A mutation in the fifth transmembrane domain of the cystic fibrosis transmembrane conductance regulator (CFTR) Cl− channel (V317E) resulted in whole cell currents that exhibited marked outward rectification on expression in Xenopus oocytes. However, the single-channel unitary current (i)-voltage (V) relationship failed to account for the rectification of whole cell currents. In excised patches containing one to a few channels, the time-averaged single-channel current (I)-V relationship (I = N ×P o × i, where N is the number of active channels and P o is open probability) of V317E CFTR displayed outward rectification, whereas that of wild-type CFTR was linear, indicating that theP o of V317E CFTR is voltage dependent. The decrease in P o at negative potentials was due to both a decreased burst duration and a decreased opening rate that could not be ameliorated by a 10-fold increase in ATP concentration. This behavior appears to reflect a true voltage dependence of the gating process because the P o-V relationship did not depend on the direction of Cl− movement. The results are consistent with the introduction, by a point mutation, of a novel voltage-dependent gating mode that may provide a useful tool for probing the portions of the protein that move in response to ATP-dependent gating.
- cystic fibrosis transmembrane conductance regulator
- chloride channel
- voltage dependence
the cystic fibrosistransmembrane conductance regulator (CFTR) is a member of the ATP-binding cassette family (24) and functions as a plasma membrane Cl− channel that is expressed in many epithelial tissues (1, 3). Channel activity in CFTR is controlled by protein kinase A (PKA)-dependent phosphorylation of its regulatory domain; subsequent interactions of ATP with the nucleotide binding domains (NBDs) are coupled to channel gating. Neither gating nor conductance of single wild-type (WT) CFTR channels is thought to be voltage dependent. Similarly, macroscopic currents of WT CFTR are time independent.
We studied the impact on channel function of a mutation (V317E) at a site in the extracellular half of the fifth transmembrane domain (TM5). Mutations in this region of TM5 alter the selectivity properties of macroscopic currents (25). Mutations at a roughly homologous position in TM11 (S1118) confer voltage-jump relaxations of macroscopic current, consistent with voltage-dependent gating (31). V317E CFTR macroscopic currents display marked rectification. In principle, this behavior could result from either a change in conduction or a change in gating, but the mutation revealed only a minor effect on single-channel conductance. Instead, we found that the macroscopic rectification of V317E CFTR was due to a voltage dependence of the open probability (P o) of single channels. Hence, these results have implications for understanding the link between ATP-dependent gating and the conformational changes that gate the pore and suggest that V317 and S1118 may be at a key site of conformational change that is associated with channel gating. Portions of these data have been presented in abstract form (25, 30).
MATERIALS AND METHODS
Oocyte preparation and cRNA injection.
The methods used were similar to those described previously (19,21). Briefly, stage V-VI oocytes from Xenopus were prepared as previously described (22) and were incubated at 18°C in a modified Leibovitz's L-15 medium with the addition of HEPES (pH 7.5), gentamicin, penicillin, and streptomycin. cRNA was prepared from a construct carrying the full coding region of CFTR in the pALTER vector (Promega, Madison, WI) for WT CFTR or in pBluescript (Stratagene, La Jolla, CA) for V317E and V317Q CFTR. For most experiments, oocytes were injected with ∼5 ng of CFTR cRNA plus 0.6 ng of cRNA for the human β2-adrenergic receptor, which allows activation of PKA-regulated currents by the addition of isoproterenol (Iso) to the bath. Recordings were made 42–96 h after cRNA injection.
Site-directed mutations were engineered in one of two ways:1) with a pSelect vector containing a 1.7-kb fragment of the amino end of CFTR (SacI-SphI) and the Altered Sites protocol (Promega) or 2) with a nested PCR strategy in which the mutation was designed into antiparallel oligomers that were first-round amplified with flanking oligomers (at AflII andSphI, respectively) and second-round amplified with the flanking oligomers alone. In both cases, theAflII-SphI fragment was subcloned back into a pBluescript vector (kindly provided by M. Drumm) containing the human WT CFTR from which the native AflII-SphI fragment had been separated. All mutations were confirmed by direct sequencing of the entire fragment before being subcloned.
