The CFTR-Associated Ligand Arrests the Trafficking of the Mutant ΔF508 CFTR Channel in the ER Contributing to Cystic Fibrosis Open Access

Cystic Fibrosis (CF) is a common lethal autosomal recessive disorder caused by mutations in the Cystic Fibrosis Transmembrane Conductance Regulator (CFTR) gene [1, 2]. This gene encodes a cAMP-regulated chloride channel that is present on the apical membrane of epithelial cells lining the organs of the lungs, liver, pancreas, sweat ducts, intestines, and reproductive systems. Mutations in CFTR impair the salt and water balance of these tissues leading to host of clinical problems brought about by dehydrated secretions, and the buildup of thick mucus [3, 4]. CF airways are vulnerable to infection and inflammation, pancreatic ducts and the intestines are obstructed, sweat chloride levels are high, and male infertility is common. As a membrane protein, CFTR belongs to the ATP-binding cassette transporter family. Its correct function is dependent upon proper synthesis, maturation, and trafficking from the ER and Golgi to the plasma membrane.

There are over 2000 mutations that cause CF [5]. The most common mutation among patients is ΔF508, and approximately 90% of the patient population are carriers [6]. ΔF508 CFTR is a misfolded protein that is recognized as such by the ER quality control system. The protein is polyubiquitilated, translocated to the cytosol, and prematurely degraded by the 26S proteasome [7, 8]. Thus, the majority of ΔF508 CFTR is eliminated by the cell and cannot mature or reach the cell surface. As therapies for ΔF508 are important for the health of CF patients, targeting this protein remains the greatest clinical and research problem within the field.

CAL, also known as PIST (PDZ domain protein interacting specifically with TC10), GOPC (Golgi-associated PDZ and coiled-coil motif containing), and FIG (fused in glioblastoma) is a PDZ domain containing protein with two coiled-coil domains that interacts with the C-terminus of CFTR [9, 10]. Our lab has previously shown that CAL associates with the Golgi apparatus and reduces cell surface WT CFTR through promotion of lysosomal degradation [9, 11]. The effect of CAL on WT CFTR can be affected through a variety of regulator proteins. TC10, a small rho GTPase, binds to CAL in the region of its coiled-coil domain and promotes trafficking of WT CFTR to the cell surface when it is activated and in a GTP-bound state [12]. In contrast, Syntaxin 6 (STX6), a soluble N-ethylmaleimide-sensitive factor-activating receptor protein (SNARE), binds to CAL forming a complex that recruits E3 ubiquitin ligase MARCH2 and promotes transport of WT CFTR to the lysosome for degradation [13, 14]. Additionally, the effects of CAL on WT CFTR can be reversed through the overexpression of Na+/H+ exchanger regulatory factor 1 (NHERF1), a protein that anchors CFTR to the cytoskeleton [9].

While CAL’s interactions with WT CFTR are well characterized, very little is known about its ability to regulate ΔF508 CFTR. Given that changes in CAL have been shown to improve ΔF508 CFTR chloride current [15], we investigated the mechanism by which CAL can improve ΔF508 CFTR function and plasma membrane trafficking. In this article, we demonstrate that CAL plays a significant role in the early trafficking of ΔF508 CFTR, regulating availability of cell surface ΔF508 CFTR and influencing ER retention.

MG-132 was obtained from Sigma Aldrich. The HA-CAL plasmid and the GFP-ΔF508 CFTR plasmid were made in the Guggino lab. The KDEL-td-tomato plasmid was a generous gift from Dr. Carolyn Machamer (Johns Hopkins University School of Medicine, Baltimore, MD). The following primary antibodies were used: mouse monoclonal anti-CFTR 217/596/769 (Laboratory of Jack Riordan, University of North Carolina, Chapel Hill); mouse monoclonal anti-CFTR M3A7 (Millipore); mouse monoclonal anti-GFP (Roche); HA-probe (sc-7392), VCP (sc-133125), HSP70 (sc-66048), HSP40 (sc-59554), NHERF1, and Ezrin (sc-58758) were purchased from Santa Cruz Biotechnology; rabbit monoclonal anti-PIST (Abcam); and mouse monoclonal anti-AHSA1 (Abnova). Secondary antibodies used for immunofluorescence were conjugated to Alexa 488 and Alexa 594 (Invitrogen). CAL specific siRNA was purchased from Qiagen (GOPC_9).

