Reprograming of the ubiquitin ligase Ubr1 by intrinsically disordered Roq1 through cooperating multifunctional motifs

Abstract

One way cells control the speed and specificity of protein degradation is by regulating the activity of ubiquitin ligases. Upon proteotoxic stress in yeast, the intrinsically disordered protein Roq1 binds the ubiquitin ligase Ubr1 as a pseudosubstrate, thereby modulating the degradation of substrates of the N-degron pathway and promoting the elimination of misfolded proteins. The mechanism underlying this reprograming of Ubr1 is unknown. Here, we show that Roq1 controls Ubr1 by means of two cooperating multifunctional motifs. The N-terminal arginine and a short hydrophobic motif of Roq1 interact with Ubr1 as part of a heterobivalent binding mechanism. Via its N-terminal arginine, Roq1 regulates the ubiquitination of various N-degron substrates and folded proteins. Via its hydrophobic motif, Roq1 accelerates the ubiquitination of misfolded proteins. These findings reveal how a small, intrinsically disordered protein with a simple architecture engages parallel channels of communication to reprogram a functionally complex ubiquitin ligase.

Keywords: Protein Degradation; SHRED; Ubiquitin Ligase Regulation.

Figure 1. Roq1 reprograms Ubr1 substrate specificity in vitro.

(A) Model of stress-induced homeostatically regulated protein degradation (SHRED). Proteotoxic stress in yeast induces the synthesis of Roq1(1–104), which is cleaved by the protease Ynm3 to generate Roq1(22-104) with arginine-22 (R22) at its new N-terminus. Through R22, Roq1(22-104) binds to the Ubr1 type-1 site as a pseudosubstrate and enhances the degradation of misfolded proteins to restore cell homeostasis. In addition, binding of Roq1(22-104) to the Ubr1 type-1 site competitively inhibits recognition of type-1 N-degron substrates and allosterically promotes recognition of type-2 N-degron substrates. Numbers in unbound Ubr1 (blue) denote the known and putative substrate binding sites for type-1 N-degron substrates (1), type-2 N-degron substrates (2), folded proteins with internal degrons (3) and misfolded proteins (4). Ubⁿ = ubiquitin chain. (B, C) Western blot of Pho8 from Pho8* or Pho8 ubiquitination assays in the absence or presence of Roq1(22-60) for the times indicated. (D, E) Western blot of Luciferase from Luciferase^U or Luciferase^N ubiquitination assays in the absence or presence of Roq1(22-60) for the times indicated. (F, G) Western blot of GFP from R-GFP or F-GFP ubiquitination assays in the absence or presence of Roq1(22-60) for the times indicated. Source data are available online for this figure.

Figure 2. Roq1 binds and regulates Ubr1 via R22 and at least one other determinant.

(A) Western blot of Pho8 from Pho8* ubiquitination assays with Roq1(22-60) or Roq1(22-60)(R22A). Roq1:Ubr1 molar ratios were 10:1. t = 90 min. (B) Western blot of GFP from F-GFP ubiquitination assays without and with Roq1(22-60) or the RA dipeptide. Molar ratios were Roq1:Ubr1 = 10:1 and RA:Ubr1 = 4000:1. t = 90 min. (C) As in panel B but western blot of Pho8 from Pho8* ubiquitination assays. (D) Biolayer interferometry of Roq1-Ubr1 complex formation with immobilized Roq1(22-104) and soluble Ubr1. Data are the mean of three independent experiments. (E) Western blot of Pho8 from Pho8* ubiquitination assays with Roq1(22-60) or Roq1(22-60)(R22A). Numbers in the Roq1 rows indicate the Roq1:Ubr1 molar ratios used. t = 90 min. Source data are available online for this figure.

Figure 3. Roq1 contains a functionally essential hydrophobic motif.

