VX-809

Corrector VX-809 Promotes Interactions Between Cytoplasmic Loop One and the First Nucleotide-Binding Domain of CFTR
Tip W. Loo, David M. Clarke

PII: S0006-2952(17)30182-X
DOI: http://dx.doi.org/10.1016/j.bcp.2017.03.020
Reference: BCP 12774

To appear in: Biochemical Pharmacology

Received Date: 30 January 2017
Accepted Date: 28 March 2017

Please cite this article as: T.W. Loo, D.M. Clarke, Corrector VX-809 Promotes Interactions Between Cytoplasmic Loop One and the First Nucleotide-Binding Domain of CFTR, Biochemical Pharmacology (2017), doi: http:// dx.doi.org/10.1016/j.bcp.2017.03.020

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Corrector VX-809 Promotes Interactions Between Cytoplasmic Loop One and the First Nucleotide-Binding Domain of CFTR

Tip W. Loo and David M. Clarke

Department of Medicine and Department of Biochemistry, University of Toronto, Toronto, Ontario. M5S 1A8. Canada

Running Title: VX-809 Promotes CFTR TMD1-NBD1 Interactions.

Corresponding Author:
David M. Clarke
Department of Medicine, University of Toronto
1 King’s College Circle, Rm. 7342, Medical Sciences Building Toronto, Ontario, M5S 1A8, Canada
Tel. or FAX: 416-978-1105
E-mail: [email protected]

Abstract

A large number of correctors have been identified that can partially repair defects in folding, stability and trafficking of CFTR processing mutants that cause cystic fibrosis (CF). The best corrector, VX-809 (Lumacaftor), has shown some promise when used in combination with a potentiator (Ivacaftor). Understanding the mechanism of VX-809 is essential for development of better correctors. Here, we tested our prediction that VX-809 repairs folding and processing defects of CFTR by promoting interactions between the first cytoplasmic loop (CL1) of transmembrane domain 1 (TMD1) and the first nucleotide-binding domain (NBD1). To investigate whether VX-809 promoted CL1/NBD1 interactions, we performed cysteine mutagenesis and disulfide cross-linking analysis of Cys-less TMD1 (residues 1-436) and
TMD1 (residues 437-1480; NBD1-R-TMD2-NBD2) truncation mutants. It was found that VX- 809, but not bithiazole correctors, promoted maturation (exited endoplasmic reticulum for addition of complex carbohydrate in the Golgi) of the TMD1 truncation mutant only when it was co-expressed in the presence of TMD1. Expression in the presence of VX-809 also promoted cross-linking between R170C (in CL1 of TMD1 protein) and L475C (in NBD1 of the
TMD1 truncation protein). Expression of the TMD1 truncation mutant in the presence of TMD1 and VX-809 also increased the half-life of the mature protein in cells. The results suggest that the mechanism by which VX-809 promotes maturation and stability of CFTR is by promoting CL1/NBD1 interactions.

Key words: CFTR; cystic fibrosis; VX-809 (Lumacaftor); bithiazole correctors; transmembrane- and nucleotide-binding domains; cytoplasmic loop; cysteine cross-linking.

Chemical compounds studied in this article:

VX-770 (Ivacaftor; Kalydeco) (PubChem CID: 16220172); VX-809 (Lumacaftor) (PubChem CID: 16678941);
15Jf (PubChem CID: 11958611);
Cycloheximide (PubChem CID: 6197).

⦁ Introduction

The cystic fibrosis transmembrane conductance regulator (CFTR, ABCC7) is an ABC (ATP- binding cassette) cAMP-regulated chloride channel. It is located on the apical surface of

epithelial cells that line the lung airways and ducts of various glands where it functions to regulate salt and water homeostasis [1].
Cystic fibrosis (CF) is caused by genetic mutations that lead to reduced activity or expression of CFTR at the cell surface. Loss of CFTR function in the airway epithelia leads to obstructive airway disease and chronic bacterial infections. The majority of CF patients express CFTRs that contain processing mutations (such as F508 (most common), V232D, H1085R, and others located throughout the molecule) that impair folding, trafficking, stability and activity of the protein [1]. Processing mutations trap CFTR in the endoplasmic reticulum (ER) as a partially folded protein with incomplete domain interactions [2] and incomplete packing of the transmembrane (TM) segments [3, 4]. Partial rescue of F508-CFTR by expression at low temperature [5] or in the presence of correctors [6, 7] suggests that it should be possible to prevent CF by coaxing enough of the CFTR processing mutants to complete the folding process to yield functional channels at the cell surface using a drug-rescue approach. Clinical trials showed that monotherapy with the best corrector identified to date (VX-809) did not significantly improve lung function or sweat chloride concentration [8]. In addition, immunoblot analysis of rectal biopsy specimens from patients showed no maturation of F508-CFTR [8]. A combination of VX-809 with VX-770 (a potentiator that improves channel function) shows some promise. A problem however, is that it has been reported that VX-770 destabilized F508-CFTR [9, 10]. Another group found that VX-770 also destabilized wild-type CFTR [9]. A potential solution would be to develop better correctors to improve the efficiency of F508-CFTR maturation into a stable conformation.
The 1480 amino acids of CFTR are organized into two transmembrane domains (TMDs), two nucleotide-binding domains (NBDs), and an R domain [11] (Fig. 1A). The protein contains two N-glycosylation sites in the external loop connecting TM segments 7 and 8 of TMD2. Each

homologous half contains an N-terminal TMD followed by an NBD. The secondary structure predicts that each TMD is linked to each NBD via cytoplasmic loops (CLs) (see below). Domain interactions are predicted to be an important feature of CFTR maturation, inhibition of maturation by processing mutations, and rescue of processing mutations by correctors such as VX-809 (reviewed in [12]). CFTR is first synthesized in the ER as a core-glycosylated immature 170 kDa protein. The term “maturation” refers to the process whereby immature CFTR leaves the ER and traverses the Golgi where addition of complex carbohydrates to the pair of N- glycosylated sites in the external loop connecting TM segments 7 and 8 converts it to the 190 kDa mature protein. Mature CFTR is then delivered to the cell surface. While folding of much of CFTR occurs cotranslationally, some folding steps such as packing of the TM segments and incorporation of NBD2 into the structure appear to occur post-translationally [13, 14]. Studies on CFTR truncation mutants showed that a mutant lacking NBD2 could mature but those lacking TMD1, TMD2 or NBD1 do not [15]. Therefore, studies of the CFTR truncation mutants suggest that TMD1-TMD2 or TMD1-NBD1 interactions may be particularly important for maturation.
Although many correctors identified to date are nonspecific, VX-809 is particularly important because it is more specific (also rescued processing mutants of ABCA4 [16] but not P- glycoprotein [17] or hERG potassium channel [7] processing mutants), restores domain assembly [18], and has the ability to promote maturation of CFTR mutants with processing mutations in different domains [17] [18]. In a previous study using CFTR domains expressed as separate proteins, we found that VX-809 stabilized TMD1 (residues 1-402) but not TMD2, NBD1 or NBD2 [19]. Expression of TMD1 in the presence of VX-809 increased its half-life in intact cells from 1.5 h to 8 h. Ren et al. [20] also reported that VX-809 stabilized a TMD1 truncation mutant consisting of residues 1-380. The result suggested that the VX-809 binding site was located in TMD1.

