Correction of both NBD1 energetics and domain interface is required to restore ΔF508 CFTR folding and function

Thermodynamic destabilization of isolated NBD1 variants by the ΔF508 mutation

To determine the ΔF508-induced NBD1 energetic defect, full-length wt and ΔF508 human NBD1 variants (amino acid 389–678) with or without solubilizing (S) and/or revertant (R) mutations were purified from E. coli as described (Lewis et al., 2005). Both the R mutations (G550E, R553Q and R555K) and S mutations (F409L, F429S, F433L, F494N and H667R) could partially rescue the ΔF508 CFTR folding and functional defect (Lewis et al., 2005; Pissarra et al., 2008; Teem et al., 1993; Teem et al., 1996) and were assumed to stabilize the domain either alone or in combinations (1S, 3S, R, R1S and R4S, see Fig. 1b). The isolated NBD1s were >90-95% pure and monomeric (Fig.S1a-b, data not shown). The secondary structure composition and TNP-ATP-binding affinity of wt- and ΔF508-NBD1 variants were comparable (Fig.S1c-d and data not shown) (Qu et al., 1997; Qu and Thomas, 1996; Stratford et al., 2007).

The ΔF508 mutation reduced the apparent melting-temperature (T_m) of the wt NBD1-1S from 41.8 ± 0.2 to 33.2 ± 0.2 °C (ΔT_m ≈ 8.6 ± 0.2 °C), based on ellipticity, as well as Trp fluorescence and differential scanning fluorimetry (DSF) using different reporter dyes (Fig.1c-e and Fig.S1e-f) (Niesen et al., 2007; Senisterra et al., 2008). In the absence of second site mutations, the wt and ΔF508 NBD1-0S thermal denaturation propensity was slightly increased (T_m ~39.8 ± 0.2 and ~31.7 ± 0.1°C, respectively) relative to NBD1-1S measured at 2.5 mM ATP (Fig.1d-e). Since the protein yield was limited (Fig.S1g), only DSF scans could be performed on NBD1-0S.

The T_m difference between the wt and ΔF508 NBD1 was similar (6-8°C) for 0S, 1S, 3S, R, R4S and in ΔRI (Protasevich et al., 2010), as well as at reduced ATP concentration (Fig.1e and Fig.S1h-i). Thermal unfolding preceded domain aggregation for both wt and ΔF508-NBD1 (Supplemental Methods and Fig.S1j). Assuming a reversible, two-state folding mechanism with slow aggregation of the unfolded form (Protasevich et al.), we estimated the folding free energy (ΔG⁰) based on the DSF data. Decreasing the temperature from 37°C to 20°C lowered the ΔG⁰ of the ΔF508 and wt NBD1-0S from +1.8 to -4.2 kcal/mol and from -1.0 to -4.8 kcal/mol, respectively (Supplemental Results and Table).

The ΔF508 decreased the midpoint urea denaturation concentration (D_0.5) of the wt NBD1 variants by ~0.8-1 M at 1 mM ATP (Fig.S2e). NBD1 fractional unfolding was calculated by extrapolation of the CD data due to the oligomerization/aggregation propensity of partially unfolded ΔF508-NBD1 (Fig.1f, Fig.S2a-d and Supplemental Results) (Strickland et al., 1997; Wang et al., 2010). The estimated ΔG⁰ between the wt- and ΔF508-NBD1-1S and -3S was decreased by ~2.4 and ~1.4 kcal/mol, respectively, while the R1S and R4S mutation stabilized the ΔF508 by ~1 and ~1.6 kcal/mol at 20°C (Fig.1g). Thermal unfolding analysis yielded comparable ΔG⁰ differences between wt and ΔF508-NBD1-0S, -1S and -3S (approximately -1.7, -1.6 and -2.3 kcal/mol, respectively, Supplemental Results and Table). Thus thermal and chemical denaturation studies demonstrated the thermodynamic destabilization of the ΔF508 NBD1 with a single or no solubilizing mutation at 37°C (Fig. 2a).

