Nucleotide Binding Domain 1 Review

NUCLEOTIDE BINDING DOMAIN 1 REVIEW

FSLLGTPVLKDINFKIERGQLLAVAGSTGAGKTSLLMMIMGELEPSEGKIKHSGRISFCSQFSWI

MPGTIKENIIF*GVSYDEYRYRSVIKACQLEEDISKFAEKDNIVLGEGGITLSGGQRARISLARAVY

KDADLYLLDSPFGYLDVLTEKEI

(amino acids 433 thru 586 or possibly to 622 or 634)

NBD1 is an intracellular domain of CFTR and is a "mutational hotspot" for cystic fibrosis. This fact is taken as evidence for the central importance the NBD domains have for the correct functioning (gating, mainly) of CFTR as well as proper targeting of CFTR to the cell membrane. The delta(F508) mutation as well as many other disease-causing mutations are located in NBD1. The amino acid sequences of the NBD domains are the most conserved areas of the CFTR protein when comparing CFTR among different species, and this is yet another reason to believe that the NBDs are crucial for proper functioning of this ion channel. In the past 10 years, models attempting to explain exactly how the NBD domains function to control gating have grown exceedingly complex. In the early 1990s, before it was recognized that CFTR was itself a chloride channel in addition to being a regulator of other ion channels, the NBDs were only thought to be involved in regulation of the as yet to be discovered "true CF chloride channel". The theory has since been modified and expanded to include the notion that the NBDs serve multiple roles including opening and closing the CFTR channel as well as perhaps binding and regulating proteins in the cell by altering the gating of these other ion channels. What is known for certainty about the NBD domains of CFTR is that they bind and hydrolyze intracellular ATP. Which implies that the NBDs are capable of undergoing a conformational change during this process. It may be that this change in structure keeps the gating kinetics of CFTR (opening and closing of the channel) always "moving in the right direction" and prevents it from being reversible. Reversible channel gating would be a waste of energy, and it may be that the NBDs function to prevent this "waste". This review is divided up into the following sections: Theory and Previous Experimental Results, Structure and Mechanism of NBD1, delta(F508), and Recent Evidence.

In the September 7, 2001 issue of Science, a high-resolution crystal structure was reported for a close relative of CFTR from a bacterium. The ABC transporter, MsbA, normally transports Lipid A as a substrate against a concentration gradient and is therefore considered a pump rather than an ion channel. Unfortunately, the entire NBD domains were not resolved in the structure, however most of it was. From this picture of MsbA, we can see that the NBD domains are in direct contact with intracellular domains called ICDs, which would correspond in CFTR to the intracellular loops and the portion between TM6 and NBD1 (or TM12 and NBD2). In MsbA, amino acids 341 thru 418 are disordered. This portion contains the WalkerA motif. No ATP or ATP analog co-crystallized with MsbA. The rest of the NBD sequence is visible, with amino acids 331 to 340 forming an alpha-helix, and amino acids 418 to 564 (containing the signature motif and Walker B motif) also well-resolved. The portions of the NBD domain of MsbA which are well-resolved agree well with the other 4 NBD structures which have been solved in the past, with a rms deviation of ~1.5 Å along the alpha-carbons. Certain amino acids in the NBD domain were found to be in direct contact with the intracellular loop regions (ICDs), including residues 420-448, 500-508, and 531-556. These residues also correspond to highly conserved residues among the subfamily of ABCs called MDR-ABC transporters. Chang and Roth, Science, 293, 9/7/01 pgs 1793-1800

THEORY AND PREVIOUS EXPERIMENTAL RESULTS:

By comparing CFTR sequences from different species, Chen et al. found highly conserved sequences from amino acids 439 through 646. This may be the actual NBD1 domain sequence, as it was also found that important CF-causing mutation hotspots were highly conserved among species, essentially validating this method of domain characterization using sequence comparisons among divergent species. This new NBD1 definition also seems to agree with the crystal structure of HisP. Chen, Jian-Min et al. Mol. Biol. Evol. 18(9):1771-1788 2001

