ABCMdb

PMID: 11882668

Powe AC Jr, Al-Nakkash L, Li M, Hwang TC
Mutation of Walker-A lysine 464 in cystic fibrosis transmembrane conductance regulator reveals functional interaction between its nucleotide-binding domains.
J Physiol. 2002 Mar 1;539(Pt 2):333-46., 2002-03-01 [PubMed]

Sentences

No. Mutations Sentence Comment

12 ABCC7 p.Lys464Ala To address this issue, we examined channels with a Walker-A lysine mutation in NBD1 (K464A) using the patch clamp technique. Login to comment 13 ABCC7 p.Lys464Ala K464A mutants have an ATP dependence (EC50 ∆ 60 m) and opening rate at 2.75 m ATP (~2.1 s_1 ) similar to wild type (EC50 ∆ 97 m; ~2.0 s_1 ). Login to comment 15 ABCC7 p.Lys464Ala Delay of closing in wild type by adenylyl imidodiphosphate (AMP-PNP), a non-hydrolysable ATP analogue, is markedly diminished in K464A mutants due to reduction in AMP-PNP`s apparent on-rate and acceleration of its apparent off-rate (~2and ~10-fold, respectively). Login to comment 16 ABCC7 p.Lys1250Ala Since the delay of closing by AMP-PNP is thought to occur via NBD2, K464A`s effect on the NBD2 mutant K1250A was examined. Login to comment 17 ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala In sharp contrast to K464A, K1250A single mutants exhibit reduced opening (~0.055 s_1 ) and closing (~0.006 s_1 ) rates at millimolar [ATP], suggesting a role for K1250 in both opening and closing. Login to comment 18 ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala At millimolar [ATP], K464A-K1250A double mutants close ~5-fold faster (~0.029s_1 )thanK1250Abutopenwithasimilarrate(~0.059s_1 ),indicatinganeffectofK464Aon NBD2 function. Login to comment 22 ABCC7 p.Lys1250Ala The idea that this second functional site corresponds to NBD2 became prominent because millimolar [ATP] can prolong the opening of a Walker-A lysine mutant at NBD2 (i.e. K1250A), mimicking AMP-PNP`s effect on wild type CFTR (Carson et al. 1995; Gunderson & Kopito, 1995; Zeltwanger et al. 1999). Login to comment 24 ABCC7 p.Lys464Ala Supporting this idea, some investigators reported that a mutation in NBD1`s Walker-A motif (i.e. K464A) reduces CFTR`s opening rate (Carson et al. 1995; Gunderson & Kopito, 1995; Vergani et al. 2000). Login to comment 25 ABCC7 p.Lys464Ala However, those earlier reports are being challenged by more recent demonstrations that K464A has little effect on channel opening (Sugita et al. 1998; Ramjeesingh et al. 1999; present study). Login to comment 29 ABCC7 p.Lys464Ala To examine the contribution of NBD1 during CFTR gating, we assessed the kinetic properties of the Walker-A mutant K464A. Login to comment 31 ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala The plasmids K464A-pRBG4 and K1250A-pRBG4 were gifts from Dr R. R. Kopito (Stanford University, CA, USA), and the plasmid CFTRwt-pBQ4.7 and the retroviral vector pLJ were gifts from Dr M. Drumm (Case Western Reserve University, Cleveland, OH, USA). Login to comment 32 ABCC7 p.Lys1250Ala ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala ABCC7 p.Lys464Ala To construct K464A-pBQ and K1250A-pBQ, the 0.7 kb BspEI-BstZ171 fragment from K464A-pRBG4 and the 3.0 kb BspEI-NcoI fragment from K1250A-pRBG4, respectively, replaced the corresponding ones in CFTRwt-pBQ4.7. Login to comment 33 ABCC7 p.Lys1250Ala ABCC7 p.Lys1250Ala To create WT-pLJ and K1250A-pLJ, the 4.7 kb EcoICRI fragments from CFTRwt-pBQ4.7 and K1250A-pBQ, respectively, were ligated to Bcl I linkers, cut with BclI, then ligated into the BamHI site of the pLJ vector. Login to comment 34 ABCC7 p.Lys464Ala ABCC7 p.Lys464Ala For construction of K464A-pLJ, the 3.6 kb BspEI-SalI fragment from K464A-pBQ was exchanged with the corresponding one in WT-pLJ. Login to comment 35 ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala To generate WT-pCDNA, K464A-pCDNA and K1250A-pCDNA,the4.7 kbPstIfragmentsfromthecorresponding pBQ constructs were subcloned into the PstI site of pCDNA. Login