Standard two-electrode voltage-clamp (TEVC) techniques were used to study whole cell currents. Borosilicate glass electrodes (Sutter Instrument, Novato, CA) were pulled in four stages and filled with 3 M KCl. Pipette resistances measured 0.4–0.9 MΩ in the bath. Macroscopic data were acquired at room temperature (∼22°C) with a GeneClamp 500 amplifier (Axon Instruments, Foster City, CA) and pCLAMP software (Axon); macroscopic currents were filtered at 500 Hz and acquired by the computer at 2 kHz. Bath solution (ND96) for most TEVC experiments was nominally calcium free and contained (in mM) 96 NaCl, 2 KCl, 1 MgCl2, and 5 HEPES. The pH was adjusted to 7.5 with NaOH. The oocytes were activated by superfusion of ND96 containing Iso at a final concentration of 0.1–5.0 μM. To limit the activation of endogenous calcium-activated Cl− channels, 1 mM BaCl2 was added to all solutions used in the TEVC experiments. For characterization of permeation and rectification properties, macroscopic currents were elicited by stepping, for 75 ms, from a holding potential of −30 mV to a series of test potentials between −140 and +80 mV in +20-mV increments. Current-voltage plots were constructed from initial data at each potential (average of the first 5 ms) and from data at the end of the 75-ms pulse (average of the final 10 ms) (31). Reversal potentials (E R) were determined by linear regression around the zero-current value. The macroscopic rectification ratio (G +80/G −80) was determined by estimating the currents at E R +80 mV and E R −80 mV, using current-voltage relationships provided by pCLAMP.
All single-channel studies were performed on excised inside-out patches. Oocytes were shrunk in a hypertonic solution of (in mM) 200 monopotassium aspartate, 20 KCl, 1 MgCl2, 10 EGTA, and 10 HEPES-KOH, pH 7.2, and the vitelline membrane was then removed with fine forceps. Pipettes were pulled in four stages from borosilicate glass (Sutter) and had an average resistance of ∼8 MΩ when filled with the standard pipette solution containing (in mM) 150N-methyl-d-glucamine (NMDG)-Cl, 5 MgCl2, and 10 TES, adjusted with Tris to pH 7.4. The low-Cl− pipette solution contained (in mM) 5 NMDG-Cl, 5 MgCl2, 135 mM NMDG-aspartate, and 10 TES, adjusted with Tris to pH 7.4. With the low-Cl− pipette solution, pipette resistances were ∼30 MΩ. For these experiments, data were corrected for junction potentials of ∼9.8 mV, as determined with a free-flowing KCl electrode (n = 5 patches). Seal resistances were in the range of 200 GΩ. In most experiments, the presence of CFTR in the patch was tested by bath application of 1–10 μM Iso before excision. After excision, the patches were perfused with intracellular solution (in mM: 150 NMDG-Cl, 1.1 MgCl2, 2 Tris-EGTA, 1 MgATP, and 10 TES, pH adjusted to 7.4 with Tris). PKA (50 U/ml; Promega) was used to maintain the activity of CFTR. Channel currents were recorded at room temperature (∼22°C) with an AI2120 amplifier (Axon) and were recorded at 10 kHz to DAT tape (model DTC-790; Sony). Data were subsequently played back and filtered with a four-pole Bessel filter (Warner Instruments, Hamden, CT) at 100 Hz and acquired by the computer at 2 ms/point with the Fetchex program of pCLAMP (Axon). Some data were more heavily filtered for display (25 Hz) with the built-in software filter of Igor Pro (version 3.14; Wavemetrics, Lake Oswego, OR).
Kinetic analysis of gating.
Dwell-time analysis of open bursts was performed with Igor software. Open bursts were defined as intervals separated by closings of ≥80 ms, a value previously used in analysis of single-channel data in oocytes (32) and established to discriminate between ATP-dependent gating and intraburst blockade of CFTR in mammalian cells (29). Some of the recordings used in the analysis of burst duration contained more than one simultaneous open level (up to three). For the multiple-channel recordings, the mean burst duration was estimated with the formula tn = Σj · tj /n, wheretj is the time at which the jchannels were open at the same time and n is the total number of transitions from an open burst to an interburst closed state. Thus these multiple-channel open-burst events are transformed ton single-channel open events with burst duration tfor each event. This method was first described by Fenwick and colleagues (7) and has been used in similar types of analysis of muscarinic K+ channels (11) and, more recently, CFTR (26, 32). Given the high degree of filtering used for burst duration analysis, burst duration is equivalent to open time in these experiments.
