Proteasome subunit Rpn13 is a novel ubiquitin receptor

Abstract

Proteasomal receptors that recognize ubiquitin chains attached to substrates are key mediators of selective protein degradation in eukaryotes. Here we report the identification of a new ubiquitin receptor, Rpn13/ARM1, a known component of the proteasome. Rpn13 binds ubiquitin through a conserved amino-terminal region termed the pleckstrin-like receptor for ubiquitin (Pru) domain, which binds K48-linked diubiquitin with an affinity of approximately 90 nM. Like proteasomal ubiquitin receptor Rpn10/S5a, Rpn13 also binds ubiquitin-like (UBL) domains of UBL-ubiquitin-associated (UBA) proteins. In yeast, a synthetic phenotype results when specific mutations of the ubiquitin binding sites of Rpn10 and Rpn13 are combined, indicating functional linkage between these ubiquitin receptors. Because Rpn13 is also the proteasomal receptor for Uch37, a deubiquitinating enzyme, our findings suggest a coupling of chain recognition and disassembly at the proteasome.

Figure 1. Murine Rpn13 binds ubiquitin chains

a, mRpn13 cDNA fragments were cloned into pYTH9 vector in frame with the Gal4 DNA-binding domain, and the resulting bait vectors transformed into yeast strain Y190 with prey pACT2 vectors containing wt-ubiquitin, I44A-ubiquitin, or hRpn2 (positive binding control) cDNA in frame with Gal4 DNA-activating domain. b, Architecture of Rpn13 from various species. The N-terminal domain is generally conserved (black box) whereas the C-terminal region (grey box) is absent in S. cerevisiae and has diverged beyond recognition in one of the two S. pombe proteins (S. pombe (1)). S. pombe (1)=SPCC16A11.16, S. pombe (2)=SPBC342.04. Percent identity to the conserved hRpn13 Pru domain is provided at right. c, Alignment of Rpn13 N-terminal sequences. Residues that are invariant or conserved in at least 50% of sequences are shaded in black or grey, respectively. d, To identify the minimal region required for ubiquitin binding, mRpn13 deletion mutants were expressed as GST-fused proteins, purified, and tested for their binding to linear tetraubiquitin by immunoblotting with anti-ubiquitin antibodies. Tetraubiquitin was obtained by thrombin cleavage of GST-fused tetraubiquitin (GST 4×Ub) and equivalent amounts of GST-fused deletion mutants were used in GST pull-down assay.

Figure 2. Rpn13 contributes to recognition of ubiquitin conjugates by the proteasome

a, rpn13Δ proteasomes show defects in ubiquitin conjugate binding. Proteasomes were purified from strains (SY733, SY729, SY725, and SY722) bearing the indicated mutations. Proteasomes (4 pmol) were mixed with autoubiquitinated Cdc34 (16 pmol), resolved by native PAGE, and visualized using LLVY-AMC. Note that UBL/UBA proteins cannot contribute to ubiquitin chain binding in these experiments, since all proteasomes used in this figure are from a rad23Δ dsk2Δ ddi1Δ genetic background. ubp6Δ is also in the genetic background, to prevent chain disassembly during the assay. b, Proteasome composition is maintained in the absence of Rpn13. Proteasomes from panel A (25 μg) were resolved by SDS-PAGE and stained with Coomassie blue. An asterisk marks contaminating protein. c, Reconstitution of ubiquitin conjugate binding. A subset of proteasomes from panel A (4 pmol each) was incubated with scRpn13 (20 pmol) cleaved from the GST moiety (+ lanes) or GST alone (remaining lanes) to allow reassembly, then mixed with autoubiquitinated Cdc34 (16 pmol). Complexes were resolved by native PAGE and visualized as in (A).

