Vx-770 potentiates CFTR function by promoting decoupling between the gating cycle and ATP hydrolysis cycle

Abstract

Vx-770 (Ivacaftor), a Food and Drug Administration (FDA)-approved drug for clinical application to patients with cystic fibrosis (CF), shifts the paradigm from conventional symptomatic treatments to therapeutics directly tackling the root of the disease: functional defects of the cystic fibrosis transmembrane conductance regulator (CFTR) chloride channel caused by pathogenic mutations. The underlying mechanism for the action of Vx-770 remains elusive partly because this compound not only increases the activity of wild-type (WT) channels whose gating is primarily controlled by ATP binding/hydrolysis, but also improves the function of G551D-CFTR, a disease-associated mutation that abolishes CFTR’s responsiveness to ATP. Here we provide a unified theory to account for this dual effect of Vx-770. We found that Vx-770 enhances spontaneous, ATP-independent activity of WT-CFTR to a similar magnitude as its effects on G551D channels, a result essentially explaining Vx-770’s effect on G551D-CFTR. Furthermore, Vx-770 increases the open time of WT-CFTR in an [ATP]-dependent manner. This distinct kinetic effect is accountable with a newly proposed CFTR gating model depicting an [ATP]-dependent “reentry” mechanism that allows CFTR shuffling among different open states by undergoing multiple rounds of ATP hydrolysis. We further examined the effect of Vx-770 on R352C-CFTR, a unique mutant that allows direct observation of hydrolysis-triggered gating events. Our data corroborate that Vx-770 increases the open time of WT-CFTR by stabilizing a posthydrolytic open state and thereby fosters decoupling between the gating cycle and ATP hydrolysis cycle. The current study also suggests that this unique mechanism of drug action can be further exploited to develop strategies that enhance the function of CFTR.

Cystic fibrosis (CF), a genetic disease affecting 1 in every ∼2,500 newborns in Caucasian populations (1), results in multiorgan dysfunction that eventually leads to premature death (2). The culprit behind CF is the malfunction of the cystic fibrosis transmembrane conductance regulator (CFTR) protein, a chloride channel that plays a critical role in maintaining water and salt homeostasis across many epithelium-lining tissues (3–5). As a member of the ATP-binding cassette (ABC) protein superfamily, CFTR inherits the canonical domains found in most ABC proteins, namely the two nucleotide binding domains (NBDs) that use ATP hydrolysis as the free energy source to drive the transport/gating cycle and the two transmembrane domains (TMDs) that form the substrate translocation pathway. In addition, CFTR possesses a unique regulatory (R) domain that contains multiple consensus serine residues for protein kinase A (PKA)-dependent phosphorylation (6–8). Phosphorylation of the R domain enables CFTR to assume the function of an ATP-gated ion channel. It is generally accepted that after CFTR is phosphorylated by PKA, opening and closing of the gate in the TMDs are triggered by ATP-induced NBD dimerization and hydrolysis-elicited NBD dimer separation respectively (9–12). Mutations that disrupt this process could compromise the channel function and cause CF. For example, defects in NBD dimer formation have been proposed to account for the low open probability (Po) seen in ∆F508- and G551D-CFTR, the first and third most common CF-associated mutations (13–15). Therefore, finding small molecules (i.e., CFTR potentiators) that can improve the function of CFTR is expected to pave the way for targeted treatment in CF (16).

In the past decades, although advances in symptomatic treatments have significantly prolonged the life expectancy of patients with CF (17), little success has been made in translating CFTR potentiators to clinical use until very recently. The discovery of a CFTR potentiator, Vx-770 (Ivacaftor, structure shown in Fig. 1B), via high-throughput drug screening (18) has successfully evolved into a unique CF treatment that targets the fundamental defect in CFTR gating. Vx-770 has shown promising clinical outcomes for patients carrying the G551D mutation (19, 20) and was recently approved by the FDA. This ground-breaking invention thus opened a unique chapter for CF therapeutics as we are now able to tackle the root of this disease. However, the detailed mechanism for Vx-770’s action remains obscure.

Fig. 1.

