Cellular and Molecular Life Sciences

, Volume 66, Issue 21, pp 3469–3486

Molecular models of the open and closed states of the whole human CFTR protein


    • IMPMC, UMR7590, CNRSUniversités Pierre et Marie Curie-Paris 6 et Denis Diderot-Paris 7
  • Pierre Lehn
    • INSERM U613, IFR148 ScInBioSUniversité de Bretagne Occidentale
    • IMPMC, UMR7590, CNRSUniversités Pierre et Marie Curie-Paris 6 et Denis Diderot-Paris 7
Research Article

DOI: 10.1007/s00018-009-0133-0

Cite this article as:
Mornon, J., Lehn, P. & Callebaut, I. Cell. Mol. Life Sci. (2009) 66: 3469. doi:10.1007/s00018-009-0133-0


Cystic fibrosis transmembrane conductance regulator (CFTR), involved in cystic fibrosis (CF), is a chloride channel belonging to the ATP-binding cassette (ABC) superfamily. Using the experimental structure of Sav1866 as template, we previously modeled the human CFTR structure, including membrane-spanning domains (MSD) and nucleotide-binding domains (NBD), in an outward-facing conformation (open channel state). Here, we constructed a model of the CFTR inward-facing conformation (closed channel) on the basis of the recent corrected structures of MsbA and compared the structural features of those two states of the channel. Interestingly, the MSD:NBD coupling interfaces including F508 (ΔF508 being the most common CF mutation) are mainly left unchanged. This prediction, completed by the modeling of the regulatory R domain, is supported by experimental data and provides a molecular basis for a better understanding of the functioning of CFTR, especially of the structural features that make CFTR the unique channel among the ABC transporters.


