Identification of ubiquitin variants that inhibit the E2 ubiquitin conjugating enzyme, Ube2k

Middleton, A.J.

1 *

, Teyra, J.

, Zhu, J.

Sidhu, S.S.

, Day, C.L.

Department of Biochemistry, School of Biomedical Sciences, University of Otago, Dunedin 9054,

New Zealand

The Donnelly Centre for Cellular and Biomolecular Research, University of Toronto, 160 College

Street, Toronto, Ontario, M5S 3E1, Canada

*Author for correspondence

Department of Biochemistry

School of Biomedical Sciences

University of Otago

Dunedin 9054

New Zealand

Tel: +64 3 479 7870

email: adam.middleton@otago.ac.nz

.CC-BY-NC 4.0 International licensemade available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is

The copyright holder for this preprintthis version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.28.446107doi: bioRxiv preprint

Summary

Transfer of ubiquitin to substrate proteins regulates most processes in eukaryotic cells. E2 enzymes

are a central component of the ubiquitin machinery, and generally determine the type of ubiquitin

signal generated and thus the ultimate fate of substrate proteins. The E2, Ube2k, specifically builds

degradative ubiquitin chains on diverse substrates. Here we have identified protein-based reagents,

called ubiquitin variants (UbVs), that bind tightly and specifically to Ube2k. Crystal structures

reveal that the UbVs bind to the E2 enzyme at a hydrophobic cleft that is distinct from the active

site and previously identified ubiquitin binding sites. We demonstrate that the UbVs are potent

inhibitors of Ube2k and block both ubiquitin charging of the E2 enzyme, and E3-catalysed ubiquitin

transfer. The binding site of the UbVs suggests they directly clash with the ubiquitin activating

enzyme, while potentially disrupting interactions with E3 ligases via allosteric effects. Our data

reveal the first protein-based inhibitors of Ube2k and unveil a hydrophobic groove that could be an

effective target for inhibiting Ube2k and other E2 enzymes.

Keywords

Ubiquitin, ubiquitin-conjugating, E2 enzyme, Ube2k, HIP-2, phage display, ubiquitin variant,

proteasome, neurodegeneration

.CC-BY-NC 4.0 International licensemade available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is

The copyright holder for this preprintthis version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.28.446107doi: bioRxiv preprint

Introduction

Ubiquitin transfer is a post-translational modification that plays a critical role in almost all aspects

of eukaryotic cells, including control of protein degradation by the proteasome (Hochstrasser,

1996), modulation of gene expression (Mark and Rape, 2021), recruitment of proteins to signalling

platforms (Chen, 2005; Stewart et al., 2016), and dictation of the precise timing of cell division

(Pines, 2011). Consequently, dysregulation of ubiquitin transfer results in many diseases, such as

cancers, immune disorders, and neurodegenerative diseases (Rape, 2018). As a result, the ability to

manipulate the components of the ubiquitin system is of considerable interest for treatment of

disease.

Ubiquitin transfer is governed by three families of enzymes: the E1 ubiquitin activating

enzymes, the E2 ubiquitin conjugating enzymes, and the E3 ubiquitin ligase enzymes (Gundogdu

and Walden, 2019). Together, the machinery covalently links ubiquitin to a substrate Lysine or N-

terminal Methionine residue with an isopeptide bond. Ubiquitin itself can be a substrate as it

contains seven Lys residues, and this results in the formation of ubiquitin chains with distinct

consequences (Osborne et al., 2021). For example, chains linked by Lys63 can act as scaffolds for

recruiting proteins to signalling cascades, whereas Lys48-linked ubiquitin chains typically result in

degradation of the attached substrate by the proteasome (Osborne et al., 2021). The nature of the

ubiquitin signal is typically dictated by the E2 enzyme, of which there are ~40 in humans.

Depending on their structure and biological context, E2 enzymes can add a single ubiquitin moiety

or ubiquitin chains of various types to substrate proteins. Because the downstream effects of

ubiquitin transfer are entirely dependent on the nature of the ubiquitin signal, E2 enzymes have a

central role in ensuring substrate proteins are correctly modified by the ubiquitin machinery.

E2 enzymes are characterised by a conserved ubiquitin conjugating domain (UBC) that

interacts with an E1 enzyme and E3 ligase via conserved and partly overlapping interfaces. After

charging of the E2 enzyme with ubiquitin, the resulting E2~Ub conjugate disengages the E1 and

interacts with one of the hundreds of E3 ligases. The E3 ligase typically coordinates the choice of

substrate to be modified, while also activating the E2~Ub bond so that it is susceptible to

nucleophilic attack (Dou et al., 2012; Koliopoulos et al., 2016; Plechanovova et al., 2012; Sanchez

et al., 2016; Scott et al., 2014). When activated, the conjugated donor ubiquitin makes extensive

contacts with the E2 enzyme, including nestling of ubiquitin’s flexible C-terminal tail into a shallow

groove of the E2 enzyme. When building ubiquitin chains, the E2 enzyme must also interact with a

substrate ubiquitin, and this is referred to as the acceptor ubiquitin. Disrupting any of these protein-

protein interactions can disable the E2 enzyme.

.CC-BY-NC 4.0 International licensemade available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is

The copyright holder for this preprintthis version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.28.446107doi: bioRxiv preprint

The ubiquitin conjugating enzyme, Ube2k, produces Lys48-linked ubiquitin chains

exclusively, and promotes the degradation of its targets in cells (Chen and Pickart, 1990). A recent

report indicates that Ube2k is involved in promoting the degradation of the pro-survival Bcl-2

protein family member, Mcl-1, suggesting Ube2k may have a pro-apoptotic role (Djajawi et al.,

2020). Other studies suggest that Ube2k is involved in overcoming cell-cycle arrest in response to

DNA damage by promoting the degradation of p53 (Hong et al., 2019). In addition, Ube2k levels

contribute to multiple neurodegenerative diseases that are characterised by aggregation of proteins.

For example, Ube2k is involved in Huntingtin disease (Kalchman et al., 1996); is associated with

cell death from polyglutamine diseases (de Pril et al., 2007); is proapoptotic in response to the

accumulation of amyloid-β (Song et al., 2003); and is elevated in the brains of individuals with

schizophrenia (Meiklejohn et al., 2019). Recent work suggests that Ube2k deficiency results in

motor impairment reminiscent of Parkinson’s disease, and could act as a biomarker of the disease

(Su et al., 2018). There is also evidence that Ube2k can add Lys48-linked chains onto already-

established Lys63 ubiquitin chains, and thereby quench Lys63-induced signalling (Pluska et al.,

2021). While the roles of Ube2k are diverse, they are unified by the ability of Ube2k to promote the

degradation of substrate proteins. Development of inhibitors that specifically target Ube2k would

not only provide tools to allow a greater understanding of its function, but may also provide a

framework for small-molecule design.

Here we have used a phage-displayed ubiquitin variant (UbV) library to isolate specific

inhibitors of Ube2k. Of the six UbV binders identified, two are potent inhibitors of ubiquitin

transfer promoted by Ube2k. Biochemical experiments show that both UbVs inhibit charging of

Ube2k by the E1 enzyme, and also E3-catalysed discharge. Structures of the UbV-Ube2k complexes

show that both UbVs bind in a hydrophobic cleft distant from the active site, and this site includes

E2-E1 contacts. While this explains why charging of Ube2k is impeded, the UbV binding site does

not overlap with the E3-binding site, and the reason for decreased E3-catalysed discharge is less

certain. Our research reveals a hydrophobic groove on Ube2k that can bind ligands to block the

assembly of degradative ubiquitin chains.

.CC-BY-NC 4.0 International licensemade available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is

The copyright holder for this preprintthis version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.28.446107doi: bioRxiv preprint

Results

Selection and characterisation of UbVs that bind Ube2k

To discover specific modulators of the E2 enzyme Ube2k, we generated a stable Ube2k~Ub

conjugate and performed selection using a highly diverse phage-displayed library containing 2 x 10

unique UbVs. The library was a further iteration of libraries generated for selection against

deubiquitinases, E3 ligases, and other E2 enzymes (Ernst et al., 2013; Gabrielsen et al., 2017; Teyra

et al., 2019). In this library, residues were diversified across the surface of ubiquitin that is involved

in the vast majority of ubiquitin-protein interactions, including contacts with E2s, ubiquitin-

associating domains, and deubiquitinases. The surface comprises the beta sheet of ubiquitin and its

five flexible C-terminal residues (Figure 1A). To minimise disruptions to the fold of ubiquitin while

maximising diversity across the surface, the library was built using a ‘soft-randomisation’ approach

with degenerate codons that encode approximately 50% wild-type sequence at each diversified

position (Sidhu et al., 2000).