Energetic coupling model of CFTR gating and the structure of Vx-770. (A) Energetic coupling model for CFTR gating (modified from ref. 24). The red box marks the reentry pathway. The rate of C2 → C1 transition is extremely slow for WT-CFTR, therefore in the continuous presence of ATP, the channel rarely visits the C1 state. C1, C2, C2 ATP, and C2 ATP dimerized (C2 AD) represent closed states and O1, O2, and O2 ATP represent open states with different NBD configurations as illustrated in the cartoon. Semiquantitative analysis suggested that the O1 state is extremely stable (see ref. 24 for detail). As a result, for WT-CFTR, closure of most opening events is the consequence of ATP hydrolysis, which effectively creates a shortcut for the channel to escape from the otherwise very stable O1 state. (B) Chemical structure of Vx-770. A was modified from figure 6 in ref. 24.

Vx-770 can act on WT-CFTR as well as many disease-related CFTR mutants (21). It boosts the Po of CFTR mainly by prolonging its open time (18, 21). Perhaps the most puzzling observation is that Vx-770 not only prolongs the open time of WT-CFTR, which is controlled by ATP hydrolysis; it also increases the activity of G551D-CFTR, a mutation that likely prevents ATP-induced NBD dimerization from happening due to its unique position in the signature sequence (14) and for the same reason eliminates ATP hydrolysis (22, 23). This apparent paradox is no longer indecipherable as unique insights into CFTR’s gating mechanism have emerged (24).

The newly proposed energetic coupling model (Fig. 1A) for CFTR gating offers a unique mechanism for prolonging the open time of WT-CFTR besides delaying hydrolysis and/or NBD dimer separation implicated in other gating models (25, 26). As shown in Fig. 1A, after ATP hydrolysis and dissociation of the hydrolytic products, the gate remains open for a finite period (i.e., O2 state), wherein another ATP molecule can capture this posthydrolytic open state and shuffle the channel back to the prehydrolytic open state (O1). Because during this “reentry” transition (O2 → O2ATP → O1 in Fig. 1A), the channel stays in different open states, the open time is expected to increase when the frequency of reentry is augmented. In addition, as the first step of reentry is ATP binding to the emptied ATP binding site (O2 ↔ O2ATP in Fig. 1A), prolongation of the open time caused by an increase of reentry events is predicted to be [ATP] dependent. Indeed, such prediction was validated by our experiments with WT-CFTR treated with Vx-770. Furthermore, a recent fortuitous finding of a mutant, R352C-CFTR, which allows quantification of ATP hydrolysis-triggered open-to-open transition, imparts a straightforward approach to examine how Vx-770 affects the reentry pathway. Our data with R352C-CFTR not only support the hypothesis that Vx-770 lengthens the open time of WT-CFTR by increasing the frequency of reentry, but also spawn unique targets for CFTR potentiators that could complement the action of Vx-770, a crucial step toward the ultimate goal of curing CF.

Results

Vx-770 was reported to increase the Po of WT-CFTR mainly by prolonging its open time (18, 21). Besides slowing down ATP hydrolysis and/or delaying NBD dimer separation, which can increase the open time, an updated gating model shown in Fig. 1A depicts another means to achieve this kinetic effect, namely, promoting the reentry pathway (O2 → O2ATP → O1 in Fig. 1A). Because ATP-induced NBD dimerization and subsequent ATP hydrolysis are likely abolished by the G551D mutation, these two plausible mechanisms may not easily account for the effect of Vx-770 on G551D-CFTR. As for the third possible mechanism, promoting the reentry pathway does not seem to be a good candidate at first glance as the reentry pathway requires ATP-induced NBD dimerization as well. However, the gating model presented in Fig. 1A does offer a testable hypothesis for achieving dual effects of Vx-770 in potentiating both WT- and G551-CFTR. If Vx-770 shifts the equilibrium of the C2 ↔ O2 transition (i.e., ATP-independent gating) toward the O2 state, one expects to observe an increased activity of G551D-CFTR as ATP-independent openings are preserved in this mutant, as well as for WT-CFTR because more reentry events will occur with a stabilized O2 state.

