Whereas conjugation with ubiquitin targets proteins for degradation, conjugation with small ubiquitin-related modifiers (SUMOs), or SUMOylation, is involved in various cellular processes and required for normal growth and development in all eukaryotes (Johnson, 2004). Although there is only 18% sequence identity between ubiquitin and SUMO1 (the founding member of the SUMO family [Okura et al., 1996]), their structured domains share a common fold known as β-grasp, in which five β-strands wrap around an α-helix (Bayer et al., 1998) (Figure 1). SUMOylation, like ubiquitination, is through an isopeptide linkage between a conserved Gly-Gly motif at the C-terminus of SUMOs and a lysine sidechain on target proteins. All SUMOs are translated as precursors that are first processed by the SENP family of SUMO-specific proteases, whereby a short C-terminal extension is cleaved to expose the reactive Gly-Gly motif. Like the much better-known posttranslational modification phosphorylation, SUMOylation is reversible, and de-SUMOylation is also catalyzed by SENPs (Yeh, 2009). The catalytic activity of SENPs on short peptide substrates is low, likely to avoid off-target cleavage, and is substantially enhanced by the β-grasp domain of SUMOs (Mikolajczyk et al., 2007). The mechanism of this allosteric activation has remained poorly understood. The aim of the present study was to gain insight into the allosteric mechanism through extensive molecular dynamics (MD) simulations.

Figure 1

Download asset Open asset

The human genome encodes three SUMOs (SUMO1, SUMO2, and SUMO3) and six SUMO-specific SENPs (SENP1, SENP2, SENP3, SENP5, SENP6, and SENP7). SENPs belong to the cysteine protease superfamily, and regulate both physiological and pathological processes mediated by SUMOylated proteins (Yeh, 2009). As an example, hypoxia (i.e., low oxygen), a condition common to fetal development and tumorigenesis, induces the nuclear entry and SUMOylation of hypoxia-inducible factor 1α (HIF1α) (Cheng et al., 2007). SUMOylated HIF1α is subject to ubiquitin-dependent degradation, and hence deSUMOylation of HIF1α by SENP1 is central to hypoxia response. For this and other reasons SENP1 is overexpressed during the development of prostate cancer, and has emerged as a potential therapeutic target (Cheng et al., 2006; Bawa-Khalfe et al., 2010; Wang et al., 2013). Several types of SENP inhibitors have been identified, including SUMO variants or C-terminal fragments tethered with electrophilic traps (Hemelaar et al., 2004; Borodovsky et al., 2005; Dobrota et al., 2012) and small molecules (Albrow et al., 2011; Ponder et al., 2011; Chen et al., 2012; Madu et al., 2013). The therapeutic value of these inhibitors is limited by covalent linkage with the catalytic cysteine residue, low specificity, or low potency. Better understanding of the mechanism for the allosteric activation of SENPs may lead to new avenues for drug development.

Structural studies have shown that SENP catalytic domains only undergo localized conformational rearrangements upon binding SUMOs (as precursors, processed products, or SUMO conjugates), and that SENP-SUMO interactions are largely similar during precursor processing and deSUMOylation (Reverter and Lima, 2004, 2006; Shen et al., 2006a, 2006b; Xu et al., 2006; Alegre and Reverter, 2014). In SENP1, the catalytic triad comprises residues His533, Asp550, and Cys603 (Figure 1). The catalytic domain (residues 419–644) consists of eight α-helices and seven β-strands, divided into two subdomains. The lower subdomain contains the α1, α2, and α8 helices and the β1-β2 hairpin, while the upper subdomain contains the α4-α6 helices and the β3-β7 sheet. The two remaining helices, α3 and α7, glue the two subdomains and also form the bottom of a channel where the extended C-terminus of SUMO1 lies and the catalytic triad is situated. Three loops, including residues 464–466 between α2 and α3, residues 530–533 between β4 and β5, and residues 599–602 between β7 and α7, line the channel and likely undergo transient motions during the entrance of the SUMO C-terminus into the channel. Hereafter these loops are referred to as loopA, loopB, and loopC, respectively. One tryptophan residue, Trp534, floors the substrate while another, Trp465, forms a lid over its conserved Gly-Gly motif. The burial of the catalytic center in the channel and consequently the transient motions around the channel necessary for proper alignment of the substrate with respect to the catalytic center may contribute to the low catalytic activity of SENPs on short peptides.

The β-grasp domain of SUMO1 binds into a large cleft to the side (hereafter the exosite), and makes separate contacts with the two subdomains of SENP1: the β1-β2 hairpin in the lower subdomain and the α4-α5 helices in the upper subdomain (Figure 1). An insertion in the β1-β2 hairpin of SENP2 that extended the interface with SUMO2 resulted in no other change in the structure of the catalytic domain but nevertheless increased its catalytic activity (Alegre and Reverter, 2014). Given that SENPs generally lack major conformational changes upon binding SUMOs, there is no simple explanation for the allosteric activation of SENPs by the β-grasp domain of SUMOs. In this context, we note that, in addition to conformational changes, the possibility that changes in conformational dynamics can mediate allosteric effects has received attention in the recent literature (Guo and Zhou, 2016).

