Impact of the ΔF508 Mutation in First Nucleotide-binding Domain of Human Cystic Fibrosis Transmembrane Conductance Regulator on Domain Folding and Structure
2004; Elsevier BV; Volume: 280; Issue: 2 Linguagem: Inglês
10.1074/jbc.m410968200
ISSN1083-351X
AutoresH.A. Lewis, Xun Zhao, Chi Chiu Wang, J.M. Sauder, Isabelle Rooney, B.W. Noland, Don Lorimer, M.C. Kearins, K. Conners, Brad Condon, Peter C. Maloney, William B. Guggino, J.F. Hunt, Spencer Emtage,
Tópico(s)Tracheal and airway disorders
ResumoCystic fibrosis is caused by defects in the cystic fibrosis transmembrane conductance regulator (CFTR), commonly the deletion of residue Phe-508 (ΔF508) in the first nucleotide-binding domain (NBD1), which results in a severe reduction in the population of functional channels at the epithelial cell surface. Previous studies employing incomplete NBD1 domains have attributed this to aberrant folding of ΔF508 NBD1. We report structural and biophysical studies on complete human NBD1 domains, which fail to demonstrate significant changes of in vitro stability or folding kinetics in the presence or absence of the ΔF508 mutation. Crystal structures show minimal changes in protein conformation but substantial changes in local surface topography at the site of the mutation, which is located in the region of NBD1 believed to interact with the first membrane spanning domain of CFTR. These results raise the possibility that the primary effect of ΔF508 is a disruption of proper interdomain interactions at this site in CFTR rather than interference with the folding of NBD1. Interestingly, increases in the stability of NBD1 constructs are observed upon introduction of second-site mutations that suppress the trafficking defect caused by the ΔF508 mutation, suggesting that these suppressors might function indirectly by improving the folding efficiency of NBD1 in the context of the full-length protein. The human NBD1 structures also solidify the understanding of CFTR regulation by showing that its two protein segments that can be phosphorylated both adopt multiple conformations that modulate access to the ATPase active site and functional interdomain interfaces. Cystic fibrosis is caused by defects in the cystic fibrosis transmembrane conductance regulator (CFTR), commonly the deletion of residue Phe-508 (ΔF508) in the first nucleotide-binding domain (NBD1), which results in a severe reduction in the population of functional channels at the epithelial cell surface. Previous studies employing incomplete NBD1 domains have attributed this to aberrant folding of ΔF508 NBD1. We report structural and biophysical studies on complete human NBD1 domains, which fail to demonstrate significant changes of in vitro stability or folding kinetics in the presence or absence of the ΔF508 mutation. Crystal structures show minimal changes in protein conformation but substantial changes in local surface topography at the site of the mutation, which is located in the region of NBD1 believed to interact with the first membrane spanning domain of CFTR. These results raise the possibility that the primary effect of ΔF508 is a disruption of proper interdomain interactions at this site in CFTR rather than interference with the folding of NBD1. Interestingly, increases in the stability of NBD1 constructs are observed upon introduction of second-site mutations that suppress the trafficking defect caused by the ΔF508 mutation, suggesting that these suppressors might function indirectly by improving the folding efficiency of NBD1 in the context of the full-length protein. The human NBD1 structures also solidify the understanding of CFTR regulation by showing that its two protein segments that can be phosphorylated both adopt multiple conformations that modulate access to the ATPase active site and functional interdomain interfaces. Cystic fibrosis causes lung, liver, pancreas, and reproductive tract disorders, typically leading to death prior to middle age from deterioration in pulmonary function (1Ratjen F. Doring G. Lancet. 