ARTICLE https://doi.org/10.1038/s41467-022-29485-0 OPEN Structural basis for the SUMO protease activity of the atypical ubiquitin-specific protease USPL1 Ying Li1, Nathalia Varejão 1 & David Reverter 1✉ Post-translational protein modifications by ubiquitin and ubiquitin-like modifiers regulate many major pathways in the cell. These modifications can be reversed by de-ubiquitinating 1234567890():,; enzymes such as ubiquitin-specific proteases (USPs). Proteolytic activity towards ubiquitin- modified substrates is common to all USP family members except for USPL1, which shows a unique preference for the ubiquitin-like modifier SUMO. Here, we present the crystal structure of USPL1 bound to SUMO2, defining the key structural elements for the unusual deSUMOylase activity of USPL1. We identify specific contacts between SUMO2 and the USPL1 subdomains, including a unique hydrogen bond network of the SUMO2 C-terminal tail. In addition, we find that USPL1 lacks major structural elements present in all canonical USPs members such as the so-called blocking loops, which facilitates SUMO binding. Our data give insight into how a structural protein scaffold designed to bind ubiquitin has evolved to bind SUMO, providing an example of divergent evolution in the USP family. 1 Institut de Biotecnologia i de Biomedicina (IBB) and Dept. de Bioquímica i Biologia Molecular, Universitat Autònoma de Barcelona, 08193 Bellaterra, Spain. ✉email: david.reverter@uab.cat NATURE COMMUNICATIONS | (2022)13:1819 | https://doi.org/10.1038/s41467-022-29485-0 | www.nature.com/naturecommunications 1 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-022-29485-0 S mall ubiquitin-like modifiers (SUMO) belong to the family cell, it causes significant frizzled protein mislocalization and of ubiquitin-like proteins (UbLs), which are post- damage in cell proliferation, which interestingly does not depend translational modifiers that regulate a wide plethora of on its deSUMOylase catalytic activity25. The CBs are membrane- protein functions in cells1,2. Among them, the most preeminent less compartments involved in the biogenesis of snRNP and function of ubiquitin is the degradation of intracellular proteins snoRNP, maintenance of telomeres, and processing of histone by the ubiquitin-proteasome system (UPS)3. SUMOylation con- mRNA26. USPL1 is a low-abundant component of the CBs that tributes to many cellular pathways and is involved in many plays a role in the RNAPII transcription of snRNA that is functions, such as DNA replication, nuclear transport, or DNA essential for cell growth25. Moreover, SUMO has been involved in damage control4–6. SUMO, ubiquitin, and all other UbL modi- tumorigenesis, genetic variants of USPL1 are closely related to fiers are covalently attached to target proteins by the formation of grade-3 breast cancer27, as well as USPL1 may also be involved in an isopeptide bond with an internal lysine residue, but before the signaling pathway of multiple myeloma28. that, UbLs need to be activated by a dedicated conjugation USPL1 can interact with both SUMO1 and SUMO2/3, but it pathway through an enzymatic cascade formed by E1, E2, and E3 shows a much higher activity for the SUMO2/3 isoform18. The enzymes7,8. The covalent attachment of ubiquitin and SUMO to sequence of the catalytic domain of USPL1 is unequivocally protein targets is a reversible process through the action of de- related to USP members with deubiquitinase activity, probably ubiquitinating (DUBs) or de-SUMOylating proteases9,10. forming the canonical right hand-like structure with palm, DUBs can be divided into seven families, namely the Ubiquitin thumb, and fingers subdomains. The existing understanding of C-terminal hydrolases (UCHs), Ubiquitin-specific proteases USPL1 is not abundant, and the determinants of the particular (USPs), Machado–Joseph Disease protease family (MJDs), specificity for SUMO have not yet been determined. In order to Ovarian tumor proteases (OTUs), MINDY protease family, get insights into the interaction of USPL1 with SUMO, here we JAMM family, and ZUFSP/Mug105 family, each of which pre- show the crystal structure of the complex between human USPL1 sents a unique structural fold11–14. Most DUB families are and the human precursor of SUMO2 at 1.8 Å resolution. Struc- cysteine proteases, except the JAMM family, which are metallo- tural and biochemical analysis of the complex interface of USPL1- proteases. USPs constitute the largest DUB family constituted by SUMO2 by mutagenesis analysis sheds light on key structural 56 members in humans. The catalytic domain of USPs is com- determinants of this unusual USP family member that has posed of three subdomains, named thumb, palm, and fingers, and evolved to be specific for SUMO. the overall structure resembles the shape of a human right hand15,16. The active site of USPs is formed by a catalytic triad (Cys-His-Asp/Asn) located between the palm and thumb sub- Results domains, which together with the fingers subdomain grasps Covalent crosslinked complex between USPL1 and SUMO2 ubiquitin for proteolytic cleavage15. In addition, USPs are mul- precursor. The catalytic domain for recombinant expression of tidomain enzymes in which adjacent domains aid to achieve human USPL1 is based on the sequence alignment with members multiple functions, such as zinc finger domain (ZnF), ubiquitin- of the USP ubiquitin-specific protease family and comprises like domain (UbL), ubiquitin-related domain (UBA) or ubiquitin residues Met212 through Leu514 (Supplementary Fig. 1)18. The interaction motif (UIM)11,16. USPs are active towards ubiquitin, sequence identity of the catalytic domain among members of the but there are two exceptions: USP18, which is active towards the USP family is typically around 15–50% and most are specific for double-headed ISG15; and USPL1, a distant member of the cleaving off ubiquitin from protein targets or ubiquitin chains (or family specific for SUMO17,18. Whilst ISG15 and ubiquitin share ISG15 in the case of USP18)16. However, USPL1 sequence a 30% sequence identity and the structural analysis showed a identity is one of the lowest in the USP family14, with identities quite similar binding interface with USP18, SUMO, and ubiquitin ranging from 16% for USP7 to 14% for USP28, as examples of display little homology (16% sequence identity), foreseeing the two USP members. The homology is basically observed in the presence of a divergent binding mechanism between USPL1 secondary structure elements around the catalytic triad of the and SUMO. active site (Cys236-His456-Asp472) (Supplementary Fig. 1). This Until recently, SUMO proteases were only constituted by the sequence divergence with all members of the USP family reflects six members of the SENP protease family in humans, which are the existing relevant substitutions in USPL1, which evolved to similar to ULP1, the first discovered SUMO protease in S. cleave off SUMO rather than ubiquitin from protein targets. cerevisiae19 and all belong to the C48 cysteine protease structural The complex between USPL1 and SUMO2 was prepared by class (SENP1-SENP3 and SENP5-SENP7)10. However, two novel taking advantage of the formation of a covalent bond between the types of SUMO proteases have been described in recent years SUMO2 C-terminal glycine (Gly93) and the active site cysteine of without sequence/structural homology to the SENP/ULP family: USPL1 (Cys236), which stabilizes the complex, increases the the deSUMOylating peptidase 1 and 2 (DESI1 and DESI2)20; and binding affinity and thus the chances of crystallization. To the deubiquitinating enzyme USPL1, which is a member of the prepare this covalent bond, we used the dehydroalanine (DHA) ubiquitin USP family but is specific for SUMO, instead of strategy29, which creates an electrophilic center highly reactive ubiquitin18. An analogous finding was observed in the char- with nucleophiles such as internal