Skip to main content
NCBI home page
Protein Science : A Publication of the Protein Society logoLink to Protein Science : A Publication of the Protein Society
. 2024 Feb 15;33(3):e4904. doi: 10.1002/pro.4904

Identification of small‐molecule binding sites of a ubiquitin‐conjugating enzyme‐UBE2T through fragment‐based screening

Yong Yao Loh 1, Jothi Anantharajan 1, Qiwei Huang 1, Weijun Xu 1, Justina Fulwood 1, Hui Qi Ng 1, Elizabeth Yihui Ng 1, Chong Yu Gea 1, Meng Ling Choong 1, Qian Wen Tan 1, Xiaoying Koh 1, Wan Hsin Lim 1, Kassoum Nacro 1, Joseph Cherian 1, Nithya Baburajendran 1,✉, Zhiyuan Ke 1,✉, CongBao Kang 1,✉
PMCID: PMC10868430  PMID: 38358126

Abstract

UBE2T is an attractive target for drug development due to its linkage with several types of cancers. However, the druggability of ubiquitin‐conjugating E2 (UBE2T) is low because of the lack of a deep and hydrophobic pocket capable of forming strong binding interactions with drug‐like small molecules. Here, we performed fragment screening using 19F‐nuclear magnetic resonance (NMR) and validated the hits with 1H‐15N‐heteronuclear single quantum coherence (HSQC) experiment and X‐ray crystallographic studies. The cocrystal structures obtained revealed the binding modes of the hit fragments and allowed for the characterization of the fragment‐binding sites. Further screening of structural analogues resulted in the identification of a compound series with inhibitory effect on UBE2T activity. Our current study has identified two new binding pockets in UBE2T, which will be useful for the development of small molecules to regulate the function of this protein. In addition, the compounds identified in this study can serve as chemical starting points for the development of UBE2T modulators.

Keywords: drug design, fragment‐based drug discovery, protein structure, UBE2T, ubiquitylation

1. INTRODUCTION

Protein ubiquitylation marks a protein for degradation and plays a critical role in immune response, development, and programmed cell death. It is a finely controlled process and its dysregulation can result in serious diseases such as cancer and neurodegenerative disorders (Kolas et al., 2007). Protein ubiquitylation is achieved primarily through the sequential action of E1, E2, and E3 enzymes (Hershko & Ciechanover, 1998; Scheffner et al., 1995). E1 enzymes are ubiquitin‐activating enzymes, which activate the C terminus of ubiquitin to generate a thioester‐linked E1‐ubiquitin conjugate in an ATP‐dependent manner. The activated ubiquitin is then transferred to the catalytic cysteine residue of a ubiquitin‐conjugating E2 enzyme (Keszei & Sicheri, 2017; Stewart et al., 2016). Finally, ubiquitin is then transferred to the side chain of lysine residues in a target protein through the action of a ubiquitin ligase E3 enzyme. This sequential transfer of ubiquitin involves specific protein–protein interactions to ensure the target protein is properly ubiquitylated, and any molecule capable of affecting the activities of these enzymes or disrupting their protein–protein interactions can have an effect on the ubiquitylation of the target protein, which will in turn affect corresponding downstream signal transduction in response to stimuli such as DNA damage (Jackson & Durocher, 2013; Popovic et al., 2014).

Out of the approximately 40 E2 enzymes in humans, UBE2T has emerged as a promising target for drug discovery because of its linkage with a variety of diseases. It is a key component of the Fanconi anemia (FA) pathway, where, along with the Really interesting new gene (RING)‐type E3 ligase FA complementation group L (FANCL), it is responsible for catalyzing the monoubiquitylation of the FANCI/FA group D2 protein (FANCD2) complex (Ma et al., 2023; Machida et al., 2006). The FA pathway is indispensable for the repair of DNA interstrand crosslinks (Ceccaldi et al., 2016; D'Andrea & Grompe, 2003), and its dysregulation has been linked to chemotherapy resistance, where cancer cells can overcome drug‐induced DNA interstrand crosslink formation. Therefore, regulating the activity of UBE2T can play a role in modulating the FA pathway to restore the function of the drug (Cornwell et al., 2019). In addition, UBE2T is an oncogenic protein that has been found to be overexpressed in several types of cancers (Tao et al., 2023; Yu et al., 2021; Zheng et al., 2020). It is critical for the initiation and progression of pancreatic cancer (Zheng et al., 2020) and has been demonstrated to promote the proliferation of renal cell carcinoma cells through regulating phosphoinositide 3‐kinase (PI3K)/protein kinase B (AKT) signaling (Hao et al., 2019). Suppression of UBE2T activity has also exhibited an inhibitory effect on the proliferation and invasion of ovarian cancer cells (Cui et al., 2022), lung adenocarcinoma (Li et al., 2022), and other cancers (Liu et al., 2017). Finally, UBE2T also plays important roles in primary cellular processes and participates in protein–protein interactions (Hodson et al., 2014; Lim et al., 2016; Longerich et al., 2014). Although the mechanism of action of UBE2T in various cancers remains to be confirmed, its correlation with cancer development highlights the potential of UBE2T as a target for many cancers. Therefore, inhibitors that can suppress its enzymatic activity or affect its interactions with other proteins have the potential to be developed as drugs against different diseases (Morreale et al., 2017; Yu et al., 2021).