2003; 361: 681-689Abstract Full Text Full Text PDF PubMed Scopus (893) Google Scholar). CFTR 1The abbreviations used are: CFTR, cystic fibrosis transmembrane conductance regulator; NBD1 and NBD2, nucleotide-binding domains 1 and 2; MSD1 and MSD2, membrane-spanning domains 1 and 2; R, regulatory region; TM, transmembrane; ABC, ATP-binding cassette; RI, regulatory insertion; RE, regulatory extension; CD, circular dichroism; rmsd, root mean square deviation; m, mouse; h, human; TCEP, Tris(2-carboxymethyl)phosphine. protein is composed of two membrane spanning domains (MSD1 and MSD2), two nucleotide-binding domains (NBD1 and NBD2), and a regulatory region (R). Although it functions as an ATP-gated anion channel, CFTR is a member of the ATP-binding cassette (ABC) transporter superfamily (2Dean M. Rzhetsky A. Allikmets R. Genome. Res. 2001; 11: 1156-1166Crossref PubMed Scopus (1499) Google Scholar) based on high sequence similarity between the NBDs and canonical ABC domains. Understanding the exact molecular pathology caused by the ΔF508 mutation in CFTR is of great importance in the development of drugs to treat cystic fibrosis because of the prevalence of this mutation in the human population. ΔF508 CFTR fails to mature appropriately in the endoplasmic reticulum and is poorly populated in the epithelial membrane (3Cheng S.H. Gregory R.J. Marshall J. Paul S. Souza D.W. White G.A. O'Riordan C.R. Smith A.E. Cell. 1990; 63: 827-834Abstract Full Text PDF PubMed Scopus (1423) Google Scholar, 4Dalemans W. Barbry P. Champigny G. Jallat S. Dott K. Dreyer D. Crystal R.G. Pavirani A. Lecocq J.P. Lazdunski M. Nature. 1991; 354: 526-528Crossref PubMed Scopus (569) Google Scholar, 5Lukacs G.L. Chang X.B. Bear C. Kartner N. Mohamed A. Riordan J.R. Grinstein S. J. Biol. Chem. 1993; 268: 21592-21598Abstract Full Text PDF PubMed Google Scholar, 6Ward C.L. Kopito R.R. J. Biol. Chem. 1994; 269: 25710-25718Abstract Full Text PDF PubMed Google Scholar). It has been proposed that the primary effect of the ΔF508 mutation is to cause misfolding of NBD1, which leads to aberrant transport and ultimately targeted proteolytic degradation of CFTR (7Thomas P.J. Ko Y.H. Pedersen P.L. FEBS Lett. 1992; 312: 7-9Crossref PubMed Scopus (60) Google Scholar, 8Thomas P.J. Shenbagamurthi P. Sondek J. Hullihen J.M. Pedersen P.L. J. Biol. Chem. 1992; 267: 5727-5730Abstract Full Text PDF PubMed Google Scholar). Channels harboring the deletion show enhanced sensitivity to proteolytic degradation (9Zhang F. Kartner N. Lukacs G.L. Nat. Struct. Biol. 1998; 5: 180-183Crossref PubMed Scopus (129) Google Scholar) but have at least partial wild-type chloride conductance properties (4Dalemans W. Barbry P. Champigny G. Jallat S. Dott K. Dreyer D. Crystal R.G. Pavirani A. Lecocq J.P. Lazdunski M. Nature. 1991; 354: 526-528Crossref PubMed Scopus (569) Google Scholar, 10Denning G.M. Anderson M.P. Amara J.F. Marshall J. Smith A.E. Welsh M.J. Nature. 1992; 358: 761-764Crossref PubMed Scopus (1061) Google Scholar). Canonical ABC domain structures are composed of three subdomains, a central F1-type ATP-binding core subdomain, an antiparallel β-sheet (ABCβ) subdomain, and an α-helical (ABCα) subdomain. One surface of the latter subdomain contains the stringently conserved LSGGQ signature sequence, whereas the opposite surface is of high sequence variability and is known to mediate contact of the ABC domain to the transmembrane α-helices in the structures of two different prokaryotic ABC transporter membrane proteins (11Chang G. FEBS Lett. 2003; 555: 102-105Crossref PubMed Scopus (153) Google Scholar, 12Bass R.B. Locher K.P. Borths E. Poon Y. Strop P. Lee A. Rees D.C. FEBS Lett. 2003; 555: 111-115Crossref PubMed Scopus (24) Google Scholar). Residue Phe-508 in the NBD1 of CFTR occurs near the C terminus of the first helix in the ABCα subdomain in this putative MSD1-interacting region. Recently, we reported the crystal structure of NBD1 from mouse CFTR (mNBD1) (13Lewis H.A. Buchanan S.G. Burley S.K. Conners K. Dickey M. Dorwart M. Fowler R. Gao X. Guggino W.B. Hendrickson W.A. Hunt J.F. Kearins M.C. Lorimer D. Maloney P.C. Post K.W. Rajashankar K.R. Rutter M.E. Sauder J.M. Shriver S. Thibodeau P.H. Thomas P.J. Zhang M. Zhao X. Emtage S. EMBO J. 2004; 23: 282-293Crossref PubMed Scopus (338) Google Scholar). The structure is similar to those previously reported for other ABC transporter NBDs (Fig. 1A and Ref. 14Jones