This hypothesis predicts three experimentally verifiable outcomes for WT-CFTR: first, ATP-independent activity upon ATP washout should increase as the C2 ↔ O2 equilibrium is shifted toward the O2 state by Vx-770. Second, because ATP is required to shuffle the channel from O2 to O1, higher [ATP] is expected to increase the reentry frequency to a larger extent, resulting in a longer open time. Third, stabilization of the O2 state by Vx-770 will enhance the effect of pyrophosphate (PPi) as this nonhydrolyzable ligand can bind to the vacated ATP binding site. These predictions were tested by the following experiments.

Vx-770 Increases ATP-Independent Gating.

To quantitatively examine ATP-independent activity of WT-CFTR, we measured macroscopic currents before and after washout of ATP in excised inside-out patches wherein CFTR channels were prephosphorylated by PKA (Fig. 2). Upon ATP application, the channels were rapidly activated, represented by a sharp increase of the current. After ATP washout, whereas the macroscopic current decayed rapidly with a time constant of hundreds of milliseconds, some residual activity that represents infrequent ATP-independent openings remained discernible even seconds after removal of ATP (Fig. 2 A and B). We measured the ratio of the postwashout ATP-independent current to the robust ATP-induced current to assess the ATP-independent activity and found that this ratio for WT-CFTR is approximately four times higher when treated with 200 nM Vx-770 (Fig. 2C). Notably, because Vx-770 treatment nearly doubles the Po of the WT-CFTR (18), the ratio presented in Fig. 2C underestimates the true effect of Vx-770 on ATP-independent gating. After taking into consideration the difference in Po between control and drug-treated channels, Vx-770 actually increases the ATP-independent Po by approximately ninefold (Fig. 2D). Similarly, we found that Vx-770 increases the mean macroscopic current of G551D-CFTR by 12.8 ± 1.9 fold (n = 12; Fig. S1). Taking account of the basal Po of G551D-CFTR (∼120 times lower than WT) (14), we estimated a Po of ∼0.04 for Vx-770-treated G551D-CFTR. Our results are somewhat different from those reported previously (an eightfold increase by Vx-770 with a rectified Po of ∼0.15) (18), and thus demand further mechanistic studies of Vx-770 on G551D.

Fig. 2.

Effects of Vx-770 on ATP-independent activity of WT-CFTR. (A and B) Representative traces for ATP-induced macroscopic current of WT-CFTR channels in the absence (A) or presence (B) of Vx-770. Patches were perfused with ATP until the current reaches the steady state to measure ATP-induced current. ATP-independent activity was measured after ATP washout (expanded in the red and blue boxes). Traces shown in A and B were chosen because they exhibit similar amplitude of ATP-independent current, whereas the ATP-induced current is much greater in A, indicating a higher ATP-independent activity in the presence of Vx-770. (C) Effects of Vx-770 on the ratio of ATP-independent currents to ATP-induced currents. (D) Estimated Po for ATP-independent activity in the presence or absence of Vx-770. Bars above each trace mark applications of the indicated ligand(s) to the patch (same in all figures). *P < 0.05.

We next analyzed microscopic kinetics of spontaneous ATP-independent gating in the presence or absence of Vx-770 and found that Vx-770 increases the mean open time by approximately twofold (332 ± 39 ms, n = 18, vs. 656 ± 136 ms, n = 12, P < 0.05). The much larger increase of the Po of spontaneous ATP-independent gating by Vx-770 indicates that Vx-770 also raises the spontaneous opening rate of WT-CFTR.

Effects of Vx-770 on Microscopic Kinetics of ATP-Gated WT-CFTR.

As described above, the model shown in Fig. 1A dictates that stabilizing the O2 state (see below) should present a higher probability for ATP to capture the O2 state and bring it back to the O1 state. Because this reentry pathway is an ATP-dependent process (O2 → O2 ATP requires ATP binding), it is predicted that more reentry events will occur at higher [ATP]. The increased reentry frequency makes the channel shuttling between different open states and thereby effectively delays channel closure. As shown in Fig. 3A, WT-CFTR channels open into bursts that last for hundreds of milliseconds. These opening bursts are separated by interburst closures with a time constant of hundreds of milliseconds to seconds depending on [ATP]. Notably, there are brief closings buried within each opening burst known as flickers. However, because these flickering closings were traditionally considered ATP-independent events (12, 26–28), a computer program had been developed by Csanady et al. (29) to isolate the ATP-dependent opening bursts for single-channel kinetic analysis.

