Discovery and Characterization of Inhibitors of Ubiquitin Like Modifier Activating Enzyme 5 (UBA5): Targeting the UFMylation Pathway in Cancer and Neurodegenerative Disease

Abstract

UBA5 is the E1 enzyme that initiates UFMylation, a ubiquitin-like modification implicated in proteostasis, neurodegeneration, and cancer. Here, we report a chemical biology approach to interrogate UBA5 function through small-molecule inhibition. We developed a robust high-throughput screening assay based on AMP-Glo^™ and screened a library of blood–brain barrier–permeable compounds, identifying five inhibitors spanning three distinct scaffolds with low micromolar potency. Orthogonal biochemical and gel-based assays confirmed that these compounds directly inhibit UBA5-mediated UFM1 activation and conjugation, with selectivity over other E1 enzymes, including UBA1. In cells, these compounds suppressed endogenous UFMylation without significantly affecting global polyubiquitination, supporting pathway and target engagement. Together, these studies establish UBA5 as a tractable enzymatic target and provide first-in-class chemical tools to probe UFMylation. Given the emerging role of UFMylation in disease, these inhibitors offer a foundation for developing therapeutics targeting proteostasis pathways.

Graphical Abstract

INTRODUCTION

The ubiquitin-fold modifier 1 (UFM1) conjugation system, known as UFMylation, is an essential post-translational modification pathway that regulates a broad array of cellular processes, including protein folding, stress response, and DNA damage repair.^1–6 UFMylation is catalyzed by three core enzymes: the E1-activating enzyme UBA5, the E2-conjugating enzyme UFC1, and the E3-ligase UFL1, which work sequentially to attach UFM1 to target substrates. Disruption of this pathway has been linked to several diseases, including cancer,^{4, 7–11} neurodegenerative disorders,^12–16 and developmental diseases,^17–19 emphasizing the critical need to understand and potentially target this system therapeutically. Among these enzymes, UBA5 plays a pivotal role in initiating the UFMylation process by activating UFM1 through ATP-dependent adenylation and subsequent transfer to UFC1.

UBA5, as an E1-activating enzyme, occupies a unique position in the UFMylation cascade by catalyzing the first step of UFM1 activation. This function is crucial because it represents the commitment step toward UFMylation, controlling the overall flux of this modification in cells. Due to its central role, any perturbation in UBA5 function can have profound consequences on cellular physiology. For instance, UBA5 alterations have been implicated in a range of human diseases, including Alzheimer’s disease^{12, 13} and various cancers,^{4, 7} where abnormal UFMylation either contributes to the pathogenesis or progression of these conditions. Hence, UBA5 has emerged as a potential therapeutic target in drug discovery efforts aimed at modulating the UFMylation pathway.

In recent years, evidence has accumulated pointing to UBA5’s involvement in neurodegenerative diseases, particularly Alzheimer’s disease.^12–15 A CRISPRi screen in human tauopathy neurons identified the UFMylation cascade as a key regulator of seeding-induced tau propagation²⁰ and oligomeric tau accumulation.²¹ UFMylation is elevated in tangle-bearing neurons in PSP and AD, and its inhibition suppresses tau spread in vitro and in vivo.²⁰ Altered UFMylation dynamics, often due to UBA5 mutations and dysregulation, have been linked to impaired protein homeostasis, which is a hallmark of Alzheimer’s disease.^22–25 For instance, studies have identified mutations in the UBA5 gene in patients suffering from early-onset neurodegeneration,²⁶ further supporting the idea that UFMylation contributes to the development of neurological disorders. Moreover, UBA5’s role in cellular stress response pathways, including endoplasmic reticulum (ER) stress,^{1–3, 27–32} highlights its importance in the maintenance of neuronal function under stress conditions, which is another critical factor in Alzheimer’s pathology.^22–25

UBA5 has also been implicated in the pathogenesis of cancer, with studies showing that abnormal UFMylation can influence oncogenic signaling pathways. For example, dysregulated UFMylation has been linked to the development and progression of various cancers, including breast, colon, gastric, glioblastoma, liver, lung, oral, pancreatic, and renal cancers.^{4, 11, 33–38} UBA5 overexpression has been observed in tumor samples from patients with these cancers, correlating with poor prognosis^{4, 7, 11} and increased metastatic potential.^{11, 33} The oncogenic role of UBA5 is thought to stem from its influence on key cellular processes, such as apoptosis, autophagy, and DNA repair,^{1–3, 27–32} all of which are often altered in cancer cells. Thus, UBA5 represents a promising target for the development of novel anti-cancer therapies.

Despite the significant role of UBA5 in human disease, only a limited number of inhibitors have been identified.^{11, 39–41} One identified inhibitor, adenosine 5’-sulfamate (ADS), is an ATP-competitive analogue with an IC₅₀ of 13 μM.³⁹ However, ADS is a pan-E1 inhibitor, limiting its utility.³⁹ Another inhibitor, named compound 8.5, was developed as a selective UBA5 inhibitor with an IC₅₀ of 4 μM.⁴⁰ Despite its selectivity, compound 8.5 is a Zn²⁺-chelator, which may reduce its therapeutic potential. A third compound, DKM 2-93 (DKM), was discovered as a covalent UBA5 inhibitor, but it has a relatively high IC₅₀ of 430 μM.⁴¹ The natural product Usenamine A has been found to influence UBA5 expression;¹¹ however, it remains unconfirmed whether UBA5 is its direct target. Currently, there is no evidence of a direct interaction between Usenamine A and UBA5, nor of any inhibition of UBA5 enzymatic activity by the compound. To date, no high-throughput screening (HTS) efforts have been reported for UBA5 inhibitors.

In this manuscript, we report the development of an HTS assay to identify UBA5 inhibitors by using the commercially available AMP-Glo^™ kit. Our HTS assay demonstrated high robustness, meeting the NIH HTS Assay Validation guidelines’ acceptance criteria: a Z’-factor ≥ 0.4 or a signal window (SW) ≥ 2, and a coefficient of variation (CV) ≤ 20%.⁴² We also present the results of a pilot high-throughput screening of a Core compound library from the Arizona Center for Drug Discovery (ACDD), comprising 5,120 blood-brain-barrier (BBB)-permeable compounds across sixteen 384-well plates. Several novel inhibitors of UBA5 were identified and characterized by using biochemical, biophysical, and cellular assays.

Results

AMP-Glo^™-based bioluminescence HTS assay

To develop an in vitro UFMylation assay suitable for HTS, we performed codon optimization, synthesized, cloned, expressed and purified UFM1 and UBA5_37-346 encoding UBA5 amino acids (aa) 37-346, representing the functional fragment of UBA5.^43–48

We initially explored a few different assay platforms. In the UFM1 activating reaction catalyzed by UBA5 (UFM1+UBA5+ATP →UFM1-UBA5+AMP+PPi) (Figure 1A), ATP is converted to AMP and pyrophosphate (PPi), and UFM1 is covalently conjugated to UBA5 (UFM1-UBA5) through several intermediate steps.³⁹ Our initial efforts to use commercial kits to quantify ATP depletion or to quantify PPi production were unsuccessful. For reasons unknown, the reagents from the PPi kits from commercial vendors produced high background readings when mixed with UBA5 protein alone, making it impossible to develop a meaningful HTS assay. Additionally, attempts to directly quantify ATP depletion using the Promega Kinase-Glo^™ kit were also unsuccessful due to a low signal-to-background (S/B) ratio of only approximately 1.5-fold.

