Identification of mammalian protein quality control factors by high-throughput cellular imaging

Abstract

Protein Quality Control (PQC) pathways are essential to maintain the equilibrium between protein folding and the clearance of misfolded proteins. In order to discover novel human PQC factors, we developed a high-content, high-throughput cell-based assay to assess PQC activity. The assay is based on a fluorescently tagged, temperature sensitive PQC substrate and measures its degradation relative to a temperature insensitive internal control. In a targeted screen of 1591 siRNA genes involved in the Ubiquitin-Proteasome System (UPS) we identified 25 of the 33 genes encoding for 26S proteasome subunits and discovered several novel PQC factors. An unbiased genome-wide siRNA screen revealed the protein translation machinery, and in particular the EIF3 translation initiation complex, as a novel key modulator of misfolded protein stability. These results represent a comprehensive unbiased survey of human PQC components and establish an experimental tool for the discovery of genes that are required for the degradation of misfolded proteins under conditions of proteotoxic stress.

Figure 1. A temperature sensitive (ts) allele of the SV40 Large T antigen (LTag) fused to EGFP engages the PQC machinery.

a) Outline of the fluorescent PQC reporter system in mammalian cells. LTR: Retroviral Long Terminal Repeat, CMV: immediate early cytomegalovirus promoter, LTag; SV40 Large T antigen, ts: temperature sensitive, IRES: internal ribosome entry site, NLS: nuclear localization signal. b) U2OS cells stably expressing either a wild-type (WT) or a temperature-sensitive (ts) allele of LTag fused to EGFP were grown at 33.5°C and then shifted to 38.5°C for the indicated amount of time. Total cell lysates were probed in Western Blotting with the indicated antibodies. c) Same as b), except that cells were either treated with DMSO or the proteasome inhibitor MG132 at a final concentration of 2 µM.

Figure 2. The PQC assay quantitatively measures the degradation of LTag(ts)-EGFP at the restrictive temperature.

a) U2OS cells stably expressing LTag(ts)-EGFP and NLS-DsRedExpress2 (NLS-Red) were grown in 384 well-plates for 48 hrs at 33.5°C and then shifted at the indicated temperatures for 24 hrs. DRAQ5 stains the DNA and is used to identify the nuclei. Cells were fixed in paraformaldehyde and imaged sequentially at 488 nm, 568 nm, and 640 nm (40× magnification, 1 imaging field). Scale bar: 20 µm b) Montage of 12 40× fields representing the entire population of cells in a 384-well. Scale bar: 100 µm. c) Same as b), except that cells were treated with MG132 (250 nM). (-) indicates the untreated control. d) Same as b) except that U2OS cells stably expressing the LTag(WT)-EGFP and NLS-DsRedExpress2 were used. e) Histograms representing the quantification of nuclear fluorescence intensities and the EGFP/DsRedExpress2 ratio. Typically >300 cells were imaged per well. Values represent averages +/− S.E.M of 4 experiments.

Figure 3. Primary hits from the Ubiquitin Proteasome System (UPS) siRNA screen.

a) U2OS cells expressing LTag(ts)-EGFP and NLS-DsRedExpress2 were screened in triplicate at 38.5°C using a library containing 991 siRNA pools targeting human genes belonging to the Ubiquitin Proteasome System (UPS). siRNA pools were ranked according to their Z-score in the EGFP/DsRedExpress2 ratio when compared to the median value of the population. Positive control wells (red) were treated with MG132 (250 nM). The horizontal dashed line represents the threshold (Z = 2) used for the selection of the positive hits. b) U2OS cells expressing LTag(ts)-EGFP and NLS-DsRedExpress2 were treated with an siRNA pool against the PSMD8 subunit of the 19S proteasome, the 20S proteasome chaperone C13ORF12, or the 20S proteasome chaperone DSCR2/PAC1 for 48 hrs at 33.5°C and then shifted to 38.5°C for 24 hrs. Images shown here are montages of 12 imaging fields at 40× magnification. Scale bar: 100 µm. c) Histogram representing the Z-scores of the EGFP/DsRedExpress2 ratio obtained in the secondary screen at 38.5°C using a validation library containing 48 On-Target Plus siRNA pools targeting selected primary hits from the UPS and Ubi123 screens. U2OS cells expressing LTag(ts)-EGFP and NLS-DsRedExpress2 were used in this experiment. Typically >300 cells were imaged per well. Values represent averages +/− S.E.M of 4 experiments.

Figure 4. PQC Validation of primary hits.

a) U2OS cells expressing either LTag(WT)-EGFP or LTag(ts)-EGFP and NLS-DsRedExpress2 were screened in quadruplicate at 33.5°C or 38.5°C using the validation library described in Figure 3C. The ΔZ-score is calculated as the difference between the EGFP/DsRedExpress2 ratio Z-score obtained using LTag(ts)-EGFP cells and the EGFP/DsRedExpress2 ratio Z-score for LTag(WT)-EGFP expressing cells. C13ORF12 is also known as POMP. b) Deconvolution and measurement by quantitative RT-PCR of the RNA silencing activity of the original siGenome pool of 4 siRNA oligos targeting the POMP gene. c) Measurement of the biological activity of the 4 siRNA oligos directed against the POMP gene in the PQC activity assay. Red bars represent siRNA oligos possessing siRNA silencing activity (see b)). Values represent averages +/− S.E.M of 4 experiments.

Figure 5. Genome-wide siRNA screen for PQC factors.

a) U2OS cells expressing LTag(ts)-EGFP and NLS-DsRedExpress2 were screened against a library containing ∼18, 000 siRNA pools targeting human protein-coding genes. The histogram represents the distribution of the EGFP/DsRedExpress2 ratio robust Z-score measured for each siRNA pool. The location of the EGFP/DsRedExpress2 ratio robust Z-score threshold used to select positive siRNA pools is indicated. b) Analysis of hits from the genome-wide PQC screen reveals significant enrichment for translation/translation initiation processes. Enrichment analysis was conducted using GeneGo Process Networks with a false discovery rate (FDR) of 0.05. 9 of the top 84 genes are found within this network, including EIF3A, EIF3F, NHP2-like protein 1 (NH2L1), NOP56/NOL5A, RPS13, RPS24, RPS28, RPS4X and RPS8. c) U2OS cells expressing LTag(ts)-EGFP or LTag(WT)-EGFP and NLS-DsRedExpress2 were independently transfected for validation with 4 siRNAs targeting 71 of the genes that were scored as positive in the primary PQC screen. Positive hits in this secondary screen were classified as genes whose silencing by at least two independent oligo siRNAs had a EGFP/DsRedExpress2 ratio larger than 140% of the value obtained by transfection of a non-targeting siRNA oligo. The table shows such positive genes in the LTag(ts)-EGFP or LTag(WT)-EGFP expressing cell lines, respectively.