Fig. 3.

Vx-770 prolongs the open time of WT-CFTR in an ATP-dependent manner. (A) Representative single-channel traces for WT-CFTR in conditions depicted above each trace. (B) Summary of the mean open time at different [ATP] in the presence (black markers) or absence (blue markers) of 200 nM Vx-770. Mean Po for conditions denoted in A–C are 0.38 ± 0.03, n = 13; 0.71 ± 0.02, n = 25 and 0.45 ± 0.02, n = 17, respectively. (C) Summary of the opening rate of WT-CFTR at different [ATP] in the presence (black markers) or absence (blue markers) of 200 nM Vx-770. (D) In the presence of 200 nM Vx-770, the relaxation time constants of macroscopic CFTR current after removing specified concentrations of ATP. *P < 0.05 compared with 10 mM ATP. ^#P < 0.05 compared with the same [ATP] but without Vx-770.

In the presence of Vx-770, the burst duration of WT-CFTR was indeed visibly longer than in its absence (Fig. 3A). Furthermore, such prolongation of the opening burst became even more prominent at higher [ATP] (Fig. 3B), an observation not only corroborating the hypothesis that Vx-770 promotes the ATP-dependent reentry mechanism, but also distinguishing itself from the well-known increase of the open time by abolishing ATP hydrolysis (12, 28, 30–32). It is noteworthy that such [ATP]-dependent open time was not observed with WT-CFTR in the absence of Vx-770 (refs. 9, 12, 33, cf. 34), presumably due to a relatively low reentry frequency even in the presence of millimolar [ATP] so that minute changes in the open time were masked by a large patch-to-patch variation (see ref. 24 for details).

We also measured the relaxation time constant of macroscopic current after removing different concentrations of ATP. In the presence of 200 nM Vx-770, these relaxation time constants are not significantly different over a wide range of [ATP] (Fig. 3D), suggesting that the open time measured with the macroscopic relaxation analysis remains fairly constant once the reentry pathway is eliminated by removing ATP. It is, however, noted that the relaxation time constant is slightly longer than the mean open time measured at low [ATP] in the presence of same concentration of Vx-770 (Fig. 3B and SI Discussion).

Besides prolonging the open time, Vx-770 also increases the opening rate of WT-CFTR (Fig. 3C). This result suggests that kinetic steps other than the C2 ↔ O2 transitions are affected by Vx-770 (Discussion).

Vx-770 Enhances the Effect of PPi.

Whereas the O2 state can bind ATP to initiate another round of ATP hydrolysis (i.e., O2 → O2ATP → O1 → O2 in Fig. 1A), it is expected that when the O2 state captures a nonhydrolyzable ATP analog such as PPi, the channel will sojourn to a locked-open state with a lifetime in tens of seconds (30, 35, 36). In a previous study, we have developed a protocol to quantify such effect of PPi (36). As shown in Fig. S2, when the ligand was switched from ATP to a 1-s pulse of PPi, the current dropped immediately but halted at the halfway point, followed by a slow decay that reflects the gradual closure of PPi locked-open channels. In patches treated with Vx-770, the locked-open current (indicated by the red mark) relative to ATP-induced current is significantly (P < 0.01) higher than that of control (Fig. S2 and ref. 36).

Effects of Vx-770 on R352C-CFTR.

The experimental results described so far are consistent with the idea that Vx-770 stabilizes the O2 state and thereby confers a longer window of time for the channel to respond to ATP (Fig. 3) or PPi (Fig. S2). However, measurements of the mean open time for spontaneous ATP-independent gating described above may be subjected to some technical deficiency (SI Discussion). Strictly speaking, the interpretation of an increased frequency of reentry events by Vx-770 is an inferred one as open-to-open transitions of WT-CFTR during reentry are “invisible.” Fortunately, our recent discovery of a CFTR mutant, R352C, that exhibits different single-channel conductance for the O1 and O2 states grants us a unique opportunity to directly assess the effects of Vx-770 on the lifetime of the O2 state. Furthermore, as the O1 to O2 transition can be visually discerned with this mutant, the effect of Vx-770 on the frequency of reentry events can be quantified by simply counting the number of openings that embody more than one round of O1 → O2 transition (see ref. 24 for detailed explanation).