Figure 1. Development and optimization of AMP-Glo^™-based UBA5 UFMylation assay.

(A) Schematic representation of the assay principle. UBA5 catalyzes ATP-dependent activation of UFM1, generating AMP, which is quantified using the AMP-Glo^™ detection system. (B) AMP titration demonstrating the linear detection range of the assay in a 384-well format. (C) AMP production as a function of UBA5 concentration. (D-E) Substrate titration analyses of the AMP-Glo^™ UBA5 assay using varying concentrations of ATP (D) and UFM1 substrate (E), using 384-well plate. For each experiment, only the indicated substrate was varied while all other assay components were maintained at fixed concentrations. RLU, relative fluorescence unit. (F) HTS assay performance parameters for the optimized AMP-Glo^™ UBA5 assay in a 384-well plate, comparing the RFU generated with and without UBA5. N=16. Data are presented as mean ± SD. ****, p<0.0001. Unless otherwise indicated, assay conditions were: UBA5, 0.8 μM; UFM1, 5 μM; ATP, 5 μM.

Finally, we developed a luminescence-based coupling UBA5 assay to quantify AMP production by using the Promega AMP-Glo^™ kit (Figure 1A). The kit’s Reagent I first depletes the input ATP and converts AMP to ADP. Then, Reagent II converts ADP back to ATP, which is detected by the firefly luciferase in the kit, generating luminescence. Thus, the UFM1-activation reaction catalyzed by UBA5 can be monitored by measuring the luminescence.

We first established the linear detection range of AMP using the kit, ensuring that the generated luminescence corresponds directly and linearly to AMP production (Figure 1A, 1B). This step is crucial to guarantee that the inhibitor’s effectiveness will have a linear relationship with the luminescence generated. As shown in Figures 1A and 1B, AMP concentrations within the 0–10 μM range showed a linear relationship with luminescence production. The addition of up to 100 μM ATP did not significantly alter this trend or the signal (data not shown), indicating that Reagent I in the kit effectively depleted the input ATP up to 100 μM without any issues. Next, we demonstrated that AMP production was dependent on UBA5 concentration. Increasing UBA5 concentrations resulted in a corresponding increase in luminescence signal, consistent with UBA5-catalyzed UFM1 activation and AMP generation (Figure 1C). To further optimize the assay, we evaluated the effects of ATP and UFM1 concentrations, buffer conditions, DMSO and detergent tolerance, and incubation time (Figure 1D–1E). These studies were performed to establish assay conditions that remained a robust signal while maximizing assay performance metrics recommended by the NIH HTS Assay Validation Guidelines, including signal window (SW), signal-to-background (S/B) ratio, and Z’ factor.⁴² ATP and UFM1 titration experiments produced concentration-dependent increases in luminescence that approached saturation (Figure 1D, 1E). Fitting the titration data yielded apparent K_m,app values of 11.9 μM for ATP and 0.13 μM for UFM1, which were used to guide selection of operational substrate concentrations for HTS. Under the optimized conditions, comparison of endpoint luminescence signals in the presence and absence of UBA5 yielded excellent assay performance, with a Z’ factor of 0.89, S/B ratio of 7.4, SW of 169, and CV of 0.5% (Figure 1F).

Inhibition of UBA5 activity by positive control inhibitor DKM 2-93

We next evaluated the previously reported UBA5 inhibitor DKM 2-93 (DKM)⁴¹ using the AMP-Glo^™ assay. DKM dose-dependently inhibited UBA5 activity, with an IC_50-UBA5-Glo value of 547 μM (Figure 2A), comparable to the previously reported IC₅₀ of 430 μM.⁴¹

Figure 2. HTS AMP-Glo^™ UBA5 assay.

(A) Dose response curve of inhibition of the UBA5 activity by concentration series of DKM 2-93. N=3. Data were normalized using DMSO-treated wells with UBA5 set as 100%, and wells without UBA5 set as 0%. (B) DKM (3 mM) inhibited UBA5 activity in a 384-well plate. Data are presented as mean ± SD, analyzed by t-test. N=96. ****, P<0.0001. (C) DKM2-93 does not significantly interfere with the AMP-Glo^™ assay system. Reactions containing ATP (5 μM) in the presence or absence of AMP (5 μM) were incubated with either DMSO or DKM2–93 (3 mM) prior to detection using the AMP-Glo^™ Assay Kit according to the manufacturer’s protocol. Data are presented as mean ± SD (N = 3) and were analyzed using an unpaired two-tailed t-test. (D) Well-to-well relative luminescence in 384-well format with and without UBA5; N=96. (E) Z’ factor over plates and dates. Each spot represents Z’ score calculated from the results of 384-well plates on 4 different plates/day on 3 dates.

To determine whether DKM directly interferes with the AMP-Glo^™ detection system, we performed the assay in the absence of UBA5 using reactions containing ATP with or without AMP. As shown in Figure 2B, AMP generated a strong luminescence signal in both the DMSO- and DKM-treated samples, whereas only background signals were detected in reactions lacking AMP. No significant differences in luminescence were observed between the DKM- and DMSO-treated groups under either condition, indicating that DKM does not significantly interfere with the AMP-Glo^™ coupling enzymes or detection chemistry. These results demonstrate that the inhibition observed in the UBA5 AMP-Glo^™ assay is unlikely to arise from direct effects on the assay reagents.

We then used DKM as a control compound to evaluate the robustness of the AMP-Glo^™ UBA5 assay in a 384-well plate format. The optimized assay exhibited excellent HTS performance, with a Z’ factor of 0.85, S/B ratio of 5.8, SW of 37, and CV of 2% (Figure 2C). These performance metrics demonstrated excellent well-to-well consistency (Figure 2D), were reproducible across multiple plates and independent experiments (Figure 2E), and showed no detectable plate drift or edge effects (data not shown).

Pilot HTS

Next, we conducted a pilot HTS against a small library of 5,120 blood-brain-barrier (BBB)-permeable compounds, distributed in sixteen 384-well plates, from the Arizona Center for Drug Discovery Core. DKM served as a positive control inhibitor. The results demonstrated the robustness of the AMP-Glo^™ UBA5 assay, with an average Z’ score of 0.79, S/B ratio of 6.4, SW of 58, CV of 5.3% (Figure 3A). These metrics far exceeded the NIH HTS Assay acceptance criteria⁴² for a reliable HTS assay, which are Z’ ≥ 0.4, SW ≥ 2, and CV ≤ 20%.

Figure 3. Pilot HTS.

(A) Z’ factor, S/B ratio, and CV over plates of a pilot HTS against a CNS compound library in sixteen 384-well plates. (B) Screening flowchart. (C) Compound structures of hits. (D) Representative kinetic readings of dose-response inhibition of UBA5 by Compound 2. (E) Dose response inhibition of UBA5 activity by 5 hit compounds. Data were normalized using DMSO-treated wells with UBA5 set as 100%, and wells without UBA5 set as 0%. N=3.

We identified 13 compounds at 20 μM that showed significant inhibition (>30%), yielding a hit rate of 0.25% (Figure 3B, 3C). After cheminformatic analysis, three hits were removed as pan-assay interference compounds (PAINS),^{49, 50} leaving 10 compounds. Powders for these remaining compounds were re-ordered for further analysis.