to comment 36 ABCC7 p.Lys1250Ala ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala ABCC7 p.Lys464Ala For the creation of K464A-K1250A-pCDNA, the 2.7 kb BspEI-PflMI fragment from K464A-pCDNA was used to substitute the corresponding region in K1250A-pCDNA. Login to comment 40 ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala To obtain recordings with few channels per patch, we used NIH3T3 cells stably transfected with wild type CFTR (Berger et al. 1991), CFTR-K1250A and CFTR-K464A (Zeltwanger et al. 1999; present study). Login to comment 54 ABCC7 p.Lys464Ala Macroscopicdose-responseanalysisofK464Achannels To quantify the ATP dependence of K464A channels, we normalized baseline-subtracted mean current levels elicited by test [ATP] against mean current levels at 2.75 m ATP as previously described (Zeltwanger et al. 1999; see Fig. 1). Login to comment 60 ABCC7 p.Lys464Ala An analysis of flickery events in K464A channels using 100 Hz filtering and 500 Hz sampling justified the 50 ms cut-off (data not shown). Login to comment 68 ABCC7 p.Lys464Ala However, since the solution exchange in our system is relatively slow (~5 s; cf. Zeltwanger et al. 1999), relaxation times for K464A, but not wild type, are probably rate limited more by nucleotide removal rather than actual channel closing (see Results). Login to comment 69 ABCC7 p.Lys464Ala To estimate more accurately the AMP-PNP-dependent open time for K464A channels, we analysed recordings with only one or two channels; open events where two channels were open simultaneously were averaged and included as described previously (Fenwicketal.1982;Wangetal.1998).Dwelltimeswererankedin order of decreasing duration, normalized by the number of events and displayed as survivor plots (i.e. the probability of still being open at time t given that the channel was open at time zero versus time). Login to comment 72 ABCC7 p.Lys464Ala ABCC7 p.Lys464Ala Based on our open time analysis of K464A channels in the presence of 250 µ ATP and 1 m AMP-PNP (see above), we used a cut-off of 842 ms for both wild type and K464A channels. Login to comment 73 ABCC7 p.Lys464Ala Use of this cut-off for both types of channels is justifiable, since both K464A and wild type have similar open times (i.e. dwell times in O) at 250 µ ATP alone. Login to comment 75 ABCC7 p.Lys1250Ala ABCC7 p.Lys1250Ala ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala ABCC7 p.Lys464Ala EstimationofkineticparametersforK1250Aand K464A-K1250A For recordings of quasi-macroscopic K1250A and K464A-K1250A channel currents, open probability was estimated by means of variance analysis (Sigworth, 1980): Po = (1 _ (s2 /Ii)), where Po represents open probability, s2 the variance of steady-state current, I the mean steady-state current and i the single channel amplitude. Login to comment 80 ABCC7 p.Lys1250Ala ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala ABCC7 p.Lys464Ala RESULTS As a step towards understanding how CFTR`s NBDs participate in gating, we examined the kinetic behaviour of the NBD1 mutant K464A, the NBD2 mutant K1250A and double mutant K464A-K1250A. Login to comment 81 ABCC7 p.Lys464Ala K464A mutants were examined first. Login to comment 82 ABCC7 p.Lys464Ala To determine whether K464A affected CFTR`s ATP dependence, we performed a dose-response analysis. Login to comment 83 ABCC7 p.Lys464Ala Some earlier reports indicated that K464A reduced ATP affinity (Anderson & Welsh, 1992; Vergani et al. 2000). Login to comment 89 ABCC7 p.Lys464Ala From the fit, the Km for K464A channels was 59 ± 9 µ. Login to comment 91 ABCC7 p.Lys464Ala This comparison indicates that the K464A mutation had little effect on ATP sensitivity. Login to comment 92 ABCC7 p.Lys464Ala To estimate the ATP dependence of open probability in K464A mutants, the Po was measured in patches with one to four channels. Login to comment 97 ABCC7 p.Lys464Ala The fit results closely matched the measured