Genome-wide CRISPR Analysis Identifies Substrate-Specific Conjugation Modules in ER-Associated Degradation

Abstract

The ubiquitin proteasome system (UPS) maintains the integrity of the proteome by selectively degrading misfolded or mis-assembled proteins, but the rules that govern how conformationally defective proteins in the secretory pathway are selected from the structurally and topologically diverse constellation of correctly folded membrane and secretory proteins for efficient degradation by cytosolic proteasomes is not well understood. Here, we combine parallel pooled genome-wide CRISPR-Cas9 forward genetic screening with a highly quantitative and sensitive protein turnover assay to discover a previously undescribed collaboration between membrane-embedded cytoplasmic ubiquitin E3 ligases to conjugate heterotypic branched or mixed ubiquitin (Ub) chains on substrates of endoplasmic-reticulum-associated degradation (ERAD). These findings demonstrate that parallel CRISPR analysis can be used to deconvolve highly complex cell biological processes and identify new biochemical pathways in protein quality control.

Figure 1.. Parallel forward genetic analysis maps substrate-selective ERAD pathways.

(A) Schematic of the reporters used in this study. Triangles represent N-glycosylation sequons; -SH represents cysteines. (B) Fluorescence histograms of the indicated cell lines treated +/− doxycycline (dox) for 16 hr before adding inhibitors for 3 hr; representative of three independent experiments. See also Fig. S1B–C. (C) Cells were treated with inhibitors for 3 hr and median fluorescence intensity (MFI) was measured by flow cytometry analysis at the indicated times after inhibiting protein synthesis with emetine; representative of two independent experiments. (D) Genetic analysis workflow. (E) Screen metrics are well correlated with substrate degradation rates. Linear least-squares regression of gene effects from all four screens plotted against normalized degradation rates, calculated as the slope of the log-transformed decay kinetics, obtained by translation shut-off analysis in cells expressing single sgRNAs. Each data point represents the disruption of a specific gene for a given substrate indicated by color. Grey, control sgRNA. (F) Hierarchical clustering of genes by gene score. Gene score signs were assigned based on guide enrichment, with a negative sign indicating guide disenrichment in the GFP^high population relative to the GFP^low population and a positive sign indicating guide enrichment in the GFP^high population relative to the GFP^low population. Analysis includes genes in the indicated categories at ≤1% FDR in at least one screen. (G) Bubble plot of ERAD genes. Cytosolic E3 ligases identified in the genetic screens are in the dashed box. See also Figures S1–3 and Tables S1-4.

Figure 2.. Degradation of GFP^u* is promoted by Ub chain diversification.

(A) Plots of GFP^u* turnover assessed by protein synthesis shut-off and immunoblotting of lysates from cell expressing the indicated sgRNAs; data are the mean of three independent experiments ± SEM. *P<0.05; **P<0.001 determined by Student’s t-test. See also Fig. S4A–C. (B) Immunoblots of TX-100 soluble (S) and insoluble (I) cell fractions. Treatments were 3 hr; representative of two independent experiments. (C) Immunoblots of cytosolic (cyto) and membrane (memb) cell fractions; representative of three independent experiments. Dashed line indicates removal of irrelevant gel lanes. (D) Plots of GFP^u* turnover assessed at the indicated times after protein synthesis shut-off by flow cytometry analysis of cells expressing the indicated single or double sgRNAs. (E) GFP^u* turnover in UBE3C^KO cells expressing indicated UBE3C variants. Top: Immunoblots; bottom: GFP^u* quantification; representative of two independent experiments. (F-G) Ub_n-GFP^u* conjugates affinity captured from lysates treated with UPS inhibitors (F-G) and expressing exogenous UBE3C variants (G); representative of three (F) or two (G) independent experiments. See also Fig. S4D. (H) Schematic of (I). (I) Ub_n -GFP^u* conjugates affinity captured with immobilized Halo-TRABID NZF1 and treated with the indicated DUBs for 1 hr. Arrows indicate vOTU-resistant Ub_n -GFP^u*; representative of two independent experiments. See also Fig. S4F. See also Figure S4.

Figure 3.. INSIG1-GFP and A1AT^NHK-GFP are modified with heterotypic K11/K48 polyUb chains.

