The Protease Ste24 Clears Clogged Translocons

Abstract

Translocation into the endoplasmic reticulum (ER) is the first step in the biogenesis of thousands of eukaryotic endomembrane proteins. Although functional ER translocation has been avidly studied, little is known about the quality control mechanisms that resolve faulty translocational states. One such faulty state is translocon clogging, in which the substrate fails to properly translocate and obstructs the translocon pore. To shed light on the machinery required to resolve clogging, we carried out a systematic screen in Saccharomyces cerevisiae that highlighted a role for the ER metalloprotease Ste24. We could demonstrate that Ste24 approaches the translocon upon clogging, and it interacts with and generates cleavage fragments of the clogged protein. Importantly, these functions are conserved in the human homolog, ZMPSTE24, although disease-associated mutant forms of ZMPSTE24 fail to clear the translocon. These results shed light on a new and critical task of Ste24, which safeguards the essential process of translocation.

Figure 1. A folding prone SRP independent protein can cause translocon clogging

A. When translation and translocation are not tightly coupled, substrates can fold in the cytosol during translocation, stopping further insertion and clogging the translocon. B. A folding prone “clogger” construct can both induce and quantitate clogging. Under regulation of the galactose inducible promoter we fused the SRP independent substrate Pdi1 that bears 5 glycosylation sites, to the stably folding dihydrofolate reductase (DHFR) enzyme followed by 3 glycosylation sites and a hemagglutinin (HA) tag (left). When clogged this fusion protein can be easily recognized on a western blot as it runs as an intermediate hemi-glycosylated form (right). C. In WT cells, most of the clogger is fully glycosylated, while in mutants of the SRP independent translocon (sec62-DAmP, sec63-DAmP, Δsec66, Δsec72) the cytosolic unglycosylated band is prevalent. Treatment of WT cells with tunicamycin, blocking N-linked glycosylation, generates only the fastest migrating form of the clogger. Split panes represent different lanes on one gel, and are shown as such hereafter. D. The clogger construct was routed to the SRP dependent pathway by appending the hydrophobic core of the DPAP B signal anchor. 26% the SRP independent clogging construct becomes clogged, while only 5% of the SRP dependent counterpart does (n=3, Mean±SEM, p-value<0.001). See also Figure S1.

Figure 2. A systematic screen highlights a role for Ste24 in resolving translocon clogging

A. The clogging construct was inserted into the yeast deletion and hypomorphic allele collections via SGA and its effect on growth was measured relative to a control protein (mCherry). B. ~110 mutants showed differential growth in the presence of the control protein or the clogger (FDR corrected p-value<0.05). Mutants attenuated for the SRP independent translocon, the unfolded protein response (UPR) and the ER protease Ste24 showed notably slower growth with the clogger (for a list of all strains see Table S1). C. Liquid growth was measured for WT and Δste24 cells either expressing the clogger or not (n=42, Mean±SD). D. Clogging in WT or Δste24 cells expressing the clogger was quantified by western blot, probing for the clogging construct with anti-HA. The clogged form of the construct accumulates in Δste24 cells. As a control, WT cells expressing the clogger were treated with tunicamycin. E. The localization of the endogenous SRP independent substrates Gas1, Kar2 and Pdi1 were assayed by western blot. Less translocation takes place when expressing the clogger in Δste24 cells. F. WT and mutant cells were analyzed by western blot for the localizations of Gas1, Kar2 and Pdi1. Mutant cells were either depleted of Ste24 (using a repressible Gal promoter on glucose), deleted for the auxillary translocon component Sec66 or both. SRP independent substrates accumulate in the cytosol when translocation becomes inefficient and the quality control actions of Ste24 are missing. See also Figure S2 and Tables S1–S4.

Figure 3. The role of Ste24 in relieving clogging is not an indirect effect of its previously described functions

A. Western blotting with anti-HA to quantify clogging in CAAX processing mutants- Δste24, Δram1 or Δaxl1 demonstrates that only Δste24, but not other CAAX mutants, generates clogging B. WT cells expressing the clogger were treated with the ER stress inducing agent, DTT, and subsequent clogging was assessed by anti-HA western blot. (n=3, Mean±SEM) showing that clogging is not exacerbated by ER stress. C. A serial dilution growth assay was carried out for WT or Δste24 cells, bearing either an empty plasmid, a plasmid encoding for functional STE24 or for the proteolytically dead mutant, STE24_E296G. These strains were examined for growth when the clogger is either repressed or expressed, showing that the metalloprotease activity of Ste24 is required to restore growth with the clogger. D. Clogging was assayed by western blot with anti-HA in Δste24 cells expressing the clogger and a plasmid encoding for either WT STE24 or STE24_E296G. While Δste24 bearing the STE24 plasmid demonstrates complete rescue of the phenotype, the STE24_E296G plasmid fails to rectify clogging and may even act as a dominant negative. See also Figure S3.

