BACK TO DIAGRAM CFTR REVIEW PAGE
TRANSMEMBRANE HELIX 1
(amino acids 81 thru 102)
FFWRFMFYGIFLYLGEVTKAVQPLLL
Ion channels form pores in membranes. The two basic methods employed by ion channels to form pores are: 1) formation of transmembrane helices (most likely the case with CFTR), and 2) formation of beta-sheets (bacterial ion channels use this method).
Interactions can take place among different helices (interhelical interactions) of membrane proteins and usually take the form of polar-polar interactions. The amino acids Ser, Thr, Tyr, and Trp make up most of the hydrogen-bonding (~55%). Ser alone makes up 19%. It is likely that most helices form at least one hydrogen bond with another helix, and may provide specific orientations of helices and stabilize the protein structure. Kosynkina, U.Ill Chicago
The only known high-resolution x-ray crystal structure of an ABC transporter family member is of MsbA. It reveals a central chamber formed within the membrane by contacts between transmembrane helices 2 and 5. This is due the the fact that MsbA, like all bacterial ABC transporters functions as a homodimer (while CFTR functions as a single protein), with TM2 of MsbA contacting TM5 in the opposite half of the protein. These four transmembrane helices can be seen as forming a kind of "hinge" (2 each) on either side of the chamber. All 6 of the transmembrane helices in the MsbA monomer are tilted with respect to the plane of the membrane by between 30 and 40 degrees. It has been predicted by the authors that during gating of MsbA, the transmembrane helices either flip or twist to expose newly accessible areas within the chamber. The transmembrane helix in MsbA which corresponds to TM1 of CFTR is from amino acids 22-52. Chang and Roth, Science 293, 9/7/01 pgs 1793-1800.
PORE: Experimental evidence has been suggesting for some time that the pore of CFTR is made up of primarily membrane-spanning helices of the N-terminal half of the protein (i.e. TM1 thru 6). There is still no direct structure of the pore, however. CD spectroscopy has been done on 6 peptides (TM1-6). There appears to be a shift in secondary structure of TM6 from helix to beta-sheet in different organic solvents. These peptides have also been shown to form a chloride channel when added to pure phospholipid bilayers. ATP hydrolysis by the NBDs may be involved in changing the structure of the pore. The diameter of the pore is predicted to be 5.3 Angstroms as suggested by blocking experiments, but narrower near the cytoplasmic end of the channel. Studies using water-soluble MTS reagents which modify cysteine (cystine can be deliberately mutated into the predicted pore region) provide evidence that the primary site for determining anion selectivity occurs near cytoplasmic end of channel, and favors anions over cations by 25-fold. Evidence also suggests CFTR has a "multi-ion pore" which is able to accommodate more than one chloride ion at a time. Amino acid substitutions in TM1 and TM6 have been reported to cause changes in anion selectivity. CFTR may have a large inner vestibule for ion binding. In CFTR, 4 prolines are conserved across species and two are associated with CF. Note: prolines are important in membrane-spanning domains of transport proteins. Proline 99 in CFTR could be part of the pore. When mutated to leucine, there was no longer any selection for Cl- over I-. Sheppard and Welsh propose that P99 is involved in forming a kink in M1 in the channel, but doesn't itself line the pore. There are 4 predicted in the TMs: P99, P205, P324, P1021. Since mutants which lack the C-terminal half of the protein (no helices 7-12 or NBD2) appear nearly normal, they could be forming dimers with each other. CFTR appears selective for anions which are more easily dehydrated and follow the Lyotropic series. It is also capable of discriminating based on ion size. (ex: gluconate won't pass thru) . K95 and K335 were the first two mutations engineered into CFTR to show importance in anion selectivity. Changing them to negatively charged residues results in a preference for I- over Cl-. The apparent anion selectivity of wild-type CFTR exhibits an asymmetry that may involve either or both of these two residues. Normally, when I- is outside the cell, Chloride is preferred over Iodine (almost double) but when I- is inside, it is the preferred ion by about 1.5 times. Perhaps when iodine enters pore from outside, it binds to these lysine residues which are near the outer pore. The inner side of the chamber consists of a cluster of positively-charged amino acids (4 args, 2 lys), while the outer side is more hydrophobic. In this article, the authors speculate on possible mechanisms for lipid transport thru the chamber.
In the recently published structure of an ABC superfamily member, MsbA, it can be seen that there is a central chamber within the membrane bilayer. It should be remembered that MsbA is a lipid pump rather than an ion channel (like CFTR), however it is interesting to observe that its transmembrane helices, which form the chamber, are all tilted like a "teepee", with a tilt of between 30 and 40 degrees per helix. There are also two openings in the chamber which appear to allow lipid from the inner leaflet of the membrane into the chamber, persumably to be pumped out to the other side of the bilayer once the NBD domains hydrolyze ATP. Chang and Roth, Science 293, 9/7/01 pgs 1793-1800.
