Glycerol reverses the misfolding phenotype of the most common cystic fibrosis mutation.

The common ΔF508 mutation in the cystic fibrosis transmembrane conductance regulator (CFTR) interferes with the biosynthetic folding of nascent CFTR polypeptides, leading to their retention and rapid degradation in an intracellular compartment proximal to the Golgi apparatus. Neither the pathway by which wild-type CFTR folds nor the mechanism by which the Phe deletion interferes with this process is well understood. We have investigated the effect of glycerol, a polyhydric alcohol known to stabilize protein conformation, on the folding of CFTR and ΔF508 in vivo. Incubation of transient and stable ΔF508 tranfectants with 10% glycerol induced a significant accumulation of ΔF508 protein bearing complex N-linked oligosaccharides, indicative of their transit to a compartment distal to the endoplasmic reticulum (ER). This accumulation was accompanied by an increase in mean whole cell cAMP activated chloride conductance, suggesting that the glycerol-rescued ΔF508 polypeptides form functional plasma membrane CFTR channels. These effects were dose- and time-dependent and fully reversible. Glycerol treatment also stabilized immature (core-glycosylated) ΔF508 and CFTR molecules that are normally degraded rapidly. These effects of glycerol were not due to a general disruption of ER quality control processes but appeared to correlate with the degree of temperature sensitivity of specific CFTR mutations. These data suggest a model in which glycerol serves to stabilize an otherwise unstable intermediate in CFTR biosynthesis, maintaining it in a conformation that is competent for folding and subsequent release from the ER quality control apparatus.

The common ⌬F508 mutation in the cystic fibrosis transmembrane conductance regulator (CFTR) interferes with the biosynthetic folding of nascent CFTR polypeptides, leading to their retention and rapid degradation in an intracellular compartment proximal to the Golgi apparatus. Neither the pathway by which wild-type CFTR folds nor the mechanism by which the Phe 508 deletion interferes with this process is well understood. We have investigated the effect of glycerol, a polyhydric alcohol known to stabilize protein conformation, on the folding of CFTR and ⌬F508 in vivo. Incubation of transient and stable ⌬F508 tranfectants with 10% glycerol induced a significant accumulation of ⌬F508 protein bearing complex N-linked oligosaccharides, indicative of their transit to a compartment distal to the endoplasmic reticulum (ER). This accumulation was accompanied by an increase in mean whole cell cAMP activated chloride conductance, suggesting that the glycerol-rescued ⌬F508 polypeptides form functional plasma membrane CFTR channels. These effects were dose-and time-dependent and fully reversible. Glycerol treatment also stabilized immature (core-glycosylated) ⌬F508 and CFTR molecules that are normally degraded rapidly. These effects of glycerol were not due to a general disruption of ER quality control processes but appeared to correlate with the degree of temperature sensitivity of specific CFTR mutations. These data suggest a model in which glycerol serves to stabilize an otherwise unstable intermediate in CFTR biosynthesis, maintaining it in a conformation that is competent for folding and subsequent release from the ER quality control apparatus.
Cystic fibrosis (CF), 1 a lethal hereditary exocrinopathy af-fecting approximately one in two thousand live births among populations of Caucasian or northern European descent, is caused by the functional absence of a plasma membrane chloride channel, designated cystic fibrosis transmembrane conductance regulator (CFTR) (1). The vast majority of severe CF cases in these populations is linked to a single genetic lesion, deletion of a phenylalanine codon (⌬F508) (2,3), which interferes with the folding of newly synthesized CFTR polypeptides. Nascent ⌬F508 molecules fail to traffic to the plasma membrane (4) but rather are retained by the ER quality control mechanism that prevents unfolded or misfolded proteins and unassociated subunits from exiting the ER. Instead, these retained immature ⌬F508 molecules are rapidly degraded (5,6) in a pre-Golgi compartment by a process that appears to require covalent modification by ubiquitin (7). Moreover, plasma membrane CFTR-like Cl Ϫ channel activity can be detected when ⌬F508 is overexpressed (8) or synthesized at reduced temperature (9), suggesting that Phe 508 does not play an essential role in CFTR function and raising the possibility of therapeutic intervention in CF by increasing the efficiency of ⌬F508 folding.
