Protein quality control at the inner nuclear membrane

Yeast methods and plasmids

Yeast genome manipulations (gene deletions and tagging) were performed using conventional PCR targeting, as described³⁰. Yeast strains and plasmids used in this study are listed in Supplementary Tables 4 and 5, respectively.

β-galactosidase activity assay

Cells were grown in synthetic minimal medium and β-galactosidase activity was measured in N-lauroyl-sarcosine-permeabilized cells as described³¹.

RNA isolation and quantitative reverse transcriptase-PCR

Strains with auxotrophies complemented by plasmids pRS316 (URA3), pRS317 (LYS2) and pAB1 (HIS3, MET15, LEU2) were grown in synthetic minimal medium to 10⁷ cells/ml and harvested by centrifugation. RNA was isolated using RiboPure™ Yeast Kit and treated with Turbo-DNase (Ambion). The quality of RNA preparations was assessed by electrophoresis on a 1% agarose gel with 10 mM guanidine thiocyanate, and the lack of DNA contamination was confirmed by PCR. One μg of RNA was used for cDNA synthesis with oligo (dT)12-19 (Invitrogen) using Super Script® III Reverse Transcriptase (Life Technologies). Quantitative PCR reactions were prepared using Kapa Sybr®Fast qPCR Master Mix (KapaBiosystems). cDNA mixtures were diluted 1/40 and 5 μl were used in a reaction volume of 20 μl with the following primer pairs:

• AGP1fwd(CTGCCGTGCGTAGGTTTT) and

• AGP1rev(AGAAGAAGGTGAGATAGCCGA);

• GNP1fwd(CACCACAAGAACAAGAACAGAAAC) and

• GNP1rev(ACCGACCAGCAAACCAGTA);

• TAF10fwd(ATATTCCAGGATCAGGTCTTCCGTAGC) and

• TAF10rev(CAACAACAACATCAACAGAATGAGAAGACTAC).

The levels of gene expression in three biological replicates were determined in two separate amplifications with triplicate technical replicates of each of the three genes analysed using the comparative ΔC_T method (Rotor Gene 6000, Corbett Life Science). Relative levels of AGP1 and GNP1 mRNA were normalized with respect to the levels of the invariant reference geneTAF10; the levels of AGP1 and GNP1 in strains carrying the indicated mutations were subsequently averaged and normalized to the levels of expression in the corresponding isogenic wild type strains.

Purification of decahistidine-Ub protein conjugates

Ubiquitylated proteins were purified from100 OD₆₀₀ equivalents of exponentially growing yeast cells expressing 10 His-tagged ubiquitin using a protocol adapted from Becuwe et al. (2012)³². Cell pellets were resuspended in 2 ml 20% trichloroacetic acid and lysed for 2 min using glass beads in a Disrupter Genie homogenizer (Scientific Industries). After precipitation, proteins were resuspended in 3 ml guanidium buffer (6 M GndHCl, 100 mM Tris-HCl pH 9, 300 mM NaCl, 10 mM imidazole, 0.2% Triton X-100, 5 mM chloroacetamide), clarified at 30,000 g and incubated for 1.5 h at room temperature withTALON® Metal Affinity Resin (Clontech). The beads were then washed with wash buffer (8 M Urea, 100 mM sodium phosphate pH 7.0, 300 mM NaCl, 5 mM imidazole, 0.2% Triton X-100, 5 mM chloroacetamide) containing 0.2% SDS (twice) and lacking SDS (twice).10His-ubiquitin conjugates were finally eluted with 200 μl elution buffer (8 M Urea, 100 mM sodium phosphate pH 7.0, 300 mM NaCl, 250 mM imidazole, 0.2% Triton X-100, 5 mM chloroacetamide). Total extracts (1% of the amount used for purification) and ubiquitin conjugate eluates were analyzed by SDS-PAGE and immunoblotting with antibodies against the TAP tag (PAP, 1:1,000, Sigma). As controls, levels of ubiquitin conjugates and Pgk1 were assessed with anti-ubiquitin (P4D1 HRP conjugate, 1:1,000, Santa Cruz) and anti-Pgk1 antibodies (clone 22C5D8, 1:10,000, Invitrogen), respectively. Immunogenic proteins were detected by chemiluminescence using SuperSignal West Femto Substrate (Thermo Scientific) and recorded using autoradiographic films (CP-BU, Agfa) processed with a Curix 60 developing machine (Agfa).