Human UBE2T is composed of 197 amino acids and contains a core catalytic ubiquitin‐conjugating (UBC) domain of approximately 150 residues (Stewart et al., 2016; Valimberti et al., 2015). The structure of its UBC domain has been determined to be made up of an α/β‐fold with four helices and a β‐sheet formed by four strands. To date, several inhibitors of UBE2T have been identified through fragment‐based screening and high‐throughput screening (Anantharajan et al., 2023; Hodge et al., 2015; Morreale et al., 2017). These inhibitors either form a covalent bond with the cysteine residue in the active site or modulate UBE2T function through an allosteric mechanism (Figure S1). Despite this progress, further work is required for the development of more potent UBE2T inhibitors. In this study, we carried out fragment screening of a library using 19F‐NMR spectroscopy to identify (1) new chemical starting points for further development and (2) novel pockets on the protein surface suitable for binding to small molecules (Figure 1a). The binding of two fragments to UBE2T were characterized via X‐ray crystallographic studies, where their distinct binding pockets were identified. Further screening of analogues containing a pyrimidinone core was performed, resulting in the identification of a compound with an IC50 of 288.2 μM against UBE2T activity. Our current study has successfully identified two new binding pockets in UBE2T feasible for the binding of small molecules and provides a strategy to identify inhibitors through fragment screening.

FIGURE 1.

FIGURE 1

Open in a new tab

Fragment screening to identify UBE2T binders. (a) The process of fragment‐based drug discovery used in this study. (b) Confirmed hits for further studies. The dissociation constants (Kd) were determined using solution NMR spectroscopy. For some fragments, the Kd can not be accurately determined due to the small chemical shift or signal overlap. NA indicates that the accurate Kd cannot be obtained. LE, ligand efficiencies; MOA, mechanism of action.

2. RESULTS

2.1. Strategy used in fragment screening

Druggability analysis of UBE2T based on a previously published cocrystal structure (Morreale et al., 2017) indicated that it lacked a deep and hydrophobic pocket that favors binding to a drug‐like small molecule. The binding of a compound (EM04) to UBE2T did not induce the generation of a druggable pocket (Figure S2). Sitemap calculation (Halgren, 2009) suggested that the overall druggability score (DScore) is 0.708, which is considered to be likely undruggable as only DScore >1 is considered druggable. To further probe alternative drug‐binding sites to allow for the development of small molecules capable of regulating the enzymatic activity and the function of UBE2T, a fragment‐based drug discovery approach was applied. This strategy consisted of (1) fragment screening using 19F‐NMR spectroscopy, (2) hit confirmation using protein‐observed NMR spectroscopy, (3) determining hit binding mode using X‐ray crystallography, and (4) fragment growth through the screening of structural analogues of the fragment hits from a compound library composed of drug‐like molecules (Figure 1b). To facilitate this, a protein construct without the C‐terminal flexible region was expressed, purified, and referred to as UBE2T (Figure S3) (Huang et al., 2023). Purified UBE2T exhibited well‐dispersed cross peaks in the 1H‐15N‐HSQC spectrum, demonstrating its folding in solution (Figure S3).

2.2. Fragment screening using 19F‐NMR

19F‐NMR spectroscopy is a powerful tool for fragment screening, where the low molecular weight fragments binding weakly to a target can be identified (Ayotte et al., 2022; Kang, 2019). First, the wide chemical shift range of fluorinated molecules makes it possible to screen mixtures of compounds in a single experiment to increase the efficiency of screening (Buchholz & Pomerantz, 2021). Second, the high sensitivity of 19F‐NMR allows for the identification of ligands even if the binding to a target is weak. In this study, a 19F‐NMR screen of a library of 582 fluorinated fragments was carried out, and a total of 65 fragments were identified to interact with UBE2T (Figure S4). These hits were subjected to further confirmation using chemical shift perturbation (CSP) analysis by comparing the 1H‐15N HSQC spectra of 15N‐labeled UBE2T in the absence and presence of the individual fragments.

2.3. Fragments confirmation with 1H‐15N‐HSQC experiment

Comparison of the 1H‐15N‐HSQC spectra of UBE2T in the absence and presence of a fragment was used to confirm the hits from 19F‐NMR (Figures S5–S19). CSPs were caused by 15 fragments, which were then selected for cocrystal structural studies. These fragments bind weakly to UBE2T with the dissociation constants (Kd) in mM range (Figure 1b). For fragments where it was possible to obtain binding affinities, ligand efficiencies (LEs) were calculated to understand their potential for further growth (Figure 1). Based on the CSP experiment, the fragment hits can be classified into two groups. One group of fragments  only caused CSP for residue I74, suggesting that these fragments might bind to the active site. The other group of fragments caused CSPs for numerous residues, such as residues in α2, suggesting that these fragments might bind to a different site (Figures S5–S19). Therefore, the binding modes of these fragments were further determined using X‐ray crystallography, and only fragments with cocrystal structures available were chosen for further studies.

