A small-molecule inhibitor of the ubiquitin activating enzyme for cancer treatment
Marc L Hyer1–3, Michael A Milhollen1,3, Jeff Ciavarri1, Paul Fleming1, Tary Traore1, Darshan Sappal1, Jessica Huck1, Judy Shi1, James Gavin1, Jim Brownell1, Yu Yang1, Bradley Stringer1, Robert Griffin1, Frank Bruzzese1, Teresa Soucy1, Jennifer Duffy1, Claudia Rabino1, Jessica Riceberg1, Kara Hoar1,
Anya Lublinsky1, Saurabh Menon1, Michael Sintchak1, Nancy Bump1, Sai M Pulukuri1, Steve Langston1, Stephen Tirrell1, Mike Kuranda1, Petter Veiby1, John Newcomb1, Ping Li1, Jing Tao Wu1, Josh Powe1, Lawrence R Dick1 , Paul Greenspan1, Katherine Galvin1, Mark Manfredi1, Chris Claiborne1, Benjamin S Amidon1 & Neil F Bence1,2

The ubiquitin–proteasome system (UPS) comprises a network of enzymes that is responsible for maintaining cellular protein homeostasis. The therapeutic potential of this pathway has been validated by the clinical successes of a number of UPS modulators, including proteasome inhibitors and immunomodulatory imide drugs (IMiDs). Here we identified TAK-243 (formerly known as MLN7243) as a potent, mechanism-based small-molecule inhibitor of the ubiquitin activating enzyme (UAE), the primary mammalian E1 enzyme that regulates the ubiquitin conjugation cascade. TAK-243 treatment caused depletion of cellular ubiquitin conjugates, resulting in disruption of signaling events, induction of proteotoxic stress, and impairment of cell cycle progression and DNA damage repair pathways. TAK-243 treatment caused death of cancer cells and, in primary human xenograft studies, demonstrated antitumor activity at tolerated doses. Due to its specificity and potency, TAK-243 allows for interrogation of ubiquitin biology and for assessment of UAE inhibition as a new approach for cancer treatment.

Ubiquitin conjugation in mammals is initiated by two key enzymes, the ubiquitin activating enzymes UAE (also known as UBE1) and UBA6, which are collectively referred to as E1 enzymes. UAE (encoded by the UBA1 gene) is responsible for charging an estimated
>99% of cellular ubiquitin, whereas UBA6 is responsible for charging
<1% of ubiquitin. UAE catalyzes ubiquitin-charging of all (~35) E2 cellular ubiquitin-conjugating enzymes, except for the E2 USE1, whose charging is catalyzed by UBA6 (ref. 1). Ubiquitin-charged E2 enzymes cooperate with cellular E3 ligases (for example, cullin, Ring, HECT, RBR or U-box) to direct specific cellular target protein ubiquitylation modifications, which are classified as monomeric or polyubiquitin; moreover, polyubiquitin chains can be of the Lys11 (K11), Lys29 (K29), Lys48 (K48) or Lys63 (K63) type. These varied ubiquitin modifications are read by ubiquitin-binding proteins and can dictate outcomes in which the target proteins are either degraded or not. For example, K48-linked polyubiquitin is associated with proteasome-mediated degradation, K63-linked polyubiquitin can mediate autophagy and signal transduction, polyubiquitylation has been shown to be important for signal transduction mediated by the transcription factor NF-B2,3, mono-ubiquitylation of histones can alter gene regulation, and mono-ubiquitylation of surface receptors can modulate their internalization and lysosomal proteolysis.

The clinical success of the proteasome inhibitor bortezomib4 has piqued interest in targeting other components of the UPS for cancer therapy. Although the structural and mechanistic diversity of UPS enzymes has presented challenges for the small-molecule interrogation of E1, E2, E3 and deubiquitinating (DUB) enzymes in cancer biology, progress is emerging with ongoing clinical evaluation of pevonedistat (also known as TAK-924 and MLN4924; an inhibitor of the E1 for the ubiquitin-like molecule NEDD8)5, second-generation IMiDs, inhibitors of the E3 ligases for inhibitor of apoptosis protein (IAP) and murine double minute 2 (MDM2), and the valosin-containing protein (VCP) inhibitor CB-5083 (refs. 6–9).
UBA1 is an essential gene in yeast10,11. Although no data has been published on the effects of genetic ablation of Uba1 in mice, it is likely to cause lethality in this species as well. Here we report iden- tification of TAK-243 (MLN7243), a first-in-class inhibitor of UAE. TAK-243 is a mechanism-based inhibitor that potently inhibits UAE via formation of a TAK-243–ubiquitin adduct. Treatment of cells in vitro with TAK-243 led to loss of cellular ubiquitin conjugates, resulting in defective ubiquitin-dependent protein turnover and sig- naling, impaired cell cycle progression and defective DNA repair, increased proteotoxic stress, and ultimately cancer cell death. TAK- 243 treatment of tumor cells in vivo caused a dramatic reduction of

1Takeda Pharmaceuticals Inc., Cambridge, Massachusetts, USA. 2Present addresses: Agios Pharmaceuticals, Cambridge, Massachusetts, USA (M.L.H.) and Nurix, Inc., San Francisco, California, USA (N.F.B.). 3These authors contributed equally to this work. Correspondence should be addressed to M.A.M. ([email protected]).
Received 30 September 2016; accepted 19 December 2017; published online 15 January 2018; doi:10.1038/nm.4474

cellular polyubiquitination and induced pronounced antitumor activ- ity in mice bearing human xenograft tumors.
TAK-243 forms a TAK-243–ubiquitin adduct and potently inhibits UAE in vitro
We undertook a medicinal-chemistry-based effort to generate small- molecule inhibitors of UAE, which yielded >700 potential candidates. We then screened these candidates using an in vitro biochemical assay that measured UAE-mediated transfer of ubiquitin to an E2 substrate (UBCH10) and identified a potent, mechanism-based UAE inhibitor, TAK-243 (Fig. 1a). TAK-243 had a half-maximal inhibitory concentration (IC50) value of 1  0.2 nM (Fig. 1b) in the UBCH10 E2 thioester assay. Using similar biochemical assays, we found that TAK-243 had weaker inhibitory activity against other closely related E1 ubiquitin-like activating enzymes such as Fat10- activating enzyme (UBA6; 7  3 nM), NEDD8-activating enzyme (NAE; 28  11 nM), SUMO-activating enzyme (SAE; 850  180 nM), ISG15-activating enzyme (UBA7; 5,300  2,100 nM) and autophagy- activating enzyme (ATG7; >10,000 nM) than it did against UAE. Consistent with the concept that TAK-243 inhibits UAE by using a

substrate-assisted mechanism of action similar to that described previously for the small-molecule NAE inhibitor pevonedistat5, X-ray crystallographic studies revealed that TAK-243 was bound to a site of UAE (humanized yeast UAE) normally occupied by AMP (Fig. 1c). Notably, a continuous electron-density unit was observed to con- nect the C terminus of ubiquitin to the sulfamate moiety of TAK-243 (Fig. 1c and Supplementary Table 1), indicating the presence of a covalent TAK-243–ubiquitin adduct. We attempted to crystallize human UAE bound to TAK-243; however, crystals were difficult to obtain, and they diffracted to a lower resolution as compared to that with a humanized yeast UAE version, which has recently been used by others12. Additional biochemical assays (Supplementary Fig. 1) dem- onstrated that TAK-243 is a time-dependent human UAE inhibitor (Supplementary Fig. 1a,b) that has a mechanism of action consist- ent with a substrate-assisted mechanism previously described for the small-molecule NAE inhibitor pevonedistat5. The TAK-243–ubiquitin adduct (Fig. 1c), once formed, remained tightly bound to UAE and blocked UAE catalytic activity (Supplementary Fig. 1a). A transthi- olation assay showed that TAK-243 inhibited UAE from transfer- ring ubiquitin to an E2 enzyme (Supplementary Fig. 1c). TAK-243 showed high selectivity, as indicated by minimal inhibitory activity

a F F


b d TAK-243 (M)
100 UBCH10Ub

Mr (kDa) 17

75 USE1Ub
USE1 38
50 UBC12 21











UBC9 18
ATG7 78

[TAK-243] (M)

Figure 1 TAK-243 is a mechanism-based, cell-active inhibitor of UAE. (a) Chemical structure of the adenosine sulfamate UAE inhibitor TAK-243.
(b) Dose-response inhibition profiles of TAK-243 for the E1 enzymes UAE, UBA6, NAE, SAE, UBA7 and ATG7. Data represent the mean  s.d. of
n = 2 experiments run with duplicate samples. (c) Structure, as shown using a ribbon diagram (left) and an electron density map (right), of the TAK- 243–ubiquitin adduct species in which TAK-243 forms a covalent bond with the C terminus of ubiquitin. The electron density map shows that the TAK- 243–ubiquitin adduct occupies the adenylate (AMP)-binding site of UAE. (d) Representative western blot analysis (of n = 2 independent experiments) for the response of HCT-116 cells to different concentrations of TAK-243. The status of ubiquitin and ubiquitin-like protein charging on the E1 and E2 enzymes UBCH10 (UAE-specific ubiquitin E2), USE1 (UBA6-specific ubiquitin E2), UBC12 (NEDD8 E2), UBC9 (SUMO E2) and ATG7 (autophagy UBL E1) was assessed.


