Structural basis for the human SENP5's SUMO isoform discrimination

Abstract

Post-translational SUMO modification is a widespread mechanism for regulating protein function within cells. In humans, SUMO-conjugated proteins are partially regulated by the deconjugating activity of six SENP family members. The proteolytic activity of these enzymes resides within a conserved catalytic domain that exhibits specificity for the two primary SUMO isoforms: SUMO1 and SUMO2/3. SENP5, along with SENP3, are nucleolar proteins involved in ribosome biogenesis and preferentially target SUMO2/3 isoforms. Here, we present the crystal structures of human SENP5 in complex with both SUMO1 and SUMO2 isoforms. These structures reveal a minimal complex interface and elucidate the molecular basis for SENP5's preference for the SUMO2 isoform. This preference can be attributed to a basic patch surrounding SENP5 Arg624 at the interface. Swapping mutagenesis and structural analysis demonstrate that Arg624 is favorably oriented to interact with Asp63 in SUMO2/3, while its interaction with the equivalent Glu67 in SUMO1 is less favorable. These results suggest that subtle structural differences within SUMO isoforms can significantly influence their deconjugation by SENP enzymes, opening new avenues for exploring the regulation of SUMOylation in various cellular processes and for developing therapeutic agents targeting SUMOylation pathways.

Fig. 1. SUMO2 isoform preference of SENP5_CD catalytic domain.

a SDS-PAGE of the endpoint assays of SENP5 using RanGAP1-SUMO1 and RanGAP1-SUMO2 substrates. (Right) Plot of the fraction of the RanGAP1-SUMO2 substrate after 30 min reaction. Data values represent the mean ± SEM, n = 3 technical replicates. Significance was measured by a two-tailed unpaired t test relative to wild-type. All data were analyzed with a 95% confidence interval. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001. Exact P value: 0.0003. b Plot of the proteolytic cleavage of SUMO1-AMC and SUMO2-AMC to test the activity of SENP5. SUMO2-AMC was incubated with SENP5, and released AMC was identified by fluorescence. c Plot of a competition assay by measuring the interference of the SENP5 activity for SUMO2-Rhodamine substrate by competition with either RanGAP1-SUMO1 and RanGAP1-SUMO2 at different concentrations. (Right) Inhibitory constant values for RanGAP1-SUMO1 and RanGAP-SUMO2 substrates on the SENP5 activity on SUMO2-Rho. Data values represent the mean ± SEM, n = 3 technical replicates. Source data are provided as a Source Data file.

Fig. 2. Crystal structure of the complex of human SENP5_CD with SUMO1 and SUMO2.

a Cartoon representations of the SENP5_CD-SUMO1 complex structure. SENP5 catalytic domain and SUMO1 are shown in orange and blue, respectively. The catalytic residues are labelled and depicted in stick representation. Secondary structure elements are labelled. N-terminal and C-terminal are marked. b Cartoon representations of the SENP5_CD-SUMO2 complex structure. SENP5 catalytic domain and SUMO1 are shown in blue and green, respectively. The catalytic residues are labelled and depicted in stick representation. Secondary structure elements are labelled. N-terminal and C-terminal are marked. c Structural superimposition between SENP5_CD-SUMO1 (blue) and SENP5_CD-SUMO2 (green) complexes. SENP5_CD is depicted as an electrostatic surface representation. d Structural alignment of sequences corresponding to the catalytic domains for human SENP2, SENP5 and SENP7. Red asterisks indicate interface residues with SUMO2. Catalytic triad residues are shown in red. Secondary structure cartoon is depicted above for SENP5 (green), SENP2 (violet) and SENP7 (brown). e Superposition of the SENP5_CD structures (blue surface) in complex with SUMO1 (red) or SUMO2 (blue). Side chain interface residues are depicted in a stick representation.

Fig. 3. Structural details of the SENP5_CD complex with SUMO1 and SUMO2.

a Close-up view of the C-terminal tail of human SUMO2 (yellow) in complex with SENP5_CD (blue). Binding side-chain residues in the contact area of substrate and enzyme are shown in stick representation and labelled. Hydrogen bonds are represented by dashed red lines. b Similar close-up views of the C-terminal of SUMO2 (gray) in complex with SENP2 (upper) and SENP7 (below). c (Upper) Close-up view of the ribbon representation of SENP5_CD-SUMO2 interface in two orthogonal views. Binding side-chain residues in the contact area are labelled and shown in stick representation. Hydrogen bonds and charged interactions are represented by dashed red lines. (Below) Similar close-up view of the ribbon representation of SENP5_CD-SUMO1 interface in two orthogonal views.