Glycerol and other polyols are known to stabilize protein conformation (10), increase the rate of in vitro protein refolding (11), and increase the kinetics of oligomeric assembly (12). We report here that treatment of ⌬F508-expressing cells with glycerol dramatically stabilizes newly synthesized ⌬F508 polypeptides and leads to the accumulation, in the plasma membrane, of stable, functional CFTR Cl Ϫ channels.

EXPERIMENTAL PROCEDURES
Cell Culture and Transient Transfections-Cells were cultured and transfected exactly as described previously (5). Mutant cDNA constructs were engineered as described in (13).
Pulse-Chase Experiments and Immunoprecipitation and Immunoblotting-Pulse-chase and immunoblotting experiments were performed as described previously (5), with the following modifications in the presence of glycerol. Transfected HEK cells were first treated with 5% glycerol in methionine and cysteine-free DMEM supplemented with 5% dialyzed FCS for 30 min. HEK cells, which were floating after this treatment, were spun down and suspended in 10% glycerol in methionine and cysteine-free DMEM-FCS. Cells were pulse-labeled with 35 Sprotein labeling mix (1175 Ci/mmol, DuPont NEN) at a concentration of 0.5 mCi/ml for 15 min and chased with DMEM-FCS with 10 mM methionine and 4 mM cysteine.
Kinetics of Glycerol Uptake-Untransfected HEK cells were treated with 7.5% glycerol in DMEM-FCS for 30 min, and cells were spun down and resuspended in 3 ml of DMEM-FCS supplemented with 7.5% glycerol. Glycerol uptake was measured by incubating untransfected HEK cells at 37°C with [ 3 H]glycerol (0.1 mCi/ml, DuPont NEN) for various times. Cell-associated [ 3 H]glycerol was separated from free [ 3 H]glycerol by rapid centrifugation (25 s at 16,000 ϫ g) through an 80-l cushion of silicon oil (Versalube, General Electric). After removing the supernatant, the bottom of the tube was cut, and the cell pellets were extracted in 1% Triton X-100 for determination of cell-associated radioactivity.
Electrophysiological Recording-After treatment with 10% glycerol for 24 h, glycerol was removed from HEK cells expressing CFTR or ⌬F508 by diluting the glycerol-containing medium slowly with fresh medium over the course of 1 h. After removal of the glycerol, cells were allowed to recover for at least 1 h prior to analysis by whole cell patch clamp recording. The electrodes were made from Dagan LA-16 glass, Sylgard TM coated and fire polished to give a final resistance of 2-5 megaohms when filled with pipette solution. The bath solution contained in mM: 150 NaCl, 5 KCl, 2.5 CaCl 2 , 2.5 MgCl 2 , and 10 HEPES * This work was supported by National Institutes of Health Grant DK43994. This work was done during the tenure of an established investigatorship of the American Heart Association (to R. R. K.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ Supported by the Cystic Fibrosis Foundation and a Human Frontiers Science Program long term fellowship.
ʈ To whom correspondence should be addressed. Tel.: 415-723-7581; Fax: 415-723-8475; E-mail: kopito@leland.stanford.edu. 1 The abbreviations used are: CF, cystic fibrosis; CFTR, cystic fibrosis buffered to 7.3. The osmolarity was adjusted to 300 mosm/liter. The pipette solution contained in mM: 125 CsCl, 2.5 MgCl 2 , 10 EDTA, 3.5 Mg-ATP, 0.5 cAMP, and 10 HEPES buffered to 7.3. The osmolarity was adjusted to 250 mosm/liter to avoid inducing swelling currents. In some experiments 80 units/liter protein kinase A were added to the pipette solution. The experiments were conducted at 22°C. Whole cell currents were measured using the Axopatch TM 1-C, and the data were digitized, stored to disc, and analyzed using the program pClamp TM . The current was filtered at 500 Hz and digitized at 2 KHz.