P.M. George A.M. Cell Mol. Life Sci. 2004; 61: 682-699Crossref PubMed Scopus (457) Google Scholar) with the primary exception of two segments that undergo regulatory phosphorylation. One of these segments, which we label the regulatory insertion (RI), is an ∼30-residue insert between the first two β-strands in the ABCβ subdomain. The other segment, which we label the regulatory extension (RE), occurs at the C terminus of mNBD1 where it extends ∼20 residues longer than canonical ABC domains. The RI and RE protein segments both have elevated B-factors in crystal structures of mNBD1, suggesting that they might be conformationally dynamic as would be required to allow formation of the heteromeric NBD1/NBD2 ATP-sandwich complex believed to important in CFTR channel gating (13Lewis H.A. Buchanan S.G. Burley S.K. Conners K. Dickey M. Dorwart M. Fowler R. Gao X. Guggino W.B. Hendrickson W.A. Hunt J.F. Kearins M.C. Lorimer D. Maloney P.C. Post K.W. Rajashankar K.R. Rutter M.E. Sauder J.M. Shriver S. Thibodeau P.H. Thomas P.J. Zhang M. Zhao X. Emtage S. EMBO J. 2004; 23: 282-293Crossref PubMed Scopus (338) Google Scholar). We hypothesized that phosphorylation might control channel activation by altering their conformational preferences and thereby modulating steric interference with formation of the NBD1/NBD2 ATP-sandwich complex. Here, we report the production of soluble and monomeric domains of human NBD1 (hNBD1) both without and with the ΔF508 mutation for crystal structure determination and biophysical folding studies. We observe no significant differences in the folding properties of the two versions of hNBD1 but do observe changes in the surface at the putative site of interaction with MSD1. Therefore, our results provide new insights into the molecular pathology of the predominant disease-causing cystic fibrosis mutation, which is of particular relevance to drug discovery efforts. Cloning, Protein Expression, and Purification—hCFTR NBD1 (residues 389-673 and 389-678, both with and without the ΔF508 mutation; see Fig. 1A and Table I) was expressed in Escherichia coli as an N-terminal, His6-Smt3 fusion protein (see Ref. 13Lewis H.A. Buchanan S.G. Burley S.K. Conners K. Dickey M. Dorwart M. Fowler R. Gao X. Guggino W.B. Hendrickson W.A. Hunt J.F. Kearins M.C. Lorimer D. Maloney P.C. Post K.W. Rajashankar K.R. Rutter M.E. Sauder J.M. Shriver S. Thibodeau P.H. Thomas P.J. Zhang M. Zhao X. Emtage S. EMBO J. 2004; 23: 282-293Crossref PubMed Scopus (338) Google Scholar). Cells were grown overnight at 20 °C, harvested by centrifugation, and lysed by sonication on ice. Recombinant protein was initially purified by nickel ion affinity chromatography followed by removal of the tag using Ulp1 protease (15Li S.J. Hochstrasser M. Nature. 1999; 398: 246-251Crossref PubMed Scopus (605) Google Scholar). The sample was then passed through an S200 gel filtration column followed by a second nickel ion affinity column to remove residual His6-Smt3 tag. The protein was concentrated to 5-20 mg/ml in buffer containing 10% glycerol, 10% ethylene glycol, 100 mm arginine, 2-5 mm ATP, 3-7.5 mm MgCl2, 2 mm TCEP, and 50 mm Tris, pH 8.0. As reported for mNBD1, the ATPase activity of hNBD1 was indistinguishable from background in the coupled pyruvate kinase/lactase dehydrogenase assay (data not shown).Table IHuman NBD1 proteins Thermodynamic values are listed for those proteins analyzed in equilibrium denaturation experiments. **, proteins for which crystal structures were determined and reported herein.a Mutations in blue indicate human to mouse sequence changes, in green human to fish, and in red the three suppressor mutants.b High yields are >2 mg of purified, Smt3-cleaved protein p/g of fermentation pellet, medium is 0.7–2.0 mg/g, low is 0.1–0.6 mg/g, and none is 2 mg of purified, Smt3-cleaved protein p/g of fermentation pellet, medium is 0.7–2.0 mg/g, low is 0.1–0.6 mg/g, and none is 1.0). The refined model of hNBD1-2b-F508A includes residues 388-411 and 429-671 in molecule A; 389-413, 429-532, 539-541, and 547-671 in molecule B; 389-415 and 429-671 in molecule C; 388-410 and 426-672 in molecule D; 389-411 and 429-671 in molecule E; 1 ATP/protomer and 704 water molecules. The refined model of