Fig. 4 shows representative single-channel current traces in the absence (Fig. 4A) or presence (Fig. 4B) of Vx-770. Table 1 summarizes the number of events with different characteristic gating patterns with or without Vx-770. As shown in our recent report (24), most of the opening bursts in R352C-CFTR contain one O1 → O2 transition; only ∼10% of the openings are categorized as events with more than one O1 → O2 transition. However, the percentage of this latter event was nearly doubled in the presence of 200 nM Vx-770, indicating that Vx-770 increases the frequency of reentry. Furthermore, the single-channel current traces in Fig. 4 A and B also suggest that the lifetime of the opening bursts is increased by Vx-770. Indeed, by constructing the open time histograms by pooling all opening events from seven patches, we obtained a mean opening burst time of ∼153 ms in the absence of Vx-770 and ∼188 ms in its presence (Fig. S3A and ref. 24). This increase of the overall bursting time by Vx-770 is mostly due to a 2.5-fold increase of the O2 dwell time (110 ms vs. 43 ms; Fig. S3C) with a minor effect on the O1 dwell time (103 ms vs. 84 ms; Fig. S3B).

Fig. 4.

Effects of Vx-770 on R352C-CFTR. (A and B) Representative single-channel traces for R352C-CFTR in the absence (A) or presence (B) of 200 nM Vx-770. When treated by Vx-770, the percentage of opening events contain more than one O1 → O2 transition is increased (summarized in Table 1) and the mean open time is prolonged (Fig. S3).

Table 1.

Summary of the effects of Vx-770 on the gating patterns exhibited in R352C-CFTR and W401F/R352C-CFTR

Five different categories of the gating pattern are illustrated. Number of opening events and their percentage for each category are displayed. [ATP] was kept at 2.75 mM for all experiments. #, the (O1-O2)n category includes events that contain at least one O2 → O1 transition (such as C → O1 → O2 → O1 → O2 → C, C → O2 → O1 → O2 → C, C → O1 → O2 → O1 → C, etc.).

*Data for R352C-CFTR and R352C/W401F-CFTR in the absence of Vx-770 were taken from ref. 24.

Discussion

Vx-770 is the first FDA-approved CF medication that directly targets CFTR. In vitro studies showed that Vx-770 significantly increases the Po of WT as well as many disease-related CFTR mutants (18, 21). This effect is likely to occur in vivo as clinical studies showed symptomatic improvements in patients with CF treated with Vx-770 (19, 20). Ironically, the detailed mechanism of Vx-770 remains mostly unknown.

As a CFTR potentiator, Vx-770 is unique in that it increases the open time of WT-CFTR, which is mainly controlled by ATP hydrolysis (11, 28, 31, 32) and the activity of G551D-CFTR, a mutant that does not respond to ATP (14). Indeed, the classical methods used to lengthen the open time of WT-CFTR by abolishing ATP hydrolysis with either mutations or nonhydrolyzable ATP analogs fail to work on the G551D channels (14, 37). One possible explanation for this apparent dichotomy is that Vx-770 works on WT- and G551D-CFTR channels via two unrelated mechanisms. However, the observation that the magnitude of Vx-770’s effects on ATP-independent activity of WT-CFTR (∼9×, Fig. 2D) is similar to the fold increase of G551D-CFTR (∼8×, ref. 18, ∼13× in the current study) suggests a unified theory may explain Vx-770’s dual actions.