Similar to SENPs, deubiquitinating enzymes (DUBs) in the cysteine protease class also have very low catalytic activity on short peptide substrates (Dang et al., 1998). Although the catalytic domains of DUBs in different families have different structural scaffolds, they all contain a catalytic channel that bears resemblance to that in SENPs (Hu et al., 2002; Misaghi et al., 2005; Messick et al., 2008). The resemblance further extends to the positioning of the ubiquitin domain in the exosite. It may thus be expected that allosteric activation mechanisms of SENPs and DUBs share certain similarities.

A major step toward elucidating the allosteric activation of SENP1 by the β-grasp domain of SUMO1 was taken in a recent study (Chen et al., 2014). These authors confirmed the ability of a pre-bound β-grasp domain, as contained in a truncated SUMO1 construct (residues 1–92, hereafter referred to as trunSUMO1), to enhance the SENP1 catalytic activity on a short peptide substrate, and further traced the enhancement solely to an increase in k_cat. Their NMR data showed that trunSUMO1 binding yielded a gradient in backbone amide chemical shift perturbations (CSPs) emanating from the exosite cleft to the catalytic center. Significant sidechain methyl CSPs were also observed on six residues that dotted the region from the exosite cleft to the catalytic center, which indicated to the authors that the allosteric effect was propagated through the hydrophobic core. Their Carr-Purcell-Meiboom-Gill (CPMG) relaxation dispersion data showed that trunSUMO1 binding enhanced microsecond-millisecond (μs-ms) dynamics for residues around the catalytic center, suggesting a realignment of these residues. Chen et al.’s NMR data provided residue-level information, but their sparseness precluded a full picture on how the allosteric communication occurred in SENP1 upon binding trunSUMO1.

Here we present an atomistic picture for the allosteric communication derived from MD simulations. The simulations showed that trunSUMO1 binding induces significant conformational and dynamic changes in SENP1, including widening of the exosite cleft and quenching of nanosecond (ns) dynamics in all but a distal region. Calculated backbone amide and sidechain methyl CSPs are in broad agreement with the experimental data of Chen et al. (2014) but more pronounced, and our sidechain CSPs clearly identify two hydrophobic pathways, each within a subdomain, for allosteric communication from the exosite cleft to the catalytic center. The β-grasp domain, by serving as a bridge that links the two exosite interface regions, strengthens intra- and inter-subdomain motional coupling, which in turn may be the underlying reason for the quenching of ns dynamics and potentially may also be the instigator of inter-subdomain μs-ms dynamics (Guo et al., 2015). The concerted action of the β-grasp induced conformational and dynamic changes is captured by a dock-and-coalesce mechanism for SENP-catalyzed SUMO cleavage, whereby the wedging of the β-grasp domain into the exosite cleft enables the docking of the proximal portion of the C-terminus, and the strengthened cross-channel coupling initiates inter-subdomain correlated motions to allow for the distal portion to coalesce around the catalytic center.

We carried out three replicate explicit-solvent MD simulations for apo SENP1, for SENP1 bound with trunSUMO1, and for SENP1 bound with the full-length SUMO1 precursor (referred to as preSUMO1 hereafter). For each system, the three simulations accumulated 1.7 μs of total time and are denoted as sim1, sim2, and sim3.

β-Grasp binding induces wider exosite cleft and stronger inter-subdomain contact

To compare the conformational sampling by the apo, trunSUMO1-bound, and preSUMO1-bound forms of SENP1, we carried out principal component analysis on the replicate simulations of the three systems. The distributions of conformations in the plane of the first two principal components (PC1 and PC2), translated to free energy surfaces according to the Boltzmann relation, are displayed in Figure 2—figure supplement 1a. The three systems cover overlapping as well as distinct regions in conformational space. The apo form samples disconnected free energy basins, in contrast to the two bound forms, indicating a decrease in flexibility upon β-grasp binding (see below).

Both PC1 and PC2 feature prominent displacements in the two exosite interface regions and in the loops lining the catalytic channel (Figure 2—figure supplement 1b,c). Relative to the apo form, the two bound forms move along positive PC1 and PC2, both of which involve the moving apart between the two interface regions and the opposite movements of loopA and loopB with respect to the β-grasp domain. We hence directly monitored the relative motions between the two interface regions and between the three channel-lining loops (Figure 2). The distributions of the cleft distance, defined as between the centers of heavy atoms of the interface residues 448–453 and 506–513 (shown as mauve in Figure 2a), are displayed in Figure 2b for the three forms of SENP1. The mean and standard deviation of the cleft distance change from 22.4 ± 1.5 Å for the apo form to 23.1 ± 0.5 Å and 22.6 ± 1.5 Å for the trunSUMO1-bound and preSUMO1-bound forms, respectively, indicating widening of the exosite cleft upon β-grasp binding (Figure 2—figure supplement 2a,b).