Confirmation experiments revealed that 5 out 10 compounds showed dose-dependent inhibition of UBA5 activity with IC_50-UBA5-Glo values in low micromolar range (Figure 3C–3E, Table 1). Notably, all identified hits possess low molecular weight (<400 Da), which is a good size for the hit-to-lead discovery, and have CNS multiparameter optimization (MPO) scores greater than 4.0, which is the desired threshold for compounds suitable for treating CNS diseases.⁵¹ The five hits identified possess three distinct scaffolds. Hits 1 and 4 contain “1,2,4-oxadiazole” ring as the central piece flanked by a fused bicycle – “1-alkyl-1H-benzo[d]imidazol-5-yl-“ – and a phenyl ring in 3 and 5 position, hit 2 possesses a fused tricyclic “tetrahydronaphtho[2,3-b]furan-4-amine” core substituted with a phenyl ring, and 3 and 5 are amide derivatives prepared with a common fused bicyclic amine, 4,5,6,7-tetrahydrobenzo[b]thiophen-2-yl-amine, with 2-furan- and thiofuran carboxylic acids. The structural features of all 5 hits suggest that these may or may not occupy the same site of the enzyme or the coupling enzymes, especially as 1, 3, 4, and 5 are likely to form a curved shape, whereas hit 2 seems to form a rod-like structure. All hits could potentially generate sound hydrophobic contacts and hydrogen bonding(s) through various rings and the hydrogen donating and accepting groups.

Table 1.

Compounds properties.

	MW (Da)	LogD	MPO	TPSA (Å)	IC50-UBA5-Glo (μM)	IC50-UBA5-FP (μM)	IC50-UBA1-Glo (μM)	SIE1	IC50-coup-Glo (μM)	SICoupling	CC50 (μM)
											MCF-7	STTG1	HEK
1	348.36	3	4.8	75.2	11.7	16.6	>100 (42%*)	>8.5	23.7	2.0	104	>200	115
2	264.33	2.1	4.8	52	7.2	11.7	80 (54%*)	11	8.6	1.2	22	168	33.1
3	248.31	3.2	5	83.4	8.3	14.0	95 (57%*)	11	51.5	6.2	46.8	>200	150
4	355.19	4	4.2	56.7	8.0	9.7	>100 (18%*)	>12	35.1	4.4	>100	>200	180
5	264.38	3.3	4.8	98.5	12.2	13.1	59 (73%*)	4.8	95.6	7.8	>100	>200	193

Note: MW, molecular weight; SI, selectivity index; SI_E1 =IC_50-UBA1-Glo/IC_50-UBA5-Glo; SI_Coupling= IC_50-coup-Glo/IC_50-UBA5-Glo;

^*,percent of inhibition at 100 μM.

UBA5-free coupling enzyme assay

To determine if the identified hits inhibit the coupling enzymes in the AMP-Glo^™ kit, we performed a UBA5-free coupling enzyme assay. In this assay, AMP served as the direct substrate for the coupling enzymes in the AMP-Glo^™ kit, along with all other components such as ATP, UFM1, and the reagents from the coupling kit, but without UBA5. The UBA5-free AMP-Glo^™ assay demonstrated sensitivity comparable to the primary AMP-Glo^™ assay in the presence of UBA5 (Figure 4A). When AMP was added to the kit, it directly generated significant luminescence compared to the reaction without AMP, resulting in a Z’ factor of 0.81, an S/B ratio of 17.7, an SW of 37.0, and a CV of 4.1%. These parameters also exceeded the NIH HTS Assay Validation guidelines⁴².

Figure 4. UBA5-free AMP-Glo^™ assay.

(A) AMP-Glo assay in the absence of UBA5. AMP (5 μM) was used as a direct substrate. UFM1, 5 μM. (B) Fitting of dose-response curve of inhibition of the coupling enzymes in the AMP-Glo kit by compounds, using the coupling enzyme tertiary assay. AMP, 5 μM; UFM1, 5 μM. Data were normalized using DMSO-treated wells with AMP set as 100%, and wells without AMP set as 0%. N=3.

We next evaluated the five hit compounds using a UBA5-free AMP-Glo^™ counterscreen to assess potential interference with the coupling enzymes. As shown in Figure 4B and Table 1, all five compounds exhibited varying degrees of inhibition of the AMP-Glo detection system. Compound 2 showed the greatest interference, with an IC_50-coup-Glo value (8.6 μM) comparable to its IC_50-UBA5-Glo value (7.2 μM). Compound 1 also exhibited measurable interference, although its IC_50-coup-Glo (23.7 μM) remained approximately twofold higher than its IC_50-UBA5-Glo (11.7 μM). In contrast, compounds 3–5 displayed substantially weaker inhibition of the coupling enzymes, with IC_50-coup-Glo values 4.3- to 7.8-fold higher than their corresponding IC_50-UBA5-Glo values. These results highlight an inherent limitation of coupled luminescence assays, in which inhibition of the detection enzymes may contribute to the observed activity for some compounds. Therefore, all hit compounds were advanced to orthogonal biochemical assays that do not rely on the AMP-Glo detection chemistry.

Development of a secondary assay based on fluorescence polarization (FP).

To further validate the screening hits, we developed a secondary assay using fluorescence polarization (FP) with the Transcreener AMP²/GMP² FP kit (Bellbrook Labs) (Figure 5A). Unlike the AMP-Glo assay, the FP assay quantifies AMP production through competitive displacement of an Alexa Fluor 633-labeled AMP tracer from an anti-AMP antibody, eliminating the coupling enzymes and luciferase used in the luminescence assay. In this assay, AMP generated by the UBA5-UFM1 reaction competes with the fluorescent tracer for antibody binding, resulting in a dose-dependent decrease in FP signal (Figure 5A).

Figure 5. Secondary FP-based assay.

(A) Principle of FP-based secondary assay. (B) AMP titration using the FP assay based on the Transcreener^® AMP/GMP FP Assay Kit (Bellbrook), following the manufactory protocol. (C) DKM titration against UBA5 using the FP assay. N=3. For all FP assays here and below (unless specified): UBA5, 1.0 μM; UFM1, 2.5 μM; ATP, 20 μM. (D) DKM (4 mM) inhibited UBA5 activity using the FP assay in a 384-well plate. N=16. ***, P<0.001. (E) Fitting of dose-response curve of inhibition of UBA5 by compounds, using the FP-based secondary assay. Data were normalized using DMSO-treated wells with UBA5 set as 100%, and wells without UBA5 set as 0%. N=3.

As expected, AMP dose-dependently reduced FP values, confirming efficient displacement of the fluorescent tracer (Figure 5B). UBA5-mediated UFMylation similarly produced a marked decrease in FP compared with reactions lacking UBA5, demonstrating robust AMP production by the enzyme (Figure 5C). DKM, a previously reported UBA5 inhibitor, dose-dependently inhibited the reaction and was used as a positive control to validate the assay (Figure 5C). The optimized FP assay achieved a Z’-factor of 0.5, an S/B ratio of 3.7, an SW of 5.7, and a CV of 8.2%, satisfying the NIH HTS Assay Validation guidelines⁴² (Figure 5D).

