A Cdc48 "Retrochaperone" Function Is Required for the Solubility of Retrotranslocated, Integral Membrane Endoplasmic Reticulum-associated Degradation (ERAD-M) Substrates

Abstract

A surprising feature of endoplasmic reticulum (ER)-associated degradation (ERAD) is the movement, or retrotranslocation, of ubiquitinated substrates from the ER lumen or membrane to the cytosol where they are degraded by the 26S proteasome. Multispanning ER membrane proteins, called ERAD-M substrates, are retrotranslocated to the cytosol as full-length intermediates during ERAD, and we have investigated how they maintain substrate solubility. Using an in vivo assay, we show that retrotranslocated ERAD-M substrates are moved to the cytoplasm as part of the normal ERAD pathway, where they are part of a solely proteinaceous complex. Using proteomics and direct biochemical confirmation, we found that Cdc48 serves as a critical "retrochaperone" for these ERAD-M substrates. Cdc48 binding to retrotranslocated, ubiquitinated ERAD-M substrates is required for their solubility; removal of the polyubiquitin chains or competition for binding by addition of free polyubiquitin liberated Cdc48 from retrotranslocated proteins and rendered them insoluble. All components of the canonical Cdc48 complex Cdc48-Npl4-Ufd1 were present in solubilized ERAD-M substrates. This function of the complex was observed for both HRD and DOA pathway substrates. Thus, in addition to the long known ATP-dependent extraction of ERAD substrates during retrotranslocation, the Cdc48 complex is generally and critically needed for the solubility of retrotranslocated ERAD-M intermediates.

Keywords: Cdc48; E3 ubiquitin ligase; Hmg2; Npl4; Retrotranslocation; Ufd1; endoplasmic reticulum (ER); endoplasmic-reticulum-associated protein degradation (ERAD); ubiquitin; ubiquitin ligase.

FIGURE 1.

In vitro retrotranslocated Hmg2 does not associate with lipids. A, in vitro retrotranslocation of Hmg2-GFP. Microsomes were prepared from Δhrd1 strains expressing TDH3_pr-Hmg2-GFP and TDH3_pr-Hrd1. Cytosol was prepared from strains expressing TDH3_pr-UBC7 or no Ubc7 (Δubc7). To initiate ubiquitination and retrotranslocation of Hmg2-GFP, microsomes and cytosol were mixed, and the reaction was centrifuged to discern ubiquitinated Hmg2-GFP that either has been retrotranslocated into the S20 S or remained in the membrane P20 P. T, total. Following fractionation, Hmg2-GFP was immunoprecipitated from both fractions, resolved on 8% SDS-PAGE, and immunoblotted for ubiquitin or Hmg2-GFP. B, full-length Hmg2-GFP retrotranslocated in vitro. Supernatant fraction isolated from in vitro retrotranslocation assay as described above was incubated with buffer or Usp2Core for 1 h at 37 °C. Full-length Hmg2-GFP was immunoprecipitated and immunoblotted for ubiquitin or Hmg2-GFP. C, in vitro retrotranslocation of Hmg2-GFP required Cdc48. In vitro retrotranslocation was a carried out as described above. Different combinations of WT of cdc48-2 prepared from microsomes (MIC) and cytosol (CYT) were mixed to evaluate the Cdc48 contribution to Hmg2-GFP retrotranslocation. D, in vitro retrotranslocated Hmg2-GFP does not associate with lipids. Crude lysate or S100 supernatant containing retrotranslocated Hmg2-GFP was prepared, layered at the bottom of the centrifuge tube, and subjected to ultracentrifugation as described under “Experimental Procedures.” Aliquots were removed from the top to the bottom of the sucrose gradient, and each fraction was directly immunoblotted for Sec61 and PGK with anti-Sec61 or anti-PGK, respectively, or immunoprecipitated with anti-GFP and immunoblotted for ubiquitinated Hmg2-GFP with anti-ubiquitin. WB, Western blot.

FIGURE 2.

In vivo retrotranslocated soluble Hmg2-GFP was full length. A, GGPP induced ubiquitination and retrotranslocation of Hmg2-GFP. WT and cdc48-2 strains were grown to log phase and treated with different combinations of MG132 (25 μg/ml) and GGPP (11 μm). Crude lysate was prepared from each strain and ultracentrifuged to discern ubiquitinated Hmg2-GFP that either has been retrotranslocated into the soluble fraction (S) or remained in the membrane (P). Following fractionation, Hmg2-GFP was immunoprecipitated from both fractions, resolved on 8% SDS-PAGE, and immunoblotted for ubiquitin and Hmg2-GFP. B, full-length Hmg2-GFP retrotranslocates into soluble fraction in vivo. S100 supernatant containing retrotranslocated Hmg2-GFP was isolated from in vivo retrotranslocation assay and incubated in the presence or absence of Usp2Core for 1 h at 37 °C. Full-length Hmg2-GFP was immunoprecipitated and immunoblotted for ubiquitin or Hmg2-GFP. C, in vivo retrotranslocation of Hmg2-GFP requires Cdc48. Same as A except WT, cdc48-2, hrd2-1, and Δubc7 were used and treated with vehicle or MG132 (25 μg/ml). D, in vivo retrotranslocation of Hmg2-GFP does not require proteasome shuttle factors Rad23/Dsk2. Same experiment as C except Δrad23Δdsk2 strain was included.