2.4. Fragment‐binding modes revealed by X‐ray crystallography

Well‐diffracting UBE2T apo crystals were used to soak the fragments to determine their modes of binding. We obtained the crystal structures of UBE2T in complex with fragments ETC‐0097 and ETC‐6705 at 1.7 and 1.5 Å resolutions, respectively (Figure 2 and Table S1). The structures revealed two unique compound‐binding pockets different from the previously reported UBE2T binder, EM04 (Morreale et al., 2017). ETC‐6705 was observed to bind to an allosteric site further away from the catalytic pocket. The cocrystal structure demonstrated that ETC‐6705 sat in a pocket between α1 and α2. Residues including R6, R9, E10, M13, L14, E17, N103, and A105 formed the fragment‐binding pocket. Direct hydrogen bond interaction was observed between the fragment ETC‐6705 and residue R9, and water‐mediated hydrogen bond interactions were observed between fragment and residues E10 and A105 (Figures 2, 3 and Figure S20). ETC‐6705 also engaged in hydrophobic interactions with UBE2T residues including M13, L14, E17, N103, and A105 (Figures 2, 3 and Figures S20 and S21). One of the key interactions between UBE2T and E3 ligase, FANCL RING, is the ‐stacking interaction in the interface between R6 and R9 of UBE2T and Y311 of FANCL RING. Overlay of ETC‐6705‐UBE2T complex with the FANCL RING–UBE2T complex (PDB ID 4CCG) showed the R9 residue of UBE2T interacting with the fragment ETC‐6705 (Figure S21). Further optimization of ETC‐6705 might result in compounds with potential to disrupt the positioning of Y311 of FANCL RING (E3) (Hodson et al., 2014), which indicates it may have the potential to be developed into an inhibitor to affect E3 and UBE2T interactions (Figure 3 and Figure S21). Unlike ETC‐6705, ETC‐0097 is binding to a solvent‐exposed pocket closer to the active site catalytic residue (C86). The binding pocket of ETC‐0097 included residues R84, I85, C86, L87, D88, K91, K95, and D122. A hydrogen bond (3.0 Å) between side chain of R84 and ETC‐0097 was identified (Figures 2 and 3). Water‐mediated hydrogen bond interactions were observed between ETC‐0097 and residues I85 and L87. In addition, position of ETC‐0097 was stabilized by a hydrogen bond interaction with carbonyl oxygen of L124 from the symmetry‐related molecule (Figures S20 and S22).

FIGURE 2.

FIGURE 2

Open in a new tab

Mechanisms of action of the identified fragments. (a) Structural element of UBE2T. (b) Overlay of the crystal structures determined in this study. Composite omit map of ETC‐0097 and ETC‐6705 (contoured at 1.0 σ) is shown in blue. (c) Crystal structure of UBE2T in complex with ETC‐6705. Upper panel shows residues critical for binding to ETC‐6705, and lower panel shows surface charge of the fragment‐binding site. (d) Crystal structure of UBE2T in complex with ETC‐0097. Upper panel shows residues critical for binding to ETC‐0097, and lower panel shows surface charge of the fragment‐binding site. The structure of UBE2T is shown in white. Residues within 4 Å of the fragment‐binding pocket are shown in sticks and labeled with residue name and residue number. Water molecules are shown as spheres and highlighted in light green. Hydrogen bond interactions are indicated as dashed lines. Upper panel shows the residues that may be critical for binding. Lower panel shows the surface charge analysis of the fragment‐binding pockets. Negative and positive charges are indicated in red and blue, respectively.

FIGURE 3.

FIGURE 3

Open in a new tab

Structure insights into developing UBE2T inhibitors. Different views of electron density maps for ETC‐0097 (a and b) and ETC‐6705 (c and d) are shown. (a) Fo‐Fc map for ETC‐0097 shown in green contoured at 3σ and rotated by 90°. (b) Composite omit map for ETC‐0097 shown in blue contoured at 1σ and rotated by 90°. (c) Fo‐Fc map for ETC‐6705 shown in green contoured at 3σ and rotated by 90°. (d) Composite omit map for ETC‐6705 shown in blue contoured at 1σ and rotated by 90°.The occupancy for ETC‐0097 and ETC‐6705 are 92% and 68%, respectively, in the final refined structures. Hydrogen bond interactions are indicated as dashed lines. (e) Cocrystal structures of several fragments in complexes with UBE2T. In addition to the structures determined in this study, the binding site of EM04 (PDB ID 5NGZ) is shown. Only one UBE2T structure is shown for clarity. (f) Superposition of ETC‐6705‐UBE2T complex structure with the Fanconi anemia complementation group L (FANCL) Really interesting new gene (RING)–UBE2T complex structure (PDB ID 4CCG).