cysteines in proteins (Fig. 1a). acterization of SENP8/DEN1, which is another member of the First, in the human SUMO2 precursor, Gly93 was substituted for SUMO protease family (SENP/ULP family), but specific for the cysteine, and Cys48 was substituted for serine, leaving only one Nedd8, another type of UbL modifier21–24. internal cysteine at the position of the reactive Gly93. Next, the Full-length of USPL1 is composed of 1.092 residues containing SUMO2-C48S/G93C precursor was incubated with 2,5-dibromo- a USP-like catalytic domain in a middle region and long, dis- hexanediamide, which can desulfurize the Cys93 of the SUMO ordered protein extensions without the presence of evident precursor into DHA at mild conditions. Finally, the SUMO2 globular domains. So far, the only function described for USPL1 precursor bearing DHA at position of Gly93 was incubated with takes place in the nucleus, where it is a component of the Cajal USPL1 to create a covalent bond between active site nucleophile Bodies (CBs), co-localizing with coilin18. USPL1 seems to have an cysteine of USPL1 and the DHA electrophilic trap on the SUMO2 impact on the formation and dynamics of CBs and in cell pro- precursor (Fig. 1a). liferation, as observed in the USPL1-depleted cells18. Whilst A major concern of the DHA strategy in SUMO proteases is USPL1 deletion does not affect the overall SUMOylation in the the possible steric hindrances in the C-terminal di-glycine motif 2 NATURE COMMUNICATIONS | (2022)13:1819 | https://doi.org/10.1038/s41467-022-29485-0 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-022-29485-0 ARTICLE a b SH USPL1 SENP2 SENP7 2,5-dibromo Y Y hexane diamide MW 0 1 3 0 1 3 0 1 3 h V V (KDa) SUMO2 SUMO2 N pH 8 N H O H O 50- - USPL1-SUMO2 40- 30- USPL1 25- 20- 15- - SUMO2 DHA S + USPL1 pH 8, 30 ºC V Y SUMO2 SH N H O Fig. 1 Covalent thioether bond formation of SUMO2-DHA with USPL1 catalytic domain. a Schematic representation of the reaction to form the covalent thioether between the SUMO2-C48S/G93C precursor and USPL1 catalytic domain. b Active site probe assay for USPL1, SENP2, and SENP7 with SUMO2- DHA. Time course assay of USPL1, SENP2, and SENP7 at 2 μM using the SUMO2-DHA substrate at 6 μM at 30 °C for 3 h. n = 3 technical replicates. Source data are provided as a Source Data file. after the substitution of Gly93 for DHA, which might clash with the structure of the finger domain, as occurs in several USPs, narrow protease binding pockets. Therefore, we initially checked forming a zinc-finger motif (ZnF). In addition to the general the formation of the covalent crosslink after incubation of structural role of the ZnF motif, in USP21 it is also important for SUMO2 DHA precursor with USPL1, and two other well- the binding of the distal ubiquitin in a linear diUb substrate33. characterized SUMO proteases, SENP2 and SENP7 (Fig. 1b). SUMO2 is grabbed by the USPL1-like right hand in the complex, Interestingly, only USPL1 is able to form the covalent crosslink with the C-terminal tail of the SUMO2 precursor extended with the SUMO2 DHA precursor, indicating a correct and towards the catalytic triad of the USPL1 active site formed by specific binding between SUMO2 and USPL1. However, the two Cys236, His456, and Asp472 (Fig. 2b, c). Structural comparison other well-characterized SUMO proteases, SENP2 and SENP7, between USPL1 in complex with SUMO2 with the apo form of are unable to form a covalent bond between the active site USPL1 from the AlphaFold model indicates a very good cysteine and DHA (Fig. 1b). As known in the SENP/ULP protease overlapping of both structures, with a main-chain rmsd (root family, the integrity of the C-terminal di-glycine motif is essential mean square deviation) of 0.8 Å (Fig. 2c). The structural for C-terminal binding, which must be placed in a shallow tunnel overlapping also revealed a similar location of residues forming formed by two tryptophan residues and the simple substitution of the active site catalytic triad, with a 3.8 and 3.6 Å distance glycine for alanine precludes the interaction30,31, as probably between the Cys236 Sγ and the His456 Nδ1; and 2.8 and 2.9 Å occurs in SENP2 and SENP7. On the other hand, since the between His456 Nε2 and Asp472 Oδ1, in the unbound and reaction between SUMO2 DHA precursor and USPL1 is highly SUMO2-bound USPL1, respectively (Fig. 2c). Such distances efficient, we envision a different binding mechanism for the suggest that the USPL1 catalytic triad is already preformed and interaction of the C-terminal tail of SUMO2 with the active site the protease might be active in the absence of the SUMO groove of USPL1, compared to the SENP protease family. substrate, in contrast to the apo structures of other UPSs, such as USP7, USP15, and USP18, in which binding of the ubiquitin Overall structure of the USPL1-SUMO2 complex. The USPL1- substrate is necessary to rearrange the catalytic triad to an active SUMO2 complex was formed by incubation of USPL1 with the conformation15,17,34. SUMO2 DHA precursor at 37 °C for 2 h to form the covalent bond between the DHA group with the active site cysteine of Structural comparison with ubiquitin-specific USP members. USPL1 (Supplementary Fig. 2). A few diffraction quality crystals As mentioned above, the sequence identity between USPL1 and of the USPL1-SUMO2 complex were obtained in the initial the other USP members with deubiquitinase activity is only screening in a condition containing 0.2 M potassium thiocyanate, around 15%, which differs from the higher sequence identities 0.1 M sodium acetate pH 5.5, 8% w/v PEG20000, and 8% w/v shown among the ubiquitin-specific USP members (Fig. 3a and PEG500MME, which was hard to reproduce in subsequent Supplementary Fig. 1). This might be a consequence of the screenings. Molecular replacement with available USP models did deSUMOylase activity displayed by USPL1 in contrast to the not work due to the lower sequence identity with USPL1 (<15%), deubiquitinase activity for all other USPs. Structural overlapping but fortunately, we were able to solve the structure by using the between USPL1 with USP7 (rmsd 3.02 Å, 197 aligned, 14.21% recently reported USPL1 model from the alpha-fold server32, identity) and with USP28 (rmsd 2.61 Å, 184 aligned, 14.67% which unambiguously resulted in a correct final solution. The identity) are low, only showing a good superposition in the sec- crystals contained one USPL1-SUMO2 complex per asymmetric ondary structure elements encompassing the catalytic triad in the unit and diffracted to a resolution of 1.8 Å. The final electron active site in the palm subdomain, namely the α1 helix (Cys236) density map model of the USPL1 catalytic domain includes most and the β9 and β10 strands (His456 and Asp472) (Fig. 3a, b). In of the residues (Ser225 to Ile501), with the only disruption of a general, the thumb and fingers subdomains of USPL1 show very disordered loop connecting two alpha helices in the thumb sub- little homology with USP7 and USP28, particularly in the dif- domain (Leu285 to Lys295). ferent length and orientation of α3, α4, and α5 helices in the USPL1 adopts the typical fold of the catalytic domain of the thumb subdomain, and in the beta sheet containing the ZnF USP family, resembling the shape of a right-hand containing motif in the fingers subdomain15,35. palm, fingers, and thumb subdomains (Fig. 2b and Supplemen- Interestingly, a major difference in USPL1 is the lack of two tary Fig. 3). A Zn2+ atom coordinated by four cysteines stabilizes loops in the palm subdomain, namely so-called blocking loops, NATURE COMMUNICATIONS | (2022)13:1819 | https://doi.org/10.1038/s41467-022-29485-0 | www.nature.com/naturecommunications 3 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-022-29485-0 a Nt Nt hSUMO2 hSUMO2 Zn2+ Zn2+ fingers thumb 180º thumb fingers hUSPL1 C236 Ct Ct H456 palm D472 Nt Nt palm Ct hUSPL1 hSUMO2 Ct Y95 b c T91 V94 hSUMO2 G92 DHA93 