Mr (kDa) 188


1 2 4 8 16 24 Time (h) b

TAK-243 (M)


TAK-243–Ub adduct


DMSO control 2n 4n

50 100 150

0.25 M MLN4924


1 M TAK-243
2n 4n

50 100 150

0.5 M TAK-243

43 c-Jun
65 c-Myc
40 MCL1

53 p53
55 Tubulin
c 1 2 4 8 16 24 Time (h)

TAK243 (M)




2n 4n >4n

50 100 150
FL2 (DNA content)

TAK-243 (M)



2n 4n

50 100 150

TAK-243 (M)

Mr (kDa)

p-IRE1 (Ser724) XBP-1S

(kDa) 17



Time (h)



(kDa) 23








p-eIF2A (Ser51) ATF4
Cleaved PARP

Cleaved caspase-3


20 J/m2 UV


1 M TAK-243


1 M TAK-243

55 Tubulin
Figure 2 TAK-243 inhibits cellular ubiquitin conjugation, which leads to substrate stabilization, cell cycle arrest, ER stress and an impaired DNA damage response. (a) Representative western blot analysis (of n = 2 independent experiments) showing dose response and time course of TAK-243 target engagement in HCT-116 cells, as assessed by immunoblotting for the TAK-243–ubiquitin adduct (MIL90), polyubiquitin (polyUb), ubiquitylated histone H2B, c-Jun, c-Myc, MCL1, XIAP and p53. Tubulin was used as a loading control; UAE (UBE1) levels are also shown. (b) FACS analysis showing DNA content in HCT-116 cells at 24 h after TAK-243 treatment. FL2 represents PI+ cells. The NAE inhibitor pevonedistat was used for comparison.
Plots are representative of n = 2 independent experiments. (c) Representative western blot analysis (of n = 2 independent experiments) showing dose response and time course of the effects of TAK-243 on the UPR and apoptosis in HCT-116 cells, as assessed by immunoblotting for BIP, ATF6, PERK, phosphorylated ERN1 (p-IRE1a), XBP1s, phosphorylated eIF2a (p-eIF2a), ATF4, CHOP and GADD34, cleaved caspase-3 and cleaved PARP. The PERK-specific antibody recognizes phosphorylated and unphosphorylated PERK; the high-molecular-weight band stained by anti-PERK corresponds to phosphorylated PERK. Tubulin was used as a loading control. (d) E2–ubiquitin enzyme charging in HCT-116 cells at 8 h after TAK-243 treatment, as
assessed by western blotting. The following UAE-specific E2 enzymes were evaluated: UBE2A, UBE2B, UBE2T, PCNA and FANCD2. UAE was included as a loading control. Western blots are representative of n = 2 independent experiments. (e) Representative images (of n = 2 independent experiments) showing DNA damage repair, as assessed by COMET assays, in Calu-6 cells that were treated with UV radiation and then treated with DMSO or TAK-243 (1 M) for 0 or 6 h. Scale bar, 10 m. See Supplementary Figure 7 for quantification and dose response.

in a panel of kinase and receptor assays, as well as on human car- bonic anhydrase type I and type II (Supplementary Table 2). Taken together, these data indicate that UAE and UBA6, the ubiquitin E1 enzymes, are the primary biochemical targets of TAK-243.
Next we characterized the selectivity of TAK-243 in a well-character- ized xenograft-amenable HCT-116 colon cancer cell line using a panel of ubiquitin-like (UBL)-thioester charging assays. These assays meas- ure the efficiency of an E1 enzyme, such as UAE, to transfer ubiquitin or a UBL-like moiety to an E2 enzyme. Consistent with the in vitro results using purified UAE, TAK-243 showed strong selectivity over the Sumo and autophagy UBL pathways, as indicated by the negligible effect of TAK-243 treatment on the UBC9SUMO:UBC9 charging ratio

and by a slight increase in the ATG7UBL:ATG7 charging ratio (Fig. 1d). The latter effect may reflect induction of autophagy due to activation of the integrated stress response. TAK-243 was tenfold more selec- tive against UAE than against NAE, as indicated by its effects on the UBCH10Ub and UBC12NEDD8 charging ratios. TAK-243 was showed equally potent inhibition of the two E1 enzymes capable of activating ubiquitin (UBA6 and UAE), as indicated by comparable decreases in levels of charged USE1 (USE1Ub) and UBCH10 (UBCH10Ub), using the cell-based E2~UBL thioester assay (Fig. 1d).
We also generated a UBA6-specific small-molecule inhibitor (US patent US9593121B2; section 067, figure I-01)13 as a tool to inter- rogate the consequences of selectively inhibiting UBA6 activity.

Table 1 Anti-proliferative EC50 values of TAK-243 in cell lines
Cell line Origin TAK-243 (M)
HCT116 Colon 0.020  0.01 (16)
HT29 Colon 0.308  0.08 (4)
LOVO Colon 0.450  0.14 (3)
SW480 Colon 0.044  0.01 (3)
SW48 Colon 0.056  0.02 (10)
H460 Lung 1.31  0.19 (9)
Calu6 Lung 0.174  0.17 (7)
A549 Lung 0.816  0.38 (6)
NCI-H520 Lung 0.013  0.01 (4)
NCI-H69 Lung 0.011  0.01 (3)
NCI-H82 Lung 0.006  0.00 (3)
MDA-MB-231 Breast 0.172  0.14 (4)
HCC1954 Breast 0.123  0.03 (4)
A375 Melanoma 0.381  0.28 (10)
WM266.4 Melanoma 0.063  0.01 (3)
MIA-PACA-2 Pancreatic 0.045  0.02 (3)
HL60 Leukemia 0.017  0.01 (6)
MV-411 Leukemia 0.015  0.01 (4)
THP-1 Leukemia 0.019  0.01 (4)
OCI-M2 Leukemia 0.132  0.04 (4)
MM1.S Multiple myeloma 0.012  0.01 (4)
RPMI-8226 Multiple myeloma 1.099  0.22 (4)
NCI-H929 Multiple myeloma 0.109  0.01 (3)
WSU-DLCL2 Lymphoma 0.016  0.01 (5)
MCF10A cycling Breast epithelial 0.033  0.01 (3)
MCF10A non-cycling Breast epithelial 0.021  0.01 (3)
Normal dermal fibroblast Human fibroblast 0.934  0.05 (2)
UBA6 (WT) MEFs Mouse embryonic fibroblast 0.792  0.39 (3)
UBA6-null MEFs Mouse embryonic fibroblast 2.580  1.06 (3)
UBE1 overexpression HCT116 Colon 0.021  0.01 (3)
UBE1 vector control HCT116 Colon 0.036  0.02 (3)
The indicated tumor cell lines, as well as normal dermal human fibroblasts and MEFs that were either wild type (WT) or a knockout for UBA6, were treated as described in the Online Methods, and the concentration at half maximal viability (EC50) value  s.d. is reported. UBE1-overexpressing HCT-116 cells are further described in Supplemen- tary Figure 2b. Numbers in parentheses represent the number of times the assay was performed.

Treatment of HCT-116 cells with this UBA6 inhibitor had negligible effects on the levels of bulk polyubiquitin chains, NEDDylated cul- lins and high-molecular-weight SUMOylated species, despite showing robust, specific inhibition of the UBA6-specific charging of the E2 enzyme USE1 (Supplementary Fig. 2a–d).
To further characterize TAK-243, we evaluated its effects on a panel of UAE-specific E2 enzymes (UBCH10, UBC1, UBC2, UBC3, UBC5, UBC13, UBE2T, UBE2A and UBE2B) using the cellular E2~UBL thioester assay (Supplementary Fig. 2e,f). In these experi- ments, HCT-116 colon cancer cells and WSU-DLCL2 lymphoma cells were treated with TAK-243, and western blotting analysis was used to measure ubiquitin-charged E2s. TAK-243 treatment resulted in a uniform, dose-dependent loss of ubiquitin-charged E2~UBL thioesters (Supplementary Fig. 2e,f), consistent with the conclu- sion that TAK-243 treatment leads to uniform disruption of cellular ubiquitin-conjugation events. Moreover, TAK-243 treatment reduced the global levels of K48-linked polyubiquitylated chains in HCT- 116 cells in a dose-dependent manner (Supplementary Fig. 2g). As expected, TAK-243 modulated E2 thioester charging and K48-linked

chain formation at comparable half-maximal effective concentration (EC50) values.
TAK-243 inhibits the turnover of short-lived proteins and disrupts cell cycle progression
Because ubiquitin is known to target proteins for degradation, we evaluated the effect of UAE inhibition on the turnover of short- lived proteins by using a cellular assay that monitors the release of 35S-methionine from metabolically labeled proteins. Side-by-side controls in this experiment included inhibitors of the proteasome (ixazomib), NAE (pevonedistat) and UBA6 (UBA6i) (Supplementary Fig. 3). TAK-243 treatment resulted in protein turnover inhibition similar to that observed with a proteasome inhibitor. In contrast, UBA6 inhibition had no effect on protein turnover, and inhibition of NAE had a modest effect (20%) (Supplementary Fig. 3), consist- ent with previous work14. Inhibition of the proteasome and UAE decreased protein turnover to similar extents, suggesting the bulk of proteasome-dependent protein turnover is ubiquitin dependent (Supplementary Fig. 3).
We further explored the effect of TAK-243 on protein turnover by measuring the effects of TAK-243 treatment on the stability of specific short-lived oncoproteins and tumor suppressors (c-Jun, c- Myc, MCL1, XIAP and p53), which are dependent on ubiquitin for degradation. The solid tumor cell line, HCT-116, was selected for this experiment because, unlike the proteasome inhibitor bortezomib, TAK-243 demonstrated broad antitumor activity in models of solid and hematological tumors (see below). An antibody (MIL90) raised against the TAK-243–ubiquitin adduct species was used to detect adduct formation and target engagement in cells12 (Supplementary Fig. 4). At the earliest time point examined (1 h), TAK-243–ubiquitin adduct formation was clearly evident in HCT-116 cells (Fig. 2a). In addition, downstream UAE pathway inhibition by TAK-243 was evi- dent, as shown by a dose- and time-dependent loss of both polyubiq- uitin chains and mono-ubiquitylated histone H2B; however, TAK-243 treatment did not affect UAE (UBE1) protein levels (Fig. 2a). TAK- 243 treatment also caused accumulation of short-lived proteins such as c-Jun, c-Myc, MCL1 and p53 (Fig. 2a), consistent with the concept that loss of polyubiquitylated chains impairs the targeting of proteins for proteasomal degradation. For example, TAK-243 treatment led to accumulation of both p53 and c-Myc, both of which have been reported to have half-lives of <60 min in HCT-116 cells15,16.
Ubiquitin is known to have an essential role in cell cycle regula- tion15. To examine the effect of UAE inhibition on cell cycle regula- tion, we used flow cytometric analysis to measure the DNA content in TAK-243-treated HCT-116 cells. TAK-243 treatment resulted in a complete G2/M arrest at concentrations near its EC50 value (0.05 M); however, at higher concentrations (>EC90, 1 M), TAK-243 treat- ment resulted in arrest at both the G1 and G2/M phases of the cell cycle, which was suggestive of the importance of ubiquitylation on proteins that are critical for proper passage through the cell cycle (Fig. 2b). Protein levels of the cell cycle markers cyclin B and cyclin D1 and of the cyclin-dependent kinase (CDK) inhibitors p21 and p27 correlated with the cell cycle DNA profiles observed in TAK- 243-treated cells (Supplementary Fig. 5a). Notably, the dominant cell cycle phenotype observed after TAK-243 treatment was different than that after treatment with an NAE inhibitor but was more similar to that after treatment with a proteasome inhibitor (ixazomib)16. In particular, NAE inhibition (using pevonedistat) has been reported to induce an S-phase re-replication phenotype (DNA content  4N) through stabilization and dysregulation of chromatin licensing and