Kinetic destabilization of NBD1 by the ΔF508 mutation

To determine the unfolding energy barrier between the native and unfolded states, the rate of NBD1 unfolding was monitored as a function of urea concentration by CD spectroscopy (Fig.2b and Fig. S2a-b). The NBD1-1S initial unfolding kinetics exhibited mono-exponential behavior and was found to be independent of protein concentration between 3-14 μM (data not shown). The ΔF508-NBD1-1S extrapolated unfolding rate in water (k_u^H₂0) was >30 fold (0.73 sec^-1) and ~17-fold (0.04 sec^-1) faster than its wt counterpart at 20°C and 37°C, respectively (Fig. 2c-d). The k_u^H₂0 at 37°C was determined by the extrapolation of k_u^H₂0 obtained at 16-30°C (Fig.S2f). The unfolding activation energy (ΔG_u^‡) of the wt NBD1-1S was reduced in the ΔF508 from ~5.0 ± 0.11 to ~2.9 ± 0.04 kcal/mol (±SEM, n=6) at 20°C and from ~1.8 to ~0.2 kcal/mol at 37°C (Fig.S2g and Supplemental Table 1). In contrast, the R1S and R4S mutations partially restored kinetic stability of the ΔF508 NBD1-1S by reducing the k_u^H₂0 ~3-4 fold at 20°C (Fig.2c). Jointly, these results showed for the first time that the ΔF508 reduces both thermodynamic and kinetic components of the ΔF508-NBD1-1S energetic defect at 37°C, which could be reversed by second site suppressor mutations.

The increased conformational flexibility of the ΔF508-NBD1-1S was verified by a modest, but significant enhancement in deuterium uptake, determined by hydrogen/deuterium exchange (HDX) mass spectrometry (MS) after 15 min incubation at 24°C, but not at 0°C (Fig.S3). Due to thermodynamic stabilization, the extent of HDX for ΔF508-NBD1-1S was profoundly reduced by the R1S mutation, consistent with the increased backbone dynamics of the ΔF508-NBD1-3S, localized to residues 509-511 (Lewis et al., 2010).

The ΔF508 NBD1 conformational defect is recognizable by protein quality control

To ascertain that the ΔF508-NBD1 structural defect may serve as a degradation signal in vivo, the metabolic turnover of NBD1 fusion proteins was determined in prokaryotes and eukaryotes. The degradation rate of the ΔF508-NBD1-1S was ~4-fold faster than the wt-NBD1-1S in E. coli at 37°C (data not shown). To examine the effect of NBD1 conformational stability on the ER export efficiency, NBD1s were tethered to the C-terminally truncated CD4 (CD4T-NBD1) (Fig.3a). The PM density of the chimeras was monitored by cell surface ELISA as a surrogate measure of ER export efficiency at comparable translational rates in COS7 cells at 37°C (Supplementary Results, Fig.S4a-c and (Du and Lukacs, 2009)). The wt and ΔF508 CD4T-NBD1 PM expression levels were 20% and 10%, respectively, of CD4T (Fig.3c). The ΔF508 mutation decreased the PM density of CD4T-NBD1 variants containing second site mutations, while conformational stabilization by the R4S reversed the phenotype (Fig.3c). The low level of ΔF508 CD4T-NBD1-R1S expression could be attributed to the partially normalized unfolding kinetics of the domain (Fig.2c). Stabilization of wt NBD1 by second site mutation, however, elevated the chimera PM expression from 20% to 60% (Fig.3c). These results in concert with the effect of F508E, F508R, F508G, F508S, F508D and F508N mutations revealed that the CD4T-NBD1 PM density was proportional to the domain stability if the NBD1 T_m was >38°C (Fig.3d and Fig.S4d). Chimeras containing NBD1s with T_m <38°C exhibited ≤10% expression of CD4T in COS7 cells at 37°C (Fig.3d), consistent with the domain instability and the consequential Ub-dependent ER retention/degradation as shown below. These findings are reminiscent of the folding energy-dependent biosynthetic secretion of transthyretin and BSEP (Sekijima et al., 2005) and suggest that proteostatic mechanisms can recognize and eliminate structurally damaged proteins with impaired folding energetics not only in the lumen (Wiseman et al., 2007), but at the cytoplasmic surface of the ER as well.