hSUMO2 Y95 Zn2+ 3.6 C236 T91 H456 Q89 Ct 2.9 V94 Ct C236 D472 G92 DHA93 hUSPL1 H456 Q88 C236 Q90 D472 hUSPL1-AF H456 D472 hUSPL1/AF hUSPL1 hUSPL1-SUMO2 3.8 C236 H456 2.8 D472 Fig. 2 Crystal structure of the complex of human USPL1 with SUMO2 precursor. a Two views of the USPL1-SUMO2 complex structure are shown in cartoon representation. USPL1 catalytic domain and SUMO2 precursor are shown in purple and red, respectively. USP right hand-like domains are labeled. The catalytic residues are labeled and depicted in stick representation. Zinc atom is shown in yellow. b Detailed view of the 2Fo-Fc electron density map contoured at 1σ of the C-terminal tail of SUMO2 bound to USPL1. Same color code as in (a). c Structural superimposition between USPL1-SUMO2 (blue) complex and the USPL1 AlphaFold (AF) model (yellow). Close-up views of the superimposition of active sites of USPL1 and AF model are shown in the right panel. Catalytic triad residues from USPL1 are labeled and shown in stick representation, whereas the corresponding residues from AF model are shown below. Hydrogen bonds are represented by dashed lines and the distances (Å) are also depicted. between strands β6 and β7, and between β8 and β9 next to the The electron density maps clearly show the covalent bond formed catalytic triad (Fig. 3a–d). These two loops are highly conserved between DHA93 in SUMO2 and Cys236 in USPL1 (Fig. 2b), in in the USP family due to their major role in ubiquitin binding which the sidechain and not the C-terminal carboxylate is (Supplementary Figs. 1 and 6) and they are structurally crosslinked to Cys236 Sγ. This fact constrains the geometry of rearranged upon ubiquitin binding to trap ubiquitin within the this region, as observed by the cis configuration of the Gly92- right hand-like structure15. However, in USPL1 both loops are DHA93 peptide bond in the USPL1 complex (Fig. 4), in which not required for binding to SUMO2. In fact, the superposition of strong hydrogen bonds between the nitrogen and carbonyl oxy- USPL1 with USP28 displays a ∼15° angle rotation of SUMO with gen of the Gly92-DHA93 peptide bond are established with the respect to ubiquitin, which would result in a collision with the USPL1 backbone (each 2.9 Å distance) (Fig. 4). larger Blocking Loop (Fig. 3e). Thus, in addtion to the unique The extended conformation of the SUMO2 tail (Gln90-Thr91- contacts with fingers and thumb subdomains, the lack of the Gly92) is comparable to ubiquitin, but the specific backbone blocking loops in USPL1 contributes to the different position of hydrogen bonds are different. Whereas Gly92 engages similar SUMO on the surface of USPL1 compared to ubiquitin, which is hydrogen bonds with USPL1 as in the ubiquitin-USP28 complex, particularly evident in the different orientation of C-terminal tail Thr91 forms two hydrogen bonds with the side chain of Tyr451 backbone (Fig. 3f), and in the comparison of the electrostatic Oζ (3.2 and 2.7 Å distance for the backbone N and O, potential surfaces between USPL1 and USP28 (Fig. 3g). respectively) (Fig. 4). In ubiquitin USPs Tyr451 is occupied by a highly conserved histidine (Supplementary Fig. 4), forming a SUMO2 interface of the C-terminal tail with USPL1. As similar hydrogen bond with the backbone oxygen of ubiquitin expected by the divergent C-terminal sequences of SUMO Arg74, but in this case, the backbone nitrogen interacts with the (-FQQQTGG) and ubiquitin (-VLRLRGG), the contacts engaged β8-β9 blocking loop, absent in USPL1 (Supplementary Fig. 4). by the SUMO C-terminal tail with the active site cavity of USPL1 The backbone oxygen of Gln90 forms a hydrogen bond with a constitute key signatures for the specificity of USPL1 for SUMO2. water molecule that is fixed by contacts with His493 and Ser332, 4 NATURE COMMUNICATIONS | (2022)13:1819 | https://doi.org/10.1038/s41467-022-29485-0 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-022-29485-0 ARTICLE a b 1 2 3 225 240 260 280 300 Blocking hUSPL1 Loops hUSP28 2 4 5 6 1 2 5 320 340 360 1 5 3 4 2 6 1 3 4 3 4 5 6 7 7 AS 380 400 420 440 8 9 10 Blocking Loop 11 8 9 10 11 12 460 480 500 Blocking Loop c d e Zn2+ ubiquitin hSUMO2 hSUMO2 ubiquitin Blocking Loops AS hUSP28 hUSPL1 hUSP28 Nt Nt f g R72 fingers hUSPL1 fingers hUSP28 ubiquitin Blocking R74 Loops L71 hSUMO2 ubiquitin G75 C236 T91 G92 thumb D472 thumb Q88 Q89 H456 AS AS hSUMO2 Fig. 3 Comparative analysis of USPL1 with the ubiquitin USP28. a Alignment of sequences corresponding to the catalytic domains for human USPL1, USP7, and USP28 based on the structural alignment of human USP7 (PDB code 4WPI) and USP28 (PDB code 6HEJ)35,44. Secondary structure elements are numbered (β-strands and α-helices) and indicated above the alignment for USPL1 (blue) and below the alignment for USP28 (gray). Gaps are denoted by --- and the large sequence insertion within USP28 is depicted ≠ to indicate that the sequence is missing from the alignment. Side chain identity (100% conservation) is denoted in the alignment by a yellow background. Three catalytic residues are depicted in red. Active site-blocking loops are denoted by red squares. All images were prepared with PyMOL45. b Cartoon representation of the structural overlapping of the catalytic domains of USPL1 with USP28 (PDB code 6HEK)35. The blocking loops connect β-strand 6 and β-strand 7, and β-strand 8 and β-strand 9 of the palm domain are missing in USPL1 and shown in yellow. Active site residues (AS) are shown in stick representation. Zinc atom is shown in yellow. c Surface representation of SUMO2 (red) in complex with USPL1, shown in cartoon representation (blue). Active site residues (AS) are shown in stick representation. Zinc atom is shown in yellow. d Surface representation of ubiquitin (dark green) in complex with USP28 (PDB code 6HEK)35, shown in cartoon representation (gray). Active site residues (AS) are shown in stick representation. Zinc atom is shown in yellow. Blocking loops (yellow) are indicated. e Structural comparison of C-terminal tails of SUMO2 (red) and ubiquitin (green) in complex USPL1 and USP28, respectively. f Structural overlapping of SUMO2-USPL1 and ubiquitin-USP28 displaying the collision of the blocking loop (yellow) of USP28 (gray) with SUMO2 (red). g Electrostatic potential surface representation for USPL1-SUMO2 and USP28-ubiquitin to highlight the differences between USPL2 and USP28 in the analogous surface. SUMO2 (red) and ubiquitin (dark green) are shown in a line representation. both highly conserved in USPL1 (Fig. 4 and Supplementary Such tyrosine, located one position after the active site histidine, Fig. 1). In ubiquitin the backbone oxygen of Leu73 (equivalent to is highly conserved in all ubiquitin-specific USPs (Fig. 3 and Gln90 in USPL1) forms a hydrogen bond with the side chain of a Supplementary Fig. 1), but has been substituted by phenylalanine tyrosine (replaced by Phe457 in USPL1) (Supplementary Fig. 4). in all USPL1 orthologs, which forms hydrophobic contacts with NATURE COMMUNICATIONS | (2022)13:1819 | https://doi.org/10.1038/s41467-022-29485-0 | www.nature.com/naturecommunications 5 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-022-29485-0 hSUMO2 E331 Y95 Y233 M330 A234 Q89 T91 Ct G92 Q88 Q90 V94 F457 H421 C236 N455 H493 Y451 H456 V423 D472 hUSPL1 hSUMO2 E331 W237 L235 Y233 S332 H421 T91 A234 H493 G92 C236 DHA93 Y95 M330 Q90 Q89 Q88 Y451 H456 Ct V94 D472 hUSPL1 N455 Fig. 4 Atomic details of the C-terminal tail interaction of SUMO2 with USPL1. Two views of the stick representation of the main contacts of the C-terminal of SUMO2 in complex with the active site groove of USPL1. USPL1 catalytic domain and SUMO2 are shown in blue and red, respectively. SUMO2 and USPL1 interface residues are labeled. Dashed lines indicate hydrogen bond contacts. On the left side, surface representation of SUMO2 (red) in complex with USPL1, shown in a blue cartoon. Interface residues are shown in stick