The FP assay was subsequently used as an orthogonal secondary assay to characterize the screening hits. All five compounds dose-dependently inhibited UBA5 activity, yielding IC_50-UBA5-FP values of 16.6 μM, 11.7 μM, 14.0 μM, 9.7 μM, and 13.1 μM for compounds 1–5, respectively (Figure 5E, Table 1). These values were in good agreement with those obtained in the primary AMP-Glo assay, supporting direct inhibition of UBA5 rather than solely interference with the AMP-Glo detection system. The slight discrepancy between IC₅₀ values measured using AMP-Glo^™ and FP assays may be due to assay sensitivity, with the FP assay being slightly less sensitive, requiring a slightly higher UBA5 concentration than the AMP-Glo^™ assay.

As described in the following sections, all five compounds were further confirmed using a gel-based UBA5~UFM1 charging assay, providing an additional orthogonal validation independent of both AMP-Glo and FP detection chemistries.

Selectivity assay

UFM1 is a member of the ubiquitin-like protein (UBL) family, which is covalently attached to target proteins to regulate their activities. This process occurs via a general E1-E2-E3 multienzyme cascade. A total of 17 human UBLs from 9 phylogenetic classes have been identified as being conjugated to various molecules.^{52, 53} While all UBLs share a similar overall structural fold, each UBL typically operates through its own unique E1–E2–E3 enzyme cascade and exerts specific effects on its respective targets. In humans, eight E1 enzymes, UBA1, UBA2/SAE1, UBA3/NAE1, UBA4-7, and ATG7, have been identified as key initiators responsible for the conjugation of specific UBLs.^{52, 53} E1 enzymes are divided into canonical and noncanonical families. Canonical E1 enzymes, such as UBA1, UBA2/SAE1, UBA3/NAE1, UBA6, and UBA7, are responsible for activating Ub, the SUMO protein family, NEDD8, FAT10, and ISG15, respectively. They possess two pseudosymmetric adenylation domains, which are encoded by either one or two genes. In contrast, noncanonical E1 enzymes, such as ATG7, UBA4, and UBA5, which respectively activate the ATG8 and ATG12 protein families, URM1, and UFM1, form homodimers to perform the E1 function. Despite having distinct structures, all E1s share a common catalytic mechanism, which involves the conversion of ATP to AMP during the UBL activation process.^{52, 53} Therefore, identifying inhibitors that are specific to UBA5, without affecting other E1 enzymes, is crucial.

Given the shared ATP-to-AMP conversion mechanism among E1s, we hypothesized that the AMP-Glo^™ assay could be adapted for use with other E1s as a specificity assay for identified hits. To demonstrate this, we cloned, expressed, and purified His-tagged UBA1 in a manner similar to UBA5. The AMP-Glo assay successfully detected UBA1 activity (Figure 6A). In the presence of UBA1, we observed a 6.6-fold increase in luminescence compared to the control without E1, indicating that UBA1, like UBA5 in its UFMylation reaction, converts ATP to AMP during Ub activation, leading to luminescence increase (Figure 6A). We optimized the assay conditions similarly to those for UBA5, observing a dose-dependent increase in luminescence against UBA1 (Figure 6B), ATP (Figure 6C), and other components. Using this specificity assay, we found that these hits only moderately inhibited UBA1 (Table 1, Figure 6D,6E). Specifically, compounds 1 and 4 exhibited IC_50-UBA1-Glo values larger than 100 μM, while compounds 2, 3 and 5 showed moderate IC_50-UBA1-Glo values between 59 to 95 μM. These values are 4.8- to 12-fold higher than those for UBA5 (Table 1, SI_E1), suggesting that the hits selectively target UBA5.

Figure 6. Quaternary E1 specificity assay.

(A) UBA1 AMP-Glo^™ specificity assay with and without UBA1. For all UBA1 AMP-Glo assays here and below (unless specified): UBA1, 0.3 μM; Ub, 10 μM; ATP, 5 μM. ****, p<0.0001. (B) AMP production as a function of UBA1 concentration. (C) Substrate titration analyses of the AMP-Glo^™ UBA1 assay using varying concentrations of ATP, using 384-well plate. (D-E) Titration of identified Compound (1-5) against UBA1. Data were normalized using DMSO-treated wells with UBA1 set as 100%, and wells without UBA1 set as 0%. N=3.

Identified hits inhibited UFMylation in vitro in gel-based assay

To independently validate inhibition of UBA5 activity, we employed a gel-based UFMylation assay that directly monitors UFM1 activation through formation of the UBA5–UFM1 thioester intermediate. Assay specificity was first confirmed using the catalytically inactive UBA5 C250A mutant and the reported UBA5 inhibitor DKM (Figure 7A). Wild-type UBA5 efficiently formed the UBA5–UFM1 intermediate, whereas the C250A mutation completely abolished UFM1 conjugation. Consistent with prior reports, DKM (4 mM) significantly reduced UFM1 activation, validating the assay.

Figure 7. Inhibition of UFMylation by hit compounds in vitro using a gel-based UBA5-charging assay with recombinant proteins.

(A) Validation of the in vitro gel-based UFMylation assay using wild type UBA5, the catalytically inactive C250A mutant, and the control UBA5 inhibitor DKM. Formation of the UFM1 UBA5 thioester complex was detected by SDS PAGE. (B-F) Evaluation of five hit compounds, #1 (B), #2 (C), #3 (D), #4 (E), #5 (F), using the gel based UFMylation assay. Representative Coomassie stained SDS PAGE gels are shown (left) with corresponding quantification of UFMylation levels (right). Reactions contained UBA5 (12 μM), UFM1 (25 μM), and ATP (2 mM) and were incubated for the indicated times in the presence of DMSO or test compounds. Samples were resolved on 12% SDS-PAGE gels and stained with Coomassie Blue. UFMylation activity was quantified as the ratio of the UFM1 UBA5 thioester complex to the combined intensities of free UFM1 and UBA5, measured using ImageJ. Data were normalized to the DMSO treated reaction at 2 hours, which was set to 100 percent, and the 0-hour time point, which was set to 0 percent. N=2.

We next assessed five representative hit compounds identified from biochemical screening (Figure 7B–F). All compounds reduced UBA5–UFM1 thioester formation in a dose-dependent manner, demonstrating direct inhibition of UBA5-mediated UFM1 activation. The relative inhibitory effects were consistent with compound potencies measured in the AMP-Glo- and FP-based assays, indicating strong concordance across orthogonal platforms.

Together, these data provide rigorous biochemical validation that the identified hits directly target UBA5 and suppress UFM1 activation. The gel-based assay therefore serves as a stringent orthogonal confirmation of UBA5 inhibition and supports further mechanistic and cellular evaluation of these compounds.

Hits selectively inhibited UFMylation but not ubiquitination in cells

Subsequently, we selected three representative compounds, 1, 2, and 3, each from a different scaffold class, to evaluate their ability to inhibit the endogenous UFMylation process in HEK293T cells. Typically, HEK293T cells were seeded in 12-well plates, cultured to approximately 80% confluent, and incubated with compounds at concentrations of 10 μM and 60 μM for 24 hours. After incubation, the cell media were removed, and cells were lysed with 100 μL RIPA lysis buffer. Protein concentrations were quantified using the BCA assay, and western blotting was performed using an anti-UFM1 antibody (cat # 109305, Abcam) at 1:2000 dilution, to assess the UFMylation levels of UBA5 and its downstream E2, UFC1, using an anti-UFM1 antibody (Figure 8A). Results indicated that all three compounds inhibited endogenous UFMylation of both UBA5 and UFC1 in a dose-dependent manner. Notably, at 60 μM, compound 3 reduced UFMylation of both UBA5 and UFC1 by nearly 50%.

