CFTR Mechanisms- Further Genetics PDF

Preview

Summary

Lecture, tutorial and textbook notes per lecture material (level 7)...

Description

PART 3 (slides 31-39) CFTR: evidence for NBD-dimerization driven opening These slides describe experiments we carried out to verify whether a molecular mechanism involving NBD dimerization, could be used by the CFTR channel, just as it is used by other members of the ABC transporter family. We looked closely at prokaryotic ATP-bound crystals. We noticed a H-bond spanning the composite ATP binding site, from the head of one monomer to the tail of the other. Conservation in both regions is high so we could easily identify the corresponding residues in composite site 2 of CFTR: the acceptor is equivalent to a threonine (T1246) in the head of NBD2; the donor is equivalent to an arginine (R555) close to the signature sequence in the NBD1 tail. Therefore, an equivalent H-bond could well span CFTR’s composite site 2. How can we study whether two positions are interacting in a functional protein? One way is by constructing a mutant cycle (slides 33-34). The wild-type (WT), the two single mutants (one at the donor, R555K, and one at the acceptor position, T1246N) and the double mutant (R555K/T1246N) form the corners of a thermodynamic cycle. One can measure the changes (“s”) each mutation causes. If the 2 target residues are not interacting, then the effects of mutating e.g, T to N should not depend on whether the other residue is R or K, i.e. effects should not depend on whether the mutation is done in WT or mutant background. Therefore 3=1 and 3-1= 0. However, if the two residues indeed form a H-bond in WT, the effects of mutating the acceptor from T to N will depend strongly on whether the donor is R or K. In this case 3 will not be equal to 1. The difference ( 3-1) can be used as a measure of energetic coupling between the two target residues ( Gint). In other words, a difference between mutation-linked changes on parallel sides of a mutant cycle indicates – and to some extent quantifies – energetic coupling between the two target residues. Several kinetic parameters can be used to characterize WT and mutants, and to obtain useful mutation-linked changes (“s”) at different stages of the gating cycle. We first used the energetics of ATP binding to probe the interaction between the two target sites in the states that bind the ATP (we know that ATP binds to closed channels: see slide 13). Because the opening step is very slow compared to binding/unbinding steps, the binding step is likely to reach a steady state that is not too far from equilibrium. Therefore the apparent dissociation constants, obtained from ATP-dependence of opening rate, are reasonable estimates of the dissociation constant in the closed states preceding opening. We therefore used apparent affinity for ATP to characterize WT and each mutant, and build mutant cycles. For R555K, there is virtually no effect: half maximal opening rate is obtained with 50 M ATP, like for WT. However, after introducing the T1246N mutation, opening rate at 50 M ATP is much less than half maximal. Dose-response curve (opening rate vs [ATP]) is shifted to higher [ATP]. Introducing the same T to N mutation in a background already containing the R555K mutation shifts the dose-response curve to overlap the curve for the single T1246N mutant. The effects on parallel sides of mutant cycle are similar in this case yielding a coupling energy that is not significantly different from zero. We then used the free energy of activation for the opening reaction (G‡opening), to probe the interaction between the two target sites in states approaching the open state. The single R555K mutation slows opening, (=increases closed dwell-time); the single T1246N mutation also slows opening. Both single mutations increase the energetic barrier the protein needs to overcome to open. But introducing the T to N mutation in a background already mutated at the donor position (R555K) partially restores fast opening (double mutant has shorter closed dwell-time). Different

effects of the mutation on parallel sides of the mutant cycle indicate energetic coupling. The negative sign of the coupling energy is consistent with R and T forming a stabilizing H-bond in the transition state for the opening reaction, a stabilizing bond that was not there in the closed ground state preceding opening. We have repeated a similar mutant cycle analysis to study how interactions between these two residues change during the C1O1 transition, and the results are consistent with the hydrogen

bond being present in the open ground state too (not just the transition state for opening), but not in the closed state preceding opening. In summary, we found that the residues show a statedependence of energetic coupling: they are not coupled in the closed states that bind ATP, but they are coupled in the open state and in the transition state approaching the open state. The results support a molecular mechanism in which opening of the channel pore is coupled to the ATPase cycle via formation of NBD dimers.