2.5. Comparison of ETC‐0097 and ETC‐6705 binding to UBE2T in solution

Having identified the binding modes for ETC‐0097 and ETC‐6705 (Figure 2), we further compared their interactions with UBE2T in solution. ETC‐6705 interacted with UBE2T with an affinity of approximately 3.2 mM based on a titration experiment monitoring the CSP in the 1H‐15N‐HSQC spectra of UBE2T. The binding affinity of ETC‐0097 was not obtained due to signal overlaps in the NMR spectra. In addition, the fragments binding to UBE2T caused different changes in the 1H‐15N‐HSQC spectra (Figure 4 and Figures S18 and S19). ETC‐0097 binding to UBE2T only caused CSPs and line broadening of residues including I74, H76, I85, I104, D122, and D144 close to the active site (Figure 4). In contrast, ETC‐6705 binding to UBE2T caused CSPs of numerous residues close to its binding site as well as residues such as I85 close to the catalytic C86 (Figure 4). The solution NMR study is consistent with the results from the cocrystal structures. In addition, compounds binding to the ETC‐6705‐binding pocket might have dual functions, not only preventing UBE2T from associating with E3 ligase but also affecting the chemical environment of residues close to the active site to affect enzymatic activity. The effect of these fragments on the activity of UBE2T was evaluated using the developed biochemical assay, which was carried out in a mixture containing Cy5‐labeled ubiquitin, the E1 enzyme, UBE2T, and the RING domain containing amino acids 275–375 of the E3 FANCL RING (Cornwell et al., 2019). The IC50 value of a compound can be obtained if it inhibits the activity of UBE2T (Figure S23). No measurable IC50 values (only values of >900 μM were observed) were obtained for these fragments in the biochemical assay, which is not surprising because these fragments bind weakly to UBE2T.

FIGURE 4.

FIGURE 4

Open in a new tab

UBE2T and fragment binding in solution. The superimposed 1H‐15N‐HSQC spectra of 0.3 mM UBE2T1–154 in the presence of dimethyl sulfoxide (DMSO) (black) and 3 mM ETC‐6705 (red) (a) and ETC‐0097 (red) (b) are shown, respectively. Cross peaks exhibiting chemical shift perturbations or line broadening observed upon addition of a fragment are labeled with residue name and sequence number. (c) Chemical shift perturbations (CSPs) induced by ETC‐6705 (top) and ETC‐0097 (bottom) binding to UBE2T are plotted against residue number. For ETC‐6705, the line corresponding to 2× SD is indicated. (d) Residues affected upon binding to ETC‐6705. Some affected residues upon binding to ETC‐6705 are mapped onto the structure of UBE2T (PDB ID 5NGZ). Affected residues upon fragment binding are shown as green spheres. The catalytic residue C86 is shown as a sphere. (e) Residues from UBE2T affected upon ETC‐0097 binding. Residue I74 exhibited significant CSP and several residues exhibited line broadening upon UBE2T binding to ETC‐0097. Residues exhibited CSP or line broadening upon binding to fragments are labeled in blue and brown, respectively. C86 is shown as a sphere. The structures of fragments are shown.

2.6. Fragment growth through a database searching strategy

Although 15 fragments were identified in this study, we focused on the fragments for which we were able to obtain cocrystal structures with UBE2T for further studies. Based on the crystal structure obtained for ETC‐6705 (IC50 >900 μM, Kd = 3.20 ± 1.21 mM), we proceeded to explore whether this fragment could be grown to obtain more potent inhibitors (Figure 5). We were interested in replacing the —CF3 at the 2‐position to gain additional hydrogen bonding interactions with N103, R6, and R9. For reasons of synthetic tractability, a nitrogen atom was incorporated to scaffold hop from a pyridinone core to a pyrimidinone core, and analogues containing acid and amide moieties extending out from the 2‐position were synthesized (ETC‐2783 and ETC‐2784, respectively). Unfortunately, no improvement in potency was observed. Instead, we managed to obtain a cocrystal structure for the matched pair retaining the —CF3 substituent (ETC‐2443) and found that it was bound to a completely different site compared with the original fragment. Although the structure was not reported in this study due to low resolution, this result showed the importance of the core structure in fragment binding. Encouraged by this result, we proceeded to screen further structural analogues of this pyrimidinone core in our compound database and were able to identify ETC‐2889, a compound with approximately 10‐fold improvement in binding affinity and 3‐fold improvement in potency (IC50 = 288.2 μM, Kd = 0.311 mM). Chemical shift perturbation experiments confirmed its association with UBE2T in solution (Figure S24). Residues close to both the active site and the ETC‐6705‐binding site exhibited CSP upon binding to ETC‐2889. Further experiments are needed to confirm the binding modes. Nonetheless, ETC‐2889 is active against UBE2T enzymatic activity, and other derivatives including ETC‐7231, ETC‐7393 and ETC‐2966 possessing the same structural motif also exhibited measurable IC50 values. These compounds can be utilized as chemical starting points for further development of potent inhibitors of UBE2T.

FIGURE 5.

FIGURE 5

Open in a new tab

Fragment growth to improve the potency. (a) The chemical structures of ETC‐6705 and compounds in fragment growth are shown. The IC50s of the identified compounds are shown in the figure. (b) A strategy that can be applied in fragment‐based drug design. Fragment growth can be carried out through searching a compound library. LE, ligand efficiencies; MOA, mechanism of action.