Purification of hexahistidine-Ub protein conjugates

Ubiquitylated proteins were purified from 50 OD₆₀₀ equivalents of exponentially growing yeast cells expressing 6His-tagged ubiquitin as previously described³³. 6His-ubiquitin conjugates were retained on nickel–nitrilotriacetic acid Sepharose beads (Qiagen) and eluted (eluate, E) in the presence of 300 mM (Stp1-HA, Stp1-RI_17-33-HA) or 500 mM (Stp2-HA, Stp2Δ_2-13-HA) imidazole. Total extracts (T), flow through (F) and eluate fractions were precipitated with 10% trichloroacetic acid, analysed by SDS-PAGE and immunoblotting with antibodies against the HA tag (1:5,000, Roche) and the signals were recorded using autoradiographic film (CL-Xposure, Thermo Scientific). As controls, levels of ubiquitin conjugates and Pgk1 were assessed with anti-His₅ (1:5,000, Qiagen) and anti-Pgk1 antibodies (1:10,000, In Vitrogen), respectively, and detected by chemiluminescence using SuperSignal West Dura Extended Duration Substrate (Thermo Scientific) and a Molecular Imager ChemDoc XRS+ with Image Lab v3 build 11 software (BioRad). Loaded T and F fractions correspond to 2% (Stp1-HA or Stp1-RI_17-33-HA) and 0.7% (Stp2-HA or Stp2Δ_2-13-HA) of the amount used for purification of ubiquitin conjugates.

Bimolecular fluorescence complementation

BiFC interaction assays were performed using E2 and E3 proteins tagged with the VC173 and VN155 fragments (VC and VN) of the Venus fluorescent protein, respectively³⁴. All E2 and E3 proteins were tagged C-terminally, with the following exceptions that were N-terminally tagged: Ubc6, since the C-terminus of Ubc6 faces the ER lumen³⁵; Ubc7, to preserve its interaction with Cue1³⁶; Ubc1, because the growth of strains expressing Ubc1 endogenously tagged at the C-terminus with VC appears compromised; the E3 proteins Far1, Mot2, Nam7, Prp19, Ste5 and Tfb3, since they all have their E2 binding domain at the N-terminus. All fusions were expressed from their endogenous chromosomal loci, with the exception of Rsp5-VN, which was expressed from its endogenous promoter on the centromeric plasmid pGR703 (Supplementary Table 5).

Strains expressing VC-tagged E2 proteins were constructed in the scEB115 background. scEB115 carries markers for selection of haploid progeny in automated crosses (can1∷STE2pr-spHIS5 and lyp1∷STE3pr-HPH) and expresses the proteasomal subunit Rpn7 fused to the red fluorescent protein tDimer2 as nuclear marker (Supplementary Table 4). Strains expressing VN-tagged E3 proteins were either obtained from a commercially available collection (Bioneer Corporation) or constructed by homologous recombination in the BY4741 background. Expression of VC- and TAP-tagged fusions was validated by immunoblotting with mouse anti-GFP (clones 7.1 and 13.1, Roche) and peroxidase anti-peroxidase (Sigma) antibodies to detect VC and TAP tag, respectively, and mouse anti-action (clone c4, Merck Millipore) for loading controls.

Strains expressing individual E2 and E3 protein fusions were crossed to produce an array of yeast strains each expressing Rpn7-tDimer2 and a unique combination of tagged E2 and E3 proteins, as described²⁰. The resulting strains were cultivated overnight at 20°C in YPD medium and diluted in low fluorescence medium³⁷ 3-4 h before imaging. Imaging was performed in 8-well LabTek chambers or 96-well plates (Imaging plates CG, Zell-Kontakt) using an inverted Leica SP8 confocal microscope. Images of the BiFC signal were collected using a 514 nm laser and a narrow band-pass filter (525-538 nm) around the emission peak of the Venus fluorescent protein to reduce the contribution of cellular autofluorescence. Rpn7-tDimer2 was imaged simultaneously using a 580-630 nm filter. Cellular auto fluorescence was imaged separately using the same band-pass filter as for BiFC images, but with a458 nm excitation. Rpn7 localises to the nucleus throughout the cell cycle in growing cells and relocalises to cytoplasmic structures when cells enter quiescence³⁸. Rpn7-imer2 images were visually inspected before image processing to verify that cells are not quiescent. Rpn7-tDime and auto fluorescence images were used to segment the BiFC images into nuclear and cytoplasmic regions and to unmix the BiFC signal. Image segmentation and single cell fluorescence measurements were performed using custom plugins in ImageJ³⁹ (available upon request). To enable comparison of data from different experiments, the quantification results were rescaled so that BiFC signals of control cells had a mean of zero and a standard deviation of one. Statistical analysis and graphical representation were performed with GraphPad Prism software. Statistically significant differences from control cells were identified by one-way ANOVA followed by Bonferroni post-hoc tests to correct for multiple comparisons.