CFBE41o- cells (derived by Dieter Gruenert; a cell line from bronchial epithelial cells of a cystic fibrosis patient), also called parental cells, were cultured in minimal essential medium Eagle (MEM, Invitrogen) with 10% fetal bovine serum (FBS, Invitrogen), penicillin/streptomycin (Invitrogen), and glutamine (Invitrogen). CFBE-ΔF508CFTR cells (CFBE41o- cells stably expressing ΔF508 CFTR) were cultured in MEM with 10% FBS, penicillin/streptomycin, glutamine, and puromycin. CFBE-WTCFTR cells (CFBE41o- cells stably expressing wildtype CFTR) were cultured in MEM with 10% FBS, penicillin/streptomycin, glutamine, and hygromycin B. African green monkey kidney cell line COS-7 was maintained in DMEM with 10% FBS, penicillin/streptomycin, and L-glutamine. All cell lines were stored in a humidified 37°C CO₂ incubator. For plasmid transfection, cells were transiently transfected using 4μg of plasmid DNA and lipofectamine 2000 (Invitrogen) according to the manufacturer’s instruction, and incubated for 48 hours. For knockdown, cells were transfected with 20nM siRNA using Interferin (Polyplus Transfection) and incubated for 48 hours.

Cells were washed twice with cold Dulbecco’s phosphate buffered saline (DPBS, Invitrogen) and then resuspended in lysis buffer (150 mM Tris HCl pH 7.5, 50 mM NaCl, 1% Nonidet P-40) with Halt Protease Inhibitor Cocktail and EDTA Solution (Thermo Fisher) added. After 30 minutes rotating at 4°C, the lysates were cleared by centrifugation at 15, 000 g for 15 minutes. Pellets were discarded, and supernatants were collected for further experimental use. Protein A/G Plus-Agarose beads (Santa Cruz Biotechnology) were mixed with lysates, and primary antibodies, and rotated overnight at 4°C. Beads were then washed five times with lysis buffer and incubated in Laemmli sample buffer at 37°C for 20 minutes. Total lysates were denatured in Laemmli sample buffer at 37°C for 20 minutes. The protein concentration of each sample was measured and equal amounts of total proteins were loaded onto 10% HCl-Tris gels. Proteins were separated by SDS-PAGE and transferred to PVDF membranes (BioRad). After transfer, membranes were blocked with 5% milk in tris-buffered saline with .05% Tween (TBST) for 1 hour. The membranes were incubated with primary antibodies in 5% milk in TBST at 4°C overnight. Membranes were then washed 3 times in TBST and incubated with the appropriate horseradish peroxidase-conjugated secondary antibody in 5% milk in TBST for an hour. Following secondary incubation, membranes were washed 3 times and incubated with Super Signal West Dura Extended Duration Substrate (Thermo Fisher) for two minutes. Results were visualized using a Fujifilm LAS4000 developer.

Cells were washed twice with cold DPBS containing 1 mM Ca²⁺ and 1 mM Mg²⁺. Following the wash, cells were incubated with 0.5 mg/ml sulfo-NHS-SS-biotin (Thermo Scientific) in DPBS with Ca²⁺ and Mg²⁺ at 4°C for 30 minutes. Free sulfo-NHS-SS-biotin was then quenched by washing with cold 50 mM glycine in DPBS three times. Cells were subsequently washed twice with DPBS containing 1 mM Ca²⁺ and 1 mM Mg²⁺ and twice with cold DPBS. Following these washes, cells were lysed with lysis buffer, rotated for 30 minutes, and the lysates were centrifuged at 15, 000 g at 4°C for 15 minutes. The supernatants were incubated with NeutrAvidin UltraLink Resin (Thermo Scientific) at 4°C for 1 hour. The beads were washed four times with lysis buffer, and proteins were eluted with Laemmli sample buffer at 37°C for 20 minutes. Biotinylated protein samples were analyzed by western blotting as described above.