DNA replication factor 1 (CDT1) turnover14,17; in contrast, a re- replication phenotype was not observed in TAK-243-treated cells, as no cell population with DNA content  4N was detected (Fig. 2b). Unlike pevonedistat, TAK-243 treatment stabilized both CDT1 and its endogenous inhibitor geminin (Supplementary Fig. 5b), which would presumably prevent re-replication. Additionally, unlike pev- onedistat11,18, TAK-243 treatment did not cause loss of the mitotic cell cycle protein histone H3 that is phosphorylated on Ser10 (pH3 (Ser10)) (Supplementary Fig. 5a).
TAK-243 induces irresolvable endoplasmic reticulum stress Impaired ubiquitylation is anticipated to profoundly affect protein quality control, especially within the endoplasmic reticulum (ER). Ubiquitin is known to be important for proper retro-translocation of misfolded proteins from the ER to the cytosol, and thus, unlike proteasome inhibition, inhibition of ubiquitylation by TAK-243 would be expected to enhance accumulation of misfolded proteins within the ER membrane and lumen19. Consistent with this hypoth- esis, TAK-243 treatment of HCT-116 cells led to robust activation of the unfolded protein response (UPR), as evidenced by increased phosphorylation of the kinase PERK (PERK mobility shift) and of the protein endoplasmic reticulum to nucleus signaling 1 (ERN1; also known as IRE-1) as well as accumulation of ‘activating transcription factor 6’ (ATF6) (Fig. 2c). Moreover, TAK-243 treatment induced rapid expansion of the ER surface area and led to hyper-vesiculization within this organelle (Supplementary Fig. 6), strongly suggesting the induction of ER stress. TAK-243 treatment also led to activa- tion of the downstream ER stress signals XBP1s and phosphorylated eIF2, and ATF4 (suggestive of prolonged ER stress), induced acti- vation of the mediators of ER-stress-induced apoptosis CHOP and GADD34, and ultimately resulted in cellular apoptosis, as indicated by the accumulation of cleaved PARP (poly(ADP-ribose) polymerase
1) and cleaved caspase-3 (Fig. 2c).

TAK-243 treatment alters DNA damage repair
Ubiquitylation is critical for coordinated regulation of the DNA damage response18,20,21. TAK-243 treatment led to impairment in the charging of multiple E2 enzymes associated with DNA damage repair, including UBC13, UBE2T, UBE2A and UBE2B (Fig. 2d and Supplementary Fig. 2e). Therefore TAK-243 treatment would be expected to impair multiple DNA repair pathways, including homol- ogous recombination, interstrand cross-link repair and post-repli- cative repair. Consistent with its effects on the charging of UBE2T, UBE2A and UBE2B, TAK-243 inhibited the mono-ubiquitylation of PCNA and FANCD2, two key substrates for interstrand cross-linking and post-replicative repair, with no effects on UBE1 protein levels (Fig. 2d). To test for a functional consequence of TAK-243 on DNA repair, we directly assessed DNA damage following ultraviolet (UV) irradiation using the COMET assay22. The lung carcinoma cell line Calu-6 was used for improved COMET assay reproducibility. Calu-6 cells were pretreated for 1 h with 1 M TAK-243 or DMSO (control) and then exposed to UV irradiation (20 J/m2). Immediately following irradiation (time t = 0), damaged DNA (comet tails) was observed in both the DMSO-treated and TAK-243-treated cells (Fig. 2e). At 6 h after UV irradiation, there were far fewer comet tails in cells that were treated with DMSO (control), suggesting that the DNA damage was repaired; however, TAK-243-treated cells showed persistent, unre- solved DNA damage (Fig. 2e and Supplementary Fig. 7). TAK-243 treatment alone (in the absence of UV irradiation) did not induce DNA damage (Supplementary Fig. 7). These results are consistent

with previous work linking ubiquitin to DNA damage repair23 and suggest that TAK-243 could enhance the antitumor activity of clini- cally validated DNA damaging agents. To test this hypothesis in vivo, we combined TAK-243 treatment with beam-focused radiation in two primary xenograft (PDX) models of cancer (breast and non-small-cell lung cancer (NSCLC)). Consistent with our hypothesis, combined treatment led to complete tumor responses in both of these models (Supplementary Fig. 8).
TAK-243 has anti-proliferative activity in human cancer cells Given our findings that UAE inhibition leads to proteotoxic stress (ER stress and UPR), impaired cell cycle progression and inhibition of DNA repair, we reasoned that TAK-243 would display a broad anti-prolif- erative activity against cancer cells. We assessed the anti-proliferative effect of TAK-243 on a panel of cell lines derived from hematologic and solid tumors. TAK-243 treatment induced complete cell killing of all of the lines tested, with variable EC50 values that ranged from
0.006 M to 1.31 M (Table 1). Many of the tumor cell lines demon-
strated EC50 values in the double-digit nanomolar range. In contrast, TAK-243 showed weaker cytotoxic effects (1 M) on normal human dermal fibroblasts, suggesting that a therapeutic window might exist for TAK-243 treatment in vivo. The anti-proliferative effects of TAK- 243 are likely to be driven by apoptosis, as apoptotic cell death mark- ers (cleaved PARP and cleaved caspase-3) were detectable following treatment (Fig. 2c). Of note, TAK-243 demonstrated substantial anti- proliferative effects in vitro in multiple-myeloma-derived cell lines (Table 1), an indication for which proteosome inhibitors have shown clinical benefit. For example, bortezomib and TAK-243 had similar EC50 values in vitro for MM1.S multiple-myeloma cells (0.011  0.001
M (n = 4) and 0.012  0.001 M (n = 4), respectively). The potency of TAK-243 showed no correlation with UBA1 mRNA levels across a panel of tumor cell lines and was not significantly affected by UAE overexpression (Table 1 and Supplementary Fig. 9a,b). A lack of correlation between potency and target levels is not uncommon, as similar observations have been noted for other drugs that target the kinase PLK or MEK, the proto-oncoprotein b-RAF and NAE24–27. Moreover, TAK-243 potency did not correlate with proliferation rates of the tumor cell lines tested (Supplementary Fig. 9c).
In addition to inhibiting UAE, TAK-243 was also found to inhibit UBA6 in cells (Fig. 1d). We therefore attempted to tease apart the contribution of UBA6 inhibition on cell viability. UBA6-knockout cells were used to test the hypothesis that lack of UBA6 confers sen- sitivity of mouse embryonic fibroblasts (MEFs) to TAK-243. Viability experiments comparing TAK-243-treated wild-type (WT) and UBA6- knockout MEFs indicated that, if anything, UBA6-knockout cells were more resistant to TAK-243 treatment than WT MEFs (approximately threefold decrease in sensitivity) (Table 1). Additionally, anti-pro- liferative effects of UBA6-specific inhibitors were observed only at concentrations 10- to 100-fold higher than that required for TAK-243 in the cell lines tested (data not shown). Collectively these findings support the idea that TAK-243 exerts its cell viability effect through inhibition of UAE.
Pharmacokinetic and pharmacodynamic analysis of TAK-243 in tumor bearing mice
We characterized the pharmacokinetic (PK) parameters of TAK-243 in immunocompromised CB-17 SCID mice that bore a subcutane- ous diffuse, large B cell lymphoma (WSU-DLCL2) tumor following an acute, intravenously administered dose of TAK-243. Both plasma and WSU-DLCL2 tumor tissues from these mice were analyzed for

exposure to TAK-243. In plasma, the TAK-243 parent compound was characterized by a high clearance (CL) rate (3.99 to 4.99 liter per h per kg body weight (liter/h/kg)) and a short terminal half-life (t1/2) (0.2–0.4 h) (Supplementary Fig. 10a). TAK-243 rapidly forms a TAK- 243–ubiquitin adduct, and our PK measurements did not account for TAK-243 that was sequestered as a ubiquitin adduct. Despite the high plasma clearance rate observed for TAK-243, the drug showed a high volume of distribution at steady state (Vss) (1.13–1.74 liter/h/kg) with a 5.8- to 8-fold higher area-under-the-curve (AUC) drug exposure in tumor tissue relative to that in plasma, and a prolonged tumor t1/2 of
16.9–20.6 h (Supplementary Fig. 10b).


b 120






0 10 20 30 40 50
Time (h)

Having demonstrated substantial TAK-243 exposure in tumors, c
we next evaluated TAK-243 target engagement with UAE and down- stream pathway modulation in tumor samples. We developed immu- nohistochemistry (IHC) assays to analyze several pharmacodynamic (PD) biomarkers associated with UAE inhibition. The monoclonal antibody raised against the TAK-243–ubiquitin adduct (MIL90) was used to monitor UAE target engagement in tumors. Downstream pathway inhibition was evaluated using several biomarkers; antibodies were used to detect both polyubiquitylated and mono-ubiquitylated proteins (using the FK2 antibody, which detects all cellular forms of ubiquitin conjugates), mono-ubiquitylated histone H2B and cleaved caspase-3 (as a measure of apoptosis). Mice bearing either WSU- DLCL2 lymphoma or PHTX-132Lu (primary NSCLC) xenograft tumor tissues were dosed using a single intravenous injection of TAK- 243 at its maximum tolerated dose. Both UAE target engagement, as assessed by the presence of the TAK-243–ubiquitin adduct, and downstream pathway inhibition were clearly evident in tumor tis- sue (Fig. 3a). Following a single dose, TAK-243 induced a rapid and prolonged PD response (Fig. 3b,c), consistent with pronounced UAE inhibition in tumor tissue. Of the PD biomarkers evaluated, adduct formation was detected most sensitively, and the adduct could be