PART 4 (slides 40-48) ABC multidrug transporters: possible structural underpinning for polyspecificity A few efflux transporters of the ABC transporter superfamily (ABCB1, ABCG2, ABCCs) are very important in mediating the movement through the body of therapeutic drugs. These transporters, located in the apical plasma membrane of several “barrier” epithelia (gut, liver, bloodbrain barrier, placenta), catalyze active, ATP-dependent export of drugs. They recognize and transport many structurally diverse compounds (this characteristic is known as “polyspecificity”), indeed they interact with most of the drugs in use today. Therefore, they play important roles in determining the pharmacokinetic profile of drugs. Here, we will concentrate on their molecular structures and how they work. First we will focus on the TMDs (slide 42). These domains provide the drug binding site and conformational changes alternatively expose the binding site to the cytosol or the external solution (alternating access). These conformational changes also alter how tightly the drug binds to the transporter. A complete export cycle includes a minimum of 4 steps: 1) drug enters binding cavity from inside, binds to high affinity site 2) conformational change: cavity exposed to outside + drug binding site has lower affinity 3) drug is released to outside 4) empty TMDs reset: return to original inward-facing conformation, with high affinity binding. In the absence of an energy input, microscopic reversibility would ensure that cycles in the clockwise direction are as frequent as cycles in anticlockwise direction (i.e. no net transport of drug across the membrane occurs). But a free energy releasing process (ATP binding and hydrolysis catalyzed by the NBDs) is coupled to the transport process, resulting in more frequent clockwise cycling (= net transport of drugs out of cell). The exact way conformational changes in NBD and TMD are coupled is still unclear (and probably is different for different ABC transporters). We know that ATP binding favours NBD dimer formation and ATP hydrolysis triggers dimer dissociation. Consistent with recent cryo-EM structural data, flipping of drug-binding site from inward to outward exposure (and from high to low affinity) is driven by ATP binding and NBD dimer formation; while resetting of the empty transporter to the inward facing conformation is driven by ATP hydrolysis and NBD dimer dissociation (slide 43). Thus a shared molecular mechanism couples ATPase cycles of ABC proteins to distinct TMD conformational changes. In CFTR TMD movements result in channel opening and closing. In transporters TMD movements result in alternate exposure of binding site.

The most striking characteristic of these transporters (here we use ABCB1/P-glycoprotein as an example) is the diversity of transported substrates: structurally very diverse organic molecules, ranging in size from less than 200 Da to ~ 2kDa. Many contain aromatic groups, but aliphatic linear or circular molecules are also transported. Most substrates are quite hydrophobic. Indeed, the only common denominator identified so far for ABCB1/PgP substrates is their amphipathic nature, suggesting that they probably concentrate in membrane leaflets. Several crystal structures of ABCB1 have been solved, at low resolution, in the inward facing conformation, with no ATP bound at the NBDs. One is shown in slide 46 (on the left). The NBDs are far apart, the TMDs form a chamber, open to the cytoplasm but also accessible from the inner membrane leaflet, which allows access to a hydrophobic (hydrophobic=red; hydrophilic=blue) binding surface in the protein, approximately located at the centre of the membrane. In ABCB1 in the outward facing conformation (slide 47 right) the NBDs are tightly dimerized and the TMDs form a chamber giving access to the outer membrane leaflet. For ABCB1/P-gP drugs probably reach and leave the binding site from the membrane leaflets. But how can a single protein recognize so many different drugs? One possibility is that many hydrophobic ligands (Phe, Val, Leu, Ile) are redundantly available for drug binding. Hydrophobic interactions are less specific than hydrogen bonds or salt bridges, having no directional or sign requirements. Different drugs could contact different sets of ligands. Another possible explanation resides in the unusual flexibility in the TMDs of ABCB1. The different crystal structures solved have a range of distances between the NBDs, and EPR spectroscopy has confirmed that this occurs also in vivo. Slide 48 illustrates the results of a molecular dynamics (MD) study. These are computer simulations, using the atom coordinates from a crystal structure and mathematical equations describing forces between atoms to constrain simulations. Such simulations can model how the protein structure changes when subjected to thermal motions. Concentrating on the TMD regions, Wen et al (4) noted a great conformational flexibility in the backbone, with bending and kinking occurring continuously in the -helices, especially around the large number of glycines (at glycines, breakage of the -helix is not energetically costly). Non- -helical regions of the TMs are present in various experimental structures (slide 49). Therefore ABCB1 might “recognize” so many different drugs, as different drugs can contact many different hydrophobic ligands in the transporter, and these ligands can assume many different positions. Thus redundancy and flexibility of available drug-transporter contacts might underpin broad substrate specificity.

PART 5 (slides 51-59) Molecular pharmacology of small molecules targeting CFTR Mutations in the gene encoding CFTR cause cystic fibrosis (CF). Thousands of different mutations are known to cause CF. The most common mutation is deletion of F508 (F508del) present in ~90% of patients. F508del-CFTR misfolds, is retained in the ER and degraded by the ubiquitin/proteasome system. In addition, F508del-CFTR molecules that do reach the plasma membrane show greatly impaired channel function and reduced stability there. Other mutations, such as G551D, present in < 5% of the CF population, do not impair processing and only affect the ion channel function.

Recently, compounds have been identified which bind to mutant CFTR and partially rescue their defects. Compounds targeting CFTR fall in two classes: potentiators, capable of increasing channel activity and correctors, increasing plasma membrane density. Potentiator drug VX-770 causes a dramatic improvement in the health of patients carrying gating mutations (e.g. G551D) demonstrating that restoring CFTR-mediated anion permeability through pharmacological means ultimately results in significant clinical benefit. A cryo-EM structure of VX-770-bound CFTR has recently been published. However the drug’s mechanism of action remains unclear. However, no single drug identified so far can effectively treat CF caused by the F508del mutation. VX-770 and leading corrector VX-809 proved ineffective, alone, and only marginally effective in combination (approved in the US since 2015, very recently approved in UK). Small molecules with more powerful effects on CFTR are currently being sought intensively....