A typical human cell contains about 20.000 protein-encoding genes of which 26% are predicted to produce integral membrane proteins (1). These are embedded within the lipid bilayer of biological membranes and include ion channels, receptors and transporters. The abundance of membrane proteins within the human proteome reflects their crucial biological role. For example, membrane proteins are required for transport of ions and larger molecules, signal transduction, cell-cell communication, cell-cell interaction and to organize and maintain cellular shape. Despite the biophysical complexity of the lipid bilayer, integral membrane proteins are anchored into the membrane through two simple structural domains, namely: the 0α-helix bundle and β-barrel (2). β-barrel proteins are made up of antiparallel β-sheets folded into a cylindrical structure. These proteins are restricted to the outer membranes of Gram-negative bacteria, mitochondria and chloroplasts. However, helix bundle proteins (figure 1) are found in all biological membranes and are vastly more abundant than β-barrel proteins as well as structurally and functionally more diverse. These proteins are easily recognized by their transmembrane domains (TMDs) comprising a stretch of 20 hydrophobic residues such as leucine, isoleucine, alanine, valine and phenylalanine that form a tightly packed bundle in the membrane (figure 1). TMDs are not only responsible for membrane anchoring but they also mediate oligomerization as well as protein-protein interactions. The available crystal structures of helix bundle proteins reveal a variety of sizes and shapes, ranging from proteins with one TMD to a complex protein with a cluster of over 20 TMDs that are connected by loops and contain large soluble domains on either side of the membrane (2). While TMDs are primarily made up of hydrophobic residues, loops located on the cytoplasmic side of the membrane are enriched in positively charged residues which are typically absent from loops located on the extra-cytoplasmic side of the membrane. The preference for positively charged residues at the cytoplasmic face of the membrane is known as the positive inside rule and has been observed in nearly all helix bundle proteins irrespective of organism and membrane system (3). Based on their key biological role, it is not surprising that integral membrane proteins are also of medical importance. In fact, these proteins make up more than 60% of current drug targets with G protein-coupled receptors (GPCRs), ion channels and transporters as important examples (4). Moreover, dysfunctional membrane proteins are responsible for a range of diseases. In fact, many of the known ~16000 disease-linked missense mutations in protein sequences are found in membrane proteins with their TMDs as frequent targets for these mutations (5). Molecular analysis of a subset of these mutant proteins revealed that most of these mutations don’t affect function, but rather interfere with normal protein folding or trafficking (5). This results in decreased amounts of functional protein at the destination membrane, thereby ultimately causing disease. Thus, protein misassembly is an important factor underlying the aetiology of various heritable pathologies in humans. These include cystic fibrosis (CF), the most common lethal genetic disorder found in populations of northern European ancestry with a prevalence of 1 out of 2500 (6). CF is caused by mutations in the gene encoding the cystic fibrosis transmembrane regulator (CFTR). This is a chloride-conducting transmembrane channel which regulates anion transport and mucus clearance of the lungs. Its loss of activity results in mucus retention, chronic infection and subsequent inflammation that is damaging to the lungs (6). The development of potent therapeutics has been impaired by the absence of structural information. It should therefore be noted that high resolution structures of human CFTR with or without small molecule drugs were reported recently (7,8). These further the molecular understanding of CFTR and provide a basis for drug development. Here, I will discuss these structures as well as our current understanding of CFTR.
Role of CFTR in cystic fibrosis
It is now well established that CF is life threatening genetic disorder that is predominantly found in Caucasian populations. Clinically, CF is mainly characterized by lung inflammation due to the inability to clear the airways from mucus and microorganisms, resulting in chronic infection and maintenance of a pro-inflammatory environment. In 1938, CF was first recognized as a disease of the pancreas, while it was later also associated with lung infection and abnormal salt loss (9). It was demonstrated in 1946 that CF is a genetic disorder, although the CFTR gene, encoding a transmembrane chloride channel, was discovered in 1989. Subsequently, mutations in this gene were linked with CF that typically result in a lack of functional CFTR in epithelial cells. This in turn affects chloride and bicarbonate transport, thereby impairing a range of physiological processes in the lung, including salt homeostasis, mucociliary function and innate immunity. Currently, the cystic fibrosis mutation database contains more than 2000 mutations of the CFTR gene of which many are associated with the onset of CF. The most common mutation in Caucasians is ΔF508, which accounts for about 50% of the CF cases in Europe, and results in a protein that is unable to fold properly. Sequence analysis of mutant CFTR alleles revealed that most alterations are missense mutations, while frameshifts, splicing and nonsense mutations have been described as well. Likewise, deletions and insertions in the CFTR gene are also known. CFTR mutations can be divided into six classes based on their effect on protein function (figure 2 adapted from 6).
