Protein quality control at the inner nuclear membrane

Abstract

The nuclear envelope is a double membrane that separates the nucleus from the cytoplasm. The inner nuclear membrane (INM) functions in essential nuclear processes including chromatin organization and regulation of gene expression. The outer nuclear membrane is continuous with the endoplasmic reticulum and is the site of membrane protein synthesis. Protein homeostasis in this compartment is ensured by endoplasmic-reticulum-associated protein degradation (ERAD) pathways that in yeast involve the integral membrane E3 ubiquitin ligases Hrd1 and Doa10 operating with the E2 ubiquitin-conjugating enzymes Ubc6 and Ubc7 (refs 2, 3). However, little is known about protein quality control at the INM. Here we describe a protein degradation pathway at the INM in yeast (Saccharomyces cerevisiae) mediated by the Asi complex consisting of the RING domain proteins Asi1 and Asi3 (ref. 4). We report that the Asi complex functions together with the ubiquitin-conjugating enzymes Ubc6 and Ubc7 to degrade soluble and integral membrane proteins. Genetic evidence suggests that the Asi ubiquitin ligase defines a pathway distinct from, but complementary to, ERAD. Using unbiased screening with a novel genome-wide yeast library based on a tandem fluorescent protein timer, we identify more than 50 substrates of the Asi, Hrd1 and Doa10 E3 ubiquitin ligases. We show that the Asi ubiquitin ligase is involved in degradation of mislocalized integral membrane proteins, thus acting to maintain and safeguard the identity of the INM.

Extended Data Figure 1

Identification of Ubc6 and Ubc7 ubiquitin conjugating enzymes as functional interacting partners of Asi1 and Asi3. a, Quantification of BiFC signals in cells expressing VC-Ubc6 and all tested E3 ubiquitin ligases. BiFC signals were measured in the cytoplasm and nucleus of individual cells (n as indicated in the figure). Whiskers extend from 10th to 90th percentiles. The same representation is used in c and d. b, Immunoblot showing expression levels of VC-tagged E2 ubiquitin conjugating enzymes. Ubc11-VC could not be detected in the growth condition of the BiFC assay. c, Quantification of BiFC signals in cells co-expressing VC-tagged E2 ubiquitin conjugating enzymesand Asi1-VN or Asi3-VN (n as indicated in the figure). d, Detection of a significant BiFC signal between Asi1-VN and Ubc4-VCin cells lacking UBC6 (n as indicated in the figure). e, Coomassie-stained gels of recombinant proteins used in microscale thermophoresis experiments. f, mRNA levels of AGP1 and GNP1 measured with qRT-PCR in the indicated strains (mean ± s.d., n = 3 clones). The signal was normalized to wild type (dashed line). g, Ubiquitylation of Stp1-HA or Stp1-RI_17-33-HA (left panel) and Stp2-HA or Stp2Δ_2-13-HA (right panel) in strains expressing 6His-ubiquitin. Stp1-RI_17-33 and Stp2Δ_2-13 variants exhibit compromised cytoplasmic retention and enhanced Asi-dependent degradation, whereas full length Stp1 is degraded primarily in the cytoplasm in SCF^Grr1-dependent manner. Total cell extracts (T), flow through (F) and ubiquitin conjugates (E) eluted after immobilized-metal affinity chromatography were separated by SDS-PAGE followed by immunoblotting with antibodies against the HA tag, Pgk1 and the His tag. Representative immunoblots from 3 technical replicates. Statistics: (a, c, d) One-way ANOVA with Bonferroni correction for multiple testing. *P<10^-4. (f) Two-tailed t-test. *P<0.05 **P<0.10.

Extended Data Figure 2

Lack of genetic interaction between ASI1 and HRD1 or DOA10 at 37°C. 10-fold serial dilutions of strains grown on synthetic complete medium for 2 days at 30 or 37°C.