data; the measured mean Po for K464A channels was 0.37 ± 0.02 (n = 18) at 2.75 m ATP, 0.21 ± 0.05 (n = 5) at 100 µ ATP, and 0.19 ± 0.02 (n = 3) at 50 µ ATP, indicating that the EC50 for ATP lies between 50 and 100 µ. Login to comment 100 ABCC7 p.Lys464Ala Furthermore, a comparison of K464A to wild type (Zeltwanger et al. 1999; present study, Fig. 1B) shows little change in ATP sensitivity of Po. Login to comment 101 ABCC7 p.Lys464Ala A Michaelis-Menten fit to our previous wild type data gave a Km of 137 µ with a maximum Po of 0.41, slightly higher than K464A. Login to comment 102 ABCC7 p.Lys464Ala We next tested how [ATP] affected channel opening and closing rates in K464A mutants. Login to comment 105 ABCC7 p.Lys464Ala ABCC7 p.Lys464Ala ATP dose-response relationships for CFTR-K464A A, representative trace of macroscopic CFTR-K464A channel current stimulated by 2.75 m and 25 µ ATP after steady-state activation by PKA phosphorylation (not shown). Login to comment 106 ABCC7 p.Lys464Ala B, trace of CFTR-K464A channels exposed to different [ATP] after steady-state activation by PKA and ATP. Login to comment 107 ABCC7 p.Lys464Ala C, macroscopic dose-response relationship for CFTR-K464A (•; present study) and wild type CFTR (ª; data taken from Zeltwanger et al. 1999). Login to comment 108 ABCC7 p.Lys464Ala D, open probability versus [ATP] for CFTR-K464A (0; present study) and wild type (1; data taken from Zeltwanger et al. 1999). Login to comment 109 ABCC7 p.Lys464Ala The maximum Po for CFTR-K464A at 2.75 m ATP is ~0.37. Login to comment 112 ABCC7 p.Lys464Ala Dashed lines are fits of the Michaelis-Menten equation to the CFTR-K464A data (see Methods). Login to comment 113 ABCC7 p.Lys464Ala shows sweeps from a recording of a single CFTR-K464A channel exposed to 2.75 m and 100 µ ATP. Login to comment 115 ABCC7 p.Lys464Ala In 100 µ ATP the K464A channel exhibits longer closures, while the duration of openings at both concentrations appears similar. Login to comment 116 ABCC7 p.Lys464Ala Figure 2B shows the closed and open time distributions from the K464A recordings in Fig. 2A. Login to comment 121 ABCC7 p.Lys464Ala The relationship between rates and [ATP] for K464A mutants is shown in Fig. 3A; the data shown represent the mean rates pooled from 18 experiments. Login to comment 124 ABCC7 p.Lys464Ala To quantify the ATP sensitivity of opening in K464A mutants, the data were fitted to a Michaelis-Menten equation (Fig. 3A). Login to comment 127 ABCC7 p.Lys464Ala Our findings confirm that ATP-dependent opening is not impaired by the K464A mutation, qualitatively consistent with the data of Sugita et al. (1998) and Ramjeesingh et al. (1999). Login to comment 128 ABCC7 p.Lys464Ala Unlike opening, K464A channel closing exhibits little, if any, dependence on ATP. Login to comment 131 ABCC7 p.Lys464Ala ABCC7 p.Lys464Ala Single-channel kinetics of CFTR-K464A A, representative sweeps from experiments with single CFTR-wild type (left-hand sweeps; taken from Zeltwanger et al. 1999) and CFTR-K464A channels (right-hand sweeps) exposed to 2.75 m (top sweeps) and 100 µ MgATP (bottom sweeps) subsequent to activation by PKA (40 U ml_1 ) and ATP (2.75 m). Login to comment 133 ABCC7 p.Lys464Ala B, survivor plots of closed (left-hand panels) and open dwell times (right-hand panels) at 2.75 m (top panels) and 100 µ MgATP (bottom panels) for the single CFTR-K464A channel shown in A (cf. wild type distributions in Zeltwanger et al. 1999). Login to comment 135 ABCC7 p.Lys464Ala hand, a comparison of closing rates for K464A and wild type at 2.75 m ATP does reveal a difference (3.6 ± 0.3 s_1 , n = 18 versus 2.1 ± 0.3 s_1 , n = 6, respectively; P < 0.005; Fig. 3B). Login to comment 136 ABCC7 p.Lys464Ala Thus, K464A appears to abolish the ATP dependence of the closing rate seen in wild type. Login to comment 137 ABCC7 p.Lys464Ala From this result, we suspected that K464A might affect the second functional