Knowledge of the mechanism of VX-809 is needed to develop better correctors. One advantage of VX-809 compared to other correctors is that it promotes maturation of F508- CFTR and stabilizes the protein [21]. We previously observed that human ABC proteins defective in processing showed defects in TMD-NBD interactions [2]. Here, we used CFTR truncation mutants to test our prediction that VX-809 promotes maturation and stability of CFTR by promoting domain interactions between the first cytoplasmic loop (CL1) of TMD1 and NBD1. This is the major TMD1/NBD1 contact point identified recently in the atomic structure of CFTR [22].

⦁ Materials and Methods

⦁ Chemicals.

Corrector 3-(6-{[1-(2,2-Difluoro-benzo[1,3]dioxol-5-yl)-cyclopropanecarbonyl]-amino} -3- methyl-pyridin-2-yl)-benzoic acid (VX-809, LumacaftorTM) was obtained from Selleck Chemicals LLC (Houston, TX). Corrector N-(2-(5-chloro-2-methoxyphenylamino-4′-methyl-

4,5′-bithiazol-2′-ylpivalamide (15Jf) and potentiator N-(5-Hydroxy-2,4-bis(2-methyl-2- propanyl)phenyl]-4-oxo-1,4-dihydro-3-quinolinecarboxamide (VX-770, IvacaftorTM, KalydecoTM) were obtained from Cystic Fibrosis Foundation Therapeutics, Inc. and Dr. Robert Bridges (Rosalind Franklin University, Chicago, IL, USA). Dulbecco’s modified Eagle’s media and calf serum were obtained from Wisent Inc. (St. Bruno, Quebec). Cycloheximide was obtained from BioShop Canada (Burlington, ON). Monoclonal antibody against GADPH (1:2500 dilution of a
100 g/ml stock) was obtained from Santa Cruz Biotechnology (Dallas, TX). 1, 10- phenanthroline and sodium butyrate were obtained from Sigma Aldrich Canada (Oakville, ON). Mouse monoclonal antibody A52 (1:1000 dilution of a 100 g/ml stock) and rabbit polyclonal antibody against NBD2 of CFTR (1:5000 dilution of serum) were generated as described previously [3, 23]. Endoglycosidase Hf (used at 20 Units/l) and endoglycosidase F (PNGase F; used at 10 Units/l) were obtained from New England Biolabs (Whitby, Ontario, Canada). Peroxidase-labeled secondary antibodies to mouse or rabbit IgGs were obtained from KPL (Gaithersberg, MD). Luminata Forte Western HRP substrate was obtained from Millipore.

⦁ Construction of mutants.

Mutations were introduced into CFTR cDNA (residues 1-1480) by site-directed mutagenesis as described by Kunkel [24]. Cys-less constructs were derived from a Cys-less CFTR cDNA that was constructed by replacing Cys-590 and Cys-592 with leucines and changing the other 16 cysteines to alanines [25, 26]. Construction of the V510C/A1067C mutant in a full-length Cys- less CFTR background was described previously [2]. A truncation mutant lacking TMD1 (TMD1) contained the initiating methionine residue 1 followed by residues 437-1480 [19]. A

L475C mutation was introduced into NBD1 of the Cys-less TMD1. The TMD1 truncation mutant consisted of residues 1-436 and an A52 epitope tag at its C-terminal end (TMD1-A52) [19]. The epitope for monoclonal antibody A52 was derived from the cDNA of SERCA1 Ca2+- ATPase [27] and attached to the C-terminal end of TMD1 cDNA to yield a sequence of CFTR- Leu436–NSASPEFDDLPLAEQREAARRGDPRQ (A52 epitope underlined). A R170C mutation was introduced into CL1 of Cys-less TMD1-A52. The TMD1+TMD2 construct consisted of residues 1-388 (TMD1) joined to residues 847-1196 (TMD2) together with an A52 tag at the C-terminal end as described previously [19].

⦁ Expression of mutants.

The mutant CFTRs were transiently expressed in HEK 293 cells as described previously [2]. HEK 293 cells were transfected with the cDNAs and the medium was changed four hours later to fresh medium (Dulbecco’s modified Eagle’s medium containing 10% (v/v) calf serum) with or without correctors VX-809 (5 M) or 15Jf (5 M) or DMSO vehicle (< 1% (v/v)). These concentrations of correctors were used because they were effective for rescuing processing and truncation mutants of CFTR [28]. Cells were harvested 18 to 48 h after the change in medium. SDS extracts from whole cells expressing the CFTR truncation mutants were subjected to immunoblot analysis using 7% (for TMD1) or 10% (for TMD1-A52) (w/v) acrylamide gels and monoclonal antibody A52 (for TMD1-A52) or a rabbit polyclonal antibody (to detect TMD1) followed by peroxidase-labeled anti-mouse or anti-rabbit secondary antibodies respectively, and the chemiluminescence signals were recorded on film or using the ChemiDocTM XRS+ system (BioRad, Mississauga, ON.). An equivalent amount of the sample was loaded onto 10% (v/v) SDS-PAGE gels and subjected to immunoblot analysis with a monoclonal antibody against

glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (internal control). In some cases, SDS extracts were treated with endoglycosidases H or F as described previously [19].
To test the effects of VX-809 on the stability of TMD1, cells were co-transfected with plasmids encoding TMD1 or TMD1-A52 and media changed to contain 5 mM sodium butyrate with or without 5 M VX-809 or the equivalent volume of DMSO vehicle (< 1% (v/v)). Sodium butyrate is a histone deacetylase inhibitor that increases expression of plasmid encoded proteins in HEK 293 cells [29]. After 18 h at 37 oC, protein synthesis was stopped by addition of medium containing 0.5 mg/ml cycloheximide with or without 5 M VX-809 but no sodium butyrate. The cells were then incubated at 37 oC for various time periods (0-8 h) [17, 19]. Whole cell SDS extracts were subjected to immunoblot analysis as described above.

⦁ Cross-linking analysis.

