Proc. Natl. Acad. Sci. USA Vol. 90, pp. 9963-9967, November 1993 Physiology Functional roles of the -binding folds in the activation of the cystic fibrosis transmembrane conductance regulator (chloride channel/traffic ) LISA S. SMIT*, DANIEL J. WILKINSONt, MONIQUE K. MANSOURAf, FRANCIS S. COLLINS*§1, AND DAVID C. DAWSONt Departments of *Human Genetics and tPhysiology, tBioengineering Program, and §Department of Internal Medicine, University of Michigan, Ann Arbor, MI 48109; and lNational Center for Human Genome Research, Bethesda, MD 20892 Communicated by Gerhard Giebish, July 2, 1993

ABSTRACT The cystic fibrosis transmembrane conduc- A tance regulator (CFTR), a member of the traffic ATPase NBF1 FSLLGTPVLKDINFKIERGQLLAVA STGAGKT SLLMMIMGE superfamily, possesses two putative nudeotide-binding folds 1 11 11 11 (NBFs). The NBFs are sufficiently similar that sequence align- NBF2 ytegGnaiLenIsFsIspGQrvgllrTGsGKs tLLsafl-r ment of highly conserved regions can be used to identify analogous residues in the two domains. To determine whether NBF1 LEPSEGKIKHSGRISFCSQFSWIMP-GTIKENI-IFGVSY-- this structural homology is paralleled in function, we compared the activation of chloride conductance by forskolin and NBF2 LlntEGeIqidGwdSitlQ-qWrkafGvIpqkvfIFsgtfrk 3-isobutyl-l-methylxanthine in Xenopus oocytes expressing CFTRs bearing mutations in NBF1 or NBF2. Mutation of a NBF1 --DEYR------YRSVIKACQLEEDISKFAEKDNIVLGEGGI conserved in the putative linker domain in either NBF lIIl 11 11 produced virtually identical changes in the sensitivity of chlo- NBF2 nlDpYeqwsdqeiwkVadevgLrsvIeqFpgKldfVLvdGGc ride conductance to activating conditions, and mutation of this I, B site in both NBFs produced additive effects, suggesting that in NBF1 TLSG QRARISLARAVYKDADILYLLDSPFGYLDVLTEKEIFE the two NBFs this region plays a similar and critical role in the 11 III 1 111 I 11 activation process. In contrast, substitutions in the NBF2 vLSh hkqlmcLARsVlskAklilLLDePsahLDpvTyqiIrr Walker A and B motifs, thought to form an integral part of the FIG. 1. Amino acid sequence alignment of NBF1 (amino termi- nucleotide-binding pockets, produced strikingly different ef- nus) and NBF2 (carboxyl terminus). Amino acid sequences ofNBF1 fects in NBF1 and NBF2. Substitutions for the conserved (residues 433-588) and NBF2 (residues 1219-1386) are shown in (Walker A) or aspartate (Walker B) in NBF1 resulted in a single-letter code. The bracket indicates the putative linker motif. marked decrease in sensitivity to activation, whereas the same Boxed residues indicate the Walker A motif, glycine-551 and -1349 changes in NBF2 produced an increase in sensitivity. These (located within the linker motif), and the Walker B motif. results are consistent with a model for the activation of CFTR in which both NBF1 and NBF2 are required for normal pocket to other domains of the (12, 13). The impor- function but in which either the nature or the exact conse- tance of this putative linker motif to CFTR function is quences of nucleotide binding differ for the two domains. evidenced by the seven CF missense mutations reported within this region of NBF1 and NBF2 (14). In a variety of cells expressing CFTR, experimental ma- Cystic fibrosis (CF) is caused by mutations in the gene neuvers that elevate cytosolic cAMP activate a Cl--selective encoding the cystic fibrosis transmembrane conductance conductance (gc) (15-24), and it is now generally accepted regulator (CFTR). Homology analysis revealed that CFTR is that CFTR functions as a Cl--selective ion channel (17, 25) a member of a superfamily of referred to as traffic gated by cAMP-dependent protein (PKA) and ATP ATPases (1) or ATP binding cassette (ABC) proteins (2, 3) (19, 26-29). In Xenopus oocytes expressing wild-type CFTR that includes various prokaryotic permeases as well as eu- we demonstrated (18) dose-dependent activation of Cl- con- karyotic proteins such as P-glycoprotein (4, 5). A consistent ductance by a phosphodiesterase inhibitor, 3-isobutyl-1- feature of this family is the presence of hydrophilic domains methylxanthine (IBMX), in the presence of forskolin, an having amino acid sequences characteristic of