Zeiss LSM 510 laser scanning system and 63x oil-immersion lens were used. Cells were seeded onto cover glasses and later transiently transfected with various plasmids, as well as a CAL specific siRNA for 48 hours. The following steps were done at room temperature: cells were fixed with 4% formaldehyde for 15 minutes. Subsequently, cells were permeabilized with 0.3% Triton X-100 in DPBS for 5-7 minutes and blocked with 3% bovine albumin serum (BSA) for 45 minutes. After blocking, cells were incubated with primary antibodies and then with secondary antibodies conjugated to Alexa488 or Alexa 594. Cells were incubated with a 1: 1000 dilution of DAPI for 5 minutes and mounted using ProLong Gold Antifade (Invitrogen). Images were cropped, adjusted, and exported as TIFs using Imaris Imaging Software. Figs. for publishing were put together using Adobe Photoshop. For quantification of colocalization, Pearson’s correlation coefficients were calculated using Imaris Imaging Software.

Short-circuit current (I_SC) measurements were performed in a six-channel Easy-Mount chamber system (Physiologic Instruments, San Diego, CA) that accepts Snapwell filters (Corning Costar, Acton, MA; 3407). I_SC was measured with a VCCMC6 multichannel voltage-current clamp amplifier (Physiologic Instruments) in the voltage-clamp mode. Data were acquired on an 1.71-GHz PC running Windows XP (Microsoft, Redmond, WA) and equipped with DI-720 (DATAQ Instruments, Akron, OH), with Acquire and Analyze version 2.3.159 (Physiologic Instruments) software. Cells were cultured to confluence on Snapwell filters before measurement. The cell monolayers were bathed on both sides with solution containing 115 mM NaCl, 25 mM sodium gluconate, 5 mM potassium gluconate, 1.2 mM MgCl₂, 1.2 mM CaCl₂, 10 mM D-glucose and 10 mM HEPES (pH 7.4 with NaOH). The mucosal side was replaced with a low Cl^– solution containing 139 mM sodium gluconate, 1.2 mM NaCl, 5 mM potassium gluconate, 1.2 mM MgCl₂, 1.2 mM CaCl₂, 10 mM D-glucose, and 10 mM HEPES (pH 7.4 with NaOH). The solution was constantly circulated, maintained at 37°C, and bubbled gently with air. Amiloride (20 µM) was added to the mucosal solution, and after stabilization, forskolin (10 µM) was added to the serosal chamber, followed by the CFTR channel inhibitor CFTRinh-172 (5 µM). The short circuit current is measured with a basolateral to apical chloride gradient to enhance the Cl^- transport out of the cells across the apical membrane to measure CFTR-dependent current more accurately.

Data are displayed in bar graphs using Graphpad software. All quantitative results are given as means ± standard error (SE) which are represented by lines extending from each bar. Means and SEs are calculated from at least three independent experiments. Values of p<0.05 were considered significant.

Changes in intracellular levels of ΔF508 CFTR can be quantified through measurement of the CFTR B band and C band seen on western blots. The B band, which is approximately 150 kDa in size, represents the immature core-glycosylated form of ΔF508 CFTR. The C band, which is 170 kDa, represents the mature complex glycosylated form of ΔF508 CFTR [16]. A previous study reported that CAL knockdown enhanced ΔF508 CFTR maturation and chloride currents. The effect of CAL silencing was shown to be specific for ΔF508-CFTR and did not affect the expression of BCRP, another ABC transporter encoded by the abcg2 gene or the α-subunit of Na/K ATPase [15]. We extended these observations by analyzing changes in the B band and C band of ΔF508 CFTR when CAL is inhibited in ΔF508-CFBE cells.

ΔF508-CFBE cells were transiently transfected with CAL specific siRNA and collected lysates were analyzed by western blot. The results of the western blot in Fig. 1A and summarized in Fig. 1 B-D confirm that after transfection with CAL specific siRNA, the expression of CAL protein is successfully reduced while ΔF508 CFTR B band and C band increase in the total lysate (Fig. 1A-D). A negative scrambled control and a lipofectamine mock transfection control were also tested and shown to have no effect on expression levels of ΔF508 CFTR (Fig. 1F-G).

Results were then quantified for statistical analysis. After CAL knockdown, the immature B band of ΔF508 CFTR increases about 6 times compared to the control (Fig. 1B). The mature C band ΔF508 CFTR increases similarly compared to the control (Fig. 1C). These results were shown to be statistically significant (Fig. 1B, C). On the other hand, the ratio of B or C bands to the total CFTR did not change indicating that both changed in the same proportion (1D) CAL knockdown by siRNA was determined to be sufficient and effective via quantification, with percentage of median knockdown being 76% (Fig. 1E). Our data confirm that CAL knockdown in ΔF508-CFBE cells increases the immature B band and mature C band of ΔF508 CFTR to statistically significant levels (Fig. 1).