RESULTS AND DISCUSSION
The effect of glycerol treatment on steady-state expression of ⌬F508 was initially evaluated by immunoblot analysis of detergent extracts of HEK 293 cells transfected with ⌬F508 cDNA (Fig. 1). In untreated cells, only the immature (core glycosylated, 140 kDa) form was detected in cells incubated at 37°C (Fig. 1A), as previously observed (5). However, a diffuse immunoreactive band, corresponding to the mobility of mature (complex glycosylated, 165 kDa) CFTR, was apparent in extracts of cells treated with glycerol for 24 h. The steady-state levels of mature ⌬F508 induced by glycerol or by incubation at reduced temperature in these cells were similar, but the effects appear to be additive. The small difference in mobility between mature ⌬F508 rescued by reduced temperature and that rescued by glycerol is probably due to differences in terminal glycosylation (15). The effect of glycerol on the maturation of ⌬F508 was also observed in a stable line of C127 mammary carcinoma cells expressing ⌬F508 cDNA (16) (Fig. 1B), indicating that this phenomenon is not unique to transiently transfected HEK cells. Glycerol significantly increased expression of mature ⌬F508 in these cells above basal levels and above the levels induced by incubation at 26°C. The effect on steady-state expression of mature ⌬F508 in C127 cells was maximal at 10% glycerol; concentrations above or below this level did not support ⌬F508 maturation. By contrast, similar levels of ⌬F508 maturation were observed in HEK cells between 8 and 15% (data not shown). Glycerol did not appear to be acutely toxic to either cell type. Cell viability (determined by trypan blue exclusion) after 24-h exposure to 10% glycerol was between 75% in HEK cells and 90% in C127 cells. Mature ⌬F508 accumulated to clearly detectable levels in C127 cells 6 h following the addition of glycerol and continued to accumulate up to 48 h (Fig. 1C). This effect was reversible; the level of mature ⌬F508 in glycerol-treated C127 cells decreased with time following glycerol removal at a rate consistent with the half-life of mature ⌬F508 protein estimated from pulse-chase experiments (see below). Taken together, these data demonstrate that exposure of ⌬F508-expressing cells to 10% glycerol partially rescues the "misprocessing" phenotype of the mutant protein.
The effect of glycerol on ⌬F508 processing could not be replicated by incubating the cells with similar concentrations of other structurally related polyols (1,2-propanediol or 1,3-propanediol), perhaps because of lower permeability of cell membranes to these agents. Similarly, dimethyl sulfoxide (2%) did not support ⌬F508 maturation; higher concentrations were not tested because of its toxicity. Glycerol, a small (M r ϭ 98) polyhydric alcohol, is highly permeant across the plasma membrane of animal cells (17,18). In HEK cells incubated at 7.5% glycerol, [ 3 H]glycerol equilibrated rapidly (t 1 ⁄2, ϳ5 min) across the plasma membrane (data not shown). As glycerol is uncharged its distribution across the plasma membrane is independent of membrane potential. Thus, it is likely that intracellular and extracellular glycerol concentrations rapidly equilibrate in HEK cells and that glycerol's effect on ⌬F508 processing is due to its high intracellular concentration.