The in vitro ubiquitination propensities of wt and ΔF508 NBD1-1S, -3S and -R1S were assessed following the native NBD1 unfolding at 25-50°C for 5 min. Subsequent ubiquitination was performed in the presence of E1 enzyme, UbcH5c (an E2 Ub-conjugating enzyme), CHIP (C-terminal Hsp70 interacting proteins, an E3 Ub-ligase) and ATP (Meacham et al., 2001) for 30 min at 26°C in vitro. Ubiquitination was monitored by the appearance of high molecular mass adducts in anti-NBD1 immunoblots (Fig.3e). Half-maximal ubiquitination was achieved at temperatures (T_0.5) corresponding to the T_m of NBD1-1S and -3S (Fig.3f). The unfolding temperature required for 50% ubiquitination was increased proportionally to the ΔF508-NBD1-R1S stability (Fig.3f). This result indicates that the NBD1 conformational defect correlates with the domain ubiquitination propensity that may signal ER degradation.

Conformational stabilization of the ΔF508-NBD1 modestly improves ΔF508-CFTR folding

As a next step, we assessed whether ΔF508-CFTR global misfolding can be solely attributed to impaired NBD1 energetics. To this end, CFTR folding behavior was determined as a function of NBD1 thermal stability (T_m) by measuring the conversion of radioactively pulse-labeled core-glycosylatyed CFTR into complex-glycosylated form in BHK cells (Fig.4a) (Du et al., 2005). The 1S, 3S, R or R1S mutations increased the wt CFTR folding efficiency from ~27% up to ~52% (Fig.4b). The cellular and PM density of wt CFTR were similarly increased, as evidenced by immunoblotting and cell surface ELISA, respectively, in BHK and COS7 cells (Fig.4c-d). In contrast, stabilization of the ΔF508-NBD1 only marginally enhanced the mutant folding efficiency (from 0.4% to 3.5%), expression and PM density (Fig.4a-d). These and additional results discussed in the following section show that both folding efficiency and PM density of wt variants increased at ~37 and ~14 fold steeper slopes, respectively, than their ΔF508 or F508X counterparts as a function of NBD1 T_m (Fig.5a-b) and suggest that NBD1 folding energetics can define the wt but not ΔF508 CFTR domain-domain assembly. A possible explanation for this phenomenon could be the permissive effect of NBD1-CL4 interface stabilization by the F508 residue for coupled domain folding, an assumption supported by the following results.

1. While most missense mutations of the F508 (Q508, S508, D508 and N508) largely preserved the NBD1 thermal stability, monomeric state, secondary structure composition and TNP-ATP-affinity, they severely compromised CFTR folding (Fig.5a, Fig.S4d and data not shown). Cys and Met substitutions of F508 retained wt CFTR folding conceivably by preserving intra- and inter-domain (CL4) side chain interactions (Du et al., 2005).

2. Although certain combinations of F508N and R or S mutations conferred thermodynamic and kinetic stability comparable with or exceeding that of the wt NBD1-1S (e.g. F508N-R; ΔG⁰ = -5.1 ± 0.1 kcal/mol, T_m ≈47°C and k_u^H₂0 = 0.16 10^-3 sec^-1, Fig.S5), they failed to restore wt-like folding and expression of the ΔF508 CFTR (Fig.5a-b). Similar to the 3S and R4S mutations (Fig.5a-b), the stabilizing ΔRI mutation (Protasevich et al., 2010) was unable to reinstate wt-like folding and expression in ΔF508 CFTR (Fig.4b and d).

3. To demonstrate that the hydrophobic cluster of the NBD1-CL4 interface (Fig.1b) stabilizes the CFTR native fold, the conformation of individual domains was probed by limited trypsinolysis in isolated microsomes (Du et al., 2005). The proteolytic fragmentation patterns were visualized by immunoblotting using domain-specific antibodies (Fig.5c). The NBD1, NBD2, MSD1 and MSD2-containing fragments, represented by 29-33, 29-31, 35-37 and 53-76 kDa immunoreactive bands, have been validated (Cui et al., 2007; Du et al., 2005) and are indicated by dashed boxes (Fig.5c).

In spite of the R1S-induced stabilization of the ΔF508-NBD1, the conformational flexibility of MSD1, MSD2 and NBD2-containing fragments could not be reversed, and it remained proteolytically more sensitive than their wt counterparts (Fig.5c). This observation implies that correction of the NBD1 energetic defect is insufficient to reinstate domain assembly. Remarkably, NBD1, MSD1 and NBD2 were susceptible to comparable misfolding upon disrupting the NBD1-CL4 interface by the R1070W substitution in wt CFTR. This reduced the channel folding efficiency by ~87% (Fig.S6a and Fig.6a-c). Conversely, stabilization of the NBD1-CL4 interface by the V510D substitution (see bellow) increased the wt CFTR folding efficiency by ~2-fold, supporting the critical role of the NBD1-CL4 interface in the coupled domain folding of CFTR (Fig.6b and see below).