mononucle- activator of adenylate cyclase. We used the IBMX dose- otide-binding proteins. CFTR has two putative nucleotide- response to compare the conductance associated with wild- binding folds (NBFs) (6), and the amino acid sequences are type CFTR to that associated with CFTR variants bearing shown aligned in Fig. 1. Each predicted NBF contains three mutations in NBF1 (18). Certain CF-associated NBF1 mu- sequence motifs that are well conserved across the traffic tations reduced the sensitivity of CFTR to activating condi- ATPase superfamily. Two of these motifs, Walker A and B tions, and the reduction in sensitivity was correlated with the (7), are thought to form a portion of a nucleotide-binding severity of CF in patients carrying the mutations. These pocket. The conserved lysine in Walker A is predicted to findings suggested that the integrity of NBF1 was essential interact with the phosphoryl moiety of the bound nucleotide for normal activation of Cl- conductance by cAMP and that (8, 9) and the in Walker B to coordinate Mg2+ the reduced sensitivity to activation could be a causative of the MgATP complex (10, 11). The third motif lies between factorin disease. The IBMX dose-response appears to reflect the Walker A and B sequences and in a prokaryotic homo- the sensitivity of CFTR to increases in the activity of PKA logue, histidine permease, has been proposed to function as because reduced sensitivity of CFTR variants to IBMX is a linker that transmits signals from the nucleotide-binding Abbreviations: CF, cystic fibrosis; CFTR, cystic fibrosis transmem- The publication costs of this article were defrayed in part by page charge brane conductance regulator; NBF, nucleotide-binding fold; IBMX, payment. This article must therefore be hereby marked "advertisement" 3-isobutyl-1-methylxanthine; PKA, cAMP-dependent protein ki- in accordance with 18 U.S.C. §1734 solely to indicate this fact. nase. 9963 Downloaded by guest on September 30, 2021 9964 Physiology: Smit et al. Proc. Natl. Acad. Sci. USA 90 (1993) paralleled by reduced sensitivity to injected cAMP or cata- comparisons of activation sensitivities, CFTR variants were lytic subunit of PKA (30). expressed in the same oocyte populations. Wild-type CFTR The purpose of the present study was to examine the was tested in each population, and the results were combined functional consequences of corresponding mutations in to construct the average dose-response relationships shown NBF1 and NBF2 of CFTR by comparing the effects of these in Figs. 2-5. The greatest source of variability in determina- mutations on the sensitivity of Cl- conductance to activation tion of the normalized IBMX dose-response was the diffi- by forskolin and IBMX. The Xenopus oocyte expression culty in estimating the true value of the maximum cAMP- system offered two potential advantages: (i) dose-dependent induced Cl- conductance in oocytes expressing high levels of activation permitted a direct, quantitative comparison of the activatable CFTR. For this reason, expression levels for sensitivities of CFTR variants and (ii) mutant CFTR protein easily activated constructs (wild type, G551A, G1349A, appears less likely to be subject to intracellular processing K1250Q, and D1370N) were adjusted by reducing the amount problems in oocytes than in other cell types. For example, of injected RNA so that the maximum Cl- conductance was AF508, a mutant that exhibits a severe processing defect at similar to that attained by less sensitive constructs. The 37°C in mammalian cells (31-33), is associated with a robust IBMX dose-response relationships presented here were con- cAMP-activated Cl- conductance in oocytes (18). The results structed from experiments in which the maximum obtained with this expression system are consistent with the stimulated idea that both NBFs are required for normal activation of Cl- conductance was <100 ,LS. CFTR, but suggest that the two domains function differently in this process. RESULTS Analogous Mutations in the Putative Linker Domain. We MATERIALS AND METHODS began our functional comparison of NBF1 and NBF2 by In Vitro Mutagenesis Procedures, Plasmid