Figure 2 with 2 supplements see all

Download asset Open asset

To characterize the movements of the three channel-lining loops, we defined a coordinate system attached to the α3 and α7 helices, which form the bottom of the channel (Figure 2a). In the coordinate system, the z axis is along the helical axis of α7 (as defined by the vector from the Cα center of residues 612–615 to the Cα center of residues 604–607); the x axis goes through the Cα center of α3 residues 469–480 and hence points to the exosite cleft; and the y axis points into the upper subdomain. None of the three loops exhibits any overt movement along z, so we focus on the differences among the three forms of SENP1 in the distributions of the Cα center x and y coordinates of the three loops (Figure 2c and Figure 2—figure supplement 2a,b). Relative to the apo form, the two bound form show similar movements for both loopA and loopB, the former away from the exosite cleft (i.e., decreasing x) and toward the upper subdomain (i.e., increasing y) while the latter toward the cleft and away from the channel (i.e., increasing x and y); loopC moves in opposite directions in the two bound forms but perhaps that is accidental due to the long distance from the binding interface. The movement of loopA (toward α7) leads to stronger inter-subdomain contact; the x component of loopB’s movement has a similar effect, while the y component may create space for the docking of the proximal portion of a C-terminus when tethered to the β-grasp domain.

The stronger inter-subdomain contact induced by β-grasp binding can be further illustrated by the movement of the sidechain of the channel lid residue, W465 (Figure 2d). In apo SENP1 this sidechain samples a broad range of positions, but in preSUMO1-bound SENP1 it stays stably near the top of α7 (on the sub-μs timescale); the situation is intermediate in trunSUMO1-bound SENP1.

Nanosecond dynamics of SENP1 is quenched upon β-grasp binding

As noted above, the narrowing of accessible conformational regions upon β-grasp binding suggests a decrease in flexibility, i.e., quenching of ns dynamics. To compare the residue-specific flexibilities among the three forms of SENP1, we calculated their Cα root-mean-square fluctuations (RMSFs). Relative to apo SENP1, the two bound forms both show decreased flexibility throughout most of the amino acid sequence, except for the N- and C-terminal segments (Figure 3a).

Figure 3

Download asset Open asset

To visualize how the changes in RMSF are distributed spatially, we display them according to a color scale on the structures of the two bound forms (Figure 3b,c). In both systems, rigidification propagates from the exosite cleft to the entire upper subdomain and to most of the lower subdomain. The rigidification is compensated to some extent by higher flexibility in the distal region of the lower subdomain, comprising the N- and C-terminal segments. Therefore, the allosteric effects elicited by the β-grasp domain include both the stronger inter-subdomain contact and the quenching of ns dynamics in all but a distal region.

Calculated sidechain methyl CSPs identify two hydrophobic pathways for allosteric communication

Chen et al. (2014) measured the backbone amide CSPs of SENP1 upon trunSUMO1 binding. Their CSPs can be described as a gradient emanating from the exosite cleft to the catalytic center (Figure 4a), a pattern that is somewhat similar to that of the changes in RMSF (Figure 3b). In addition, they observed significant sidechain methyl CSPs on six residues, including Leu450, Val501, Val 509, V516, Val518, and ValV532, which sparsely occupy the region from the exosite cleft to the catalytic center (Figure 4a).

Figure 4

Download asset Open asset

To compare with these NMR results, we calculated the chemical shifts of backbone amides and sidechain methyls on the apo and trunSUMO1-bound SENP1 simulations. The resulting CSPs are displayed in Figure 4b. Overall, they agree well with the experimental data, but are more pronounced. The calculated backbone amide CSPs spread all the way to the back, in both subdomains. The calculation identifies three of the six experimentally detected residues as well as 12 others as having significant sidechain methyl CSPs (>0.08 ppm). Interestingly, with the 12 additional residues, two hydrophobic pathways emerge, each connecting an exosite interface region in one subdomain to the catalytic center. Chemical shifts are a measure of both conformational and dynamics effects. Hence the sidechain methyl CSPs are a manifestation of the conformational and dynamic changes induced by β-grasp binding, and the hydrophobic pathways may represent short paths over which allosteric effects are propagated from the exosite cleft to the catalytic center.

Strengthened intra- and inter-subdomain motional coupling underlies quenching of ns dynamics

Community analysis is a way to reveal the pattern of motional coupling within a protein, based on residue-residue physical contact and positional correlation during an MD simulation (Sethi et al., 2009). A protein structure is partitioned into communities, within which residues form dense contacts but between which residues form sparse contacts. The strength of coupling, or betweenness, between two communities is determined by the magnitudes of positional correlations within networks of contacting residues.