hNBD1-7a-ΔF508 includes residues 391-402 and 438-675, 1 ATP, and 135 water molecules. (see Table II for refinement statistics.) Equilibrium Denaturations—Experiments were performed as described by Pace et al. (20Pace C.N. Shirley B.A. Thompson J.A. Creighton T.E. Measuring the Conformational Stability of a Protein. IRL Press, New York1989: 311-330Google Scholar) using 20 mm sodium-potassium phosphate, pH 8.0, and 1 mm TCEP as the protein buffer. Proteins were denatured in 7 m urea that contained phosphate buffer. Samples were mixed by inversion and equilibrated at room temperature for 2 h prior to spectroscopic measurements. Circular dichroism measurements at 222 nm for equilibrium denaturation experiments were obtained using an AVIV 215 spectropolarimeter at 22 °C and a 0.1-cm quartz cuvette. Equilibrium denaturation data were analyzed with a mathematical model assuming a 2-state unfolding/refolding mechanism using all of the experimentally determined unfolding data. Nonlinear least-squares fitting was done using Kaleidagraph v.3.52 (Synergy Software, Reading PA). Tryptophan Fluorescence—Fluorescence emission for kinetic refolding measurements was monitored using a Cary Eclipse scanning spectrofluorometer with a 1-cm small volume quartz cuvette. The excitation wavelength was 295 nm, and the emission wavelength was 324 nm with a 5-nm slit width. All measurements were made at ambient temperature. Refolding Kinetics—Unfolded stocks of protein (4 mg/ml) were made at 20-times the desired final protein concentration in the refolding measurement and allowed to unfold for 2 h at room temperature in 7 m urea. Refolding was initiated by diluting the unfolded stock 20-fold into buffer lacking urea. Refolding as a function of time was followed either by CD at 222 nm or by tryptophan fluorescence as described above. The resulting data were analyzed by a simple single exponential decay using Kaleidagraph v.3.52. Solubility-enhancing Mutations Yield Well Behaved Forms of hNBD1 and the ΔF508 Mutant—Our work on mNBD1 established the boundaries of the first ABC fold within the sequence of CFTR and provided soluble murine protein. However, equivalent domains in hCFTR could not be recovered in soluble form following expression in E. coli despite substantial expression levels. Isolated NBDs from ABC transporters have frequently exhibited low solubility when expressed in the absence of their cognate TM domains. These solubility limitations have been shown, in some cases, to be caused by a tendency of the protein to self-associate and precipitate in the native conformation rather than from instability in protein folding (21Yuan Y.R. Blecker S. Martsinkevich O. Millen L. Thomas P.J. Hunt J.F. J. Biol. Chem. 2001; 276: 32313-32321Abstract Full Text Full Text PDF PubMed Scopus (224) Google Scholar, 22Smith P.C. Karpowich N. Millen L. Moody J.E. Rosen J. Thomas P.J. Hunt J.F. Mol. Cell. 2002; 10: 139-149Abstract Full Text Full Text PDF PubMed Scopus (681) Google Scholar). Therefore, we expected that engineered hNBD1 proteins with improved solubility could be produced by introducing mutations in surface residues to make them more hydrophilic. Many of the changes that yielded soluble protein were to residues that are naturally present in mNBD1. We also investigated the effect of three suppressor mutations (G550E, R553Q, R555K) that have been observed to improve in vivo trafficking efficiency of STE6-CFTR chimeras containing the ΔF508 mutation expressed in yeast (23Teem J.L. Berger H.A. Ostedgaard L.S. Rich D.P. Tsui L.C. Welsh M.J. Cell. 1993; 73: 335-346Abstract Full Text PDF PubMed Scopus (142) Google Scholar, 24Teem J.L. Carson M.R. Welsh M.J. Receptors Channels. 