representation. Zinc atom is shown in yellow. Gly92 but is unable to establish a hydrogen bond as in ubiquitin. SUMO2 Interface with the thumb and fingers subdomains. In addition, the side chain of Gln90 forms a hydrogen bond with Two major-specific contacts stand out in the interaction between a conserved His421 (3.0 Å distance) and with the backbone the USPL1 thumb subdomain and the SUMO2 surface, namely oxygen of Val423. Finally, the C-terminal tail of SUMO2 (Gly92) Arg324 and Phe335. Arg324 is engaged in a strong well-oriented is sandwiched between Phe457 and Met330, substituted by a electrostatic interaction with Asp71 of SUMO2, with 2.7 and highly conserved glutamine in ubiquitin-specific USPs. 3.1 Å distances between the NH1 and NH2 of the Arg324 gua- A relevant difference compared to the ubiquitin-specific USPs nidinium group and the OD1 and OD2 of the Asp71 carboxylate complexes is the extensive interface established by the two group, respectively (Fig. 5a). Arg324 is well conserved in all arginines and the two leucines in the ubiquitin tail (Leu71-Arg72- USPL1 orthologs and this interaction is unique to USPL1, not Leu73-Arg74), which participate in electrostatic and hydrophobic present in ubiquitin-specific USPs. Interestingly, SUMO2 Asn68 contacts on opposite sides of the ubiquitin tail. Particularly and Asp72 were important determinants for the specificity of relevant is the interaction of the two ubiquitin leucines with the SENP7 for SUMO2/3 over SUMO136,37, where they are sub- blocking loops, absent in USPL1, and the ubiquitin Arg72 and stituted by alanine and histidine, respectively. In USPL1, the Arg74 with Glu258 and Gln254 in the ubiquitin-USP28 complex proteolytic activity against SUMO1 is low compared to SUMO2 (Supplementary Fig. 4) (PDB:6HEK)35. In contrast, in USPL1 (Fig. 6), however, removal of the side chain of Arg324 did not only Gln90 from the equivalent Gln88-Gln89-Gln90-Thr91 tail is increase the activity for SUMO1 substrates, as observed in activity engaged in a specific interaction with His421, indicating that this assays with the USPL1 R324A point mutant (Supplementary interface might be less relevant in USPL1 for SUMO binding. Fig. 5). The complex with USPL1 was formed with the SUMO2 Phe335 is buried in a hydrophobic groove in the precursor, which contains the complete or immature C-terminal SUMO2 surface formed by Pro66, Phe87, and the aliphatic tail formed by Val-Tyr extension after the proteolytic cleavage site chains of Arg59 and Arg61 (Fig. 5a). Phe335 is highly conserved (di-Glycine motif). Both residues at the SUMO2 C-terminal tail in all USPL1 orthologs and not observed in other USPs. In are well observed in the electron density maps, but probably their ubiquitin-specific USPs this location is occupied by a conserved structural conformation is constrained by the covalent crosslink glutamate (Glu258 in USP28), which is engaged in a strong between DHA93 and the USPL1 Cys236. electrostatic bridge interaction with Arg72, from the ubiquitin 6 NATURE COMMUNICATIONS | (2022)13:1819 | https://doi.org/10.1038/s41467-022-29485-0 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-022-29485-0 ARTICLE a hSUMO2 Table 1 Crystallographic statistics of the USPL1-SUMO2 complex. USPL1 D71 R324 USPL1-SUMO2 N68 Data collection Space group P21 P66 Unit cell parameters (Å) 50.71, 69.88, 53.64 R61 Wavelength (nm) 0.97625 Resolution range (Å) 41.04–1.79 F87 F335 Rmerge 0.08 (0.47)a Rpim 0.06 (0.36)a R59 (I/σ(I)) 7.8 (2.1)a Completeness (%) 96.5 (97.8)a Multiplicity 2.5 (2.5)a b C364 C400 N398 CC (1/2) 0.99 (0.72)a Zn2+ Structure refinement Resolution range (Å) 41.04–1.80 C397 P396 No. of unique reflections 33540 C361 S28 G27 Rwork/Rfree (%) 17.6/19.7 D26 Q25 No. of atoms Protein 2823 Q88 Water molecules 200 T380 Zn2+ 1 V423 Overall B factors (Å2) 33.99 USPL1 H421 USPL1 (Å2) 31.10 SUMO2 (Å2) 42.66 Fig. 5 SUMO2 interface with the thumb and fingers subdomains of Zn2+ (Å2) 24.22 USPL1. a Stick representation of the main contacts of the thumb subdomain Water molecules (Å2) 38.29 of USPL1 with SUMO2. USPL1 catalytic domain and SUMO2 are shown in Rms deviations blue and red, respectively. SUMO2 and USPL1 interface residues are Bonds (A) 0.007 labeled. Dashed lines indicate hydrogen bond contacts. On the left side, Angles (°) 0.845 surface representation of SUMO2 (red) in complex with USPL1, shown in a Ramachandran favored (%) 97.94 blue cartoon. Interface residues are shown in stick representation. Zinc Ramachandran allowed (%) 1.77 atom is shown in yellow. b Stick representation of the main contacts of the Ramachandran outliers (%) 0.29 PDB code 7P99 fingers subdomain of USPL1 with SUMO2. Same color code as in (a). aData from the last shell in parenthesis (1.70–1.79 Å). C-terminal tail (Supplementary Fig. 4). Both Phe335 in USPL1 and Glu258 in USP28 probably have an important impact on the binding and specificity of SUMO2 or ubiquitin, respectively. which shows similar activities for SUMO1 and SUMO2-AMC In the USPL1 fingers subdomain, a Zn2+ atom stabilizes the substrates (Fig. 6a)31. As expected, USPL1 does not show any structure by the coordination of four conserved cysteine residues activity against ubiquitin-AMC substrate (Fig. 6a)18. To char- (Fig. 5b), and as observed in other USPs, the cysteine acterize the binding interface in this kinetically trapped inter- coordination to Zn2+ is essential for the USP catalytic activity. mediate complex, specific interactions in the USPL1-SUMO2 In USPL1, the contacts are basically engaged by the β1-β2 hairpin interface have been mutated. The overall structural integrity of all loop of SUMO2, which nicely fits in the fingers subdomain USPL1 mutants is similar to the wild-type form, as observed by surface. The only specific side-chain contact is engaged by comparing their purification profiles by ion-exchange chroma- Asn398, located next to the Zn2+ site and conserved in all USPL1 tography and by the similar spectra displayed in the intrinsic Trp- orthologs, which forms two hydrogen bonds with the main chain fluorescence emission analysis (Supplementary Fig. 7). The oxygen and nitrogen of Gly27 and Ser28 in SUMO2, respectively USPL1 mutant interface analysis has been conducted in the (2.8 and 3.0 Å distances). In ubiquitin-USP complex structures, C-terminal tail, thumb, and fingers subdomains (Fig. 6c). Binding the fingers subdomain normally display a rigid body adjustment to SUMO2 has been checked by using a SUMO2 DHA chemical to fix ubiquitin, sometimes having an impact in chain specificity, trap substrate, and the proteolytic activity has been checked as occurs in USP2133. However, in the USPL1-SUMO2 complex, against the SUMO2-AMC fluorescent substrate, as well as by the orientation of the fingers subdomain is similar to the RanGAP-SUMO2 and diSUMO2 substrates (Table 1). AlphaFold-2 model of the apo form (Fig. 2) and displays a In the C-terminal tail interface, Phe457 has been replaced either different orientation compared to all known structures of by tyrosine, which is present in all USP members, or by leucine, ubiquitin USPs, showing an average 14 Å displacement between which maintains the hydrophobic character of the interaction. In the Zn2+ atoms of the ZnF motifs (Supplementary Fig. 6). As a both F457Y and F457L mutants the binding reaction with SUMO2 consequence, the contacts observed at the interface of the fingers DHA is strongly diminished and the catalytic activity against all subdomain with SUMO2 are not observed in the other USP- tested substrates is reduced (Fig. 6b–f). Interestingly, despite that all ubiquitin complexes, basically due to the different orientation of other USP members contain a tyrosine in that position, the only the fingers subdomain with respect to SUMO2. presence of a Oζ from the tyrosine side chain seriously compromises the interaction of USPL1 with SUMO2 substrates. The SUMO C-terminus is sandwiched between Phe457 