The mean time-averaged currents (I) and their variance (ς2) were calculated from traces at least 15 s in duration. The P o and the number of active channels in the patch (N) were calculated from variance data with the equations ς2 = I ×i(1 − P o) and N= I/(i × P o), wherei is the unitary current. In our patches, we could directly measure i and I; thus these equations yield values for N and P o that are independent of the number of open and closed states.
All reported values are presented as means ± SE unless otherwise noted. Statistical analysis was performed with the t-test for paired or unpaired measurements (SigmaStat; Jandel Scientific, San Rafael, CA) as appropriate for each set of experiments, withP ≤ 0.05 considered indicative of significance.
Whole cell currents exhibit strong rectification.
Exposure of oocytes expressing CFTR and β2-adrenergic receptors to Iso resulted in the development of Cl−current that peaked within 6–7 min. On reaching peak stimulation, oocytes were voltage clamped to a series of potentials (from −140 to +80 mV) to study properties of rectification and the time dependence of macroscopic currents. Figure 1 shows families of whole cell currents from TEVC experiments in WT and V317E CFTR (Fig. 1, A and B, respectively). In this series of experiments, whole cell current-voltage (I-V) relationships were determined by monitoring the current during a 75-ms voltage pulse. Consistent with previous results under identical conditions (32), WT CFTR exhibited time-independent currents over most of the voltage range, aside from a brief relaxation visible at the most hyperpolarizing potentials.I-V relationships for WT CFTR currents (Fig. 1 C) were linear for outward currents but showed significant rectification at strongly hyperpolarizing potentials. As in previous studies, the degree of rectification of WT CFTR macroscopic currents varied between oocytes. Both the rectification and brief relaxation of inward currents may have been due to blockage of the pore by constituents of the oocyte cytoplasm (19, 32). In contrast, V317E CFTR currents exhibited time-dependent behavior at all voltages (Fig. 1 B); inward currents relaxed toward zero, while outward currents increased in conductance through time. This behavior is evident in Fig.1 D, which shows the discrepancy between initial currents (average of the first 5 ms after the voltage jump) and currents recorded after 75 ms (late currents, average of the last 10 ms of the voltage jump). The initial I-V relationship exhibited linear currents at depolarizing potentials. In contrast, the I-Vrelationship constructed from currents at the end of the 75-ms voltage pulse exhibited pronounced outward rectification across the entire voltage range. The outward current amplitude did not approach a constant value during the short voltage pulse, indicating that the underlying time- and voltage-dependent process may be extremely slow. V317E CFTR also exhibited modest tail currents (Fig. 1 B) not seen in WT CFTR with this voltage protocol.
E R in ND96 bath solution did not differ between WT CFTR (−28.1 ± 0.59 mV; n = 12) and V317E CFTR (−27.1 ± 0.58 mV; n = 10). Despite the strong change in rectification during the voltage pulse (Fig. 1 D), the E R in V317E CFTR was not time dependent. Hence, selectivity for Cl− over Na+ did not appear to differ between WT and the mutant and did not change during the time-dependent behavior of V317E CFTR.
To quantify the change in rectification through time, we calculatedG +80/G −80(16) using data from experiments such as those shown in Fig. 1, A and B, as the whole cell chord conductance (G) at E R +80 mV divided by G at E R −80 mV.G +80/G −80 was 1.67 ± 0.13 for WT CFTR initial and late currents and was 1.52 ± 0.08 for V317E CFTR initial currents (P = 0.004 compared with WT CFTR), and 2.24 ± 0.20 for V317E CFTR late currents (P < 0.001 compared with WT CFTR). The smaller degree of rectification of initial currents in V317E CFTR compared with WT CFTR may reflect reduced susceptibility of V317E CFTR currents to blockage by the cytoplasmic constituent(s) or may reflect an impact of holding potential on the process underlying the rectification (seeMechanism of the voltage dependence of V317E CFTR). Hence, the mutant starts out exhibiting less outward rectification than the WT but undergoes a process or processes that both reduces inward current and increases outward current during the 75-ms pulse, leading to pronounced outward rectification in the macroscopic I-Vrelationship. The kinetics of macroscopic currents were not sensitive to a 10-fold increase in bath magnesium concentration (data not shown).