Results of our community analysis are displayed in Figure 5 for apo and trunSUMO1-bound sim1, and in Figure 5—figure supplement 1 and Figure 5—figure supplement 2 for the other two replicate simulations. Very similar community structures are obtained from the replicate simulations of each system, so here we focus on sim1. In all cases, five major communities (numbered 1 to 5) can be recognized for SENP1, anchored by β1β2 and α2; α3; α4α5; α6 and various parts of β4–β7; and α7, respectively, and possibly containing other segments. In the apo form, in addition to these five major communities, four minor communities (numbered 1′, 2′, 4′, and 4″) are also formed. In the trunSUMO1-SENP1 complex, SENP1 is partitioned only to the five major communities, with the minor communities absorbed. In particular, communities 4, 4′, and 4″ in apo SENP1 coalesce into a single community upon trunSUMO1 binding.

Figure 5 with 2 supplements see all

Download asset Open asset

Moreover, the inter-community betweennesses are also strengthened, especially between the two communities, 2 and 5, anchored by the two central helices (α3 and α7, respectively) and community 1 in the lower subdomain and community 4 in the upper subdomain. This strengthened inter-community coupling is in line with the stronger inter-subdomain contact noted above. It comes about because the β-grasp domain couples strongly to both subdomains. Therefore, the β-grasp domain, by serving as a bridge linking the two exosite interface regions, strengthens the motional coupling both within and between the subdomains of SENP1. A similar bridging role was identified for a peptide in inducing inter-domain allosteric communication in Pin1 (Guo et al., 2015).

With the strengthened intra-subdomain coupling upon trunSUMO1 binding, it can be expected that allosteric communication from the exosite cleft to the catalytic center becomes more effective. The effectiveness of allosteric communication between two sites can be measured by the lengths, defined using the residue-residue positional correlations (i.e., stronger correlations correspond to shorter lengths), of paths connecting the sites. Indeed, the path lengths from the two exosite interface regions to the catalytic center are shorter in the trunSUMO1-bound form than in the apo form. This result is in line with the two hydrophobic pathways identified by sidechain methyl CSPs.

The simultaneous occurrence of strengthened intra- and inter-subdomain coupling and quenching of ns dynamics observed here on SENP1 upon binding an allosteric activator conforms to a pattern previously recognized from a number of other proteins (including Pin1) in which conformational dynamics plays a prominent role in mediating allosteric communication (Guo and Zhou, 2015). This pattern was explained by a dynamic model of allostery. According to this model, fast motions (e.g., those on a ns timescale) are uncorrelated; an allosteric activator strengths inter-community coupling, which in turn leads to quenching of fast motions. The strengthened coupling here comes about because the β-grasp domain serves as a bridge that links the two interfacial regions. The model further predicts another dynamic effect, i.e., slower (e.g., μs-ms), cross-community correlated motions are initiated. Indeed, Chen et al.’s CPMG data for residues around the catalytic center indicated enhanced μs-ms conformational exchange upon trunSUMO1 binding (Chen et al., 2014).

Allosteric effects can be reproduced by restraining the two exosite interface regions at an widened separation

The results presented above lead to the hypothesis that the conformational and dynamic effects of β-grasp binding are achieved through increasing the exosite cleft distance and then rigidly holding the two interface regions at the widened separation. As demonstrated previously for Pin1 (Guo et al., 2015), this type of hypotheses can be tested by restrained MD simulations. Specifically, we ran an apo simulation but restrained the two exosite interface regions (i.e., the β1β2 hairpin in the lower subdomain and the α4α5 helices in the upper subdomain) to their conformation from a snapshot in trunSUMO1-bound sim1, in which the exosite cleft distance is at a widened 23.9 Å (Figure 6a).

Figure 6 with 1 supplement see all

Download asset Open asset

The simulation of this control system reproduces the main conformational and dynamic effects of β-grasp binding. The three channel-lining loops in the control system move in essentially the same ways as those in the trunSUMO1-bound form (compare Figure 6—figure supplement 1a,b with Figure 2c and Figure 2—figure supplement 2b). The RMSF of the control system (Figure 6b) also shows a similar albeit slightly more pronounced decrease, indicating quenching of ns dynamics, for most of the SENP1 structure, including all the three channel-lining loops. Lastly the calculated backbone amide CSPs of the control system have a similar gradient to that of the trunSUMO1-bound form, and the calculated sidechain methyl CSPs again identify two hydrophobic pathways that may propagate allosteric effects from the exosite cleft to the catalytic center (Figure 6c).

Through extensive MD simulations, we have found that SUMO1 β-grasp binding to SENP1 induces significant conformational and dynamic effects, including the widening of the exosite cleft and the quenching of ns dynamics in all but a distal region. The calculated CSPs are in broad agreement with the NMR data of Chen et al. (2014) but our results overall present a deeper, atomistic picture for the allosteric communication from the exosite cleft to the catalytic center. We conclude that the wedging into the exosite cleft and the bridging between the two exosite interface regions by the β-grasp domain are the underlying reason for the widened exosite cleft and the quenched fast dynamics, and may also be the instigator of inter-subdomain slow motions. Enhanced μs-ms conformational exchange has indeed been observed in CPMG experiments on trunSUMO1-bound SENP1 (Chen et al., 2014). Our findings have broad implications for the mechanism of allosteric activation of SENPs by SUMOs, the SUMO paralogue specificity of SENPs, and the development of drugs targeting SENPs.