Construction, and examining the effects of mutations at glycine-551 in NBF1 RNA Synthesis. To generate the mutation constructs, a 1.7-kb and the corresponding residue in NBF2, glycine-1349. These (includes NBF1 Walker A) and a 3.0-kb (includes G551, residues were chosen because both loci are associated with NBF1 Walker B, and NBF2) Sph I-Sac I CFTR cDNA naturally occurring CF mutations (14), they lie within a fragment were cloned into pSelect vector (Promega). Oligo- putative linker , and the glycine is absolutely nucleotide-mediated site-directed mutagenesis was per- conserved throughout the traffic ATPase family (6). The formed (Altered Sites manual, Promega) using oligonucleo- effects of , aspartic acid, and alanine substitutions for tides corresponding to the desired mutations. The mutated glycine-551 and -1349 were compared, and the results are fragments were transferred into a full-length CFTR cDNA plotted in Fig. 2 as stimulated conductance (gcl) normalized construct (23) in pBluescript (Stratagene) and the presence of to maximum stimulated gcl versus IBMX dose. As previously the mutations was verified by sequencing. The full-length reported (18), G551S, a mutation associated with mild disease constructs contained a valine-for-methionine substitution at (34), exhibited a moderate reduction in sensitivity (K1l2 = 1.1 amino acid 1475 (V1475M). This substitution does not affect mM IBMX) compared to wild-type (Kl2 = 0.3 mM). The the sensitivity of CFTR to forskolin and IBMX (unpublished analogous substitution in NBF2 (G1349S), although not iden- data). RNA was synthesized by in vitro transcription (18). tified in patients, produced a virtually identical reduction in Oocytes and RNA Injection. Female toads (Xenopus laevis) sensitivity (K1l2 = 1.1 mM IBMX). G551D, associated with were anesthetized by immersion in ice water containing severe CF (35, 36), and G1349D, also a CF mutation (37), 3-aminobenzoic acid ethyl ester (2 mg/ml; Sigma), and both exhibited a dramatic reduction in sensitivity (K1l2 = 2.5 oocytes were removed via a small abdominal incision. The NBF1 NBF2 follicular membranes were removed by blunt dissection after 0 0 wt (12) A A G551A (8) A-A G1349A (4) incubation for 2-6 hr in a nominally Ca2+-free solution c O G551S (9) * * G1349S (6) containing collagenase (2.5 mg/ml; GIBCO). Defolliculated V v G551 D (6) '-V G1349D (8) oocytes were injected with 0.1-15 ng of RNA in 50 nl of 100 diethyl pyrocarbonate-treated water. Two-Electrode Voltage Clamp. Oocytes injected 3-6 days previously were placed in a perfusion chamber, impaled with E two electrodes, and voltage clamped (TEV-200; Dagen In- struments, Minneapolis). The perfusion solution was an amphibian Ringer's solution containing 100.5 mM Na+, 2 mM CO) K+, 1.8 mM Ca2+, 1 mM Mg2+, 105.6 mM Cl-, and 5 mM Hepes. Membrane currents were assayed at a holding po- .E tential of 60 mV, inside negative, and the CFTR-mediated current was defined as that activated by a stimulatory cock- tail containing 10 AM forskolin and 0.02-5 mM IBMX. Membrane currents in uninjected or water-injected oocytes were typically unresponsive, whereas oocytes expressing 0.2 0.5 1 CFTR variants exhibited quasi-steady-state inward currents IBMX, mM with maximum values ranging from 200 nA to several micro- FIG. 2. IBMX dose-response relationships for wild-type (wt) amperes. Current-voltage records were acquired by means of CFTR and variants with a single mutation at either the 551 (NBF1) a computer-generated ramp command that changed the hold- or the 1349 (NBF2) locus. Glycine (G) was replaced by alanine (A), ing potential from -120 mV to +40 mV at a rate of 100 serine (S), or aspartate (D). Individual oocytes were exposed to 10 mV/sec. These records were corrected for a small compo- ,uM forskolin and various doses of IBMX. Values for stimulated nent of capacitative current. Membrane conductances were conductance (g) were calculated by subtracting the membrane calculated from the conductance measured in the absence of drugs from the conductance slopes of the current-voltage relations measured with each dose of IBMX, and each value was normalized between -50 and -70 