1996; 4: 63-72PubMed Google Scholar, 25DeCarvalho A.C. Gansheroff L.J. Teem J.L. J. Biol. Chem. 2002; 277: 35896-35905Abstract Full Text Full Text PDF PubMed Scopus (86) Google Scholar). Although the mechanistic basis has not been characterized, these suppressor mutations could act in improving the folding efficiency, stability, or solubility of NBD1. Table I summarizes the yield of soluble hNBD1 protein obtained from expression constructs containing 1-7 potentially solubilizing mutations chosen based on natural variations in mouse or fish orthologs, either in the presence or absence of the ΔF508 mutation. Recombinant proteins harboring the F508A mutation gave higher yields than the equivalent proteins with phenylalanine at position 508, whereas constructs with the ΔF508 mutation consistently gave lower yields. Although the yield of purified protein from in vivo expression procedures can be influenced either by folding efficiency or solubility in its native conformational state, the data presented below indicate that the ΔF508 mutation produces minimal perturbation of the equilibrium folding properties of hNBD1 in vitro. We conclude, therefore, that the effect of the Phe-508 deletion on the yield of soluble protein is most likely attributable to reductions in protein solubility. This surface mutagenesis strategy, together with improvements in the purification procedure (see "Materials and Methods"), yielded large amounts of stable and soluble, monodisperse hNBD1 protein in buffers containing ATP. Static light-scattering measurements and elution profiles in gel filtration experiments show these proteins to be monomeric and monodisperse (data not shown). Importantly, their availability provides for the first time the opportunity to characterize the biochemical and biophysical properties of hNBD1 and the effect of the disease-causing ΔF508 mutation on the structure of this domain. Crystal Structure of hNBD1 Shows That Regulatory Protein Segments Adopt Multiple Conformations Altering Access to the Active Site—High-resolution diffraction data were obtained for hNBD1-2b-F508A, containing two solubilizing mutations (F429S and H667R) in addition to the F508A substitution (Table II). These mutations are either surface-exposed in the crystal structure or located in conformationally dynamic regions of the domain. One (F429S) participates in intermolecular packing interactions stabilizing the lattice (data not shown). With the exception of the RI and RE segments, the structure of hNBD1-2b-F508A closely matches that of mNBD1. Least squares superposition of the remainder of the F1-type core and ABCβ subdomains yields a 0.46-Å root mean square deviation (rmsd) for 127 C-α atoms, only slightly exceeding the 0.39-Å rmsd observed after superposition of the different molecules within the asymmetric unit of the crystal structure of hNBD1-2b-F508A. The ABCα subdomains are also similar in structure, with an rmsd of 0.71 Å for superposition of 49 C-α atoms, although it exhibits a 10° rotation in hNBD1 relative to the orientation observed in mNBD1. Even so, residue Gln-493 in the Q-loop, which is considered important for hydrolysis of the ATP, maintains hydrogen-bonding contact with the γ-phosphate of ATP. In contrast, dramatic differences are observed when comparing the conformations of the RI and RE in the crystal structures of hNBD1 and mNBD1 (Fig. 1B). Both segments undergo ∼180° reorientations. We concluded that these differences reflect the dynamic flexibility in the regulatory regions and that their exact conformations are determined by the different intermolecular packing in the two crystal structures. Significantly, the conformational change in RI observed in the structure of hNBD1 exposes the conserved aromatic residue Trp-401, and a canonical base-stacking interaction now occurs between its side chain and the adenine base of ATP as observed in almost all other NBD structures but not in that of mNBD1 (26Schmitt L. Tampe R. Curr. Opin. Struct. Biol. 2002; 12: 754-760Crossref PubMed Scopus (276) Google Scholar). Interactions with the phosphate group of the nucleotide are the same in both the hNBD1 and mNBD1 structures. Although the conformation of the regulatory segments observed in the crystal structure of hNBD1-2b-F508A would preclude formation of a canonical ATP-sandwich complex with NBD2 because of a steric overlap at the interface, their dramatic change in conformation compared with the crystal structure of mNBD1 confirms our prediction that these segments of NBD1 are conformationally dynamic. Presumably, additional conformational adjustments can occur to allow formation of an NBD1-NBD2 ATP-sandwich complex with canonical geometry, perhaps in a manner modulated by regulatory phosphorylation at the