site for ATP, which is presumed responsible for prolonging open time in wild type channels at millimolar [ATP] (Zeltwanger et al. 1999). Login to comment 138 ABCC7 p.Lys464Ala Since this second functional site for ATP is thought to be the site for AMP-PNP action (Hwang et al. 1994; Carson et al. 1995; Mathews et al. 1998b; Zeltwanger et al. 1999), we examined whether K464A affected AMP-PNP`s prolongation of channel opening. Login to comment 144 ABCC7 p.Lys464Ala Prolonged openings can also be seen in K464A channels exposed to AMP-PNP, but with lower frequency and shorter duration. Login to comment 145 ABCC7 p.Lys464Ala For example, Fig. 4A (lower trace) shows a single K464A channel activated by PKA (40 U ml_1 ) and ATP (250 µ), then exposed to AMP-PNP (1 m). Login to comment 146 ABCC7 p.Lys464Ala Unlike the wild type example, the K464A channel took much longer to become locked open. Login to comment 148 ABCC7 p.Lys464Ala These two examples suggest that the K464A mutation A. C. Powe, Jr, L. Al-Nakkash, M. Li and T.-C. Hwang338 J. Physiol. 539.2 Figure 3. Login to comment 149 ABCC7 p.Lys464Ala ABCC7 p.Lys464Ala Dependence of CFTR-K464A closing and opening rates on [ATP] A, plot of mean opening rates (reciprocals of mean closed times; 0) and closing rates (reciprocals of mean opened times; ª) versus [ATP] for CFTR-K464A. Login to comment 151 ABCC7 p.Lys464Ala B, comparison of opening and closing rates for CFTR-K464A and wild type at 2.75 m ATP. Login to comment 152 ABCC7 p.Lys464Ala Asterisks indicate a significant difference between closing rates for wild type and K464A (P < 0.005). Login to comment 154 ABCC7 p.Lys464Ala ABCC7 p.Lys464Ala AMP-PNP weakly locks open CFTR-K464A A, representative sweeps from experiments with wild type (top trace) and CFTR-K464A single channels (bottom trace) exposed first to PKA (40 U ml_1 ) and MgATP (250 µ), then with the addition of AMP-PNP (1 m). Login to comment 155 ABCC7 p.Lys464Ala Arrows indicate the baseline, downward deflections channel openings. B, survivor plot of open dwell times for CFTR-K464A. Login to comment 158 ABCC7 p.Lys464Ala reduces the ability to become locked open and to stay locked open by AMP-PNP, further substantiating the idea that K464A affects the second functional site for ATP. Login to comment 160 ABCC7 p.Lys464Ala The opening of K464A channels is slowed by AMP-PNP as well. Login to comment 161 ABCC7 p.Lys464Ala The mean closed time for K464A channels exposed to 250 µ ATP and 1 m AMP-PNP was 1366 ± 260 ms (n = 6), approximately threefold longer than in 250 µ ATP alone (~465 ms; Fig. 3A). Login to comment 163 ABCC7 p.Lys464Ala To quantify the extent to which K464A affects the second functional site, we first estimated the duration of AMP-PNP-dependent locked open events using current relaxation time courses upon removal of AMP-PNP. Login to comment 165 ABCC7 p.Lys464Ala Figure 5A shows a comparison of relaxations from wild type (top trace) and K464A channels (bottom trace). Login to comment 167 ABCC7 p.Lys464Ala ABCC7 p.Lys464Ala Wild type channels close slowly, while K464A channels rapidly shut, indicating a shortened locked open duration in the mutant. On average, wild type channels show a mean relaxation time constant of 105 ± 22 s (n = 5) and K464A channels 12 ± 3 s (n = 5; P < 0.01) (Fig. 5B). Login to comment 168 ABCC7 p.Lys464Ala However, given that the solution exchange in our system is relatively slow (~5 s; cf. Zeltwanger et al. 1999), relaxation times for K464A are probably rate limited more by nucleotide removal than actual channel closing. Login to comment 170 ABCC7 p.Lys464Ala To obtain a more accurate estimate for locked open event duration in K464A channels, we analysed the open time distribution from patches with one or two channels in the presence of PKA, ATP and AMP-PNP (Fig. 4B; see Methods). Login to comment 172 ABCC7 p.Lys464Ala One component was ~300 ms, consistent with the mean open