and Met330 Mutagenesis analysis of the specific interface contacts in (Fig. 6c), however, removal of Met330 side chain has a minor effect USPL1. USPL1 shows a proteolytic isoform preference for human on the catalytic activity and even seems to increase the binding SUMO2, in contrast to other SUMO peptidases such as SENP2 affinity for SUMO2, as observed in the M330A mutant (Fig. 6b). NATURE COMMUNICATIONS | (2022)13:1819 | https://doi.org/10.1038/s41467-022-29485-0 | www.nature.com/naturecommunications 7 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-022-29485-0 a USPL1 wt Ubiquitin - AMC SENP2 USP28 USPL1 SUMO2 - AMC SUMO2 - AMC SUMO1 - AMC SUMO1 - AMC b c thumb MW (KDa) R324 50- - USPL1-SUMO2 40- fingers N398 F335 E331 30- - USPL1 M330 Zn2+ H421 F457 H493 H456 Y451 D472 d SUMO2 - AMC SUMO2 - AMC WT WT M330A F335A WT M330A Y451F H421N F457Y F457L Y451F R324A 1.38±0.27 0.55±0.16 0.38±0.09 0.20±0.07 0.15±0.03 0.06±0.02 H421N N398A * ** ** ** ** F457Y F457L H493N E331P N398A R324A F335A H493N 0.00±0.00 0.00±0.00 0.18±0.05 0.29±0.08 0.73±0.40 E331P *** *** ** ** ns MW - *** *** *** *** e (KDa) *** ** *** *** 40- *** - RanGAP1-SUMO2 *** 30- 25- 20- - RanGAP1 f MW (KDa) - 40- ** 30- - di-SUMO2 *** *** 25- *** *** *** 20- *** *** *** - SUMO2 While Met330, which is normally substituted by glutamine in Finally, two conserved histidine residues were analyzed: H421N ubiquitin USPs, may still be involved in regulating USPL1 activity, it removes the specific hydrogen bond contact with Gln90, and does not seem to be essential for binding and catalysis. However, H493N removes the interaction to a fixed water (Fig. 4). the backbone interactions of the adjacent Glu331 with the SUMO Interestingly, whereas the H421N mutant affects partially binding C-terminal tail are essential for binding and catalysis, as observed in and catalysis, the integrity of the His493 side chain, which bridges the E331P mutant, in which a proline substitution distorts the the C-terminal SUMO backbone through a water molecule, seems backbone orientation and removes a critical hydrogen bond essential for binding and catalysis. The overall structural integrity (Fig. 6b–f). of the H493N mutant seems correct, as observed in Trp- Removal of the Oζ from the tyrosine side chain in the Y451F fluorescence stability analysis compared to wild-type (Supple- mutant decreases binding and activity by fourfold compared to mentary Fig. 7), thus highlighting the relevance of this water- wild-type (Fig. 6b–f), highlighting the role of the two hydrogen bridged interaction in the overall hydrogen bond network of the bonds with Thr91 of the C-terminal tail of SUMO2 (Fig. 4). C-terminal tail. 8 NATURE COMMUNICATIONS | (2022)13:1819 | https://doi.org/10.1038/s41467-022-29485-0 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-022-29485-0 ARTICLE Fig. 6 Functional analysis of the SUMO2-USPL1 interface. a Left, activity assay of USPL1 with SUMO1-AMC and SUMO2-AMC. Middle, activity assay of Ubiquitin-AMC with USPL1 and USP28. Right, activity assay of SENP2 with SUMO1-AMC and SUMO2-AMC. Reactions were conducted in triplicate and the average curve is displayed. b Binding interaction of USPL1 point mutants with SUMO2-DHA. Reaction assays were conducted with USPL1 wild-type and mutants for 3 h. n = 3 technical replicates. c USPL1-SUMO2 complex structure shown in cartoon representation. USPL1 catalytic domain and SUMO2 precursor are shown in purple and red, respectively. USP right hand-like domains are labeled. Interface residues are labeled and shown in stick representation. Zinc atom is shown in yellow. d Left, activity assays of USPL1 wild type and mutants of the C-terminal tail interface using SUMO2-AMC. Middle, similar activity assays of USPL1 wild type and mutants of the thumb and fingers subdomains. Right, table indicating the mean slope values plus/ minus the standard deviation of the activity assays. n = 3 technical replicates. Significance was measured by a two-tailed unpaired t-test relative to wild- type. *P < 0.05, **P < 0.01, ***P < 0.001. e Left, endpoint assays of USPL1 wild type and mutants using the RanGAP1-SUMO2 substrate. Right, plot of the RanGAP1-SUMO2 fraction after 10 min reaction. Data values represent the mean ± SD, n = 3 technical replicates. Significance was measured by a two- tailed unpaired t-test relative to wild-type. *P < 0.05, **P < 0.01, ***P < 0.001. Exact P values from left to right: <0.0001, 0.0868, <0.0001, 0.0002, <0.0001, 0.0001, 0.0065, <0.0001, <0.0001, <0.0001. f Left, endpoint assays of USPL1 wild type and mutants using di-SUMO2 substrate. Right, plot representation of the product SUMO2 fraction after 20 min reaction. Data values represent the mean ± SD, n = 3 technical replicates. Significance was measured by a two-tailed unpaired t-test relative to wild-type. *P < 0.05, **P < 0.01, ***P < 0.001. Exact P values from left to right: <0.0001, <0.0001, 0.0003, <0.0001, 0.0021, <0.0001, 0.0009, <0.0001, <0.0001, <0.0001. Source data for Fig. 6a, b, d, e, and f are provided as a Source Data file. Two specific contacts to SUMO2 have been analyzed in the a b ubiquitin fingers USPL1 thumb subdomain, Arg324, and Phe335, both highly fingers conserved in the USPL1 orthologs (Supplementary Fig. 1). The N398 hSUMO2 R342A mutant removes a strong and well-oriented electrostatic R324 thumb bridge with Asp71 of SUMO2 (Fig. 5a), and the results indicate F335 thumb that binding and catalysis with SUMO2 substrates are seriously L73 R72 compromised, highlighting a major role for Arg324 (Fig. 6b–f). Q88 Q89 T91 Blocking Loop L71 Q90 Likewise, in a lesser degree compared to Arg324, the F335A M330 R74 mutant also disturbs binding and catalysis for SUMO2 substrates. Thus our in vitro proteolytic activities as well as the strong conservation in the family indicate that both Arg324 and Phe335 hUSPL1 hUSP28 play a major role in the specific interaction of SUMO2 Fig. 7 Comparison of the C-terminal tails of SUMO2 and ubiquitin. with USPL1. a Electrostatic potential surface representation for the USPL1 in complex The structural architecture of the fingers subdomain is maintained with SUMO2 (red line). The C-terminal tail of SUMO (QQQTGG) is labeled by four conserved cysteine coordination residues forming a ZnF and shown in stick representation. Main interface contacts of thumb and motif (Fig. 5b). Despite the extended interface of the fingers fingers subdomains are labeled. b Electrostatic potential surface subdomain with SUMO, only the highly conserved Asn398 seems to representation for the USP28 (PDB code 6HEK) in complex with ubiquitin establish specific contacts by the formation of two hydrogen bonds (green lines) from complexes with USP2 (PDB code 2hd5), USP7 (PDB with SUMO2 backbone atoms (Fig. 5b). Interestingly, removal of code 5JTV), and USP30 (PDB code 5OHK)35,46–48. The C-terminal tail of these interactions in the N398A mutant seriously compromises ubiquitin (LRLRGG) is labeled and shown in stick representation. binding and catalysis with SUMO2 substrates. Altogether, the specific contacts of the thumb and fingers subdomain, Arg324, Phe335, and Asn398, have an essential contribution to fix the globular domain of by the backbone, the presence of two arginines and two leucines SUMO2 in a correct orientation for catalysis. in opposite sides of the ubiquitin tail generates unique interac- tions with acidic and hydrophobic surface patches in all ubiquitin Discussion USPs (Fig. 7), including USP18, in which ISG15 has a similar tail All members of the human USP family of DUBs show specificity as ubiquitin. Interestingly, such hydrophobic patch in all towards ubiquitinated substrates, with the exception of USP18, ubiquitin-specific USPs is formed by the so-called blocking loops, which is specific towards ISG15, a double-headed UbL modifier which are structurally rearranged upon binding to the ubiquitin similar to ubiquitin, and USPL1, which shows