Since our data indicate that CAL can regulate levels of ΔF508 CFTR in total lysate, we next examined if CAL knockdown directly or indirectly affected ΔF508 CFTR trafficking through analysis of the binding relationship between the two proteins. A previous study has shown that CAL can bind to WT CFTR in order to regulate it’s trafficking to the lysosome, as opposed to the plasma membrane [9]. Here, COS-7 cells were transiently transfected with various GFP-tagged CFTR plasmids or HA-CAL and collected total lysates were subjected to co-immunoprecipitation. Samples from total lysate and the binding fractions were then run on western blots for analysis.

In Fig. 2A, there are five lanes for assessment. Lane 1 is the untransfected control, and as such, no CFTR appears in the total lysate (top) or Co-IP fractions (bottom). Since CFTR is not present, CAL is not observed in the Co-IP fraction (bottom). Given that an anti-CAL antibody was used, there is some endogenous level of CAL in COS-7 cells so a faint band is observed in the total lysate (top). Lane 2 has been transfected with GFP-WTCFTR plasmid and pulled down with an anti-GFP antibody. Thus, WT-CFTR is observed in the total lysate (top) and Co-IP fractions (bottom) as detected by an anti-GFP antibody. Levels of endogenous CAL are so low that no binding is visible.

Lane 3 has been co-transfected with GFP-WTCFTR and HA-CAL. Enhancing intracellular levels of CAL through transfection results in a visible band for CAL in the Co-IP fraction indicating that WT CFTR and HA-CAL are bound. Lane 4 has been transfected with GFP-ΔF508CFTR plasmid. Thus, ΔF508 CFTR is observed in the total lysate (top) and Co-IP fractions (bottom) but no binding to CAL is visible. Finally, lane 5 has been co-transfected with GFP-ΔF508CFTR and HA-CAL. Here, binding of ΔF508 CFTR and CAL is visible in the Co-IP fraction (bottom). These results indicate that not only does CAL bind to wt-CFTR as was shown previously [9] but also to ΔF508 CFTR (Fig. 2A).

To establish the cell compartment in which CAL binds ΔF508 CFTR in ΔF508-CFBE cells, a series of co-immunofluorescence experiments were performed (Fig. 2B). ΔF508-CFBE cells were transfected with KDEL-td-tomato plasmid (a gift from Dr. Carolyn Machamer) and examined by fluorescence microscopy. CAL is shown to colocalize with KDEL, indicating that CAL localizes at the ER (Fig. 2B). These results suggest that some CAL interacts with ΔF508 CFTR at the level of the ER.

Next, to establish the site of CAL degradation in ΔF508-CFBE cells, we investigated various organelles linked to protein degradation. Previous studies focused on WT CFTR tightly couple CAL degradation to the lysosome, as it directs WT CFTR there for degradation [9, 11]. In Fig. 2C, ΔF508-CFBE cells were transfected with HA-tagged CAL and treated with MG-132, an inhibitor of the 26S proteasome. MG-132 visibly increased the B band and C band of ΔF508 CFTR to as similar amount in CAL transfected and untransfected cells (Fig. 2D & E). Treatment with MG-132 increased levels of CAL to 1.5 times that of the control, indicating that CAL is degraded in the 26S proteasome (Fig. 2F). Following treatment with bafilomycin and tubacin, no changes in CAL were observed (Fig. 3). This result indicates that CAL degradation is not lysosomal or aggresomal in ΔF508-CFBE cells.

Since surface expression of mature CFTR is key to correction of the mutant protein, we sought to determine if CAL knockdown could produce mature cell surface expressed ΔF508 CFTR. It is thought that very little ΔF508 CFTR reaches the cell surface, as most of it is prematurely degraded in the proteasome. However, a small fraction of ΔF508 CFTR is able to exit the ER and escape the ER-associated degradation (ERAD) pathway but is not stable at the cell periphery [17]. To determine if CAL knockdown affects ΔF508 CFTR surface expression, we preformed several cell surface biotinylation.

In Fig. 4A, mature cell surface C band was observed in the biotinylation fraction of the WT sample (top panel; Lane 1). Additionally, a very faint band representing immature cell surface B band was observed in the untransfected control ΔF508-CFBE sample (top panel; Lane 2).