time for K464A channels in 250 µ ATP alone (Fig. 4B; cf. Fig. 3). Login to comment 178 ABCC7 p.Lys464Ala In addition to reducing the time spent locked open, the K464A mutation also impairs the ability to become locked open, as exemplified by Fig. 4A. Login to comment 179 ABCC7 p.Lys464Ala In that example, the K464A channel opens and shuts many times before being locked open, while wild type opens and shuts only a few times before locking open. Login to comment 181 ABCC7 p.Lys464Ala ABCC7 p.Lys464Ala K464A enhances the dissociation of AMP-PNP A, representative traces from experiments with macroscopic currents from CFTR-wild type (top trace) and CFTR-K464A channels (bottom trace) exposed to PKA (40 U ml_1 ), MgATP (250 µ) and AMP-PNP (1 m). Login to comment 183 ABCC7 p.Lys464Ala For the traces shown, the mean relaxation time constant for wild type is 64.8 ± 0.1 s and for CFTR-K464A is 9.1 ± 0.1 s. B, mean relaxation time constants (± ...) Login to comment 184 ABCC7 p.Lys464Ala for wild type and K464A channels. Login to comment 185 ABCC7 p.Lys464Ala Asterisks indicate a significant difference between wild type and K464A (P < 0.01). Login to comment 190 ABCC7 p.Lys464Ala Using the open duration data from K464A channels exposed to AMP-PNP (Fig. 4B), we calculated a cut-off of 842 ms; openings longer than 842 ms were L state and shorter ones O. Login to comment 191 ABCC7 p.Lys464Ala This cut-off was used to classify open events for both wild type and K464A channels. Login to comment 193 ABCC7 p.Lys464Ala Figure 6A shows the distributions obtained from two experiments, one with wild type channels and the other with K464A. Login to comment 194 ABCC7 p.Lys464Ala In the experiments shown, the locking rate for wild type (ª) was 1110 ± 70 s_1 _1 (21 events) and for K464A (1) 370 ± 10 s_1 _1 (49 events), qualitatively consistent with the examples in Fig. 4A. Login to comment 195 ABCC7 p.Lys464Ala On average, wild type channels exhibited a faster locking rate (890 ± 140 s_1 _1 , n = 3) than K464A (560 ± 110 s_1 _1 , n = 6; P < 0.05; Fig. 6B), indicating that the mutation impairs AMP-PNP`s ability to lock open CFTR. Login to comment 197 ABCC7 p.Lys1250Ala Thus, channels bearing K1250A, the lysine-to-alanine mutation in the Walker-A motif of NBD2, exhibit a prolonged open state, similar to the AMP-PNP-dependent locked open state (Carson et al. 1995; Gunderson & Kopito, 1995; Ramjeesingh et al. 1999; Zeltwanger et al. 1999). Login to comment 198 ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala We wondered whether K464A reduces channel open time in K1250A mutants as it does with AMP-PNP. Login to comment 199 ABCC7 p.Lys1250Ala ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala To test this idea, we determined the mean open time for both K1250A single mutant and K464A-K1250A double mutant channels using relaxation time courses upon ATP withdrawal (Fig. 7A). Login to comment 200 ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala In the examples shown, the double mutant relaxes more rapidly than the single mutant. On average, K464A-K1250A channel currents decay fivefold morequicklythanK1250A(34 ± 7 s,n = 5versus167 ± 37 s, n = 6, respectively; Fig. 7C; cf. Zeltwanger et al.1999). Login to comment 202 ABCC7 p.Lys1250Ala ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala We also examined the open probability of K1250A and K464A-K1250A. Login to comment 204 ABCC7 p.Lys464Ala K464A reduces the apparent on-rate of AMP-PNP A, Semilog plot of a probability function determined by measuring the cumulative time channels spent in the O state before entering the L state (see Methods). Login to comment 205 ABCC7 p.Lys464Ala The dwell times are from individual experiments for CFTR-K464A (1) and CFTR-wild type (ª). Login to comment 207 ABCC7 p.Lys464Ala B, comparison of the mean locking rates for CFTR-wild type and CFTR-K464A. Login to comment 208 ABCC7 p.Lys464Ala Asterisk