(A) Workflow of the Roq1 mutagenesis screen. Roq1 variants mutagenized by error-prone PCR were introduced into a reporter strain lacking endogenous Roq1. Cells were allowed to form colonies on assay plates, SHRED-deficient colonies were picked and their defective Roq1 variants were sequenced. (B) Cellular levels of the SHRED reporter Rtn1Pho8*-GFP after tunicamycin treatment for 5 h relative to levels in untreated cells, as measured by flow cytometry. The roq1 mutant cells contained an empty plasmid (no Roq1), a plasmid encoding wild-type Roq1 (WT Roq1) or plasmids encoding variants of ubiquitin-Roq1(22-104) fusions. Ubiquitin-Roq1 fusions are processed by cells to yield Roq1(22-104). WT, wild-type. Bars are the mean ± s.e.m.; n = 3 biological replicates. (C) Schematic of the Roq1(22-104) sequence showing the point mutations identified in the Roq1 mutagenesis screen. (D, E) As in (B). 4A = Y55A,Y56A,F57A,V58A. (F) Western blot of HA tag from cells expressing Ub-Roq1(22-104)-HA(74) or mutant versions containing a Y55D or a 4 A mutation. Cells were treated with cycloheximide to block protein synthesis for the time indicated. CHX cycloheximide. (G) Quantification of western blots as shown in (F). For each Roq1 variant, data are normalized to t = 0 min. Data are the mean of three technical replicates. Source data are available online for this figure.

Figure 4. Roq1 interacts with Ubr1 through a heterobivalent binding mechanism.

(A) Western blot of FLAG and HA tags from cell lysates (input) or anti-HA immunoprecipitates of roq1 knockout cells expressing FLAG-Ubr1 and HA-tagged Roq1 variants as indicated (Roq1 immunoprecipitate). All Roq1 variants were expressed as ubiquitin-Roq1(22-104)-HA(74) fusions, which are processed by cells to yield Roq1(22-104)-HA(74) with an internal HA tag. Note that Roq1(22-104)-HA(74) expressed under the endogenous ROQ1 promoter is undetectable due to low expression levels. Roq1(22-104) naturally runs as a double band. The nature of the slower migrating band is unknown, as is the reason for the strong impact of the F57D mutation on Roq1 electrophoretic mobility. P_ROQ1 = ROQ1 promoter, P_GPD = GPD promoter, WT = wild-type. (B) As in (A). All Roq1 variants were expressed under the strong GPD promoter. (C) Biolayer interferometry of Roq1-Ubr1 complex formation with immobilized Roq1(22-104) and soluble Ubr1. Data are the mean of two independent experiments. 4A = Y55A,Y56A,F57A,V58A. Source data are available online for this figure.

Figure 5. The Roq1 hydrophobic motif selectively promotes the ubiquitination of misfolded proteins.

(A) Biolayer interferometry of Roq1-Ubr1 complex formation with immobilized Roq1(22-104) and soluble Ubr1. Data are the mean of two independent experiments. The data for wild-type Roq1 are the same as in Fig. 4C. (B) Western blot of Pho8 from Pho8* ubiquitination assays with Roq1(22-60) or Roq1(22-60)(V58E). Numbers in the Roq1 rows indicate the Roq1:Ubr1 molar ratios used. t = 90 min. (C) As in (B) but western blot of Luciferase from Luciferase^U assays. t = 30 min. (D) As in (B) but western blot of GFP from R-GFP assays. t = 15 min. (E) As in panel B but western blot of GFP from F-GFP assays. t = 15 min. (F) As in (B) but western blot of Strep tag from Cup9-Strep assays. t = 15 min. Source data are available online for this figure.

Figure 6. Roq1 promotes ubiquitin chain initiation on Ubr1 substrate proteins.

(A) Western blot of Pho8 from Pho8* ubiquitination assays with lysine-free ubiquitin (no-K-ubiquitin) and Roq1(22-60), Roq1(22-60)(V58E), or RA dipeptide. Numbers to the left of the gel indicate distinct conjugates of no-K-ubiquitin and Pho8*. (B) As in (A) but western blot of GFP from F-GFP assays. (C) As in (A) but western blot of Strep tag from Cup9-Strep assays. Source data are available online for this figure.

Figure 7. Delineation of distinct activities of the Roq1 hydrophobic motif in vivo.