Extended Data Figure 3

tFT screens for substrates of Asi and ERAD E3 ubiquitin ligases. a, Tagging approach used to construct the tFT libraryin a strain carrying the I-SceI meganuclease under an inducible promoter. First, a module for seamless C-terminal protein tagging with the mCherry-sfGFP timer is integrated into a genomic locus of interest using conventional PCR targeting. Subsequent I-SceI expression leads to excision of the heterologous terminator and the URA3 selection marker, followed by repair of the double strand break by homologous recombination between the mCherry and mCherry^ΔN sequences. A tFT fusion protein is expressed under control of endogenous promoter and terminator in the final strain. b, Workflow of screens for substrates of E3 ubiquitin ligases involved in protein degradation. Each tFT query strain is crossed to an array of mutants carrying different gene deletion alleles. The resulting strains are imaged with a fluorescence plate reader to identify proteins with altered stability in each mutant. c, Volcano plots of the screens for proteins with altered stability in the indicated mutants. Plots show z-scores for changes in protein stability on the x-axis and the negative logarithm of p-values adjusted for multiple testing on the y-axis. The number of proteins with increased (red) or decreased(blue) stability at 1% false discovery rate is indicated. d, Fraction of proteins in the tFT library and in the three clusters in Fig. 3b mapped to the full Yeast Slim set of Component GO terms. Note that the GO term Cytoplasm contains all cellular contents except the nucleus and the plasma membrane. e, The three clusters in Fig. 3b are enriched for proteins in the indicated Component GO terms. Bar plot shows -log₁₀ transformed p-values of significant enrichments.

Extended Data Figure 4

Analysis of integral membrane protein substrates of the Asi E3 ubiquitin ligase. a, Differences in log₁₀mCherry/sfGFP intensity ratio between the indicated mutants and the wild type (mean ± s.d., n = 4) for tFT-tagged proteins from the Asi cluster in Fig. 3b. b, Quantification of BiFC signals in strains co-expressing VC-Ubc6 and Asi3-VN (top panel). BiFC signals were measured in the cytoplasm and nucleus of individual cells (n as indicated in the figure). Whiskers extend from 10th to 90th percentiles. A substantial BiFC signal is retained in the asi2Δ mutant, despite reduced expression of Asi3 (immunoblot, bottom panel). c, Quantification of sfGFP signals in strains expressing tFT-tagged proteins from the Asi cluster in Fig. 3b. Fluorescence microscopy examples representative of five fields of view (top panel). Scale bar, 5 μm. sfGFP intensities were measured in individual cells (middle panel) and at the nuclear rim (bottom panel); a.u., arbitrary units. For each protein, measurements were normalized to the mean of the respective wild type. Whiskers extend from minimum to maximum values.

Extended Data Figure 5

Cycloheximide chase experiments with substrates of the Asi E3 ubiquitin ligase. Degradation of 3HA-tagged proteins after blocking translation with cycloheximide. Whole cell extracts were separated by SDS-PAGE followed by immunoblotting with antibodies against the HA tag and Pgk1 as loading control. Representative immunoblots from 2 technical replicates. Left panel wild type and asi1Δ immunoblots are reproduced in Fig. 3f.

Extended Data Figure 6

Influence of tagging and expression levels on localization of Vtc1 and Vtc4. Fluorescence microscopy of strains expressing Vtc1 or Vtc4 tagged endogenously with myeGFP at the C-terminus or tagged with sfGFP at the N-terminus and expressed under control of endogenous or TEF1 promoters. Representative de convolved images of five fields of view with ∼100 cells each. Arrowheads indicate nuclear rim localisation. Scale bar, 5 μm. Statistics: (a, c) Two-tailed t-test. *P<0.05. (b) One-way ANOVA with Bonferroni correction for multiple testing. *P<10^-4.