Recombinant protein expression and purification

E. coli BL21 (DE3) were transformed with plasmids encoding MBP-Hrd1^CT(Hrd1 residues 321-551), MBP-Asi1^RING(Asi1 residues 559-624), MBP-Asi3^RING(Asi3 residues 613-676), GST-Ubc6^Δ™ (Ubc6 residues 1-230), GST-Ubc7or Cue1^U7BR (Cue1 residues 151-203) and were cultivated in LB medium. Cue1^U7BR was co expressed with GST-Ubc7. Protein expression was induced by addition of 1 mM IPTG during 4 h at 25°C. Cells were pelleted, resuspended in PBS, and lysed by sonication. Lysates were rotated with glutathione (GE Healthcare) or amylose beads (New England Biolab) for 1 h at 4°C. Beads were washed with PBS, 1 mM DTT. E2s were cleaved from GST using thrombin (Stago). MBP-E3s were eluted using 10 mM amylose and dialyzed against PBS, 1 mM DTT. All recombinant proteins were concentrated using spin filters (3 kDa, Amicon). Protein purity was tested by Coomassie staining after SDS-PAGE. Protein concentration was estimated by absorbance measurement at 280 nm.

Microscale thermophoresis

Microscale thermophoresis analysis was performed essentially as described¹² using MBP-Asi1^RING, MBP-Asi3^RING or MBP-Hrd1^CT fluorescently labelled with the fluorescent dye NT-647 (labelling was performed with the Monolith™ Protein Labeling Kit RED-NHS according to the instructions of the supplier) and high precision standard treated capillaries. MBP-E3s were diluted to 100 nM in PBS, 5% glycerol, 0.1% Tween 20, 1 mM DTT, 10 μM ZnAc and titrated with varying concentrations of unlabelled E2s before loading into capillaries. The difference of the thermophoretic properties of MBP-E3s were measured using a Monolith NT.115 instrument (NanoTemper Technologies GmbH) and a laser power of 60%. A non-linear fit with the law of mass action was used to derive the dissociation constant(K_d) of the interaction as well as the theoretical thermophoretic properties of the MBP-E3 in its fully bound and unbound states. Those values were then used to normalize the measurements and calculate the fraction of E3 bound at each E2 concentration. Data were plotted and fitted with the GraphPad Prism software.

tFT library construction

4081 verified or uncharacterized S. cerevisiae open reading frames were selected for tagging based on structural and functional criteria (detailed in the Supplementary Methods) to increase the probability that the C-terminal tFT tag would not affect protein functionality, and to avoid exposing the tag to an environment that could impact folding and maturation of the fluorescent proteins. Protocols for strain construction and validation are described in the Supplementary Methods. Briefly, strain manipulations were automated and performed in 96-well format whenever possible. Using conventional PCR targeting³⁰ and lithium acetate transformation of yeast⁴⁰, the module for seamless protein tagging with the mCherry-sfGFP timer (pMaM168 in Supplementary Table 5) was integrated into each selected genomic locus in the strain yMaM330 (Supplementary Table 4), a strain compatible with automated yeast genetics that carried a construct for conditional expression of the I-SceI meganuclease from the GAL1 promoter integrated into the leu2 locus. Correct integration of the tagging module into each locus and expression of tFT protein fusions was verified by PCR and whole colony fluorescence measurements for 4044 open reading frames, with two independent clones validated for 3952 open reading frames (Supplementary Table 2).