Correction of both ΔF508-NBD1 energetics and NBD1-CL4 interaction is required to synergistically restore ΔF508-CFTR folding and processing

Since CFTR biogenesis appears to be associated with the formation of NBD1-CL4 interface (He et al., 2010; Serohijos et al., 2008a), we tested whether stabilization of the NBD1-CL4 interface alone or in combination with that of NBD1 were also required to increase the ΔF508 CFTR folding efficiency. R1070W or V510D substitutions at the interfaces restored the proximity of the ΔF508 NBD1 and CL4 as shown by Cys-crosslinking. The NBD1-CL4 interface stabilization can be accomplished by filling the cavity created by the ΔF508 with the bulky hydrophobic side chain of R1070W or by salt bridge formation between V510D in NBD1 and R1070 in CL4 (Fig.S6b) (He et al., 2010; Loo et al., 2008, 2010; Thibodeau et al., 2010).

R1070W similar to V510D alone modestly increased the ΔF508-CFTR folding efficiency, cellular and PM expression (Fig.6a-e) in part confirming previous reports (He et al., 2010; Thibodeau et al., 2010). In sharp contrast, combining an interface mutation with NBD1 stabilization by 3S, R or R1S enhanced the ΔF508-CFTR folding efficiency by 25-30-fold and led to 50-80% of wt CFTR expression in both COS7 and BHK cells (Fig.6a-e and Fig.S7a). At the individual domain level, the combination of R1S and R1070W, but not R1S alone, largely restored domain assembly, as indicated by the wt-like trypsin resistance of the MSD1, MSD2 and NBD2 in ΔF508-CFTR-R1S-R1070W (Fig.5c). Similarly, the ΔF508-CFTR folding and expression were synergistically rescued by combining the ΔF508-NBD1 stabilizing mutation ΔRI with R1070W or V510D (Fig.6f-h and Fig.S7b) ruling out non-specific effects of second site mutations. Direct energetic stabilization of the ΔF508-NBD1-0S and -3S by the V510D mutation was marginal (Fig.S5).

The four-domain minimal folding unit of CFTR (CFTR-1218X) that lacks the NBD2, mimics the cooperative domain folding of wt CFTR (Du and Lukacs, 2009). Accordingly, substantially increased folding efficiency of ΔF508-CFTR-1218X was only achieved by synergistic stabilization of the NBD1-CL4 interface (R1070W or V510D) and NBD1 (3S, -3R or -R1S) (Fig.7a-c and Fig.S7c).

The cooperative domain-folding model predicts that destabilizing NBD1 or the NBD1-CL4 interface would limit the CFTR folding efficiency as reported for several CF-causing mutations in the CLs (Riordan, 2008). This was indeed the case. CF-associated non-conserved amino acid substitutions (A559E or R600G) that are solvent-exposed and distant from the NBD1-CL4 interface prevented the expression of soluble NBD1 in E. coli and CFTR processing in mammalian cells (Fig.S7d and data not shown). Likewise, disrupting the NBD1-CL4 interface by the R1070W mutation compromised both wt CFTR and CFTR-1218X folding (Fig.6b-c, Fig.6e left panel, and Fig.7c), as well as the NBD1, NBD2 and MSD1 conformation, which was partly rescued by stabilizing the NBD1 with 3S, R or R1S (Fig.S6a). Together, these observations suggest that NBD1-CL4 association results in mutually stabilized NBDs and MSDs. This inference is supported by the partially regained protease resistance of NBD1, MSD1 and NBD2 (Fig.S6a) as well as restored folding and processing upon NBD1 stabilization by 3S, R, R1S or ΔRI in the CFTR-R1070W and CFTR-1218X-R1070W (Fig.6b-c, e and Fig.7c).