The Ube2k~Ub conjugate was used for the screen in an attempt to isolate UbVs that could

bind to either Ube2k, ubiquitin, or the E2-ubiquitin interface. Ube2k has a C-terminal extension that

contains a ubiquitin associating (UBA) domain (Merkley and Shaw, 2004), which binds ubiquitin

but is dispensable for ubiquitin transfer in vitro. As our goal was to identify modulators of ubiquitin

transfer, we generated a truncated construct that comprised only the UBC domain of Ube2k

(Ube2k-UBC). In addition, we included an active site mutation (C92K) to enable formation of a

stable Ube2k~Ub conjugate, and a C-terminal AVI tag to enable specific biotinylation of Ube2k

(Figure 1B). After purification and biotinylation of Ube2k, the E2 enzyme was charged with

ubiquitin, further purified, then immobilised to streptavidin/neutravidin-coated wells for phage-

display selections.

After five rounds of phage-display selections against Ube2k~Ub, clones were sequenced and

six unique UbVs were chosen for closer analysis (Figure 1C). Each UbV (including an N-terminal

FLAG tag) was subcloned into a vector that encoded a cleavable N-terminal His tag, and the

resulting UbVs were expressed in Escherchia coli. After purification, interaction of each UbV with

Ube2k~Ub was confirmed using an ELISA experiment (Figure S1). Importantly, the ELISAs

showed that each UbV bound to Ube2k and Ube2k~Ub with comparable affinity, but did not bind

ubiquitin alone. This indicated that the UbVs targeted Ube2k. The effect of each UbV on ubiquitin

transfer was then assessed using a chain-building assay in the presence of Ube2k (Figure 1D). For

this assay, each component of the ubiquitin cascade (E1, E2, E3, ubiquitin, and ATP) was mixed

together and incubated at 37 °C prior to analysis. In this experiment, five of the UbVs (UbV.k.1-5

.CC-BY-NC 4.0 International licensemade available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is

The copyright holder for this preprintthis version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.28.446107doi: bioRxiv preprint

Figure 1C, 1D) decreased ubiquitin chain assembly by Ube2k, with UbV.k.1 and UbV.k.2 being the

most potent inhibitors. As a result, we focussed our attention on these two UbVs.

The inhibitory UbVs bind tightly and specifically to Ube2k

The UBC domain of E2 enzymes is structurally conserved, and while there is less conservation on a

sequence level, there is the possibility that the UbVs that bind and inhibit Ube2k might cross-react

with other E2s. To measure specificity, we assessed binding of UbV.k.1 and UbV.k.2 to a

representative panel of 18 E2 enzymes. Our results revealed a preferential binding of both UbVs for

Ube2k, with minimal binding to other E2 enzymes (Figure 2A). In support of this, no inhibition was

seen when chain building assays were performed with two other E2 enzymes, Ube2d2 and

Ube2n/Ube2v2 in the presence of the UbVs (Figure S2A).

Next, we performed ITC to quantitatively measure the interaction between the UbVs and

Ube2k. Each UbV was titrated against full-length Ube2k as well as Ube2k-UBC. The K

UbV.k.1 for Ube2k was calculated to be 40 nM (Figure 2B), while UbV.k.2 bound to Ube2k

approximately ten-fold weaker, with a calculated K

of 420 nM (Figure 2B). Similar dissociation

constants were observed for both UbVs with Ube2k-UBC (Figure S2B, S2C), suggesting that the

UbVs interacted with the UBC domain and not the UBA domain. In further support of a stable

interaction, thermal denaturation experiments demonstrated that the melting temperature of Ube2k

was increased by 4 or 2 °C in the presence of UbV.k.1 or UbV.k.2, respectively (Figure S2D).

Together, these results demonstrate that the inhibitory UbVs are highly specific for Ube2k, bind

tightly to the E2 enzyme, and increase its stability.

The UbVs inhibit charging and discharging of Ube2k

What remained unclear was how the UbVs can block ubiquitin transfer from Ube2k. To promote

ubiquitin transfer, E2 enzymes must first be loaded with ubiquitin by an E1 enzyme, which transfers

ubiquitin to the E2 in a transthiolation reaction. The E2~Ub conjugate then disengages from the E1

enzyme and binds an E3 ligase that catalyses ubiquitin transfer (Rennie et al., 2020). The chain-

building assay (Figure 1D) involves both charging of the E2 by an E1 enzyme, and E3-dependent

ubiquitin transfer. To tease out the mechanism of inhibition, we analysed E1-dependent charging

and E3-catalysed transfer separately. First, to determine if the UbVs affected charging of Ube2k, we

monitored the formation of an E2~Ub thioester linked conjugate in the absence or presence of the

UbVs. As observed in Figure 3A, both UbVs impeded charging of Ube2k as judged by slowed

formation of an E2~Ub species over time, with UbV.k.1 being a more potent inhibitor than UbV.k.2

.CC-BY-NC 4.0 International licensemade available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is

The copyright holder for this preprintthis version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.28.446107doi: bioRxiv preprint

(Figure 3A). We also confirmed that the UbVs did not affect the upstream E1 activation of ubiquitin

(Figure S3). These results suggest that the UbVs impede loading of Ube2k by the E1 enzyme.

To assess whether the UbVs could also inhibit E3-catalysed ubiquitin transfer from Ube2k,

we prepared a fluorescently-tagged thioester-linked Ube2k~Ub-Cy3

conjugate and monitored the

disappearance of the conjugate and appearance of di-ubiquitin in the presence of an E3 ligase and

excess UbVs (Figure 3B). Our results show that the E3-dependent discharge was slowed by the

UbVs when using either RNF12 or RNF125 E3 ligases (Figure 3B, 3C). However, in the absence of

an E3 ligase, the UbVs did not affect basal ubiquitin transfer activity of Ube2k (Figure S4). These

results suggest that, as well as inhibiting charging of the E2 enzymes by the E1 enzyme, the UbVs

also disrupted E3 ligase catalysed ubiquitin discharge.

Crystal structures of the UbV-Ube2k complexes reveal the mechanism of inhibition

To establish exactly how the two UbVs bind and inhibit the activity of Ube2k, we solved their

structures in complex with the E2 enzyme. Following copurification of UbV.k.2 and Ube2k, crystals

were obtained for the complex and its structure was solved by molecular replacement using Ube2k

and ubiquitin. The structure was refined to a resolution of 2.4 Å with final R

work

free

of 20.6 / 24.9

(Table 1). Crystals of a similarly purified UbV.k.1-Ube2k complex could not be obtained. Instead,

we purified an isopeptide-linked Ube2k~Ub conjugate and copurified it with UbV.k.1. With this

complex, crystals were obtained and the structure of UbV.k.1-Ube2k~Ub was solved to 3.0 Å with

final R

work

free

of 23.4 / 29.3 (Table 1). In both cases, the asymmetric unit contained only one

complex.

In both structures, Ube2k shows the expected fold comprising a UBC domain bridged by a

short linker to a C-terminal UBA domain (Figure S5A). A C-alpha overlay of Ube2k from the two

structures has an RMSD of 1.2 Å, and they both overlay well with a published structure of Ube2k

(PDB: 5DFL)(Middleton and Day, 2015) with C-alpha RMSDs of 1.3 and 0.65 Å for Ube2k-

UbV.k.1 and -UbV.k.2, respectively. In the Ube2k-UbV.k.2 complex, residues 32-34 of Ube2k and

7-12 of UbV.k.2 are shifted towards each other, and it appears that a slight conformational change

of both proteins is necessary to form the complex.