tFT library screening

Haploid array strains carrying deletions of individual components of the ubiquitin-proteasome system were obtained from the genome-wide heterozygous diploid yeast deletion library⁴¹ by sporulation and tetrad dissection. Screens were conducted in 1536-colony format. Using pinning robots (BioMatrix, S&P Robotics), tFT query strains (prior to marker excision) were mated with array mutants. Selection of diploids, sporulation and selection of haploids carrying simultaneously a tFT protein fusion and a gene deletion were performed by sequential pinning on appropriate selective media, as described²⁰, followed by seamless marker excision¹⁹. In each screen, a single tFT strain was crossed to a set of mutants in the ubiquitin-proteasome pathway (including the asi1Δ, asi3Δ, hrd1Δ, doa10Δ, ubc6Δ and ubc7Δ mutants) (A.K., B.F., J.D.B., M.M., D.K., C.B., W.H. and M.K., manuscript in preparation) with four technical replicates of each cross. Technical replicates were arranged next to each other. Fluorescence intensities of the final colonies were measured after 24 h of growth on synthetic complete medium lacking histidine at 30°C using Infinite M1000 or Infinite M1000 Pro plate readers equipped with stackers for automated plate loading (Tecan) and custom temperature control chambers. Measurements in mCherry (587/10 nm excitation, 610/10 nm emission, optimal detector gain) and sfGFP (488/10 nm excitation, 510/10 nm emission, optimal detector gain) channels were performed at 400 Hz frequency of the flash lamp, with 10 flashes averaged for each measurement.

Measurements were filtered for potentially failed crosses based on colony size after haploid selection. Fluorescence intensity measurements were log-transformed and the data was normalized for spatial effects on plates by local regression. To estimate the changes from normal protein stability, median effects for tFT and deletion strains were subtracted from log-ratios of mCherry and sfGFP intensities. To avoid variance-mean dependences, standard deviations were regressed against the absolute fluorescence intensities. Changes in protein stability were divided by the regressed standard deviations, yielding a measurement comparable to a z-score, and tested against the hypothesis of zero change. A moderated t-test implemented in the R/Bioconductor package limma was used to compute p-values. p-values were adjusted for multiple testing by controlling the false discovery rate using the method of Benjamini-Hochberg.

Crosses with additional mutants were performed with independently constructed deletion strains using identical procedures on a RoToR pinning robot (Singer). Whole colony fluorescence intensities were corrected for auto fluorescence using measurements of corresponding mutant colonies crossed to strain yMaM344-2 expressing a truncated non-fluorescent mCherry^ΔN protein. For each tFT fusion, mCherry/sfGFP intensity ratios in each mutant were compared to a control cross with a wild type strain carrying the kanMX selection marker in the his3Δ locus.

Fluorescence microscopy

Strains were grown at 30°C in low fluorescence medium (synthetic complete medium prepared with yeast nitrogen base lacking folic acid and riboflavin (CYN6501, ForMedium)) to OD₆₀₀ of 0.4-1.2 and attached to glass-bottom 96-well plates (MGB096-1-2-LG-L, Matrical) using Concanavalin A (C7275, Sigma) as described⁴². Single plane images were acquired on a Delta Vision Elite system (Applied Precision) consisting of an inverted epifluorescence microscope (IX71; Olympus) equipped with an LED light engine (SpectraX, Lumencor), 475/28 and 575/25 excitation, and 525/50 and 624/40 emission filters (Semrock), a dual-band beam splitter 89021 (Chroma Technology), using either a 100× NA 1.4 UPlanSApo or a 60× NA 1.42 Plan ApoN oil immersion objective (Olympus), an sCMOS camera (pco.edge 4.2, PCO) and a motorized stage contained in a temperature-controlled chamber. Image correction and quantification were performed in ImageJ³⁹. Dark signal and flat field corrections were applied to all images as described⁴². Image deconvolution was performed with Soft worx software (Applied Precision) using the conservative ratio algorithm with default parameter settings. Individual cell, perinuclear region and cytoplasm segmentation masks were manually defined in deconvolved images and applied to non-deconvolved images. Mean single cell fluorescence measurements were corrected for cellular auto fluorescence. Mean perinuclear fluorescence measurements were corrected for cytoplasmic fluorescence of each individual cell.

Strains expressing N- and C-terminally tagged Vtc1 and Vtc4 were imaged with exposure setting adjusted to the expression levels: 3.3-fold longer exposure time for C-terminally tagged fusions. Representative deconvolved images were scaled identically.

Cycloheximide chases

Strains were grown at 30°C in synthetic complete medium to OD₆₀₀ of ∼0.8 before addition of cycloheximide to 100 μg/ml final concentration. One ml samples taken at each time point were immediately mixed with 150 μl of 1.85 M NaOH and 10 μl beta-mercaptoethanol, and flash frozen in liquid nitrogen. Whole cell extracts were prepared as previously described⁴⁰, separated by SDS-PAGE followed by semi-dry blotting and probed sequentially with mouse anti-HA (12CA5) and mouse anti-Pgk1 (22C5D8, Molecular Probes) antibodies. A secondary goat anti-mouse antibody (IgG (H+L)-HRP, Dianova) was used for detection on a LAS-4000 system (Fuji).