mV. IBMX dose-response relation- to the maximum stimulated conductance (gm.) observed in the same ships were determined as described in the legend to Fig. 2. oocyte. In this figure and the following figures, the plotted values are For each CFTR construct, two or more batches of RNA were means ± SEM for several oocytes and the number of oocytes is expressed in oocytes from two or more donors. For direct shown in parentheses. Downloaded by guest on September 30, 2021 Physiology: Smit et al. Proc. Natl. Acad. Sci. USA 90 (1993) 9965 mM IBMX). The alanine substitution in either NBF produced O-O wt (12) a moderate reduction in sensitivity, roughly equivalent to that V-V K464Q ( 4) observed with the serine substitutions. 1OOT YF-V K1250Q ( 4) NBF1/NBF2 Double Mutants. To explore further the rel- the we measured the 9 . ative contributions of two domains, E 80 0 T~ ~ ~~~ //, activation of Cl- currents in oocytes expressing the double 0I-) I I mutants G551S/G1349S and G551D/G1349D. The combina- I I I00160-+ T/T / tion of two serine substitutions had an additive effect; the -0 . V10 T/ double mutant was markedly less susceptible to activation a) v ~~~I than either of the singly substituted constructs (Fig. 3). 40 I~~~~~~~~~ Similarly, the double aspartate substitution, compared to E O ° ~~~~~T/ 20- Q .y 1 either single substitution, reduced the sensitivity to activa- (1'- V V tion. Serine/aspartate double mutants were indistinguishable from the single aspartate constructs (data not shown). It ()v -r ii mm should be noted that for the least sensitive mutants the 0.02 0.05 0.2 0.5 1 2 5 dose-response never reached a plateau value, so that the IBMX, mM sensitivity to activation for those constructs was, ifanything, FIG. 4. IBMX dose-response relationships for CFTR variants overestimated due to normalization to a submaximum con- with a single mutation at the Walker A lysine (K) of either NBF1 ductance. (K464) or NBF2 (K1250). Lysine was replaced by glutamine (Q), and Substitutions in the ATP-Binding Pocket. To determine dose-response relationships were constructed as in Fig. 2. whether the functional homology between NBF1 and NBF2 extends beyond the putative linker region surrounding gly- response curves for these mutants show evidence of a small cine-551 and -1349, the effects of mutations in the Walker A decline in Cl- conductance at the higher doses of IBMX (2-5 and B motifs were explored. For the purpose of comparison, mM). This may reflect a small degree of IBMX block of one residue was chosen in each region; for Walker A, the CFTR that becomes apparent when the channels are com- highly conserved lysine (K464 and K1250) and for Walker B, pletely activated. the invariant aspartic acid (D572 and D1370). Substitutions for the Walker B aspartic acid also produced Fig. 4 shows the results obtained with constructs in which different results in the two NBFs. Fig. 5 shows the results glutamine was substituted for the Walker A . In NBF1 obtained when asparagine was substituted for the aspartic the substitution resulted in a moderate reduction in the acid at 572 (NBF1) or 1370 (NBF2). In NBF1, this substitu- sensitivity of CFTR to activation (K1l2 = 0.8 mM), roughly tion resulted in a dramatic reduction in sensitivity (K1l2 = 2.5 equivalent to that seen with G551S. The result of the substi- mM), roughly equivalent to that seen with AF508 (18) or tution at the 1250 codon in NBF2 was very different. Here, G551D, whereas in NBF2, the identical substitution pro- the same substitution produced an increase in the sensitivity duced an increase in sensitivity to forskolin and IBMX (Kql2 of Cl- conductance to activation over that observed with = 0.02 mM). wild-type CFTR (K1l2 < 0.02 mM). Alanine and arginine Macroscopic Conduction Properties Are Not Altered by NBF substitutions at lysine-464 and -1250 were associated with Mutations. For each mutation, we also compared the mac- sensitivities similar to those observed with the glutamine roscopic conduction properties with those seen in oocytes substitutions (K464A or K464R, Kil2 = 0.8 mM; K1250A or expressing wild-type CFTR. Current-voltage plots were rap- K1250R, K,12 < 0.02 