At least two mechanisms could account for the effect of glycerol on the accumulation of mature ⌬F508. One possibility is that glycerol stabilizes a ⌬F508 folding intermediate, which is normally rapidly diverted to the degradation apparatus. The stabilized folding intermediate would remain competent to fold into a conformation resistant to proteolysis and permissive for maturation beyond the ER. A second possibility is that glycerol acts by increasing the stability of a mature but unstable ⌬F508 polypeptide that has escaped ER retention. To discriminate between these models, the kinetics of ⌬F508 maturation and degradation were evaluated by pulse-chase labeling and immunoprecipitation in ⌬F508-transfected HEK cells (Fig. 2). In control cells not treated with glycerol, label in the band corresponding to immature ⌬F508 decayed rapidly and was nearly undetectable after 6 h of chase ( Fig. 2A). The t 1 ⁄2 of this decay was estimated to be 45 min (Fig. 2, A and D), similar to previously reported values (5). No label was detected at the mobility corresponding to the mature protein. By contrast, in the presence of glycerol, the kinetics of immature CFTR degradation were significantly slowed (t 1 ⁄2 ϭ 87 min); some of this label was clearly chased into mature ⌬F508 (Fig. 2, A and D). The fractional conversion of immature ⌬F508 in glyceroltreated (Fig. 2E) cells ranged between 3 and 8% in separate experiments, which is considerable when compared with the ϳ20 -25% efficiency of wild-type CFTR processing. The kinetics of wild-type CFTR degradation were also slowed by glycerol, suggesting that the effect of glycerol is not unique to the folding of ⌬F508 molecules (Fig. 2, C and D). Glycerol had no measurable effect on the stability of mature ⌬F508 or CFTR (Fig. 2, B  and E). These data suggest that accumulation of mature ⌬F508 in glycerol-treated cells is not the result of an effect on the mature protein and that glycerol stabilizes the immature form of ⌬F508. However, inhibition of ⌬F508 degradation, either Glycerol and CFTR Folding 636 directly with protease inhibitors or indirectly by blocking ubiquitination, also stabilizes the immature form of the protein but, unlike glycerol treatment, does not result in any accumulation of mature forms (7). This disparity in the fate of stabilized immature ⌬F508 molecules suggests that glycerol maintains immature ⌬F508 in a maturation-competent state, either by inhibiting reactions that are off pathway or by enhancing reactions that are on the folding pathway.
To rule out the possibility that the effects of glycerol on ⌬F508 processing and degradation are due to a general disruption of the ER quality control machinery, we examined the effect of glycerol on the maturation and degradation of other CFTR mutants that, like ⌬F508, are unable to escape the ER (Fig. 3). Cells expressing the missense mutants D572A and S1251A and a mutant harboring a deletion of exon 13 (⌬EX13) were pulse-labeled with [ 35 S]Met and chased for 5 h in the presence or absence of glycerol (Fig. 3A). These mutants were synthesized as immature polypeptides that were degraded and, unlike ⌬F508, failed to mature even in the presence of glycerol. Thus, some mutations in all three major cytoplasmic domains of CFTR, including the first and second nucleotide binding domains (D572A and S1251A, respectively) as well as the "R" domain, can lead to a glycerol-insensitive ER retention phenotype, suggesting that glycerol treatment does not induce a general suppression of ER retention mechanisms. The efficiency of processing and the ability to be rescued by glycerol are also highly dependent upon the nature of the substituted amino acid in CFTR Lys 464 missense mutants (Fig. 3B). Processing of mutants K464R and K464A was inefficient by comparison with wild type and was enhanced by incubation in the presence of 10% glycerol, even after accounting for the unequal label present in the immature precursor in the presence of glycerol (Fig.  3B). By contrast, no maturation was detectable for the mutant K464W in the presence or absence of glycerol. These data support the argument that glycerol rescue of CFTR maturation is not the result of a general suppression of ER quality control and suggest a correlation between the "leakiness" of the mutation and its ability to be remediated by glycerol.