(A) Cellular levels of the SHRED reporter Rtn1Pho8*-GFP after tunicamycin treatment for 5 h relative to the levels in untreated cells, as measured by flow cytometry. The indicated point mutations were introduced into the endogenous ROQ1 gene. 4A = Y55A,Y56A,F57A,V58A, WT wild-type. Bars are the mean ± s.e.m.; n = 3 biological replicates. (B) Flow cytometry of mCherry/sfGFP fluorescence of untreated and estradiol-treated cells expressing the ubiquitin-R-mCherry-sfGFP type-1 N-degron substrate and Roq1 variants under an estradiol-inducible promoter system. Plotted on a log2 scale is the fold change of mCherry/sfGFP fluorescence upon Roq1 expression by estradiol treatment for 6 h. An increased mCherry/sfGFP ratio indicates stabilization of ubiquitin-R-mCherry-sfGFP upon Roq1 expression. Bars are the mean ± s.e.m.; n = 3 biological replicates. (C) As in panel B but with the ubiquitin-F-mCherry-sfGFP type-2 N-degron substrate. A decreased mCherry/sfGFP ratio indicates destabilization of Ubiquitin-F-mCherry-sfGFP upon Roq1 expression. Bars are the mean ± s.e.m.; n = 3 biological replicates. Source data are available online for this figure.

Figure 8. Functional Roq1 minimally consists of R22, a short linker, and the hydrophobic motif.

(A) Schematic of Roq1 linker variants for the shortening and replacement of the sequence between R22 and the hydrophobic motif in Roq1(22-104). (B) Cellular levels of the SHRED reporter Rtn1Pho8*-GFP after tunicamycin treatment for 5 h relative to levels in untreated cells, as measured by flow cytometry. The roq1 mutant cells contained an empty plasmid (no Roq1), a plasmid encoding wild-type Roq1 (WT Roq1) or plasmids encoding variants of ubiquitin-Roq1 fusions in which the linker between R22 and the hydrophobic motif was replaced, shortened, or both. Ubiquitin-Roq1 fusions are processed by cells to yield Roq1 starting with R22. Bars are the mean ± s.e.m.; n = 3 biological replicates. (C) Schematic of Roq1 constructs for the replacement of the linker between R22 and the hydrophobic motif in shortened Roq1(22aa). (D) Western blot of Pho8 from Pho8* ubiquitination assays with Roq1 variants with shortened or artificial linkers between R22 and the hydrophobic motif. (E) Model of Ubr1 reprograming by Roq1. Cleaved Roq1(22-104) interacts with Ubr1 through a heterobivalent binding mechanism that creates avidity. Binding involves physical contact of R22 with the Ubr1 type-1 site and of the YYFV hydrophobic motif with an undefined region in Ubr1. Occupancy of the Ubr1 type-1 site activates the type-2 substrate binding site and a binding site for folded substrates with internal degrons, such as Cup9. In parallel, the hydrophobic motif activates a putative fourth substrate binding site to enhance the degradation of misfolded proteins. Numbers in unbound Ubr1 (blue) denote the known and putative substrate binding sites for type-1 N-degron substrates (1), type-2 N-degron substrates (2), folded proteins with internal degrons (3) and misfolded proteins (4). The box above Roq1(22-104) highlights the minimally required functional elements of Roq1. Ubⁿ = ubiquitin chain. Source data are available online for this figure.

Figure EV1. In vitro reconstitution of Ubr1 regulation by Roq1.