Figure 8. Effects of compound treatment on UFMylation and ubiquitination in cells.

(A) Inhibition of UFMylation of UBA5 and UFC1 by representative compounds in cell. Inhibition of UFMylation of UBA5 was assessed using a cell-based UFMylation assay (Left panel), with accompanying band quantification (Right panel). HEK293T cells (80% confluent) in 12-well plates incubated with 1-3 (10 μM and 60 μM) for 24 hours and media-free cells lysed with 100 μL RIPA lysis buffer. Cell lysates separated using a 4-12% Bis-Tris Gel (Invitrogen) (left panel) and probed using anti-UFM1 antibody (cat # 109305, Abcam) at 1:2000 dilution. The UBA5-UFM1 and UFC1-UFM1 bands were quantified using ImageJ (Right panel), with band intensities normalized to β-actin. The DMSO-treated sample average was set to 1.0 as the baseline for comparison across treatment groups. N=2 (treatment groups). N=4 (DMSO). ns, no significance; ***, p<0.001; *, p<0.05. (B) Compounds #1–#5 do not affect global cellular ubiquitination. HEK293T cells were treated with compounds #1–#5 (60 μM) or the UBA1 inhibitor TAK-243 (125 or 250 nM) for 5 h. Whole-cell lysates were analyzed by Western blot using anti-ubiquitin and anti-GAPDH antibodies (Left panel). Global cellular ubiquitination was quantified as the integrated density of the ubiquitin-positive high-molecular-weight smear and normalized to the corresponding GAPDH signal. Ubiquitination levels are presented as fold change relative to the DMSO control. Data represent the mean ± SD from three independent experiments (n = 3), except for TAK-243 at 125 nM (n = 1) (Right panel). Statistical significance was determined by one-way ANOVA using the DMSO group as the reference. **, p < 0.01.

To determine whether these compounds selectively target the UFMylation pathway without affecting the ubiquitination pathway, we next examined their effects on global cellular ubiquitination. HEK293T cells were treated with compounds 1–5 (60 μM) for 5 h, and whole-cell lysates were analyzed by anti-ubiquitin Western blotting (Figure 8B). TAK-243, a potent UBA1 inhibitor,⁵⁴ markedly reduced global cellular ubiquitination and served as a positive control. In contrast, compounds 1–5 produced only minor variations in ubiquitination levels, none of which were statistically significant compared with the DMSO-treated control. These results indicate that the compounds do not broadly inhibit the ubiquitin-proteasome pathway and further support their selectivity for the UFMylation pathway.

Cytotoxicity analysis

To evaluate cytotoxicity, ER positive breast cancer MCF-7 cells were treated with increasing concentrations of the five compounds and cell viability was assessed using the WST assay.^55–58 As shown in Figure 9, only compounds 2 and 3 exhibited moderate cytotoxicity toward MCF-7 cells, with CC₅₀ values of 22 μM and 36.8 μM, respectively. In contrast, the remaining three compounds showed minimal cytotoxicity, with CC₅₀ values greater than 100 μM, the highest concentration tested. Assessment of non-malignant cells revealed that compound 2 also moderately reduced HEK293T cell viability (CC₅₀ = 33.1 μM), whereas the other compounds were substantially less toxic, with CC₅₀ values ranging from 115 μM to 193 μM. Collectively, these results indicate that most compounds display low cytotoxicity and a favourable selectivity profile between cancerous and non-cancerous cells.

Figure 9. Cell viability.

Cell viability. ER+ breast cancer MCF-7 cells, astrocyte-like CCF-STTG1 cells, and HEK293T cells were incubated with various concentrations of compounds, and then viability was assayed at 48 h of incubation, using the WST assay. Data were normalized by setting the absorption readings of cells treated with DMSO to 100% and the readings from cell-free wells treated with DMSO to 0%. N = 2.

Activation of ER Stress in mammalian Cells

Previous studies have shown that UFMylation is critical for maintaining endoplasmic reticulum (ER) homeostasis and that disruption of this pathway, including inhibition of UBA5, induces ER stress–associated apoptosis and autophagy, which play significant roles in cancer and neurodegenerative diseases.^59–63 Given that the compounds suppressed UBA5 expression in a dose-dependent manner (Figure 10), we next examined whether ER stress was induced. HEK293T and MCF-7 cells were treated with compounds at 10 μM and 60 μM, and ER stress activation was assessed by monitoring BiP expression. Consistent with prior reports showing maximal BiP induction at 24 h following ER stress activation,¹¹ subsequent analyses were performed at this time point.

Figure 10. UBA5 inhibition differentially modulates BiP and ERα expression in HEK293T and MCF-7 cells.

Inhibition of UFMylation of UBA5 was assessed using a cell-based UFMylation assay (top panel), with accompanying band quantification (bottom panel). HEK293T and MCF-7 cells were treated with UBA5 inhibitors at the indicated concentrations (10 and 60 μM) or vehicle control (DMSO). Whole-cell lysates were prepared after treatment and equal amounts of total protein were subjected to SDS–PAGE followed by immunoblotting was conducted with an anti-BiP and anti-ERα antibody for the ER stress marker BiP (A) and ERα (B). GAPDH for both experiments was used as a loading control. The bands were quantified using ImageJ, with band intensities normalized to GAPDH. N=2 . **p>0.001; ***p>0.0001; ****p<0.0001 and ns, non significant.

As shown in Figure 10A, compounds 1–4 induced a clear dose-dependent increase in BiP expression in MCF-7 cells, with compounds 3 and 4 producing the strongest upregulation relative to the DMSO control. Compound 5 increased BiP at 10 μM; however, at 60 μM, BiP levels were comparable to those of the DMSO control. In contrast, HEK293T cells exhibited a response markedly different from that observed in MCF-7 cells. Compound 1 did not significantly alter BiP expression at either concentration. Compound 2 modestly increased BiP, although the signal at 60 μM was lower than at 10 μM. Conversely, compounds3-5 reduced BiP expression relative to the DMSO, without clear dose-dependence.

We next examined estrogen receptor-α (ERα), a key driver of ERα-positive breast cancer cell proliferation (Figure 10B). In MCF-7 cells, compounds 1, 2, 3, and 5 reduced ERα protein levels at both 10 μM and 60 μM, with greater suppression observed at the higher concentration, indicating a dose-dependent effect. Compound 4 reduced ERα expression to a similar extent at both concentrations, without clear dose dependence. As expected, ERα expression in HEK293T cells was intrinsically low.^{64, 65} Although reductions were detected following compound treatment, the changes were variable and not consistently dose dependent. Compounds 1 and 2 showed greater ERα reduction at 60 μM than at 10 μM, whereas compounds 3 and 4 exhibited slightly less reduction at the higher concentration. Compound 5 reduced ERα expression to a similar extent at both doses.