To test for the effects of VX-809 on CL1-NBD1 cross-linking, the L475C TMD1 mutant was co-transfected with R170C TMD1-A52 or Cys-less TMD1-A52 at a ratio of 1:2 TMD1 to TMD1 cDNA. The mutants were transiently expressed in HEK 293 cells in the absence or presence of 5 M VX-809. Membranes were prepared from the transfected cells as described previously [2] and suspended in TBS, pH 7.4. Samples were then treated with or without oxidant (1 mM copper phenanthroline) for 5 min at 0 oC. The reactions were stopped by addition of EDTA (10 mM final concentration). The samples were then mixed with 1 volume of 2X SDS sample buffer with no reducing agent and samples subjected to immunoblot analysis using 7% (to detect TMD1) or 10% (to detect TMD1) (w/v) SDS polyacrylamide gels.
To test if VX-809 or VX-770 could inhibit NBD-TMD cross-linking, membranes prepared from cells expressing full-length V510C/A1067C CFTR or mutant R170C-TMD1-A52 plus

L475C TMD1 were pre-treated for 30 min at 0 oC with or without (contained equivalent amount of DMSO vehicle) 20 M VX-809 (for V510C/A1067C) or 20 M VX-809 plus 20 M VX-770 (for R170C-TMD1-A52 plus L475C TMD1). Samples were then treated with or without 1 mM copper phenanthroline for 5 min at 0 oC. The reactions were stopped by addition of 2X SDS sample buffer containing 25 mM EDTA but no thiol reducing agent. Samples were then run on 7% (w/v) SDS polyacrylamide gels and subjected to immunoblot analysis with rabbit polyclonal antibody against NBD2.

⦁ Data analysis.

The Western blots were either recorded on films or ChemiDocTM XRS+ Imager and analyzed using ChemiDocTM XRS+ with Image LabTM software (version 3.0) (Bio-Rad Lab. Inc., Mississauga, Ontario). The results were expressed as an average of quadruple experiments + standard deviation. The Paired Student T-test (GraphPad software; https://www.graphpad.com) was used to test for statistical significance (P < 0.001; n=4) when comparing two samples. A one-way ANOVA (Figs. 2 and 4) with a post-hoc Tukey HSD test (Fig. 2) was used to test for differences between groups. Significance was accepted at P < 0.01.

⦁ Results.

⦁ The mechanisms of CFTR rescue by bithiazoles and VX-809 are different.

VX-809 appears to bind to TMD1 [19, 20]. Does VX-809 interactions with TMD1 promote interactions with the other domains? To test if VX-809 interactions with TMD1 promote

interactions with other domains, we tested if a truncation mutant lacking TMD1 (residues 437- 1480; NBD1-R-TMD2-NBD2 (TMD1 in Fig. 1A)) would mature if co-expressed with TMD1 in the presence of VX-809 or the bithiazole corrector 15Jf. Bithiazoles were selected for study since they are another important class of CFTR correctors [6, 28, 30]. Bithiazoles may also interact with the TMDs of CFTR because they block cross-linking between TM segments 6 and 12 [31].
To test if VX-809 or 15Jf would promote maturation of TMD1, HEK 293 cells were transfected with TMD1 alone or together with TMD1-A52 and expressed with or without 5 M VX-809 or 15Jf for 48h. An A52 epitope was added to the C-terminal end of TMD1 for detection because the anti-CFTR rabbit polyclonal antibody reacts with NBD2 [3]. Whole cell SDS extracts were subjected to immunoblot analysis. Fig. 1B shows that correctors did not promote maturation of mutant TMD1 when it was expressed alone. It was expressed as a 120 kDa immature protein when expressed alone in the absence or presence of VX-809 or 15Jf (Fig. 1B). Maturation of TMD1 however, was induced when it was co-expressed with TMD1-A52 in the presence of VX-809 but not 15Jf. When TMD1 was incubated in the presence of TMD1-A52 and VX-809 for 48 h, about equivalent levels of 120 kDa (immature) and 140 kDa (mature) forms of TMD1 were observed. The 140 kDa protein was mature TMD1 protein because it was sensitive to endoglycosidase F and resistant to endoglycosidase H (Fig. 1C). These results suggest that VX-809 promoted TMD1-TMD1 interactions to yield a TMD1 structure that can leave the ER for processing in the Golgi.
The 15Jf corrector appears to rescue CFTR by a different mechanism than VX-809 because it did not rescue TMD1 when it was co-expressed with TMD1 (Fig. 1B). To determine if 15Jf differed from VX-809, we tested whether TMD1 alone could be stabilized with 15Jf using the

cycloheximide chase approach [19]. Accordingly, HEK 293 cells were transfected with the TMD1-A52 cDNA in the absence or presence of 15Jf or VX-809. The next day, the media was replaced with fresh media containing cycloheximide to inhibit protein synthesis. Whole cell SDS extracts were then collected at various intervals and subjected to immunoblot analysis. In the absence of correctors, the 45kDa TMD1-A52 protein showed a half-life of about 1 hour (Fig. 2A). Similarly, the presence of 15Jf did not significantly affect the stability of the TMD1-A52 protein as its half-life was also about 1h. These results are consistent with the observation that 15Jf did not rescue TMD1 when co-expressed with TMD1-A52 (Fig. 1B). By contrast, the presence of VX-809 stabilized TMD1-A52 since its half-life increased to more than 8 hours. The results suggest that VX-809 and 15Jf rescue CFTR by different mechanisms.
We previously found that corrector 15Jf appeared to directly interact with the CFTR TM domains because it would promote core-glycosylation of a TMD1 (residues 1-388) + TMD2 (residues 847-1196) truncation mutant (Fig. 1A) [17]. The TMD1+TMD2 mutant lacks the NBDs and the R domain and is likely inserted incorrectly into the membrane. Incorrect insertion of the TM segments of ABC truncation mutants (and correction with drug substrates) has previously been observed with CFTRs’ sister protein – the P-glycoprotein drug pump [32]. To test if VX-809 would promote glycosylation of A52-tagged TMD1+TMD2, the mutant was expressed in HEK 293 cells in the presence or absence of 5 M VX-809 or 15Jf. Whole cell SDS extracts were subjected to immunoblot analysis (Fig. 2B). In the absence of correctors, (TMD1+TMD2)-A52 was expressed as an 80 kDa unglycosylated protein. The presence of VX- 809 did not increase the yield of glycosylated protein. Expression in the presence of 15Jf however, resulted in about equivalent amounts of 85 and 80 kDa (TMD1+TMD2)-A52 proteins. The 85 kDa protein was core-glycosylated as it was sensitive to treatment with endoglycosidase H (Fig. 2C). Since the glycosylation sites are in TMD2, the results suggest that 15Jf rescues

CFTR through interactions with TMD2 whereas VX-809 rescues through interactions with TMD1.

⦁ VX-809 promotes interactions between CL1 and NBD1.

The atomic structure of zebrafish CFTR shows that TMD1 contacts NBD1 at CL1 [22] as shown in Fig. 3A. The zebrafish CFTR structure is generally consistent with the predicted structural model for human CFTR [33]. Other TMD1 domain contacts are CL2 with NBD2 and packing of TM segments with TMD2 (Fig. 3A). Since NBD2 is not needed for maturation [15] and VX-809 did not promote glycosylation of the TMD1+TMD2 truncation mutant (Fig. 2B), we predicted that stabilization of the TMD1 protein with VX-809 increases maturation of the
TMD1 protein (Fig. 1B) by promoting CL1-NBD1 interactions.