3. DISCUSSION

UBE2T is a promising drug target due to its relationship with various cancers through its enzymatic activity in the ubiquitylation pathway and its participation in protein–protein interactions (Leung et al., 2021; Stewart et al., 2016). The structures of UBE2T reveal the low druggability at the active site due to the lack of a deep and hydrophobic pocket (Figure S2). The pockets that are important for interacting with other proteins are shallow, which makes developing small molecules to disrupt protein–protein interactions very challenging (Hodson et al., 2014). Despite these challenges, considerable effort has been spent to develop small molecules that can suppress the enzymatic activity or disrupt its interactions with other proteins (Cornwell et al., 2019; Strickson et al., 2013). To overcome the challenges of low druggability, reports in the literature have applied a few different strategies. First, inhibitors that can form a covalent bond with UBE2T were developed to exploit the reactivity of C86 in the active site. NSC697923 was shown to be a cell‐permeable and selective inhibitor of E2 enzymes including UBE2T (Alpi et al., 2016). Similar to BAY 11–0782, these compounds are able to form a covalent bond with the active site residue (C86) to suppress the activity of E2 enzymes (Alpi et al., 2016; Hodge et al., 2015). We also identified covalent inhibitors from a high‐throughput screening campaign (Anantharajan et al., 2023). Second, conventional high‐throughput assays were applied to identify potent inhibitors by setting up a sensitive assay. A homogeneous time‐resolved fluorescence (HTRF) assay was developed to identify compounds that can affect the proteins involved in the FA pathway (Cornwell et al., 2019). The identified inhibitor in the study was able to suppress UBE2T/FANCL RING‐mediated FANCD2 monoubiquitylation and was suggested to play a role in the sensitization of cells toward DNA cross‐linking agents (Cornwell et al., 2019). Lastly, assays using biophysical methods, such as thermal shift assay, have been applied to identify compounds that can affect the protein's activity through competitive or allosteric mechanisms (Morreale et al., 2017).

Fragment‐based drug discovery approaches have been widely used in the discovery of small molecules against different types of targets, including the modulation of protein–protein interactions (Li & Kang, 2021; Norton et al., 2016). Several fragments binding to UBE2T have been identified, which can serve as a starting point for the development of more potent inhibitors (Figure 1 and Figure S1) (Morreale et al., 2017). Compared with biochemical assays, fragment‐based drug discovery using biophysical methods can often provide more candidates for further development. In this study, we used a 19F‐NMR approach to identify fragments that bind weakly to UBE2T (Figure 1). Then, the binding sites of the identified fragments were obtained by solving the cocrystal structures. Two novel binding pockets were identified, one site close to the catalytic residue C86 and the other allosteric pocket distal to the active site (Figure 2). Although these fragments bind weakly to UBE2T with a dissociation constant in millimolar range and low LE values, the cocrystal structures determined in this study and previous studies provide important information for developing different types of inhibitors to modulate the functions of UBE2T (Kang & Keller, 2020; Li & Kang, 2020) (Figure 3). First, an inhibitor binding to the site close to C86 can affect the transportation of ubiquitin, which will result in an inhibitory effect on the enzymatic activity. In this case, ETC‐0097 will be a promising candidate for growing into a competitive inhibitor. Second, allosteric inhibitors can be useful for affecting the function of UBE2T through affecting its binding to its substrate or its partner E3 ligase. We demonstrated that ETC‐6705 was bound to a pocket between α1 and α2. This pocket is close to residues important for interacting with FANCL RING (Figure 3), implying that small molecules binding to this pocket can be developed into a compound to disrupt protein–protein interactions. Although the predicted drug score of UBE2T was not improved in the obtained cocrystal structures, potent compounds suitable for further development are still possible after careful chemical design. It is noted that a counter screen against other E2 enzymes will be useful to further understand the specificity of the inhibitors.

Based on the structural information, we grew one of the fragments to explore the possibility of improving the potency of the fragments. We found that a slight modification of the fragment resulted in a change in the binding pocket. This is not surprising as the interaction between the fragment and UBE2T may be driven by a few atoms as suggested by the low ligand efficiency of the fragment. Our current study suggests that structural investigation of the compounds during fragment growth is critical for understanding structure–activity relationships. Fragments with low ligand efficiencies may alter their binding sites even when a slight change of the fragment is made. Although the initial fragment growth did not result in potent inhibitors, we also demonstrated that fragment growth can be carried out by screening a library of analogues that possess the core structure of the fragment. Although fragment linking is a powerful strategy for fragment growth, this strategy requires suitable hits for linking (Frank et al., 2013; Li, 2020). Therefore, searching for structural analogues in a compound library based on the identified hits may be a complementary strategy for efficient fragment growth (Figure 5). Using this strategy, four compounds with  the same core structure exhibited measurable IC50 values, implying the possibility of developing potent inhibitors with improved chemical properties. Further study on these identified compounds will be useful for developing UBE2T inhibitors. It is noted that in this study, although we have focused only on fragments for which we were able to obtain cocrystal structures for further development, the other identified fragment hits can still be used for further development, and more studies are needed to characterize their binding modes with UBE2T.

In summary, we performed fragment screening using solution phase NMR spectroscopy to identify compounds that bind weakly to UBE2T. Crystal structures of UBE2T in complex with the hits, ETC‐0097 and ETC‐6705, revealed two distinct binding pockets on UBE2T. ETC‐0097 binds close to the catalytic pocket, whereas ETC‐6705 binds at the protein‐protein interaction (PPI) interface between UBE2T and E3, offering an unprecedented opportunity for screening PPI libraries to discover novel UBE2T inhibitors. Further studies demonstrated that the activity of the fragments can be improved. The novel binding pockets and the new fragments hits identified in this study will be useful for the future development of UBE2T inhibitors against both its catalytic activity and its protein–protein interactions.