FIGURE 3.

Cdc48 is required for extraction of ubiquitinated Hmg2-GFP from the ER membrane. A, microsomes were analyzed by alkali extraction (Na₂CO₃) with 0.2 m carbonate at the indicated pH values. Equal volumes of the P and TCA-precipitated S fractions were resolved by SDS-PAGE and analyzed by immunoblotting. B, for analysis of ubiquitinated Hmg2-GFP, pellet and supernatant fractions were immunoprecipitated for Hmg2-GFP and immunoblotted for ubiquitinated Hmg2-GFP with anti-Ub.

FIGURE 4.

In vivo retrotranslocated Hmg2 does not associate with lipids. Crude lysate or S100 supernatant and P100 pellet were prepared, layered at the bottom of a centrifuge tube, and subjected to ultracentrifugation as described under “Experimental Procedures.” Aliquots were removed from the top to the bottom of the sucrose gradient, and each fraction was either directly immunoblotted for Sec61 and PGK with anti-Sec61 or anti-PGK, respectively, or immunoprecipitated with anti-GFP and immunoblotted for ubiquitin and Hmg2-GFP. WB, Western blot.

FIGURE 5.

Ste6-166 retrotranslocates identically as Hmg2-GFP. A, retrotranslocated Ste6* remains soluble in the S100 fraction. Crude lysate from WT and cdc48-2 strains expressing Ste6*-3HA were centrifuged at 20,000 × g at 4 °C to yield membrane fraction (P20) and supernatant fraction (S20). S20 supernatant was further ultracentrifuged at 100,000 × g at 4 °C to yield membrane fraction (P100) and supernatant fraction (S100). Ste6*-3HA was immunoprecipitated from all fractions with anti-HA (Covance), resolved on 8% SDS-PAGE, and immunoblotted with anti-HA for Ste6*-3HA or anti-ubiquitin for ubiquitinated Ste6*-3HA. Retrotranslocated Hmg2-GFP remains in the S100 fraction. B, same as A except WT and cdc48-2 strains expressing Hmg2-GFP were used instead. C, full-length Ste6* retrotranslocates into the S100 fraction. S100 supernatant was incubated in the presence or absence of Usp2Core for 1 h at 37 °C. Ste6*-3HA was immunoprecipitated with polyclonal anti-HA and immunoblotted with anti-HA and anti-Ub.

FIGURE 6.

Cdc48 is the main interacting protein of retrotranslocated Hmg2-GFP. A, S100 supernatant containing retrotranslocated Hmg2-GFP and control strains not expressing Hmg2-GFP or expressing cytosolic GFP were precipitated with GFP-Trap beads. Following Co-IP, low pH-eluted proteins were digested with trypsin and subjected to LC-MS/MS analysis. A representative bar graph from the first two runs of mass spectrometry analysis was generated to include fold enrichment for corresponding proteins, which reflects the relative enrichment of spectral counts for interacting proteins versus the beads only and cytosolic GFP control. Proteins with ≥1.5-fold enrichment over controls in the mass spectra were included. Note: Hmg2-GFP is highlighted in green; Cdc48 is highlighted in red; proteasome subunits are highlighted blue, and Cdc48 cofactors are highlighted in purple. B, Hmg2-GFP Co-IPs with Cdc48. Strains expressing the indicated proteins were grown, and equal amounts were harvested. S100 supernatant was prepared as described under “Experimental Procedures” and immunoprecipitated with GFP-Trap to pull down retrotranslocated Hmg2-GFP. The precipitates were resolved on 8% SDS-PAGE and immunoblotted for ubiquitinated Hmg2-GFP and Cdc48. C, Hmg2-GFP Co-IPs with Npl4 and Ufd1. Same as B except Hmg2-GFP precipitates were resolved on 8% SDS-PAGE and immunoblotted for Npl4 and Ufd1. D, Ste6*-3HA-GFP Co-IPs with Cdc48. Same as B except strains expressing Ste6*-3HA-GFP were used for immunoprecipitation with GFP-Trap followed by immunoblotting for Cdc48. E, same as D except Ste6*-3HA-GFP precipitates were resolved and immunoblotted for Npl4 and Ufd1-SBP.