A dock-and-coalescence mechanism for SENP-catalyzed SUMO cleavage

The concerted action of the β-grasp induced conformational and dynamic changes is expected to help both the binding step and the subsequent catalytic step for trunSUMO1-bound SENP1 reacting with an isolated peptide substrate. First, widening of the exosite cleft may enhance the rate at which the substrate binds to the catalytic center, although this effect may not be detectable in an enzymatic assay if the binding does not rate-limit the overall enzymatic reaction. Second, initiation of slow conformational dynamics may facilitate the proper alignment of the substrate around the catalytic residues. These putative effects can explain the observation of Chen et al. (2014) that trunSUMO1 enhances the SENP1 catalytic activity on a peptide substrate through increasing k_cat.

In SENP-catalyzed SUMO cleavage for processing of SUMO precursors and de-SUMOylation of target proteins, the C-terminal substrate is tethered to the β-grasp domain via an extended linker. Tethering can also help both the binding step and the subsequent catalytic step. The binding of a substrate-containing SUMO (i.e., a precursor or SUMO conjugate) likely occurs in a sequential manner: the β-grasp domain first wedges into the exosite cleft and the C-terminus then docks into the catalytic channel. Tethering can facilitate this binding by correctly orienting the C-terminal substrate. This effect and the β-grasp induced widening of the exosite cleft together may speed up the docking of the C-terminus into the catalytic channel sufficiently as to make the full binding step rate-limited by the initial wedging of the β-grasp into the exosite. Tethering can also reinforce the motional coupling between the exosite interface regions and the channel-lining loops. This may in turn intensify the slow conformational dynamics that facilitates the proper alignment of the substrate around the catalytic residues, thereby accelerating the catalytic step.

In previous studies we have put forward a docking-and-coalesce mechanism to describe the binding of intrinsically disordered proteins, which, similar to the disordered SUMO C-terminus, usually form extended conformations on their targets (Qin et al., 2011; Zhou et al., 2012). Here we adapt this mechanism to describe the steps of SENP-catalyzed SUMO cleavage reaction just before the actual cut of the (iso)peptide bond (Figure 7). The process starts with the wedging of the SUMO β-grasp domain into the SENP exosite cleft. This wedging widens the separation between and also tightly links the two exosite interface regions. The widening of the exosite cleft enables the docking of the proximal portion of the C-terminus into the catalytic channel. Moreover, the wedged β-grasp domain strengthens the motional coupling within and between the two subdomains of SENP, and the newly docked proximal portion of the SUMO C-terminus reinforces this coupling. According to a theoretical model (Guo and Zhou, 2015), the strengthened coupling sets up the condition for the emergence of inter-subdomain correlated motions, which finally allow for the distal portion of the SUMO C-terminus to align properly around the catalytic center for bond cleavage.

Figure 7

Download asset Open asset

The preceding dock-and-coalesce mechanism differs from the previous generic version (Qin et al., 2011; Zhou et al., 2012) by the cooperation between a structure domain (i.e., the β-grasp domain) and a disordered region (i.e., the C-terminus) of the substrate protein and by the prominent role of allosteric communication within the target protein. In addition, the end result here is the transition-state complex for a catalytic reaction as opposed to a ground-state complex. Certain aspects of the mechanism presented here have been suggested previously by Shen et al. (2006a), including the stimulation of catalytic channel opening by SUMO binding and the facilitation of substrate alignment by cross-channel dynamics, without elaboration. Lastly the dock-and-coalesce mechanism may not only provide a qualitative description of the individual steps but also serve as a framework for quantitative calculations of kinetic rate constants (see next).

1. 1. Albrow VE
   2. Ponder EL
   3. Fasci D
   4. Békés M
   5. Deu E
   6. Salvesen GS
   7. Bogyo M

   (2011) Development of small molecule inhibitors and probes of human SUMO deconjugating proteases

   Chemistry & Biology 18:722–732.

   https://doi.org/10.1016/j.chembiol.2011.05.008

   • Google Scholar
2. 1. Alegre KO
   2. Reverter D

   (2014) Structural insights into the SENP6 Loop1 structure in complex with SUMO2

   Protein Science 23:433–441.

   https://doi.org/10.1002/pro.2425

   • Google Scholar
3. 1. Bawa-Khalfe T
   2. Cheng J
   3. Lin SH
   4. Ittmann MM
   5. Yeh ET

   (2010) SENP1 induces prostatic intraepithelial neoplasia through multiple mechanisms

   Journal of Biological Chemistry 285:25859–25866.