Cystic fibrosisStructureModelChannelP-gp

Supplementary material

18_2009_133_MOESM1_ESM.tif (1000 kb)
Figure S1: Modeling of the inward-facing conformation of CFTR (closed channel). The passage from the MSD configuration found in the MsbA closed apo model to a “tight” closed apo conformation, likely representing the closed form of the CFTR channel, is illustrated here. This occurs through a small rigid body rotation of the two MSDs around the ICLs pivots. (TIFF 999 kb)
18_2009_133_MOESM2_ESM.tif (433 kb)
Figure S2: Superimposition of the MsbA open apo conformer with the tight conformation of the CFTR inward-facing model, in the vicinity of the channel entrance (extracellular side) and showing the transmembrane helices. The transmembrane helices of the MsbA open apo conformer are depicted in grey and are labeled TM1 to TM12, whereas those of the tight conformer (colored) are labeled TM1t to TM12t. One can note the good overall correspondence between the two conformers, noticeably for helices TM6 and TM12 at the center. (TIFF 433 kb)
18_2009_133_MOESM3_ESM.tif (740 kb)
Figure S3: Detailed views of the CFTR chloride channel in the “tight”, inward-facing conformation. A) Side view (left) and top view (right); B) Stereo views from the top. The orientations are similar to those shown in Figure 3. The potential chloride translocation pathway is symbolized by colored spheres (the colors ranging from blue to red while moving from the intracellular side to the extracellular side). (TIFF 740 kb)
18_2009_133_MOESM4_ESM.tif (517 kb)
Figure S4:The R2 domain: Sequence alignment deduced from the HCA comparison of the C-terminal part of the CFTR R domain (here named the R2 domain), preceding the second membrane-spanning domain (starting at transmembrane helix TM7), with the N-terminal tail of the same protein, preceding the first membrane-spanning domain (MSD1, starting at transmembrane helix TM1). The CFTR sequences of four species are shown: Homo sapiens (h, genbank identifier (gi): 147744553), Salmo salar (s, gi:12746235), Danio rerio (d, gi:117380075), Fundulus heteroclitus (f, gi:3015540), illustrating the presently known most divergent CFTR sequences. At top are presented a few N-terminal sequences of other ABC proteins, sharing significant similarities with the CFTR N-terminal sequence (hABCCB: Homo sapiens ABCC11, gi:74762666; hMRP9: Homo sapiens MRP9, gi:161788999; hMRP5: Homo sapiens MRP5, gi:8928547; dGD10651: Drosophila simulans GD10651, gi:195581868, hMRP4: Homo sapiens MRP4, gi:206729914). Between the CFTR N-terminal and R2 sequences are shown the N-terminal sequence of Saccharomyces cerevisiae YOR1 (yYOR1, gi:1730876) and the R sequence of Homo sapiens MRP3 (hMRP3: gi:6920069), which provide intermediate links. The alanine content of the region aligned with the N-ter helix is in good agreement with such a secondary structure. Conserved hydrophobic amino acids (V, I, L, F, M, Y, W) are colored dark green, and residues that can substitute them (A, C, T) are colored light green. Loop-forming amino acids (P, G, D, N, S) are colored yellow. Conserved aromatic residues (Y, W, F) are colored purple, conserved basic and acidic residues are colored blue and pink, respectively. The secondary structures predicted using HCA for the associated hydrophobic clusters (1 stands for V, I, L, F, M, Y, W; 0 for other residues (Eudes et al., BMC Struct Biol (2007) 7: 2) are indicated above and below the CFTR N-terminal and R2 sequence, respectively. Black stars indicate the phosphorylation sites. The amino acid ranges are shown at the end of the sequences. Below the alignment are shown the secondary structure propensities estimated for the R2 domain, based on NMR experimental data, for the non-phosphorylated and phosphorylated forms of CFTR (based on the data presented in Figure 2 of the Baker’s article (Baker et al., Nat Struct Mol Biol (2007) 14: 738-745). (TIFF 517 kb)
18_2009_133_MOESM5_ESM.tif (357 kb)
Figure S5: 3D structures of the regulatory domains (TOBE domains) of three ABC proteins. From left to right: the Pyrococcus horikoshii multiple sugar binding transport ATP-binding protein (pdb 2d62), the Archaeoglobus fulgidus molybdate/tungstate ABC transporter ModC (pdb 3d31) and the Sulfolobus solfataricus glucose transporter (pdb 1oxs). Strands constituting the “Greek-Key” motifs are labeled according to Figure S6. The fifth strand of each TOBE domain is depicted in pink. The N- and C-termini of the regulatory domains are indicated. The WO4 groups, which bind to the ModC regulatory domains and allow a trans-inhibition of the molybdate/tungstate ABC transporter, are shown. (TIFF 356 kb)
18_2009_133_MOESM6_ESM.tif (407 kb)
Figure S6: The R1 domain:Top) Sequence alignment deduced from the HCA comparison of the N-terminal part of the CFTR R domain (here named the R1 domain) with the regulatory domains found in some bacterial ABC transporters immediately after their NBDs (TOBE domains). The sequences of six of these proteins are shown, distributed into two subfamilies. The similarities between amino acids are reported as in Figure S4, and phosphorylation sites are indicated by black stars. Grey stars indicate the C-terminal amino acid. The position of the regulatory extension (RE), observed in the NBD1 crystal structure, is indicated. Bottom) Secondary structure propensities estimated for the R1 domain, based on NMR experimental data, for the non-phosphorylated and phosphorylated forms of CFTR (based on the data presented in Figure 2 of the Baker’s article (Baker et al., Nat Struct Mol Biol (2007) 14: 738-745). (TIFF 407 kb)
18_2009_133_MOESM7_ESM.tif (197 kb)