mM). In some oocytes expressing idly acquired by using a ramp voltage-clamp command at the constructs bearing mutations at residue 1250, a Cl- conduc- peak of activation, and the effect of reducing extracellular tance was detected in the absence of forskolin and IBMX. Cl- was noted. A comparison ofplots obtained in this manner This baseline conductance was subtracted in the computation disclosed that there were no readily discernible differences in of the activated conductance, so the apparent K1l2 for acti- the shape of the plot or the extent to which the reversal vation is probably an underestimate of sensitivity. The dose- potential was shifted by reducing extracellular Cl- (data not shown). Even in the case of a mutation that severely com- NBF1 NBF1 + NBF2 promised activation, e.g. G1349D, current-voltage plots O O G551S (9) O--O G551S + G1349S (7) were indistinguishable from wild type. v- v G551D (6) *--* G551 D + G 1 349D (5) 100- 0 0 wt (12) 1 A -A rl% - 7 k t rd:\ 0 Owt (12) 0 100 T z-Za uab/ZN k b) Oz A-- 6 AK A D1370N ( 5) T/ z g 80- E 0 / 80+ T C3) T/ E A O ° 60- * /1 T I I * , + 0 V 60 ai) T X 40- a) f I~0 ° T 40+- A I~~~~~~ E 0 r / T co 20 E t u A~~~~ 6- o ._ -- CO / / 20 + 3 _ ...... aiAi~~~~~~~~~~~ . 0.02 0.05 0.2 0.5 1 2 45 ti im A ,! IBMX, mM 0.02 0.05 0.2.0.2 , 0.5 2 45 IBMX, mM FIG. 3. IBMX dose-response relationships for CFTR variants with mutations at both the 551 (NBF1) and 1349 (NBF2) loci. Both FIG. 5. IBMX dose-response relationships for CFTR variants were replaced by (G551S + G1349S) or aspartates with a single mutation at the Walker B aspartate of either NBF1 (G551D + G1349D), and the dose-response relationships were (D572) or NBF2 (D1370). Aspartate (D) was replaced by asparagine constructed as in Fig. 2. (N), and dose-response relationships were constructed as in Fig. 2. Downloaded by guest on September 30, 2021 9966 Physiology: Smit et al. Proc. Natl. Acad. Sci. USA 90 (1993) DISCUSSION observations reported here are consistent with the idea that modifications of the Walker motifs in NBF2 could increase Both NBF1 and NBF2 Are Required for Normal Activation sensitivity to activation by attenuating the inhibitory influ- of CFTR. The effects of mutating glycine-551 and -1349 in ence of ADP. NBF1 and NBF2 on the sensitivity of C1- currents to CFTR Activation: and NBF Transduction activation by forskolin and IBMX provide strong evidence Processes. To understand how NBF mutations could alter the that both of these domains play a critical role in the trans- sensitivity of CFTR to activation by cAMP, it is necessary to duction mechanism that produces Cl- channel activation in place the results of these experiments in the context of a response to increases in cytosolic cAMP. Although the kinetic model for the activation process, such as that shown precise role of the NBFs in channel activation has not been in Fig. 6. In this simplified scheme, activation requires established, the available evidence (19, 26-29) can be ac- phosphorylation of the R domain and the binding of ATP at commodated by a kinetic model that envisions channel the NBFs followed by a further transition of the phosphor- activation as requiring at least two events: (i) phosphoryla- ylated, ATP-bound form of CFTR to the conducting (open) tion of one or more serine residues in the R domain in the state. No assumption is made about the order of the first two presence of the catalytic subunit of protein kinase A and (ii) events in this model, but the rapid activation and inactivation the binding and/or hydrolysis of ATP at one or both of the of phosphorylated CFTR by ATP in detached patches (27) NBFs. Experiments utilizing detached patches from NIH suggest that phosphorylation followed by ATP binding is one 3T3 cells expressing wild-type CFTR showed that, after mode ofactivation. The model is simplified in that it does not exposure to a phosphorylating environment, channel opening include explicitly multiple phosphorylated states (26, 29, 40, required the presence of a readily hydrolyzable analog of 41), differential binding ofATP at the two NBFs, or inhibition ATP (27); this observation prompted the speculation