(A) Plots of INSIG1-GFP turnover assessed by protein synthesis shut-off and flow cytometry analysis of cells expressing the indicated sgRNAs; data are the mean ± SEM of three independent experiments. *P<0.05; **P<0.001 determined by Student’s t-test. See also Figs. S5A–B. (B-C) Plots of INSIG1-GFP turnover assessed by protein synthesis shut-off and flow cytometry analysis of cells expressing the indicated single or double sgRNAs. (B) data are representative of two independent experiments. (C) data are the mean ± of three independent experiments. *P<0.05 determined by Student’s t-test. See also Fig. S5C. (D), (G) Ub and Ub KGG peptides from GFP immunoisolates from urea-denatured lysates were analyzed by LC-MS/MS. Data are the integrated MS ion peak intensity of Ub or Ub-KGG normalized to the integrated MS ion peak intensity of GFP; representative of two independent experiments. (D) INSIG1-GFP, (G) A1AT^NHK-GFP. (E), (H) GFP immunoisolates from urea-denatured lysates were immunoblotted with linkage- or architecture-specific Ub antibodies; (E) INSIG1-GFP, (H) A1AT^NHK-GFP; representative of three independent experiments. (F), (I) Ub conjugates were immunoisolated using linkage- or architecture-specific Ub antibodies from denatured lysates; (F) INSIG1-GFP from NMS-873-treated lysates. Dashed line indicates removal of irrelevant gel lanes, (I) A1AT^NHK-GFP from MG132-treated lysates. See also Figures S5–6 and Table S5.

Figure 4.. GFP-RTA^E177Q is rapidly degraded by evading glycan quality control.

(A) Bubble plot of N-glycan biosynthesis genes. (B) Structure of the core high-mannose N-glycan. The α1,2-glycosidic bonds recognized by EDEMs are highlighted in red and the two C-branch α1,6-glycosidic bonds that contact the binding pocket of OS9 are highlighted in yellow. (C) Top: Schematic of GFPRTA^E177Q N-glycosylation sequons. Bottom: Effect of endo H (H) or PNGase F (F) endoglycosidases on GFP-RTA^E177Q mutant mobility. Filled and open arrowed are glycosylated (GP) and nonglycosylated (nGP) GFP-RTA^E177Q, asterisks indicate proteolytic cleavage products; representative of three independent experiments. (D), (F) GFP-RTA^E177Q (D) or GFP-RTA^{N11Q, E177Q} (F) turnover in cells treated for 6 hr with kifunensine or for 4.5 hr with tunicamycin. Arrows indicate the glycosylated (GP) and non-glycosylated (nGP) forms of GFP-RTA. Arrowheads indicate the core glycosylated (CG), mannose trimmed (deM), and nonglycosylated (nGP) forms of CD147. Right: Quantification of GFP-RTA turnover; data are the mean ± SEM of three independent experiments. *P<0.05; **P<0.001 determined by Student’s t-test. (E) Comparison of GFP-RTA^E177Q and GFP-RTA^{N11Q, E177Q} turnover; data are the mean ± SEM of three independent experiments. **P<0.01 determined by Student’s t-test. See also Figure S7.

Figure 5.. Model of ERAD-L, -M, and -C.

(A) Glycosylated ERAD-L substrates utilize the HRD1 complex for ER dislocation and ubiquitylation. Substrate-specific heterogeneity in ERAD-L is revealed by GFP-RTA^E177Q, which can engage luminal recognition factors for slow delivery to the HRD1 complex or can be rapidly dislocated and degraded by bypassing glycan recognition machinery. Dislocated ERAD-L substrates are modified with K48 and K11 Ub linkages and targeted for destruction by cytosolic proteasomes. (B) The ERAD-M substrate INSIG1-GFP utilizes GP78 for ER dislocation and ubiquitylation. Efficient delivery to cytosolic degradation machinery may be facilitated by conjugation of K11/K48 chains that increase affinity for p97/VCP and the proteasome. (C) Efficient degradation of the ERAD-C substrate GFP^u* requires conjugation of heterotypic K48/K29 chains via the ER-embedded E3 ligase TRC8 and the cytosolic E3 UBE3C.