0 20 40 60 80
Time (h)
0 10 20 30 40 50
Time (h)

detected at very low TAK-243 doses (3 mg per kg body weight (mg/kg) was the lowest dose examined) that were insufficient to modulate downstream PD readouts (polyubiquitin or mono-ubiquitylated histone H2B) or antitumor efficacy (data not shown). The adduct could be detected 0.5–1.0 h before downstream biomarker modula- tion, a lag which could be due to a combination of differences in assay sensitivity, de novo UAE synthesis and drug washout in the tumor over time. Of note, at its maximally tolerated dose, TAK-243 showed negligible inhibition of NAE-dependent neddylated cullin levels in HCT-116 xenografts, despite profound inhibition of polyubiquitin formation and ubiquitylated histone H2B levels (Supplementary Fig. 11a,b). These data indicate TAK-243 profoundly affects down- stream UAE biomarkers while having little-to-no impact on NAE- associated biomarkers in vivo.
TAK-243 has UAE-specific antitumor efficacy in vivo
The antitumor activity of TAK-243 was determined with a panel of human-patient-derived and cell-line-derived xenograft (PDX and CDX, respectively) tumor models that represented both solid and hematological cancers. We established subcutaneous tumors in mice using the following CDX models: WSU-DLCL2 (diffuse, large B cell lymphoma), HCT-116 (colon carcinoma), THP-1 (acute myeloid leukemia), CWR22 (prostate cancer), Calu-6 (lung, non-small-cell adenocarcinoma (NSCLC)), HCC70 (triple-negative breast cancer) and MM1.S (multiple myeloma), as well as the following human pri- mary PDX models: PHTX-24c (colon cancer), PHTX-132Lu (lung, NSCLC), PHTX-55B (triple-negative breast cancer), PHTX-235O (ovarian cancer), and HNM626 (neck cancer). Tumor-bearing mice

Figure 3 TAK-243 induces a pharmacodynamic (PD) response in xenograft tumors. Mice bearing WSU-DLCL2 lymphoma or PHTX-132Lu primary human NSCLC tumors were treated with a single intravenous dose of TAK-243 (25 mg/kg) or vehicle control. (a) Representative IHC images (of n = 3 mice per time point) of the TAK-243–ubiquitin adduct (4 h post-dosing, total cellular ubiquitin conjugates (FK2 antibody; 8 h after dosing), monoubiquitylated H2B (anti-H2BUb; 8 h after dosing), and apoptosis (cleaved caspase-3; 24 h after dosing) in tumor tissue. Scale bar, 100 mm. (b) Quantification of the area that stained positive for presence of the TAK-243–ubiquitin adduct in WSU-DLCL2 (blue line) and PHTX-132Lu (red line) tumor tissue over time. (c) Quantification
of the area that stained positive for polyubiquitin conjugates (red line) and H2BUb (black dashed line) in WSU-DLCL2 (top) and PHTX-132Lu (bottom) tumor tissue over time. For b,c, data points indicate mean 
s.e.m. (n = 3 tumors/time point).

were dosed for 3 weeks with TAK-243,which was administered intra- venously on a twice-per-week schedule (for example, Monday and Thursday; denoted BIW), and tumor growth and animal body weight were monitored. TAK-243 treatment induced a marked and robust antitumor activity response in all of the models examined (Fig. 4, Table 2 and Supplementary Fig. 12a). The human cancer cell lines that were treated with TAK-243 both in vitro and in vivo retained a similar rank order in sensitivity. TAK-243 inhibited both mouse and human UAE, as both TAK-243–ubiquitin adduct formation and downstream PD readouts were detectable in mouse tissues (data not shown); therefore, mice could be used to preliminarily evaluate the therapeutic window for TAK-243. The dose-limiting toxicity observed in mice was a decrease in body weight. At the maximum tolerated

a c




0 5 10 15 20 25
Time (d)




0 5 10 15 20 25
Time (d)







0 5 10 15 20 25
Time (d)







0 5 10 15 20 25
Time (d)

dose in vivo (23–26 mg/kg), mean maximal body weight loss was
<12% of the animals’ body weight relative to that before the start of treatment (Supplementary Fig. 12b).
Two additional in vivo experiments were conducted to demonstrate that TAK-243 antitumor activity is driven through UAE inhibition. First, genetic knockdown of UAE expression, using a doxycycline (Dox)-inducible short hairpin RNA (shRNA) system, demonstrated comparable antitumor activity to that of TAK-243 in the HCT-116 xenograft model (Fig. 4b and Supplementary Fig. 13). In the sec- ond experiment, we used a xenograft mouse model with derivative of THP-1 cells (in which there is an Ala171Thr substitution in NAE that renders it resistant to pevonedistat) and demonstrated that TAK- 243 retained antitumor activity (Supplementary Fig. 14)28. Taken together, these data indicate that TAK-243 is UAE specific and has efficacy in a broad range of tumor models for both solid and hema- tological human cancers.

The diverse roles of ubiquitin in regulating cellular protein home- ostasis and ubiquitin signaling highlight the possibility of targeting the UPS to modulate human disease. Despite this potential, only a small fraction of the >500 enzymes involved in UPS have been tar- geted with agents that have entered human clinical studies. Here we report the identification of a first-in-class small-molecule inhibitor of UAE, TAK-243. TAK-243 is a potent inhibitor of UAE that results in complete inhibition of cellular ubiquitylation, leading to impaired ubiquitin-dependent proteolysis, ER stress, and impaired cell cycle progression and DNA damage repair. TAK-243 exposure caused robust loss of monoubiquitylation and polyubiquitylation of proteins and had substantial antitumor activity in preclinical models of human cancer. Previous small-molecule UAE inhibitors have demonstrated liabilities associated with much weaker potency and off-target activ- ity29,30. The data here validate TAK-243 as a new research tool for the acute modulation of cellular ubiquitylation activity both in vitro and

Figure 4 TAK-243 treatment has antitumor activity in mice bearing subcutaneous xenograft tumors. (a–d) TAK-243 antitumor activity, as assessed by tumor volume over time, was evaluated in mouse xenograft models bearing diffuse large B cell lymphoma (WSU-DLCL2) (a), colon cancer (HCT-116) (b), NSCLC (primary PHTX-132Lu model) (c) or multiple myeloma (MM1.S) (d) tumors. Mice with established tumors were dosed intravenously (starting on day 0) using either TAK-243 at the indicated doses or vehicle. Bortezomib (0.8 mg/kg) was used in the MM1. S model. Animals were treated biweekly on days 0, 3, 7, 10, 14 and 17. Data are mean tumor volume values (n = 10 mice per group); error bars represent the s.e.m. A two-tailed Welch’s t-test was used to compare antitumor activity between TAK-243-treated mice and vehicle-treated mice on day 21 (P < 0.001 for all models). All of the mice were included in the analysis.

in vivo, and they support the assessment of TAK-243 in clinical trials in patients with advanced malignancies.
Our cellular and in vivo data indicate that TAK-243 exerts its effects through UAE inhibition. Although TAK-243 inhibits UBA6 with equal potency in cells, UBA6 inhibition had no discernible effects on total ubiq- uitin conjugation or E2 thioester formation, with the expected excep- tion of the E2 enzyme USE1. Moreover, genetic knockout studies for UBA6 expression have failed to demonstrate significant effects on cancer cell viability31,32. In addition, any inhibition of NAE and the NEDD8- dependent cullin ring E3 ligases by TAK-243 would be superseded by the dominant role of UAE in supplying ubiquitin to all ligases. Consistent with this assertion, TAK-243 treatment did not result in the loss-of-func- tion phenotype characterized by DNA re-replication seen after inhibi- tion of NAE activity (as previously demonstrated with pevonedistat5). The findings that TAK-243 was efficacious in a model for NAE that was resistant to pevonedistat and was inactive when UAE expression was genetically knocked down support the conclusion that UAE inhibition is the mechanistic driver of TAK-243 antitumor activity.
Bortezomib, ixazomib and other compounds that target the fun- damental process of protein homeostasis demonstrate that a thera- peutic window can exist for patient benefit. Similarly, our preclinical data with TAK-243 suggest that a potential therapeutic window may exist for UAE inhibition. Mice tolerate TAK-243 with repeated dos- ing up to 25 mg/kg, doses that yield robust antitumor activity with minimal body weight loss or general organ toxicity. Additional comparative toxicological studies in rats and dogs (data not shown)

have demonstrated that TAK-243 is safe to test in a human clinical trial (NCT02045095).
Loss of ubiquitylation activity, like proteasome inhibition, is expected to impair the degradation of short-lived proteins. Beyond its role in pro- teasome-related biology, additional functions of ubiquitin have become increasingly elucidated over the last several decades3,33. Receptor inter- nalization and lysosomal degradation, autophagy, signal transduction, transcription and DNA repair all involve ubiquitin modifications that are independent of K48-linked polyubiquitin chains. As a result, UAE inhibition, while sharing a number of mechanistic similarities with pro- teasome inhibition, also presents profound differences. For example, TAK-243 demonstrates substantial activity in a wide range of preclinical models of solid tumors, raising the hope that TAK-243 may demonstrate clinical benefit in solid tumor malignancies, whereas bortezomib and other proteasome inhibitors have yet to demonstrate consistent antitu- mor activity, both preclinically and clinically. For this reason, our pre- clinical investigation of TAK-243 focused mostly on solid tumors. It has been speculated that more profound disruption of protein homeostasis, above and beyond proteasome disruption, may be required for clinical activity in solid tumors; both TAK-243 and the recently disclosed VCP inhibitor CB-5083 (ref. 8), both upstream inhibitors of the ubiquitin proteasome system, represent new opportunities for the evaluation of this hypothesis. How differences between modulation of ubiquitin biol- ogy and proteasome inhibition might be exploited clinically is a matter of ongoing investigation. Notably, although we found that bortezomib and TAK-243 have similar EC50 values for inhibition of the cell viabil- ity of multiple-myeloma MM1.S cells in vitro, these compounds have marked differences in their antitumor efficacy in vivo when tested with MM1.S xenografts. Although these differences might be due to differ- ing PK properties of the compounds, they may also reflect the differing biological roles of their targets in the UPS pathway. The discovery of TAK-243 provides a valuable tool for the study of protein ubiquityla- tion and provides a new opportunity to study the modulation of protein homeostasis and ubiquitin signaling for the treatment of cancer.
Methods, including statements of data availability and any associated accession codes and references, are available in the online version of the paper.
Note: Any Supplementary Information and Source Data files are available in the online version of the paper.

The authors would like to thank W. Harper (Harvard Medical School) for the UBA6-knockout and control MEFs, A. Berger for critical review of the manuscript, and E. Koenig and P. Shah for genomic data analysis. We would also like to thank J. Afroze, I. Bharathan, J. Gaulin, M. Girad, C. McIntyre, F. Soucy, T.T. Wong and Y. Ye for performing the chemical synthesis of TAK-243. All activities were completed and funded through Takeda Pharmaceuticals Inc.