(A) Western blot of HA tag from ubiquitination assays with Roq1(22-104)-HA and Roq1(22-60)-HA in the absence of a designated Ubr1 substrate. Ubiquitination reactions were stopped after 30 min with buffer containing dithiothreitol and, where indicated, treated with NaOH to hydrolyze ester bonds between Roq1 and ubiquitin. Ubr1 catalyzes the formation of NaOH-sensitive conjugates of Roq1(22-104) and ubiquitin, which is almost completely lost by shortening Roq1 to Roq1(22-60). Note that the bands of unmodified Roq1(22-60) are weaker than those of unmodified Roq1(22-104) even though equal molar amounts of the two proteins were used for the assays. The reason for this difference is that the small Roq1(22-60) is only weakly bound by the nitrocellulose membrane used for western blotting so that its amounts are lower than they should be. Less ubiquitinated Roq1 is generated from Roq1(22-60) than from Roq1(22-104). WT, wild-type. (B) Cellular levels of the SHRED reporter Rtn1Pho8*-GFP after tunicamycin treatment for 5 h relative to levels in untreated cells, as measured by flow cytometry. The roq1 mutant cells contained an empty plasmid (no Roq1), a plasmid encoding wild-type Roq1 (WT Roq1), ubiquitin-Roq1(22-104) or ubiquitin-Roq1(22-60). The ubiquitin fusions are processed by cells to yield Roq1(22-104) or Roq1(22-60) starting with R22. Bars are the mean ± s.e.m.; n = 3 biological replicates. (C) Western blot of Pho8 from solubility assays of Pho8*. Ubiquitination assays including Roq1(22-60) were carried out for 0 or 90 min and soluble and insoluble Pho8* were separated by centrifugation. T = total; S = supernatant; P = pellet. (D) Western blot of Pho8 from Pho8* ubiquitination assays with and without Ubr1 and Roq1(22-60). No Pho8* ubiquitination occurs in the absence of Ubr1. (E). Western blot of Pho8 from Pho8* ubiquitination assays with and without Roq1(22-60). Ubiquitination reactions were stopped after 90 min and, where indicated, treated with NaOH to hydrolyze ester bonds between Pho8* and ubiquitin. Ubiquitin-Pho8* conjugates were resistant to alkaline hydrolysis, showing that they consisted of amide rather than oxyester bonds. (F) Western blot of Pho8 from Pho8* ubiquitination assays with different concentrations of Roq1(22-60). Roq1(22-60) was omitted or used at molar ratios of 1:1, 2:1, 5:1 and 10:1 relative to Ubr1. (G) Western blot of ALFA tag from untreated and tunicamycin-treated control cells (strain SSY122) that do not express an ALFA-tagged protein or cells with chromosomally ALFA-tagged Ubr1 and Roq1 (strain SSY4598). The levels of Ubr1 and Roq1 are similar in cells exposed to tunicamycin-induced proteotoxic stress. Note that proteolytic cleavage of ALFA-tagged Roq1 is inefficient, likely because of the position of the tag at the extreme C-terminus (Szoradi et al, 2018). Asterisks mark non-specific bands. n.t., no treatment; Tm, tunicamycin. (H) Western blot of Pho8 from Pho8* ubiquitination assays without and with Roq1(22-104) or Roq1(22-60) for the times indicated.Source data are available online for this figure.

Figure EV2. Roq1 stimulates ubiquitination of folded Ubr1 substrate proteins with internal degrons.

(A) Western blot of Strep tag from Cup9-Strep ubiquitination assays without and with Roq1(22-60) for the times indicated. (B) Western blot of maltose-binding protein (MBP) tag from Mgt1-MBP ubiquitination assays without and with Roq1(22-60) for the times indicated. (C) Western blot of FLAG tag from Chk1-ALFA-FLAG ubiquitination assays without and with Roq1(22-60) for the times indicated. The asterisk denotes truncated Chk1 that arose during the expression and purification of Chk1-ALFA-FLAG.Source data are available online for this figure.

Figure EV3. Overexpressed Roq1(R22A) does not activate Ubr1.

Cellular levels of the SHRED reporter Rtn1Pho8*-GFP after tunicamycin treatment for 5 h relative to levels in untreated cells, as measured by flow cytometry. Cells were wild-type or lacked chromosomal ROQ1. The roq1 mutants contained plasmids encoding Roq1 variants that were expressed under the endogenous ROQ1 promoter (P_ROQ1) or the strong GPD promoter (P_GPD), were full-length or lacked the first 21 residues (22-104), and were otherwise wild-type or contained the R22A mutation. Roq1 with the R22A mutation did not support SHRED, even when overexpressed. Full-length Roq1 and ubiquitin-fused Roq1(22-104) are processed in cells by Ynm3 or ubiquitin proteases to yield Roq1(22-104). Ubiquitin-fused Roq1(22-104) was used when Roq1 was expressed under the GPD promoter to avoid that cleavage of Roq1 by Ynm3 becomes limiting for the amounts of Roq1(22-104) available for Ubr1 activation. Bars are the mean ± s.e.m.; n = 3 biological replicates.Source data are available online for this figure.

Figure EV4. Roq1 contains a functionally essential hydrophobic motif.