indicates a significant difference between wild type and K464A (P < 0.05). Login to comment 210 ABCC7 p.Lys1250Ala ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala In this example, all four K1250A channels remain open for most of the sweep, whereas only two of the three double mutants channels are open most of the time, suggesting a lower Po for K464A-K1250A. Login to comment 211 ABCC7 p.Lys1250Ala As expected, K1250A channels exhibit a steady-state Po of 0.89 ± 0.02 (n = 6; Fig. 7C) and double mutants 0.67 ± 0.05 (n = 5; P < 0.005). Login to comment 212 ABCC7 p.Lys1250Ala ABCC7 p.Lys1250Ala ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala Our estimates of Po for both mutants are much higher than previously reported (~ 0.2_0.34 for K1250A and ~ 0.25 for K464A-K1250A; Carson et al. 1995; Ramjeesingh et al. 1999; but cf. ~0.9 for K1250A; Gunderson & Kopito, 1995). Login to comment 213 ABCC7 p.Lys1250Ala The differences probably arise from measurements of presteady-state phosphorylated K1250A channels which exhibit a lower Po than those at the steady state (A.C.Powe & T.-C.Hwang, unpublished observations). Login to comment 214 ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala To see whether the reduced Po of K464A-K1250A mutants arises only from shorter open times, we calculated mean closed times from steady-state Po and relaxation time constants (see Methods). Login to comment 215 ABCC7 p.Lys1250Ala The calculated mean closed time for K1250A at 2.75 m ATP was 18 ± 4 s (n = 6; Fig. 7C). Login to comment 216 ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala Thus, K1250A prolongs closed time >30-fold compared either to wild type or to the K464A single mutant (Fig. 3B). Login to comment 218 ABCC7 p.Lys1250Ala ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala The mean closed time for K464A-K1250A (17 ± 4 s, n = 5; Fig. 7B) is similar to that for K1250A (P ∆ 0.42). Login to comment 219 ABCC7 p.Lys464Ala Our results show that the lower Po in the double mutant is mostly due to shortening of K1250A`s long open time by K464A. Login to comment 220 ABCC7 p.Lys1250Ala Although K1250A exhibits long open times at millimolar [ATP], the mutant channel opens only briefly at micromolar [ATP] (Zeltwanger et al. 1999). Login to comment 221 ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala We tested whether K464A-K1250A behaved in a similar manner. Login to comment 222 ABCC7 p.Lys464Ala A typical example of K464A-K1250A`s gating behaviour is shown in Fig. 8A. Login to comment 226 ABCC7 p.Lys1250Ala This result is qualitatively similar to that shown for K1250A (Zeltwanger et al. 1999). Login to comment 227 ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala We then examined the open time distribution of K464A-K1250A channels in the presence Functional interaction between nucleotide binding domains of CFTRJ. Physiol. 539.2 341 Figure 7. Login to comment 228 ABCC7 p.Lys1250Ala ABCC7 p.Lys1250Ala ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala ABCC7 p.Lys464Ala K464A shortens K1250A relaxation A, representative trace of macroscopic current relaxations from CFTR-K1250A (top trace) and CFTR-K464A-K1250A double mutant channels upon withdrawal of PKA (40 U ml_1 ) and ATP (1 m). Login to comment 229 ABCC7 p.Lys1250Ala ABCC7 p.Lys1250Ala ABCC7 p.Lys1250Ala ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala ABCC7 p.Lys464Ala Mean relaxation time constant for the CFTR-K1250A trace shown is 110 ± 1 s and for CFTR-K464A-K1250A is 30 ± 1 s. B, few-channel traces of CFTR-K1250A and CFTR-K464A-K1250A at the steady state in 2.75 m ATP. Login to comment 230 ABCC7 p.Lys1250Ala ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala Dashed lines indicate baseline (all channels closed); marks at the left indicate open channel current levels (a total of 4 channels for K1250A and 3 for K464A-K1250A). Login to comment 231 ABCC7 p.Lys1250Ala ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala C, comparison of steady-state Po, mean open (relaxation) times and mean closed times for CFTR-K1250A and CFTR-K464A-K1250A. Login to comment 232 ABCC7 p.Lys1250Ala ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala Asterisks indicate significant differences between CFTR-K1250A and CFTR-K464A-K1250A (**P < 0.01; ***P < 0.005). Login to comment 234 ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala A single exponential fit to the distribution provides an estimated open time of 241 ± 3 ms, similar to the open time of wild type, K464A and K1250A at 10 µ ATP (~250 ms; Zeltwanger et al. 1999; present study, Fig. 3A). Login to comment 235 ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala Thus, K464A had little effect on brief openings seen in K1250A at micromolar [ATP]. Login to comment 241 ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala We demonstrate that K1250A, but not K464A, affects the opening rate. Login to comment 242 ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala We also show that both K464A and K1250A affect closing at millimolar [ATP] but in opposite ways. Login to comment 243 ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala K464A accelerates closing whereas K1250A delays it. Login to comment 249 ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala Ramjeesingh et al. (1999) showed that K464A only partly reduced CFTR`s ATPase activity while K1250A eliminates it altogether. Login to comment 250 ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala Aleksandrov et al. (2001) showed that K1250A had no effect on 8-azido- [a-32 P]-ATP labelling of CFTR whereas K464A drastically reduced it. Login to comment 255 ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala It was further hypothesized that hydrolysis is the main pathway for closing under normal conditions and that blocking hydrolysis with K464A or K1250A permits closing only through the slow unbinding pathway, resulting in prolonged openings. Login to comment 256 ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala Based on that model, one would predict that the double mutant K464A-K1250A should exhibit long openings at all ATP concentrations, since the hydrolysis pathway at both A. C. Powe, Jr, L. Al-Nakkash, M. Li and T.-C. Hwang342 J. Physiol. 539.2 Figure 8. Login to comment 257 ABCC7 p.Lys1250Ala ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala ABCC7 p.Lys464Ala K464A-K1250A gating at millimolar and micromolar [ATP] A, representative sweep from experiments with CFTR-K464A-K1250A channels exposed first to 1 m and then to 10 µ MgATP. Login to comment 258 ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala Arrow indicates the baseline, downward deflections channel openings. B, survivor plot of open dwell times for CFTR-K464A-K1250A at 10 µ MgATP. Login to comment 262 ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala We find, however, that the mean open time for K464A-K1250A is ~250 ms at 10 µ ATP and ~30 s at 2.75 m ATP (Figs 7 and 8B). Login to comment 263 ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala The double mutant open time at 10 µ ATP is similar to that of wild type, K464A and K1250A at the same [ATP] (Zeltwanger et al. 1999; present study). Login to comment 264 ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala This finding, together with the dramatic differences between K464A and K1250A mutants, casts considerable doubt on the idea that the NBDs function identically. Login to comment 270 ABCC7 p.Lys464Ala The frequency and duration of these locked open events in the presence of AMP-PNP are reduced by the NBD1 mutation K464A (Figs 4, 5 and 6). Login to comment 272 ABCC7 p.Lys464Ala The reduction of AMP-PNP`s lock open effect by K464A then suggests that lysine 464 in NBD1 regulates nucleotide action at NBD2. Login to comment 276 ABCC7 p.Lys464Ala Because K464A reduces the