As described in Results, the newly proposed CFTR gating model (Fig. 1A) offers a reasonable mechanism for the action of Vx-770. A distinctive feature of this model is that ATP-independent gating (C2 ↔ O2) is an integral part of the gating transitions observed for WT-CFTR in the presence of ATP. The model also features a nonstrict coupling between ATP hydrolysis and each opening/closing event, allowing hydrolysis of more than one ATP molecule within an opening burst. Based on this model, one can explain Vx-770’s effects on both WT- and G551D-CFTR by proposing that the drug shifts the equilibrium of C2 ↔ O2 transition toward the O2 state. By doing so, Vx-770 can increase the open time of WT-CFTR because now more opening bursts embody multiple rounds of ATP hydrolysis. Our demonstration of [ATP]-dependent increase of the open time (Fig. 3B) in the presence of Vx-770 corroborates this idea. Furthermore, by examining the effects of Vx-770 on R352C-CFTR, we were able to show that this compound indeed increases the frequency of reentry events (Fig. 4 and Table 1).

Notably, the reentry pathway was first suggested by a report using PPi as a tool to fish out a unique posthydrolytic state (36) and later established by scrutinizing the R352C mutant channel that exhibits hydrolysis-dependent open channel conductance: unequal single-channel current amplitudes between pre- and posthydrolytic states (24). Direct evidence for the existence of reentry events in WT-CFTR was lacking because most previous studies failed to observe [ATP]-dependent open time for WT-CFTR (refs. 9, 12, 33 cf. 34). Therefore, the observation that WT-CFTR’s open time can be prolonged by increasing [ATP] in the presence of Vx-770 indicates that the reentry pathway is indeed an integral part of WT-CFTR gating scheme.

Perhaps one of the most interesting questions yet to be answered for Vx-770 pharmacology is its binding site(s). The R domain is not likely to be the target because Vx-770 effectively potentiates the activity of a CFTR construct with the whole R domain removed (∆R-CFTR, ref. 9; Fig. S4). Several clues lead us to propose that Vx-770 may bind to the TMDs of CFTR. First, Vx-770 is extremely hydrophobic. (Precipitates were seen when [Vx-770] > 5 μM.) Thus, a significant partition of Vx-770 into the lipid bilayer is expected. Second, the kinetic steps (i.e., opening and closing of the gate in the absence of ATP, C2 ↔ O2) primarily affected by Vx-770 are supposed to take place in CFTR’s TMDs (24). The observation that Vx-770 increases the opening rate of ATP-gated WT-CFTR (Fig. 3C) can also be explained if Vx-770 shifts the equilibrium of C2 ATP dimerized (C2 AD) ↔ O1 (opening and closing of the gate in TMDs with dimerized NBDs). Third, our previous studies on CFTR’s TMDs have demonstrated a similar and yet more robust phenotype—an increase of spontaneous ATP-independent gating accompanied by a prolongation of ATP-dependent open time as Vx-770 by chemical modifications of engineered cysteine in the sixth transmembrane segment of CFTR (38). If Vx-770 binds to CFTR’s TMDs, applying the drug from the intracellular side or the extracellular side should yield similar effects. Indeed, as shown in Fig. S5A, the open time for WT-CFTR is visibly lengthened by pipette application of Vx-770. Furthermore, the effects of Vx-770 on the open time are very similar regardless of whether it is applied from the cytoplasmic or extracellular side of the membrane (Fig. S5C). Although more studies are needed to establish exactly where Vx-770 binds, this result does underscore the potential of CFTR’s TMDs as a drug target.

Our result (Fig. 2) indicates that the spontaneous ATP-independent Po in the presence of Vx-770 is still less than 1/10th of the ATP-gated WT-CFTR activity. In contrast, Eckford et al. (39) recently reported that Vx-770 increases ATP-independent Po of WT-CFTR to a level of ∼0.4 (vs. ∼0.03 in Fig. 2D). Because the ATP-independent Po for CFTR is ∼0.004 (28), such result would mean that Vx-770 increases the ATP-independent activity by ∼100-fold, a magnitude similar to the maximal effect of CFTR’s natural ligand ATP! Whereas the exact reason for the difference between current results and those reported by Eckford et al. (39) awaits further studies, it is important to point out that an accurate estimation of the Po for G551D-CFTR from microscopic kinetic analysis is challenging because the number of channels cannot be reliably determined.