known protein kinase A sites in these segments. Crystal Structure of ΔF508 hNBD1 Shows Minimal Conformational Changes but Substantive Changes in Surface Topography at the Putative Site of MSD1 Interaction—Crystals diffracting to a resolution of 2.3 Å were obtained for hNBD1-7a-ΔF508, which contains seven mutations (F409L, F429S, F433L, G550E, R553Q, R555K, H667R) in addition to the deletion of Phe-508 (see Table II). It shows only minor differences compared with hNBD1-2b-F508A except in the immediate vicinity of the deletion of Phe-508 (Fig. 2) and at the regulatory segments. Superposition of the F1-type core and ABCβ subdomains with those in hNBD1-2b-F508A gives an rmsd of 0.51 Å for 127 C-α atoms, similar to the deviations observed between the different protomers in that structure. The ABCα subdomain is rotated by 6° relative to its position in hNBD1-2b-F508A but is largely conserved in structure, exhibiting an rmsd of 0.87Å for the superposition of 49 C-α atoms. Furthermore, all of the contacts to bound ATP molecules are the same in the two hNBD1 crystal structures, independent of the presence of the ΔF508 mutation and the other solubilizing mutations. However, the position of α-helix 9b in the RE is very similar in the structures of hNBD1-7a-ΔF508 and mNBD1 (and different from the position observed in the structure of hNBD1-2b-F508A), suggesting that this may represent a preferred conformation of the dynamically flexible RE. The structure of hNBD1-7a-ΔF508 shows that deletion of Phe-508 is accommodated through a simple shortening of the loop connecting α-helices 3 and 4 in the ABCα subdomain (Fig.2, A-C). The superposition of the three available NBD1 crystal structures (mNBD1, hNBD1-2b-F508A, and hNBD1-7a-ΔF508) based on least squares alignment of α-helices 3 and 4 demonstrates that the conformation is extremely similar even in the immediate vicinity of the deletion, consistent with the quantitatively similar folding parameters observed either in the absence or presence of the ΔF508 mutation. However, the surface topography of NBD1 is dramatically altered at the site of the mutation (Fig. 2, compare D with E), which represents the presumed region of binding to the MSD1 of CFTR based on the interdomain interactions previously observed in the crystal structures of two ABC transporter integral membrane proteins (11Chang G. FEBS Lett. 2003; 555: 102-105Crossref PubMed Scopus (153) Google Scholar, 12Bass R.B. Locher K.P. Borths E. Poon Y. Strop P. Lee A. Rees D.C. FEBS Lett. 2003; 555: 111-115Crossref PubMed Scopus (24) Google Scholar). The deletion causes Gly-509 to move into the position occupied by Phe-508, resulting in a >90° rotation of the side chain of Val-510 as it moves into the center of the interhelical loop as well as reorganization of the backbone of residues 509-510 (including reorientation of their carbonyl groups). These changes result in a significant change in both the topography and chemical properties of the surface at this critical interdomain interaction site (Fig. 2, D and E). Thus, the structure of hNBD1-7a-ΔF508 suggests that interdomain interactions are likely to be substantially altered by the ΔF508 mutation. Of the seven solubilizing mutations present in the ΔF508 form of hNBD1, three (F409L, F429S, F433L) occur in disordered regions and therefore likely interact with solvent, whereas residue H667R is only minimally solvent-exposed on the surface of α-helix 9b in the RE. The three suppressor mutations (G550E, R553Q, R555K) occur either in or immediately following the LSGGQ signature sequence at the N terminus of α-helix 5. There are no backbone conformational differences between the three different NBD1 crystal structures in this immediate region (Fig. 1B), indicating that suppressor mutations do not change the tertiary structure of the domain. Furthermore, the side chains of these residues adopt similar conformations to those adopted by the wild-type residues in the other NBD1 structures (data not shown). Future study will be required to determine how these mutations improve the stability of NBD1 and the yield of native recombinant protein in vivo and whether these phenomena are related. However, the remote location of the suppressor