4. METHODS

4.1. Recombinant protein production

The codon‐optimized complementary DNAs (cDNAs) encoding residues 1–154 of UBE2T were synthesized (Genscript) and cloned into pET15b and pGEX‐6P‐1 plasmids, respectively. The resulting plasmids were used for expressing proteins with a histidine tag and glutathione‐S‐transferase (GST) tag at the N terminus, respectively. To express a recombinant FANCL RING, the cDNA encoding amino acids of 275–375 of human FANCL RING was cloned into pGEX‐6P‐1 vector. The resulting plasmid will express GST‐fused FANCL RING protein for biochemical assay. The plasmids were transformed into Escherichia coli BL21(DE3) competent cells and the recombinant proteins were produced using the methods described previously (Anantharajan et al., 2023; Huang et al., 2023; Li et al., 2014). Please refer to Supplementary Information for details.

4.2. Fragment screening

Fragment screening was carried out using 19F‐NMR on a Bruker magnet with a proton frequency of 400 MHz equipped with a cryo‐probe. The fragment library contains 582 fluorinated fragments with molecular weight less than 300 Da. It was purchased from Enamine, and the fragments exhibited good solubility and covered diverse structures. A mixture of 10 fluorinated fragments (100 μM each) were prepared in a buffer that contained 20 mM sodium phosphate, 150 mM NaCl, and 1 mM Dithiothreitol. An internal control‐1,1,1‐trifluoroacetone was added in the sample. Normal 1D 19F NMR spectra of the mixtures in the absence and presence of UBE2T were recorded at 298 K. Please refer to Supplementary Information for details.

4.3. Titration experiments

To determine the binding affinity of the fragments with UBE2T, titration experiments were performed using a 15N‐labeled protein sample (Kang et al., 2013; Li et al., 2018; Zhang et al., 2016). The assignments for UBE2T have been deposited in the Biological Magnetic Resonance Data Bank under accession code 52051, and used for mapping the binding site (Huang et al., 2023). LE was calculated using the equation: LE = (−2.303RT)/N×log(Kd), where R is the ideal gas constant, T is the temperature in Kelvin, and N is the number of nonhydrogen atoms in the fragment (Murray et al., 2014). Please refer to Supplementary Information for details.

4.4. Crystallization and structure determination of the complexes

The apo crystals of UBE2T were soaked overnight (19–22 h) with crystallization buffer containing 5 mM fragments. The crystals were cryoprotected using 20% ethylene glycol and flash frozen in liquid nitrogen. The structures have been deposited into protein data bank with accession IDs of 8JVD and 8JUC, respectively. Please refer to Supplementary Information for details.

4.5. HTRF assay

The biochemical assay was modified from HTRF assay from a previous study (Cornwell et al., 2019). More details can be found in the Supplementary Information.

AUTHOR CONTRIBUTIONS

CongBao Kang: Conceptualization; methodology; data curation; writing – original draft. Yong Yao Loh: Methodology; data curation; writing – review and editing. Jothi Anantharajan: Methodology; data curation; writing – review and editing. Qiwei Huang: Data curation. Weijun Xu: Data curation; investigation. Justina Fulwood: Data curation. Hui Qi Ng: Data curation. Elizabeth Yihui Ng: Data curation. Chong Yu Gea: Data curation. Meng Ling Choong: Supervision. Qian Wen Tan: Data curation; supervision. Xiaoying Koh: Data curation; supervision. Wan Hsin Lim: Supervision; methodology. Kassoum Nacro: Methodology; supervision. Joseph Cherian: Supervision; investigation. Nithya Baburajendran: Supervision; methodology; writing ‐ review and editing. Zhiyuan Ke: Data curation; supervision; writing – review and editing.

Supporting information

Data S1. Supporting Information.

Click here for additional data file. (3.8MB, docx)

ACKNOWLEDGMENTS

This research was undertaken in part using the MX2 beamline at the Australian Synchrotron, part of ANSTO, and made use of the Australian Cancer Research Foundation (ACRF) detector. The synchrotron radiation experiments were perfomed at SPring‐8 with the approval of the Japan Synchrotron Radiation Research Institute (JASRI). The authors thank Subramanyam Vankadara for the help. The authors also appreciate the valuable suggestions from colleagues at EDDC.

Loh YY, Anantharajan J, Huang Q, Xu W, Fulwood J, Ng HQ, et al. Identification of small‐molecule binding sites of a ubiquitin‐conjugating enzyme‐UBE2T through fragment‐based screening. Protein Science. 2024;33(3):e4904. 10.1002/pro.4904

Yong Yao Loh and Jothi Anantharajan contributed equally to this work.

Review Editor: Jeanine Amacher

Contributor Information

Nithya Baburajendran, Email: nithya_baburajendran@eddc.sg.

Zhiyuan Ke, Email: ke_zhiyuan@eddc.sg.

CongBao Kang, Email: kang_congbao@eddc.sg.

DATA AVAILABILITY STATEMENT

The structures have been deposited into protein data bank with accession IDs of 8JVD and 8JUC, respectively.