Figure 4. Clogging recruits Ste24 to the SRP independent translocon

A. The proximity of Ste24 to translocon subunits was visualized with a split Venus assay- fusing the C terminus of the fluorophore Venus (VC) to Ste24 and the N terminus (VN) to the auxiliary translocon subunits Sec63, Sec66, Sec72 or the SRP receptor subunit Srp102. Upon expression of the clogger, a ring-like ER signal is seen with Ste24 and subunits of the SRP independent translocon. Scale bar- 5μm. B. Fluorescence levels were measured for all split-Venus pairs in the presence or absence of the clogger by flow cytometry vs an untagged strain (n=3, Mean±SEM). Fluorescence in the presence of the clogger was increased when UPR was compromised (Δhac1), indicating that this is not an indirect increase due to this stress pathway. C. The amounts of the non essential subunits of the translocon, Sec66-GFP and Sec72-GFP, were examined with anti-GFP western blot in the presence or absence of the clogger in either WT or Δste24 strains and were unaltered. Pgk1 was used as a loading control. D. Translation was blocked by cycloheximide (CHX) in a WT strain lacking or expressing the clogger and samples were taken at the indicated time points. The samples were probed by western blot for endogenous Sec61, Sec62 and Sec63, showing that the half-life of these essential SRP independent translocon subunits is not altered upon clogger expression. See also Figure S4.

Figure 5. Ste24 directly interacts with the clogged protein and cleaves it

A. Cells expressing the clogger and an empty vector or a FLAG tagged version of the proteolytically dead STE24_E296G were subjected to anti-flag immunoprecipitation. A 1/200 fraction of the sample input and the precipitate were subjected to WB, probing for both the clogger construct (Anti-HA) as well as STE24_E296G (Anti-Flag). Ste24 binds to the clogged and unglycosylated forms of the construct. B. Cycloheximide (CHX) treatment of WT or Δste24 cells expressing the clogger was analyzed by anti-HA western blot (n=3, Mean±SEM), showed that only the clogged fraction is stabilized in Δste24 cells (p-value<0.05). Pgk1 was used as a loading control. C. Anti-HA western blot was carried out to examine fragments generated from the clogger in WT or Δste24 cells, without or with a genetic attenuation of the proteasome (scl1-DAmP). Clogger fragments not present in the absence of Ste24 are marked with an arrowhead. See also Figures S5 and S6.

Figure 6. The unclogging activity of Ste24 is conserved in the human homologue, ZMPSTE24

A. The fraction of the clogger found in each translocational form was quantified in Δste24 cells bearing plasmids encoding for functional ZMPSTE24 or disease-associated mutations (n=3, Mean±SEM). Lack of Ste24 doubles clogging (p-value<0.05). Mammalian ZMPSTE24, but not all disease-associated mutants, can restore clogging to WT levels. B. The fraction of clogged protein found in yeast expressing the WT or mutant forms of ZMPSTE24 (Fig 6A) was plotted against the known residual proteolitic activity of each ZMPSTE24 protease (Barrowman et al., 2012), showing a correlation (R²- 0.98). The outlier- mutation L438F is the only mutation located in the peptide-binding groove of ZMPSTE24 (Quigley et al., 2013). C. A human folding prone construct, made up of the luminal co-chaperone ERdj3, C terminally fused to GFP and 3 glycosylation sites (3Gly) was constructed. D. HEK293 cells were transfected with the human clogging prone construct and treated with the proteasome inhibitors MG132 or Bortezomib to stabilize cytosolic forms of the clogger. Additionally, cells were exposed to lopinavir (LPV) or tipranavir (TPV) to inhibit ZMPSTE24 function. Protein samples were extracted and analyzed by an anti-GFP western blot, showing an the accumulation of the hemiglycosylated clogged protein when ZMPSTE24 is inhibited. As a control, cells were treated with tunicamycin. See also Figure S7.

Figure 7. Ste24 approaches the translocon and cleaves clogged substrates during faulty translocation

The SRP independent pathway is intrinsically more prone to translocon clogging, as translation and translocation are not tightly coupled. Thus partial folding of a cytosolic domain while the luminal domain has already engaged chaperones, can generate a scenario whereby the protein can neither proceed to insert nor be retrotranslocated. To resolve this stalemate, Ste24 engages with the SRP independent translocon, and cleaves the clogged substrate, whose fragments can then be degraded by the proteasome. Thus Ste24 function is essential for freeing clogged translocons for further rounds of insertion.