Small heat-shock proteins select deltaF508-CFTR for endoplasmic reticulum-associated degradation

Abstract

Secreted proteins that fail to achieve their native conformations, such as cystic fibrosis transmembrane conductance regulator (CFTR) and particularly the DeltaF508-CFTR variant can be selected for endoplasmic reticulum (ER)-associated degradation (ERAD) by molecular chaperones. Because the message corresponding to HSP26, which encodes a small heat-shock protein (sHsp) in yeast was up-regulated in response to CFTR expression, we examined the impact of sHsps on ERAD. First, we observed that CFTR was completely stabilized in cells lacking two partially redundant sHsps, Hsp26p and Hsp42p. Interestingly, the ERAD of a soluble and a related integral membrane protein were unaffected in yeast deleted for the genes encoding these sHsps, and CFTR polyubiquitination was also unaltered, suggesting that Hsp26p/Hsp42p are not essential for polyubiquitination. Next, we discovered that DeltaF508-CFTR degradation was enhanced when a mammalian sHsp, alphaA-crystallin, was overexpressed in human embryonic kidney 293 cells, but wild-type CFTR biogenesis was unchanged. Because alphaA-crystallin interacted preferentially with DeltaF508-CFTR and because purified alphaA-crystallin suppressed the aggregation of the first nucleotide-binding domain of CFTR, we suggest that sHsps maintain the solubility of DeltaF508-CFTR during the ERAD of this polypeptide.

Figure 1.

HSP26, HSP82, and FES1 mRNA levels are elevated in yeast expressing CFTR. (A) The indicated mRNAs were detected by Northern blot analysis from yeast strain JN516 transformed with a plasmid expressing CFTR and yeast transformed with a vector control (–; pRS426). The -fold up-regulation upon CFTR expression as determined by Northern blot analysis is indicated, and when standardized to the levels of SEC61 message, they are HSP26, 2.0; HSP82, 1.6; and FES1, 1.5. By comparison, the average -fold up-regulation from the microarray analyses (n = 6) are HSP26, 1.4; HSP82, 1.3; and FES1, 1.4. (B) Total cell extracts were prepared from yeast either lacking or containing the CFTR expression vector, as in A, and proteins were resolved by SDS-PAGE, transferred to nitrocellulose membranes, and analyzed with the indicated antisera. The ratios of protein in each sample from two independent sets of experiments were quantified as described in Materials and Methods.

Figure 2.

Deletion of the small heat-shock protein-encoding genes slows CFTR degradation in yeast. CFTR degradation was assessed in wild-type and hsp26Δ yeast (A) and in wild type and hsp26Δhsp42Δ yeast (B). CFTR protein levels were quantified and averaged from at least three independent sets of experiments and were standardized to the levels detected directly after cycloheximide addition (0 min). Closed circles represent the CFTR protein levels in wild-type yeast, and open circles represent the CFTR protein levels in the mutant strains. Vertical bars indicate the SEs of the mean. p values in A at the 40, 60, and 90-min time points are 0.011, 0.022, and 0.062, respectively, and in part B are <0.03. CPY* (C) and Ste6p* (D) degradation rates were determined by pulse-chase immunoprecipitation experiments in HSP26HSP42 wild-type yeast and in the hsp26Δhsp42Δ double mutant strain after addition of cycloheximide. Protein levels were quantified from two independent sets of experiments and averaged. Standardization of the data and symbols are the same as described above. Vertical bars indicate the range of the data.

Figure 3.

Deletion of the small heat-shock protein-encoding genes does not impact CFTR polyubiquitination in yeast. CFTR polyubiquitination was assessed in wild-type and hsp26Δhsp42Δ yeast and in wild-type and hrd1Δdoa10Δ yeast as described in Materials and Methods.

Figure 4.

αA-crystallin overexpression decreases the steady state levels of ΔF508-CFTR. HEK293 cells were transfected with 1.5 μg of pcDNA3.1-CFTR or pcDNA3.1-ΔF508-CFTR and 0.5 μg of pcDNA3.1 or pcDNA3.1-αA-crystallin per 60-mm dish. (A) A representative immunoblot of the indicated proteins is shown after protein concentrations were normalized in all samples. Because Hsp70 levels remained constant, the amount of this chaperone served as an additional loading control. (B) CFTR protein levels were quantified from three independent sets of experiments and averaged. All values were obtained after standardization to the levels of Hsp70. The black bar represents the -fold decrease of wild-type CFTR total levels upon αA-crystallin overexpression, and the gray bar depicts the -fold decrease of ΔF508-CFTR levels upon αA-crystallin overexpression. Vertical bars indicate the SEs of the means (p = 0.013; t test).

Figure 5.

αA-crystallin overexpression has no effect on the biogenesis of wild-type CFTR. HEK293 cells were transfected with 1.5 μg of pcDNA3.1-CFTR and 0.5 μg of pcDNA3.1 or pcDNA3.1-αA-crystallin per 60-mm dish. Rates of CFTR maturation (C-band) and CFTR B-band degradation as well as αA-crystallin expression levels were determined by pulse-chase immunoprecipitation in cells transfected with CFTR and containing a vector control or the αA-crystallin expressing vector.

Figure 6.

αA-crystallin overexpression accelerates the degradation of ΔF508-CFTR. (A) HEK293 cells were transfected with 1.5 μg of pcDNA3.1-ΔF508-CFTR and 0.5 μg of pcDNA3.1 or pcDNA3.1-αA-crystallin per 60-mm dish. Rates of ΔF508-CFTR degradation as well as αA-crystallin expression levels were determined by pulse-chase immunoprecipitation. (B) ΔF508-CFTR degradation was determined from four independent sets of experiments and averaged. All values were obtained after standardization to the levels detected at the beginning of the chase period (0 h). Closed circles represent ΔF508-CFTR protein levels in HEK293 cells with the vector control, and open circles represent ΔF508-CFTR protein levels in HEK293 cells overexpressing αA-crystallin. Vertical bars indicate the SEs of the mean. *p = 0.076, **p = 0.0001, ***p = 0.003, and ****p = 0.5.

Figure 7.

αA-crystallin coprecipitates preferentially with ΔF508-CFTR. (A) HEK293 cells were transfected with 6 μg of pcDNA3.1-CFTR or pcDNA3.1-ΔF508-CFTR and 2 μg of pcDNA3.1 or pcDNA3. 1-αA-crystallin per 100-mm dish. The interaction between CFTR or ΔF508-CFTR and αA-crystallin was assessed by immunoblot analysis for αA-crystallin after CFTR immunoprecipitation. The ER luminal chaperone BiP served as a negative control. (B) Relative binding efficiency was determined from 3 independent sets of experiments and averaged. All values were obtained after standardization to the amount of CFTR immunoprecipitated. The black bar represents αA-crystallin protein levels coprecipitated with wild-type CFTR, and the gray bar represents αA-crystallin protein levels coprecipitated with ΔF508-CFTR. Vertical bars indicate the SEs of the mean. *p = 0.012 (one-way analysis of variance for 2 correlated samples).

Figure 8.

αA-crystallin inhibits NBD1 aggregation. NBD1 aggregation was measured as a change in absorbance at 400 nm after its dilution from guanidinium hydrochloride in the absence or presence of αA-crystallin. Data were standardized to the amount of aggregation in the absence of chaperone at the 600-s time point. Open squares represent endogenous NBD1 aggregation, and closed symbols represent NBD1 aggregation in the presence of αA-crystallin. Triangles, αA-crystallin:NBD1 molar ratio of 0.5:1; squares, αA-crystallin:NBD1 molar ratio of 1:1.