REFERENCES

  1. Alpi AF, Chaugule V, Walden H. Mechanism and disease association of E2‐conjugating enzymes: lessons from UBE2T and UBE2L3. Biochem J. 2016;473:3401–3419. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Anantharajan J, Tan QW, Fulwood J, Sifang W, Huang Q, Ng HQ, et al. Identification and characterization of inhibitors covalently modifying catalytic cysteine of UBE2T and blocking ubiquitin transfer. Biochem Biophys Res Commun. 2023;149238:149238. [DOI] [PubMed] [Google Scholar]
  3. Ayotte Y, Woo S, LaPlante SR. Practical considerations and guidelines for spectral referencing for fluorine NMR ligand screening. ACS Omega. 2022;7:13155–13163. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Buchholz CR, Pomerantz WCK. 19F NMR viewed through two different lenses: ligand‐observed and protein‐observed 19F NMR applications for fragment‐based drug discovery. RSC Chem Biol. 2021;2:1312–1330. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Ceccaldi R, Sarangi P, D'Andrea AD. The Fanconi anaemia pathway: new players and new functions. Nat Rev Mol Cell Biol. 2016;17:337–349. [DOI] [PubMed] [Google Scholar]
  6. Cornwell MJ, Thomson GJ, Coates J, Belotserkovskaya R, Waddell ID, Jackson SP, et al. Small‐molecule inhibition of UBE2T/FANCL‐mediated ubiquitylation in the Fanconi anemia pathway. ACS Chem Biol. 2019;14:2148–2154. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Cui P, Li H, Wang C, Liu Y, Zhang M, Yin Y, et al. UBE2T regulates epithelial–mesenchymal transition through the PI3K‐AKT pathway and plays a carcinogenic role in ovarian cancer. J Ovarian Res. 2022;15:103. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. D'Andrea AD, Grompe M. The Fanconi anaemia/BRCA pathway. Nat Rev Cancer. 2003;3:23–34. [DOI] [PubMed] [Google Scholar]
  9. Frank AO, Feldkamp MD, Kennedy JP, Waterson AG, Pelz NF, Patrone JD, et al. Discovery of a potent inhibitor of replication protein a protein–protein interactions using a fragment‐linking approach. J Med Chem. 2013;56:9242–9250. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Halgren TA. Identifying and characterizing binding sites and assessing druggability. J Chem Inf Model. 2009;49:377–389. [DOI] [PubMed] [Google Scholar]
  11. Hao P, Kang B, Li Y, Hao W, Ma F. UBE2T promotes proliferation and regulates PI3K/Akt signaling in renal cell carcinoma. Mol Med Rep. 2019;20:1212–1220. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Hershko A, Ciechanover A. The ubiquitin system. Annu Rev Biochem. 1998;67:425–479. [DOI] [PubMed] [Google Scholar]
  13. Hodge CD, Edwards RA, Markin CJ, McDonald D, Pulvino M, Huen MSY, et al. Covalent inhibition of Ubc13 affects ubiquitin signaling and reveals active site elements important for targeting. ACS Chem Biol. 2015;10:1718–1728. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Hodson C, Purkiss A, Miles JA, Walden H. Structure of the human FANCL RING‐Ube2T complex reveals determinants of cognate E3‐E2 selection. Structure. 2014;22:337–344. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Huang Q, Ng HQ, Loh YY, Ke Z, Lim WH, Kang C. Backbone (1)H, (15)N and (13)C resonance assignments for an E2 ubiquitin conjugating enzyme‐UBE2T. Biomol NMR Assign. 2023;17:269–274. [DOI] [PubMed] [Google Scholar]
  16. Jackson SP, Durocher D. Regulation of DNA damage responses by ubiquitin and SUMO. Mol Cell. 2013;49:795–807. [DOI] [PubMed] [Google Scholar]
  17. Kang C, Keller TH. Probing biological mechanisms with chemical tools. Pharmacol Res. 2020;153:104656. [DOI] [PubMed] [Google Scholar]
  18. Kang C, Ye H, Chia J, Choi BH, Dhe‐Paganon S, Simon B, et al. Functional role of the flexible N‐terminal extension of FKBP38 in catalysis. Sci Rep. 2013;3:2985. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Kang CB. 19F‐NMR in target‐based drug discovery. Curr Med Chem. 2019;26:4964–4983. [DOI] [PubMed] [Google Scholar]
  20. Keszei AFA, Sicheri F. Mechanism of catalysis, E2 recognition, and autoinhibition for the IpaH family of bacterial E3 ubiquitin ligases. Proc Natl Acad Sci U S A. 2017;114:1311–1316. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Kolas NK, Chapman JR, Nakada S, Ylanko J, Chahwan R, Sweeney FD, et al. Orchestration of the DNA‐damage response by the RNF8 ubiquitin ligase. Science. 2007;318:1637–1640. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Leung RWH, Ho NPY, Leung CON, Lee TKW. UBE2T: a new molecular regulator of cancer stemness in hepatocellular carcinoma. Oncotarget. 2021;12:1727–1728. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Li Q. Application of fragment‐based drug discovery to versatile targets. Front Mol Biosci. 2020;7:180. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Li Q, Kang C. Mechanisms of action for small molecules revealed by structural biology in drug discovery. Int J Mol Sci. 2020;21:5262. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Li Q, Kang C. Perspectives on fragment‐based drug discovery: a strategy applicable to diverse targets. Curr Top Med Chem. 2021;21:1099–1112. [DOI] [PubMed] [Google Scholar]