specificity towards substrate. Interestingly, in all USPL1 orthologs, the two blocking SUMO, a different UbL modifier. USPs are multidomain proteins loops sequences have been deleted and are absent in the structure. that utilize different strategies to ensure their specific functions in Thus, in contrast to the high number of specific contacts engaged the cell. In some cases, the presence of additional domains by the -LRLR- motif of the ubiquitin C-terminal tail, in the adjacent to the USP catalytic domain enhances substrate binding USPL1-SUMO2 structure only the Gln90 side chain of the or provides specificity towards a particular type of polyubiquitin equivalent -QQQT- motif forms specific interactions with USPL1, chain. Commonly, in all members of the USP family, the struc- and all other contacts in the C-terminal tail of SUMO are tural fold of the catalytic domain, formed by the characteristic established by backbone atoms (Fig. 4). palm, thumb, and fingers subdomains reminiscent of a human It is interesting to note how USPL1 has evolved to interact with right hand, is maintained, exposing all the key surface elements SUMO. Key USP major residues for specific interaction with necessary for interaction with ubiquitin. However, as described ubiquitin have been deleted or, in some cases, replaced by resi- here, USPL1 is a distant member of the family with the canonical dues specific for interacting with SUMO2 (Fig. 3 and Supple- right-hand scaffold of a ubiquitin USP catalytic domain, but mentary Fig. 1). The di-glycine motif in the C-terminal tail of exposing key surface elements unique for SUMO binding, a low- SUMO, also conserved in ubiquitin, is sandwiched by a narrow identity UbL in comparison to ubiquitin. cleft formed by Phe457 and Met330, both located on opposite Possibly, a major difference between USPL1 and the other USP sides of Gly92. Whilst Phe457 is highly conserved in all USPL1 members can be found in the interface formed by the dissimilar orthologs, it is replaced by tyrosine in all ubiquitin USPs, which C-terminal tail of ubiquitin (LRLRGG) and SUMO (QQQTGG) forms a hydrogen bond with the ubiquitin C-terminal (Fig. 7). In addition to different hydrogen bond contacts engaged backbone15. Interestingly, a phenylalanine substitution by NATURE COMMUNICATIONS | (2022)13:1819 | https://doi.org/10.1038/s41467-022-29485-0 | www.nature.com/naturecommunications 9 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-022-29485-0 tyrosine in USPL1, F457Y, which mimics ubiquitin USPs, shows for these long non-catalytic disordered extensions of the full- decreased binding and reduced catalysis of SUMO2 substrates length USPL1, such as in the recruitment of SUMO substrates or (Fig. 6), underscoring the relevance of single contacts at the in the regulation of the proteolytic activity of the catalytic USPL1-SUMO2 interface. Another remarkable difference is domain, perhaps by internal protein–protein interactions. USPL1 Tyr451, conserved in all USPL1 orthologs and replaced by his- has been reported to be involved in the regulation of the RNA tidine in all ubiquitin USPs (Figs. 3a, 4 and Supplementary Fig. 1), polymerase-II-mediated snRNA transcription in the Cajal Bodies and both forming hydrogen bonds with the C-terminal tail (CBs), a membrane-less compartment in the nucleus where backbone. While removal of this hydrogen bond interaction is USPL1 has been localized. However, the essential function of crucial in ubiquitin USPs15, binding and catalysis of USPL1 in the CBs biology seems to be independent of the SUMO SUMO2 substrates with the Y451F mutant on USPL1 is reduced protease activity, perhaps the long-disordered extensions of by approximately fourfold compared to wild-type (Fig. 6). Both USPL1 participate in protein–protein interactions of this residues, F457 and Y451, are partly responsible for the different membrane-less compartment. In any case, the relevance of the hydrogen bond networks established in the C-terminal tails of SUMO protease activity of USPL1 in the cellular context still SUMO and ubiquitin. needs to be disclosed. Probably in USPL1, an ultimate adaptation to bind SUMO2 In summary, we have unraveled the structural determinants for has occurred through specific changes in the fingers and thumb the unique specificity of USPL1 for SUMO2, a paradigmatic subdomains. Arg324 and Phe335 in the thumb subdomain, and example of divergent evolution in the ubiquitin USP family to Asn398 in the fingers subdomain, are engaged in specific inter- interact with a distant UbL family member. actions with SUMO2 residues. Interestingly, all three residues are highly conserved in all USPL1 orthologs and their substitution by Methods alanine seriously compromises the binding and catalytic activity Plasmids, cloning and point mutation. HA-USPL1-pcDNA3.1 was a gift from of SUMO2 substrates (Fig. 6). These residues have no equivalent Frauke Melchior (Addgene plasmid #85760; http://n2t.net/addgene:85760; in any ubiquitin-specific USPs, basically, as a consequence of the RRID:Addgene_85760)18. The catalytic domain construct pET28a-USPL1CD was amplified by PCR using Phusion polymerase and cloned into the BamHI/NotI low sequence identity in those subdomains (Fig. 3), resulting in restriction enzymes sites of pET28a vector using ligation. The USPL1 point an ~10–15° rotation of the SUMO globular domain with respect mutants constructs were designed by different primers and were created by the to ubiquitin in the complex. As a consequence, the SUMO QuickChange site-directed mutagenesis kit (Stratagene). All primers are shown in interface with the fingers and thumb subdomains does not share Supplementary Table 1. pET28a-Δ14-human SUMO2 (14 amino acids deletion) was constructed at the Sloan-Kettering Institute in New York by David Reverter. any similar interaction compared to the other ubiquitin-specific The plasmid of pET28a-Δ14-SUMO2GCVY(C48S) has been generated by USP complexes, which involve different contacts at these Restriction Enzyme Free PCR39. subdomains. In this manuscript, we want to emphasize the plasticity of the Protein expression and purification. The USPL1 CD and Δ14SUMO2 expression conserved catalytic domain of USPs, which despite being constructs were transformed into E. coli Rosetta (DE3) cells (Novagen). Bacteria designed to bind ubiquitin with high specificity, has evolved to were grown at 37 °C to OD600 = 0.7~0.8, and IPTG was added to a final con- centration of 0.5 mM. Bacteria were grown an additional 16 h at 20 °C and har- interact with SUMO in USPL1. SUMO and ubiquitin share a 16% vested by centrifugation. Cell suspensions were equilibrated in 350 mM NaCl, sequence identity, basically observed in the structural elements of 20 mM Tris-HCl (pH 8.0), 10 mM imidazole, 20% sucrose, 1 mM DTT, and 0.1% the globular domain. However, the non-conserved residues in IGEPAL CA-630, and cells were broken by sonication. After removing cell debris their surface participate in unique and specific protein–protein by centrifugation, proteins were separated from lysate by nickel affinity chroma- tography using Ni Sepharose 6 Fast Flow (GE Healthcare) and eluted with lysis interactions with USP family members. USPL1 is a paradigmatic buffer including 20 mM Tris-HCl (pH 8.0), 350 mM NaCl, 300 mM imidazole, and example of divergent evolution in the USP family, in which the 1 mM DTT. Fractions containing the target protein were collected, diluted to aforementioned surface substitutions facilitate the interaction 50 mM NaCl, applied to an anion exchange resin (Resource Q; GE Healthcare), and with SUMO but maintain the right hand-like subdomain scaffold eluted with a 0–1 M NaCl gradient from 0 to 35% in 20 mM Tris-HCl (pH 8.0) and of the catalytic domain. 