Cross-linking has been successfully used to study the impact of processing mutations on domain interactions in ABC proteins [2]. For example, cysteine mutagenesis and cross-linking analysis was a useful approach to study CFTR NBD1-CL4 (TMD2) interactions because cross- linking between the N- and C-terminal halves of the protein caused it to migrate slower on SDS- PAGE gels [2, 34]. It has been difficult however, to examine NBD1-CL1 interactions using cross-linking analysis because cross-linking between domains in the same half of CFTR causes very little change in migration of the protein on SDS-PAGE gels. For example, He et al. [35] could only detect cysteine NBD1-TMD1 cross-linking with a 3,6-dioxaoctane-1,8-diyl bismethanethiosulfonate (M8M) cross-linker in mutants V171C/E407C or V171C/L408C after performing limited trypsin digestion.
Although mutants V171C/L408C or V171C/E407C showed cross-linking, the Val171 (NBD1) - Glu407/Leu408 (TMD1) contact point was not critical for folding or activity since

maturation of CFTR was not reduced when residues 404-435 were deleted [36] and cross-linking had no effect on channel gating. We propose that an important NBD1-CL1 contact point involves segments containing residues Arg170 in CL1 and Leu475 in NBD1. This contact point is predicted to be important because Leu475 is located on a segment of NBD1 that forms a ball- and-socket connection with the intracellular helix at the end of CL1 [22].
To test for the effect of VX-809 on NBD1-CL1 interactions, we introduced the R170C mutation into Cys-less TMD1-A52 and L475C mutation into Cys-less TMD1. Cysteine mutagenesis and cross-linking studies work best in Cys-less backgrounds because CFTR contains 18 endogenous cysteines that can interfere with cross-linking analysis of introduced cysteines. The rationale was that cross-linking between cysteines in TMD1-A52 and TMD1 would be readily detectable because the linked proteins would migrate slower on SDS-PAGE gels. Although Cys-less CFTR is a processing mutant it is suitable for cysteine mutagenesis and cross-linking studies because a Cys-less mutant retains channel activity after rescue [31].
We examined the effects of VX-809 on cross-linking by co-transfecting cells with L475C

TMD1 and A52-tagged R170C or Cys-less TMD1 and incubation with or without the corrector. The cells were transfected with a 2:1 ratio of TMD1-A52 to TMD1 cDNA to promote maturation of TMD1 protein. Membranes prepared from the transfected cells were then treated with or without oxidant (copper phenanthroline) and samples subjected to immunoblot analysis with rabbit polyclonal antibody to NBD2 of CFTR (Fig. 3B). In the absence of VX-809, L475C
TMD1 did not mature (Fig. 3B, middle panel). The immature form of L475C TMD1 was not cross-linked with copper phenanthroline as there was no change in migration pattern on SDS- PAGE gels (Fig. 3B, middle panel). Mature L475C TMD1 was observed when it was co- expressed with Cys-less (Fig. 3B, left panel) or R170C TMD1-A52 (Fig. 3B, right panel). When the samples were treated with copper phenanthroline, about 50% of the mature form of L475C

TMD1 showed a shift to a slower migrating protein only when it was expressed with R170C TMD1 (Fig. 3B, right panel). These results suggest that the slower migrating protein represented L475C TMD1 protein cross-linked to R170C TMD1-A52 protein.
The R170C TMD1-A52 mutant contained an A52 epitope tag to distinguish its expression from the TMD1 mutant. The samples in Fig. 3B (middle and right panels) were also subjected to immunoblot analysis with monoclonal antibody A52. A slow migrating protein was only observed when the membranes prepared from cells incubated with VX-809 were treated with oxidant (Fig. 3C, middle panel). The slow migrating species represented R170C TMD1-A52 cross-linked to L475C TMD1 because it was sensitive to treatment with dithiothreitol (Fig. 3C, right panel). These results suggest that the mechanism of VX-809 rescue of the TMD1 CFTR is to promote interactions between NBD1 and CL1 of TMD1.

⦁ TMD1 partially stabilizes TMD1 when expressed in the presence of VX-809.

Corrector VX-809 stabilizes TMD1 (Fig. 2A). We predicted that rescue of TMD1 by TMD1 plus VX-809 also promotes maturation of the TMD1 construct into a more stable protein. To test for stabilization of TMD1, the cycloheximide chase approach was used with Cys-less versions of TMD1 and TMD1. The Cys-less constructs are useful as they serve as good tools for understanding how correctors promote maturation of processing mutants since full-length Cys-less CFTR is a processing mutant [31].
HEK 293 cells were transfected with TMD1-A52 plus TMD1 cDNAs. One set of transfected cells was incubated at 37 oC in the presence of VX-809. Another set of cells was incubated in the absence of VX-809 at 30 oC to promote low temperature maturation of TMD1-A52 plus TMD1.

The next day, the media was replaced with that containing cycloheximide to inhibit protein synthesis and the cells were incubated at 37 oC. Whole cell SDS extracts were collected at various time points and samples subjected to immunoblot analysis. It was found that VX-809 increased the stability of the mature form of TMD1 (Fig. 4). In the absence of VX-809, mature
TMD1 had a half-life of less than 2 h (Fig. 4). The presence of VX-809 increased the half-life of TMD1 to more than 8 h. The results show that rescue of TMD1 with TMD1 plus VX-809 induces the protein to fold into a more stable structure.

⦁ VX-770 or VX-809 do not block NBD-TMD cross-linking.

The ability of VX-809 to stabilize TMD1 expressed as a separate protein suggests that TMD1 contains a VX-809 binding site [19, 20]. By contrast, modeling studies suggest that the VX-809 binding site is located at the NBD1-CL4 interface [5, 35, 37].
To determine if binding of VX-809 modulates the NBD1-CL4 contact point, we tested whether it could inhibit cross-linking between cysteines V510C(NBD1) and A1067C(CL4). The V510C/A1067C site has been shown to be an important contact point (Fig. 1A) because a V510D mutation rescued CFTR processing mutants [38] and the V510C/A1067C cysteines formed a disulfide bond when the full-length mutant was treated with oxidant [2]. Cross-linking of V510C/A1067C could be easily detected because the cross-linked protein migrated slower on SDS-PAGE gels [2].
Accordingly, the V510C/A1067C full-length mutant (in a Cys-less background) was expressed in HEK 293 cells. Membranes were prepared and samples were pre-incubated with or without 20 M VX-809. Samples were then treated with copper phenanthroline (oxidant) for 5