Both UbVs bind at the same position of Ube2k, which comprises a hydrophobic cleft

formed between α-helix 1 and β-sheet 1 of Ube2k (Figure 4A, 4B). Surprisingly, an overlay of the

two complexes shows that the UbVs are rotated by approximately 15° with respect to one another

and Ube2k (S5B), even though the contact residues from both are largely from the same sequence

positions. Notably, the contact residues from both UbVs are part of the diversified surface, with

.CC-BY-NC 4.0 International licensemade available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is

The copyright holder for this preprintthis version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.28.446107doi: bioRxiv preprint

major contributions from those mutated from wild-type ubiquitin (Figure 1B, Figure 4C, 4D, 4E).

An analysis using PISA (Protein, Interfaces, Structures and Assemblies) (Krissinel and Henrick,

2007) reveals that the interface between UbV.k.1 and Ube2k buries approximately 470 Å

, while

UbV.k.2 buries approximately 530 Å

. While UbV.k.2 buries a greater surface area of Ube2k than

UbV.k.1, the smaller hydrophobic patch at the core of the UbV.k.2-Ube2k interaction may explain

its reduced affinity for Ube2k (Figure 4E, 4F).

The more potent UbV, UbV.k.1, uses an array of four Phe residues (Phe2, Phe4*, Phe64, and

Phe66; wild-type ubiquitin residues are indicated with asterisks) and one Tyr residue (Tyr12) to

form a hydrophobic surface that packs against Ube2k residues Phe13, Leu17, Leu31, and Phe36

(Figure 4C, 4F). This patch is complemented by a hydrophilic ‘shell’ made up of the hydroxyls of

Tyr12, Thr14*, and the side chain Glu63 of UbV.k.1; and Lys10, Lys14, Asp30, and Glu34 of

Ube2k (Figure 4C, 4F). In a similar manner, the UbV.k.2-Ube2k interaction has a hydrophobic core,

comprising Phe4*, Phe8, and Ile66 of UbV.k.2, which interacts with the same patch on Ube2k as for

UbV.k.1 (Phe13, Leu17, Leu31, and Phe36). This is supported by an extensive polar shell nucleated

by Ser6 of UbV.k.2, which acts to stabilise contacts between Thr12* of UbV.k.2 with Asp30 and

Leu31 of Ube2k (Figure 4D, 4F). In addition, the side chain of Lys14 of Ube2k contacts Gln2* and

Asp64 of UbV.k.2, while the side chain of Lys28 of Ube2k makes contacts with Gly10*.

To test the importance of the diversified residues, we introduced single point mutations to

revert both UbVs to the wild-type ubiquitin sequence and assessed binding of the mutants to Ube2k

by ELISA EC

measurements. Interestingly, each revertant disrupted the EC

by at least an order

of magnitude, while some of the mutations almost eliminated binding (Figure 5A, 5B). For

UbV.k.1, mutation of Phe64 and Glu63 to their wild-type amino acid (Glu and Lys, respectively)

almost completely disrupted binding to Ube2k, while for UbV.k.2, mutating Ser6 to Lys, as well as

Ile66 to Thr were highly disruptive to binding. In support of the importance of these residues, the

mutations, F64E and E63K made to UbV.k.1, as well as S6K and I66T on UbV.k.2, did not inhibit

the formation of ubiquitin chains (Figure 5C). These data suggest that the network of diversified

residues is highly interdependent and disruption of any of them results in a notable loss of binding

and inhibition.

Architecture of UbV-Ube2k interactions

Ubiquitin activation by an E1 enzyme sets the stage for subsequent transfer to E2 enzymes and E3-

ligase catalysed ubiquitin transfer. The E1 enzyme engages E2s using two major interactions: first

the E2 enzyme is recruited by a ubiquitin-fold domain (UFD) that interacts with a semi-conserved

.CC-BY-NC 4.0 International licensemade available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is

The copyright holder for this preprintthis version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.28.446107doi: bioRxiv preprint

surface on E2s; subsequently, the UFD domain rotates and positions the active site domain of the

E1 next to the catalytic Cys of the E2 enzyme to allow a transthiolation reaction to occur (Olsen and

Lima, 2013; Williams et al., 2019). Because there is no structure of Ube2k bound to the E1, we

generated a molecular model by overlaying Ube2k from our two structures with Cdc23 from a

recent E2-E1 complex structure and mapped the predicted UFD contacts on Ube2k (Figure 6A). In

our model, both the UbVs clash with the predicted binding site of the UFD, likely explaining why

ubiquitin charging of Ube2k by the E1 enzyme is reduced in the presence of the inhibitory UbVs.

The UbVs also disrupt the activity of E3-catalysed discharge of Ube2k (Figure 3B, 3C). To assess

whether the UbVs clash with the likely E3 binding site, we generated a model of an E3-engaged

Ube2k molecule. For this, we overlayed Ube2k from both complexes with Ube2d2 from a crystal

structure of Ube2d2-RNF12 (Figure 6B) (Middleton et al., 2020). In this model there is no overlap

between RNF12 and the UbVs, suggesting that inhibition of E3-catalysed discharge of Ube2k is

likely due to allosteric effects. Together, the architecture of both the UbV-Ube2k interactions

suggests a direct steric clash with the E1 enzyme, while the mechanism of inhibition of the E3

ligase by the UbVs is less clear, and appears to be due to allosteric effects.

.CC-BY-NC 4.0 International licensemade available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is

The copyright holder for this preprintthis version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.28.446107doi: bioRxiv preprint

Discussion

As the central enzymes in the ubiquitin cascade, E2 enzymes play a critical role in determining the

exact nature of the ubiquitin code (Stewart et al., 2016). Here we have focused on developing tools

to modulate targeting of proteins to the proteasome by identifying regulators of the E2 enzyme,

Ube2k, which only produces Lys48-linked degradative ubiquitin chains. Using phage display, we

discovered two UbVs that are potent inhibitors of ubiquitin transfer by Ube2k. The UbVs bind

tightly and highly specifically to Ube2k, and they appear to isolate the E2 from the ubiquitin system

by disrupting interactions with both E3 ligases and the E1 enzyme. The fact that both UbVs bind at

the same hydrophobic cleft on Ube2k suggests this site may prove to be an effective target for small

molecule binding. By inhibiting the synthesis of a degradative signal, these UbVs will provide tools

for research, and may suggest approaches for design of small molecule Ube2k inhibitors. For

example, the UbVs may prove useful for discovery of small-molecules that bind at this same

hydrophobic cleft by displacement assays, as demonstrated by others (Maculins et al., 2020).

To date, approximately twenty inhibitors of ubiquitin-conjugating E2 enzymes have been

discovered (Ardecky et al., 2010; Ceccarelli et al., 2011; Chen et al., 2017; Cheng et al., 2014;

Cornwell et al., 2019; Garg et al., 2020; Helms et al., 2011; Kothayer et al., 2016; Liu et al., 2014;

Pulvino et al., 2012; Sanders et al., 2013; Strickson et al., 2013; Wang et al., 2018, 2011). Many of

the reported inhibitors target the active site Cys, others act by disrupting protein-protein

interactions, and some have an allosteric effect on ubiquitin transfer. Because of the conservation

and lack of a distinctive active site pocket on E2 enzymes, many of these orthosteric inhibitors have

low specificity and relatively high IC

values (Osborne et al., 2021). Display technologies offer

powerful ways of isolating protein or peptide-based inhibitors of proteins (including E2 enzymes)

that can circumvent these problems. For example, phage display was used to discover UbVs that

bound at the ‘backside’ ubiquitin binding site found on some E2 enzymes (Garg et al., 2020). These

variants bound more tightly than many other inhibitors of E2 enzymes, with IC

measurements

ranging from 65-280 nM. Importantly, they could all inhibit ubiquitin transfer by either directly

blocking interactions with other components of the ubiquitin cascade (similar to what we observed

here) or by allosteric effects that disrupt the enhanced ubiquitin transfer normally provided by

backside-bound ubiquitin.