Characterization of unitary currents.
To identify the mechanism responsible for the pronounced outward rectification of V317E CFTR whole cell currents, single-channel studies were performed. Unitary currents (i) were determined from all-points histograms (data not shown) of multiple closed-to-open transitions at a particular potential. In agreement with previous studies that utilized excised inside-out patches (2, 4), we found the i-V relationship of WT CFTR to be linear, yielding a single-channel conductance (g) of 7.26 ± 0.09 pS (n = 13 patches) when measured over the range of membrane potential (V m) of −100 to +100 mV in symmetrical Cl− (Fig. 2,A and B). Single-channel current (i) amplitude for WT CFTR was 0.73 ± 0.01 pA at +100 mV and 0.75 ± 0.01 pA at −100 mV (P > 0.7). In contrast, thei-V relationship for V317E CFTR was found to exhibit slight voltage dependence, with g = 7.65 ± 0.14 pS at negative potentials and g = 6.34 ± 0.14 pS at positive potentials (n = 7 patches; Fig. 2,C and D). i Amplitude for V317E CFTR was 0.64 ± 0.02 pA at +100 mV and 0.77 ± 0.03 pA at −100 mV (P < 0.02). Thus i at positive potentials for V317E CFTR exhibited only 87% of the conductance of WT CFTR. Most importantly, rectification of i was directed inward, opposite to that of the whole cell currents.
To clarify the mechanism leading to reduced g, we studied V317Q. Glutamine and glutamic acid have similar-sized side chains but differ in that glutamine is not charged. The i-Vrelationship of V317Q CFTR was found to have slight voltage dependence, with g = 7.28 ± 0.16 pS at negative potentials and g = 5.91 ± 0.06 pS at positive potentials (n = 5 patches; Fig. 2, E and F).i for V317Q CFTR was 0.59 ± 0.01 pA at +100 mV and 0.73 ± 0.01 pA at −100 mV (P < 0.001). Hence, introduction of a bulky side chain at this position, regardless of charge, appears to result in a slight but significant reduction ofi at positive potentials. Furthermore, the lack of effect of charge at position 317 on the translocation rate of Cl−through the permeation pathway suggests that residue 317 does not directly interact with permeating anions.
The rectification of V317E unitary currents was slight and in the opposite direction of the rectification observed in whole cell experiments. Thus rectification of V317E whole cell currents cannot be explained by voltage-dependent alterations in single-channel conductance.
Characterization of rundown.
To compare the time-averaged, steady-state currents in patches containing either WT or V317E CFTR at multiple holding potentials, it was necessary to control for differences in channel number (N) between patches. We constructed I-Vrelationships by normalizing currents at various potentials to currents at a single potential. This required accounting for time-dependent rundown of channel activity to ensure that all differences inI between WT and the mutant CFTR were due to voltage-dependent, not time-dependent, phenomena.