M.L.H., M.A.M., N.F.B., B.S.A., J.G. and L.R.D. participated in writing, reviewing and editing of the manuscript; M.L.H., M.A.M. and N.F.B. participated in the planning, initiation, data generation and analysis of biological experiments; J.C.,
S.L. and P.F. participated in the planning, initiation, design and execution of chemical synthesis; M.S. and N.B. performed crystallography experiments; J.P. was the toxicology representative on the program; J.G., T.S., F.B. and J.B. performed biochemical analyses; M.A.M., D.S., J.D., C.R., J.R. and K.H. performed in vitro cell culture experiments; M.L.H., T.T., J.H., J.S. and S.M.P. performed in vivo antitumor activity and pharmacodynamic experiments; A.L. and S.M. evaluated compound potencies in cell-based assays; Y.Y. and B.S. performed immunohistochemistry experiments; R.G. performed pharmacokinetic analyses; B.S.A., S.T., M.K., P.V., J.N., P.L., J.T.W., P.G., K.G., M.M. and C.C. provided project oversight and review.

The authors declare competing financial interests: details are available in the online version of the paper.

Reprints and permissions information is available online at reprints/index.html. Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

1. Jin, J., Li, X., Gygi, S.P. & Harper, J.W. Dual E1 activation systems for ubiquitin differentially regulate E2 enzyme charging. Nature 447, 1135–1138 (2007).
2. Swatek, K.N. & Komander, D. Ubiquitin modifications. Cell Res. 26, 399–422 (2016).
3. Komander, D. & Rape, M. The ubiquitin code. Annu. Rev. Biochem. 81, 203–229 (2012).
4. Richardson, P.G. et al. A phase 2 study of bortezomib in relapsed, refractory myeloma. N. Engl. J. Med. 348, 2609–2617 (2003).
5. Brownell, J.E. et al. Substrate-assisted inhibition of ubiquitin-like protein-activating enzymes: the NEDD8 E1 inhibitor MLN4924 forms a NEDD8–AMP mimetic in situ. Mol. Cell 37, 102–111 (2010).
6. Huang, X. & Dixit, V.M. Drugging the undruggables: exploring the ubiquitin system for drug development. Cell Res. 26, 484–498 (2016).
7. Bedford, L., Lowe, J., Dick, L.R., Mayer, R.J. & Brownell, J.E. Ubiquitin-like protein conjugation and the ubiquitin–proteasome system as drug targets. Nat. Rev. Drug Discov. 10, 29–46 (2011).
8. Anderson, D.J. et al. Targeting the AAA ATPase p97 as an approach to treat cancer through disruption of protein homeostasis. Cancer Cell 28, 653–665 (2015).
9. Krönke, J. et al. Lenalidomide induces ubiquitination and degradation of CK1 in del(5q) MDS. Nature 523, 183–188 (2015).
10. McGrath, J.P., Jentsch, S. & Varshavsky, A. UBA1: an essential yeast gene encoding ubiquitin-activating enzyme. EMBO J. 10, 227–236 (1991).
11. Kulkarni, M. & Smith, H.E. E1 ubiquitin-activating enzyme UBA1 plays multiple roles throughout C. elegans development. PLoS Genet. 4, e1000131 (2008).
12. Misra, M. et al. Dissecting the specificity of adenosyl sulfamate inhibitors targeting the ubiquitin-activating enzyme. Structure 25, 1120–1129 (2017).
13. Amidon, B.S. et al. Indole-substituted pyrrolopyrimidinyl inhibitors of Uba6. US Patent 9593121B2. (2017).
14. Soucy, T.A. et al. An inhibitor of NEDD8-activating enzyme as a new approach to
treat cancer. Nature 458, 732–736 (2009).
15. Hershko, A. The ubiquitin system for protein degradation and some of its roles in the control of the cell division cycle. Cell Death Differ. 12, 1191–1197 (2005).
16. Gu, J.J. et al. MLN2238, a proteasome inhibitor, induces caspase-dependent cell death, cell cycle arrest, and potentiates the cytotoxic activity of chemotherapy agents in rituximab-chemotherapy-sensitive or rituximab-chemotherapy-resistant B cell lymphoma preclinical models. Anticancer Drugs 24, 1030–1038 (2013).
17. Milhollen, M.A. et al. Inhibition of NEDD8-activating enzyme induces re-replication and apoptosis in human tumor cells consistent with de-regulating CDT1 turnover. Cancer Res. 71, 3042–3051 (2011).
18. Ulrich, H.D. Ubiquitin and SUMO in DNA repair at a glance. J. Cell Sci. 125, 249–254 (2012).
19. Christianson, J.C. & Ye, Y. Cleaning up in the endoplasmic reticulum: ubiquitin in charge. Nat. Struct. Mol. Biol. 21, 325–335 (2014).
20. Moudry, P. et al. Ubiquitin-activating enzyme UBA1 is required for cellular response to DNA damage. Cell Cycle 11, 1573–1582 (2012).
21. Ungermannova, D. et al. Identification and mechanistic studies of a novel ubiquitin E1 inhibitor. J. Biomol. Screen. 17, 421–434 (2012).
22. Collins, A.R. The comet assay for DNA damage and repair: principles, applications and limitations. Mol. Biotechnol. 26, 249–261 (2004).
23. Vlachostergios, P.J., Patrikidou, A., Daliani, D.D. & Papandreou, C.N. The ubiquitin– proteasome system in cancer, a major player in DNA repair. Part 2: transcriptional regulation. J. Cell. Mol. Med. 13 9B, 3019–3031 (2009).
24. Wildey, G. et al. Pharmacogenomic approach to identify drug sensitivity in small- cell-lung cancer. PLoS One 9, e106784 (2014).
25. Wilmott, J.S. et al. BRAFV600E protein expression and outcome from BRAF inhibitor treatment in BRAFV600E metastatic melanoma. Br. J. Cancer 108, 924–931 (2013).
26. Barretina, J. et al. The Cancer Cell Line Encyclopedia enables predictive modeling of anticancer drug sensitivity. Nature 483, 603–607 (2012).
27. Huang, J. et al. NEDD8 inhibition overcomes CKS1B-induced drug resistance by upregulation of p21 in multiple myeloma. Clin. Cancer Res. 21, 5532–5542 (2015).
28. Milhollen, M.A. et al. Treatment-emergent mutations in NAE confer resistance to the NEDD8-activating enzyme inhibitor MLN4924. Cancer Cell 21, 388–401 (2012).
29. Yang, Y. et al. Inhibitors of ubiquitin-activating enzyme (E1), a new class of potential cancer therapeutics. Cancer Res. 67, 9472–9481 (2007).
30. Xu, G.W. et al. The ubiquitin-activating enzyme E1 as a therapeutic target for the treatment of leukemia and multiple myeloma. Blood 115, 2251–2259 (2010).
31. Gilbert, L.A. et al. Genome-scale CRISPR-mediated control of gene repression and activation. Cell 159, 647–661 (2014).
32. Wang, T., Wei, J.J., Sabatini, D.M. & Lander, E.S. Genetic screens in human cells using the CRISPR–Cas9 system. Science 343, 80–84 (2014).
33. Yau, R. & Rape, M. The increasing complexity of the ubiquitin code. Nat. Cell Biol.
18, 579–586 (2016).

Reagents. Pevonedistat, ixazomib, bortezomib and the UBA6 inhibitor (see compound I-01 in US patent 9593121 B2 (refs. 13,14) were synthesized at Takeda.

Protein reagents. All E1and E2 enzymes used in the E1–E2 transthiolation assays were expressed as N-terminally hexahistidine-tagged full-length human proteins in the Sf9 baculovirus expression system9. N-terminal Flag-tagged UBLs were generated after gene synthesis and subcloning in a pDEST14 vec- tor, expressed in Escherichia coli and then purified as described previously9. Glutathione-S-transferase (GST)-tagged E2 enzymes were used for detec- tion in the homogeneous time-resolved fluorescence (HTRF) assay for the UAE, NAE, SAE and ATG7 reactions. UBC2, UBC12, UBC9 and ATG3 were expressed as amino-terminal GST-tagged fusion proteins in E. coli and puri- fied as described9. USE1 was expressed and purified as described27. UBC8 was expressed as N-terminally hexahistidine-tagged full-length human protein in
E. coli and was purified over nickel–agarose followed by a S200 sizing column. For both USE1 and UBC8, an N-terminal GST tag negatively affected activ- ity, and so both used a biotin modification for detection. USE1 and UBC8 were biotinylated using Sulfo-NHS-LC-Biotin (Pierce, Rockford, catalog no. 21335) following the manufacturer’s standard protocol. The antibody against the TAK-243–ubiquitin adduct was generated by immunizing rabbits with a coupled cysteinyl Gly-Gly-TAK-243 species in a manner similar to that previously published for the generation of an antibody to the pevonedis- tat–NEDD8 adduct5.
Biochemical assays. The E1–UBL-dependent pyrophosphate exchange (PPiX) activity was monitored over time in the presence of different concentrations of TAK-243 to assess the rate of E1 inactivation. Reactions were run in the presence of 1 mM ATP, and both the reaction protocol and curve fit analysis were performed similarly as that described previously34. The rates of recovery of UAE, NAE and SAE after TAK-243–UBL formation were assessed using the E1–E2 HTRF transthiolation assay, using a modified version of the recovery method described previously5. The HTRF E2 transthiolation assay was used to determine the IC50 value of the UBA6 inhibitor against recombinantly puri- fied UAE, NAE and SAE. This assay was run as previously described5,14,35. An AlphaScreen E2 transthiolation assay format using similar conditions was used to measure the IC50 value of the UBA6 inhibitor against recombinant purified UBA6. Biotin-tagged USE1 enzyme was used instead of GST-tagged E2s. For the antibodies used in HTRF-based quantification, 6 g/ml strepta- vidin-coupled donor beads (PerkinElmer Life Science, catalog no. 6760002) and 15 g/ml of anti-Flag-coupled acceptor beads (PerkinElmer Life Science, catalog no. 6760613M) were used for detection. UBA6i IC50 titrations were determined in the presence of concentrations of ATP at the KM for ATP for each respective E1 enzyme.
In vitro E2~UBL thioester assay. The HTRF E2 transthiolation assay was used for TAK-243 IC50 studies against recombinant, purified UAE, NAE, SAE and ATG7 enzymes. This assay was run as previously described5,14,35. An AlphaScreen E2 transthiolation assay format using similar conditions was used for UBA6 and UBA7. Biotinylated tagged E2 enzymes were used instead of GST-tagged E2s. Instead of HTRF detection antibodies, 6 g/ml streptavidin-coupled donor beads (PerkinElmer Life Science, catalog no. 6760002), and 15 g/ml of anti-Flag- coupled acceptor beads (PerkinElmer Life Science, catalog no. 6760613M) were used for detection. TAK-243 IC50 titrations were determined in the presence of concentrations of ATP at the KM for ATP for each respective E1 enzyme.
Crystallization and structure determination. All protein constructs and reagents for crystallization studies were generated in a manner similar to that described previously5. Due to difficulties in crystallizing the human form of UAE, a humanized yeast version of UAE (also known as Uba1 from Saccharomyces cerevisiae, UniProtKB P22515) was used for all studies. Although there are sequence differences between yeast and human UAE, the active site residues are highly conserved. For example, 21 of 25 residues within 5 Å of the TAK-243-binding site are identical. In addition, the use of yeast UAE crystal structures as model systems to describe the enzymatic mechanism of human