Surprisingly, the UbVs we discovered did not bind at a predicted ubiquitin-binding site

(Figure 6C), nor did they block the active site Cys. Instead they bound at a hydrophobic groove

between α-helix 1 and β-sheet 1 on Ube2k that overlaps considerably with the binding site of the

UFD domain of E1 enzymes (Figure 6A) (Lv et al., 2017; Olsen and Lima, 2013; Tokgoz et al.,

.CC-BY-NC 4.0 International licensemade available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is

The copyright holder for this preprintthis version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.28.446107doi: bioRxiv preprint

2011; Williams et al., 2019), but binds in a distinct manner. The UFD-E2 interface appears to be

largely governed by polar contacts between acidic residues on the E1 and a set of basic residues

conserved on helix-1 of E2 enzymes. By contrast, both of the UbVs reported here rely on a

hydrophobic patch, as well as critical polar contacts between the molecules (Figure 4C, 4D, 4F,

Figure 5). A sequence comparison of the E2 enzymes used in our representative screen suggests that

the residues of Ube2k that interact with the UbVs are not highly conserved (Figure S6). In

particular, this is true for the hydrophobic residues, Phe13, Leu17, Leu31, and Phe36, which

provide the core of both of the interactions. This low conservation likely explains the specificity of

the interaction with Ube2k. How the UbVs block engagement by the E3 ligase is not as clear,

because in our model (Figure 6B) the UbVs bind away from the predicted RNF12 binding site on

Ube2k. E3 ligase-E2 interactions are highly conserved, and it is unlikely that the E3-Ube2k

interaction will differ greatly from our model. As a result, it is probable that the inhibition of E3

ligase activity is due to allosteric effects caused by the UbV-Ube2k interactions.

Currently, the role of Ube2k in biology has not been extensively investigated, but studies

have shown that the enzyme appears to have diverse roles in cells, such as regulating cell division

(Bae et al., 2013), controlling the stability of p53 (Hong et al., 2019), and is likely to play roles in

the development of neurodegenerative diseases (Kalchman et al., 1996; Meiklejohn et al., 2019; de

Pril et al., 2007; Song et al., 2003; Su et al., 2018). The UbVs may prove to be productive in

narrowing down the exact importance of Ube2k in the cell. While there is considerable redundancy

in the cell, Ube2k is the only E2 enzyme that contains a ubiquitin-binding domain, which may act to

increase the local concentration of the enzyme to promote rapid synthesis of extensive ubiquitin

chains, and degradation. An interesting recent report suggests the UBA domain may be involved in

recruiting Ube2k to Lys63-linked ubiquitin chains to potentially quench signalling pathways

promoted by Lys63 chains (Pluska et al., 2021). Further work is needed to understand more about

the role of Ube2k in cells, and the UbVs may prove essential to these studies.

.CC-BY-NC 4.0 International licensemade available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is

The copyright holder for this preprintthis version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.28.446107doi: bioRxiv preprint

Materials and Methods

DNA Constructs

For phage selection, a DNA construct encoding residues 1-157 of Ube2k (Ube2k-UBC) containing

the mutations C92K and K97R fused to a short GSGS linker and an AVI tag was obtained from

GeneArt (Invitrogen). This construct was used to generate a stable isopeptide-linked Ube2k~Ub

conjugate. A DNA construct encoding the biotin ligase BirA was obtained from GeneArt. Both of

these constructs were cloned into a pET-LIC vector encoding a 6xHis tag followed by a 3C protease

cleavage site N-terminal to the construct. For the E2 specificity screen, a modified pET-LIC vector

was constructed to allow co-expression of C-terminal AVI-tagged proteins with the biotin ligase

BirA. This vector was named pET-LIC-3Avi-BirA-Duet. For cloning of the E2s, primers containing

LIC overhangs were used to amplify a representative panel of E2 enzymes, and these were then

cloned into pET-LIC-3Avi-BirA-Duet. Mutations were generated using single-step site-directed

mutagenesis (Liu and Naismith, 2008). After phage selection, the DNA sequences encoding the

UbVs were amplified using universal primers and cloned into pET-LIC. Cloned UbVs included an

N-terminal FLAG tag and a C-terminal GGSGG tag. An RNF125 construct encoding residues 16-

125, as well as Ube2k were cloned into pGEX-6P3 using BamHI and EcoRI restriction digestion

and ligation.

Protein expression and purification

Each protein was produced in the E. coli BL21-star cell line in LB broth by incubation at 37 °C

followed by 18 °C for overnight expression. Protein expression was either induced by 0.2 mM

IPTG, or by autoinduction. After centrifugation, cell pellets were resuspended in 1x PBS, sonicated,

and centrifuged at 15,000x g. For constructs in the pET-LIC vector, the His-tagged protein was

immobilised to nickel beads (Macherey-Nagel), eluted with 300 mM imidazole, and either dialysed

or desalted into PBS. Subsequently, this was mixed with in-house produced His-tagged 3C for

overnight digestion to remove the His tag. For purification of BirA, the 3C digestion was not

performed. After confirmation of digestion, the protein was further purified using size-exclusion

chromatography over a Superdex S75 13/300 or 16/600 column (GE Healthcare). For Ube2k-UBC

containing the Avi tag, the purified protein was incubated at a final concentration of 100 μM with

300 μM biotin, 2 μM BirA, 2.5 mM ATP, and 2.5 mM MgCl

for 1 h at room temperature. To

confirm biotinylation, the purified proteins were incubated with excess streptavidin (New England

Biolabs) for 10 minutes before analysis via gel-shift using SDS-PAGE. The protein was then

desalted to remove excess biotin, and the BirA removed using Ni

affinity chromatography. For the

.CC-BY-NC 4.0 International licensemade available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is

The copyright holder for this preprintthis version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.28.446107doi: bioRxiv preprint

constructs in pET-LIC-3Avi-BirA-Duet, a final concentration of 200 μM biotin was added to the E.

coli cultures before transfer to 18 °C. Ubiquitin, E1, Ube2k, Ube2d2, Ube2n, Ube2v2, RNF12,

TRAF6 and 3C protease were purified using previously established methods (Berndsen and

Wolberger, 2011; Middleton and Day, 2015; Middleton et al., 2017, 2020; Sato et al., 2009).

RNF125 was purified as described for RNF12. Ubiquitin with all Lys residues mutated to Arg (K0)

was purified identically to wild-type ubiquitin. Thioester and isopeptide linked Ube2k~Ub were

purified using techniques previously described(Middleton et al., 2014). Fluorescently labelled

ubiquitin for chain-building assays (Ub-5AIF), ubiquitin discharge assays (Ub-Cy3

) and E1

activation assays (Ub-Cy3) were produced as previously described. For the pulldown experiment,

GST-Ube2k and GST-3C were bound to GSH resin and washed extensively. GST-Ubiquitin and

GST alone were purified by binding to GSH resin, and elution from the resin with 10 mM GSH for

30 min. Subsequently, they were purified by size-exclusion chromatography. All proteins were

frozen in liquid nitrogen and kept at -80 °C.

Phage-display selections

Five rounds of phage-display selections were performed as previously described using a library

highly similar to the ones used in earlier studies (Ernst et al., 2013; Garg et al., 2020). For selection,

biotinylated Ube2k-UBC~Ub was immobilised to Nunc Maxisorp 96-well plates (Fisher Scientific,

Nepean, ON, Canada) coated alternately with neutravidin (New England BioLabs) or streptavidin

(ThermoFisher Scientific). After blocking, the UbV library was added to immobilised Ube2k~Ub

for 1 h at 4 °C followed by washing with PBS containing 0.02% Tween-20 and elution at low pH.

Subsequently, E. coli TOP10 cells (ThermoFisher Scientific) were infected with the eluted phage

before being grown overnight at 37 °C. The E. coli TOP10 cells were plated and 96 individual

clones (48 from round four and 48 from round five) were grown and sequenced. From the

sequences, six distinct UbVs were analysed in subsequent experiments.