(A) Roq1 structure prediction by AlphaFold. (B) Roq1 disorder prediction. The plot shows the average disorder tendency across the Roq1(1–104) sequence on a scale of 0 to 1. Data are the mean of the predictions of twelve different algorithms for disorder prediction. Error bars show the standard deviation. (C) Multiple sequence alignment of Roq1 homologs from fourteen yeast species. The numbering of the residues corresponds to the S. cerevisiae sequence. * fully conserved residue, : strongly conserved residue. weakly conserved residue. (D) Western blot of HA tag and Pgk1 from roq1 mutant strains expressing ubiquitin-Roq1(22-104)-HA(74) variants under the control of the strong GPD promoter. Roq1(22-104) naturally runs as a double band. The nature of the slower migrating band is unknown. Pgk1 served as a loading control. WT, wild-type. (E, F) Cellular levels of the SHRED reporter Rtn1Pho8*-GFP after tunicamycin treatment for 5 h relative to the levels in untreated cells, as measured by flow cytometry. The roq1 mutant cells contained an empty plasmid (no Roq1), a plasmid encoding wild-type Roq1 (WT Roq1) or plasmids encoding variants of ubiquitin-Roq1(22-104). Ubiquitin-Roq1 fusions are processed by cells to yield Roq1(22-104). Bars are the mean ± s.e.m.; n = 3 biological replicates. Source data are available online for this figure.

Figure EV5. Effect of Roq1 on Pho8* recognition and Rad6 recruitment, effect of separating R22 and the hydrophobic motif different molecules, and expression levels of Roq1 linker variants.

(A) Western blots of FLAG tag and Pho8 from input and Pho8 precipitate of an in vitro pulldown assays with FLAG-Ubr1, Pho8/Pho8* fused to maltose-binding protein (MBP), and Roq1(22-60)-HA as indicated. Pho8 or Pho8*-MBP were precipitated with amylose resin. The asterisk denotes truncated Ubr1 that arose during the expression and purification of FLAG-Ubr1. (B) Biolayer interferometry of Rad6-Ubr1 complex formation with immobilized Rad6 and soluble Ubr1. Complex formation was tested in the absence of Roq1(22-104) and with Roq1(22-104):Ubr1 molar ratios of 2:1 or 10:1. Data are the mean of two independent experiments. (C) Western blot of GFP from R-GFP ubiquitination assays with Roq1(22-60), Roq1(22-60)(R22A), Roq1(22-60)(4A) or an RA dipeptide. Wild-type Roq1, the RA dipeptide and Roq1(4A) are able to bind the Ubr1 type-1 site and suppress ubiquitination of R-GFP, whereas Roq1(R22A) is not. 4A = Y55A,Y56A,F57A,V58A. (D) Western blot of Pho8 from Pho8* ubiquitination assays with combinations of Roq1(22-60), Roq1(22-60)(R22A), Roq1(22-60)(4A) and an RA dipeptide. Combining a type-1 site binder, Roq1(4A) or the RA dipeptide, with the hydrophobic motif-containing Roq1(R22A) does not reconstitute Roq1 activity. 4A = Y55A,Y56A,F57A,V58A. (E) Western blot of HA tag and Pgk1 from cells expressing Roq1(22-104)-HA(74) or Roq1(83aa_GSP)-HA(74), in which the sequence between S23 and P51 of Roq1 was replaced with a GSP-based linker. The arrow indicates unprocessed ubiquitin-Roq1, in which the N-terminal ubiquitin has not been removed by ubiquitin proteases. The asterisk marks truncated Roq1. (F) Western blot of HA tag and Pgk1 from cells expressing Roq1(22-104)-HA(74) or Roq1 variants in which the sequence between S23 and P51 was successively shortened. The arrow indicates unprocessed ubiquitin-Roq1, in which the N-terminal ubiquitin has not been removed by ubiquitin proteases. The asterisks marks truncated Roq1. (G) Western blot of HA tag and Pgk1 from cells expressing Roq1(22-104)-HA(74), Roq1(63aa)-HA(74) with a shortened sequence between S23 and P51, Roq1(63aa_GS)-HA(74) with a GS-based linker replacing the shortened sequence or Roq1(63aa_GSP)-HA(74) with a GSP-based linker replacing the shortened sequence. The arrow indicates unprocessed ubiquitin-Roq1, in which the N-terminal ubiquitin has not been removed by ubiquitin proteases. The asterisk marks truncated Roq1. Source data are available online for this figure.