UAE is well documented in the literature12. Based on a two-residue amino acid difference in the active site between human and yeast UAE, the substitutions Asn471Met and Lys519Arg (yeast numbering) were introduced into a trun- cated form of yeast UAE (residues 10–1,024). These two substitutions were required to bring the potency of TAK-243 in line with the human enzyme (data not shown). Crystals were formed by mixing humanized yeast UAE, TAK-243, ubiquitin and ATP–MgCl2. Crystals were briefly transferred into cryoprotectant (held on ice), which was comprised of 80% reservoir solution (0.2 M magne- sium formate, 0.1 M Bis-Tris pH 6.5) and 20% polyethylene glycol (PEG)-400, and flash-cooled in liquid nitrogen. Data were collected at the Lilly Research Laboratories Collaborative Access Team (LRLCAT) beamline of the Advanced Photon Source (Argonne, IL, USA) and processed using iMOSFLM36 and SCALA37. The structure was solved by molecular replacement using existing yeast UAE coordinates (Protein Data Bank (PDB) entry 3CMM) as a starting model. Manual rebuilding of the model was accomplished using the program Coot38, and refinement was carried out with the CCP4i39 graphical interface to Refmac40. Coordinates have been deposited in the PDB with structure code 5TR4. Refer to Supplementary Table 1 for crystallographic data collection and refinement statistics.

Kinase panel. Inhibitory activities of TAK-243 (at 1 M) against 319 kinases were measured in studies conducted at the Reaction Biology Corporation (Malvem). Results are expressed as percentage inhibition of kinase activity (percent of con- trol). Briefly, kinase-specific substrates were prepared in the following reaction buffer: 20 mM 4-(2-hydroxyethyl) piperazine-1-ethanesulfonic acid (HEPES) (pH 7.5), 10 mM MgCl2, 1 mM ethylene glycol tetra-acetic acid (EGTA), 0.02% Brij-35, 0.02 mg/ml bovine serum albumin (BSA), 0.1 mM Na3VO4, 2 mM dithiothreitol (DTT) and 1% DMSO. Cofactors unique to each assay were added to the substrate solution. The kinase to be tested was added to the reaction mix- ture, and [-32P]ATP (specific activity 10 Ci/l) was added to the reaction and mixed. The reaction was incubated for 120 min at room temperature, and the reactions were then spotted onto P81 ion-exchange paper (Whatman, catalog no. 3698-915). Filters were washed in 0.75% phosphoric acid, and radioactivity was measured and analyzed. Staurosporine was included as a control compound. Refer to Supplementary Table 2 for a full list of the kinases tested.

Carbonic anhydrase assay. TAK-243 IC50 values were determined by titrating TAK-243 into reactions containing 25 nM human carbonic anhydrase (HCA)- I or 2.5 nM HCA-II with 25 M fluorescein diacetate (assay buffer: 25 mM MOPS, pH 7.5 and 0.02% Triton-X100). The activity of the carbonic anhydrase enzymes were measured over time using a BMG Labtech Polarstar reader at an excitation wavelength of 485 nm and an emission wavelength of 520 nm. The assay is formatted as a gain-of-signal assay based on the release of acetate from fluorescein (relieving the quenching of fluorescein). The IC50 values were calculated by fitting the titration curves to a standard logistic regression model. Refer to Supplementary Table 2 for the HCA-I and HCA-II data.

Novascreen assay. Methods were conducted according the manufacturer’s instructions (Caliper Life Sciences) (Supplementary Table 2).

Cell lines, culturing conditions and transfections. All cell lines were obtained from the ATCC (except for WSU-DLCL2, which was obtained from the DSMZ), tested negative for mycoplasma and were authenticated using SNPchip analy- sis. Cell lines were cultured according to manufacturer’s recommendations. Normal dermal fibroblasts were obtained from Lonza (catalog no. CC-2511) and cultured according to the supplier’s instructions. UBA6-knockout and control MEFs were generous gifts from Dr. Wade Harper, Ph.D., Harvard Medical School. For the UAE overexpression experiments, plasmid pDEST40 (or empty vector) carrying the sequence for full-length human UBE1 (Takeda) was transfected (Lipo3000, Fisher Scientific) into HCT-116 cells for 24 h before TAK-243 treatment.
Western blot analysis. HCT-116 and WSU-DLCL2 cells were maintained in log-phase growth in McCoy’s 5A modified or RPMI-1460 medium, respec- tively (Life Technologies, catalog no. 16600-082 and 11875-093, respec- tively) supplemented with 10% fetal bovine serum (Thermo Scientific)

at 37 °C in a 5% CO2 incubator. Cells were grown in 6-well cell culture dishes and treated with DMSO (0.1%) or with 0.01, 0.10 or 1.00 M TAK- 243 for the times indicated. Whole-cell extracts were prepared using RIPA buffer (Pierce, catalog no. 89901). Immunoblotting (all antibodies were used at 1:1,000 dilution unless otherwise noted) after SDS–PAGE con- ducted under nonreducing (E2-thioester) or reducing conditions was done as previously described3; 30 g total protein was fractionated by SDS– PAGE and immunoblotted with the following primary antibodies to the fol- lowing proteins (a complete list of antibodies and their information is pro- vided in Supplementary Table 3): polyubiquitin (P4D1), p53 (Santa Cruz, 1:2,000 dilution), tubulin (Santa Cruz, 1:4,000 dilution)), ATF6 (Imgenex), Ub-H2B, c-Jun, c-Myc, MCL1, XIAP, Bip (GRP78), PERK, phospho-eIF2a (Ser51), ATF4, CHOP, cleaved PARP, cleaved caspase-3 (Cell Signaling), the TAK-243–ubiquitin adduct (MIL90, Takeda), UBE1 (UAE), phospho-IRE1a (Ser724) (Abcam), XBP1s (BioLegend) and GADD34 (ProteinTech Group). Horseradish peroxidase (HRP)-labeled secondary antibodies to rabbit or mouse IgG (Santa Cruz) were used as appropriate at a dilution of 1:2,000. Immunoblots were developed with ECL reagent (GE Healthcare, catalog no. RPN2106). For Supplementary Figures 2e, 4, 9b, 11 and 13, Alexa-Fluor-680- conjugated secondary antibodies to rabbit or mouse IgG (Invitrogen) were used (1:1,000 dilution), and detection and quantification were performed using the Li-Cor Odyssey Infrared Imaging system. Refer to Supplementary Table 3 for the full antibody list and to Supplementary Fig. 15 for uncropped western blot images.

Cellular thioester assays. HCT-116 cells grown in 6-well cell culture dishes were treated with DMSO (0.1%) or increasing concentrations of TAK-243 for 4 h, and whole-cell extracts were prepared using RIPA buffer (Pierce, catalog no. 89901). Cellular lysates for E2–UBL thioester detection were fractionated by SDS–PAGE under nonreducing conditions and immu- noblotted with primary antibodies to UBCH10 (Boston Biochem), USE1 (Takeda), UBC12 (Takeda), UBC9 (Epitomics) and ATG7 (Epitomics). Mouse- or rabbit-specific Alexa-Fluor-680-conjugated secondary anti- bodies (Life Technologies; for information on all antibodies used, see Supplementary Table 3) were used, and blots were imaged using the Li- Cor Odyssey Infrared Imaging system. To analyze E2 thioesters involved in DNA damage repair, HCT-116 cells were cultured with DMSO or TAK- 243 for 8 h. The following antibodies were used to detect E2–ubiquitin thioesters on immunoblots: UBE2A and UBE2B (Cell Signaling), ubiquitin- conjugating enzyme E2 T (UBE2T; Boston Biochem), proliferating cell nuclear antigen (PCNA; Santa Cruz), Fanconi anemia complementation group D2 (FANCD2; Santa Cruz) and UBE1 (Abcam). All primary and secondary anti- bodies were used at 1:1,000 dilution. Refer to Supplementary Table 3 for a full antibody list and to Supplementary Figure 15 for uncropped western blot images.
Cell cycle analysis. Logarithmically growing HCT-116 cells (plated at 6 ×105) were incubated with DMSO, pevonedistat (0.25 M) or TAK-243 (0.05 M or 1 M) for 16 h. The cells were collected, fixed with 70% ethanol and stored overnight at 4 °C. Fixed cells were centrifuged and washed with PBS to remove the ethanol. The pellets were resuspended in propidium iodide (PI; Invitrogen, catalog no. P3566) and RNAse A (Sigma, catalog no. R-4642) in PBS for 1 h on ice, protected from light. Cell-cycle distribution was determined using flow cytometry (FACS Canto II, Becton Dickinson), and the analysis was completed using BD FACSDiva software version 6.1.1 (Becton Dickinson). The gating strategy was defined by forward (FSC) or side (SSC) scatter of untreated con- trol cells. Gating on live, single PI-positive cells were used for cell cycle analy- sis (Supplementary Fig. 16). 10,000 total events and 5,000 gated events were recorded for each sample analyzed.
Lys48 immunofluorescence assay. HCT-116 cells were plated (1.5 × 104 cells/ well) in 96-well cell culture plates and incubated for 24 h at 37 °C and 5% CO2 for attachment. Increasing concentrations of TAK-243 were added to the plates (with a starting concentration of 20 M), and the plates were incubated for
3.5 h at 37 °C and 5% CO2. Cells were fixed in 2% paraformaldehyde (catalog no. 30525-89-4, VWR) for 10 min followed by 10 min of permeabilization in