Ubiquitin transfer assays

For ubiquitin chain-building experiments, 5 μM E3 (RNF125, TRAF6, or RNF12), 0.1 μM E1, 8-10

μM E2 (Ube2k, Ube2d2, or Ube2n/Ube2v2), 50 μM ubiquitin, and 5 μM Ub-5AIF were incubated

with or without 10-30 μM UbVs in buffer containing 20 mM Tris-HCl (pH 7.5), 50 mM NaCl, 2

mM TCEP, 2 mM ATP, and 2 mM MgCl

. Reactions were incubated at 37 °C and mixed with SDS

loading dye containing 2-mercaptoethanol at the indicated time points. The resulting SDS-PAGE

gels were imaged on a Las-3000 (Fuji Film) imager for approximately 60 s. For ubiquitin discharge

.CC-BY-NC 4.0 International licensemade available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is

The copyright holder for this preprintthis version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.28.446107doi: bioRxiv preprint

assays, purified thioester-linked Ube2k~Ub-Cy3

conjugate (approximately 10 μM) was mixed

with 250 nM RNF125 or 2.5 μM RNF12 and 50 μM ubiquitin with or without UbVs, and incubated

at 20 °C. At the indicated time points, the samples were mixed with SDS loading dye containing no

reducing agent and resolved with SDS-PAGE. Samples were imaged on an Odyssey FC imaging

system (LI-COR) at 600 nm with a 2 min exposure. Charging assays of Ube2k were performed in

PBS with 0.1 μM E1, 10 μM Ube2k, 50 μM ubiquitin, +/- 10 μM UbV.k.1 or UbV.k.2, 2 mM

MgCl

, and 2 mM ATP. Reactions were incubated at 37 °C before being mixed with non-reducing

SDS dye and proteins were visualised with Coomassie blue staining of the gels. E1 activation

experiments were performed in PBS, 2 mM MgCl

, 2 mM ATP with 0.5 μM E1, 50 μM Ub-Cy3,

with or without 30 μM UbV.k.1 or UbV.k.2. The reactions were incubated at 20 °C for 5 min before

being mixed with non-reducing SDS dye, and gels resolved via fluorescence of Cy3 as for discharge

experiments.

ELISA experiments

For epitope mapping, 384-well high-binding plates (Greiner Bio-One) were coated overnight with

streptavidin, GST, or GST-Ubiquitin. Biotinylated Ube2k-UBC~Ub, Ube2k, or PBS was mixed

with the immobilised streptavidin for ~ 15 min. Purified UbVs were then added at 0.5 μM and

binding was detected via reaction of 3,3',5,5'-tetramethylbenzidine (TMB) with an HRP-FLAG

antibody (1:4000, ThermoFisher Scientific). The colorimetric reaction was quenched with either

phosphoric or sulfuric acid after 5-10 min. Absorbance at 450 nm was measured with a ClarioSTAR

Plus (BMG Labtech). Background was subtracted and values were normalised where the highest

signal was 1.0. Similar approaches were used for the E2 specificity and EC

ELISA experiments.

For the EC

measurements, the UbVs were diluted as indicated before being mixed with

immobilised Ube2k. The EC

value is the concentration at which the signal of absorbance is 50%

of total binding. These data were analysed and plotted in R version 4.0.4.

Binding experiments

Isothermal titration calorimetry was performed with a VP-ITC (MicroCal) at 30 °C. Ube2k and

Ube2k-UBC were added to the cell at 8 and 8-10 μM, respectively, while UbV.k.1 (at 150 or 85

μM) and UbV.k.2 (at 110 μM) were in the syringe. All samples were either dialysed against or

purified with a common stock of PBS. Analysis was performed using NITPIC, SEDPHAT, and

GUSSI (Scheuermann and Brautigam, 2015). Thermal denaturation of Ube2k was performed at a

final protein concentration of 5 μM. SYPRO Orange (ThermoFischer) dye was added to the protein

.CC-BY-NC 4.0 International licensemade available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is

The copyright holder for this preprintthis version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.28.446107doi: bioRxiv preprint

in a white 96-well PCR plate (Lab Supply, Dunedin, New Zealand) and measured in a Roche

LightCycler 480 II instrument using the SYPRO Orange program. Data were analysed and plotted

with R version 4.0.4.

Crystallography and structure solution

For crystallography, the copurified complexes were mixed with the crystal screens PACT Premier

and JCSG plus (Molecular Dimensions) at 200:200 nL and 200:100 nL protein:well solution drop

ratios in Swissci 3-well sitting drop plates using a mosquito (TTP Labtech). Diffraction-quality

crystals of the UbV.k.2-Ube2k complex were produced in 0.2 M sodium citrate tribasic trihydrate,

0.1 M Bis-Tris propane pH 7.5, and 20% PEG 3350 and the dataset was collected at the MX1

beamline, Australian Synchrotron. UbV.k.1-Ube2k~Ub crystals were grown in 0.2-0.3 M

ammonium citrate dibasic, 20-25% PEG 3350 and data were collected at the MX2 beamline,

Australian Synchrotron. Data were processed and scaled with XDS (Kabsch, 2010), and datasets

were merged with Aimless (CCP4, 1994). Subsequently, Phaser-MR (McCoy et al., 2007) was used

to solve the structures using a single Ube2k molecule from PDB 5DFL (Middleton and Day, 2015),

and the core domain of ubiquitin (PDB: 1UBQ) (Vijay-Kumar et al., 1987). For the UbV.k.1-

Ube2k~Ub dataset, the ice ring at 3.4 Å was excluded from the processing. Both structures were

refined with Phenix refine (Adams et al., 2010), and manually corrected iteratively using Coot

(Emsley et al., 2010). All images were generated using ChimeraX (Pettersen et al., 2021). Crystal

coordinates for UbV.k.1-Ube2k~Ub and UbV.k.2-Ube2k have been deposited in the Protein Data

Base under accession codes 7MYF or 7MYH, respectively.

.CC-BY-NC 4.0 International licensemade available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is

The copyright holder for this preprintthis version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.28.446107doi: bioRxiv preprint

Acknowledgements

We acknowledge Jun Gu for assistance with phage display, and Khanh Nguyen for help with the

thermal denaturation experiments. This research was undertaken in part using the MX1 and MX2

beamline at the Australian Synchrotron, part of ANSTO, and made use of the Australian Cancer

Research Foundation (ACRF) detector. We thank the New Zealand Synchrotron Group for

facilitating access to the Australian Synchrotron. A.J.M. was supported by a Marsden Fund Fast-

start Grant administered by the Royal Society of New Zealand, and a Sir Charles Hercus Health

Research Fellowship awarded by the Health Research Council of New Zealand.

Author Contributions

Conceptualization, A.J.M., C.L.D., S.S.S.; Investigation, A.J.M., J.Z., J.T.; Resources, A.J.M.,

S.S.S., C.L.D.; Writing – Original Draft, A.J.M.; Writing – Review & Editing, C.L.D., S.S.S., J.T.,

A.J.M.; Visualization, A.J.M., J.T.; Funding Acquisition, A.J.M., S.S.S., C.L.D.

Declaration of Interests

The authors declare no competing interests.

.CC-BY-NC 4.0 International licensemade available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is

The copyright holder for this preprintthis version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.28.446107doi: bioRxiv preprint

Figure 1. Selection for UbVs that bind to Ube2k (A) The structure of wild-type ubiquitin (1UBQ)

with the diversified residues shown as numbered spheres. Ubiquitin shown as a cartoon against a

semi-transparent surface with acidic, basic, and hydrophobic groups coloured red, blue, or white,

respectively. (B) Domain structure of full-length Ube2k and Ube2k-UBC used in this study. UBC:

ubiquitin conjugating domain; UBA: ubiquitin associated domain; AVI: biotin-specific tag. (C)

Sequence alignment of wild-type ubiquitin and the UbVs isolated from the selection. Only regions

diversified in the library are shown. In the alignment, the amino acids indicate changes relative to

the wild-type ubiquitin sequence, while dashes represent no change. (D) Chain-building ubiquitin

(Ub) transfer assay performed without (-E3 and +E3) and with the six UbVs. For this assay, the E3

ligase RNF125 was used to promote ubiquitin transfer. Imaged as fluorescence from 5AIF-tagged

ubiquitin. Assay performed in triplicate with similar results. See also Figure S1.

.CC-BY-NC 4.0 International licensemade available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is

The copyright holder for this preprintthis version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.28.446107doi: bioRxiv preprint

Figure 2. The UbVs bind to Ube2k specifically and tightly. (A) E2 enzyme specificity screen was

performed using ELISA. A representative panel of E2 enzymes were immobilised to a plate before

the UbVs were added and binding was detected spectrophotometrically by absorbance at 450 nm.

Readings were normalised, and four replicates were performed. (B) Thermograms (top) and fitted

isotherms (bottom) of ITC measurements performed in technical triplicate (shown in blue, red, and

orange) where either UbV.k.1 or UbV.k.2 was injected into Ube2k. Binding constants (K

) are

shown. See also Figure S2.