3 mM DTT. Concentrated the protein using Amicon Ultra-30K ultrafiltration device (Millipore) and snap-frozen in liquid nitrogen prior to storage at −80 °C. It is also interesting to compare the SUMO interface between USPL1 and members of the SUMO-specific SENP/ULP protease Preparation of the USPL1CD- Δ14-SUMO2 DHA complex. After purification of family. Interestingly, in this family, all members are specific for Δ14-SUMO2GCVY (C48S) protein, 2 mM DTT was added as a solid to a 500 µl SUMO except SENP8/DEN1, which has proteolytic activity for aliquot of protein solution to reduce any contaminant disulfide and gently shaking Nedd8-conjugated substrates21–24. The total interface with 15 min. The buffer was changed to 20 mM Tris8, 150 mM NaCl by PD-10 (GE SUMO2 is quite large in both cases, 1169 and 1070 A2 for USPL1 Healthcare) and concentrated to 10 mg/ml and kept at room temperature for the next step. A stock solution of 2,5-dibromo hexanediamide (DHA) was prepared by and SENP2 (PDB 2IO0)38, respectively. However, excluding the dissolving 35.5 mg in 418μl DMF. A 10-fold molar DHA (25 µmol) was added into particular contacts with the extended C-terminal tail of SUMO2 Δ14-SUMO2GCVY (2.5 µmol). The reaction mixture was incubated at room in each family, almost no similarities are observed at the interface temperature for 30 min and then at 37 °C for 10 h. The insoluble dibromide was with the globular domain of SUMO. As an example, in a similar removed using centrifugation and further purified using a Resource Q column to get Δ14-Sumo2 DHA protein. The mixture of USPL1CD and Δ14-Sumo2 DHA interface region in both USPL1 and SENP2, Arg59, Arg61, and (1:3 molar ratio) was incubated at 30 °C for 3 h. Anion exchange chromatography Phe87 are engaged in salt bridge and hydrophobic contacts with (Resource Q; GE Healthcare) and gel filtration chromatography (Superdex 75; GE SENP238. However, in USPL1, those salt bridges are not present, Healthcare) were carried out to purify the USPL1CD-Δ14SUMO2DHA complex. but the two arginines, together with Phe87, form a hydrophobic pocket that interacts with Phe330 on USPL1, which is essential Crystallization and data collection. The complex USPL1CD-Δ14SUMO2DHA for binding and inhibition (Figs. 5a and 6). was concentrated to 8 mg/mL for crystallization screening in a buffer containing The catalytic domain of USPL1 is embedded in the middle of a 20 mM Tris 8.0, 170 mM NaCl and 1 mM DTT. Crystallization experiments were performed at 18 °C by the sitting drop vapor diffusion method and crystals grew up structurally disordered full-length protein, which does not appear in a protein mixture with an equal volume of a condition solution containing 0.2 M to have any other obvious well-folded domain in the sequence, as potassium thiocyanate, 0.1 M sodium acetate 5.5, 8% w/v PEG20000, and 8% w/v seen in the predicted model from Alphafold-232. This type of PEG500MME. Crystals were harvested after 3 days and soaked 5–10 s in the domain organization of the full-length protein is often present in crystallization buffer supplemented with 15% ethylene glycol, and then snap-frozen in liquid nitrogen to storage. the SENP/ULP protease family, in which long-disordered protein Diffraction data were collected to 1.8 Å resolution at beamline ID30B at the extensions are usually found adjacent to the catalytic domain10. ESRF (Grenoble, France). Data processing was conducted by AutoProcessing with As for the SENP/ULP family, we cannot rule out a functional role MxCUBE40,41. The space group was P21 and there was one complex per 10 NATURE COMMUNICATIONS | (2022)13:1819 | https://doi.org/10.1038/s41467-022-29485-0 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-022-29485-0 ARTICLE asymmetric unit. The structure of USPL1CD-Δ14SUMO2DHA was solved by 13. Amerik, A. Y. & Hochstrasser, M. Mechanism and function of molecular replacement with USPL1 Alphafold2 model as a search mode32. deubiquitinating enzymes. Biochim. Biophys. Acta 1695, 189–207 (2004). Following rounds of model building and refinement were carried out with Coot 14. Clague, M. J., Urbé, S. & Komander, D. Breaking the chains: deubiquitylating and Phenix42,43. The structure of USPL1CD-SUMO2 has been deposited in the enzyme specificity begets function. Nat. Rev. Mol. Cell Biol. 20, 338–352 Protein Data Bank under accession codes PDB 7P99. (2019). 15. Hu, M. et al. Crystal structure of a UBP-family deubiquitinating enzyme in SUMO-AMC hydrolysis assays. USPL1 wild type and mutants were incubated isolation and in complex with ubiquitin aldehyde. Cell 111, 1041–1054 (2002). with ubiquitin-, SUMO1- or SUMO2-AMC at 30 °C and measured the fluorescence 16. Mevissen, T. E. T. & Komander, D. Mechanisms of Deubiquitinase Specificity emission using 345 nm excitation and 445 nm emission wavelengths using a Jasco and Regulation. Annu. Rev. Biochem. 86, 159–192 (2017). FP-8200 spectrofluorometer. All measurements were carried out in triplicate with 17. Basters, A. et al. Structural basis of the specificity of USP18 toward ISG15. Nat. 1 nM USPL1 and 0.1 μM SUMO-AMC in a buffer containing 100 mM NaCl, Struct. Mol. Biol. 24, 270–278 (2017). 20 mM Tris-HCl pH 8, 10 mM DTT. 18. Schulz, S. et al. Ubiquitin-specific protease-like 1 (USPL1) is a SUMO isopeptidase with essential, non-catalytic functions. EMBO Rep. 13, 930–938 (2012). In vitro de-SUMOyation assays. Protease activity was measured by incubating di- 19. Li, S. J. & Hochstrasser, M. A new protease required for cell-cycle progression SUMO2 and NΔ419RanGAP1-SUMO2 (1 μm) with purified 20 nM of USPL1 wt in yeast. Nature 398, 246–251 (1999). and mutants at 37 °C in a buffer containing 40 mM Tris-HCl (pH 8.0), 250 mM NaCl, and 2 mM DTT. Reactions were stopped after 0, 5, 10, and 20 min with SDS- 20. Shin, E. J. et al. DeSUMOylating isopeptidase: a second class of SUMO BME loading buffer and analyzed by gel electrophoresis (PAGE). Proteins were protease. EMBO Rep. 13, 339–346 (2012). detected by staining with SYPRO Ruby protein gel stain (Invitrogen). Products 21. Gan-Erdene, T. et al. Identification and characterization of DEN1, a were detected using a Gel-Doc XR system (BioRad) and quantified with ImageLab deneddylase of the ULP family. J. Biol. Chem. 278, 28892–28900 (2003). software (Bio-Rad). The fraction of analyzed bands were plotted as mean ± SD. 22. Reverter, D. et al. Structure of a complex between Nedd8 and the Ulp/Senp Statistical analyses were performed with Graphpad Prism 5.0. Significances were protease family member Den1. J. Mol. Biol. 345, 141–151 (2005). measured by a two-tailed unpaired t-test relative to wild type. All data were ana- 23. Mendoza, H. M. et al. NEDP1, a highly conserved cysteine protease that lyzed with a 95% confidence interval. deNEDDylates cullins. J. Biol. Chem. 278, 25637–25643 (2003). 24. Wu, K. et al. DEN1 is a dual function protease capable of processing the C terminus of Nedd8 and deconjugating hyper-neddylated CUL1. J. Biol. Chem. Intrinsic fluorescence measurements. Intrinsic fluorescence spectra were 278, 28882–28891 (2003). recorded using a Jasco FP-8200 spectrofluorometer. Tryptophan emission spectra 25. Hutten, S., Chachami, G., Winter, U., Melchior, F. & Lamond, A. I. A role for were obtained by setting the excitation wavelength at 295 nm and collecting the Cajal-body-associated SUMO isopeptidase USPL1 in snRNA transcription emission in the 315–400 nm range. USPL1 (wild-type and mutants) were diluted to mediated by RNA polymerase II. J. Cell Sci. 127, 1065–1078 (2014). achieve 1 μM in 20 mM HEPES pH 7.5 buffer containing 250 mM NaCl and 26. Hebert, M. D. & Poole, A. R. Towards an understanding of regulating Cajal 10 mM DTT. The temperature was set at 30 °C. For thermostability at 30 °C pro- teins were kept at this temperature for up to 60 min and Trp spectra were recorded body activity by protein modification. RNA Biol. 14, 761–778 (2017). at different times. 