Although Vx-770 is now an established treatment for patients with CF, its clinical impacts are somewhat limited as the G551D mutation only accounts for ∼3% of CF-associated mutations (www.genet.sickkids.on.ca/cftr/app). Furthermore, the sweat chloride level, a parameter directly reflecting CFTR activity in vivo and thus a widely used diagnostic tool for CF, in patients taking Vx-770 (19, 20) were still higher than those found in patients with mild-form CF (19, 20, 40), suggesting that Vx-770 alone is insufficient to completely rectify the dysfunction associated with the G551D mutation in vivo. Furthermore, recent clinical trials suggest that treatments with drugs that can improve the trafficking of ∆F508-CFTR may not be sufficient to ameliorate the clinical symptoms in patient carrying the ∆F508 mutation (1) presumably because this most common pathogenic mutation causes deficits in both membrane expression and gating (41–46). Thus, discovering new ways to enhance CFTR activity remains an outstanding goal in the field. If one accepts the reentry mechanism as a potential route to improve CFTR function, the model in Fig. 1 actually offers another potential target: accelerating the O2 → O2ATP transition could enhance the reentry pathway and such maneuver is expected to complement the effect of Vx-770.

Lately, we discovered that W401F, a conserved mutation in NBD1, increases the reentry frequency without significantly altering the O2 → C2 transition (24). Although how the W401F mutation modifies the function of NBDs is unclear, this finding does raise the possibility that manipulating NBD function can complement the action of Vx-770. As shown in Fig. S3 and Table 1, the W401F mutation almost doubles the reentry frequency of R352C-CFTR and also significantly enhances the effect of Vx-770 (Fig. 5, Table 1 and Fig. S3). Future studies along this line may shed light on rational drug design for CF treatment.

Fig. 5.

W401F mutation and Vx-770 work additively on the reentry pathway. Representative single-channel traces for R352C/W401F-CFTR treated with 2.75 mM ATP in the absence (A) or presence (B) of 200 nM Vx-770. Frequency of opening bursts containing multiple rounds of O1 → O2 transition (summarized in Table 1) is increased and the overall open time is prolonged (Fig. S3D) compared with R352C-CFTR recorded in the same condition.

Materials and Methods

Cell Culture and Transient Expression System.

We used Dulbecco’s modified Eagle’s medium supplemented with 10% (vol/vol) FBS to grow Chinese hamster ovary (CHO) cells at 37 °C. One day before performing the transfection, the cells were trypsinized, transferred to 35-mm tissue culture dishes, and cotransfected with the cDNA of CFTR and green fluorescence protein (GFP) encoding pEGFP-C3 (Clontech) using PolyFect transfection reagent (Qiagen). After transfection, culture dishes containing the transfected cells were grown at 27 °C for 2 d before patch-clamp experiments.

Mutagenesis.

All mutations used in this study were constructed using QuikChange XL kit (Stratagene) according to the manufacturer’s protocols. The DNA constructs were sequenced (DNA core, University of Missouri) to confirm the mutation made on cDNA.

Electrophysiological Recordings.

An EPC10 amplifier (HEKA) was used for all of the electrophysiological experiments carried out at room temperature. When performing patch-clamp experiments, glass chips carrying the transfected cells were transferred to a chamber located on the stage of an inverted microscope (IX51; Olympus). Flaming/Brown-type micropipette puller (P97; Sutter Instrument) was used to pull borosilicate capillary glasses into pipettes. Then, a homemade microforge was used to polish the pipette tip before experiments. The resistance of polished pipettes was 2–4 MΩ in the bath solution. Membrane patches were excised into an inside-out mode after observing the seal resistance > 40 GΩ. 25 IU PKA and 2.75 mM ATP were then perfused to the patch membrane until the CFTR current reached a steady state. To maintain the phosphorylation level, 10 IU PKA was added to all other ATP-containing solutions applied thereafter. Throughout the recording, the membrane potential was kept at −60 mV. An eight-pole Bessel filter (LPF-8; Warner Instruments) was used to filter the data with a 100-Hz cutoff frequency and digitized to a computer at a sampling rate of 500 Hz. For a better visual effect of our data presentation, the inward current was inverted so that upward deflections represent channel openings. A fast solution exchange perfusion system (SF-77B; Warner Instruments) was used for recordings that require ligand exchange. The dead time of solution change is ∼30 ms (30).