More Pore: Conduction and permeation are the two characteristics of CFTR which are most affected my mutations in the membrane spanning regions. It is still possible that 2 or more molecules of CFTR must come together to form the pore. However, there is no biochemical evidence for this, so it is not generally believed to multimerize. The "symmetry" of the GABA(A) ligand-gated chloride channel's pore appears to be missing with the pore in CFTR. The model which so far has described CFTR the best is one based on a channel with 3 binding sites for chloride. In addition to a large vestibule on the intracellular side of the channel, there is evidence that a second vestibule exists near the extracellular side, since large MTS reagents are accessible when added to the extracellular side. The narrowest part of the pore is most likely the part between the vestibules. It has been suggested that a "pore loop" formed by amino acids from helix 6 near the intracellular side is responsible for the apparent narrowing.
In October, 2001, McCarty and Zhang of Emory University reported on a study of the possible 3D structure of the pore of CFTR. They compared the relative importance of various sites previously studied and identified new sites that contribute strongly to anion selectivity, using chloride and substitute anions in oocytes expressing wild-type cystic fibrosis transmembrane conductance regulator or 12-pore-domain mutants, and determined relative permeability and relative conductance for 9 monovalent anions and 1 divalent anion. Their data indicate that ".. a region of strong discrimination resides between T338 and S341 in transmembrane 6, where mutations affected selectivity between chloride and both large and small anions. Mutations further toward the extracellular end of the pore only strongly affected selectivity between chloride and larger anions. Only mutations at S341 affected selectivity between monovalent and divalent anions. The data are consistent with a narrowing of the pore between the extracellular end and a constriction near the middle of the pore." Am J Physiol Lung Cell Mol Physiol 2001 Oct;281
Channel-lining water accessible residues have been found in transmembrane helices 1, 3, and 6 using the substituted cysteine accessibility method. Using the accessibility-accessibility method, it was found that three residues, G91, K95, and Q98 were lining the pore. The sequence suggests definite alpha-helical secondary structure.
Like all anion channels, CFTR doesn't discriminate very much against cations (~1:10). This is probably due to a higher hydration energy for water around anions than cations. Usually, CFTR conducts chloride at room temperature at around 8-10 pS. The pore of CFTR seems to allow large organic ions in when present at the intracellular side only. This could be due to a need for ATP hydrolysis. This indicates that the processes of gating and permeation may be linked in CFTR.
Anion selectivity has been shown by mutagenesis to be determined in part by Lysine95 in transmembrane helix 1 as well as Lysine 335 and 347 in helix 6. When these amino acids are mutated to either Asp or Glu, CFTR let pass Iodine rather than Chloride, however different investigators have gotten dissimilar results. When CFTR was first discovered in1989, the presence of positively charged residues in the predicted transmembane helices was taken as evidence CFTR could possibly be a chloride channel as well as a regulator.
There are 6 positively charged amino acid residues in the transmembrane helices of CFTR (K95 in M1, R134 in M2, R334 and K335 as well as R347 in M6, R1030 in M10) and they seem to be highly conserved between species.
Mutation of Lys95 and Lys335 (in TM6) to negatively charged amino acids changes CFTR anion selectivity from Br>Cl>I>F to I>Br>Cl>F.
Cysteine accessibility studies has shown that helices 1 and 6 line the pore.
The boundaries of the transmembrane helices are clearly marked by amino acids which have low propensity for alpha helix or beta sheet formation. The boundary amino acids tend to have a high propensity for beta turn or irregular secondary structures.
Amino acid sequence similarities among the ABC protein family are less conserved in the transmembrane helices.
A series of truncation mutants were constructed in 1995. It was found that CFTR missing the first 118 amino acids functioned similarly to wild-type, but had a smaller conductance and open probability. Mutants without helix one can therefore form chloride channels.
Anderson, in 1991, found that halide selectivity in CFTR pore is governed by positive charge(s) on TM1 as well as TM6.
Renthal, et. al (from U.Texas, San Antonio), following up on evidence that areas of transmembrane helices which make contact with other helices in the membrane tend to be smooth, while parts of helices contacting lipid membrane tend to be rough, developed a prediction method that indicates it may be possible to predict using only amino acid sequence which parts interact with membrane and which with other parts of the protein. They speculate that the entropy of the lipid acyl chains may be increased by interactions with the rough surface. Biophysical Society Meeting, 2001
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.
MOPS blocks CFTR channel depending on whether the channel is in a highly activated state or a lower activated state. This carries with it the important implication that the conformational state of the gating process helps determine permeation properties.
Relative permeability ratios of ions in CFTR indicate permeation is probably a combination of low and high field strength interactions. It is not simply a matter of solvation energies, even though they are involved at least partly. This is because of binding sites (probably 2 or 3 in pore) as well as CFTR being a multi-ion pore. It appears that residues in TM6 determine relative permeability sequences. Smith et al. uses block by gluconate anions to predict three or four energy barriers and two or three wells for CFTR's pore.