Channel gating in ΔF508-CFTR is corrected by interfacial and NBD1 stabilizing mutations

Synergistic correction of the ΔF508-CFTR channel gating defect could be observed by the interfacial and NBD1 stabilizing mutation, measured by the open probability of reconstituted and phosphorylated CFTR in a phospholipid planar bilayer (Fig.7d). The low-temperature rescued ΔF508-CFTR open probability (P_o= 0.11 ±0.01, n=6) was reduced to 30% of the wt at -60 mV holding potential. The 3S and R1070W mutations increased the ΔF508-CFTR P_o by 27% and 54%, respectively. Combination of the suppressor mutations resulted in an P_o (0.28 ±0.03, n=8) that was 154% higher than the control (Fig.7e), indicating that ΔF508 CFTR functional rescue requires the stabilization of both NBD1 energetics and the NBD1-CL4 interface.

ΔF508 destabilizes the NBD1 both thermodynamically and kinetically

The small difference in thermal denaturation profile of ΔF508-NBD1-0S and -1S (T_m~32 and 33°C, respectively) and folding efficiency of CD4T-NBD1-ΔF508, as well as ΔF508 CFTR containing 0S or 1S, demonstrate that the 1S mutation marginally reverts the ΔF508-NBD1 structural defect and justified the use of NBD1-1S as a surrogate domain for NBD1-0S. Thermal and chemical denaturation scans revealed that the ΔF508 mutation thermodynamically destabilizes the NBD1-0S and -1S at 37°C (Fig.1d-g) in line with recently published data using variable numbers of second site mutations (Hoelen et al., 2011; Protasevich et al., 2010; Wang et al., 2010), but at variance with earlier findings (Lewis et al., 2005; Qu and Thomas, 1996; Thibodeau et al., 2005).

The thermodynamics of the NBD1-1S folding profile is profoundly altered by the ΔF508 mutation. The attenuated (re)folding kinetics of the ΔF508 NBD1 (Qu and Thomas, 1996; Serohijos et al., 2008b; Thibodeau et al., 2005) could be partly explained by the ~4-fold increased folding activation energy (ΔG_f^‡) of the isolated ΔF508 NBD1-1S at 37°C (Fig. 2a). In addition, we discovered that ΔF508 NBD1-1S has ~10-fold reduced unfolding activation energy (ΔG_u^‡) and a ~17-fold faster unfolding rate relative to its wt counterpart at 37°C (Fig.2a, d and S2g). The combination of ΔF508 NBD1-1S rapid unfolding (T_1/2 ~1 sec), reduced ΔG⁰ and impeded folding rate constitute the complex energetic folding defect and are primarily responsible for ΔF508 CFTR (Denning et al., 1992), CD4T-ΔF508-NBD1 and SUMO-ΔF508-NBD1-1S misprocessing, as well as for the 2-10 fold enhanced NBD1 protease susceptibility in ΔF508 CFTR (Cui et al., 2007; Du et al., 2005; Roy et al., 2010; Thibodeau et al., 2010). Energetic stabilization of ΔF508-NBD1 variants at reduced temperature (Fig.1e) also offers a reasonable explanation for the native-like crystal structure of ΔF508 NBD1-2S and -3S obtained at 4-20°C (Lewis et al., 2010; Lewis et al., 2005).

We demonstrated that multiple R and S mutations restore the ΔF508 NBD1 conformational stability at 37°C (Fig.1-2), but modestly increase the folding efficiency (Fig.4b), expression and PM function of ΔF508 CFTR (Pissarra et al., 2008; Teem et al., 1993). Although some of the R and S mutations are solvent exposed or far from F508, they can stabilize NBD1 through long-range, intra-domain interactions among the three NBD1 subdomains with high contact order (Khushoo et al., 2011), accounting for the structural basis of cooperative NBD1 folding and misfolding. Similar effects were observed in the solvent exposed W62G mutation of lysozyme that leads to unfolding and fibril formation (Zhou et al., 2007).

Biophysical characterization also revealed that the wt NBD1-0S and -1S are thermally unstable at 37°C (T_m~40-42°C). This manifests in the wt SUMO-NBD1and CD4T-NBD1 recognition by the protein quality control machinery in bacteria and mammalian cells, respectively, and the domain ubiquitination in vitro. While inter-domain interactions likely stabilize NBD1 conformationally in wt CFTR, as suggested by the enhanced NBD1 protease susceptibility upon destabilizing the MSD1 or NBD2 by CF causing point mutation (Cui et al., 2007; Du and Lukacs, 2009; Du et al., 2005; Xiong et al., 1997), or the NBD1-CL4 interface by R1070W (Fig.S6a), these stabilizing factors are compromised due to coupled misfolding of MSD1, MSD2 and NBD2 in ΔF508 CFTR and other mutants (Cui et al., 2007; Du and Lukacs, 2009; Du et al., 2005; Xiong et al., 1997). Conversely, stabilizing NBD1 by second site mutations is sufficient to increase wt channel folding efficiency by ~2-fold (Fig.S7e).