   https://doi.org/10.1074/jbc.M110.134874

   • Google Scholar
4. 1. Bayer P
   2. Arndt A
   3. Metzger S
   4. Mahajan R
   5. Melchior F
   6. Jaenicke R
   7. Becker J

   (1998) Structure determination of the small ubiquitin-related modifier SUMO-1

   Journal of Molecular Biology 280:275–286.

   https://doi.org/10.1006/jmbi.1998.1839

   • Google Scholar
5. 1. Borodovsky A
   2. Ovaa H
   3. Meester WJ
   4. Venanzi ES
   5. Bogyo MS
   6. Hekking BG
   7. Ploegh HL
   8. Kessler BM
   9. Overkleeft HS

   (2005) Small-molecule inhibitors and probes for ubiquitin- and ubiquitin-like-specific proteases

   ChemBioChem 6:287–291.

   https://doi.org/10.1002/cbic.200400236

   • Google Scholar
6. 1. Brooks BR
   2. Brooks CL
   3. Mackerell AD
   4. Nilsson L
   5. Petrella RJ
   6. Roux B
   7. Won Y
   8. Archontis G
   9. Bartels C
   10. Boresch S
   11. Caflisch A
   12. Caves L
   13. Cui Q
   14. Dinner AR
   15. Feig M
   16. Fischer S
   17. Gao J
   18. Hodoscek M
   19. Im W
   20. Kuczera K
   21. Lazaridis T
   22. Ma J
   23. Ovchinnikov V
   24. Paci E
   25. Pastor RW
   26. Post CB
   27. Pu JZ
   28. Schaefer M
   29. Tidor B
   30. Venable RM
   31. Woodcock HL
   32. Wu X
   33. Yang W
   34. York DM
   35. Karplus M

   (2009) CHARMM: the biomolecular simulation program

   Journal of Computational Chemistry 30:1545–1614.

   https://doi.org/10.1002/jcc.21287

   • Google Scholar
7. 1. Chen CH
   2. Namanja AT
   3. Chen Y

   (2014) Conformational flexibility and changes underlying activation of the SUMO-specific protease SENP1 by remote substrate binding

   Nature Communications 5:4968.

   https://doi.org/10.1038/ncomms5968

   • Google Scholar
8. 1. Chen Y
   2. Wen D
   3. Huang Z
   4. Huang M
   5. Luo Y
   6. Liu B
   7. Lu H
   8. Wu Y
   9. Peng Y
   10. Zhang J

   (2012) 2-(4-Chlorophenyl)-2-oxoethyl 4-benzamidobenzoate derivatives, a novel class of SENP1 inhibitors: Virtual screening, synthesis and biological evaluation

   Bioorganic & Medicinal Chemistry Letters 22:6867–6870.

   https://doi.org/10.1016/j.bmcl.2012.09.037

   • Google Scholar
9. 1. Cheng J
   2. Bawa T
   3. Lee P
   4. Gong L
   5. Yeh ET

   (2006) Role of desumoylation in the development of prostate cancer

   Neoplasia 8:667–676.

   https://doi.org/10.1593/neo.06445

   • Google Scholar
10. 1. Cheng J
    2. Kang X
    3. Zhang S
    4. Yeh ET

    (2007) SUMO-specific protease 1 is essential for stabilization of HIF1alpha during hypoxia

    Cell 131:584–595.

    https://doi.org/10.1016/j.cell.2007.08.045

    • Google Scholar
11. 1. Dang LC
    2. Melandri FD
    3. Stein RL

    (1998) Kinetic and mechanistic studies on the hydrolysis of ubiquitin C-terminal 7-amido-4-methylcoumarin by deubiquitinating enzymes

    Biochemistry 37:1868–1879.

    https://doi.org/10.1021/bi9723360

    • Google Scholar
12. 1. Dobrotă C
    2. Fasci D
    3. Hădade ND
    4. Roiban GD
    5. Pop C
    6. Meier VM
    7. Dumitru I
    8. Matache M
    9. Salvesen GS
    10. Funeriu DP

    (2012) Glycine fluoromethylketones as SENP-specific activity based probes

    ChemBioChem 13:80–84.

    https://doi.org/10.1002/cbic.201100645

    • Google Scholar
13. 1. Guo J
    2. Pang X
    3. Zhou HX

    (2015) Two pathways mediate interdomain allosteric regulation in pin1

    Structure 23:237–247.

    https://doi.org/10.1016/j.str.2014.11.009

    • Google Scholar
14. 1. Guo J
    2. Zhou HX

    (2015) Dynamically driven protein allostery exhibits disparate responses for fast and slow motions

    Biophysical Journal 108:2771–2774.

    https://doi.org/10.1016/j.bpj.2015.04.035

    • Google Scholar
15. 1. Guo J
    2. Zhou HX

    (2016) Protein allostery and conformational dynamics

    Chemical Reviews 116:6503–6515.