that by ADP. The model does provide, however, a basis for ATP hydrolysis may provide free energy necessary to effect understanding how an alteration in the ATP-binding or trans- a conformational change that leads to the conducting state. duction steps can change the kinetics of the dose-dependent Although the mechanism for the ATP effect remains obscure, activation of Cl- conductance by increasing concentrations the present results indicate that normal activation requires of cAMP. two functional NBFs. That substitutions for the invariant The steady-state solution for this kinetic scheme can be glycine in the putative linker region produce comparable found in standard textbooks on kinetics (42). A effects in either NBF1 or NBF2 suggests that there is a basic detailed exposition is beyond the scope of this paper, but similarity in the mechanism by which each NBF contributes several qualitative results are pertinent. In the context of the to the activation of Cl- conductance, and is consistent with present experiments the parameter of interest is the apparent the observation that substitution in either ofthese regions can Kl2 for the activation of Cl- conductance as cAMP is produce severe CF. Anderson and Welsh (28) found that in increased. The most important result of the analysis is that detached patches from transfected cells, mutation of the the sequential nature of the reactions makes them intrinsi- conserved glycine (G551S or G1349D) dramatically reduced cally interdependent. That is, the apparent Kl2 for the the value of the open probability (PO) in the presence of PKA activation of Cl- conductance by PKA will be a function of and MgATP. The requirement for two intact NBFs has been the kinetic constants for all of the intermediate reactions demonstrated for some other traffic ATPases, including the leading to the opening of the Cl- channel. For example, mammalian P-glycoprotein (38) and the bacterial protein regardless ofthe order of events, a mutation that reduces the UvrA (39). affinity ofthe NBFs for ATP, or renders the transduction step NBF1 and NBF2 Function Differently in Activation. In less likely, will result in an increase in the level of PKA contrast to the symmetric effect of alterations in the putative required to achieve a half-maximal Cl- conductance. These linker regions of NBF1 and NBF2, substitutions in the effects occur because altering any single step in the reaction presumed nucleotide-binding pockets produced markedly sequence affects the probability of finding the protein in its different effects. Substitutions for the conserved lysine in different states. For example, if the predominant order of Walker A or the conserved aspartate in Walker B, when made reactions were phosphorylation followed by ATP binding, a in NBF1, produced a discernible reduction in the sensitivity defect in the latter step would reduce the rate offormation of of CFTR to activating conditions. In contrast, the analogous the ATP bound form so that achieving half-maximal conduc- substitutions in NBF2 had the opposite effect, actually in- tance would require an increased abundance of the phos- creasing sensitivity over that seen with wild-type CFTR. This phorylated form of CFTR. Likewise, for the reverse order, striking difference suggests that, despite the fact that both reduced efficacy of ATP binding would dictate that higher NBFs are required for normal activation, they either function levels of PKA are required to overcome the reduced abun- differently in the activation of CFTR or differ in their dance of CFTR-ATP. Regardless of the preferred order of structural constraints on the transduction process. In de- the reaction, or even if there is none, a defect in the tached patches from mouse mammary epithelial cells ex- ATP-binding step or the subsequent transduction process is pressing CFTR, Anderson and Welsh (28) found that Walker to shift the IBMX curve to A mutations in the putative nucleotide-binding pocket of expected dose-response the right. either NBF1 or NBF2 increased the apparent Kl2 for the activation of phosphorylated CFTR by MgATP, as would be expected if the binding and/or