Overall, compounds 1–5 differentially modulated BiP and ERα expression in a compound-, dose-, and cell-type–dependent manner. MCF-7 cells exhibited pronounced induction of BiP alongside divergent effects on ERα, whereas HEK cells showed a general reduction in ERα expression and a more limited or suppressive response of BiP. These results highlight distinct cellular responses to the compounds across epithelial and cancer-derived cell contexts.

Competitive inhibition of the UBA5

To elucidate the mode of inhibition of compounds, enzyme kinetic analyses were performed on UBA5 in the presence and absence of each inhibitor (Figure 11A). Relative to the DMSO control, all five compounds caused an increase in the apparent K_m,app values while V_max remained largely unchanged. This kinetic pattern, characterized by an elevated K_m,app with no significant alteration in V_max, indicates that compounds inhibit UBA5 through a competitive inhibition mechanism, as defined by classical Michaelis–Menten kinetics.

Figure 11. Mechanism of inhibition and binding affinity of UBA5 with compounds.

(A) Inhibition of UBA5 by a competitive mechanism. Lineweaver–Burk plots of kinetics experimental data for inhibition of the UBA5 by compound (1-5). The UBA5 at 0.8μM and 5μM UFM1 were mixed with DMSO, compound (1-5) (10 μM and 20 μM) for 30 min. Substrate ATP was added at various concentrations. AMP Glo^™ assay kit was used to perform this experiment. N=2. (B) Protein thermal shift assay. Protein TSA using real-time PCR system. N=3.

Protein thermal shift assay

To further investigate the mode of action of the compounds, we examined whether they directly bind to UBA5 using a protein thermal shift assay (TSA), as previously described.⁶⁶ The TSA results revealed differential thermal stabilization of UBA5 by the compounds. While compound 4 and 5 induced a modest increase in the melting temperature (T_m) of UBA5 by 0.6 °C, compound 1 and 2 at 30 μM caused a more pronounced shift of 0.7 and 1 °C, respectively (Figure 11B). Given that a change of ≥0.5 °C in T_m is generally considered indicative of significant binding, these results suggest that compounds interact directly with UBA5, with compound 2 exhibiting stronger binding affinity to UBA5.

Discussion

UBA5 plays a crucial role in the UFMylation pathway, a post-translational modification process. There is growing evidence linking the dysregulation of UFMylation to various human diseases, including neurodegenerative disorders and cancers.^{4, 7–19} Given its central role in cellular homeostasis, targeting UBA5 with specific inhibitors presents a promising therapeutic strategy.

In this study, we developed an HTS assay to identify novel inhibitors of UBA5, the E1-activating enzyme critical to the UFMylation cascade. Our HTS assay, based on the AMP-Glo^™ kit (Promega), demonstrated robustness and reproducibility, allowing the successful screening of a small BBB-permeable compound library. Through this screening, we identified several novel UBA5 inhibitors, which were further characterized by secondary biochemical, biophysical, and cellular assays.

One key finding was the identification of five novel UBA5 inhibitors, belonging to three distinct chemical scaffolds, with low micromolar IC₅₀ values. Importantly, these compounds demonstrated good selectivity for UBA5 over other E1 enzymes, including UBA1, which activates ubiquitin, underscoring their potential for selective targeting of the UFMylation pathway. Selectivity is crucial since many E1 enzymes share common structural and mechanistic features,^{52, 53} making off-target effects a significant concern in the development of E1 inhibitors. Consistent with their biochemical selectivity, the compounds did not significantly alter global cellular ubiquitination despite effectively suppressing UFMylation in cells, whereas the UBA1-selective inhibitor TAK-243 markedly reduced ubiquitination. These findings demonstrate that the compounds selectively inhibit the UFMylation pathway without broadly disrupting the ubiquitin-proteasome system, providing additional evidence that their cellular activity is mediated through UBA5 inhibition rather than nonspecific inhibition of ubiquitin E1 enzymes.

The kinetic behavior of the identified compounds is consistent with classical competitive inhibition of UBA5, suggesting that they interfere with substrate or ATP binding at or near the enzyme’s active site. Such a mode of inhibition is well documented for ubiquitin-activating (E1) enzymes, where small molecules often target the adenylation domain or adjacent substrate-recognition regions.^{52, 67, 68} This interpretation is further supported by ligand-induced stabilization of UBA5 observed in thermal shift assays, indicating direct protein–compound engagement and reduced conformational flexibility upon binding.^{69, 70} Together with prior structural and biochemical studies of E1 enzymes, these findings support a model in which the compounds directly engage the UBA5 active site to competitively block UFM1 activation, with compound 2 emerging as a particularly promising lead for further structural and biophysical optimization.⁵

Moreover, these inhibitors showed promising activity in both in vitro assays and cellular models. Three compounds effectively inhibited endogenous UFMylation in HEK293T cells, reducing the conjugation of UFM1 to UBA5 and UFC1, the downstream E2 enzyme in the pathway. Importantly, this inhibition occurred without detectable changes in global cellular ubiquitination, further confirming the cellular selectivity of these compounds for the UFMylation pathway. These findings indicate that the identified inhibitors can effectively suppress UFMylation while sparing the ubiquitin conjugation machinery, an important characteristic for minimizing potential off-target effects during therapeutic development.

Selective vulnerability of ERα-positive breast cancer cells to UBA5 inhibition suggests a therapeutic window driven by an increased reliance on UFMylation-mediated proteostasis, a pathway critical for maintaining ER homeostasis.⁷¹ Consistent with this model, UBA5 inhibition in Erα-positive MCF-7 breast cancer cells induced robust BiP upregulation, indicative of ER stress activation, whereas non-cancerous HEK293T cells exhibited attenuated or divergent BiP responses. Disruption of UBA5 function is therefore expected to impair UFMylation-dependent ER function, activate unfolded protein response, and shift cells from adaptive proteostasis toward stress-induced signaling programs.^{72, 73} Importantly, in ERα-positive breast cancer MCF-7 cells, compound-induced ER stress correlated with a reduction in ERα protein levels, particularly under conditions associated with strong BiP induction. Sustained ER stress has been shown to suppress ERα expression and estrogen-dependent transcription, thereby attenuating hormone-driven survival signaling. In contrast, the intrinsically low ERα expression in HEK293T cells and the absence of robust ER stress activation likely limit such downstream effects. Together, these findings support a mechanistic model in which pharmacological inhibition of UBA5 disrupts UFMylation, triggers ER stress, and preferentially suppresses ERα-driven signaling in hormone receptor–positive breast cancer cells. This differential stress and signaling response highlights UBA5 as a promising therapeutic target in diseases characterized by UFMylation dependence and ER proteostasis vulnerability.

The ability to target UBA5 in diseases like Alzheimer’s and cancer could open new avenues for therapy. UBA5 has been implicated in protein homeostasis and stress response, both of which are crucial in neurodegenerative diseases.^{1–3, 22–25, 27–32} Additionally, aberrant UFMylation has been linked to oncogenesis, with UBA5 overexpression associated with tumor progression and poor prognosis in multiple cancer types.^{4, 11, 33–38} Our work supports the hypothesis that inhibiting UBA5 could modulate these pathways, providing therapeutic benefits.

Despite the promising results, there are still challenges and limitations to address. While our hits demonstrated reasonable selectivity for UBA5, further optimization is necessary to improve their potency and pharmacokinetic properties. Structural studies, including co-crystallization with UBA5, could provide insight into the binding modes of these inhibitors and guide rational drug design for enhanced efficacy. Additionally, detailed in vivo studies are needed to evaluate the therapeutic potential of these inhibitors in disease models, particularly neurodegenerative and cancerous conditions where UFMylation dysregulation plays a key role.