min at 0 oC. The reactions were stopped by addition of 2X SDS sample buffer that contained 25 mM EDTA but no thiol reducing agent. Immunoblot analysis of the samples showed that the presence of corrector VX-809 did not significantly inhibit (P > 0.1) cross-linking when compared to a sample that was cross-linked without corrector (Fig. 5A). The results suggest that VX-809 does not bind to an NBD1-CL4 site that could block V510C/A1067C cross-linking. It is possible however, that the introduction of cysteines into this pocket might have changed the original pocket configuration.
It has also been reported that the potentiator VX-770 in combination with corrector VX-809 could destabilize wild-type CFTR or mutant F508 [9, 10]. To determine if VX-770 plus VX- 809 destabilized CFTR by inhibiting NBD1-CL1 interactions, we tested if VX-770 plus VX-809 would inhibit cross-linking of the R170C TMD1 protein to the L475C TMD1 polypeptide. HEK 293 cells were co-transfected with R170C TMD1 and the L475C TMD1 constructs and rescued with VX-809. Membranes were then prepared and pre-incubated with or without 20 M VX-770 plus 20 M VX-809. The samples were then treated with 1 mM copper phenanthroline for 5 min at 0 oC. Immunoblot analysis of the cross-linked samples showed that the potentiator VX-770 plus corrector VX-809 did not significantly inhibit (P > 0.1) cross-linking when compared to a sample that was cross-linked without the potentiator plus corrector (Fig. 5B). The results suggest that VX-770 plus VX-809 do not destabilize CFTR by inhibiting NBD1-CL1 interactions.

⦁ Discussion.

Processing mutations disrupt TMD-NBD interactions in human ABC proteins [2]. The results of this study show that VX-809 promotes TMD1-NBD1 interactions in CFTR. Stabilization of the TMD1 domain by VX-809 promoted interactions between CL1 of TMD1 with NBD1 to induce maturation and enhance stability of TMD1. The results are consistent with the observation that VX-809 increased the stability of full-length CFTR at the cell surface [21]. It was recently reported however, that VX-809 restored global conformational maturation of F508 CFTR but did not thermally stabilize the protein [18].
Recent studies show that the CL1-NBD1 contact point is also important for function because it plays a critical role in ATP hydrolysis and channel function [39]. It was reported that intrinsic ATPase activity and channel gating were severely inhibited by the presence of a CL1 peptide (residues Phe157 to Lys166). The CL1 peptide was shown to directly interact with NBD1 because it could precipitate the NBD1 domain expressed as a separate protein. By contrast, an earlier study [35] suggested that the CL1-NBD1 contact point was not critical for function because cross-linking with M8M did not affect channel activity. An explanation for the difference is that M8M is a flexible cross-linker that can span large range of distances (3.9–13 Å) [40]. Since the cross-linker is flexible it may not interfere with any conformational changes required for channel activity.
The recent atomic structure of CFTR from zebrafish showed that the CL1-NBD1 contact point shares a common ball-and-socket structure found in all ABC transporters [22]. The CL1 loop contains a short intracellular helix (the “ball”) that is bound to a surface cleft (the “socket”) of NBD1. The NBD1/TMD contacts like CL1-NBD1 may be particularly important for folding of CFTR because NBD1 lacks a segment containing a  strand and an -helix that form an important helical component of the socket [22, 41]. Therefore, the NBD1/TMD interface has a weaker contact point compared to the NBD2/TMD interface. It was proposed that the

NBD1/TMD1 interface was a potential target for potentiators or correctors as it would be more exposed to the solvent [22].
It was recently reported that trimethylangelicin resembled VX-809 because it could also stabilize TMD1 expressed as a separate protein [42]. Trimethylangelicin is structurally quite different from VX-809 and has both corrector and potentiator activity. It has been shown to act as a corrector to promote maturation of F508 CFTR [43] and act as a potentiator of wild-type CFTR activity [44]. Trimethylangelicin resembles the VX-770 potentiator as it directly modifies the ATP-independent channel activity of CFTR [42]. Future studies will be needed to test if trimethylangelicin binds to a different site than VX-809 or if it also promotes CL1-NBD1 interactions.
The first TM domain appears to be an important target for rescue of other mammalian ABC proteins. For example, the most potent corrector for rescue of processing mutants of the P- glycoprotein drug pump (tariquidar) also stabilizes TMD1 of P-glycoprotein expressed as a separate polypeptide [2].
Correctors like VX-809 and trimethylangelicin appear to directly bind to TMD1 of CFTR to promote maturation and stabilize the protein. Bithiazoles also appear to directly bind to the TMDs of CFTR because they inhibited cross-linking between cysteines introduced into TMD1 and TMD2 [31]. They differ from VX-809 and trimethylangelicin however, as we found that the bithiazole 15Jf did not stabilize the TMD1 protein (Fig. 2A). Instead, bithiazoles appear to stabilize TMD2 [45]. Further evidence that bithiazoles interact with TMD2 was the finding that they enhance glycosylation of a TMD1+TMD2 truncation mutant [17]. Here we found that VX- 809 was different from 15Jf because it did not promote glycosylation of the TMD1+TMD2 mutant (Fig. 2B). The different target sites would explain why rescue of F508-CFTR with VX- 809 and bithiazoles was additive [7].

In summary, there are three major conclusions derived from these studies. First, the results suggest that the mechanisms of rescue by VX-809 and the 15Jf are different. Only VX-809 stabilized TMD1 and only 15Jf promoted glycosylation of a TMD1+TMD2 mutant. Second, VX-809 promotes TMD1-NBD1 interactions at the CL1-NBD1 interface. Third, restoration of CL1-NBD1 interactions by VX-809 enhances the stability of TMD1.
Conflict of Interest

The authors declare no competing financial interests.

Acknowledgements.

This study was supported by grants from Cystic Fibrosis Canada (3014) and the Canadian Institutes for Health Research (Grant 62832).

References

⦁ Riordan, J.R. CFTR function and prospects for therapy. Annu. Rev. Biochem. 2008; 77:701- 726.
⦁ Loo, T.W., Bartlett, M.C., Clarke, D.M. Processing mutations disrupt interactions between the nucleotide binding and transmembrane domains of P-glycoprotein and the cystic fibrosis transmembrane conductance regulator (CFTR). J. Biol. Chem. 2008; 283:28190-28197.
⦁ Chen, E.Y., Bartlett, M.C., Loo, T.W., Clarke, D.M. The FF550088 mmuuttaattiioonn ddiissrruuppttss ppaacckkiinngg ooff the transmembrane segments of the cystic fibrosis transmembrane conductance regulator. J. Biol. Chem. 2004; 279:39620-39627.
⦁ Loo, T.W., Clarke, D.M. The cystic fibrosis V232D mutation inhibits CFTR maturation by disrupting a hydrophobic pocket rather than formation of aberrant interhelical hydrogen bonds. Biochem Pharmacol 2014; 88:46-57.
⦁ Farinha, C.M., King-Underwood, J., Sousa, M., Correia, A.R., Henriques, B.J., Roxo-Rosa, M., Da Paula, A.C., Williams, J., Hirst, S., Gomes, C.M., Amaral, M.D. Revertants, low temperature, and correctors reveal the mechanism of F508del-CFTR rescue by VX-809 and suggest multiple agents for full correction. Chem Biol 2013; 20:943-55.