Mutations in NBD1 do not affect conductance, but undoubtedly do affect channel opening. And there is evidence for "cross-talk" between the two NBDs as well. NBD1 has been expressed as a peptide and shown conclusively to hydrolyze ATP. Each NBD has 3 conserved areas: Walker A, Walker B, and LSGGQ motifs (see "structure" below). The physical structure of the NBD domains from bacteria (the E.coli ribose ABC transporter) has been solved indicates that it is likely that NBD1 is actually 60 residues into R domain as originally defined. This probably means the acid-end of NBD2 will also need to be redefined. Most models suggest that ATP binding and hydrolysis at NBD1 opens the channel, while NBD2 is able to close the channel upon binding and hydrolysis of ATP. Another model suggests that CFTR only interacts with its NBD1 when partially phosphorylated at the R-domain, and that this interaction causes a brief change to the open state, which is further stabilized by full phosphorylation as well as ATP binding at NBD2. From the large number of conflicting models about NBD function, it is clear that much experimental work still needs to be done before any conclusions can be safely drawn. Also recently solved are the NBDs from MalK and Rad50.

The exact boundary between the R-domain and NBD1 still hasn't been mapped precisely. The carboxy terminal boundary of NBD1 could be from residue 589 to 670.

Overall, the sequence similarity between NBD1 and NBD2 is only ~30%, but both have conserved Walker A and B motifs.

Using blocking agents like MOPS along with non-hydrolyzable ATP analogs and mutagenesis, it has been suggested that there are two distinct open states (O1 and O2) to the pore. Patch clamp recordings revel that bursts into the open state from the closed state put CFTR into the O1 conformation first. It then switches to the O2 state. This asymmetry suggests that the various steps in the gating process is not reversible, which makes sense since energy is being put into the system by hydrolysis of ATP. It should be noted that sometimes, there was reversibility, with sequences other than Closed to O1 to O2 to Closed. But the majority followed this sequence of events. The bottom line being that ATP hydrolysis at the NBDs changes the conformation of the pore in a nonreversible asymmetric way. To date, there is no single model which is able to explain all the data generated from these CFTR kinetics studies.

There is evidence NBD 1 forms part of pore. It appears NBD1 may be partially localized in the membrane, depending on the model used. In 1992 Arispe found that purified NBD1 domains incorporate into lipid bilayers forming a channel which was specific for anions verses cations. On the other hand, there is no evidence to suggest that mutations in the NBDs or R domain change conductance and therefore form part of pore.

Normally, ATP binding occurs at the NBDs following phosphorylation of the R-domain. And hydrolysis of ATP appears to be necessary for normal channel opening. The first NBD is involved in channel opening, while the second one in closing of channel, and CFTR can be locked into various states using nonhydrolyzable analogs of ATP. Most likely, steps corresponding to ATP hydrolysis also correspond to the irreversible steps in gating which keep the process moving in the right direction. The fact that mutations in NBD1 also affect channel closing has been used to infer "cross-talk" between the two NBDs. The NBD from bacteria is known to form a dimer both in the crystal structure as well as in vivo. This is shown to occur along the edge of Arm I. The Km of pure CFTR for ATP is 0.3mM and the channel hydrolyzes it at a rate of ~ 1 molecule ATP/second, which is similar to rate of channel opening and closing. NBDs have Walker A and B motifs as do other ATPases.

A peptide based on just the core 67 amino acid residues of NBD1 as well as the entire NBD1 domain have been expressed as a soluble protein that binds ATP in vitro.

ATPase activity is increased in CFTR 2- to 3-fold when PKA is present, and is due to a decreased Km (which changes from 1 to 0.3). Vmax stays the same, however. When PKA sites are removed by mutagenesis, an increase in the ATP concentration is needed for half-maximal activity. These facts have been used to imply that the R-domain interacts directly with the NBDs. In fact, coprecipitation studies have shown the R-domain and NBD1 bind each other when coexpressed in baculovirus system.