Although these data establish that glycerol treatment facilitates the maturation of ⌬F508 molecules to a post-ER compartment, they do not establish that the "rescued" ⌬F508 molecules actually move to and are functional at the plasma membrane. To test the functional surface expression of glycerol-rescued ⌬F508 molecules, whole cell CFTR currents were Glycerol and CFTR Folding 637 examined in HEK cells expressing CFTR or ⌬F508 by the patch-clamp technique (Fig. 4). Large (56 Ϯ 4 pA/picofarad) rapidly activated, cAMP-dependent Cl Ϫ selective currents were readily observed in cells expressing CFTR but not in control (not glycerol-treated) cells expressing ⌬F508 (1.54 Ϯ 0.13 pA/ picofarad). By contrast, significant (13.31 Ϯ 2.3 pA/picofarad; p Ͻ 0.002 compared with untreated; Student's t test) Cl Ϫ currents, although slower activating than wild type, were observed in glycerol-treated cells expressing ⌬F508 cDNA. These currents were not observed in the absence of cAMP, as expected of CFTR. The difference in maximal current level and activation kinetics between CFTR and glycerol-treated ⌬F508 expressing cells is likely due both to the lower steady-state expression of mature ⌬F508 and to the lower open probability of the mutant channels (19). Collectively, these data suggest a model in which glycerol, a short chain polyhydric alcohol, serves to stabilize an otherwise unstable intermediate in CFTR biosynthesis, maintaining it in a conformation that is competent for folding and its subsequent release from the quality control apparatus. As degradation of immature CFTR and ⌬F508 is initiated without an apparent lag following translation (5), we propose that glycerol serves to stabilize nascent ⌬F508 chains soon after or during translation. In this role, glycerol would function as a chemical chaperone, much as HSP70 serves as a molecular chaperone. Our data suggest that these effects are not due to a generalized breakdown of ER quality control nor to stabilization of cell surface mature ⌬F508 molecules that have escaped quality control surveillance. We hypothesize that glycerol stabilizes an early intermediate in CFTR folding that lies at a branch point between productive folding (on pathway) and competing nonproductive (off pathway) steps. In this respect the effect of glycerol on ⌬F508 is similar to "osmotic remedial" mutants previously observed in yeast (20) and Escherichia coli (21). These mutations are temperature-sensitive and can be reversed by increasing the osmotic potential of the incubation medium causing the microorganisms to synthesize and accumulate high concentrations of intracellular osmolytes such as glycerol (22). Interestingly, we observe a strong correlation between the temperature sensitivity of CFTR mutations like ⌬F508, K464R, and K464A (data not shown) and their ability to be remediated by glycerol. Glycerol may provide a useful tool to manipulate the temperature-sensitive phenotypes of CFTR and perhaps other genes at non-permissive temperatures.
Our data establish the precedent that both the intracellular processing and the membrane Cl Ϫ transport phenotypes of the ⌬F508 mutation can be remediated by chemical means. These data should stimulate a search for other small membranepermeant molecules, which may be more effective or more easily delivered than glycerol at enhancing ⌬F508 processing. Finally, these data may have implications for the study or treatment of other diseases, including Alzheimer's, retinitis pigmentosa, and proteinase inhibitor deficiency that are associated with protein misfolding. FIG. 4. Glycerol treatment induces the expression of plasma membrane ⌬F508 CFTR Cl ؊ channels in the HEK cells. A, the mean whole cell maximum current density measured at Ϫ50 mV. The currents were divided by the cell capacitance in order to compare results from cell to cell. The currents in ⌬F508 cells not exposed to cAMP (n ϭ 8, column 1) or not glycerol-treated (n ϭ 6, column 2) had maximum current at t ϭ 0 after going whole cell. The currents in glycerol-treated ⌬F508 cells exposed to protein kinase A and/or cAMP (n ϭ 9, column 3) increased slowly and peaked at ϳ 3 min. CFTRexpressing cells responded rapidly to intracellular cAMP (n ϭ 3, column 4), peaking at less than 1 min. pF, picofarad. B, typical whole cell currents seen in 3 cells exposed to intracellular cAMP. The currents on the left are the currents seen upon break in, and the currents on the right are currents near the peak of the response. The currents shown are in response to voltage steps from ϩ100 mV to Ϫ75 mV in steps of Ϫ25 mV from a holding potential of Ϫ50 mV. The currents showed linear current-voltage (IV) behavior and no time dependence.