After excision of a patch from an oocyte expressing WT CFTR, the addition of ATP in the absence of PKA resulted in minimal channel activity, suggesting low phosphorylation (data not shown). Application of PKA (50 U/ml) and ATP (1 mM) resulted in an increase in Ithat reached a steady state of 15.20 ± 2.17 pA in the patch shown in Fig. 3 A. With variance analysis, the P o and N values were estimated to be 0.59 and 34, respectively, for the record shown. Washout of ATP resulted in a dramatic reduction in channel activity that was apparently due to a marked reduction in the opening rate (Fig.3 B). On readdition of 1 mM ATP in the continuing presence of PKA, the steady-state I was 7.59 ± 2.01 pA (Fig.3 C), indicating rundown approaching 50%. Variance analysis of this current yielded P o and Nvalues of 0.29 and 35, consistent with the decrease in current being due to a decline in P o rather than a change inN. The effect of a second washout of ATP was very similar to that of the first (Fig. 3 D). A third application of 1 mM ATP (with PKA) resulted in a steady-state I of 7.72 ± 2.04 pA (Fig. 3 E). Variance analysis of this current yielded aP o of 0.30 and an N of 34. These results suggest that in our system and for a given patch 1) the initial decline in CFTR steady-state activity observed in the presence of 1 mM ATP (Fig. 3, A vs. C) was due to a decline in the P o and not in N, and2) after PKA application, phosphorylation levels were fairly constant within a given patch over several minutes after the initial rundown of activity. To ensure consistent phosphorylation levels in the experiments described in Steady-state I-V relationships, 50 U/ml of PKA were present in all excised-patch experiments. Furthermore, all recordings were performed after the initial rundown in activity. This is important because alteration in phosphorylation levels can affect CFTR gating (5, 12, 17, 23, 27). To account for differences in phosphorylation levels between patches, data at multiple potentials in each patch were referenced to data at −100 mV in the same patch.
Steady-state I-V relationships.
Computed steady-state time-averaged current (I) values were compared in excised inside-out patches containing multiple WT or mutant channels. WT CFTR exhibited a linear I-V relationship, with a E R of ∼0 mV in the presence of symmetrical Cl− (Fig. 4 A). Observation of WT CFTR channel activity at both positive and negative potentials revealed currents of similar magnitude in a given patch (Fig. 4 B). However, this was not the case for V317E CFTR. The I-V relationship for this mutant was markedly rectified in the outward direction (Fig. 4 C) due to the reducedP o at negative membrane potentials (Fig.4 D). Outward currents exhibited a linear I-Vrelationship. Hence the I-V relationship from single channels recapitulated the rectification seen in the true macroscopicI-V relationship from whole cell recording (Fig.1 D).
The behavior of V317Q CFTR channels was also examined at multiple potentials. The single-channel activity of this mutant was found to be voltage independent (Fig. 4 F), the I-Vrelationship being similar to that of WT CFTR (Fig. 4 E). Thus the voltage dependence of gating seems to require introduction of a negative charge.
Mechanism of the voltage dependence of V317E CFTR.
Comparison of V317E and WT CFTR revealed markedly different single-channel kinetics (Fig.5 A). Despite the fact that WT CFTR macroscopic currents showed no voltage dependence, we found that WT CFTR single channels exhibited significantly longer burst durations at negative membrane potentials than at positive membrane potentials (P < 0.05). We suspect that this prolongation of the open-burst duration at negative potentials was due to the voltage-dependent intraburst flicker (not evident in Figs. 2, 4, and 5because of filtering) that has been previously reported (8, 9,17, 19, 28) and likely reflects blockage of the pore by the pH buffer TES in our intracellular solutions (13). If this intraburst flickery behavior is truly indicative of a blocking substrate occluding the pore, then prolongation of the burst duration is to be expected according to the concepts of classic open-channel block (see Ref. 6). However, because the individual dwell time for each blocking event is quite brief (typically <0.25 ms at −100 mV) (19), no significant difference was observed in the magnitude of time-averaged currents at positive and negative potentials (Fig. 4 A). This result is not considered further here. For V317E CFTR, burst durations were significantly reduced compared with WT CFTR over the entire voltage range (Fig.5 B). However, the difference between V317E and WT CFTR burst durations cannot explain the voltage dependence ofP o observed in the mutant, because the burst durations of V317E CFTR at negative potentials were not significantly shorter than the burst durations at positive potentials.
Closed times for either WT or V317E CFTR were fairly consistent in any single experiment, but they varied from patch to patch, probably as a result of variable phosphorylation levels. However, comparison of mean closed times in four different patches containing V317E CFTR revealed significant voltage dependence in the closed duration, with closed times being 4.50 ± 0.93-fold longer (P < 0.05) at V m = −100 mV compared with those atV m = +100 mV in the same patch (n = 4). In WT CFTR, no significant voltage dependence was detected in the closed durations in the same patch, with the closed times at V m = −100 mV being 0.94 ± 0.09-fold of those at V m = +100 mV (n = 5 patches). These data are summarized in Fig. 5,C and D. This result is consistent with the reduced opening rate at V m = −100 mV that is apparent in Fig. 4 D.