Figure 3. Rpn13 uses loops to bind ubiquitin

a, Stereoview of scRpn13, spanning residues T6-L101, in which β-strands are indicated in blue and hydrophobic sidechains in yellow. b, NMR titration experiments reveal scRpn13 residues that contact ubiquitin. The data were prepared as described in Methods and plotted. c, ScRpn13 residues that bind ubiquitin are within the S2-S3, S4-S5, and S6-S7 loops. Residues significantly affected by ubiquitin addition are displayed and labeled in red with their secondary structures in blue.

Figure 4. Rpn13 binds to ubiquitin and UBLs of proteasomal receptors

a, b, hRpn13 Pru binds K48-linked diubiquitin and monoubiquitin with 1:1 stoichiometry whereas two hRpn13 Pru molecules bind one K48-linked tetraubiquitin. Normalized chemical shift perturbation values are plotted against varying molar ratios of Rpn13 to tetraubiquitin (a, shades of blue), diubiquitin (b, shades of red), or monoubiquitin (a, shades of green) to reveal respective Rpn13:ubiquitin binding stoichiometries of 2:1, 1:1 or 1:1. Each data line represents a specific amino acid as indicated in the figure, with values determined as described in the Supplement. c, Binding curves for hRpn13 Pru binding to monoubiquitin or K48-linked diubiquitin as determined by intrinsic tryptophan fluorescence. Normalized fluorescence intensity values are plotted for two data sets against varying concentration of monoubiquitin (red and orange) or diubiquitin (blue and light blue). The data were fit by assuming 1:1 binding for hRpn13 Pru and monoubiquitin (orange) or diubiquitin (light blue). d, Table of hRpn13 Pru binding affinities for K48-linked diubiquitin, monoubiquitin, and hHR23a's UBL domain. Values for ubiquitin binding were determined by using the fluorescence data of (c). NMR titration curves were used to determine the value for hHR23a's UBL domain. e, mRpn13 Pru binds to the hydrophobic patch of ubiquitin containing I44. mRpn13 Pru domain was used in GST pull-down assays to assess binding to GST-tagged monoubiquitin and its mutant derivatives (I44A and triple mutant [3M*] L8-I44-V70). f, mRpn13 Pru domain was used in GST pulldown assays (as in Figure 4e) to assess its binding to GST-fused ubiquitin-like protein modifiers. g, Rpn13 binds to the hPLIC2 and hHR23A UBL domains. ¹H, ¹⁵N HSQC spectra of ¹⁵N labeled hRpn13 Pru alone (black) and with 2-fold molar excess hPLIC2 (red) or hHR23A (blue) UBL domain indicates their direct interaction. Although the effect is greater for hPLIC2, these two UBL domains affect common residues in hRpn13 Pru, suggesting that they bind the same surface.