  26. Li Q, Ng HQ, Yoon HS, Kang C. Solution structure of the cyclic‐nucleotide binding homology domain of a KCNH channel. J Struct Biol. 2014;186:68–74. [DOI] [PubMed] [Google Scholar]
  27. Li Y, Loh YR, Hung AW, Kang C. Characterization of molecular interactions between Zika virus protease and peptides derived from the C‐terminus of NS2B. Biochem Biophys Res Commun. 2018;503:691–696. [DOI] [PubMed] [Google Scholar]
  28. Li Y, Yang X, Lu D. Knockdown of ubiquitin‐conjugating enzyme E2T (UBE2T) suppresses lung adenocarcinoma progression via targeting fibulin‐5 (FBLN5). Bioengineered. 2022;13:11867–11880. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Lim K‐H, Song M‐H, Baek K‐H. Decision for cell fate: deubiquitinating enzymes in cell cycle checkpoint. Cell Mol Life Sci. 2016;73:1439–1455. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Liu L‐P, Yang M, Peng Q‐Z, Li M‐Y, Zhang Y‐S, Guo Y‐H, et al. UBE2T promotes hepatocellular carcinoma cell growth via ubiquitination of p53. Biochem Biophys Res Commun. 2017;493:20–27. [DOI] [PubMed] [Google Scholar]
  31. Longerich S, Kwon Y, Tsai M‐S, Hlaing AS, Kupfer GM, Sung P. Regulation of FANCD2 and FANCI monoubiquitination by their interaction and by DNA. Nucleic Acids Res. 2014;42:5657–5670. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Ma N, Li Z, Yan J, Liu X, He L, Xie R, et al. Diverse roles of UBE2T in cancer (review). Oncol Rep. 2023;49:69. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Machida YJ, Machida Y, Chen Y, Gurtan AM, Kupfer GM, D'Andrea AD, et al. UBE2T is the E2 in the Fanconi anemia pathway and undergoes negative autoregulation. Mol Cell. 2006;23:589–596. [DOI] [PubMed] [Google Scholar]
  34. Morreale FE, Bortoluzzi A, Chaugule VK, Arkinson C, Walden H, Ciulli A. Allosteric targeting of the Fanconi anemia ubiquitin‐conjugating enzyme Ube2T by fragment screening. J Med Chem. 2017;60:4093–4098. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Murray CW, Erlanson DA, Hopkins AL, Keserü GM, Leeson PD, Rees DC, et al. Validity of ligand efficiency metrics. ACS Med Chem Lett. 2014;5:616–618. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Norton RS, Leung EWW, Chandrashekaran IR, MacRaild CA. Applications of 19F‐NMR in fragment‐based drug discovery. Molecules. 2016;21:860. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Popovic D, Vucic D, Dikic I. Ubiquitination in disease pathogenesis and treatment. Nat Med. 2014;20:1242–1253. [DOI] [PubMed] [Google Scholar]
  38. Scheffner M, Nuber U, Huibregtse JM. Protein ubiquitination involving an E1–E2–E3 enzyme ubiquitin thioester cascade. Nature. 1995;373:81–83. [DOI] [PubMed] [Google Scholar]
  39. Stewart MD, Ritterhoff T, Klevit RE, Brzovic PS. E2 enzymes: more than just middle men. Cell Res. 2016;26:423–440. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Strickson S, Campbell DG, Emmerich CH, Knebel A, Plater L, Ritorto MS, et al. The anti‐inflammatory drug BAY 11‐7082 suppresses the MyD88‐dependent signalling network by targeting the ubiquitin system. Biochem J. 2013;451:427–437. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Tao X, Wu X, Zhou P, Yu X, Zhao C, Peng X, et al. UBE2T promotes glioblastoma malignancy through ubiquitination‐mediated degradation of RPL6. Cancer Sci. 2023;114:521–532. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Valimberti I, Tiberti M, Lambrughi M, Sarcevic B, Papaleo E. E2 superfamily of ubiquitin‐conjugating enzymes: constitutively active or activated through phosphorylation in the catalytic cleft. Sci Rep. 2015;5:14849. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Yu Z, Jiang X, Qin L, Deng H, Wang J, Ren W, et al. A novel UBE2T inhibitor suppresses Wnt/β‐catenin signaling hyperactivation and gastric cancer progression by blocking RACK1 ubiquitination. Oncogene. 2021;40:1027–1042. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Zhang Z, Li Y, Loh YR, Phoo WW, Hung AW, Kang C, et al. Crystal structure of unlinked NS2B‐NS3 protease from Zika virus. Science. 2016;354:1597–1600. [DOI] [PubMed] [Google Scholar]
  45. Zheng YW, Gao PF, Ma MZ, Chen Y, Li CY. Role of ubiquitin‐conjugating enzyme E2T in the carcinogenesis and progression of pancreatic cancer. Oncol Lett. 2020;20:1462–1468. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Data S1. Supporting Information.

Click here for additional data file. (3.8MB, docx)

Data Availability Statement

The structures have been deposited into protein data bank with accession IDs of 8JVD and 8JUC, respectively.

Articles from Protein Science : A Publication of the Protein Society are provided here courtesy of The Protein Society

RESOURCES