One potential explanation for a prolonged closed time is a reduced rate of ATP binding to NBD1 at negative potentials. To test this possibility, we observed the voltage dependence of V317E CFTR activity at 1 and 10 mM ATP. Raising the ATP concentration 10-fold had no effect on i (data not shown) and did not modify the voltage dependence of channel activity observed with 1 mM ATP in the same patch (Fig. 6 A). Furthermore, the multiple of increase in activity that was observed when transitioning from V m = −100 mV toV m = +100 mV was not significantly different (P > 0.5) in paired experiments (2.76 ± 0.37- and 3.04 ± 0.80-fold for 1 and 10 mM ATP, respectively;n = 4 patches; Fig. 6 B). Thus the effect of mutation V317E on channel gating does not appear to change the ATP dependence at concentrations in excess of the reported dissociation constant (29).
Characterization of i and I in asymmetrical Cl−.
To determine whether the voltage dependence ofP o in V317E CFTR was influenced by an anion gradient, we repeated our experiments using a low-Cl− (15 mM) solution in the pipette. Comparing WT CFTR channel behavior at +100 and −100 mV (Fig. 7 A,top and bottom, respectively), one can see a reduced i at +100 mV, due to the decreased driving force for inward Cl− movement, but no drastic difference in the overall P o. Examination of the i-Vfor WT CFTR revealed a 50-mV rightward shift in theE R (59.8 mV after correction for junction potential) in the presence of the low-Cl− pipette solution (Fig. 7 B). As expected for recordings with low Cl− concentration ([Cl−]) in the pipette, the time-averaged I-V for WT CFTR in our asymmetrical Cl− solutions exhibited slight inward rectification.
Comparing V317E CFTR channel behavior at +100 and −100 mV (Fig.7 D, top and bottom, respectively), one can see a reduced i at +100 mV, much like that seen in WT CFTR. However, a large reduction in P o was observed at −100 mV compared with +100 mV. The i-Vrelationship for V317E CFTR was shifted to the right by 49 mV (58.8 mV after corrections) in the presence of our low-Cl−pipette solution (Fig. 7 E). The i-V of V317E CFTR in these conditions showed modestly increased inward rectification compared with that of the WT channel due to the reduced single-channel amplitude at positive potentials (Fig. 2 D), V317E CFTRi amplitude at the most depolarized voltage being only 85% of that seen in WT CFTR. This is very similar to the reduction in outward currents seen in the presence of symmetrical [Cl−] (Fig. 2).
The time-averaged I-V relationship for V317E CFTR in the asymmetrical Cl− solutions showed a pronounced outward rectification (Fig. 7 F), whereas that for WT CFTR remained nearly linear (Fig. 7 C). Neither Goldman-type rectification nor inward rectification of i conferred by the amino acid substitution can account for the rectification observed inI. Hence, as the inhibition of channel opening exhibited at −100 mV was overcome by depolarization, Iincreased. Therefore, similar to the V317E CFTR behavior characterized in symmetrical Cl−, we conclude that outward rectification of I in the presence of a Cl− gradient is likely due to a voltage dependence of P o.
NPo vs. voltage relationship.
We determined the product of N and P o(NP o) in all of our patches, and because bothN and P o varied greatly from patch to patch (but were consistent within a patch after allowing for rundown), we normalized these values to the NP o at −100 mV in each patch. In symmetrical Cl− there is an obvious voltage dependence in NP o for V317E CFTR (Fig.8 A) that was most pronounced at V m values more positive than 0 mV, where net Cl− movement is inward. To determine whether this voltage-dependent increase in NP o could be due to some form of open-channel block at negativeV m, we also performed this analysis on recordings obtained with low-[Cl−] pipette solutions. By shifting the E R for Cl− flow ∼50 mV to the right, we dissociated V m from the direction of anion flow between 0 and approximately +50 mV. With symmetrical Cl− (Fig. 8 A), currents measured between V m = 0 and +50 mV reflected anion entry from the pipette. With reduced bath Cl− (Fig.8 B), currents in the same voltage range reflected anion exit into the pipette. As shown in Fig. 8 B, the increase inNP o was still observed at positive potentials (V m = +10–40 mV), where net Cl− movement is in the same direction (outward) as Cl− movement at −100 mV. This insensitivity (to the direction of ion movement) of the voltage dependence ofNP o for V317E CFTR argues strongly against voltage-dependent blockade of the pore.