Figure 5. An Rpn13 mutant specifically defective in ubiquitin chain binding

a, Reconstitution of proteasomes with recombinant GST-Rpn13. Proteasomes were purified from strains containing or lacking Rpn13 (SY775 and SY723). GST-Rpn13 (40 pmol) or buffer was mixed with proteasomes (5 pmol), which were resolved on native PAGE and visualized using suc-LLVY-AMC. The mobility shift caused by GST-Rpn13 is an indicator of its assembly into proteasomes. The presence of GST on Rpn13 is required to cause a mobility shift. All proteasomes used in this figure are from an rpn10-uim ubp6Δ genetic background. b, Mutations in Rpn13 do not attenuate assembly of Rpn13 into the proteasome. Reconstitution assays were carried out as in panel A, but using a four-fold molar excess of GST-Rpn13. c, Structural model with mutations. Mutated residues (see panels D-F) were mapped onto a model structure of scRpn13 (dark grey) complexed with monoubiquitin (light grey). E41, E42 and S93 are displayed in red, ubiquitin-binding loops in blue. These residues map to the S2-S3 (E41K, E42K) and S6-S7 (S93D) loops. d, Mutations in single loops of Rpn13 attenuate its proteasomal ubiquitin receptor function. 12 pmol of Rpn13 variants cleaved from GST were incubated with proteasomes (3 pmol) to allow reassembly. Autoubiquitinated Cdc34 (18 pmol) was then added. After 15 min at 30°C, complexes were resolved by native PAGE and visualized using suc-LLVY-AMC. e, Rpn13 mutant E41K, E42K, S93D (Rpn13-KKD) abrogates the ubiquitin receptor activity of Rpn13. Experiment performed as in panel D. f, scRpn13-KKD affinity for monoubiquitin is significantly reduced compared to wild-type. Normalized chemical shift perturbation values are plotted against molar ratios of monoubiquitin to wild-type scRpn13 (WT, red, green) or monoubiquitin to scRpn13-KKD (KKD, blue, purple). Each data line represents a specific amino acid, namely F45 (red and blue) and E72 (green and purple). Using Matlab v. 7.2, the data were fit to determine a binding constant of 65 μM for wild-type scRpn13 and an 8-fold reduction in scRpn13-KKD's affinity for monoubiquitin. g, Superposition of ¹H, ¹⁵N HSQC spectra of wild-type Rpn13 (black) and Rpn13-KKD (red). Shifted resonances are labeled in grey and those corresponding to E41, E42 and S93 in black. Chemical shift assignments are only available for the wild-type protein.

Figure 6. Phenotypic effects of the loss of ubiquitin receptor function by Rpn13

a, Canavanine sensitivity of single and double mutants in ubiquitin receptor genes. Cells in late log phase (top: SY998b, SY980f, SY1004c, and SY920b; middle: SY1076, SY1073a, SY1012a, and SY1080a; and bottom: SY1076, SY1074a, SY1012a, and SY1082a) were serially diluted and stamped on plates using a pin array. Plates were incubated at 30°C for either 2 (left) or 3 (right) days. b, Endogenous ubiquitin conjugate levels in proteasomal ubiquitin receptor mutants. Cells (SY998a, SY980a, SY1004a, and SY920a) were grown to log phase, and whole-cell extracts prepared. Proteins were resolved by 4-12% gradient SDS-PAGE, transferred to PVDF, and probed with antibody against ubiquitin. The membrane was stripped and probed with antibody against eIF5a. c, Substrate stabilization in rpn13-KKD mutants. Cells (SY992b, SY1004b) expressing Ub^V76-Val-e^ΔK-ßgal from a GAL promoter were grown to mid-log phase under inducing conditions. Protein synthesis was quenched at time zero by adding cycloheximide. Aliquots were withdrawn at the time points indicated, and lysates prepared. Proteins were visualized by SDS-PAGE/immunoblot analysis, using an antibody to ß-galactosidase, and quantified with imaging software (Kodak EDAS 290). The rate of degradation of Ub^V76-Val-e^ΔK-ßgal was reduced approximately 2 fold in the rpn13-KKD mutant as compared with wild type. Asterisks indicate distinct ß-galactosidase-derived partial breakdown products, whose relative intensities differ between wild-type and mutant. d, rpn13-KKD mutants are not deficient in proteasome levels. Cells (SY933, SY936, SY950 and SY952) were grown to late log phase at 30°C in YPD, harvested, and lysed as described (see Supplementary Information). 150 μg of extract were resolved by native PAGE, and proteasome complexes visualized using suc-LLVY-AMC (left) and Coomassie as a loading control (middle). The asterisk indicates the expected position of the proteasome CP, which is not visualized due to low levels. Extracts were also subject to a quantitative proteasome assay, using suc-LLVY-AMC (right). e, Proteasomes from rpn13-KKD mutants are loaded with Rpn13-KKD protein. 100 μg of extract prepared for panel B was incubated with 1 μg of either GST-Rpn13 (+ lanes) or GST only (samples where GST-Rpn13 is absent) on ice. Proteasome complexes were resolved by native PAGE and visualized by suc-LLVY-AMC overlay assay.