NBD1 has been expressed alone. Ko and Pedersen found that it binds ATP, forming an ion-selective channel. NBD-maltose fusion proteins (in 1995) were found to hydrolyze ATP by 3 different assays. It was determined that the Km for ATP was ~86uM and the Vmax equal to ~ 0.31/min. ADP inhibited ATPase activity. Vmax was found in 1997 to be 30 nmol/mg/min. For comparison, the P-gp protein has a Vmax of ~ 1650. Li, C., et al. JBC 271: 28463-68 1996

The fact that CFTR with a stop codon at D836X (which lacks the C-terminal half of the protein and therefore has no TM 7-12 or NBD2) is able to become activated by PKA says that the first NBD and helices 1-6 are able to interact with the R domain. And since this construct was also active without PKA suggests that the R domain also interacts with NBD2.

G-proteins (which bind and hydrolyze GTP, functioning as "on and off" switches) show some sequence similarity found in NBD1. This has led to speculation that CFTR regulates other proteins by using it's NBD1 in a manner similar to that of a G-protein. However, "threading" of a G-protein sequence using a computer onto the known structure of an NBD was not considered very successful, implying a lack of structural similarity.

Mutation of the aspartate responsible for Mg++ binding increases channel open time. NBD2 seems to have a clearer role than NBD1. NBD1 with a substitution mutation in the Walker A lysine is not able to completely abolish channel activity, as would be expected if it was solely responsible for channel opening. There are 2 schools of thought on this: first, ATP hydrolysis at NBD1 may result in channel opening, and second, ATP hydrolysis at NBD1 may serve to prime CFTR into an active state which can then be opened by ATP binding at NBD2. It is possible that mutating the Walker A lysine in NBD1 does now wipe out ATP hydrolysis completely.

NBD1 is necessary for regulation of the sodium ion channel ENaC (and possibly other ion channels) in the upper airway of the lungs. Perhaps ENaC and CFTR are brought into close proximity via scaffolding proteins called PDZs (see c-terminal). Schreiber et al. PNAS 96, 5310-5315 1999

STRUCTURE AND MECHANISM: Until 1998, no structure at atomic resolution existed for any NBD-like domains. Then, in 1998 the x-ray crystal structure for the subunit of the pump Histadine Permease from a bacteria was solved (Nature 12/17/98). This structure revealed what appears to be dual "arms" composing the bacterial NBD, having the overall shape of the letter "L". One arm contains the Walker motif and binds ATP by the "P" loop. The other arm has an aspartic acid residue. The authors speculated that somehow, upon ATP hydrolysis, the first arm moves in such a way as to expose the highly conserved sequence LSGGQ motif and that this allows for interactions to occur with the NBD and the membrane-spanning domains. They went on to speculate that the reason for two NBDs is formation of a dimer with each other. The ATP hydrolysis therefore has the effect of producing a "structural amplification" because new protein-protein binding sites are produced when this occurs. ATP hydrolysis by the NBDs may be involved in changing the structure of the pore. It still isn't known which is more important in gating behavior: binding ATP or hydrolyzing of ATP. It is known that if ATP is removed, CFTR closes. And if PKA is removed, channels remain open as long as no phosphatases are present. Each NBD has a highly conserved Walker A, Walker B, and LSGGQ motif. The alpha or gamma phosphate of ATP probably interacts with the Walker A motif and this motif is necessary for ATP hydrolysis. There is evidence that the glycine in the LSGGQ motif is also involved in hydrolysis because in G proteins, it is also present and necessary for GTP hydrolysis. The Walker B motif has an aspartate which helps coordinate the Mg++ ion. Without this divalent cation cofactor, ATP doesn't bind well at all.