Figure S7: Comparison of the HCA plots of the human CFTR R domain (aa 610 to 885) and of the yeast Yor1p corresponding region (aa 775 to 925). The sequence is shown on a duplicated alpha-helical net, and the strong hydrophobic amino acids (V I L F M Y W) are contoured: these form hydrophobic clusters, which mainly correspond to regular secondary structures. Guidelines to the use of the method can be found in Gaboriaud et al.FEBS Lett (1987) 224: 149-155 and in Callebaut et al. (1997) Cell Mol Life Sci53: 621-645. The predicted secondary structures and domain limits are reported above the human CFTR plot. Green colors highlights conservation of hydrophobic amino acids (V I L F M Y W), whereas orange is used for correspondence between hydrophobic amino acids (V I L F M Y W) and amino acids (A C T), which can substitute them in a context-dependent manner. Sequence identities are colored pink. (TIFF 197 kb)
18_2009_133_MOESM8_ESM.tif (1.9 mb)
Figure S8: A complement to Figure 6, showing the NBD1:NBD2 assembly from bottom (on the R1 sub-domain side) and from top (from the interface with membranes). (TIFF 1914 kb)
18_2009_133_MOESM9_ESM.tif (3.4 mb)
Figure S9: Putative dimeric architecture of CFTR (closed, inward-facing conformation): an alternative solution to that proposed in Figure 5. The view is shown in a similar orientation as in Figure 5. The interface between each monomer is built of the opposite and symmetric long faces, as in Figure 5 (compare the bottom views, where CFTR is seen from the extracellular side). Main contacts are between TM4 and TM4, αα2 of NBD2 and R2. However, in this alternative arrangement, R1 does not contact R1 (in the context of the proposed location of this domain). (TIFF 3510 kb)
18_2009_133_MOESM10_ESM.tif (1.2 mb)
Figure S10: Possible concerted movements of the NBD1/NBD2 domains and the R1 and R2 parts of the regulatory domain R, between the inward-facing conformation (closed state) at left and the outward-facing conformation (open state, in which the bound nucleotides are shown). The NBD1 (blue) being fixed, the NBD2 domain rubs, while moving, the R2 and R1 segments and may interact more tightly with phosphorylable segments (S660, S737, S795 and S813 are labeled in red to recall their importance (Hegedüs et al. Biochim Biophys Acta (2009) 1788: 1341-1349)) in the channel open-state through specific basic patches. To that aim, R1 (in white) may rotate around its N- and C-terminal parts toward a position (in yellow) close to the NBD2 domain (in orange). See Fig. S11 for the details of amino acids in the R1 region for the outward-facing conformation. Red stars indicate the possible area of interaction. One can note that no phosphorylable serine is present after S813 within the R2 region which provides less contact with NBD2. (TIFF 1261 kb)
18_2009_133_MOESM11_ESM.tif (790 kb)
Figure S11: Zoom on the model of the R1 domain (outward-facing conformation), in a position where it begins to interact with the two basic specific patches including K1429, R1434, R1438 (NBD2) and K1457, K1459, K1461 (C-ter domain). (TIFF 789 kb)
18_2009_133_MOESM12_ESM.tif (1.3 mb)
Figure S12: Putative alternative pathway towards the entry of the CFTR channel pore through the NBD1:NBD2 interface, leading to the bundle of intracellular loops (outward-facing configuration (Mornon et al. (2008) Cell Mol Life Sci65: 2594-2612). A) Global view with, at the center in the middle distance, the cytoplasmic ends of ICL1 and ICL2 shown in dark blue and ICL3 and ICL4 shown in red. The R1 domain, which is likely situated in front of this view, is not shown. The R1 domain, with the position shown in Figure S10, however does not impede access to this pathway. B) Closer view, in the same direction as in panel A, but showing only the putative conduction pathway of the NBD1:NBD2 interface. Anions might be attracted by the basic residues K606 and R1403 and then transported through a small pore including G576 and A1374, which are 5.1 Å apart in the model, within the Walker B loops (βc4 -αc4-5 loop). They might leave the NBD1:NBD2 interface at the levels of the Q loops (Q493 and Q1291, which are 5.6 Å apart from each other). Basic residues are shown in blue, hydrophobic ones in green and hydrophilic ones in pink. Proline residues are shown in light yellow. A1374 and G576 are shown as van der Waals surfaces, with between them a chloride ion. C) Same view as in Figure 7B, but with an alternative position of E267 at the foreground, which may weakly interact with K968 in one of its possible location. (TIFF 1.33 mb)
18_2009_133_MOESM13_ESM.ppt (528 kb)
Table 1: Comparison of the amino acid content of R “linkers” from some ABC proteins, in the increasing order of the percentage in strong hydrophobic amino acids (V, I, L, F, M, Y, W). l indicates the length of the considered segment (in amino acids). The absolute numbers (nH) and percentages (%H) in hydrophobic amino acids (V, I, L, F, M, Y, W) are given in the following columns, as well as the absolute number (nL) of loop-forming amino acids (P, G, D, N, S) and the ratio between the number of hydrophobic and loop-forming amino acids (R = nH/nL). The CFTR R1 sub-domain, the CFTR R1-R2 and the CFTR R2 sub-domain have values between those encountered for typical disordered regions (e.g. the R region observed in the crystal structure of MDR/P-gp (Aller et al., Science (2009) 323: 1718-1722) and for canonical, folded globular domains (e.g. the NBD1 and NBD2 domains, which have 33 % of strong hydrophobic amino acids). R1 is here considered to start at aa 646 (after the helix α8 observed in the crystal structure of Glcv, see Figure S6) and R2 to end at amino acid 840, before the so-called N-helix, see Figure S4). The two last lines of the table refer to the R domains of ABCA1, the phosphorylation of which also regulating the protein function (Roosbeek et al., J Biol Chem (2004) 279: 37779-37788). (PPT 528 kb)

Copyright information

© Birkhäuser Verlag, Basel/Switzerland 2009