutant demonstrates that the inhibitory component of the dual mechanism of the action of genistein on wt-CFTR is lacking. Therefore, we propose that the inhibitory site of low affinity is located within a region of NBD1 close to glycine 551. The mutations affecting the NBD1 domain of CFTR generally lead to a severe form of the disease cystic fibrosis because they prevent the trafficking of CFTR to the membrane or because they generate abnormal CFTR chloride channel activity (31). In addition, our study reveals that the G551D mutation profoundly alters the pharmacological behavior of CFTR.
The finding that CFTR activity in the presence of genistein obeys a non-Michaelis-Menten behavior indicates an alternative pathway to achieve functional desensitization of CFTR. For example, the cAMP accumulation in cultured mature brown adipocytes in response to the ␤-agonist isoprenaline could be explained by the authors as the result of two counteracting components (22): a ␤-adrenergic receptor stimulatory component and an ␣1-adrenergic receptor inhibitory component. Interestingly, the ␣-adrenergic receptor antagonist prazosin transforms the bell-shaped response into a simple Michaelis-Menten curve (22). In some enzymatic reactions, the change between Michaelis-Menten and non-Michaelis-Menten behaviors is referred to as "mechanism switching" (32, 33). We have described here a similar effect with the class III CF mutation G551D and showed that it abolishes the inhibitory response to genistein. These results finally raised important questions such as the nature of the physiological element that may control the inhibition of CFTR. Because CFTR plays a key role in the cAMP-dependent secretory pathway in epithelia and controls a variety of other channels and transporters (14,31), this inhibitory site maybe be crucial for the physiological desensitization of CFTR.
In conclusion, our results show that the high affinity stimu-latory site for genistein, located in NBD2, is dependent on the phosphorylation status of CFTR and preserved by G551D mutation. The second site is lacking in G551D-CFTR cells. This site, of low affinity for genistein, is an inhibitor in non-mutated CFTR. These data suggest that a Rovati/Nicosia model (non-Michaelis-Menten behavior (23)) describes the kinetics of the control of CFTR channel activity by genistein. Finally, it shows that glycine at position 551 in NBD1 is an important location within the binding site for genistein. Thus, not only did class III CF mutations produced protein with defective regulation and opening, but they also altered the structure and pharmacological properties of CFTR. To our knowledge, this study demonstrates for the first time how a mutation of an ion channel converts a non-Michaelis-Menten behavior (two binding sites) into a classical Michaelis-Menten model (one binding site).