The NBD domain subunit of the maltose transporter complex MalK has also had its structure determined. EMBO J. 19., 5951-61 (2001) HisP and MalK seem to have the same backbone structures. This suggests that it may be possible to infer these structures to the NBD domains of CFTR, since it too is an ABC member. There appears to be 3 layers to each domain, a top one which is antiparallel beta-sheet, a middle layer (with the conserved Walker A and B motifs) that is a mixture of alpha-helix and beta sheet, and a lower layer (which contains the signature motif and is believed to interact with the transmembrane domains) which is mostly alpha-helix. In July, 2001 Karpowich et al (Dept. Biological Sciences, Columbia University) reported the cyrstal structures of MJ1267 ATP Binding Cassette. The title of the paper is "Crystal Structures of the MJ1267 ATP Binding Cassette Reveal an Induced-Fit Effect at the ATPase Active Site of an ABC Transporter." Structure (Camb) 2001 Jul 3;9(7):571-86 For the original MJ1267 structure paper, see "The crystal structure of the MJ0796 ATP-binding cassette: Implications for the structural consequences of ATP hydrolysis in the active site of an ABC-transporter." Yuan et al., J Biol Chem 2001 Jun 11

A preliminary description of another ABC NBD domain has been reported in an abstract of the RbsA protein, an ATPase domain of the E. coli ribose transporter. Pediatr. Pulmonol. 17, 91-2 (1998).

HisP has Walker A and B in one arm of the "L". The Walker A amino acid backbone interacts with ATP using hydrogen bonds and its phosphate. Using an aspartate at the C-terminus of the Walker B motif, the NBDs bind to Mg++ ion necessary for catalysis. This corresponds to Asp572 on CFTR (NBD1). Note: for NBD2, this would correspond to Asp1370. The water molecule which is necessary for hydrolysis is hydrogen bonded to the equivalent of Ser573 and Gln493 in NBD1 and Gln1291 and Glu1371 in NBD2 of CFTR. LSGGQ in ArmII forms the core. The authors also speculate that the membrane spanning regions interact here with the armII. The common mutation site F508 can be found "on an exposed surface of an alpha-helix in Arm II." Each NBD is composed of seven alpha helices and seven beta sheet regions. The walker B motif is from amino acids 568-572 and is predicted to interact with the TM domains. In addition to the NBD of HisP, the NBD structure of LivF is available now.

Chimeras of CFTR with yeast mating factor STE6 (another ABC transporter family member) indicates there is no functional difference between the different NBDs within ABC transporter family. A CFTR NBD was able to replace the STE6 ones, and yeast mating factor was still transported out (i.e. the transporter still functioned normally). It is possible that the NBDs form a "standard engine" specialized for extracting energy from ATP and coupling it via a conformational change into changes in the membrane spanning domains.

A455E is a CF-causing mutation and results in a less severe form of the disease than deltaF508. It and P574E are located within NBD1 and appear to be in places important for binding nucleotides. Mutant CFTR with these amino acid changes have reduced chloride currents, but do not disrupt the function of the channel itself. Both have pore properties that appear normal and are regulated by cAMP-dependent phosphorylation (PKA). The open probability of A455E is indistinguishable from wild-type while P574E has an open probability even higher than normal.

P574H is a naturally occurring mutation which has the effect of compensating for another mutation by increasing the open probability by 60% over wild-type.

The naturally occurring mutations at K464 (in Walker A motif) slow the opening of CFTR channels probably by reducing the ability of NBD1 to bind the phosphates of ATP. G551 and D572 also have similar effects.

A nonsense (stop) mutation which changes gly542 to a stop codon accounts for 13% of CF mutations found in Ashkenazi Jews.

It now appears that NBD1 is not only capable of opening the channel, but may also be able to close it perhap upon the release of hydrolysis products. NBD2, therefore, may only function to prolong the open state. NBD2 may be functioning like a G-protein, i.e. as a molecular "switch", or timer.

Some researchers don't believe ATP hydrolysis is necessary at the NBD domains to open CFTR. They have been able to measure channel opening in the absence of Mg++ as well as nonhydrolyzable analogs of ATP. It may be that binding of ATP alone, as opposed to hydrolysis, may be enough to overcome the energy of activation of CFTR channel opening (~25 kcal/mol?). Note: it is possible that Mg++ is not necessarily needed for hydrolysis, and monovalent cations in the buffer are catalyzing ATP hydrolysis. Also, the largest free energy release upon a ligand binding to a protein is ~18 kcal/mol (biotin/strepavidin), suggesting ATP hydrolysis energy is needed. Biochemistry 40, 5/15/2001