Most importantly, Fig. 4A shows that in CAL siRNA transfected ΔF508-CFBE cells, both mature C band and immature B band are present on the cell surface (top panel; Lane 3, black box). The de novo C band was quantified, compared to the null, and found to be statistically significant (Fig. 4B). This result is critical, as it shows that the CAL protein is integral to regulating the cell surface expression of ΔF508 CFTR, since knockdown of CAL led to the de novo production of cell surface C band. The presence of B band on the surface when CAL is inhibited indicates that some fraction of the intracellular immature ΔF508 CFTR can traffic to the surface when freed from CAL binding at the ER. We have also observed previously (see [12] Fig. 1A) that even with wt-CFTR a small leakage of B band to the cell surface occurs particularly when processing of wt-CFTR is enhanced by activation of TC10. One suggestion is that core glycosylated, B band, form of ΔF508-CFTR traffics via a PDZ domain interaction with GRASP, a Golgi reassembly stacking protein [18]. During periods of ER stress, CFTR binds to GRASP increasing the trafficking of the B band to the plasma membrane. Perhaps in the absence of CAL, GRASP can more readily bind to CFTR. Although a possibility, precisely how the immature B band reaches the plasma membrane in our experimental circumstances will require further investigation.

To determine how knockdown of CAL can affect the progression of ΔF508 CFTR trafficking, we set out to assess colocalization of ΔF508 CFTR with the ER, Golgi, and plasma membrane. Prior studies have already established the localization of WT CFTR and CAL relative to cellular organelles [9, 11]. Control experiments show that CAL and CFTR colocalize in ΔF508-CFBE cells (Fig. 5A), while CAL knockdown results in decreased colocalization between CAL and ΔF508 CFTR as expected. Additionally, CAL knockdown is shown to affect the distribution of ΔF508 CFTR in the cell, resulting in greater spread throughout the cell and less concentration around the ER (Fig. 5A). We then used Imaris Imaging Software to calculate Pearson’s correlation coefficients to determine the quantitative colocalization of ΔF508 CFTR and CAL when CAL is inhibited compared to controls. When CAL is inhibited by siRNA, Pearson’s correlation coefficients decrease significantly (Fig. 5B).

When ΔF508 CFTR colocalization with KDEL was assessed, we observed a high level of colocalization between the two proteins (Fig. 5C). After transfection with CAL siRNA, colocalization of ΔF508 CFTR with KDEL decreased and improved the distribution of ΔF508 CFTR throughout the cell (Fig. 5C). These results demonstrate that CAL knockdown affects ΔF508 CFTR distribution and results in the release of ΔF508 CFTR from the ER, enhancing its trafficking to the cell surface. We then used Imaris to calculate Pearson’s correlation coefficients to determine the quantitative colocalization of ΔF508 CFTR and KDEL when CAL is inhibited compared to controls. When CAL is inhibited by siRNA, Pearson’s correlation coefficients decrease significantly (Fig. 5D).

We subsequently sought to establish if CAL knockdown would improve ΔF508 CFTR maturation and cell surface expression by observing how much ΔF508 CFTR colocalizes with markers for the Golgi and plasma membrane. Data from untransfected control ΔF508-CFBE cells shows that very little ΔF508 CFTR colocalizes with the Golgi (Fig. 6A). This is consistent with other studies. However, after CAL knockdown, we observed that the level of colocalization between the Golgi and ΔF508 CFTR increased (Fig. 6A).We calculated Pearson’s correlation coefficients to determine the quantitative colocalization of ΔF508 CFTR and Golgin-97 when CAL is inhibited and noted a corresponding statistically significant increase in the coefficient (Fig. 6B).

As for plasma membrane colocalization, data from untransfected control ΔF508-CFBE cells shows that no ΔF508 CFTR reaches the plasma membrane (Fig. 6C). This is consistent with the rest of the data presented in this article. Interestingly, after CAL knockdown, colocalization between plasma membrane marker cadherin and ΔF508 CFTR is observed (Fig. 6C; yellow boxes with white arrows). When Pearson’s correlation coefficients were calculated, we observed a significant increase in quantitative colocalization between ΔF508 CFTR and cadherin (Fig. 6D).