apparent on-rate (i.e. locking rate) of AMP-PNP (Fig. 6), we expected that this mutation should also impair ATP`s ability to prolong open time. Login to comment 278 ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala Furthermore, the NBD2 mutation K1250A greatly prolongs CFTR open time ~300-fold; that prolongation is then reduced ~5-fold by addition of the NBD1 mutation K464A (Fig. 7). Login to comment 288 ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala Furthermore, Ramjeesingh et al. (1999) showed that K464A reduces ATPase activity by ~80% and K1250A virtually eliminates it, suggesting that mutating one NBD affects the biochemical activity of the other. Login to comment 299 ABCC7 p.Lys1250Ala ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala ABCC7 p.Lys464Ala Channel open times at 10 µ ATP for wild type, K464A, K1250A and K464A-K1250A are all ~250 ms (Zeltwanger et al. 1999; present study, Figs 3 and 8), indicating that the NBD mutations have no effect on brief openings. Login to comment 305 ABCC7 p.Lys1250Ala The NBD2 mutation K1250A prolongs opening (Fig. 7), which suggests that blocking ATP hydrolysis at NBD2 slows exit from an open state. Login to comment 306 ABCC7 p.Lys464Ala On the other hand, the NBD1 mutation K464A shortens openings at millimolar [ATP] (Fig. 3). Login to comment 307 ABCC7 p.Lys464Ala K464A also impairs AMP-PNP`s ability to lock open CFTR and the channel`s ability to remain locked open (Figs 4, 5 and 6), which implies that blocking hydrolysis at NBD1 impairs the ability to enter and remain in the longer open state. Login to comment 311 ABCC7 p.Lys1250Ala While our results provide no evidence for involvement of NBD1 in channel opening, NBD2 seems to play a role, since the NBD2 mutation K1250A increases closed time > 30-fold (Fig. 7). Login to comment 321 ABCC7 p.Lys464Ala If so, why does the K464A mutation not affect the opening rate? Login to comment 322 ABCC7 p.Lys464Ala There are two possible explanations for this observation: (1) ATP binding, but not hydrolysis at NBD1, opens the channel; or (2) NBD1`s role in channel opening is not revealed by the K464A mutation. Login to comment 323 ABCC7 p.Lys464Ala ABCC7 p.Lys464Ala Although biochemical studies show that K464A does impair ATPase activity in isolated NBD1 (Ko & Pedersen, 1995; King & Sorscher, 1998) as well as in the whole CFTR molecule (Ramjeesingh et al. 1999), the opening rate of K464A may not be affected if opening of CFTR only requires ATP binding. Login to comment 326 ABCC7 p.Lys464Ala Another possibility is that NBD1 may be involved in channel opening, but the domain`s role is not revealed by the K464A mutation. Login to comment 333 ABCC7 p.Lys1250Ala If it turns out that NBD1 is truly involved in channel opening, then the delay of opening by NBD2 mutant K1250A would suggest that NBD1 and NBD2 interact to control opening as well as closing. Login to comment 334 ABCC7 p.Lys464Ala A. C. Powe, Jr, L. Al-Nakkash, M. Li and T.-C. Hwang J. Physiol. 539.2 Conclusions To examine the contribution of NBD1 during CFTR gating, we assessed the kinetic properties of the Walker-A mutant K464A. Login to comment 335 ABCC7 p.Lys464Ala We found that K464A had little effect on the apparent ATP dependence or opening rate of the channel. Login to comment 337 ABCC7 p.Lys464Ala K464A also diminished AMP-PNP`s ability to stabilize channel open state by affecting both apparent on- and off-rates of the non-hydrolysable analogue. Login to comment 338 ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala Finally, K464A reduces the prolongation of open time seen in the NBD2 mutant K1250A at millimolar [ATP], strongly suggesting an interaction between NBD1 and NBD2 during CFTR`s open state. Login to comment 339 ABCC7 p.Lys1250Ala ABCC7 p.Lys464Ala Although the NBD1 mutant K464A did not affect opening, the NBD2 mutant K1250A delays opening >30-fold compared to wild type. Login to comment