Figure 1

The Asi complex is a Ubc6/Ubc7-dependent E3 ubiquitin ligase of the INM. a, BiFC strategy used to assay E2-E3 interactions. E2 and E3 proteins were endogenously tagged with C- and N-terminal fragments of the Venus fluorescent protein (VC and VN). Interactions between E2 and E3 proteins enable reconstitution of functional Venus that is detected with fluorescence microscopy. Rpn7-tDimer2 served as red nuclear marker. b, Quantification of BiFC signals in cells co-expressing VC-Ubc6 and VN-tagged E3s. Fluorescence microscopy examples representative of six fields of view (top panel). Scale bar, 5 μm. BiFC signals were measured in the cytoplasm and nucleus of individual cells (bottom panel, n as indicated in the figure). Whiskers extend from 10th to 90th percentiles. c, Microscale thermophoresis analysis of interactions between recombinant maltose binding protein (MBP)-E3 fragments and the indicated E2s. Plots show the fraction of MBP-E3 bound to the E2 at each tested E2 concentration (mean ± s.d., n as indicated in the figure). Dissociation constants (K_d, mean ± s.d.) were derived from non-linear fits with the law of mass action (solid lines). d, Activity of β-galactosidase expressed from the AGP1 promoter in the indicated strains (mean ± s.d., n = 3clones); a.u., arbitrary units. e, Ubiquitylation of Stp2^N-TAP in strains expressing 10His-ubiquitin. Total cell extracts and ubiquitin conjugates eluted after immobilized-metal affinity chromatography were separated by SDS-PAGE followed by immune blotting with antibodies against the TAP tag, Pgk1 and ubiquitin. Representative immune blots from 3 technical replicates. Statistics: (b) One-way ANOVA with Bonferroni correction for multiple testing. *P<10^-4. (d) Two-tailed t-test. *P<0.05.

Figure 2

Functional overlap between Asi and ERAD E3 ubiquitin ligases. a, Histograms of Pearson correlation coefficients calculated between the genetic interaction profiles of each ASI gene and ∼75% of all yeast genes, obtained from a previously published genome-scale genetic interaction map. Asterisks mark the YMR119W-A dubious open reading frame, which overlaps with the ASI1 gene. b, 10-fold serial dilutions of strains grown on synthetic complete medium for 2 days at 30 or 37°C.

Figure 3

Systematic identification of substrates for Asi and ERAD E3 ubiquitin ligases. a, A tandem fluorescent protein timer (tFT) is composed of two fluorescent proteins: one slower maturing (e.g. the red fluorescent protein mCherry, maturation rate constant m_S) and the other faster maturing (e.g. the green fluorescent protein sfGFP, maturation rate constant m_F). When fused to a protein of interest, a tFT reports on the degradation kinetics of the fusion protein: whereas fusions undergoing fast turnover are degraded prior to mCherry maturation, resulting in a low mCherry/sfGFP intensity ratio, the relative fraction of mature mCherry increases for proteins with slower turnover. b, Summary heat map of the screens for tFT-tagged proteins with altered stability in the indicated mutants. Changes in protein stability (z-score) are color-coded from blue (decrease) to red (increase). Only proteins with a significant change in stability in at least one mutant (1% false discovery rate and z-score>4) are displayed. Clusters of potential substrates of Asi (green), Hrd1 (red) and Doa10 (blue) E3 ubiquitin ligases are indicated. c, d, Fraction of proteins in the tFT library and in the three clusters in b with a predicted transmembrane domain (TMD) or signal peptide (c) or mapped to Component GO terms (d). Each cluster is significantly enriched in proteins with a predicted TMD or signal peptide compared to the tFT library (P<0.0034, Fisher's exact test). e, Quantification of sfGFP signals in strains expressing tFT-tagged proteins from the Asi cluster in b. Fluorescence microscopy examples representative of five fields of view (top panel). Scale bar, 5 μm. sfGFPintensities were measured in individual cells and at the nuclear rim (bottom panel, nas indicated in the figure). For each protein, measurements were normalized to the mean of the respective wild type. Whiskers extend from minimum to maximum values. Statistics: Two-tailed t-test. *P<0.05. f, Degradation of 3HA-tagged proteins after blocking translation with cycloheximide. Whole cell extracts were separated by SDS-PAGE followed by immunoblotting with antibodies against the HA tag and Pgk1 as loading control. Representative immunoblots from 3 technical replicates.