The chaperone BAG6 captures dislocated glycoproteins in the cytosol

Abstract

Secretory and membrane (glyco)proteins are subject to quality control in the endoplasmic reticulum (ER) to ensure that only functional proteins reach their destination. Proteins deemed terminally misfolded and hence functionally defective may be dislocated to the cytosol, where the proteasome degrades them. What we know about this process stems mostly from overexpression of tagged misfolded proteins, or from situations where viruses have hijacked the quality control machinery to their advantage. We know of only very few endogenous substrates of ER quality control, most of which are degraded as part of a signaling pathway, such as Insig-1, but such examples do not necessarily represent terminally misfolded proteins. Here we show that endogenous dislocation clients are captured specifically in association with the cytosolic chaperone BAG6, or retrieved en masse via their glycan handle.

Figure 1. BAG6 binds dislocation substrates via its N-terminal domain.

A. Cartoon depicting the BAG6 and EBV DUB constructs used in this study. UBL: Ubiquitin-like domain. P rich: proline-rich domain. BAG: BAG-domain. EBV DUB: catalytic domain of BPFL1. UBX: a p97 interaction domain. B. HEK293 cells were transfected with either BAG6 or BAG6Δ1-472. Cells were lysed in NP40, followed by IP for Derlin2. BAG6 levels were determined by blotting with HA-antibody, where GAPDH serves as a loading control. C. HEK293 cells were transfected with Ri332 (untagged), either BAG6 or BAG6Δ1-472, and in the presence or absence of UBX-EBV. The cells were lysed in NP40, followed by an IP for BAG6 (HA). Ri332 levels were detected using a Ri332 antibody. Input levels of BAG6 and UBX-EBV were detected by immunoblotting with HA and Flag antibodies respectively. D. As in C, except that TCRα is substituted for Ri332.

Figure 2. BAG6 expression does not impair dislocation.

A. HEK293 cells were transfected with Ri332 and either BAG6, BAG6Δ1-472 or empty vector as a control. The cells were pulse-labeled for 10 min with S35 Cys/Met, after which samples were taken at the indicated chase times. The cells were lysed in 1% SDS to dissolve all protein complexes, diluted in NP40 lysis buffer to 0.1% SDS, followed by IP with HA. The right panel represents untransfected cells pulse-labeled for 1 h with S35 cys/met, after which the individual samples were subjected to IP with BAG6 antibody. B. As in A, except that Ri332 is substituted for TCRα, and the cells were lysed in NP40 lysis buffer.

Figure 3. Endogenous dislocation substrates associate with BAG6.

A. HEK293 cells were transfected with UBX-EBV and either BAG6, BAG6Δ1-472 or empty vector as a control. The cells were labeled with S35 Cys/Met for 4 h in the presence of 30 µM Z-VAD-FMK, and where indicated 2 mM DTT. The cells were collected, lysed in NP40 lysis buffer, and the samples split in two. All samples were subjected to IP for BAG6 (HA), the retrieved protein eluted and placed at 37°C for 2 h in the presence or absence of PNGaseF. A sequential IP for PDI serves as a loading control, bottom panel. B. As in A, full-length BAG6 only, but in the absence of Z-VAD-FMK. C. As in A, full-length BAG6 only, but the sample was now split in three, and treated with either no enzyme, PNGaseF, or endoH.

Figure 4. UBX-EBV and BAG6 are found in one protein complex.

A. HEK293 cells were transfected with BAG6 and either EBV DUB or UBX-EBV. Cells were lysed in NP40 lysis buffer, and subjected to IP for BAG6 (HA). The eluate was blotted for EBV DUB (FLAG). Input levels of EBV DUB, BAG6 and GAPDH were assessed with the relevant antibodies. B. HEK293 cells were transfected with UBX-EBV and either BAG6 or BAG6Δ1-472. Cells were lysed in NP40, and subjected to IP for UBX-EBV (FLAG). The eluate was blotted for BAG6 (HA). Input levels of UBX-EBV, BAG6 and GAPDH were assessed with the relevant antibodies.

Figure 5. DTT does not significantly halt protein dislocation, but affects translocation into the ER.

A. HEK293 cells were transfected with either Ri332 or TCRα, pulse-labeled for 15 min with S35 Cys/Met, after which samples were collected at the indicated chase times. 2 mM DTT was added during the pulse period when indicated. After lysis in 1% SDS, the lysate was diluted in NP40 to 0.1% SDS and subjected to IP for either Ri332 or TCRα. A sequential IP for PDI serves as a loading control. B. Quantification of the degradation of Ri332 as depicted in A (n = 3). C. Quantification of the degradation of TCRα as depicted in A (n = 3). D. Quantification of the relative amount of fully versus not fully glycosylated protein as depicted in A (n = 3). E. HEK293 cells were transfected with TCRα and BAG6. The cells were pulse-labeled for 15 min with S35 Cys/Met, after which samples were collected at the indicated chase times. 2 mM DTT was added during the pulse period when indicated. After lysis in NP40, the samples were subjected to IP for TCRα. F. Cells were transfected with Ri332 and treated with 2 mM DTT for the indicated time period. At the end of the treatment with DTT, the cells were pulse labeled for 10 min with S35 Cys/Met. The cells were lysed in 1% SDS, the lysate was diluted in NP40 to 0.1% SDS and subjected to IP for Ri332 (HA). Inpute lysate was immunoblotted for GAPDH as a loading control. G. As in F, except for a sequential IP for PDI.

Figure 6. Dislocation can be visualized by the appearance of glycoproteins in the cytosol.

HEK293 cells were transfected with either empty vector as a control, or UBX-EBV and/or p97 variants as indicated. The cells were labeled for 4 h with 35S Cys/Met, after which they were collected as a cell pellet and placed on ice. The cells were then resuspended in 100 µl HBSS containing 200 nM PFO. The cells were washed once in cold HBSS, and resuspended in 100 µl HBSS, after which they were transferred to 37°C to induce pore formation. The cells were pelleted for 5 min at 2000 rpm, after which the soluble cytosolic fraction was collected and the pellet discarded. The soluble fraction was dissolved in NP40 lysis buffer, split into two equal parts, and subjected to IP with immobilized Concanavalin A. Retrieved glycoproteins were eluted with mannose and placed at 37°C in the absence or presence of PNGaseF.