mutations from the ΔF508 site (Fig. 1B) suggest that they do not influence the hNBD1 structure in the vicinity of the ΔF508 mutation. This inference is reinforced by the fact that from our in vitro measurements the ΔF508 mutation is not observed to significantly alter folding properties of the domain either in the absence or presence of the suppressor mutation (Table I). Minimal Perturbation of the Folding Properties of ΔF508 hNBD1—Equilibrium denaturation studies were conducted on several proteins available in sufficient yield. All unfold in urea with midpoints of 4-5 m (Table I). Equivalent folding isotherms were observed using either CD at 222 nm (Fig. 3) or intrinsic tryptophan fluorescence emission intensity (data not shown). Moreover, equivalent data were obtained from either denaturation or renaturation experiments (Fig. 3A), suggesting that the folding reaction is fully reversible. hNBD1-4-F508 serves as a reference protein in these studies. Fitting of its denaturation profile gives a standard free energy of folding (ΔG0) of about -9 kcal/mol in the absence of denaturant. Although hNBD1-4 variants contain four substitutions relative to the wild-type human sequence, all are naturally occurring variations found in wild-type mNBD1. hNBD1-2f-F508, which contains two different solubilizing mutations, displayed equivalent in vitro folding characteristics including a similar standard free energy of folding. In contrast, hNBD1-7a-F508, with three suppressor mutations added to the four solubilizing mutations of hNBD1-4, shows markedly greater free energy of folding (ΔG0) of about -11 kcal/mol (Table I). Urea denaturation experiments on hNBD1-2f and hNBD1-7a proteins, either in the presence of absence of ΔF508, show that this predominant cystic fibrosis-causing mutation does not cause any apparent change in the in vitro stability of hNBD1 either in the presence or absence of the suppressor mutations (Table I). A small 0.25 m decrease in the midpoint of the urea denaturation isotherms is observed for hNBD1-7a-ΔF508 as compared with hNBD1-7a-F508 but does not reflect a difference in the stability of the domain in the absence of denaturant as revealed in the ΔG0 calculations. A change of this kind is frequently observed when hydrophobic residues are removed from proteins because of a reduction in the compactness of the denatured state (27Shortle D. FASEB J. 1996; 10: 27-34Crossref PubMed Scopus (364) Google Scholar). It is notable that the same enhancement in stability that accompanies introduction of the three suppressor mutations into a Phe-508 protein occurs as well for a ΔF508 protein. Therefore, although the predominant disease-causing ΔF508 mutation does not appear to substantially alter the in vitro stability of NBD1, other mutations affecting the efficiency of CFTR biogenesis appear to do so. Preliminary kinetic data suggest that the refolding kinetics of hNBD1 are also unaffected by the introduction of the ΔF508 mutation. Either in the absence or presence of the mutation, refolding experiments monitored by intrinsic tryptophan fluorescence intensity show a fast phase (probably <100 ms) followed by a relatively slow single exponential phase with a half-time on the order of 1.5-2.5 min (Fig. 4, A and B). Preliminary measurements of refolding kinetics using CD spectroscopy show similar results, with 80% of the total CD change in either protein occurring in the dead time of manual mixing experiments and the bulk of the remaining change occurring with a half-time similar to that observed for the slow change in tryptophan fluorescence (Fig. 4, C and D). Analysis of the products of the fast refolding reaction using analytical gel filtration chromatography shows that with both protein variants ∼30% of the refolded protein is recovered on the column with half migrating at the monomer position and the other half migrating as an aggregate in the void volume (data not shown). Preliminary investigations indicate changes in some folding parameters in the presence of Mg-ATP or upon alteration of buffer constituents, but all protein constructs have shown equivalent behavior either in the presence or absence of the ΔF508 mutation in all assays. Further experimentation will be