    https://doi.org/10.1021/acs.chemrev.5b00590

    • Google Scholar
16. 1. Han B
    2. Liu Y
    3. Ginzinger SW
    4. Wishart DS

    (2011) SHIFTX2: significantly improved protein chemical shift prediction

    Journal of Biomolecular NMR 50:43–57.

    https://doi.org/10.1007/s10858-011-9478-4

    • Google Scholar
17. 1. Hemelaar J
    2. Borodovsky A
    3. Kessler BM
    4. Reverter D
    5. Cook J
    6. Kolli N
    7. Gan-Erdene T
    8. Wilkinson KD
    9. Gill G
    10. Lima CD
    11. Ploegh HL
    12. Ovaa H

    (2004) Specific and covalent targeting of conjugating and deconjugating enzymes of ubiquitin-like proteins

    Molecular and Cellular Biology 24:84–95.

    https://doi.org/10.1128/MCB.24.1.84-95.2004

    • Google Scholar
18. 1. Hickey CM
    2. Wilson NR
    3. Hochstrasser M

    (2012) Function and regulation of SUMO proteases

    Nature Reviews Molecular Cell Biology 13:755–766.

    https://doi.org/10.1038/nrm3478

    • Google Scholar
19. 1. Hornak V
    2. Abel R
    3. Okur A
    4. Strockbine B
    5. Roitberg A
    6. Simmerling C

    (2006) Comparison of multiple Amber force fields and development of improved protein backbone parameters

    Proteins 65:712–725.

    https://doi.org/10.1002/prot.21123

    • Google Scholar
20. 1. Hu M
    2. Li P
    3. Li M
    4. Li W
    5. Yao T
    6. Wu JW
    7. Gu W
    8. Cohen RE
    9. Shi Y

    (2002) Crystal structure of a UBP-family deubiquitinating enzyme in isolation and in complex with ubiquitin aldehyde

    Cell 111:1041–1054.

    https://doi.org/10.1016/S0092-8674(02)01199-6

    • Google Scholar
21. 1. Johnson ES

    (2004) Protein modification by SUMO

    Annual Review of Biochemistry 73:355–382.

    https://doi.org/10.1146/annurev.biochem.73.011303.074118

    • Google Scholar
22. 1. Kumar A
    2. Zhang KY

    (2013) Computational investigation of SENP:SUMO protein-protein interaction for structure based drug design

    Molecular Informatics 32:267–280.

    https://doi.org/10.1002/minf.201200124

    • Google Scholar
23. 1. Madu IG
    2. Namanja AT
    3. Su Y
    4. Wong S
    5. Li YJ
    6. Chen Y

    (2013) Identification and characterization of a new chemotype of noncovalent SENP inhibitors

    ACS Chemical Biology 8:1435–1441.

    https://doi.org/10.1021/cb400177q

    • Google Scholar
24. 1. Messick TE
    2. Russell NS
    3. Iwata AJ
    4. Sarachan KL
    5. Shiekhattar R
    6. Shanks JR
    7. Reyes-Turcu FE
    8. Wilkinson KD
    9. Marmorstein R

    (2008) Structural basis for ubiquitin recognition by the Otu1 ovarian tumor domain protein

    Journal of Biological Chemistry 283:11038–11049.

    https://doi.org/10.1074/jbc.M704398200

    • Google Scholar
25. 1. Mikolajczyk J
    2. Drag M
    3. Békés M
    4. Cao JT
    5. Ronai Z
    6. Salvesen GS

    (2007) Small ubiquitin-related modifier (SUMO)-specific proteases: profiling the specificities and activities of human SENPs

    Journal of Biological Chemistry 282:26217–26224.

    https://doi.org/10.1074/jbc.M702444200

    • Google Scholar
26. 1. Misaghi S
    2. Galardy PJ
    3. Meester WJ
    4. Ovaa H
    5. Ploegh HL
    6. Gaudet R

    (2005) Structure of the ubiquitin hydrolase UCH-L3 complexed with a suicide substrate

    Journal of Biological Chemistry 280:1512–1520.

    https://doi.org/10.1074/jbc.M410770200

    • Google Scholar
27. 1. Okura T
    2. Gong L
    3. Kamitani T
    4. Wada T
    5. Okura I
    6. Wei CF
    7. Chang HM
    8. Yeh ET

    (1996)

    Protection against Fas/APO-1- and tumor necrosis factor-mediated cell death by a novel protein, sentrin

    Journal of Immunology 157:4277–4281.