Chaperone proteins are integral to proper cellular function and viability due to the various roles they play within the cell. One of the most well characterized functions of chaperones is their ability to fold and transport other immature intracellular proteins. Without chaperones, plasma membrane expressed proteins such as CFTR would not reach the cell surface.

In order to elucidate the mechanism by which CAL knockdown causes the production of mature cell surface ΔF508 CFTR, we selected a set of chaperone proteins for assessment that are known to interact with both WT CFTR and ΔF508 CFTR. Chaperone proteins of interest included the following: HSP27, HSP40, HSP70, HSP90, HDAC6, HDAC7, VCP, and Aha1. In the experiments that followed, ΔF508-CFBE cells were transfected with CAL siRNA and cell lysates were collected for co-immunoprecipitation analysis (see Fig. 7 for the negative control).

No changes in binding were observed for HSP27, and HDAC6 (Fig. 7A-C). However, when CAL is inhibited by siRNA, small but no significant changes were observed in ΔF508 CFTR binding to HSP40, Aha1, and HSP70 (Fig. 8A & B). Importantly the binding to HSP90 increased and to VCP decreased (Fig. 8A). These results were statistically significant (Fig. 8B). Our data reveals that cell surface expression of ΔF508 CFTR due to CAL knockdown is driven by increases in pro-folding chaperones such as HSP90 and decreases in anti-folding chaperones such as VCP. However, changes in HSP40 are neither pro nor anti-folding but correlate to partner protein HSP70. These results confirm that CAL knockdown improves ER to Golgi trafficking and subsequent protein maturation through enhanced binding to profolding chaperones.

Changes in NHERF1 were also measured under conditions of CAL knockdown. NHERF1 is a scaffold protein that serves to anchor CFTR to the cytoskeleton through a complex of proteins [19]. We observed a statistically significant increase in NHERF1 in the total lysate of cells that had been transfected with CAL siRNA (Fig. 7). This data shows that plasma membrane scaffolding protein NHERF1 exists in greater quantity intracellularly when CAL is inhibited, reflecting the need to stabilize a population of mature ΔF508 CFTR at the plasma membrane.

Changes in NHERF1 were also measured under conditions of CAL knockdown. NHERF1 is a scaffold protein that serves to anchor CFTR to the cytoskeleton through a complex of proteins [19]. We observed a statistically significant increase in NHERF1 in the total lysate of cells that had been transfected with CAL siRNA (Fig. 9). This data shows that plasma membrane scaffolding protein NHERF1 exists in greater quantity intracellularly when CAL is inhibited, reflecting the need to stabilize a population of mature ΔF508 CFTR at the plasma membrane.

Since we were able to show that CAL knockdown by siRNA results in cell surface expression of mature ΔF508 CFTR, we wanted to know whether this mature ΔF508 CFTR was functional. To determine if CAL knockdown increased chloride current compared to control samples, we utilized the short circuit current assay (Fig. 10). During the experiment, cells were in the presence of 10uM forskolin, which is used to raise levels of cAMP. During data collection, 30uM of genistein was applied, which has been shown to increase ΔF508 CFTR activity via potentiation. When genistein is applied, a peak in chloride current can be observed (Fig. 10A). Finally, 10uM of CFTR specific inhibitor 172 is applied at the end of the experiment as a control, to observe a decrease in chloride current for all samples (Fig. 10A). There is no change in transepithelial resistance (TER) associated with CAL silencing (Fig. 10B). Our results show that samples transfected with CAL specific siRNA have a statistically significant higher chloride current amplitude compared to the control samples (Fig. 10C). The data confirms that CAL knockdown leads to the production of mature ΔF508 CFTR that is functional at the cell surface.

Our results suggest a pathway in which CAL can arrest plasma membrane trafficking of ΔF508 CFTR in the ER through binding. When the interaction between these two proteins is inhibited, cell surface expression of functional ΔF508 CFTR occurs through a Golgi-associated trafficking pathway. We have demonstrated that this pathway is regulated by chaperone and scaffold proteins that serve to either stabilize the arrested ER pool of ΔF508 CFTR or improve its cell surface trafficking. Finally, we have established that CAL, a traditionally trans-Golgi associated protein, localizes to the ER, binds to ΔF508 CFTR, and can be degraded in the proteasome which demonstrates that establishing it as a regulator of early ΔF508 CFTR trafficking and degradation is a reasonable conclusion.