Class I, II and III mutations typically result in a lack of CFTR function due to the absence of protein production (Class I), misfolding in the ER and subsequent degradation by the proteasome (Class II) or the inability of the channel to open (Class III). Patients with these mutations have in general a severe phenotype. In contrast, Class IV, V and VI mutations result in residual CFTR activity owing to reduced conduction (Class IV), reduced expression of CFTR (Class V), or decreased protein stability (Class VI). The residual channel function associated with these mutations explains the mild phenotype of individuals with these alleles. Several therapeutic strategies have been developed, employing small molecule drugs that either restore trafficking/folding or restore channel function, so-called correctors and potentiators. Currently, two correctors (lumacaftor and tezacaftor) and one potentiator (ivacaftor) have been developed by Vertex Pharmaceuticals, which have entered the clinic. Pre-clinical studies revealed that ivacaftor restored CFTR-dependent chloride transport as well as mucociliary clearance in most class III and IV mutations (6). The efficacy of this small molecule drug in Class III mutations was subsequently confirmed in different clinical trials. To further the rational development of new therapeutics, in-depth structural information is required. With regards to this, novel structures of human CFTR with or without small molecule drugs were solved recently (7,8). These will be discussed in more detail below.
Biochemical and structural features of CFTR
The human CFTR gene is located on chromosome 7 and contains 27 exons that encode a glycosylated transmembrane channel of 1480 amino acids that transports chloride ions down their electrochemical gradient. The CFTR protein is expressed in epithelial cells of different tissues, including those of the respiratory tract, digestive tract and sweat glands (10). Moreover, CFTR belongs to the family of ATP-binding cassette (ABC) transporters. Members of this group are present in all organisms and typically function as pumps, employing the energy of ATP hydrolysis to power the translocation of substrate molecules across membranes (11). However, CFTR is the only ABC transporter that utilizes the energy of ATP hydrolysis to control the opening and closing of its ion channel. ABC transporters are multidomain proteins, comprising two nucleotide-binding domains (NBD) and two transmembrane domains. Likewise, CFTR is made up of two transmembrane domains that form the ion channel and two NBDs that bind and hydrolyze ATP with a large regulatory domain that connects the two parts of CFTR (12). Although CFTR exhibits ATPase activity, its channel function is controlled by protein kinase A (PKA)-dependent phosphorylation of the R domain.
Structural features of human CFTR
Although crystal structures of individual domains of human CFTR have been solved, no detailed structural information of the full-length protein is available. Novel technical improvements in cryo-EM enabled the structural assessment of difficult-to-crystalize proteins at atomic resolution. This technique was therefore used recently to determine the structure of human CFTR. Figure 3 shows the structure of non-phosphorylated human CFTR without ATP at 3.9 Å (7) in surface (left panel) and ribbon representation center panel), revealing an overall structure similar to other ABC transporters. The protein adopts an inward facing conformation in a closed state with two TMDS (shown in orange and green) containing 12 α helices that are packed into a bundle, and two NBDs (shown in blue and magenta) attached to the cytosolic side of the TMDs. The R-domain (in grey) is located between the TMDs at the cytosolic side near the NBDs and is made up of about 200 residues with over 10 potential phosphorylation sites.
The phosphorylated R-domain enables channel opening through dimerization of the NBDs, while the dephosphorylated R-domain inhibits channel opening by preventing dimerization of the NBDs as suggested by its location between the NBDs. It has been established that following phosphorylation of the R-domain, ATP binding by the NBDs induces channel opening and ATP hydrolysis closes it (12). Conceivably, phosphorylation of the R-domain activates the ATPase activity as the R-domain moves outward. Interactions between TMDs and NBDs maintain structural integrity and relay conformational changes from the NBDs to the TMDs, thereby coupling phosphorylation and ATP hydrolysis to channel gating. Interestingly, TMD1 contains an N-terminal extension located in the membrane, the lasso motif, that is not present in other ABC transporters (center panel). Its precise function is not well understood, although mutations in this region cause trafficking defects or abnormal channel gating (7).