hydrolysis of ATP were compromised. They also reported, however, that CFTR C1- channel activity was inhibited by ADP and that the inhibition was alleviated by Walker mutations in NBF2 but not by CFTR *ATP ( (CFTR* ATP) open mutations in NBF1, suggesting that NBF2 could be the site ~~/ ~~P of an inhibitory effect of ADP. One possibility, for example, P is that the local concentration of ADP exerts a negative ADP feedback control on CFTR activation. Of particular interest (PKA) is the observation (28) that mutation of the Walker A lysine CFTR ATP in NBF2 to methionine (K1250M) produced a 4-fold reduc- * tion in the apparent Kil2 for the activation of CFTR by ATP FIG. 6. Hypothetical model for the activation of CFTR by but also completely eliminated inhibition by 1 mM ADP. The phosphorylation and ATP binding. Downloaded by guest on September 30, 2021 Physiology: Smit et al. Proc. Natl. Acad. Sci. USA 90 (1993) 9967 Similarly, increased efficacy of the ATP-binding or transduc- Strong, T. V., Frizzell, R. A., Dawson, D. C. & Collins, F. S. tion steps, such as that which could result from reduced (1991) Science 254, 1797-1799. inhibition by ADP, will shift the apparent to the left. 19. Tabcharani, J. A., Chang, X.-B., Riordan, J. R. & Hanrahan, Kil2 J. W. (1991) Nature (London) 352, 628-631. The reactions leading to the activation of CFTR could 20. Anderson, M. P., Rich, D. P., Gregory, R. J., Smith, A. E. & exhibit a degree of interdependence beyond that required by Welsh, M. J. (1991) Science 251, 679-682. simple mass balance if there were a preferred order for the 21. Berger, H. A., Anderson, M. P., Gregory, R. J., Thompson, process-e.g., if the rate of ATP binding was significant only S., Howard, P. W., Maurer, R. A., Mulligan, R., Smith, A. E. for the phosphorylated form of CFTR, or if ATP binding & Welsh, M. J. (1991) J. Clin. Invest. 88, 1422-1431. increased the probability of phosphorylating specific sites on 22. Rich, D. P., Anderson, M. P., Gregory, R. J., Cheng, S. H., R The no evidence Paul, S., Jefferson, D. M., McCann, J. D., Klinger, K. W., the domain. present experiments provide Smith, A. E. & Welsh, M. J. (1990) Nature (London) 347, for such an additional degree of interdependence, but in view 358-363. of the potential complexity ofthe gating process, which could 23. Drumm, M. L., Pope, H. A., Cliff, W. H., Rommens, J. M., involve multiple phosphorylated states of CFTR (40) and two Marvin, S. A., Tsui, L.-C., Collins, F. S., Frizzell, R. A. & NBFs that function differently in the activation process, this Wilson, J. M. (1990) Cell 62, 1227-1233. possibility merits serious consideration. 24. Kartner, N., Hanrahan, J. W., Jensen, T. J., Naismith, A. L., Sun, S., Ackerley, C. A., Reyes, E. F., Tsui, L.-C., Rommens, We thank John Dickason and Gladys Palomeno for technical J. M., Bear, C. E. & Riordan, J. R. (1991) Cell 64, 681-691. assistance and David Gadsby for helpful discussions. This work was 25. Anderson, M. P., Gregory, R. J., Thompson, S., Souza, supported by grants from the Cystic Fibrosis Foundation (RDP MG D. W., Paul, S., Mulligan, R. C., Smith, A. E. & Welsh, M. J. 217) and the National Institutes of Health [DK45880 and DK29786) (1991) Science 253, 202-207. (D.C.D.) and DK39690 (F.S.C.)]. L.S.S. is a Howard Hughes 26. Cheng, S. H., Rich, D. P., Marshall, J., Gregory, R. J., Welsh, M. J. & Smith, A. E. (1991) Cell 66, 1027-1036. predoctoral fellow, D.J.W. is supported by a Cystic Fibrosis Foun- 27. Anderson, M. P., Berger, H. A., Rich, D. P., Gregory, R. J., dation training grant, M.K.M. was supported by a National Science Smith, A. E. & Welsh, M. J. (1991) Cell 67, 775-784. Foundation graduate research fellowship, and F.S.C. was previously 28. Anderson, M. P. & Welsh, M. J. (1992) Science 257, 1701- an Investigator of the Howard Hughes Medical Institute. 1704. 29. Picciotto, M. R., Cohn, J. A., Bertuzzi, G., Greengard, P. & 1. Ames, G. F.-L., Mimura, C. S. & Shyamala, V. (1990) FEMS Nairn, A. C. (1992) J. Biol. Chem. 267, 12742-12752. Microbiol. Rev. 75, 429-446. 