Stabilization of both NBD1 energetics and domain interface is a prerequisite for ΔF508 CFTR domain assembly

At least three mechanisms may be responsible for multi-domain protein folding. a) Individual domains can fold and function independently. b) One or more domains can facilitate the folding of the other(s) mutually and/or by acting as a folding templates or c) the domains(s) can stabilize the folded conformation by interacting in their native state to prevent unfolding. The first scenario has been ruled out (Cui et al., 2007; Du and Lukacs, 2009; Han et al., 2007). This work provides support for the second and third mechanisms by correlating NBD1 energetics and interface stabilization with CFTR folding efficiency and domain conformation.

1) Here we showed that both NBD1 and the NBD1-CL4 interface instability contribute to the intrinsically low folding efficiency of the wt CFTR in vivo (Riordan, 2008). Solubilizing, revertant or ΔRI mutations increased the wt channel folding efficiency (by 20-80%) in proportional with NBD1 stabilization, a tendency that is also observed in the background of NBD1-CL4 destabilization in R1070W-CFTR (Fig.4b and 6b). Stabilization of the NBD1-CL4 interface by salt-bridge formation between V510D and R1070 also improved wt CFTR biogenesis (Fig. 6b).

Although protease susceptibility and FRET measurements suggest that NBD1 undergoes cotranslational folding (Du et al., 2005; Khushoo et al., 2011), it is likely that NBD1 and interface stabilization can facilitate CFTR folding by favorably influencing the folding thermodynamics and/or kinetics of interacting domains. This mechanism would be consistent with the lack of measurable difference in the NBD1 protease susceptibility when comparing wt and wt-3S CFTR (data not shown) and increased folding efficiency of ΔF508 CFTR upon blocking ER-associated ubiquitination in the presence of corrector molecules (Grove et al., 2009).

2) Cysteine cross-linking studies have demonstrated that NBD1 solubilizing mutations (e.g. 3S) restored the NBD1-CL4 and NBD2-CL2 (MSD1) association (He et al., 2010). Intriguingly, the proximity of selectively cross-linked interface residues is insufficient to profoundly enhance ΔF508 CFTR and CFTR-1218X folding, presumably due to transient interactions of the domains (Fig.4b, 5c and 6a-c). Accordingly, pharmacological chaperones (e.g. RDR1, a substituted phenylhydrazone) and second site suppressor mutations used here (Fig.5a) and elsewhere (He et al., 2010; Hoelen et al., 2011; Sampson et al., 2011) failed to restore ΔF508 CFTR folding, in spite of reverting ΔF508-NBD1 stability nearly to that of the wt domain. The combination of NBD1 and NBD1-CL4 interface stabilization, however, led to robust biochemical and functional rescue, as wells as the acquisition of protease resistance in MSD1, MSD2 and NBD2 (Figs.5-6 and 7e). This last observation suggests that stabilization of MSD2 and MSD1 by NBD1 interactions is indispensable for the conformational rescue of the NBD2 and ΔF508 CFTR.

Based on the three dimensional architecture and cooperative folding model of CFTR as well as the energetic and topological defect of the ΔF508 NBD1, it is reasonable to speculate that the channel structure destabilization is initiated by the compromised NBD1-CL1 interface (He et al., 2008). This view is consistent with the increased metabolic turnover of the MSD1-NBD1-R caused by ΔF508 (Rosser et al., 2008) (Fig.7f). MSD2 translation further exacerbates the tertiary structural defect due to the disruption of the NBD1-CL4 hydrophobic cluster (He et al., 2010) and consequential destabilization of the MSD2 and MSD1-MSD2 packing (Cui et al., 2007; Du and Lukacs, 2009; Wang et al., 2007). This in turn compromises the NBD2-CL2/CL3 and NBD1-NBD2 interfaces, culminating in the global misfolding of NBD2 and ΔF508 CFTR (Cui et al., 2007; Du et al., 2005) (Fig.7f). This multi-step misfolding can be modulated by cytosolic chaperones (Qu et al., 1997; Rosser et al., 2008; Wang et al., 2006) and by strategically-placed second site suppressor mutations that may also be mimicked by small molecule correctors.