Conclusion

This study demonstrates the successful development of an HTS assay for identifying UBA5 inhibitors and the discovery of several promising candidates with selective inhibitory activity. The identified compounds inhibit endogenous UFMylation in cells while exhibiting minimal effects on global cellular ubiquitination, supporting their selectivity for the UFMylation pathway over the ubiquitin system. These findings provide valuable tools for studying UFMylation and lay the groundwork for developing potential therapies targeting UBA5 in diseases such as Alzheimer’s and cancer. Ongoing research into the identification and characterization of UBA5 inhibitors will offer crucial insights into their potential clinical applications. Further structural and pharmacological optimization of these inhibitors could lead to the development of clinically relevant compounds, opening new therapeutic avenues for modulating UFMylation in human disease.

MATERIALS AND METHODS

Materials

Ubiquitin was purchased from South Bay Bio (cat# SBB-UP0013). Anti-UFM1 antibody (Cat. #109305) was purchased from Abcam. Anti-BiP antibody (Cat. #3183S) and anti-ERα antibody (Cat. #8644S) were purchased from Cell Signaling. The anti-ubiquitin antibody (Cat. #sc-8017) was purchased from Santa Cruz Biotechnology, and the anti-GAPDH antibody (Cat. #60004-1) was obtained from Proteintech. 4-12% Bis-Tris Gel was obtained from Invitrogen.

Expression and purification of UFM1, UBA1 and UBA5_37-346

The codon-optimized gene sequence of UFM1, including a TEV protease-recognition site at the N-terminus, was synthesized and inserted between the His-tag and EcoR1 site in the pET28a vector using seamless cloning technology by GeneUniversal. The expression plasmid UFM1-p28 was transformed into E. coli Rosetta (DE3) cells, which were then cultured in Luria broth medium supplemented with 100 μg/mL kanamycin at 37 °C until the optical density (OD) at 600 nm reached 0.6. The cells were induced with 0.3 mM IPTG and further incubated with shaking at 150 rpm and 16 °C for 18 hours. After incubation, the cells were collected by centrifugation at 8000 rpm for 20 minutes. The cell pellets were resuspended in lysis buffer (20 mM HEPES, pH 7.5, 200 mM NaCl), and lysozyme along with phenyl methyl sulfonyl fluoride (PMSF) was added. Cells were lysed using sonication and then centrifuged at 15,000 rpm at 4 °C for 30 minutes. The supernatant was loaded onto a pre-charged Ni-NTA affinity column (Qiagen) and washed with the resuspension buffer containing 30 mM imidazole. His-tagged UFM1 was eluted using lysis buffer containing 50 mM and 100 mM imidazole. The UFM1 was further purified by size-exclusion chromatography using a 75 Superdex column, and peak fractions were collected and pooled. The purified UFM1 was stored in a buffer containing 25 mM HEPES, pH 7.5, 150 mM NaCl, and 2 mM DTT.

The codon-optimized gene sequence of UBA5 encoding UBA5 amino acids 37-346 was synthesized and cloned into a custom His-SUMO vector using seamless cloning technology by GeneUniversal. The expression and purification of UBA53_7–346-His-SUMO were performed similarly to UFM1, with the exception of the on-column digestion step. The supernatant was loaded onto a pre-charged Ni-NTA affinity column (Qiagen). On-column digestion was carried out using the ULP1 protease, with incubation for 2 hours at room temperature (RT). Following digestion, flowthrough fractions were collected, washed with 2 columns of the resuspension buffer, and combined. The UBA5_37–346 was then concentrated and further purified by size-exclusion chromatography using a 75 Superdex column. Peak fractions were collected and pooled, and the purified UBA5 was stored in a buffer containing 25 mM HEPES, pH 7.5, 150 mM NaCl, and 2 mM DTT.

The codon-optimized gene sequence of UBA1 was synthesized and cloned into the pET28a vector by GeneUniversal. Expression and purification of His-tagged UBA1 was carried out as described previously.⁷⁴

Steady State Kinetics Assay

Steady-state enzyme kinetic assays were performed at 30 °C in a reaction buffer containing 50 mM Bis-Tris (pH 6.5), 100 mM NaCl, and 10 mM MgCl₂. For the assay, varying concentrations of UBA5_37–346 were mixed with UFM1 (5 μM) in the presence of ATP (5 μM). Additionally, UFM1 at different concentrations was combined with UBA5 (800 nM) in the presence of ATP (5 μM). Lastly, varying concentrations of ATP were tested alongside UBA5 (800 nM) and UFM1 (5 μM). The reactions were incubated with Reagent I from the AMP Glo^™ kit (Promega) for 1 hour at room temperature (RT). Afterward, AMP Glo^™ Reagent II was added to each well, and luminescence (Gain 150) was measured using a BioTek H1 Reader. Reaction velocities for each concentration of UBA5, UFM1, and ATP were calculated and fitted to the Michaelis-Menten equation using the GraphPad Prism 9 software. The experiment was conducted in triplicate, unless otherwise specified.

AMP Detection Assay using the AMP-Glo^™ kit

The AMP Glo^™ Assay Kit from Promega was utilized to detect AMP in samples following the kit’s instructions. UBA5 (800 nM) was incubated for 1 hour with various concentrations of compounds (0-100 μM). Following this, UFM1 and ATP (5 μM each) were added and incubated for an additional 30 minutes. Then, 5 μL of AMP Glo^™ Reagent I was added to each well and incubated at RT for 1 hour, allowing for ATP depletion and conversion of AMP to ADP. The reaction was completed by adding 10 μL of AMP Glo^™ Reagent II and incubating for 30 minutes at RT to convert ADP back to ATP, enabling the enzyme reaction. Luminescence was measured in each well using a BioTek H1 Reader, and the data were analyzed relative to a standard curve to determine AMP concentration. The IC⁵⁰ values of the compounds were calculated using the GraphPad Prism 9. The experiment was performed in triplicate.

Fluorescence Polarization (FP) Assay

To measure AMP levels in samples, the Transcreener^® AMP/GMP FP Assay Kit (Bellbrook) was employed. Assay components were prepared according to the manufacturer’s instructions, with a total reaction volume of 20 μL in a 384-well black plate, consisting of 15 μL enzymatic reaction and 5 μL of kit reagent. UBA5 (1 μM) was incubated for 1 hour with varying concentrations of compounds (0-25 μM). Following this, UFM1 (2.5 μM) and ATP (20 μM) were added, and the plate was incubated for 30 minutes at RT to allow the enzymatic reaction to occur. After incubation, 5 μL of the Transcreener^® AMP/GMP FP Detection Reagent was added to each well, containing the necessary fluorescent probe. The plate was incubated for 1.5 hours as specified in the protocol to develop the fluorescent signal. Fluorescence polarization was detected using a fluorescence plate reader, with the degree of polarization correlating to the amount of AMP present in the samples. The IC₅₀ values of the compounds were calculated using GraphPad Prism 9, and the experiment was performed in triplicate.