In contrast to what is predicted by the positive inside rule, CFTR contains numerous positively charged residues throughout its transmembrane domains. The right panel of figure 3 shows that the ion conduction pathway is made up of a large cytosolic cavity, a narrow transmembrane tunnel and a gate near the extracellular surface. The wall of the cavity is lined by positively charged residues that facilitate the conductance of ions and contribute to selectivity (shown as red spheres), while the positively charged residues at the extracellular surface are involved in recruiting of anions (shown as red spheres).
Human CFTR contains several highly conserved residues that are located throughout the pore, the cytosolic parts of the TMDs and NBDs. Currently, over 50 CF-causing missense mutations are known, according to the CFTR2 database, which mainly involve these conserved residues. These are shown as red spheres in figure 4, revealing that these mutations can be divided into four groups, namely: pore construction mutations that affect ion conduction, folding mutations of hydrophobic residues in transmembrane α helices that destabilize the structure of CFTR and cause folding defects, ATPase site mutations in the NBDs that affect ATP binding and hydrolysis, resulting in defective channel gating and NBD/TMD interface mutations that impair the coupling of phosphorylation and ATP hydrolysis to channel gating also resulting in defective channel gating. The ΔF508 mutation belongs to this group. Strikingly, no CF-causing mutations are located in the R-domain.
Structural features of human CFTR in complex with potentiators
Several innovative therapeutic strategies are currently pursued that focus on correcting CFTR function. For example, one approach revolves around the use of small molecule drugs to modulate the activity of the CFTR protein. A few modulators have been discovered after laborious screening campaigns of comprehensive compound libraries to identify potential therapeutics that improve CFTR activity. These include correctors that restore folding of CFTR and potentiators that improve CFTR-mediated chloride transport (14). Currently, Vertex Pharmaceuticals has developed two correctors (lumacaftor and tezacaftor) and one potentiator (ivacaftor) that have entered the clinic. Of these, ivacaftor has shown potent efficacy in pre-clinical and clinical studies predominantly in class III mutations (6,14). To further the structure-based design of new therapeutics, in-depth structural information is required. With regards to this, novel structures of human CFTR with or without small molecule drugs were solved recently by cryo-EM (8). Figure 5 shows the structure of human CFTR in complex with ivacaftor at 3.3 Å in surface representation, revealing an overall structure that is similar to previously reported structures of CFTR (7,15) with two TMDs (shown in magenta and blue) and two NBDs (in pink and green). However, the R-domain was not resolved, while ivacaftor (in red spheres) was observed on the outer surface of TMD2 in the center of the lipid bilayer (left panel). Specifically, ivacaftor binds to CFTR at the protein-lipid interface into a cleft made up of transmembrane helices 4,5 and 8. This binding site comprises a flexible hinge of helix 8 that is conformationally active upon ATP binding. Conceivably, binding of ivacaftor stabilizes this rotation, resulting in an improved CFTR function and thereby explaining its efficacy. A close-up of the residues involved in binding of CFTR is provided in the right panel of figure 5. This drug is bound to CFTR through two hydrogen bonds, two aromatic interactions and six hydrophobic interactions. Of these interactions, the hydrogen bonds are crucial for drug recognition. Four residues appear critical for ivacaftor binding - Ser308, Phe312 and Tyr304 – because their replacement with alanine abolishes drug binding. Interestingly, another potentiator (GLPG1837) that is structurally unrelated to ivacaftor also binds to the same site within TMD2 of CFTR. This site, therefore, comprises a hotspot for the binding of potentiators. Based on the data discussed above, it seems that the presence of a potentiator in the binding pocket stabilizes the open conformation of the channel, resulting in improved chloride transport.
Mechanism of channel gating
Efficient ion transport through the CFTR pore requires that its channel is fully open, which is enabled by phosphorylation of the R-domain as well as binding of ATP to the catalytically competent NBD2 (13). Phosphorylation of the R-domain requires it to be released from the NBDs and TMDs and move out of the central cavity of the ion conduction pathway. Based on the structural data discussed above, a model for CFTR channel gating was proposed (figure 6 adapted from 7). The gating cycle is initiated (A) by binding of ATP (yellow dot) to NBD2 (in blue) and the spontaneous disengagement of the unphosphorylated R-domain (red) from its inhibitory position out of the central cavity (B), which enables its subsequent phosphorylation (P in C). This results in dimerization of the NBDs, thereby fully activating the channel (D and E) and permitting the transport of chloride ions (purple dots). Finally, ATP is hydrolyzed by the ATPase activity of NBD2. This results in disruption of the NBD dimer and channel closure (F). ADP is released and replaced with ATP, thereby priming the channel for another gating cycle.