⦁ Pedemonte, N., Lukacs, G.L., Du, K., Caci, E., Zegarra-Moran, O., Galietta, L.J., Verkman,

A.S. Small-molecule correctors of defective DeltaF508-CFTR cellular processing identified by high-throughput screening. J. Clin. Invest. 2005; 115:2564-2571.
⦁ Van Goor, F., Hadida, S., Grootenhuis, P.D., Burton, B., Stack, J.H., Straley, K.S., Decker, C.J., Miller, M., McCartney, J., Olson, E.R., Wine, J.J., Frizzell, R.A., Ashlock, M., Negulescu,
P.A. Correction of the F508del-CFTR protein processing defect in vitro by the investigational drug VX-809. Proc Natl Acad Sci U S A 2011; 108:18843-8.
⦁ Clancy, J.P., Rowe, S.M., Accurso, F.J., Aitken, M.L., Amin, R.S., Ashlock, M.A., Ballmann, M., Boyle, M.P., Bronsveld, I., Campbell, P.W., De Boeck, K., Donaldson, S.H., Dorkin, H.L., Dunitz, J.M., Durie, P.R., Jain, M., Leonard, A., McCoy, K.S., Moss, R.B., Pilewski, J.M., Rosenbluth, D.B., Rubenstein, R.C., Schechter, M.S., Botfield, M., Ordonez, C.L., Spencer-Green, G.T., Vernillet, L., Wisseh, S., Yen, K., Konstan, M.W. Results of a phase IIa study of VX-809, an investigational CFTR corrector compound, in subjects with cystic fibrosis homozygous for the F508del-CFTR mutation. Thorax 2012; 67:12-8.
⦁ Cholon, D.M., Quinney, N.L., Fulcher, M.L., Esther, C.R., Jr., Das, J., Dokholyan, N.V., Randell, S.H., Boucher, R.C., Gentzsch, M. Potentiator ivacaftor abrogates pharmacological correction of DeltaF508 CFTR in cystic fibrosis. Sci Transl Med 2014; 6:246ra96.
⦁ Veit, G., Avramescu, R.G., Perdomo, D., Phuan, P.W., Bagdany, M., Apaja, P.M., Borot, F., Szollosi, D., Wu, Y.S., Finkbeiner, W.E., Hegedus, T., Verkman, A.S., Lukacs, G.L. Some gating potentiators, including VX-770, diminish DeltaF508-CFTR functional expression. Sci Transl Med 2014; 6:246ra97.
⦁ Riordan, J.R., Rommens, J.M., Kerem, B., Alon, N., Rozmahel, R., Grzelczak, Z., Zielenski, J., Lok, S., Plavsic, N., Chou, J.L., Drumm, M.L., Iannuzzi, M.C., Collins, F.S., Tsui, L.-C.

Identification of the cystic fibrosis gene: cloning and characterization of complementary DNA. Science 1989; 245:1066-1073.
⦁ Quon, B.S., Rowe, S.M. New and emerging targeted therapies for cystic fibrosis. BMJ 2016; 352:i859.
⦁ Kleizen, B., van Vlijmen, T., de Jonge, H.R., Braakman, I. Folding of CFTR is predominantly cotranslational. Mol. Cell 2005; 20:277-287.
⦁ Du, K., Lukacs, G.L. Cooperative assembly and misfolding of CFTR domains in vivo. Mol Biol Cell 2009; 20:1903-15.
⦁ Wang, Y., Loo, T.W., Bartlett, M.C., Clarke, D.M. Modulating the folding of P- glycoprotein and cystic fibrosis transmembrane conductance regulator truncation mutants with pharmacological chaperones. Mol. Pharmacol. 2007; 71:751-758.
⦁ Sabirzhanova, I., Lopes Pacheco, M., Rapino, D., Grover, R., Handa, J.T., Guggino, W.B., Cebotaru, L. Rescuing Trafficking Mutants of the ATP-binding Cassette Protein, ABCA4, with Small Molecule Correctors as a Treatment for Stargardt Eye Disease. J Biol Chem 2015; 290:19743-55.
⦁ Loo, T.W., Bartlett, M.C., Shi, L., Clarke, D.M. Corrector-mediated rescue of misprocessed CFTR mutants can be reduced by the P-glycoprotein drug pump. Biochem Pharmacol 2012; 83:345-54.
⦁ He, L., Kota, P., Aleksandrov, A.A., Cui, L., Jensen, T., Dokholyan, N.V., Riordan, J.R. Correctors of DeltaF508 CFTR restore global conformational maturation without thermally stabilizing the mutant protein. FASEB J 2013; 27:536-45.
⦁ Loo, T.W., Bartlett, M.C., Clarke, D.M. Corrector VX-809 stabilizes the first transmembrane domain of CFTR. Biochem Pharmacol 2013; 86:612-9.

⦁ Ren, H.Y., Grove, D.E., De La Rosa, O., Houck, S.A., Sopha, P., Van Goor, F., Hoffman, B.J., Cyr, D.M. VX-809 corrects folding defects in cystic fibrosis transmembrane conductance regulator protein through action on membrane-spanning domain 1. Mol Biol Cell 2013; 24:3016- 24.
⦁ Eckford, P.D., Ramjeesingh, M., Molinski, S., Pasyk, S., Dekkers, J.F., Li, C., Ahmadi, S., Ip, W., Chung, T.E., Du, K., Yeger, H., Beekman, J., Gonska, T., Bear, C.E. VX-809 and related corrector compounds exhibit secondary activity stabilizing active F508del-CFTR after its partial rescue to the cell surface. Chem Biol 2014; 21:666-78.
⦁ Zhang, Z., Chen, J. Atomic Structure of the Cystic Fibrosis Transmembrane Conductance Regulator. Cell 2016; 167:1586-1597 e9.
⦁ Loo, T.W., Clarke, D.M. P-glycoprotein. Associations between domains and between domains and molecular chaperones. J. Biol. Chem. 1995; 270:21839-21844.
⦁ Kunkel, T.A. Rapid and efficient site-specific mutagenesis without phenotypic selection.

Proc. Natl. Acad. Sci. U. S. A. 1985; 82:488-492.

⦁ Loo, T.W., Clarke, D.M. Using a cysteine-less mutant to provide insight into the structure and mechanism of CFTR. J. Physiol. 2006; 572:312 (Commentary).
⦁ Cui, L., Aleksandrov, L.A., Hou, Y.X., Gentzsch, M., Chen, J.H., Riordan, J.R., Aleksandrov, A.A. The role of cystic fibrosis transmembrane conductance regulator phenylalanine 508 side chain in ion channel gating. J. Physiol. 2006; 572:347-358.
⦁ Loo, T.W., Clarke, D.M. Functional consequences of proline mutations in the predicted transmembrane domain of P-glycoprotein. J Biol Chem 1993; 268:3143-9.
⦁ Loo, T.W., Bartlett, M.C., Clarke, D.M. Bithiazole correctors rescue CFTR mutants by two different mechanisms. Biochemistry 2013; 52:5161-3.