DELTA (F508)

Delta(F508) is a mutant form of CFTR with a loss (delta = deletion) of a phenylalanine in the 508th position of the protein. The area around this 508th amino acid in CFTR is highly conserved among mammals. In fact, approximately 70% of CF patients have two copies of the delta(F508) mutation, while 90% have at least one. In 1992 it was determined that the dF508 CFTR protein is more subject to denaturing conditions and it was concluded that folding of the protein is probably what is altered. In 1993 it was shown the mutant protein gets turned over quicker in cell membrane. Also, CD done on peptides of this region in 1993 showed dF508 causes a decrease in beta-strand structure and an increase in random coil structure. These mutant peptide were also more subject to degredation. It has been shown using functional assays that some deltaF508 CFTR can make it out of the ER and apparently functions normally, albeit with a reduced open probability.

DeltaF508 haplotypes are very ancient and it was discovered in 1996 to be due to a high number of slippage events at microsattelites in introns surrounding the exon. It was concluded that the delta(F508) mutation occurred in a population distinct from Europe about 50,000 years ago.

In 1997, it was found that certain low molecular weight compounds like glycerol, D2O, and DMSO, known to stabilize proteins in native conformational states, rescued deltaF508 CFTR from the ER, as did low temperatures. They are called "chemical chaperones" for their ability to guide proteins into their proper configurations. In 1996, it was determined that F508 makes crucial contacts during folding. Even wild type CFTR is inefficiently processed thru ER, however. A proteosome (a cytosolic 26S complex) appears to degrade both normal and mutant CFTR protein. Since ~ 80% of CFTR is cytosolic, this isn't so surprising. 2 Chaperones may interact: cytosolic HSP70 and the resident membrane protein calnexin (one of whose functions is to retain partially folded proteins in the ER), but as of 1997 not the resident ER-lumen chaperones BiP and Grp94. In 1996, it was suggested that bacterial expressed deltaF508 NBD1 showed no unusual CD spectra, but in some studies it is slightly altered. Also in 1996 it was found that CPX can activate dF508 mutant protein but not the wild-type. In March, 1998 it was discovered that it takes 30 minutes longer for mature CFTR to be translated, but ubiquinatin takes only a mere 20 minutes, hence the rapid turnover of CFTR in the ER. It was even found that CFTR was ubiquinated while ribosomes were still attached. Polyubiquinatin occurs cotranslationally as well as post-translationally. It is a curious fact, however, that the homologus MDR (P-glycoprotein) protein is 100% translocated to CM, while CFTR is only 30% when present as wild type, suggesting that some factors other than size are important. A promising avenue of research would be to try and coax deltaF508 CFTR out of the ER using chemical chaperones. Since it appears to function normally if allowed to reach the cell membrane, this line of research has the potential to help approximately 90% of CF patients who have at least one copy of the deltaF508 protein. The only difference in NBD1 with and without deltaF508 seems to be when there is urea present. When a chimera CFTR deltaF508-STE6 was used, Teem et. al in 1993 found two mutations in NBD1: R553M and R553Q which were able to partially restore the wild-type function. At temperatures of 37 Celsius, the protein is misfolded, but between 25 and 30C, it is not. It is interesting that mutations in NBD1 result in less channels making it to the apical membrane, while mutations in NBD2 which are at equivalent places show no decrease in channel density at the surface. "The deltaF508 chloride channel formed therefrom showed a decreased half-life and reduced open probability and sensitivity to stimulation with cAMP agonists." The protein MHC class I molecule also seems to exhibit the same type of slow kinetics of folding and plasma membrane targeting that CFTR does.

For a brief discussion concerning the possible origins of this CF-causing mutation, and it's possible protective benefits, click here.........

Recent Evidence: More Pieces of the Puzzle?