    • Google Scholar
28. 1. Phillips JC
    2. Braun R
    3. Wang W
    4. Gumbart J
    5. Tajkhorshid E
    6. Villa E
    7. Chipot C
    8. Skeel RD
    9. Kalé L
    10. Schulten K

    (2005) Scalable molecular dynamics with NAMD

    Journal of Computational Chemistry 26:1781–1802.

    https://doi.org/10.1002/jcc.20289

    • Google Scholar
29. 1. Ponder EL
    2. Albrow VE
    3. Leader BA
    4. Békés M
    5. Mikolajczyk J
    6. Fonović UP
    7. Shen A
    8. Drag M
    9. Xiao J
    10. Deu E
    11. Campbell AJ
    12. Powers JC
    13. Salvesen GS
    14. Bogyo M

    (2011) Functional characterization of a SUMO deconjugating protease of Plasmodium falciparum using newly identified small molecule inhibitors

    Chemistry & Biology 18:711–721.

    https://doi.org/10.1016/j.chembiol.2011.04.010

    • Google Scholar
30. 1. Qin S
    2. Pang X
    3. Zhou HX

    (2011) Automated prediction of protein association rate constants

    Structure 19:1744–1751.

    https://doi.org/10.1016/j.str.2011.10.015

    • Google Scholar
31. 1. Reverter D
    2. Lima CD

    (2004) A basis for SUMO protease specificity provided by analysis of human Senp2 and a Senp2-SUMO complex

    Structure 12:1519–1531.

    https://doi.org/10.1016/j.str.2004.05.023

    • Google Scholar
32. 1. Reverter D
    2. Lima CD

    (2006) Structural basis for SENP2 protease interactions with SUMO precursors and conjugated substrates

    Nature Structural & Molecular Biology 13:1060–1068.

    https://doi.org/10.1038/nsmb1168

    • Google Scholar
33. 1. Sethi A
    2. Eargle J
    3. Black AA
    4. Luthey-Schulten Z

    (2009) Dynamical networks in tRNA:protein complexes

    PNAS 106:6620–6625.

    https://doi.org/10.1073/pnas.0810961106

    • Google Scholar
34. 1. Shen L
    2. Tatham MH
    3. Dong C
    4. Zagórska A
    5. Naismith JH
    6. Hay RT

    (2006a) SUMO protease SENP1 induces isomerization of the scissile peptide bond

    Nature Structural & Molecular Biology 13:1069–1077.

    https://doi.org/10.1038/nsmb1172

    • Google Scholar
35. 1. Shen LN
    2. Dong C
    3. Liu H
    4. Naismith JH
    5. Hay RT

    (2006b) The structure of SENP1-SUMO-2 complex suggests a structural basis for discrimination between SUMO paralogues during processing

    Biochemical Journal 397:279–288.

    https://doi.org/10.1042/BJ20052030

    • Google Scholar
36. 1. Vanommeslaeghe K
    2. Hatcher E
    3. Acharya C
    4. Kundu S
    5. Zhong S
    6. Shim J
    7. Darian E
    8. Guvench O
    9. Lopes P
    10. Vorobyov I
    11. Mackerell AD

    (2010) CHARMM general force field: A force field for drug-like molecules compatible with the CHARMM all-atom additive biological force fields

    Journal of Computational Chemistry 31:671–690.

    https://doi.org/10.1002/jcc.21367

    • Google Scholar
37. 1. Wang Q
    2. Xia N
    3. Li T
    4. Xu Y
    5. Zou Y
    6. Zuo Y
    7. Fan Q
    8. Bawa-Khalfe T
    9. Yeh ET
    10. Cheng J

    (2013) SUMO-specific protease 1 promotes prostate cancer progression and metastasis

    Oncogene 32:2493–2498.

    https://doi.org/10.1038/onc.2012.250

    • Google Scholar
38. 1. Xu Z
    2. Chau SF
    3. Lam KH
    4. Chan HY
    5. Ng TB
    6. Au SW

    (2006) Crystal structure of the SENP1 mutant C603S-SUMO complex reveals the hydrolytic mechanism of SUMO-specific protease

    Biochemical Journal 398:345–352.

    https://doi.org/10.1042/BJ20060526

    • Google Scholar
39. 1. Yeh ET

    (2009) SUMOylation and De-SUMOylation: wrestling with life's processes

    Journal of Biological Chemistry 284:8223–8227.

    https://doi.org/10.1074/jbc.R800050200

    • Google Scholar
40. 1. Zhou HX
    2. Pang X
    3. Lu C

    (2012) Rate constants and mechanisms of intrinsically disordered proteins binding to structured targets

    Physical Chemistry Chemical Physics 14:10466–10476.

    https://doi.org/10.1039/c2cp41196b

    • Google Scholar

Author details

1. Jingjing Guo

   Henan Engineering Research Center of Chiral Hydroxyl Pharmaceuticals, School of Chemistry and Chemical Engineering, Henan Normal University, Xinxiang, China

   Contribution

   JG, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Competing interests

   The authors declare that no competing interests exist.

2. Huan-Xiang Zhou

   Department of Physics, Institute of Molecular Biophysics, Florida State University, Tallahassee, United States

   Contribution

   H-XZ, Conception and design, Analysis and interpretation of data, Drafting or revising the article

   For correspondence

   hzhou4@fsu.edu

   Competing interests

   The authors declare that no competing interests exist.

   "This ORCID iD identifies the author of this article:" 0000-0001-9020-0302

• 854

  views

• 232

  downloads

• 17

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.