It has been estimated that 26% of the human genes encode integral membrane proteins of the alpha-helical class (1). The overrepresentation of these proteins within the human proteome is indicative of their crucial biological role. For example, membrane proteins are required for transport of ions and larger molecules, signal transduction, cell-cell communication, cell-cell interaction and to organize and maintain cellular shape (2). It is therefore not surprising that integral membrane proteins are also of medical importance. In fact, these proteins make up more than 60% of current drug targets, while many of the known ~16000 disease-linked missense mutations in protein sequences are found in transmembrane domains of membrane proteins (4,5). These mutations typically interfere with normal protein folding or trafficking, resulting in reduced amounts of functional protein at the destination membrane. Hence, misassembly of membrane proteins is an important cause of various genetic human diseases. These include CF a life-threatening genetic disorder that is predominantly found in Caucasian populations (6). CF is caused by mutations in the gene encoding the cystic fibrosis transmembrane regulator (CFTR). This is a chloride-conducting transmembrane channel which regulates anion transport across various epithelial tissues. Loss of CFTR activity affects chloride and bicarbonate transport, thereby impairing a range of physiological processes in the lung, including salt homeostasis, mucociliary function and innate immunity (6). Several innovative therapeutic strategies are currently pursued that focus on correcting CFTR function. For example, one approach revolves around the use of small molecule drugs to modulate the activity of the CFTR protein. These include correctors that restore folding of CFTR and potentiators that improve CFTR-mediated chloride transport (14). Recent structures of human CFTR in complex with potentiators show that these drugs bind to the same site within a binding pocket in the transmembrane region of CFTR at the protein-lipid interface (8). Conceivably, the presence of a potentiator in the binding pocket stabilizes the open conformation of the channel, resulting in improved chloride transport as observed in pre-clinical and clinical studies. These novel structures will improve the mechanistic insight of CFTR channel function and thereby further the molecular understanding of CF. Moreover, it can also be expected that detailed structural information will facilitate the rational design of new therapeutics.
1. Fagerberg L, Jonasson K, von Heijne G, Uhlén M, Berglund L. 2010. Prediction of the human membrane proteome. Proteomics. 10: 1141-1149.
2. von Heijne G. 2007. The membrane protein universe: what's out there and why bother? J Intern Med. 261: 543-557.
3. Nilsson J, Persson B, von Heijne G. 2005. Comparative analysis of amino acid distributions in integral membrane proteins from 107 genomes. Proteins. 60: 606-616.
4. Yin H, Flynn AD. 2016. Drugging Membrane Protein Interactions. Annu Rev Biomed Eng. 18: 51-76.
5. Sanders CR, Myers JK. 2004. Disease-related misassembly of membrane proteins. Annu Rev Biophys Biomol Struct. 33: 25-51.
6. Elborn JS. 2016. Cystic fibrosis. Lancet. 388: 2519-2531.
7. Liu F, Zhang Z, Csanády L, Gadsby DC, Chen J. 2017. Molecular Structure of the Human CFTR Ion Channel. Cell. 169: 85-95.
8. Liu F, Zhang Z, Levit A, Levring J et al. 2019. Structural identification of a hotspot on CFTR for potentiation. Science. 364: 1184-1188.
9. Davis PB. 2006. Cystic fibrosis since 1938. Am J Respir Crit Care Med. 173: 475-482.
10. Riordan JR, Rommens JM, Kerem B, Alon N et al. 1998. Identification of the cystic fibrosis gene: cloning and characterization of complementary DNA. Science. 245: 1066-1073.
11. Ford RC, Beis K. 2019. Learning the ABCs one at a time: structure and mechanism of ABC transporters. Biochem Soc Trans. 47: 23-36.
12. Callebaut I, Chong PA, Forman-Kay JD. 2018. CFTR structure. J Cyst Fibros. 17:S5-S8.
13. Sheppard DN, Welsh MJ. 1999. Structure and function of the CFTR chloride channel. Physiol Rev. 79: S23-45.
14. Gentzsch M, Mall MA. 2018. Ion Channel Modulators in Cystic Fibrosis. Chest. 154: 383-393.
15. Zhang Z, Chen J. 2016. Atomic Structure of the Cystic Fibrosis Transmembrane Conductance Regulator. Cell. 167: 1586-1597.