⦁ Gorman, C.M., Howard, B.H., Reeves, R. Expression of recombinant plasmids in mammalian cells is enhanced by sodium butyrate. Nucleic Acids Res. 1983; 11:7631-48.
⦁ Coffman, K.C., Nguyen, H.H., Phuan, P.W., Hudson, B.M., Yu, G.J., Bagdasarian, A.L., Montgomery, D., Lodewyk, M.W., Yang, B., Yoo, C.L., Verkman, A.S., Tantillo, D.J., Kurth,
M.J. Constrained bithiazoles: small molecule correctors of defective DeltaF508-CFTR protein trafficking. J Med Chem 2014; 57:6729-38.
⦁ Wang, Y., Loo, T.W., Bartlett, M.C., Clarke, D.M. Correctors Promote Maturation of Cystic Fibrosis Transmembrane Conductance Regulator (CFTR)-processing Mutants by Binding to the Protein. J. Biol. Chem. 2007; 282:33247-33251.
⦁ Loo, T.W., Clarke, D.M. The glycosylation and orientation in the membrane of the third cytoplasmic loop of human P-glycoprotein is affected by mutations and substrates. Biochemistry 1999; 38:5124-9.
⦁ Serohijos, A.W., Hegedus, T., Aleksandrov, A.A., He, L., Cui, L., Dokholyan, N.V., Riordan, J.R. Phenylalanine-508 mediates a cytoplasmic-membrane domain contact in the CFTR 3D structure crucial to assembly and channel function. Proc. Natl. Acad. Sci. U. S. A. 2008; 105:3256-3261.
⦁ Loo, T.W., Bartlett, M.C., Clarke, D.M. Correctors enhance maturation of DeltaF508 CFTR by promoting interactions between the two halves of the molecule. Biochemistry 2009; 48:9882- 9890.
⦁ He, L., Aleksandrov, A.A., Serohijos, A.W., Hegedus, T., Aleksandrov, L.A., Cui, L., Dokholyan, N.V., Riordan, J.R. Multiple membrane-cytoplasmic domain contacts in cftr mediate regulation of channel gating. J. Biol. Chem. 2008; 283:26383-26390.

⦁ Aleksandrov, A.A., Kota, P., Aleksandrov, L.A., He, L., Jensen, T., Cui, L., Gentzsch, M., Dokholyan, N.V., Riordan, J.R. Regulatory insertion removal restores maturation, stability and function of DeltaF508 CFTR. J Mol Biol 2010; 401:194-210.
⦁ Okiyoneda, T., Veit, G., Dekkers, J.F., Bagdany, M., Soya, N., Xu, H., Roldan, A., Verkman, A.S., Kurth, M., Simon, A., Hegedus, T., Beekman, J.M., Lukacs, G.L. Mechanism- based corrector combination restores DeltaF508-CFTR folding and function. Nat Chem Biol 2013; 9:444-54.
⦁ Loo, T.W., Bartlett, M.C., Clarke, D.M. The V510D suppressor mutation stabilizes DeltaF508-CFTR at the cell surface. Biochemistry 2010; 49:6352-7.
⦁ Ehrhardt, A., Chung, W.J., Pyle, L.C., Wang, W., Nowotarski, K., Mulvihill, C.M., Ramjeesingh, M., Hong, J., Velu, S.E., Lewis, H.A., Atwell, S., Aller, S., Bear, C.E., Lukacs, G.L., Kirk, K.L., Sorscher, E.J. Channel Gating Regulation by the Cystic Fibrosis Transmembrane Conductance Regulator (CFTR) First Cytosolic Loop. J Biol Chem 2016; 291:1854-65.
⦁ Loo, T.W., Clarke, D.M. Identification of the distance between the homologous halves of P- glycoprotein that triggers the high/low ATPase activity switch. J Biol Chem 2014; 289:8484-92.
⦁ Lewis, H.A., Buchanan, S.G., Burley, S.K., Conners, K., Dickey, M., Dorwart, M., Fowler, R., Gao, X., Guggino, W.B., Hendrickson, W.A., Hunt, J.F., Kearins, M.C., Lorimer, D., Maloney, P.C., Post, K.W., Rajashankar, K.R., Rutter, M.E., Sauder, J.M., Shriver, S., Thibodeau, P.H., Thomas, P.J., Zhang, M., Zhao, X., Emtage, S. Structure of nucleotide-binding domain 1 of the cystic fibrosis transmembrane conductance regulator. Embo J. 2004; 23:282- 293.

⦁ Laselva, O., Molinski, S., Casavola, V., Bear, C.E. The investigational Cystic Fibrosis drug Trimethylangelicin directly modulates CFTR by stabilizing the first membrane-spanning domain. Biochem Pharmacol 2016; 119:85-92.
⦁ Favia, M., Mancini, M.T., Bezzerri, V., Guerra, L., Laselva, O., Abbattiscianni, A.C., Debellis, L., Reshkin, S.J., Gambari, R., Cabrini, G., Casavola, V. Trimethylangelicin promotes the functional rescue of mutant F508del CFTR protein in cystic fibrosis airway cells. Am J Physiol Lung Cell Mol Physiol 2014; 307:L48-61.
⦁ Tamanini, A., Borgatti, M., Finotti, A., Piccagli, L., Bezzerri, V., Favia, M., Guerra, L., Lampronti, I., Bianchi, N., Dall’Acqua, F., Vedaldi, D., Salvador, A., Fabbri, E., Mancini, I., Nicolis, E., Casavola, V., Cabrini, G., Gambari, R. Trimethylangelicin reduces IL-8 transcription and potentiates CFTR function. Am J Physiol Lung Cell Mol Physiol 2011; 300:L380-90.
⦁ Grove, D.E., Rosser, M.F., Ren, H.Y., Naren, A.P., Cyr, D.M. Mechanisms for rescue of correctable folding defects in CFTRDelta F508. Mol Biol Cell 2009; 20:4059-69.

Figure Legends

Fig. 1. VX-809 promotes maturation of TMD1 in the presence of TMD1. (A) The boundaries of the five domains (TMD1, NBD1, R, TMD2 and NBD2) of CFTR and the composition of truncation mutants TMD1, TMD and TMD1+2 are shown. An A52 epitope tag was added to the C-terminal end of TMD1. The cysteines (R170C and L475C) introduced into Cys-less constructs used for cross-linking are shown. The numbered and filled rectangles represent TM segments. The branched lines between TM segments 7 and 8 represent N-glycosylation sites. (B) Rescue of CFTR TMD by correctors and TMD1. HEK 293 cells were transfected with cDNAs encoding TMD1 plus A52-tagged TMD1 (+TMD1) or TMD1 alone (-TMD1). The cells were incubated in the absence (None) or presence of 5 M 15JF or VX-809 for 48h. Whole cell SDS extracts were subjected to immunoblot analysis. The positions of mature and immature forms of
TMD1 are indicated. The amount of mature TMD1 relative to total (immature plus mature) was determined. Each value is the mean ± S.D. (n = 4). An asterisk indicates a significant (P < 0.001) increase when each sample treated with corrector was compared to expression in the absence of correctors. (C) To determine the glycosylation state of TMD1, it was expressed TMD1-A52 for 24h in the presence or absence of 5 M VX-809. Samples of SDS cell extracts were treated without (-) or with endoglycosidases F (F) or H (H) prior to immunoblot analysis. The positions of immature, mature and unglycosylated (Unglycos) forms of TMD1 are shown.