In August, 2001 Falcon-Perez et al from the Instituto de Investigaciones Biomedicas in Madrid, Spain reported that, in an effort to identify intramolecular interactions in ABC Transporters necessary for energy transfer mechanisms during transport, they made a series of mutant yeast cadmium factor ABC proteins (ycf1) which would suppress ycf1 mutations that affect highly conserved sites "presumably involved in ATP binding and/or hydrolysis." They found thirteen intragenic second-site suppressors ".... identified for the D777N mutation which affects the invariant Asp residue in the Walker B motif of the first nucleotide binding domain (NBD1)." They found that two of the suppressor mutations (V543I and F565L) were located in the first transmembrane domain (TMD1), nine (A1003V, A1021T, A1021V, N1027D, Q1107R, G1207D, G1207S, S1212L, and W1225C) are found within TMD2, one (S674L) in NBD1, and another one (R1415G) is in NBD2, indicating either physical proximity or functional interactions between NBD1 and the other three domains. They concluded that "....the original D777N mutant protein exhibits a strong defect in the apparent affinity for ATP and V(max) of transport. The phenotypic characterization of the suppressor mutants shows that suppression does not result from restoring these alterations but rather from a change in substrate specificity." In the article, they discuss the possible involvement of Asp777 in coupling ATPase activity to substrate binding and/or transport across the membrane. Ycf1p is a vacuolar ATP binding cassette (ABC) transporter required for heavy metal and drug detoxification and shows strong sequence similarity to CFTR. J Bacteriol 2001 Aug;183(16):4761-70

In July, 2001 Karpowich et al (Dept. Biological Sciences, Columbia University) reported the cyrstal structures of MJ1267 ATP Binding Cassette. The title of the paper is "Crystal Structures of the MJ1267 ATP Binding Cassette Reveal an Induced-Fit Effect at the ATPase Active Site of an ABC Transporter." Structure (Camb) 2001 Jul 3;9(7):571-86 For the original MJ1267 structure paper, see "The crystal structure of the MJ0796 ATP-binding cassette: Implications for the structural consequences of ATP hydrolysis in the active site of an ABC-transporter." Yuan et al., J Biol Chem 2001 Jun 11

Aleksandrov, Mengos, Chang, Aleksandrov, and Riordan (of the Mayo Clinic Scottsdale, AZ) in a March, 2001 paper entitled "Differential interactions of nucleotides at the two nucleotide binding domains of CFTR" question current views concerning models of CFTR gating regulation (i.e. that ATP hydrolysis at NBD1 and NBD2 may drive channel opening and closing, respectively) by pointing out that as yet there has been little biochemical confirmation of the predictions that these models make. They used photoaffinity labeling with 8-azido-ATP (an analog which supports channel gating as effectively as ATP) to evaluate interactions with each NBD in intact membrane bound CFTR. They stated that "mutagenesis of Walker A lysine residues crucial for azido-ATP hydrolysis to generate the azido-ADP which is trapped by vanadate indicated a greater role of NBD1 than NBD2." "Separation of the domains by limited trypsin digestion and enrichment by immunoprecipitation confirmed greater and more stable nucleotide trapping at NBD1. This asymmetry of the two domains in interactions with nucleotides was most emphatically reflected in the response to the non-hydrolyzable ATP analogue, 5?-adenylimidodiphosphate (AMP-PNP) which in the gating models was proposed to bind with high affinity to NBD2 causing inhibition of ATP hydrolysis there postulated to drive channel closing. Instead we found a strong competitive inhibition of nucleotide hydrolysis and trapping at NBD1 and a simultaneous enhancement at NBD2. This argues strongly that AMP-PNP does not inhibit ATP hydrolysis at NBD2 and thereby questions the relevance of hydrolysis at that domain to channel closing." Medline: no citation given as of 4/1/01

In March, 2000 Lu et al. expressed soluble NBD1-Rdomain and NBD2 separately in E. coli. They then combined these two proteins to assay in four of the following ways: 1) They used a fluorescence probe attached to c-end of one and saw red-shifted lambda max of fluorescence when the 2 bound together. 2) the mixture of proteins eluted first on a gel filtration column, 3) native-PAGE gel studies revealed mixture migrated as a single band w/Rf between either one separately. 4) trypsin digestion occured at a slower rate for the mixture. The interactions were stable even w/nucleotides or phosphorylation.