0.5% Triton X-100 (Sigma) at room temperature. Cells were blocked for 1 h at room temperature with Roche Blocking Buffer (Roche Diagnostics, catalog no. 11096176001). A primary antibody specific for Lys48 of ubiquitin (clone Apu2; Millipore; 1:500) was incubated with the cells for 1 h at room temperature. An Alexa-Fluor-488-conjugated goat anti-rabbit-IgG (catalog no. A-11034, Invitrogen; 1:500) was used as the secondary antibody and incubated with the cells for 1 h at room temperature. Images were analyzed using an Opera (Perkin Elmer) confocal high-content screening imaging system.
Assay for bulk protein turnover. HCT-116 cells were plated into 12-well plates at 1 × 105 cells/well and incubated overnight. The medium was exchanged with methionine-free Dulbecco’s modified Eagle’s medium (DMEM) (Invitrogen, catalog no. 21013024) containing 10% dialyzed FBS and 50 Ci/well of [35S]methionine, and the cells were incubated for 20 min to label proteins undergoing synthesis. The cells were then washed three times with DMEM sup- plemented with 2 mM methionine. Fresh medium containing 10% FBS, 2 mM methionine and the test compounds, as described in Supplementary Figure 3, were then added. At the specified time points, medium (50 l) was harvested and subjected to liquid scintillation counting. At the end of the time course, the remaining medium was removed, the cells were solubilized by addition of 1 ml of 0.2 N NaOH, and the extract was subjected to liquid scintillation counting. The percentage protein turnover at each time point was calculated as ‘(total acid soluble counts in supernatant)/(total acid soluble counts in supernatant + total counts in solubilized cells) × 100’.
CellLights microscopy for endoplasmic reticulum (ER) expansion. HCT-116 cells were transduced with CellLight ER-RFP, BacMam 2.0 (Invitrogen, C10591) at a multiplicity of infection (MOI) of 20 and treated 24 h later with DMSO or 0.1 M TAK-243 for the times indicated in Supplementary Figure 6. After treatment with the compound, the cells were added to a humidified chamber kept at 37 °C with 5% CO2 and attached to an Eclipse TE2000-U microscope (Nikon Instruments) equipped with an automated xyz stage (Prior Scientific), a filter wheel (Sutter Instruments) and an Orca-ER camera (Hamamatsu) controlled by MetaMorph software.
Comet assay. Calu-6 cells were plated into 6-well dishes at a density of 0.2 × 106 cells/well. Cells were treated with either DMSO or 1 M TAK-243 for 1 h before exposure to UV treatment (20 J/m2). Following UV treatment, cells were har- vested at time t = 0 min or incubated for 6 h in the presence or absence of 1 M TAK-243. Preparation and execution of the alkaline comet assay were performed according to the Trevigen CometAssay protocol (Trevigen, cata- log no. 4250-050-ESK). Electrophoresis was conducted with alkaline buffer (pH 13 solution containing 200 mM NaOH and 1 mM EDTA) at a constant volt- age of 21 V over a 30-min period. Following neutralization in water and fixation with 70% Ethanol, the slides were allowed to dry, and finally they were stained using a 1:5,000 dilution of SYBR green in 1× PBS for 5 min at 4 °C. After a quick rehydration step in distilled water, the slides were imaged on an imageXpress MICRO using a 4× pan-fluorescence objective lens. Comet tails were measured using Comet Assay IV software (Perspective Instruments).

Cell viability assays. Cell lines were maintained under logarithmic growth in the appropriate cell culture medium for each cell line according to the manufactur- er’s instructions. Cells were plated in appropriate complete medium in triplicate in 384-well culture plates in the presence of increasing concentrations of TAK- 243 and incubated at 37 °C for 72 h. Cell viability was assessed using the Cell Titer Glo assay kit (Promega, catalog no. G7571) according to the manufacturer’s protocol. Luminescence was measured using a PHERAstar multilabel counter (BMG Labtech) or LEADseeker system (GE Healthcare Life Sciences).
Generation of in vivo xenograft models of human tumors and efficacy studies. Eight- to 12-week-old mice were inoculated subcutaneously in the flanks with either tumor fragments or a tumor cell suspension in serum-free medium. Female CB-17 SCID mice (Charles River) were used for the follow- ing tumor models: WSU-DLCL2 (4.0 × 106 cells/mouse), MM1.S (5.0 × 106 cells/mouse), (PHTX-132Lu (primary NSCLC, 2 × 3 mm tumor fragments), THP-1 (4.0 × 106 cells/mouse), PHTX-24c (primary colon, 2 × 3 mm tumor

fragments), and THP-1 UBA3 A171T (previously described28). Male CB-17 SCID mice (Charles River) were used for the CWR22 model (2 × 3 mm tumor fragments). Female CB-17 SCID mice (Taconic) were used for the PHTX- 55B model (primary breast, 2 × 3 mm tumor fragments). Immunodeficient, female NOD SCID mice (Jackson Lab) were used for the PHTX-235O model (primary ovarian, 2 × 3 mm tumor fragments). Female nude mice (Taconic) were used for the HCT-116 (described in ref. 41) and Calu-6 models (5.0 × 106 cells/animal). CrownBio (China) performed the experiment using the HNM626 model, in which female nude mice (Shanghai SLAC Laboratory Animal Co. Ltd., Shanghai, China) were implanted in the flank with 3 × 3 × 3 mm tumor fragments. Medicilon (China) performed the experiment using the HCC-70 model, in which female nude mice (Shanghai SINO-British SIPPR/BK Lab Animal Ltd.) were implanted in the flank with 5.0 × 106 cells/mouse.
Tumor growth was monitored with vernier calipers. The mean tumor vol- ume was calculated using the formula: volume (V) = W2 × L/2, where W and L are the width and length of the tumor, respectively. When the mean tumor volume reached approximately 200 mm3, the animals were randomized into groups of n = 6–10 animals (depending on the model). Our randomiza- tion approach used a technique called minimization42 to reduce imbalances in the baseline characteristics of the mice across the groups. In particular, our approach reduced imbalances in the average tumor volume and the ini- tial average animal weight. First, we generated a large number of candidate designs using simple randomization, in which every possible design with a fixed number of animals per group was equally likely. The number of candi- date designs was not fixed; rather, new candidate designs were added until the total computation time reached 3 s. For each candidate design, the overall imbalance was calculated. The overall imbalance was defined as the sum of two terms in which the first term was the variance of the means of the tumor volumes for each group divided by the square of the overall volume mean, and the second term was the variance of the means of the animal weights for each group divided by the square of the overall weight mean. The candidate design with the lowest overall imbalance was selected as the final design. The number of mice per group was used to enable calculation of statistical significance. Mice were dosed with 20% 2-hydroxypropyl--cyclodextrin (HPbCD) or TAK-243 over a 21-d period on a BIW schedule; no blinding was used. Tumor growth and body weights were measured twice per week. The percentage TGI (mean tumor volume (MTV) of the control group – MTV of a treated group)/MTV of the control group) was calculated within 5 d of the last dose, and the statistical analysis was as described previously43. All animal studies were conducted under the approval of the Takeda Oncology Institutional Animal Care and Use Committee and complied with all relevant ethical regulations.

Pharmacodynamic analysis. Three animals bearing tumors were used per time point to enable statistical analysis, if desired. Tumor tissues were har- vested at the indicated times and placed in 10% neutral-buffered formalin. Immunofluoresence analysis was performed on 5-m paraffin-embedded tumor sections using the Discovery XT automated staining system (Ventana Medical Systems, Tucson, AZ). Sections were deparaffinized, followed by epitope unmasking with cell conditioning 1 solution (Ventana Medical Systems) for 20 min. Tumor sections were stained for the TAK-243–ubiqui- tin adduct (MIL90 antibody, Takeda), polyubiquination (FK2 antibody, EMD Millipore), histone 2b monoubiquitination (Cell Signaling, catalog no. 5546), and cleaved caspase-3 (Cell Signaling, catalog no. 9661). The DNA stain 4,6- diamidino-2-phenylindole (DAPI, Vector Laboratories, Inc., Burlingame, CA) was used to estimate the total number of cells per field. One tissue section per mouse was used for each of the three mice in a treatment group. Images were acquired using a Leica DMLB microscope (Leica Microsystems, Wetzlar, Germany) with a Photometrics Cool Snap HQ Nikon Eclipse E800 camera. Five images from each slide were captured using a 20× Leica Plan objective lens (Leica Microsystems, Wetzlar Germany) and analyzed using Metamorph 6.3r7 image processing software (Molecular Devices, Downingtown, PA) with a custom image-processing application module (Molecular Devices Inc.). The number of positive cells were counted in five fields of view, and the aver- age was calculated; DAPI vectashield HardSet Medium (Vector Laboratories, Burlingame, CA) was used as a chromatin counterstain.

TAK-243 and radiation combination efficacy experiments. In experiments were performed at WUXI (China), 6- to 9-week-old female BALB/c nude mice (Shanghai Sino-British SIPPR/BK Laboratory Animal Co., Ltd.) were implanted subcutaneously with ~30-mm3 tumor slices in the right flank, using either LU-01-0030 tumor tissue (a patient-derived xenograft (PDX) model of human NSCLC) or BR-05-0026 tumor tissue (a PDX model of breast can- cer). When the mean tumor volume reached approximately 159 mm3 (LU-01- 0030) or 177 mm3 (BR-05-0026), animals were randomized into four treatment groups. TAK-243 was administered intravenously on a BIW schedule for 3 weeks on the indicated days (Supplementary Fig. 8). Beam-focused radiation was delivered using 2 Gy on the indicated days. A Rad Source RS-2000 X-ray irradiator was used for therapy (Rad Source, serial number: 3161). Before being irradiated, the mice were anesthetized by intraperitoneal injection using 1.0% sodium pentobarbital at 80 mg/kg (Sigma, lot number SLBF7349V). Beam- focused radiation was applied to each mouse at 1 Gy/min. Tumors were meas- ured twice a week using vernier calipers. Tumor volumes were calculated using the following formula: 0.5 × (length × width2). Tumor size and body weight were measured approximately twice a week for the duration of the study. Mice were euthanized when their tumor volume reached approximately 1,000 mm3. Synergistic antitumor analysis was evaluated on days 0 to 21. All tumor volumes had a value of 1 added to them before log10 transformation. For each mouse, the log(tumor volume) at day 0 was subtracted from the log(tumor volume) on the subsequent days. This difference versus time was used to calculate an area-under-the-curve (AUC) value for each animal using the trapezoid rule. The synergy score for the combination of agents A and B was defined as
100 (mean(AUCAB)  mean(AUCA)  mean(AUCB) mean(AUCctl ))/mean(AUCctl )
Where AUCAB, AUCA, AUCB and AUCctl are the AUC values for animals in the combination group, the A group, the B group, and the control group, respectively. The standard error of the synergy score was computed based on the variation in the AUC values among the animals. A two sided t-test was used to determine whether the synergy score was significantly different from zero. If the P value was below 0.05, and the synergy score was less than zero, then the combination was considered to be synergistic.
Pharmacokinetic (PK) analysis. SCID mice bearing WSU-DLCL2 NHL xenograft tumors were administered a single intravenous dose of TAK-243 in 20% hydroxypropyl -cyclodextrin at 12.5, 18.75 and 25 mg/kg, and the tumor and plasma were harvested over a 72-h period. All studies were conducted under the approval of the Takeda Oncology Institutional Animal Care and Use Committee. TAK-243 exposures were determined using mass spectrometry (LC/MS/MS) methods with a lower limit of quantification (LLOQ) of 1 nM for plasma and 5 nM for tumors. Noncompartmental PK analysis was performed on plasma and tumor concentration versus time data using Phoenix WinNonlin version 6.3 software (Pharsight Corp, Mountain View, CA, USA).