.CC-BY-NC 4.0 International licensemade available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is

The copyright holder for this preprintthis version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.28.446107doi: bioRxiv preprint

Figure 3. The UbVs inhibit both E2 charging and E3-catalysed ubiquitin discharge. (A) E2

charging assay (see schematic on the left) performed with and without the UbVs. Samples were

quenched with non-reducing dye at the indicated time points. The gel was visualised by staining

with Coomassie blue dye. (B, C) A ubiquitin discharge experiment (schematic on the left; red star

indicates fluorescent tag) performed with and without the UbVs. Samples were quenched with non-

reducing dye at the indicated time points. The gel was visualised by fluorescence detected at 600

nm. Panels (B) and (C) show results with RNF125 or RNF12 as the E3 ligase, respectively. See also

Figure S3 and S4.

.CC-BY-NC 4.0 International licensemade available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is

The copyright holder for this preprintthis version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.28.446107doi: bioRxiv preprint

Figure 4. Crystal structures of UbV.k.1 or UbV.k.2 in complex with Ube2k. (A,B) Ribbon

representation of UbV.k.1 (A) or UbV.k.2 (B) in complex with Ube2k. N and C termini are labelled.

.CC-BY-NC 4.0 International licensemade available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is

The copyright holder for this preprintthis version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.28.446107doi: bioRxiv preprint

that contact Ube2k highlighted. Red asterisks indicate residues that are unchanged from wild-type

ubiquitin. (D) Open book representation of the UbV-Ube2k interactions. On the left is Ube2k from

both complexes, while UbV.k.1, and UbV.k.2 are shown on the right. Non-contact residues are

shown in grey, while contact residues are coloured based on their molecular lipophilicity potential:

hydrophobic atoms are coloured orange, while hydrophilic atoms are cyan. Residues unchanged

from wild-type ubiquitin are indicated with asterisks. (E,F) Close up views of the interfaces

between the UbVs and Ube2k. The main chain are shown as ribbons and side chains are shown as

sticks coloured as in panels (A) and (B). Dashed lines indicate predicted hydrogen bonds. Residues

unchanged from wild-type ubiquitin are indicated with asterisks. See also Figure S5.

.CC-BY-NC 4.0 International licensemade available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is

The copyright holder for this preprintthis version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.28.446107doi: bioRxiv preprint

Figure 5. Single point mutations to UbVs are sufficient to decrease both Ube2k binding and

inhibitory activity. (A,B) The binding of the UbVs and their mutants to Ube2k was detected using

an ELISA as in Figure 2A. UbV.k.1 and its mutants are shown in panel (A), while UbV.k.2 and its

mutants are in panel B. The EC

values were calculated as the point where the absorption signal

was 50% of maximum. (C) Ubiquitin chain-building assay performed with or without the UbVs or

revertants that fully disrupted binding. Imaged as fluorescence from 5AIF-tagged ubiquitin.

.CC-BY-NC 4.0 International licensemade available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is

The copyright holder for this preprintthis version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.28.446107doi: bioRxiv preprint

Figure 6. Molecular architecture of the Ube2k-UbV structures (A) Surface representation of

Ube2k showing residues predicted to be within 3 Å of the UFD domain of the E1 ubiquitin

activating enzyme (left, in grey), and contacts with UbV.k.1 or UbV.k.2 (middle and right in yellow

or orange, respectively). Non-contacting Ube2k residues are shown in blue. (B) Model of RNF12-

bound Ube2k was generated by overlaying the E2 molecules from the UbV.k.1-Ube2k and UbV.k.2-

Ube2k structures with Ube2d2 from the RNF12-Ube2d2 crystal structure (PDB ID: 6W7Z). Only

one Ube2k molecule with UbV.k.1 is shown represented in ribbon and semi-transparent surface.

The modelled RNF12 is shown as a green ribbon. The dotted line shows the measured distance

between RNF12 and the UbVs. (C) Ube2k in complex with UbV.k.1 with donor and acceptor

ubiquitin molecules modelled at their predicted sites (Middleton and Day, 2015).

.CC-BY-NC 4.0 International licensemade available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is

The copyright holder for this preprintthis version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.28.446107doi: bioRxiv preprint

Adams, P.D., Afonine, P.V., Bunkoczi, G., Chen, V.B., Davis, I.W., Echols, N., Headd, J.J., Hung,

L.W., Kapral, G.J., Grosse-Kunstleve, R.W., et al. (2010). PHENIX: a comprehensive Python-based

system for macromolecular structure solution. Acta Crystallogr D Biol Crystallogr 66, 213–221.

Ardecky, R., Madiraju, C., Matsuzawa, S., Zou, J., Ganji, S., Pass, I., Ngo, T.A., Pinkerton, A.B.,

Sergienko, E., Su, Y., et al. (2010). Selective UBC 13 Inhibitors. In Probe Reports from the NIH

Molecular Libraries Program, (Bethesda (MD): National Center for Biotechnology Information

(US))

Bae, Y., Jung, S.H., Kim, G.-Y., Rhim, H., and Kang, S. (2013). Hip2 ubiquitin-conjugating enzyme

overcomes radiation-induced G2/M arrest. Biochimica et Biophysica Acta (BBA) - Molecular Cell

Research 1833, 2911–2921.

Berndsen, C.E., and Wolberger, C. (2011). A spectrophotometric assay for conjugation of ubiquitin

and ubiquitin-like proteins. Analytical Biochemistry.

CCP4 (1994). The CCP4 suite: programs for protein crystallography. Acta Crystallogr D Biol

Crystallogr 50, 760–763.

Ceccarelli, D.F., Tang, X., Pelletier, B., Orlicky, S., Xie, W., Plantevin, V., Neculai, D., Chou, Y.-C.,

Ogunjimi, A., Al-Hakim, A., et al. (2011). An Allosteric Inhibitor of the Human

Cdc34&nbsp;Ubiquitin-Conjugating Enzyme. Cell 145, 1075–1087.

Chen, Z.J. (2005). Ubiquitin signalling in the NF-κB pathway. Nature Cell Biology 7, 758–765.

Chen, Z., and Pickart, C.M. (1990). A 25-kilodalton ubiquitin carrier protein (E2) catalyzes multi-

ubiquitin chain synthesis via lysine 48 of ubiquitin. J Biol Chem 265, 21835–21842.

Chen, H., Wu, G., Gao, S., Guo, R., Zhao, Z., Yuan, H., Liu, S., Wu, J., Lu, X., Yuan, X., et al.

(2017). Discovery of Potent Small-Molecule Inhibitors of Ubiquitin-Conjugating Enzyme UbcH5c

from α-Santonin Derivatives. J Med Chem 60, 6828–6852.

Cheng, J., Fan, Y.-H., Xu, X., Zhang, H., Dou, J., Tang, Y., Zhong, X., Rojas, Y., Yu, Y., Zhao, Y., et

al. (2014). A small-molecule inhibitor of UBE2N induces neuroblastoma cell death via activation of

p53 and JNK pathways. Cell Death and Disease 5, e1079-12.

Cornwell, M.J., Thomson, G.J., Coates, J., Belotserkovskaya, R., Waddell, I.D., Jackson, S.P., and

Galanty, Y. (2019). Small-Molecule Inhibition of UBE2T/FANCL-Mediated Ubiquitylation in the

Fanconi Anemia Pathway. ACS Chem Biol 14, 2148–2154.

Djajawi, T.M., Liu, L., Gong, J., Huang, A.S., Luo, M., Xu, Z., Okamoto, T., Call, M.J., Huang,

D.C.S., and van Delft, M.F. (2020). MARCH5 requires MTCH2 to coordinate proteasomal turnover

of the MCL1:NOXA complex. Cell Death Differ 27, 2484–2499.

Dou, H., Buetow, L., Sibbet, G.J., Cameron, K., and Huang, D.T. (2012). BIRC7–E2 ubiquitin

conjugate structure reveals the mechanism of ubiquitin transfer by a RING dimer. Nature Structural

and Molecular Biology 19, 876–883.

Emsley, P., Lohkamp, B., Scott, W.G., and Cowtan, K. (2010). Features and development of Coot.