Aleksandrov, Chang, Aleksandrov and Riordan in October, 2000 write that "CFTR channel gating occurs in the absence of ATP hydrolysis and hence does not depend on an input of free energy from this source." They came to this conclusion in part thru the use of ATP analogues (which are poorly or non-hydrolysable) but were sufficient to open the channel. Closing was found to occur when the analogs or their hydrolysis products were released. They continued: "Not only can channel opening occur without ATP hydrolysis but the temperature dependence of the open probability (Po) is reversed, i.e. Po increases as temperature is lowered whereas under hydrolytic conditions, Po increases as temperature is elevated. This indicates that there are different rate-limiting steps in the alternate gating pathways (hydrolytic and non-hydrolytic). These observations demonstrate that phosphorylated CFTR behaves as a conventional ligand-gated channel employing cytoplasmic ATP as a readily available cytoplasmic ligand; under physiological conditions ligand hydrolysis provides efficient reversibility of channel opening." J Physiol 2000 Oct 15;528 Pt 2:259-65

In September, 2000 Duffieux, Annereau, Boucher, Miclet, Pamlard, Schneider, Stoven, and Lallemand (from Ecole Polytechnique in France) wrote that "determination of the three dimensional structure of NBD1 is essential to better understand its structure-function relationship." Along these lines, the investigators expressed NBD1 as a soluble protein without fusion tags or need for renaturation following purification. They state that "Using tryptophan intrinsic fluorescence, we point out that the local conformation, in the region of the most frequent mutation DeltaF508, could differ from that of the nucleotide-binding subunit of histidine permease, the only available ABC structure. We have undertaken three dimensional structure determination of NBD1, and the first two dimensional 15N-1H NMR spectra demonstrate that the domain is folded. The method should be applicable to the structural studies of NBD2 or of other NBDs from different ABC proteins of major biological interest, such as multidrug resistance protein 1 or multidrug resistance associated protein 1." Eur J Biochem 2000 Sep;267(17):5306-12

It's possible that the carboxy terminus of NBD1 extends beyond amino acid 590 and is actually situated between amino acids 622 and 634. And the amino terminus of NBD1 lies between amino acids 432 and 449. Chan, Csanady, Seto-Young, Nairn, and Gadsby reported in August, 2000 this finding based on their studies using CFTR channels "severed near likely NH(2)- or COOH-terminal boundaries of NBD1." They took channel activation as proof NBD1 was intact. The sever point was systematically shifted along the primary sequence. They write "The functionally identified NBD1 boundaries are supported by Western blotting, coimmunoprecipitation, and deglycosylation studies, which showed that an NH(2)-terminal segment representing aa's 3-622 (Flag3-622) or 3-633 (Flag3-633) could physically associate with a COOH-terminal fragment representing aa's 634-1480 (634-1480); however, the latter fragment was glycosylated to the mature form only in the presence of Flag3-633. Similarly, 433-1480 could physically associate with Flag3-432 and was glycosylated to the mature form; however, 449-1480 protein seemed unstable and could hardly be detected even when expressed with Flag3-432.....Our definitions of the boundaries of the NBD1 domain in CFTR are supported by comparison with the solved NBD structures of HisP and RbsA." J Gen Physiol 2000 Aug;116(2):163-80

In September 2000, Berger and Welsh (of Howard Hughes Medical Institute) wrote that "the most conserved features of this [ABC Transporter] family are the nucleotide-binding domains. As in other members of this family, these domains bind and hydrolyze ATP; in CFTR this opens and closes the channel pore. The recent crystal structures of related bacterial transporters show that an aromatic residue interacts with the adenine ring of ATP to stabilize nucleotide binding. CFTR contains six aromatic residues that are candidates to coordinate the nucleotide base. We mutated each to cysteine and examined the functional consequences. None of the mutations disrupted channel function or the ability to discriminate between ATP, GTP, and CTP. We also applied [2-(triethylammonium)ethyl] methanethiosulfonate to covalently modify the introduced cysteines. The mutant channels CFTR-F429C, F430C, F433C, and F1232C showed no difference from wild-type CFTR, indicating that either the residues were not accessible to modification, or cysteine modification did not affect function. Although modification inactivated CFTR-Y1219C more rapidly than wild-type CFTR, and inactivation of CFTR-F446C was nucleotide-dependent; failure of these mutations to alter gating suggested that Tyr(1219) and Phe(446) were not important for nucleotide binding. The results suggest that ATP binding may not involve the coordination of the adenine ring by an aromatic residue analogous to that in some bacterial transporters. Taken together with earlier work, this study points to a model in which most of the binding energy for ATP is contributed by the phosphate groups." J Biol Chem 2000 Sep 22;275(38):29407-12