Voltage-jump relaxations indicate behavior similar to voltage-gated channels.
The single-channel experiments indicated that the rectification of macroscopic V317E CFTR currents was due to voltage-dependent kinetics of opening and closing. However, the change in macroscopic rectification and the slow relaxations during a voltage step, as shown in Fig. 1, indicate that the voltage-dependent alteration ofP o is a time-dependent process as it is in voltage-gated channels of excitable cells. Because our single-channel recordings were made several seconds after the membrane voltage was changed, they do not reflect the time dependence of this gating process. To determine whether the process evident in V317E CFTR was similar to voltage-dependent gating mechanisms observed in other channels, we asked whether the kinetics of the relaxations at hyperpolarizing potentials were sensitive to V m. Oocytes were held at −30 mV, pulsed to +50 mV for 100 ms, and then stepped for 200 ms to potentials ranging from −120 to −60 mV (data not shown). Relaxations of inward currents were fitted with a second-order exponential in V317E CFTR, whereas only a first-order exponential was required for WT CFTR currents. Time constants for these relaxations in V317E CFTR were strongly dependent on voltage (values for the longer component, τ2, were 157.0 ± 10.7 ms at −120 mV and 294.2 ± 18.2 ms at −60 mV; n = 5 patches; P < 0.001 by paired t-test). The kinetic behavior of V317Q-CFTR macroscopic currents did not differ from WT CFTR (data not shown).
The traces in Fig. 1 show a slow process at depolarizing potentials; as the membrane was held at depolarizing potentials, outward current increased as if the channels were being trapped in a new state with a high opening rate. In contrast, relaxations at hyperpolarizing potentials are consistent with the channels being trapped in a state with a low opening rate. If this interpretation is correct, the kinetics of relaxations to the low opening-rate state (at hyperpolarizing potentials) may depend on the extent to which the pool of channels has previously been trapped in the high-opening-rate state. To determine whether such a state dependency exists, we examined the kinetics of macroscopic currents at −120 mV after depolarizing prepulses that varied in duration from 20 to 250 ms (Fig.9). Oocytes expressing V317E CFTR exhibited an increased conductance at +50 mV as the prepulse duration was lengthened (Fig. 9 A). Concurrently, the peak conductance at −120 mV also increased, as did the magnitude of the relaxation (ranging from ∼25 to ∼40% of the inward current at this potential). This behavior was not evident in WT CFTR studied under identical conditions (Fig. 9 C). The slower component of the relaxation (τ2) in V317E CFTR currents contributed more to the fit as the prepulse duration was lengthened from 20 to 250 ms. Finally, values of τ2 exhibited a clear dependence on prepulse duration (Fig. 9 D). τ2 was significantly increased with prepulse durations >50 ms; this relationship became saturated with prepulses of 200 ms duration. Hence, the process at hyperpolarizing potentials is kinetically linked to the very slow process evident at depolarizing potentials. These results are reminiscent of the behavior of voltage-gated channels of excitable cells (see Ref. 10).
Channel gating in WT CFTR is traditionally thought to be voltage independent. The results presented here demonstrate that it is possible to induce significant voltage dependence in the gating mechanism of the Cl− channel by introducing negative charge at one position in TM5. The voltage-dependent gating of V317E CFTR was evident both in relaxations of macroscopic currents and in the altered gating of single channels. Thus voltage-dependent gating is the mechanism underlying the strong outward rectification of macroscopic currents carried by this channel.