27. Bermejo, J. L. et al. Exploring the association between genetic variation in the SUMO isopeptidase gene USPL1 and breast cancer through integration of data from the population-based GENICA study and external genetic databases. Int. Reporting summary. Further information on research design is available in the Nature J. Cancer 133, 362–372 (2013). Research Reporting Summary linked to this article. 28. Gao, H., Wang, H. & Yang, W. Identification of key genes and construction of microRNA-mRNA regulatory networks in multiple myeloma by integrated Data availability multiple GEO datasets using bioinformatics analysis. Int. J. Hematol. 106, The structure reported has been deposited in the Protein Data Bank under accession code 99–107 (2017). 7P99. Other Protein Data Bank accession codes used in this study: 6HEK (USP28- 29. Chalker, J. M. et al. Methods for converting cysteine to dehydroalanine on ubiquitin); 2HD5 (USP2-ubiquitin); 5JTV (USP7-ubiquitin); 5OHK (USP30-ubiquitin). peptides and proteins. Chem. Sci. 2, 1666–1676 (2011). All other data supporting the findings of this study are available within the article and its 30. Mossessova, E. & Lima, C. D. Ulp1-SUMO crystal structure and genetic supplementary information files. Source data are provided with this paper. analysis reveal conserved interactions and a regulatory element essential for cell growth in yeast. Mol. Cell 5, 865–876 (2000). 31. Reverter, D. & Lima, C. D. A basis for SUMO protease specificity provided by Received: 24 November 2021; Accepted: 17 March 2022; analysis of human Senp2 and a Senp2-SUMO complex. Structure 12, 1519–1531 (2004). 32. Jumper, J. et al. Highly accurate protein structure prediction with AlphaFold. Nature https://doi.org/10.1038/s41586-021-03819-2 (2021). 33. Ye, Y. et al. Polyubiquitin binding and cross-reactivity in the USP domain deubiquitinase USP21. EMBO Rep. 12, 350–357 (2011). References 34. Ward, S. J. et al. The structure of the deubiquitinase USP15 reveals a 1. Flotho, A. & Melchior, F. Sumoylation: a regulatory protein modification in misaligned catalytic triad and an open ubiquitin-binding channel. J. Biol. health and disease. Annu. Rev. Biochem. 82, 357–385 (2013). Chem. 293, 17362–17374 (2018). 2. van der Veen, A. G. & Ploegh, H. L. Ubiquitin-like proteins. Annu. Rev. 35. Gersch, M. et al. Distinct USP25 and USP28 oligomerization states regulate Biochem. 81, 323–357 (2012). deubiquitinating activity. Mol. Cell 74, 436–451.e7 (2019). 3. Hershko, A. & Ciechanover, A. The ubiquitin system. Annu. Rev. Biochem. 67, 36. Alegre, K. O. & Reverter, D. Swapping small ubiquitin-like modifier (SUMO) 425–479 (1998). isoform specificity of SUMO proteases SENP6 and SENP7. J. Biol. Chem. 286, 4. Jentsch, S. & Psakhye, I. Control of nuclear activities by substrate-selective and 36142–36151 (2011). protein-group SUMOylation. Annu. Rev. Genet. 47, 167–186 (2013). 37. Alegre, K. O. & Reverter, D. Structural insights into the SENP6 5. García-Rodríguez, N., Wong, R. P. & Ulrich, H. D. Functions of ubiquitin and Loop1 structure in complex with SUMO2. Protein Sci. 23, 433–441 (2014). SUMO in DNA replication and replication stress. Front. Genet. 7, 87 (2016). 38. Reverter, D. & Lima, C. D. Structural basis for SENP2 protease interactions 6. Seeler, J.-S. & Dejean, A. SUMO and the robustness of cancer. Nat. Rev. with SUMO precursors and conjugated substrates. Nat. Struct. Mol. Biol. 13, Cancer 17, 184–197 (2017). 1060–1068 (2006). 7. Varejão, N., Lascorz, J., Li, Y. & Reverter, D. Molecular mechanisms in SUMO 39. van den Ent, F. & Löwe, J. RF cloning: a restriction-free method for inserting conjugation. Biochem. Soc. Trans. 48, 123–135 (2020). target genes into plasmids. J. Biochem. Biophys. Methods 67, 67–74 (2006). 8. Cappadocia, L. & Lima, C. D. Ubiquitin-like protein conjugation: structures, 40. Monaco, S. et al. Automatic processing of macromolecular crystallography chemistry, and mechanism. Chem. Rev. 118, 889–918 (2018). X-ray diffraction data at the ESRF. J. Appl. Crystallogr. 46, 804–810 (2013). 9. Komander, D., Clague, M. J. & Urbé, S. Breaking the chains: structure and 41. Gabadinho, J. et al. MxCuBE: a synchrotron beamline control environment function of the deubiquitinases. Nat. Rev. Mol. Cell Biol. 10, 550–563 (2009). customized for macromolecular crystallography experiments. J. Synchrotron 10. Hickey, C. M., Wilson, N. R. & Hochstrasser, M. Function and regulation of Radiat. 17, 700–707 (2010). SUMO proteases. Nat. Rev. Mol. Cell Biol. 13, 755–766 (2012). 42. Emsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. Features and development 11. Li, Y. & Reverter, D. Molecular mechanisms of DUBs regulation in signaling of Coot. Acta Crystallogr. D. Biol. Crystallogr. 66, 486–501 (2010). and disease. Int. J. Mol. Sci. 22, 986 (2021). 43. Adams, P. D. et al. PHENIX: a comprehensive Python-based system for 12. Haahr, P. et al. ZUFSP deubiquitylates K63-linked polyubiquitin chains to macromolecular structure solution. Acta Crystallogr. D. Biol. Crystallogr. 66, promote genome stability. Mol. Cell 70, 165–174.e6 (2018). 213–221 (2010). NATURE COMMUNICATIONS | (2022)13:1819 | https://doi.org/10.1038/s41467-022-29485-0 | www.nature.com/naturecommunications 11 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-022-29485-0 44. Pfoh, R. et al. Crystal structure of USP7 ubiquitin-like domains with an ICP0 Additional information peptide reveals a novel mechanism used by viral and cellular proteins to target Supplementary information The online version contains supplementary material USP7. PLoS Pathog. 11, e1004950 (2015). available at https://doi.org/10.1038/s41467-022-29485-0. 45. Schrödinger, L. and W. D. PyMOL. http://www.pymol.org/pymol. (2020). 46. Renatus, M. et al. Structural basis of ubiquitin recognition by the Correspondence and requests for materials should be addressed to David Reverter. deubiquitinating protease USP2. Structure 14, 1293–1302 (2006). 47. Rougé, L. et al. Molecular understanding of USP7 substrate recognition and Peer review information Nature Communications thanks the anonymous reviewers for C-terminal activation. Structure 24, 1335–1345 (2016). their contribution to the peer review of this work. Peer reviewer reports are available. 48. Gersch, M. et al. Mechanism and regulation of the Lys6-selective deubiquitinase USP30. Nat. Struct. Mol. Biol. 24, 920–930 (2017). Reprints and permission information is available at http://www.nature.com/reprints Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Acknowledgements This work was supported by grants from the “Ministerio de Ciencia, Innovación y Universidades” PGC2018-098423-B-I00 to D.R. Y.L. acknowledges her scholarship of the China Scholarship Council program from the Chinese government. D.R. acknowledges Open Access This article is licensed under a Creative Commons support from the Serra Hunter program from Generalitat de Catalunya. We acknowledge Attribution 4.0 International License, which permits use, sharing, the European Synchrotron Radiation Facility for the provision of synchrotron radiation adaptation, distribution and reproduction in any medium or format, as long as you give facilities, and we would like to thank Andrew McCarthy for assistance in using appropriate credit to the original author(s) and the source, provide a link to the Creative beamline ID30B. Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the Author contributions article’s Creative Commons license and your intended use is not permitted by statutory Y.L. conducted the crystallization experiments. Y.L. and N.V. conducted in vitro activity regulation or exceeds the permitted use, you will need to obtain permission directly from assays. Y.L., N.V. and D.R. contributed to the correction and writing of the manuscript. the copyright holder. To view a copy of this license, visit http://creativecommons.org/ licenses/by/4.0/. Competing interests The authors declare no competing interests. © The Author(s) 2022 12 NATURE COMMUNICATIONS | (2022)13:1819 | https://doi.org/10.1038/s41467-022-29485-0 | www.nature.com/naturecommunications