Implications for correcting CF and other conformational diseases caused by multidomain membrane protein misfolding

Conformational stabilization of mutant polypeptides with energetically compromised, but partially preserved ligand-binding capacity has been achieved by pharmacological chaperones to restore their biosynthetic processing (Bernier et al., 2004). The combination of impaired energetics and domain-interactions of ΔF508 NBD1 provides a reasonable explanation for the limited rescue efficiency (<15%) of ΔF508 CFTR by individual corrector molecules isolated in HTSs of chemical libraries (Clancy et al., 2011; Van Goor et al., 2011; Wang et al., 2007). Our results suggest a rational, structure-based screening strategy to isolate combinations of corrector molecules that may preferentially stabilize the mutant NBD1 and the NBD1-CL4 interface.

At a more general level, our results highlight the potential consequence of destabilizing interface mutations in cooperatively folding multidomain membrane proteins and demonstrate that folding of neighboring domains can be compromised by interface mutations similar to a subset of multidomain soluble polypeptides (Han et al., 2007). Extension of this paradigm to multi-domain PM proteins may serve as a framework in future studies of folding diseases caused by mutations clustered at domain interfaces, as in the case of the shaker CLC and ABCC6 transporter (Feng et al., 2010; Fulop et al., 2009).

Cloning, expression and purification of CFTR and NBD1 proteins

Point mutations were introduced into wt and ΔF508 CFTR-3HA cDNA by overlapping PCR mutagenesis as described before (Du et al, 2005). CD4T-NBD1 expression constructs were generated by inserting PCR-amplified NBD1 cDNAs into the XmaI/ApaI sites of pcDNA3-CD4T (Du & Lukacs, 2009).

The pET26b-derived expression vectors with N-terminal His₆–SUMO fusion tag containing the human wt or ΔF508 NBD1-3S and -R4S with the crystallization domain boundaries were kindly provided by H. Lewis (Structural GenomiX Inc. and CFFT Inc.) and C. Lima (Columbia University) and purified as described (Lewis et al., 2005).

Circular Dichroism and Fluorescence Spectroscopy

Circular dichroism (CD) spectroscopy was performed using a Chirascan CD spectrometer (Applied Photophysics, Leatherhead, UK). CD scans were collected on proteins in 150 mM NaCl, 1 mM ATP, 3 mM MgCl₂, and 50 mM sodium phosphate buffer at pH7.5 between 260 and 190 nm at 20 or 25°C using a 0.2 mm path-length cuvette at ~14 μM NBD1 concentration. CD measurements were acquired every 0.5 nm with 0.5 s an integration time and repeated 3 times with baseline correction. Thermal unfolding scans were measured at 208 nm with a 0.5 °C step size at at 1 °C/min ramp rate with ±0.1 °C tolerance.

Tryptophan fluorescence of NBD1 (12 μM in a 1 mM ATP, 3 mM MgCl₂, 150 mM NaCl and 10 mM Hepes at pH 7.5) during thermal unfolding was monitored using a Cary Eclipse spectrofluorometer (Varian Inc., Palo Alto, CA) with a 3×3 mm micro quartz cuvette. Fluorescence emission was monitored at 340/350 nm at 290 nm excitation wavelengths at 300 nm/min scan rate. The data were fitted to a Boltzmann sigmoid function to calculate the T_m using the Excel add-on package XLfit (IDBS Ltd, Bridgewater, NJ) as described (Senisterra et al., 2008).

Differential scanning fluorimetry (DSF)

DSF of NBD1 (7-12 μM) scans were obtained in routinely in 150 mM NaCl, 20 mM MgCl₂, 10 mM HEPES and 2.5 mM ATP at pH7.5 in the presence of 2X Sypro Orange or Nano Orange, using a Stratagene Mx3005p (Agilent Technologies, La Jolla, CA) qPCR instrument (Lo et al., 2004; Niesen et al., 2007). The temperature ramp rate was 1°C/min. The data were fitted to a Boltzmann sigmoid function.

Additional methodologies are included in the Supplemental Methods.