It was previously demonstrated that CAL inhibition extends the half-life of ΔF508 CFTR and increases ΔF508 CFTR mediated chloride currents in polarized monolayers [15]. In this present study, we observed that inhibition of CAL by siRNA leads to an increase in the C band and B band of ΔF508 CFTR in the total lysate of ΔF508 CFBE cells. When we assessed cell surface expression of ΔF508 CFTR, we saw an increase in the mature cell surface C band of ΔF508 CFTR. Additionally, we showed that ΔF508 CFTR colocalizes with plasma membrane marker cadherin when CAL is inhibited in ΔF508 CFBE cells. This suggests that inhibition of CAL releases ΔF508 CFTR from being bound to CAL at the level of the ER and permits trafficking to the cell surface.

We postulate that this mature, C band of ΔF508 CFTR generated through CAL inhibition traffics to the plasma membrane via a pathway including the Golgi. This is supported by the fact that ΔF508 CFTR colocalization with the Golgi increases, and that ΔF508 CFTR has acquired complex Golgi-associated glycosylation. Additionally, we have also shown that the mature cell surface ΔF508 CFTR produced by CAL inhibition is functional. Results from short circuit current studies demonstrate that CAL inhibition increases ΔF508 CFTR mediated chloride currents to twice that of controls. We posit, due to these results, that CAL may be a useful clinical target among patients receiving combination therapy. These results are significant because currently there are no highly effective FDA-approved therapies for treating patients with the ΔF508 mutation. While combination therapy using Lumacaftor and Ivacaftor is FDA-approved and in use, the drug combination provides a modest level of clinical benefit to patients [20].

Several key chaperone proteins regulate the molecular mechanism of CAL inhibition promoting mature functional cell surface ΔF508 CFTR. Namely, VCP, HSP70, HSP40, Aha1, HSP90, and NHERF1. VCP is a Type II AAA ATPase that is a component of ER retrotranslocation machinery and is tied to the proteasomal degradation pathway [21, 22]. It was identified as a binding partner for ΔF508 CFTR in a proteomics screen [21]. It has also been shown that VCP and gp78, an autocrine motility factor, form complexes with CFTR targeted for proteasomal degradation [23]. Our study finds that VCP binding to ΔF508 CFTR in ΔF508-CFBE cells decreases when CAL is inhibited. Thus, our results reveal that inhibition of CAL leads to cell surface expression of mature functional ΔF508 CFTR most likely through suppression of the proteasomal degradation pathway. Less ΔF508 CFTR is bound to VCP, and thus even more is available to leave the ER, and traffic to the surface via the Golgi instead.

The stress-inducible HSP70 (HSPA1A) has an ATP-dependent mechanism and an important role in CFTR folding. DNAJ (HSP40) proteins are required for HSP70 to function in promoting or suppressing protein folding pathways. HSP70 links to HSP90 via HOP to complete chaperone assisted folding, but HSP70 also interacts with E3 ubiquitin ligase CHIP to drive premature protein degradation [24]. Additionally, it has been shown that overexpression of HSP70 and co-chaperone DNAJB1 induced only very modest improvements in trafficking and stabilization of ΔF508 CFTR [25, 26]. Our results show that binding of HSP70 and ΔF508 CFTR decreases to a small extent when CAL is inhibited, perhaps inhibiting the HSP70/ CHIP anti-folding interaction in order to promote cell surface expression of ΔF508 CFTR. Nevertheless, the role of HSP70 in this trafficking pathway requires greater investigation.

Molecular chaperones HSP40, HSP90, and Aha1 assist in the productive folding of CFTR. Prior studies established that overexpression of HSP40 (Hdj-1) decreases the degradation rate of immature core-glycosylated WT CFTR [26]. However, the same study noted that ΔF508 CFTR degradation rates seemed unaffected by HSP40 overexpression [26]. This is consistent with our observation that the levels of CAL are not tightly linked to those of HSP40 in stabilizing ΔF508-CFTR.