required to explore the thermodynamic and kinetic effects of potential parallel aggregation pathways and the effects of various ligands on the folding of hNBD1. We report the engineering of biochemically well behaved proteins containing the entirety of NBD1 from human CFTR. Their availability provides the opportunity to conduct structural and biochemical/biophysical studies on the domain containing the predominant disease-causing ΔF508 mutation. The altered topography of the ΔF508 mutant at the putative site of interaction with MSD1 suggests that interdomain interactions are likely to be substantially different in CFTR molecules harboring this ultimately lethal molecular defect. Disrupted interdomain interactions could play a role in altered gating properties that have previously been reported for ΔF508 CFTR channels (3Cheng S.H. Gregory R.J. Marshall J. Paul S. Souza D.W. White G.A. O'Riordan C.R. Smith A.E. Cell. 1990; 63: 827-834Abstract Full Text PDF PubMed Scopus (1423) Google Scholar, 4Dalemans W. Barbry P. Champigny G. Jallat S. Dott K. Dreyer D. Crystal R.G. Pavirani A. Lecocq J.P. Lazdunski M. Nature. 1991; 354: 526-528Crossref PubMed Scopus (569) Google Scholar, 5Lukacs G.L. Chang X.B. Bear C. Kartner N. Mohamed A. Riordan J.R. Grinstein S. J. Biol. Chem. 1993; 268: 21592-21598Abstract Full Text PDF PubMed Google Scholar). Additionally, they might be involved in the impaired intracellular trafficking observed for ΔF508 CFTR either because of enhanced sensitivity to digestion by endogenous proteases during biogenesis in the ER (9Zhang F. Kartner N. Lukacs G.L. Nat. Struct. Biol. 1998; 5: 180-183Crossref PubMed Scopus (129) Google Scholar) or because of impaired export from the ER attributable to destabilized quaternary interactions (28Ellgaard L. Helenius A. Nat. Rev. Mol. Cell. Biol. 2003; 4: 181-191Crossref PubMed Scopus (1676) Google Scholar). This suggests that an ideal small molecule "corrector" of the ΔF508 mutation might serve to reinforce interdomain interactions in CFTR, particularly at the site of the ΔF508 mutation. Our biophysical studies of intact hNBD1 domains are consistent with the structural observations showing, at most, minor differences between proteins containing wild-type Phe-508 and those with the ΔF508 mutation. Although more thorough kinetic studies and a more rigorous characterization of folding yield need to be performed in the future, the structural and thermodynamic observations reported in this paper suggest that the ΔF508 mutation causes no substantive defect in the folding of NBD1. Our characterization, however, of the biophysical properties of the suppressor mutations in NBD1 indicates that they may indeed facilitate in vitro folding of NBD1 to a significant extent. Future work will be required to determine whether this folding effect is responsible for their activity in improving maturation and the transport efficiency of CFTR in vivo. However, cosmotropic small molecule agents that promote more efficient protein folding have been shown to enhance the yield of functional CFTR in the plasma membrane for both wild-type and ΔF508 chloride channels (29Sato S. Ward C.L. Krouse M.E. Wine J.J. Kopito R.R. J. Biol. Chem. 1996; 271: 635-638Abstract Full Text Full Text PDF PubMed Scopus (466) Google Scholar, 30Zhang X.M. Wang X.T. Yue H. Leung S.W. Thibodeau P.H. Thomas P.J. Guggino S.E. J. Biol. Chem. 2003; 278: 51232-51242Abstract Full Text Full Text PDF PubMed Scopus (76) Google Scholar). On this basis, factors promoting a more efficient folding of CFTR domains in general or NBD1 in particular might have a beneficial effect on the yield of functional ΔF508 CFTR in the plasma membrane even if the mutation does not impair the folding of NBD1. We thank Drs. M. Ashlock, S. K. Burley, W. A. Hendrickson, C. Kissinger, C. Penland, P. J. Thomas, and D. Wetmore for many useful discussions, and K. Bain, J. Koss, F. Lu, L. Smyth, and Drs. S. Antonysamy and S. Wasserman for their expert contributions toward the structural characterization of hNBD1. Use of the Advanced Photon Source was supported by the United States Department of Energy, Office of Science, and Office of Basic Energy Sciences, under Contract W-31-109-Eng-38.
Referência(s)