Notably, we observed that once the patch was exposed to Vx-770, its effect cannot be washed out by continuous perfusing Vx-770 free solution. As a result, we could not bracket our Vx-770 experiments with controls. Instead, all Vx-770 experiments were done after obtaining a control in the same patch. To minimize Vx-770 contamination, following each experiment all of the devices that were in contact with Vx-770 were repeatedly washed with 50% DMSO.

Chemicals and Composition Solutions.

For WT-CFTR, we used the 150 mM Cl⁻ pipette solution containing (in millimoles): 140 methyl-D-glucamine chloride (NMDG-Cl), 2 MgCl₂, 5 CaCl₂, and 10 Hepes [pH 7.4 with N-methyl-d-glucamine (NMDG)]. Because the single-channel conductance is reduced by the R352C mutation, a pipette solution with 375 mM Cl⁻ was used (in millimoles): 360 methyl-D-glucamine chloride (NMDG-Cl), 2 MgCl₂, 5 CaCl₂, and 10 Hepes (pH 7.4 with NMDG). Cells were perfused with a bath solution containing (in millimoles): 145 NaCl, 5 KCl, 2 MgCl₂, 1 CaCl₂, 5 glucose, 5 Hepes, and 20 sucrose (pH 7.4 with NaOH). For inside-out configuration, the 150 mM Cl⁻ perfusion solution contained (in millimoles): 150 NMDG-Cl, 2 MgCl₂, 10 EGTA, and 8 Tris (pH 7.4 with NMDG), the 375 mM Cl⁻ perfusion solution used for R352C-CFTR contained (in millimoles): 375 NMDG-Cl, 2 MgCl₂, 10 EGTA, and 8 Tris (pH 7.4 with NMDG).

MgATP and PKA were purchased from Sigma-Aldrich. MgATP was stored as a 250 mM stock at −20 °C. Vx-770, a generous gift from Robert Bridges (Rosalind Franklin University, North Chicago, IL), was dissolved in DMSO and stored as a 100-µM stock at −70 °C. All chemicals were diluted to the concentration indicated in each figure using the perfusion solution and the pH was adjusted to 7.4 with NMDG. A total of 200 nM Vx-770 is considered maximally effective because increasing [Vx-770] to 500 nM did not further increase the macroscopic current amplitude (Fig. S6).

Data Analysis and Statistics.

Measurements of the steady-state mean current amplitude and plots of the dwell-time histograms were done with the Igor Pro program (Wavemetrics). Levenberg–Marquardt-based algorithm within the Igor Pro program was used to fit the current relaxation with a single exponential function. Single-channel kinetic parameters were obtained with a program developed by Csanady (29) that allows fast microscopic kinetic analysis for recordings that contain fewer than eight opening steps. We took a more conservative approach to only analyze traces with up to three opening steps. For kinetic analysis of R352C-CFTR, three current levels (C, O1, and O2) were determined from all point histograms. The dwell times of each O1 and O2 were measured manually. Because of the limited bandwidth of recordings as well as the small signal-to-noise ratio for the O1 → O2 transition, events shorter than 10 ms cannot be accurately measured; resulting missed events may be partly responsible for the apparent paucity of brief events in the dwell-time histograms shown in Fig. S3. For those ambiguous events, they were counted as real openings if the current stayed above the half amplitude threshold for 6 ms (four data points); otherwise they were considered as noises. Therefore, the minimum measurable open time is ∼10 ms (four data points above the half amplitude threshold plus one data point before and one after, a total of six data points with a 10-ms duration). With this analytic constraint, the events most affected will be the shortest O2 state, resulting in an underestimation of opening bursts with single or multiple O1 to O2 transition. Although the absolute values of the kinetic parameters may also be subject to these technical limitations, the difference between control and Vx-770–treated groups should be valid, as both were subject to the same errors.

The Igor program was used to measure the relaxation time constant of macroscopic CFTR current after ATP removal. The current relaxation phase was fit with built-in single exponential function using the Levenberg–Marquardt-based algorithm.