Western blot analysis of tumor tissue. Western blot analysis was performed as previously described14. The UAE-specific antibody used was purchased at Sigma- Aldrich (clone no. 2G2.3-5, catalog no. E 3152, 1:1,000 dilution), and the USE1 antibody (MIL21, 1:1,000 dilution) was raised at Takeda. See Supplementary Table 3 for a full list of the antibodies used in this study.

shRNA-mediated knockdown of UAE expression. A pTRIPZ-lentiviral shR- NAmir vector that specifically targeted expression of human UAE (NM_153280, XM_005272650, XM_005272648, NM_003334, XM_005272649) and a control
nontargeting pTRIPZ vector were purchased from Open Biosystems (RHS5087- EG7317, RHS4743, respectively, Thermo Fisher Scientific, Inc.). The UBA1- specific shRNA-encoding sequence was cloned in the pTRIPZ vector, in which shRNA expression is doxycycline inducible. HCT-116 cells were engineered to contain a pTRIPZ vector expressing either the UBA1-specific shRNA or a control nontargeting shRNA. The engineered cell lines were generated by stable transduction with packaged lentiviral particles. Infected cells were selected in the presence of puromycin for 2 d, left to recover for 24 h, and then shRNA expression was induced by adding 1 g/ml doxycycline (Sigma) to the culture

medium. Transduction was confirmed by examining cells for the presence of red fluorescent protein (RFP) after 72 h of induction.
Eight- to 12-week-old female nude mice (Taconic) were used in all in vivo knockdown studies. All mice were housed and handled in accordance with the Guide for the Care and Use of Laboratory Animals44. Mice were inoculated with 1 × 106 HCT-116 cells expressing a nontargeting or UBA1-specific shRNA subcutaneously in the right flank, and tumor growth was monitored with caliper measurements. When the mean tumor volume reached approximately 200 mm3 for efficacy or 600 mm3 for western blot analysis, animals were ran- domly assigned to receive either a vehicle-containing diet (standard diet) or a doxycycline-supplemented diet (LabDiet) for the duration of the study. RFP fluorescence imaging was performed on anesthetized animals using a Xenogen imager to confirm stable transduction (data not shown).
To confirm UBA1-targeting shRNA expression in vivo, tumor-bearing mice were euthanized, and non-necrotic tumor tissues were lysed in M-PER buffer (Pierce, catalog no. 78501). After sonication, samples were centrifuged at 15,000
r.p.m. for 15 min, and equal amount of proteins were resolved on 4–12% Bis-Tris gels (Invitrogen). Antibodies directed against UAE, polyubiquitin (FK2) or USE1 were used to confirm biological activity of the shRNA directed against UBA1. USE1 protein levels were used to demonstrate equal loading.
NAE-resistant HCT-116 efficacy experiments. These experiments were per- formed as previously described28. No animals were excluded from the experi- ment, and no blinding was used.

Statistical analysis. Statistical analyses were performed using Microsoft Excel (Microsoft, Inc., Redmond, WA) or XLFit (IDBS, Inc., Guildford, Surrey, UK). All data are represented as mean ± s.e.m. P values <0.05 were considered sig- nificant throughout the study. The sample size was determined on the basis of our experience with the experimental models, anticipated biological variables and previous general literature. We did not exclude samples or animals from the studies. RNA-seq and histological analyses were performed with blinding to group allocation, whereas other studies did not use blinding.

Data availability. All data generated or analyzed during this study are included in this published article (and its Supplementary Information files). Crystallographic UAE coordinates have been deposited in the PDB with structure code 5TR4.

Life Sciences Reporting Summary. Further information on experimental design is available in the Life Sciences Reporting Summary.

34. Chen, J.J. et al. Mechanistic studies of substrate-assisted inhibition of ubiquitin- activating enzyme by adenosine sulfamate analogs. J. Biol. Chem. 286, 40867– 40877 (2011).
35. Gavin, J.M. et al. Mechanistic studies on activation of ubiquitin and di-ubiquitin-like protein, FAT10, by ubiquitin-like-modifier-activating enzyme 6, Uba6. J. Biol. Chem. 287, 15512–15522 (2012).
36. Leslie, A.G. The integration of macromolecular diffraction data. Acta Crystallogr. D Biol. Crystallogr. 62, 48–57 (2006).
37. Evans, P. Scaling and assessment of data quality. Acta Crystallogr. D Biol. Crystallogr.
62, 72–82 (2006).
38. Emsley, P. & Cowtan, K. Coot: model-building tools for molecular graphics. Acta Crystallogr. D Biol. Crystallogr. 60, 2126–2132 (2004).
39. Potterton, E., Briggs, P., Turkenburg, M. & Dodson, E. A graphical user interface to the CCP4 program suite. Acta Crystallogr. D Biol. Crystallogr. 59, 1131–1137 (2003).
40. Murshudov, G.N., Vagin, A.A. & Dodson, E.J. Refinement of macromolecular structures by the maximum-likelihood method. Acta Crystallogr. D Biol. Crystallogr. 53, 240–255 (1997).
41. Manfredi, M.G. et al. Antitumor activity of MLN8054, an orally active small- molecule inhibitor of Aurora A kinase. Proc. Natl. Acad. Sci. USA 104, 4106–4111 (2007).
42. Pocock, S.J. & Simon, R. Sequential treatment assignment with balancing for prognostic factors in the controlled clinical trial. Biometrics 31, 103–115 (1975).
43. Duffey, M.O. et al. Discovery of a potent and orally bioavailable benzolactam- derived inhibitor of Polo-like kinase 1 (MLN0905). J. Med. Chem. 55, 197–208 (2012).
44. National Research Council. Guide for the Care and Use of Laboratory Animals 8th edn. (The National Academies Press, 2011).

Corresponding Author: Marc Hyer Date: 12/13/2017

Life Sciences Reporting Summary
Nature Research wishes to improve the reproducibility of the work we publish. This form is published with all life science papers and is intended to promote consistency and transparency in reporting. All life sciences submissions use this form; while some list items might not apply to an individual manuscript, all fields must be completed for clarity.
For further information on the points included in this form, see Reporting Life Sciences Research. For further information on Nature Research policies, including our data availability policy, see Authors & Referees and the Editorial Policy Checklist.

 Experimental design
1. Sample size
Describe how sample size was determined.
2. Data exclusions
Describe any data exclusions.
3. Replication
Describe whether the experimental findings were reliably reproduced.
4. Randomization
Describe how samples/organisms/participants were allocated into experimental groups.

5. Blinding
Describe whether the investigators were blinded to group allocation during data collection and/or analysis.
Note: all studies involving animals and/or human research participants must disclose whether blinding and randomization were used.

6. Statistical parameters
For all figures and tables that use statistical methods, confirm that the following items are present in relevant figure legends (or the Methods section if additional space is needed).

n/a Confirmed
The exact sample size (n) for each experimental group/condition, given as a discrete number and unit of measurement (animals, litters, cultures, etc.)
A description of how samples were collected, noting whether measurements were taken from distinct samples or whether the same sample was measured repeatedly.
A statement indicating how many times each experiment was replicated
The statistical test(s) used and whether they are one- or two-sided (note: only common tests should be described solely by name; more complex techniques should be described in the Methods section)
A description of any assumptions or corrections, such as an adjustment for multiple comparisons
The test results (e.g. p values) given as exact values whenever possible and with confidence intervals noted
A summary of the descriptive statistics, including central tendency (e.g. median, mean) and variation (e.g. standard deviation, interquartile range) Clearly defined error bars
See the web collection on statistics for biologists for further resources and guidance.
 Software
Policy information about availability of computer code
7. Software
Describe the software used to analyze the data in this study.

For all studies, we encourage code deposition in a community repository (e.g. GitHub). Authors must make computer code available to editors and reviewers upon request. The Nature Methods guidance for providing algorithms and software for publication may be useful for any submission.

 Materials and reagents
Policy information about availability of materials
8. Materials availability
Indicate whether there are restrictions on availability of unique materials or if these materials are only available for distribution by a for-profit company.
9. Antibodies
Describe the antibodies used and how they were validated for use in the system under study (i.e. assay and species).

10. Eukaryotic cell lines
a. State the source of each eukaryotic cell line used.

b. Describe the method of cell line authentication used.

c. Report whether the cell lines were tested for mycoplasma contamination.

d. If any of the cell lines used in the paper are listed in the database of commonly misidentified cell lines maintained by ICLAC, provide a scientific rationale for their use.
 Animals and human research participants
Policy information about studies involving animals; when reporting animal research, follow the ARRIVE guidelines
11. Description of research animals
Provide details on animals and/or animal-derived materials used in the study.

Policy information about studies involving human research participants
12. Description of human research participants
Describe the covariate-relevant population characteristics of the human research participants.

Corresponding author(s): Michael A. Milhollen

Initial submission Revised version Final submission
Flow Cytometry Reporting Summary
Form fields will expand as needed. Please do not leave fields blank.
 Data presentation
For all flow cytometry data, confirm that:
1. The axis labels state the marker and fluorochrome used (e.g. CD4-FITC).
2. The axis scales are clearly visible. Include numbers along axes only for bottom left plot of group (a ‘group’ is an analysis of identical markers).
3. All plots are contour plots with outliers or pseudocolor plots.
4. A numerical value for number of cells or percentage (with statistics) is provided.
 Methodological details
5. Describe the sample preparation.

6. Identify the instrument used for data collection.

7. Describe the software used to collect and analyze the flow cytometry data.

8. Describe the abundance of the relevant cell populations within post-sort fractions.

9. Describe the gating strategy used.

Tick this box to confirm that a figure exemplifying the gating strategy is provided in the Supplementary Information.

Leave a Reply

Your email address will not be published. Required fields are marked *


You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>