CFTR mutations that alter NBD1 domain reduces glibenclamide sensitivity (also called glyburide, glibenclamide contains a sulfonylurea nucleus and a cyclohexyl ring and has a Mwt. of 498. It is a weak acid w/a pKa ~6.8 and is extensively bound to plasma proteins. Another sulfonylurea called LY-295501 also inhibits CFTR. Another is DPC, or dipheny-carboxylic acid, specifically inhibits CFTR. Also, MOPS does. ). Note: glibenclamide affects channels other than CFTR (guinea pig atrial ICL). Sequence similarity between NBDs and G-proteins suggests structural and functional similarity. In 1991, the two NBDs were attempted to be aligned. It was found that no sequence alignments worked and was therefore concluded the domains were probably not due to a recent gene duplication event. In 1997 NBDs alone were expressed and found to increase beta-sheet structures and self-associate into polymers over 300,000 Daltons in size. They were very lipophilic and disrupted bilayers easily. They entered planer lipid bilayers. It was also found that a discrete domain in NBD causes membrane targeting.

In Feb, 2001 Vergani et al. (Rockefeller University, NY) suggested based on mutagenesis results of the two NBDs and using varying concentrations of ATP that "ATP hydrolysis at NBD1 is required for tight nucleotide binding at NBD2, and ATP binding at NBD2 might precede opening [of channel]." Biophysical Society Meeting 2/2001 At the same meeting, Wang, et al. (Hospital for Children, Wilmington DE) reported that their data using soluble NBDs and R domains as well as overlapping peptide libraries generated from the sequence of the intracellular region of CFTR that "ATP binding sites are in a cleft between the NBDs and that R-domain is unlikely to be the channel gate." They found that several peptides from the NBDs bind to each of the three intracellular domains. They also mapped the same peptides onto the known HisP structure to identify sites on the NBDs that interact with other domains. A third group presented their results as well. Zhao and Ma (Case Western; Cleveland, OH) studied the naturally occurring CF mutation Y569C, both in HEK 293 cells as well as purified form. They chose this mutant because it is in the Walker B motif, which may be involved in forming part of the pore. One way to show this is to determine whether any changes in ion conduction occurs, which would be an indication of pore involvement. They recorded only a ~4 pS conductance (i.e. half wild-type) for chloride. They conclude "..the Walker B motif of NBD1 may interact with the pore region of CFTR and contribute to both gating and permeation properties...Mutation of tyrosine to cysteine could have impact on the overall conformation..."

Evidence for the notion that K464 in the Walker A site in NBD1 is able to affect the nucleotide dissociation rate at NBD2 in a positive way was put forth by Powe, et al (of U. MO-Columbia) at the 45th annual meeting of the Biophysical Society in Boston, 2/2001. The authors came to this conclusion while studying the mutant K464A affects on channel gating. The dissociation of the non-hydrolyzable ATP analog AMPPNP from NBD2 was lengthened in the mutant K464A version.

A perspective on CFTR channel gating by Zou and Hwang was published in Biochemistry (5/15/2001, Vol 40) where they pointed out that recent models suggest CFTR may be gated by binding of ATP alone. These studies have been done in the absence of Mg++ ions (to inhibit hydrolysis), as well as high concentrations (~5mM) of nonhydrolyzable analogs of ATP called AMP-PMP. It is noted that this high concentration is hard to explain since the Kd for ATP binding at NBD is in the micromolar range.

Some references used for this page....