Fig. 2. The mechanisms of VX-809 and 15Jf rescue are different. (A) HEK 293 cells were transfected with a cDNA encoding A52-tagged TMD1 (residues 1-436). The cells were incubated for 18 h in the presence of DMSO (None), 5 M VX-809 or 15Jf. Protein synthesis was then inhibited by addition of 0.5 mg/ml cycloheximide with DMSO (None), VX-809 or 15Jf. At various intervals the cells were harvested and samples of whole cell SDS extracts were subjected to immunoblot analysis with monoclonal antibody A52. The amount of 45 kDa TMD1- A52 protein at each time point was quantitated and expressed relative to time 0 (% Remaining). The open circles, filled circles and open squares represent None, 15Jf and VX-809, respectively. An asterisk indicates a significant (P < 0.001) increase when each sample treated with corrector VX-809 at a given time point was compared to a similar sample not treated with corrector. A one-way ANOVA with post-hoc Tukey test showed showed that treatment with VX-809 was significantly (P < 0.01; ++) different from the group treated with 15JF or DMSO (None) (B) HEK 293 cells were transfected with A52-tagged TMD1+TMD2 truncation mutant lacking the R and NBD domains. Cells were expressed in the absence (None) or presence of 5 M VX-809 or 15Jf. Whole cell SDS extracts were subjected to immunoblot analysis. The amount of glycosylated TMD1+TMD2 protein relative to total TMD1+TMD2 (Glycos plus Unglycos) was quantitated. Each value is the mean ± S.D. (n = 4). An asterisk indicates a significant (P < 0.001) increase when compared to expression in the absence of correctors. A sample of the 15Jf cell extract was also treated without (-) or with (+) endoglycosidase H (Endo H) prior to immunoblot analysis (C). The positions of unglycosylated (Unglycos) and core-glycosylated (Glycos) TMD1+TMD2 are indicated.

Fig. 3. Corrector VX-809 promotes cross-linking between CL1/NBD1 cysteines. (A) Secondary structure of CFTR showing predicted contacts between the cytoplasmic loops (CLs) and NBDs. Branched lines between TM segments 7 and 8 indicate the glycosylation sites. Cylinders represent TM segments. (B) HEK 293 cells were transiently transfected with L475C TMD1 together with A52-tagged Cys-less (C-less) or R170C TMD1. The cells were incubated for 18 h in the absence (-) or presence (+) of 5 M VX-809. Membranes were prepared and suspended in TBS, pH 7.4. Samples were then treated with or without 1 mM copper phenanthroline (CuP) for 5 min at 0 oC and the reactions were stopped by addition of 10 mM EDTA followed by addition of SDS sample buffer containing no thiol reducing agent. Samples were subjected to immunoblot analysis with a rabbit anti-CFTR antibody. The positions of immature (I), mature (M) and cross- linked (X) proteins are indicated. The % cross-linked represents the amount of cross-linked protein relative to total (mature (M) plus cross-linked (X)). (C) The A52-tagged R170C TMD1 + L475C TMD1 samples (from Fig. 3B, middle and right panels) were also treated without (-) and with (+) copper phenanthroline (CuP) and then without (-) or with (+) dithiothreitol (DTT). They were then subjected to immunoblot analysis with A52 to detect the TMD1 protein. The % cross-linked represents the amount of cross-linked protein relative to total (X+T) protein. Each value is the mean ± S.D. (n = 4). An asterisk indicates a significant (P < 0.001) increase in cross- linked product when compared to the amount of cross-linked product after expression in the absence of VX-809.

Fig. 4. Expression in the presence of VX-809 plus TMD1 stabilizes mature TMD1. Cells were transfected with cDNAs encoding A52-tagged TMD1 (residues 1-436) and untagged TMD1 (residues 437-1480). One set of cells was incubated for 48 h at 30 oC to promote low temperature maturation of TMD1 (-809). The other set of cells (+809) was incubated for 48 h at 37 oC in the

presence of 5 M VX-809 (+809). Protein synthesis was then inhibited by addition of cycloheximide with or without VX-809 and harvested at the indicated times and whole cell SDS extracts subjected for immunoblot analysis with rabbit polyclonal antibody against NBD2 of CFTR. The positions of mature and immature forms of TMD1 are indicated. The amount of mature TMD1 protein remaining was quantitated and expressed relative to time 0. Each value is the mean ± S.D. (n = 4). An asterisk indicates a significant (P < 0.001) increase when each sample treated with VX-809 at each time point was compared to a similar sample not treated with corrector. A one-way ANOVA test showed that treatment with VX-809 was significantly (P
< 0.01; ++) different from the group treated with DMSO (-VX-809).

Fig. 5. Effect of VX-809 or VX-770 on NBD-TMD cross-linking. (A) Membranes prepared from cells expressing full-length V510C/A1067C CFTR were pre-incubated in the presence of DMSO (-) or 20 M VX-809 (+). Samples were then treated for 5 min at 0 oC with (+) or without (-) 1 mM copper phenanthroline (CuP). The reactions were stopped by addition of 2X SDS sample buffer containing 25 mM EDTA but no thiol reducing agent. Samples were subjected to immunoblot analysis. The amount of cross-linked (X-link) protein relative to total (mature plus cross-linked) of each sample was determined. Each value is the mean + S.D. (n=4). There was no significant difference (P > 0.1) in the amount of cross-linked product in the presence or absence of VX-809. (B) Membranes prepared from cells co-expressing R170C TMD1 plus L475C
TMD1 were pre-incubated in the presence of DMSO (-) or 20 M VX-770 plus 20 M VX-809 (+; VX-770/809). Samples were then treated for 5 min at 0 oC with (+) or without (-) 1 mM copper phenanthroline (CuP). The reactions were stopped by addition of 2X SDS sample buffer containing 25 mM EDTA but no thiol reducing agent. Samples were subjected to immunoblot analysis. The amount of cross-linked (X-link) protein relative to total (mature plus cross-linked)

of each sample was determined. Each value is the mean + S.D. (n=4). There was no significant difference (P > 0.1) in the level of cross-linking in the presence or absence of VX-770/809.

% Mature
None 15J
809
None 15J
809

% Remaining
% Glycos
None
VX-809 15JF

% Cross-linked
% Cross-linked