Active site alanine mutations convert deubiquitinases into high-affinity ubiquitin-binding proteins

Abstract

A common strategy for exploring the biological roles of deubiquitinating enzymes (DUBs) in different pathways is to study the effects of replacing the wild-type DUB with a catalytically inactive mutant in cells. We report here that a commonly studied DUB mutation, in which the catalytic cysteine is replaced with alanine, can dramatically increase the affinity of some DUBs for ubiquitin. Overexpression of these tight-binding mutants thus has the potential to sequester cellular pools of monoubiquitin and ubiquitin chains. As a result, cells expressing these mutants may display unpredictable dominant negative physiological effects that are not related to loss of DUB activity. The structure of the SAGA DUB module bound to free ubiquitin reveals the structural basis for the 30-fold higher affinity of Ubp8^C146A for ubiquitin. We show that an alternative option, substituting the active site cysteine with arginine, can inactivate DUBs while also decreasing the affinity for ubiquitin.

Keywords: deubiquitinating enzyme; polyubiquitin; ubiquitin binding.

Figure EV1. Catalytic activity of DUBm‐Ubp8 mutants

Progress curve of Ub‐AMC cleavage by 125 nM DUBm‐Ubp8^WT or mutants. The experiment as shown was performed once. The C146A and C146S mutants have been shown to lack any cleavage activity in at least two experiments each performed by other assay methods.

Figure 1. Isothermal titration calorimetry assays of SAGA DUB module binding to K48 diubiquitin or monoubiquitin

1. Binding of wild‐type DUBm‐Ubp8 to monoubiquitin.

2. Binding of DUBmUbp8^C146A to monoubiquitin.

3. Binding of DUBm‐Ubp8^C146S to monoubiquitin.

4. Binding of DUBm‐Ubp8^C146R to monoubiquitin.

5. Binding of DUBm‐Ubp8^C146A to K48 diubiquitin.

6. Binding of DUBm‐Ubp8^C146S to K48 diubiquitin.

Data information: Error ranges for K _d values were determined from nonlinear least squares fitting of the data to a one‐site binding model.

Figure 2. X‐ray crystal structure of SAGA DUB module mutant DUBm‐Ubp8^C146A bound to monoubiquitin

1. Overall structure of complex showing Ubp8 (green) with ubiquitin (yellow) bound to the USP domain.

2. Hydrogen bonding contacts between the C‐terminal carboxylate of ubiquitin and Ubp8.

3. In blue spheres, van der Waals radii of C146 and A146 in steric proximity of ubiquitin's C‐terminal carboxylate (yellow). DUBm‐Ubp8^WT structure is shown in teal (PDB ID 3MHH) and DUBm‐Ubp8^C146A is shown in green.

Figure EV2. Active site of DUBm‐Ubp8^C146A bound to monoubiquitin

Density is rendered from the 2F _o–F _c map and contoured at 2σ.

Figure 3. Equilibrium binding of USP4 WT and C311A to TAMRA‐labeled monoubiquitin

Binding was measured by fluorescence polarization using N‐terminally TAMRA‐labeled monoubiquitin. The dissociation constants for ubiquitin binding to USP4 WT and C311A are 92 ± 21 nM 32 and 0.60 ± 0.17 nM, respectively. Error bars are s.d. calculated on five measurements per point.

Figure EV3. Binding kinetics of USP4 WT and C311A to ubiquitin substrates

1. Dissociation kinetics of USP4 WT and C311A binding to TAMRA‐monoubiquitin, measured by stopped‐flow fluorescence polarization.

2. Association kinetics for USP4 WT and C311A binding to TAMRA‐monoubiquitin. Measured by stopped‐flow fluorescence polarization.

3. Equilibrium binding of USP4 C311A to TAMRA‐ubiquitin conjugated to lysine.

4. Equilibrium binding of USP4 C311A to TAMRA‐ubiquitin conjugated to an 18‐mer peptide derived from SMAD4.

Data information: The error bars in panels (C) and (D) are s.d. calculated on two measurements per point.

Figure 4. Enhanced binding of OTUD1 C320A to K63 diubiquitin in vitro and K63 polyubiquitin chains in cells

1. Equilibrium binding of OTUD1 C320A and C320R to K63‐linked diubiquitin was measured by fluorescence polarization using FlAsH‐tagged K63‐linked diubiquitin in which the proximal ubiquitin was fluorescently labeled. Error bars indicate s.d. and are based on three measurements per data point. One representative experiment of two is shown.

2. Whole cell lysates of HEK293 cells expressing HA‐tagged OTUD1 WT, C320R, and C320A were immunoblotted with indicated antibodies. One representative experiment of three is shown.

Figure 5. Binding of USP14 C114A to ubiquitin chains in cells

Polyubiquitin chains were co‐immunoprecipitated with FLAG‐PSMD4, an ubiquitin receptor for the 26S proteasome, from cells expressing either USP14 wild‐type, C114A, or C114R. One representative experiment of two is shown.

Figure EV4. Alignment of the active site of Ubp8^C146A+monoUb to other cysteine protease DUBs

1. A, B

   Alignment of Ubp8^C146A+monoUb active site to (A) ubiquitin‐bound UCHL1 (PDB ID 3KW5) or (B) ubiquitin‐bound Ataxin 3 (PDB ID 3O65).

2. C

   The active site of ubiquitin propargyl‐bound MINDY (PDB ID 5JQS) shown alone, as it does not align with Ubp8^C146A due to a lack of structural conservation.