Molecular Structure Influence on Cystic Fibrosis

Cystic fibrosis (CF) is an autosomal recessive lung disease that is caused by mutations in a single gene (Riorden et al., 1989). The gene encodes for a large membrane protein called the cystic fibrosis transmembrane conductance regulator (CFTR). CFTR is a member of the ATP-binding cassette (ABC) transporter family. Although it is a member of the ABC transporter family it functions as an ion channel by conducting anions (Gadsby et al., 2006). CFTR is also the largest protein that houses an R domain, which is the most unique feature of CFTR. Phosphorylation of the serine residues in the R domain is the gateway step of CFTR channel opening (Sheppard and Walsh, 1999). Many research studies indicate that the R domain must be phosphorylated before ATP can bind and channel activation can occur. While researching CFTR channel activation, several atomic structures of CFTR have been modeled, including zebrafish CFTR (Zhang and Chen, 2016). In Dr. Chen’s 2017 article (Molecular Structure of the Human CFTR Ion Channel) her research focus was to answer a question about the unresolved helix that belongs to the R domain. She proposed that PKA phosphorylation is enabled by R domain disengagement, which explains the possible gating activity of dephosphorylated CFTR.

In Figure 1, the overall structure of human CFTR in the dephosphorylated, ATP-free conformation was given. To obtain this structure human CFTR was expressed and purified from human embryonic kidney cells. The sample was treated with phosphatase and a sample of dephosphorylated CFTR was obtained. The overall resolution of this structure was 3.9 A, in which will be critiqued later in this assignment. First, I would like to report the structure of the lasso motif. The lasso motif interacts with membrane traffic machinery. When the lasso motif undergoes mutations, this can lead to abnormal channel gating. It was also hypothesized that the lasso motif interacts with the R domain, which is the reason why it can regulate channel gating. Figure 1 provides an atomic structure of the lasso motif. In a previous article, Chen describes the “lasso motif” as a rope-like structure (Jue Chen, 2017). The first 40 residues of the lasso motif are inserted into the membrane and form circular rope against TM10-11 (Jue Chen, 2017). The extended end of the lasso, residues 46–61, form the lasso helix 2, which is tucked under the elbow helix 1 (Jue Chen 2017). The Chen lab is the only lab that has ever been able to obtain a 3D structure of the lasso motif. Other labs have attempted to do so but have not been successful which is perplexing. Dr. Chen should disclose how this is done and improve the reproducibility.

To show that PKA phosphorylation stimulates CFTR channel gating Chen obtained macroscopic currents from a patch that included hundreds of human CFTR channels which were exposed to 2mM MgATP. The recordings showed few channel openings prior to phosphorylation. However, when 300 nM PKA was introduced the channel current increased (Figure 5A). In the beginning of the article, Chen stated that PKA phosphorylation of the R domain is enabled by its infrequent spontaneous disengagement, which also explains residual ATPase and gating activity of dephosphorylated CFTR. To visualize this statement, an ATPase assay was used to compare the activity of phosphorylated and dephosphorylated human CFTR. Chen included an Michaelis-Menten graph (Figure 5E) to show the difference between the Vmax values of phosphorylated and dephosphorylated CFTR. The Vmax for phosphorylated CFTR was 134 nmol/mg/min. On the other hand, the dephosphorylated CFTR had a Vmax of 23 nmol/mg/min. This study postulates that low ATPase activity of dephosphorylated CFTR is due to the release of the R domain and rare channel openings in the absence of PKA. Although this is an interesting statement, it should be inferred that if there was any residual ATPase due to the disengagement of the R domain then ATP hydrolysis could still occur regardless. The R domain is important for phosphorylation as it initiates the cascade of events that lead to channel gating. However, CFTR includes nucleotide binding domains (NBDs). The NBDs of CFTR include sequences that bind and hydrolyze intracellular MgATP. These include Walker A, Walker B, and LSGGQ motifis (Sheppard and Walsh, 1999). Walker A interacts with either the  or -phosphate of ATP, which is important for ATP hydrolysis. However, Walker B coordinates Mg2+ in MgATP which is a necessity for ATP binding. If the NBDs are present, then ATP hydrolysis will be able to efficiently take place. Dr. Li and colleagues performed an ATPase assay after purification of CFTR and confirmed that channel gating is directly affiliated with ATP hydrolysis (Li et al., 1996). In the article Dr. Chen expressed this finding as if the R domain was the sole cause of channel gating. The channel activity of the dephosphorylated R domain should have been explained in greater detail. In this section of the paper she should have indicated that there was ATPase activity due to the hydrolyzing role of the NBDs.

To obtain the 3.9 A resolution of dephosphorylated human CFTR, Chen and her colleagues had to conduct model refinement and validation. The model was refined using PHENIX and also in reciprocal space using Refmac. Structural factors were calculated from a half-map using the program Sfall. The R work and R free values were calculated using a mask constraining the model plus a 2A margin. A structure of a protein can be obtained using single particle cryoEM but it involves the calculation of an initial 3D model. However, there should be improvements made when obtaining a 3D structure of the target protein. Usually there is more noise than signal when the resolution is high. Due to this high noise-to-signal ration false interpretations can be made. However, there is a technique used to resolve false interpretations (Chen/Henderson et al., 2013). Using a mask can help to improve signal-to-noise ratio. Typically, structural biologists use various masks in their experiment to confirm if their image is at a desirable resolution or to compare the structure at different resolutions. However, Chen and her colleagues did not use different masks during their refinement procedure. In the supplemental figures section, figure D shows the Frealign output that indicates the fourier shell correlation. In this figure Chen was supposed to show the increase or decrease of resolution based on the masks used. An ideal example of the effect of masks on resolution can be found in a study conducted by Dr. Manolaridis. In this study, the cryo-EM structure of a human ABCG2 mutant was obtained (Manolaridis et al., 2018). In the extended data section, Figure 6b shows that resolution is dependent upon the presence of masks (Manolaridis et al., 2018). Using a soft mask helped to acquire a cryoEM structure of the human ABCG2 mutant at a resolution of 3.1 A. Even though Chen included the steps she used to obtain a dephosphorylated atomic model of human CFTR she did not use different masks or indicate if she used a soft mask as Manolaridis did. In the future Chen should consider using different masks in her experiments to obtain a better resolution or to make sure she is not giving a false interpretation of CFTR.

After reviewing the materials and methods section of this article, I noticed that used a specific detergent when purifying the CFTR protein. 1% 2,2-didecylpropare-1,3-bis D-maltopyranoside (LMNG) was used. My final critique of this article is the title. Although the title was concise it was slightly misleading. With the title being “Molecular Structure of the Human CFTR Ion Channel,” the reader would not expect to receive information about residual ATPase and gating activity of dephosphorylated CFTR. Even though CFTR atomic structures have been published over the years, I was under the impression that the main focus of the article was to reconfirm the architecture of human CFTR or to showcase a better resolution of human CFTR. However, it mainly focused on the structural comparison of two CFTR orthologs, channel activation by R domain phosphorylation, and the comparison of CFTR to other ABCC transporters. The title of the article should have been more descriptive. In this article, the channel gating that occurred when the R domain was dephosphorylated was interesting. However, it did not seem groundbreaking. I do not think this study will have a significant impact on the field. It basically reconfirmed the role of the NBDs and how they hydrolyze ATPs to activate channel gating.