Fig. 4. Analysis of the SENP5 point mutants.

a Close-up view of the ribbon representation of interfaces of SENP5_CD-SUMO2 (green) and SENP5_CD-SUMO1 (blue) around Asn607 and Arg624. Interface residues are labelled and shown in stick representation. b Plot of the proteolytic cleavage of SUMO1-AMC and SUMO2-AMC to test the activity of SENP5_CD wild type and SENP5_CD N607A, R624A, K627A, and R624/K627A mutants. SUMO2-AMC was incubated with SENP5_CD, and released AMC was identified by fluorescence. The plot is an average of a triplicate experiment. c SDS-PAGE of the endpoint assays of SENP5_CD wild type and SENP5_CD N607A, R624A K627A, and R624/K627A mutants using RanGAP1-SUMO1 and RanGAP1-SUMO2 substrates. (Below) Plot of the fraction of the RanGAP1-SUMO2 and RanGAP1-SUMO1 substrate after 30 min reaction. Data values represent the mean ± SEM, n = 3 technical replicates. Significance was measured by a two-tailed unpaired t test relative to wild-type. All data were analyzed with a 95% confidence interval. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001. Exact P values from left to right: 0.0005, <0.0001, <0.0001, and <0.0001. d Western Blot analysis of HEK293F cells co-transfected with Flag-PML and either the empty pcDNA3 (−) or the plasmids expressing full-length SENP5, SENP5-R624A, SENP5-C713A or SENP5-R624A/K627A. (left) Immunoblot showing the levels of endogenous SUMO1 and SUMO2 in HEK293F cells. Anti-SUMO1 and antiSUMO2 antibodies were used to detect both endogenous SUMO isoforms. Anti-tubulin antibody was used as a loading control. (right) Plot of the Western Blot triplicate showing the fraction of polySUMO conjugates for each experiment regarding to SENP5 active site mutant (C713A). Data values represent the mean ± SEM, n = 3 technical replicates. Significance was measured by a two-tailed unpaired t test relative to C713A mutant. All data were analyzed with a 95% confidence interval. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001. Exact P values from left to right: <0.0001 and 0.0032. Source data are provided as a Source Data file.

Fig. 5. Analysis of the swapping activities of SENP5 for SUMO1 and SUMO2 substrates.

a SDS-PAGE of the endpoint assays of SENP5_CD using the RanGAP1-SUMO1 and RanGAP1-SUMO2 wild types, and RanGAP1-SUMO1 E67D and RanGAP1-SUMO2 D63E point mutants as substrates. (Right) Plot of the fraction of the RanGAP1-SUMO1 wild type and point mutants after 30 min reaction. Data values represent the mean ± SEM, n = 3 technical replicates. Significance was measured by a two-tailed unpaired t test relative to wild-type. All data were analyzed with a 95% confidence interval. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001. Exact P values from left to right: 0.0001 and 0.0173. b Plot of the proteolytic cleavage of SUMO1-Rho, SUMO2-Rho wild types and SUMO1-Rho E67D and SUMO2-Rho D63E point mutants to test the activity of SENP5_CD. The plot is an average of a triplicate experiment. c Plot of a competition assay by measuring the interference of the SENP5_CD activity for SUMO2-Rhodamine substrate by competition with RanGAP1-SUMO1 and RanGAP1-SUMO2 wild types, and RanGAP1-SUMO1 E67D and RanGAP1-SUMO2 D63E point mutants at different concentrations. (Right) Inhibitory constant values (IC₅₀) for RanGAP1-SUMO1 and RanGAP1-SUMO2 wild types, and RanGAP1-SUMO1 E67D and RanGAP1-SUMO2 D63E point mutants substrates on the SENP5_CD wild type activity on SUMO2-Rhodamime fluorogenic substrate. Data values represent the mean ± SEM, n = 3 technical replicates. d Immunoblot analysis of the labeling of endogenous full length human SENP3 in HEK293T lysates with activity-based probes of either SUMO1 or SUMO2 wild type and point mutants. Anti-actin antibody was used as a loading control. n = 3 technical replicates. e Close-up view of the ribbon representation of interfaces of SENP5_CD-SUMO1 wild type depicting the basic patch region around Arg624. Interface residues are labelled and shown in stick representation. f Close-up view of the ribbon representation of interfaces of SENP5 _CD-SUMO1 E67D point mutant depicting the basic patch region around Arg624. g Superposition of the SENP5 structures (blue surface) in complex with SUMO1 wild type (blue ribbon) or SUMO1 E67D (orange ribbon). Source data are provided as a Source Data file.