The induction of voltage-dependent gating by substituting a charged residue, but not by substituting a neutral residue, suggests that this site resides in a portion of the protein that is conformationally mobile such that it can move in response to a change in the membrane electric field. The data presented are not sufficient to reveal whether this apparent movement is related to the normal gating cycle of the protein or whether, instead, it reflects a novel motion that would only be seen in this altered protein molecule. In either case, however, the results suggest that it may be possible to use the introduction of voltage-dependent gating as a tool to probe the mobility of segments in the protein and, perhaps, to identify those parts that move in response to events initiated by the hydrolysis of ATP. In the latter case, one might expect a strong interaction between ATP-dependent and voltage-dependent gating events such that, for example, the sensitivity of the channel to activating conditions (PKA + ATP) would become voltage dependent.
Other glutamate substitutions in TM domains 1, 5, 6, and 12 have been investigated [G91E, G314E, and K335E (16); S341E and T1134E (20)], but none of these exhibited voltage-dependent gating (McCarty and Dawson, unpublished observations). Although these mutants have not yet been studied at the single-channel level, no voltage-jump relaxations indicative of voltage-dependent gating were apparent in macroscopic recordings. Hence this phenomenon is thus far specific to the introduction of glutamate at position 317 in TM5. Introduction of a similarly sized but uncharged amino acid side chain at this position (V317Q CFTR) did not affect gating, indicating that it is the charge, not the bulkiness of the side chain, that confers the voltage-dependent behavior.
The voltage dependence of the whole cell I-V relationship is almost purely a gating effect, aside from a minor change in single-channel conductance at positive potentials. The small change in single-channel conductance was found for both V317E and V317Q-CFTR, indicating that it does not result from electrostatic repulsion due to introduction of the negative charge. Hence, this residue apparently has little or no impact on ion permeation, indicating that it does not lie in the permeation pathway. Rather, introduction of a bulky amino acid (glutamic acid or glutamine) at this position may induce rectification in single-channel conductance by altering protein conformation.
The voltage dependence induced by the V317E mutation impacts the CFTR gating scheme at a point distal to the binding and hydrolysis of ATP at the NBDs, as evidenced by the insensitivity of the opening rate to increased [ATP]. Alternatively, it is possible that the mutation resulted in a conformational change that altered exposure of a phosphorylation site in a voltage-dependent manner. However, the kinetic differences were observed immediately upon switching between positive and negative potentials and were observed to be immediately reversible, making it unlikely that rapid changes in phosphorylation had occurred. It also seems unlikely that a mutation in TM5 would alter the structure of the regulatory domain, which includes most of the sites of PKA-dependent phosphorylation in the CFTR. Hence, it appears that the conformational change that is induced by ATP binding and hydrolysis at the NBDs is inhibited at negativeV m in this mutant. It has recently been suggested that the pore of CFTR is not conformationally static in that permeation properties depend on gating status (13, 15, 18,31). Recent work by Kogan and colleagues (14) shows that inhibitors of permeation, such as diphenylamine-2-carboxylate, affect the rate of ATP hydrolysis, suggesting that bidirectional communication between the pore domain and cytosolic gating domains occurs. It is thus possible that the pore of the WT channel undergoes a conformational change at or in the vicinity of position 317, and we have introduced voltage dependence to this process by introduction of a charged amino acid at position 317. To date, the region that experiences this conformational change (i.e., the gate) has not been identified in CFTR. The ability to confer voltage dependence to the gating mechanism, as seen with the V317E mutation, may provide a means to localize the structures that comprise the gate.
We thank Dr. T. C. Hwang and Dr. C. Hartzell for insightful and helpful comments.
↵* Z.-R. Zhang and S. Zeltwanger contributed equally to this work.
This work was supported by American Heart Association Grant 9820032SE and by National Science Foundation Grant MCB-0077575. S. Zeltwanger was supported by National Institute of Diabetes and Digestive and Kidney Diseases 5T32-DK-07656 and by Cystic Fibrosis Foundation Fellowship ZELTWA00F0.
Address for reprint requests and other correspondence: N. A. McCarty, School of Biology, Georgia Inst. of Technology, 310 Ferst Dr., Atlanta, GA 30332 (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2002 the American Physiological Society