Acta Crystallogr D Biol Crystallogr 66, 486–501.

.CC-BY-NC 4.0 International licensemade available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is

The copyright holder for this preprintthis version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.28.446107doi: bioRxiv preprint

Ernst, A., Avvakumov, G., Tong, J., Fan, Y., Zhao, Y., Alberts, P., Persaud, A., Walker, J.R., Neculai,

A.-M., Neculai, D., et al. (2013). A Strategy for Modulation of Enzymes in the Ubiquitin System.

Science 339, 590–595.

Gabrielsen, M., Buetow, L., Nakasone, M.A., Ahmed, S.F., Sibbet, G.J., Smith, B.O., Zhang, W.,

Sidhu, S.S., and Huang, D.T. (2017). A General Strategy for Discovery of Inhibitors and Activators

of RING and U-box E3 Ligases with Ubiquitin Variants. Molecular Cell 68, 456-470.e10.

Garg, P., Ceccarelli, D.F., Keszei, A.F.A., Kurinov, I., Sicheri, F., and Sidhu, S.S. (2020). Structural

and Functional Analysis of Ubiquitin-based Inhibitors That Target the Backsides of E2 Enzymes.

Journal of Molecular Biology 432, 952–966.

Gundogdu, M., and Walden, H. (2019). Structural basis of generic versus specific E2-RING E3

interactions in protein ubiquitination. Protein Sci. 28, 1758–1770.

Helms, K.M., Wilson, R.C., Ogungbe, I.V., Setzer, W.N., and Twigg, P.D. (2011). Vitexin Inhibits

Polyubiquitin Synthesis by the Ubiquitin-conjugating Enzyme E2-25K. Natural Product

Communications 6, 1934578X1100601001.

Hochstrasser, M. (1996). Ubiquitin-dependent protein degradation. Annu Rev Genet 30, 405–439.

Hong, N.H., Tak, Y.J., Rhim, H., and Kang, S. (2019). Hip2 ubiquitin-conjugating enzyme has a

role in UV-induced G1/S arrest and re-entry. Genes Genom 41, 159–166.

Kabsch, W. (2010). Xds. Acta Crystallogr D Biol Crystallogr 66, 125–132.

Kalchman, M.A., Graham, R.K., Xia, G., Koide, H.B., Hodgson, J.G., Graham, K.C., Goldberg,

Y.P., Gietz, R.D., Pickart, C.M., and Hayden, M.R. (1996). Huntingtin is ubiquitinated and interacts

with a specific ubiquitin-conjugating enzyme. J Biol Chem 271, 19385–19394.

Koliopoulos, M.G., Esposito, D., Christodoulou, E., Taylor, I.A., and Rittinger, K. (2016).

Functional role of TRIM E3 ligase oligomerization and regulation of catalytic activity. EMBO J 35,

1204–1218.

Kothayer, H., Spencer, S.M., Tripathi, K., Westwell, A.D., and Palle, K. (2016). Synthesis and in

vitro anticancer evaluation of some 4,6-diamino-1,3,5-triazine-2-carbohydrazides as Rad6 ubiquitin

conjugating enzyme inhibitors. Bioorganic & Medicinal Chemistry Letters 26, 2030–2034.

Krissinel, E., and Henrick, K. (2007). Inference of macromolecular assemblies from crystalline

state. J Mol Biol 372, 774–797.

Liu, H., and Naismith, J.H. (2008). An efficient one-step site-directed deletion, insertion, single and

multiple-site plasmid mutagenesis protocol. BMC Biotechnol 8, 91.

Liu, L., Hua, Y., Wang, D., Shan, L., Zhang, Y., Zhu, J., Jin, H., Li, H., Hu, Z., and Zhang, W.

(2014). A Sesquiterpene Lactone from a Medicinal Herb Inhibits Proinflammatory Activity of TNF-

α by Inhibiting Ubiquitin-Conjugating Enzyme UbcH5. Chemistry & Biology 21, 1341–1350.

Lv, Z., Rickman, K.A., Yuan, L., Williams, K., Selvam, S.P., Woosley, A.N., Howe, P.H., Ogretmen,

B., Smogorzewska, A., and Olsen, S.K. (2017). S. pombe Uba1-Ubc15 structure reveals a novel

regulatory mechanism of ubiquitin E2 activity. Mol Cell 65, 699-714.e6.

.CC-BY-NC 4.0 International licensemade available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is

The copyright holder for this preprintthis version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.28.446107doi: bioRxiv preprint

Maculins, T., Garcia-Pardo, J., Skenderovic, A., Gebel, J., Putyrski, M., Vorobyov, A., Busse, P.,

Varga, G., Kuzikov, M., Zaliani, A., et al. (2020). Discovery of Protein-Protein Interaction

Inhibitors by Integrating Protein Engineering and Chemical Screening Platforms. Cell Chemical

Biology S245194562030283X.

Mark, K.G., and Rape, M. (2021). Ubiquitin-dependent regulation of transcription in development

and disease. EMBO Reports 22, e51078.

McCoy, A.J., Grosse-Kunstleve, R.W., Adams, P.D., Winn, M.D., Storoni, L.C., and Read, R.J.

(2007). Phaser crystallographic software. J Appl Crystallogr 40, 658–674.

Meiklejohn, H., Mostaid, M.S., Luza, S., Mancuso, S.G., Kang, D., Atherton, S., Rothmond, D.A.,

Weickert, C.S., Opazo, C.M., Pantelis, C., et al. (2019). Blood and brain protein levels of ubiquitin-

conjugating enzyme E2K (UBE2K) are elevated in individuals with schizophrenia. J Psychiatr Res

113, 51–57.

Merkley, N., and Shaw, G.S. (2004). Solution structure of the flexible class II ubiquitin-conjugating

enzyme Ubc1 provides insights for polyubiquitin chain assembly. The Journal of Biological

Chemistry 279, 47139–47147.

Middleton, A.J., and Day, C.L. (2015). The molecular basis of lysine 48 ubiquitin chain synthesis

by Ube2K. Scientific Reports 5, 16793.

Middleton, A.J., Budhidarmo, R., and Day, C.L. (2014). Use of E2~ubiquitin conjugates for the

characterization of ubiquitin transfer by RING E3 ligases such as the inhibitor of apoptosis proteins.

Methods in Enzymology 545, 243–263.

Middleton, A.J., Budhidarmo, R., Das, A., Zhu, J., Foglizzo, M., Mace, P.D., and Day, C.L. (2017).

The activity of TRAF RING homo- and heterodimers is regulated by zinc finger 1. Nat Commun 8,

1788.

Middleton, A.J., Zhu, J., and Day, C.L. (2020). The RING Domain of RING Finger 12 Efficiently

Builds Degradative Ubiquitin Chains. Journal of Molecular Biology 432, 3790–3801.

Olsen, S.K., and Lima, C.D. (2013). Structure of a Ubiquitin E1-E2 Complex: Insights to E1-E2

Thioester Transfer. Molecular Cell 49, 884–896.

Osborne, H.C., Irving, E., Forment, J.V., and Schmidt, C.K. (2021). E2 Enzymes in Genome

Stability: Pulling the Strings Behind the Scenes. Trends in Cell Biology.

Pettersen, E.F., Goddard, T.D., Huang, C.C., Meng, E.C., Couch, G.S., Croll, T.I., Morris, J.H., and

Ferrin, T.E. (2021). UCSF ChimeraX: Structure visualization for researchers, educators, and

developers. Protein Science 30, 70–82.

Pines, J. (2011). Cubism and the cell cycle: the many faces of the APC/C. Nat Rev Mol Cell Biol

12, 427–438.

Plechanovova, A., Jaffray, E.G., Tatham, M.H., Naismith, J.H., and Hay, R.T. (2012). Structure of a

RING E3 ligase and ubiquitin-loaded E2 primed for catalysis. Nature 489, 115–120.