FIGURE 7.

All retrotranslocated Hmg2-GFP and Ste6-166 is bound to Cdc48 complex. A, ProtA-Cdc48 complements endogenous Cdc48. Strains containing CDC48, ProtA-CDC48, or cdc48-2 were compared for growth by dilution assay. Each strain was spotted at 5-fold dilutions on YPD, and plates were incubated at 30 and 37 °C. ProteinA-Cdc48 complements endogenous Cdc48 in ERAD of Hmg2-GFP. B, indicated strains expressing Hmg2-GFP were grown into log phase, and degradation was measured by a CHX chase. After CHX addition, cells were lysed at the indicated times and analyzed by SDS-PAGE and immunoblotting for Hmg2. C, strains expressing the indicated proteins were grown, and equal amounts were harvested. S100 supernatant was prepared as described under “Experimental Procedures” and immunoprecipitated with IgG-Sepharose to pull down ProtA-Cdc48. As a control, 5% of input was withdrawn, and the rest of the lysate was incubated with IgG-Sepharose. The Sepharose beads were washed, and bound proteins (B) were eluted with 2× USB. To assess the effectiveness of binding, the unbound S was also included. Samples were resolved on 8% SDS-PAGE and immunoblotted for Cdc48 with anti-Cdc48. Blotting for Hmg2-GFP with anti-GFP required treatment of all samples with Usp2Core prior to resolving on SDS-PAGE. D, same as C except S100 supernatant was immunoprecipitated with polyclonal anti-Npl4. As a control, polyclonal antibody against PGK was also used. Samples were resolved on 8% SDS-PAGE and immunoblotted for Hmg2-GFP with anti-GFP. E, same as C except S100 supernatant was immunoprecipitated with streptavidin beads to capture Ufd1-SBP. Samples were resolved on 8% SDS-PAGE and immunoblotted for Hmg2-GFP with anti-GFP and Ufd1-SBP with anti-Ufd1. F, same as C except Ste6*-3HA-GFP was analyzed for immunoprecipitation with ProtA-Cdc48. G, S100 supernatant was immunoprecipitated with polyclonal anti-Npl4. As a control, polyclonal antibody against PGK was also used. Samples were resolved on 8% SDS-PAGE and immunoblotted for Ste6*-3HA-GFP with anti-GFP. H, same as C except S100 supernatant was immunoprecipitated with streptavidin beads to capture Ufd1-SBP. Samples were resolved on 8% SDS-PAGE and immunoblotted for Ste6*-3HA-GFP with anti-GFP and Ufd1-SBP with anti-Ufd1.

FIGURE 8.

Ubiquitin is required for Cdc48 association. A, same as (Fig. 6) except GFP-Trap-captured retrotranslocated Hmg2-GFP was resuspended in 100 μl of XL buffer, treated with buffer or (5 μg) Usp2Core, and incubated for 1 h at 37 °C. GFP-Trap-agarose beads were washed twice with IPB buffer, washed once with IPW buffer, and aspirated to dryness. The precipitates were resuspended in 2× USB, resolved on 8% SDS-PAGE, and immunoblotted for ubiquitinated Hmg2-GFP and Cdc48. Hmg2 is no longer soluble upon removal of ubiquitin chains. B, left panel, S100 supernatant was prepared as described under “Experimental Procedures” and was ultracentrifuged to discern ubiquitinated Hmg2-GFP that is either in the insoluble fraction (P) or in the soluble fraction (S). Following ultracentrifugation, both P and S fractions were treated with Usp2Core, and Hmg2-GFP was immunoprecipitated from both fractions, resolved on 8% SDS-PAGE, and immunoblotted for ubiquitin and Hmg2-GFP. Right panel, same as above except Usp2Core was added to supernatant fraction prior to ultracentrifugation. SM, starting material.

FIGURE 9.

Cdc48 is required for the solubility of Hmg2-GFP. A, in the presence of Lys-48-linked polyubiquitin chains, Cdc48 disassociates from retrotranslocated Hmg2-GFP. Same as Fig. 6 except the S100 supernatant containing retrotranslocated Hmg2-GFP was treated with 5 μm Lys-48-linked polyubiquitin chains. B, retrotranslocated Hmg2-GFP is no longer soluble upon disassociation from Cdc48. Retrotranslocated Hmg2-GFP was treated with excess Lys-48-linked polyubiquitin chains or buffer and was ultracentrifuged to discern ubiquitinated Hmg2-GFP that is either in the insoluble fraction (P) or in the soluble fraction (S). Following ultracentrifugation, both P and S fractions were treated with Usp2Core.