Cell surface rescue of ΔF508 CFTR can be initiated by partial knockdown of HSP90 co-chaperone Aha1 [27]. Aha1 is a potent stimulator of HSP90 ATPase activity and acts as a late cofactor of the HSP90 complex [28, 29]. WT CFTR that is destined for the cell surface is exchanged from HSP40/HSP70 complexes to HSP90 cargo-specific complexes [29]. Reduced levels of Aha1 binding may very well improve the trafficking and maturation of ΔF508 CFTR by enhancing CFTR-Hsp90 interactions [27, 29] however, the binding of Aha1 to ΔF508-CFTR did not change significantly when CAL is silenced. In sharp contrast, the binding Hsp90 is indeed increased subsequent to CAL silencing. HSP90 is viewed as a pro-folding decision point in the CFTR trafficking pathway. Thus, the improved CFTR-Hsp90 interactions would enhance the pro-folding pathway, improving ER to Golgi transport and diminishing ER arrest of ΔF508 CFTR. Our conclusion is further supported by the decrease we observed in the binding of ΔF508 CFTR to VCP. Taken together our results reveal that CAL knockdown leads to suppression of the anti-folding pathway for ΔF508 CFTR.

Results for NHERF1 further tie together the results of this study. NHERF1, a PDZ domain containing protein, anchors CFTR to the cytoskeleton through a protein complex involving ezrin [19]. Interaction between NHERF1 and ezrin is promoted by Rac1 activation, triggering a conformational change in NHERF1 that allows it to bind and stabilize ΔF508 CFTR [19]. Our data show that when CAL is inhibited by siRNA, there is an increase in NHERF1 in total lysate. These results suggest that steady state levels of NHERF1 increase when CAL is inhibited which then can anchor ΔF508 CFTR to the plasma membrane. We speculate that the reciprocal regulation of CAL and NHERF1 protein levels may reflect a cellular process which would ultimately determine the magnitude of CAL-dependent vs NHERF1-dependent trafficking to and residence at the plasma membrane. Overall, these results indicate that scaffold and chaperone protein levels will adjust intracellularly to support protein trafficking pathways with varying outcomes.

CAL interacts with other PDZ domain proteins such as the somatostatin subtype 5 and β1-adrenergic receptors [30]. In these PDZ domain containing proteins CAL functions in Golgi related processes. Although not studied, it is likely that given the results here, that CAL also plays a role in ER trafficking of these proteins. We speculate that CAL is located in the cytoplasm when it binds to the PDZ domain of ΔF508-CFTR and other relevant PDZ domain containing proteins. Such binding might play a role in maintaining ER morphology. Thus, alterations in CAL expression could affect the movement of ΔF508-CFTR and other relevant PDZ containing proteins out of the ER by altering its morphology. Alternately, CAL through its coiled-coil domain, might alter the trafficking of relevant PDZ containing proteins by direct interaction with another coiled-coil domain containing protein which is tethered to the ER via a transmembrane domain.

Our results provide a useful method for attaining cell surface expression of mature functional ΔF508 CFTR, by inhibiting CAL and thus blocking the interaction between CAL and CFTR. Discovering new ways to establish and enhance ΔF508 CFTR at the plasma membrane will provide more options for novel therapies.

Aha1/AHSA1 (Activator of 90 kDa heat shock protein ATPase homolog 1); CAL (CFTR-Associated Ligand); NHERF1 (Na⁺/H⁺ Exchanger Regulatory Factor); PDZ (PSD-95, Dig, and ZO1); VCP (Valosin-containing protein).

We, the authors, thank Dr. Carolyn Machamer for the KDEL-td-tomato plasmid. We acknowledge Dr. Rajini Rao, Dr. Mark Donowitz, Dr. Carolyn Machamer, and Dr. J. Marie Hardwick for helpful discussions. We also thank Dr. Julia Romano for guidance regarding immunofluorescence and microscopy. This work was supported by NIH NHLBI grant #1F31HL128116 – 01 to EASB. The research was funding by the Cystic Fibrosis Foundation and an NHLBI multi-PI grant, R01 HL122267, to LC and WBG. We thank Kathryn A. Carson ScM Senior Research Associate for her support for statistical consulting. The latter was made possible by the Johns Hopkins Institute for Clinical and Translational Research which is funded in part by Grant Number 1UL1TR001079 from the National Center for Advancing Translational Sciences of the National Institutes of Health.

The authors declare that WBG has a consultant agreement with the Vertex Corporation. LC has a license agreement with the Vertex Corporation for mutant cell lines. LC has a contract with Acetylon Pharmaceuticals. None of the authors have any financial interests that pertain directly to this work.

W.B. Guggino and L. Cebotaru contributed equally to this work.