30. Wilkinson, D. J., Drumm, M. L., Smit, L. S., Mansoura, 2. Hyde, S. C., Emsley, P., Hartshom, M. J., Mimmack, M. M., M. K., Strong, R. V., Uhler, M. D., Collins, F. S. & Dawson, Gileadi, U., Pearce, S. R., Gallagher, M. P., Gill, D. R., Hub- D. C. (1992) Pediatr. Pulmonol. Suppl. 8, 255 (abstr.). bard, R. E. & Higgins, C. F. (1990) Nature (London) 346, 31. Cheng, S. H., Gregory, R. J., Marshall, J., Paul, S., Souza, 362-365. D. W., White, G. A., O'Riordan, C. R. & Smith, A. E. (1990) 3. Higgins, C. F., Hiles, I. D., Salmond, G. P. C., Gill, D. R., Cell 63, 827-834. Downie, J. A., Evans, I. J., Holland, I. B., Gray, L., Buckel, 32. Denning, G. M., Ostedgaard, L. S. & Welsh, M. J. (1992) J. S. D., Bell, A. W. & Hermodson, M. A. (1986) Nature (Lon- Biol. Chem. 118, 551-559. don) 323, 448-450. 33. Kartner, N., Augustinas, L., Jensen, T. J., Naismith, A. L. & 4. Gros, P., Croop, J. & Housman, D. (1986) Cell 47, 371-380. Riordan, J. R. (1992) Nature Genet. 1, 321-327. 5. Chen, C.-J., Chin, J. E., Ueda, K., Clark, D. P., Pastan, I., 34. Strong, T. V., Smit, L. S., Turpin, S. V., Cole, J. L., Tom Gottesman, M. M. & Roninson, I. B. (1986) Cell 47, 381-389. Hon, C., Markiewicz, D., Petty, T. L., Craig, M. W., Rose- 6. Riordan, J. R., Rommens, J. M., Kerem, B.-S., Alon, N., now, E. C., III, Tsui, L.-C., lannuzzi, M. C., Knowles, M. R. Rozmahel, R., Grzelczak, Z., Zielenski, J., Lok, S., Plavsic, & Collins, F. S. (1991) N. Engl. J. Med. 325, 1630-1634. N., Chou, J.-L., Drumm, M. L., lannuzzi, M. C., Collins, 35. Cutting, G. R., Kasch, L. M., Rosenstein, B. J., Zielenski, J., F. S. & Tsui, L.-C. (1989) Science 245, 1066-1073. Tsui, L.-C., Antonarakis, S. E. & Kazazian, H. H., Jr. (1990) 7. Walker, J. E., Saraste, M., Runswick, M. J. & Gay, N. J. Nature (London) 346, 366-368. (1982) EMBO J. 8, 945-951. 36. Hamosh, A., King, T. M., Rosenstein, B. J., Corey, M., Le- 8. Fry, D. C., Kuby, S. A. & Mildvan, A. S. (1986) Proc. Natl. vison, H., Durie, P., Tsui, L.-C., McIntosh, I., Keston, M., Acad. Sci. USA 83, 907-911. Brock, D. J. H., Macek, M., Jr., Zemkova, D., Krasnicanova, 9. Tagaya, M., Yagami, T. & Fukui, T. (1987) J. Biol. Chem. 262, H., Vavora, V., Macek, M., Sr., Golder, N., Schwarz, M. J., 8257-8261. Super, M., Watson, E. K., Williams, C., Bush, A., O'Ma- 10. Jurnak, F. (1985) Science 230, 32-36. honey, S. M., Humphries, P., DeArce, M. A., Reis, A., 11. Schlichting, I., Almo, S. C., Rapp, G., Wilson, K., Petratos, Burger, J., Stuhrmann, M., Schmidtke, J., Wulbrand, U., K., Lentfer, A., Wittinghofer, A., Kabsch, W., Pai, E. F., Dork, T., Tummler, B. & Cutting, G. R. (1991) Am. J. Hum. Petsko, G. A. & Goody, R. S. (1990) Nature (London) 345, Genet. 51, 245-250. 309-315. 37. Beaudet, A. L., Feldman, G. L., Kobayashi, K., Lemna, 12. Mimura, C. S., Holbrook, S. R. & Ames, G. F.-L. (1991) Proc. W. K., Fernbach, S. D., Knowles, M. R., Boucher, R. C. & Natl. Acad. Sci. USA 88, 84-88. O'Brien, W. E. (1991) in The Identification of the CF (Cystic 13. Shayamala, V., Baichwal, V., Beall, E. & Ames, G. F.-L. Fibrosis) Gene: Recent Progress andNew Research Strategies, (1991) J. Biol. Chem. 266, 18714-18719. eds. Tsui, L.-C., Romeo, G., Greger, R. & Gorini, S. (Plenum, 14. Tsui, L.-C. (1992) Hum. Mutat. 1, 197-203. New York), pp. 53-54. 15. Rommens, J. M., Dho, S., Bear, C. E., Kartner, N., Kennedy, 38. Azzaria, M., Schurr, E. & Gros, P. (1989) Mol. Cell. Biol. 9, D., Riordan, J. R., Tsui, L.-C. & Foskett, J. K. (1991) Proc. 5289-5297. Natl. Acad. Sci. USA 88, 7500-7504. 39. Thiagalingam, S. & Grossman, L. (1991) J. Biol. Chem. 266, 16. Dalemans, W., Barbry, P., Champigny, G., Jallat, S., Dott, K., 11395-11403. Dreyer, D., Crystal, R. G., Pavirani, A., Lecocq, J.-P. & 40. Hwang, T.-C., Horie, M. & Gadsby, D. C. (1993) J. Gen. Lazdunski, M. (1991) Nature (London) 354, 526-528. Physiol. 101, 629-650. 17. Bear, C. E., Duguay, F., Naismith, A. L., Kartner, N., Han- 41. Strong, T. V., Wilkinson, D. J., Wood, D. L., Mansoura, M., rahan, J. W. & Riordan, J. R. (1991) J. Biol. Chem. 266, Dawson, D. C. & Collins, F. S. (1992) Pediatr. Pulmonol. 19142-19145. Suppl. 8, 249-250. 18. Drumm, M. L., Wilkinson, D. J., Smit, L. S., Worrell, R. T., 42. Segel, I. H. (1975) Enzyme Kinetics (Wiley, New York). Downloaded by guest on September 30, 2021