.CC-BY-NC 4.0 International licensemade available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is

The copyright holder for this preprintthis version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.28.446107doi: bioRxiv preprint

Pluska, L., Jarosch, E., Zauber, H., Kniss, A., Waltho, A., Bagola, K., von Delbrück, M., Löhr, F.,

Schulman, B.A., Selbach, M., et al. (2021). The UBA domain of conjugating enzyme Ubc1/Ube2K

facilitates assembly of K48/K63‐branched ubiquitin chains. EMBO J 40.

de Pril, R., Fischer, D.F., Roos, R.A., and van Leeuwen, F.W. (2007). Ubiquitin-conjugating enzyme

E2-25K increases aggregate formation and cell death in polyglutamine diseases. Mol Cell Neurosci

34, 10–19.

Pulvino, M., Liang, Y., Oleksyn, D., DeRan, M., Van Pelt, E., Shapiro, J., Sanz, I., Chen, L., and

Zhao, J. (2012). Inhibition of proliferation and survival of diffuse large B-cell lymphoma cells by a

small-molecule inhibitor of the ubiquitin-conjugating enzyme Ubc13-Uev1A. Blood 120, 1668–

1677.

Rape, M. (2018). Ubiquitylation at the crossroads of development and disease. Nat Rev Mol Cell

Biol 19, 59–70.

Rennie, M.L., Chaugule, V.K., and Walden, H. (2020). Modes of allosteric regulation of the

ubiquitination machinery. Curr. Opin. Struct. Biol. 62, 189–196.

Sanchez, J.G., Chiang, J.J., Sparrer, K.M.J., Alam, S.L., Chi, M., Roganowicz, M.D., Sankaran, B.,

Gack, M.U., and Pornillos, O. (2016). Mechanism of TRIM25 Catalytic Activation in the Antiviral

RIG-I Pathway. Cell Reports 16, 1315–1325.

Sanders, M.A., Brahemi, G., Nangia-Makker, P., Balan, V., Morelli, M., Kothayer, H., Westwell,

A.D., and Shekhar, M.P.V. (2013). Novel Inhibitors of Rad6 Ubiquitin Conjugating Enzyme:

Design, Synthesis, Identification, and Functional Characterization. Molecular Cancer Therapeutics

12, 373–383.

Sato, Y., Yoshikawa, A., Mimura, H., Yamashita, M., Yamagata, A., and Fukai, S. (2009). Structural

basis for specific recognition of Lys 63-linked polyubiquitin chains by tandem UIMs of RAP80. 28,

2461–2468.

Scheuermann, T.H., and Brautigam, C.A. (2015). High-precision, automated integration of multiple

isothermal titration calorimetric thermograms: new features of NITPIC. Methods 76, 87–98.

Scott, D.C., Sviderskiy, V.O., Monda, J.K., Lydeard, J.R., Cho, S.E., Harper, J.W., and Schulman,

B.A. (2014). Structure of a RING E3 trapped in action reveals ligation mechanism for the ubiquitin-

like protein NEDD8. Cell 157, 1671–1684.

Sidhu, S.S., Lowman, H.B., Cunningham, B.C., and Wells, J.A. (2000). [21] Phage display for

selection of novel binding peptides. In Methods in Enzymology, J. Thorner, S.D. Emr, and J.N.

Abelson, eds. (Academic Press), pp. 333-IN5.

Song, S., Kim, S.Y., Hong, Y.M., Jo, D.G., Lee, J.Y., Shim, S.M., Chung, C.W., Seo, S.J., Yoo, Y.J.,

Koh, J.Y., et al. (2003). Essential role of E2-25K/Hip-2 in mediating amyloid-beta neurotoxicity.

Mol Cell 12, 553–563.

Stewart, M.D., Ritterhoff, T., Klevit, R.E., and Brzovic, P.S. (2016). E2 enzymes: more than just

middle men. Cell Res 26, 423–440.

.CC-BY-NC 4.0 International licensemade available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is

The copyright holder for this preprintthis version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.28.446107doi: bioRxiv preprint

Strickson, S., Campbell, D.G., Emmerich, C.H., Knebel, A., Plater, L., Ritorto, M.S., Shpiro, N.,

and Cohen, P. (2013). The anti-inflammatory drug BAY 11-7082 suppresses the MyD88-dependent

signalling network by targeting the ubiquitin system. Biochemical Journal 451, 427–437.

Su, J., Huang, P., Qin, M., Lu, Q., Sang, X., Cai, Y., Wang, Y., Liu, F., Wu, R., Wang, X., et al.

(2018). Reduction of HIP2 expression causes motor function impairment and increased

vulnerability to dopaminergic degeneration in Parkinson’s disease models. Cell Death Dis 9, 1020.

Teyra, J., Singer, A.U., Schmitges, F.W., Jaynes, P., Kit Leng Lui, S., Polyak, M.J., Fodil, N.,

Krieger, J.R., Tong, J., Schwerdtfeger, C., et al. (2019). Structural and Functional Characterization

of Ubiquitin Variant Inhibitors of USP15. Structure 27, 590-605.e5.

Tokgoz, Z., Siepmann, T.J., Streich, F., Kumar, B., Klein, J.M., and Haas, A.L. (2011). E1-E2

Interactions in Ubiquitin and Nedd8 Ligation Pathways. Journal of Biological Chemistry 287, 311–

321.

Vijay-Kumar, S., Bugg, C.E., and Cook, W.J. (1987). Structure of ubiquitin refined at 1.8 A

resolution. J Mol Biol 194, 531–544.

Wang, C., Shi, G., and Ji, X. (2018). Design, synthesis, and anticancer activity evaluation of

irreversible allosteric inhibitors of the ubiquitin-conjugating enzyme Ube2g2. Med. Chem.

Commun. 9, 1818–1825.

Wang, H., Zhang, C., Rorick, A., Wu, D., Chiu, M., Thomas-Ahner, J., Chen, Z., Chen, H., Clinton,

S.K., Chan, K.K., et al. (2011). CCI-779 Inhibits Cell-Cycle G2–M Progression and Invasion of

Castration-Resistant Prostate Cancer via Attenuation of UBE2C Transcription and mRNA Stability.

Cancer Res 71, 4866–4876.

Williams, K.M., Qie, S., Atkison, J.H., Salazar-Arango, S., Alan Diehl, J., and Olsen, S.K. (2019).

Structural insights into E1 recognition and the ubiquitin-conjugating activity of the E2 enzyme

Cdc34. Nat Commun 10, 3296.

.CC-BY-NC 4.0 International licensemade available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is

The copyright holder for this preprintthis version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.28.446107doi: bioRxiv preprint

Figure

S1.

Each

UbV

binds

Ube2k.

The

binding

of the

UbVs

to the

Ube2k~Ub conjugate, Ube2k, GST-ubiquitin, and controls was detected using an ELISA as in

Figure 2A (n = 1 or 2). GST: Glutathione S transferase; GST-Ub: GST-Ubiquitin; BSA: bovine

serum albumin; SA: streptavidin.

.CC-BY-NC 4.0 International licensemade available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is

The copyright holder for this preprintthis version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.28.446107doi: bioRxiv preprint

Figure S2. Both UbVs bind tightly and specifically to Ube2k. (A) Chain-building assay

performed with and without UbV.k.1 and UbV.k.2 with (left) Ube2n/Ube2v2 and (right) Ube2d2.

Gel was imaged by fluorescence of Ub-5AIF. (B) Thermograms (top) and fitted isotherms (bottom)

of ITC measurements performed in technical triplicate (shown in blue, red, and orange) where

either UbV.k.1 or UbV.k.2 was injected into Ube2k-UBC. (C) Thermograms and isotherms of

control reactions where PBS was injected into Ube2k and Ube2k-UBC (n=1). (D) Thermal

denaturation experiment performed at increasing temperatures with SYPRO orange dye to detect

unfolding protein. Performed in technical duplicate. Melting points were calculated with R v4.0.5.

.CC-BY-NC 4.0 International licensemade available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is

The copyright holder for this preprintthis version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.28.446107doi: bioRxiv preprint

Figure S3. E1 charging assay to assess inhibition by UbVs. (top) Schematic of E1 charging

experiment. The red asterisk indicates fluorescent ubiquitin. The ubiquitin activating E1 enzyme

was thiolated with fluorescently-tagged ubiquitin with and without UbVs and analysed for

fluorescence (left) and Coomassie-stained (right). Asterisks indicate unidentified bands.

.CC-BY-NC 4.0 International licensemade available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is

The copyright holder for this preprintthis version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.28.446107doi: bioRxiv preprint