UBA5-free coupling Activity

Test compounds were evaluated at concentrations ranging from 0 to 100 μM. The AMP Glo^™ kit (Promega) was utilized for this assay with AMP as the substrate in the absence of UBA5. Following a 30-minute incubation, AMP was added to achieve a final concentration of 5 μM. Plates were read at 37 °C with a luminescence endpoint, an integration time of 2 minutes, full light emission, top optics, and gain set to 150. Appropriate controls were included. The experiment was performed in triplicate.

UBA1 specificity assay

The UBA1 specificity assay was conducted in triplicate, following a protocol similar to the AMP Detection Assay using the AMP-Glo^™ kit, which was employed for UBA5. For all UBA1 AMP-Glo assays (unless otherwise specified): UBA1, 0.3 μM; Ubiquitin, 10 μM; ATP, 5 μM.

Charging Assay

The charging assay was conducted in duplicate as reported previously⁴³. UBA5 (12 μM) was incubated with various concentrations of compounds (0-100 μM) for 1 hour in a reaction buffer comprising 50 mM Bis-Tris (pH 6.5), 100 mM NaCl, and 10 mM MgCl₂. UFM1 (25 μM) and ATP (2 mM) were added to the reaction samples, which were incubated for 2 hours at 30 °C. Zero time points were taken before ATP addition. Catalytically inactive C250A mutant (12 μM), and the UBA5 inhibitor DKM (4mM) were used as controls. Samples were analyzed via non-reducing 12% SDS-PAGE, followed by staining with Coomassie G-250. Product bands corresponding to each protein or complex were quantified using the ImageJ software, and percentage activity was calculated with GraphPad Prism 9.

Enzymatic assay

Enzyme kinetic analyses and Lineweaver–Burk plots for compounds (1-5) were performed using 0.8μM UBA5 and 5μM UFM1 in a reaction buffer containing 50 mM Bis-Tris (pH 6.5), 100 mM NaCl, and 10 mM MgCl₂ in a total volume of 40μL in 384-well plates. Reactions were initiated by adding varying concentrations of the ATP as a substrate in the presence or absence of inhibitors. Initial velocities of the enzymatic reaction with different inhibitors and DMSO were determined by linear regression of the kinetic traces and subsequently analysed in GraphPad Prism 9 using Michaelis–Menten and Lineweaver–Burk transformations.

Protein thermal shift assay (TSA)

Thermal shift assays (TSA) were performed on an Applied Biosystems 7500 Fast Real-Time PCR System (Thermo Fisher Scientific) across a temperature range of 4–90 °C, following previously established protocols.⁵⁶ Purified UBA5 was prepared at a final concentration of 5 μM in 1× PBS and incubated with compounds to achieve a final compound concentration of 30 μM. Incubations were carried out for 30 min at 25 °C. Reactions were set up in MicroAmp Fast Optical 96-Well Reaction Plates (Thermo Fisher Scientific). Protein unfolding was monitored using SYPRO Orange dye (5×; Life Technologies) as a fluorescent reporter, in accordance with the manufacturer’s instructions. Melting temperatures (Tm) were determined using the derivative method implemented in Protein Thermal Shift Software v1.0 (Thermo Fisher Scientific), and the change in melting temperature (ΔTm) was defined as the difference between Tm (DMSO control) and Tm (compound-treated sample).

Inhibition of cellular UFMylation

Briefly, HEK293T cells seeded in 12 well plates (~80% confluent) were incubated with compounds at 10 and 60 μM concentration for 24 hours, cell media were discarded, and cells were lysed using 100 μl RIPA lysis buffer. Protein concentrations were determined using BCA, western blotting experiment was performed using an anti-UFM1 antibody (Abcam, 109305; 1:2000 dilution) using 4-12% Bis-Tris Gel (Invitrogen). Quantification of the UBA5-UFM1 and UFC1-UFM1 bands was performed using ImageJ. Experiments were performed in duplicate.

Analysis of Global Cellular Ubiquitination

HEK293T cells were seeded in 12-well plates and cultured to approximately 80% confluence. Cells were treated with compounds #1–#5 at a final concentration of 60 μM for 5 h. TAK-243 (MedChemExpress, HY-100487), a selective UBA1 inhibitor known to suppress cellular ubiquitination,⁵⁴ was included as a positive control. Vehicle-treated cells served as a negative control.

Following treatment, cells were washed with cold phosphate-buffered saline (PBS) and lysed in RIPA buffer supplemented with protease inhibitors. Cell lysates were clarified by centrifugation, and total protein concentrations were determined using BCA assay.

Equal amounts of total protein (20 μg per sample) were separated by SDS-PAGE and transferred onto PVDF membranes. Membranes were blocked with 5% nonfat milk in TBST and incubated overnight at 4°C with an anti-ubiquitin primary antibody at 1:500 (Santa Cruz Biotechnology, sc-8017) After washing, membranes were incubated with an HRP-conjugated secondary antibody and visualized using enhanced chemiluminescence (ECL). Following image acquisition, membranes were stripped and incubate with an anti-GAPDH antibody at 1:2000 (Proteintech, 60004-1) as a loading control.

Global cellular ubiquitination was quantified using Bio-Rad Image Lab software. The integrated density of the ubiquitin-positive high-molecular-weight smear within each lane was measured following background subtraction and normalized to the corresponding GAPDH signal. Relative ubiquitination levels were calculated as the ratio of ubiquitin signal to GAPDH signal and normalized to the vehicle control. Data are presented as mean ± SD from three independent experiments.

Cytotoxicity

Cytotoxicity and cell viability were measured using the cell counting kit-8 (CCK-8) (GLPBIO) according to the manufacturer’s protocol, with minor modifications. Human astrocyte cells (CCF-STTG1), Human embryonic kidney epithelial cells (HEK293T) and Human breast cancer cells (MCF-7) were cultured in DMEM and RPMI-1640 media supplemented with 10% FBS at 37 °C in a humidified atmosphere of 5% CO₂. A total of 1.5 × 10⁵ CCF-STTG1, HEK293T and MCF-7 cells were seeded in duplicate in a 96-well plate and incubated for 24 hours in a CO₂ incubator. Cells were treated with UBA5 inhibitors at concentrations ranging from 0 to 200 μM for 48 hours at 37 °C. Following treatment, 10 μL of CCK-8 was added to each well and incubated at 37 °C for 1 to 4 hours. DMSO served as a control. Absorbance was measured at 460 nm using a BioTek Synergy HI microplate reader. The cytotoxic concentration (CC₅₀) was determined via nonlinear regression on dose-response curves using GraphPad Prism 9. All experiments were performed in duplicate.

Activation of ER stress

Briefly, HEK293T and MCF-7 cells seeded in 12 well plates (~80% confluent) were incubated with compounds at 10 and 60 μM concentration for 24 hours, cell media were discarded, and cells were lysed using 100 μl RIPA lysis buffer. Protein concentrations were determined using BCA. Proteins were resolved by SDS-PAGE and subsequently transferred onto a PVDF membrane. The membrane was blocked with 5% non-fat dry milk in 1xTBST for 1.5 h, washed with 1× TBST, and then incubated with the indicated primary antibodies. Western blotting experiment was performed using an anti-BiP antibody (1:2000 dilution) and anti-ERα antibody (1:2000 dilution) using 4-12% Bis-Tris Gel (Invitrogen). Quantification of the bands was performed using ImageJ. Experiments were performed in duplicate. Data were analysed using one-way analysis of variance (ANOVA).