Selective ubiquitination of drug-like small molecules by the ubiquitin ligase HUWE1

Abstract

The ubiquitin system regulates eukaryotic physiology by modifying myriad substrate proteins. Substrate specificity and the assembly of ubiquitin signals are determined by ubiquitin ligases, some of which also modify non-protein biomolecules. Here we expand this substrate realm, revealing that the human ligase HUWE1 can target drug-like small molecules. We demonstrate that compounds previously reported as HUWE1 inhibitors present substrates of their target ligase. Compound ubiquitination is driven by the canonical catalytic cascade, linking ubiquitin to the compound's primary amino group. In vitro, the modification is selectively catalyzed by HUWE1, allowing the compounds to compete with protein substrates. We establish cellular detection methods, confirming HUWE1 promotes - but does not exclusively drive - compound ubiquitination in cells. Converting the existing compounds into specific HUWE1 substrates or inhibitors thus requires enhanced specificity. More broadly, our findings open avenues for harnessing the ubiquitin system to transform exogenous small molecules into novel chemical modalities within cells.

Fig. 1. The HUWE1^HECT inhibitors require a primary amino group that is ubiquitinated.

a Multi-turnover HUWE1^HECT-driven ubiquitination reactions with varying concentrations of BI8622 and BI8626, respectively, followed by a fluorescent Ub tracer. ‘Ub_n’ denotes ubiquitination at one or more sites of the indicated protein substrate. The asterisk marks an unspecifically labeled band. For protein input, see Supplementary Fig. 1a. b Single-turnover assay, monitoring Ub thioester transfer from E2 to HUWE1^HECT in the presence of the indicated concentrations of BI8622 and BI8626 with a fluorescent Ub tracer (top); protein input (bottom). Thioester-linked (‘~’) complexes are characterized by their sensitivity to reducing agent (β-ME). Note that a truncated HECT domain variant (‘HECT^Δ4’) is used to stabilize the HUWE1^HECT~Ub product. c Quantification of assays as shown in (b). For each experiment, the amount of the HUWE1^HECT~Ub product was normalized to inhibitor-free (‘untreated’) conditions. Statistical analyses, using an unpaired, two-sided t-test; ns = not significant (compared to untreated conditions; p > 0.05). Data are represented as mean ± SD (n = 3 biological replicates). d Multi-turnover HUWE1^HECT-driven ubiquitination reactions with 20 μM of BI8626 or the indicated derivatives (Table 1), followed by a fluorescent Ub tracer; for protein input, see Supplementary Fig. 5. e Quantification of assays as shown in (d). For each experiment, the amount of HUWE1^HECT autoubiquitination was normalized to inhibitor-free conditions (‘untreated’). Data are represented as mean ± SD (n = 3 biological replicates). Statistical analyses with an unpaired, two-sided t-test, relative to untreated conditions, ****p = 0.0000023; ***p = 0.00011. Source data are provided as a Source Data file.

Fig. 2. BI8626 ubiquitination requires the catalytic cascade, including HUWE1^HECT.

a SEC-based fractionation of HUWE1^HECT-driven ubiquitination reactions ± ATP, without inhibitor (top). The assignment of the elution peaks is based on SDS-PAGE analyses, as shown in Supplementary Fig. 9a, b. The asterisk marks buffer components. SDS-PAGE analyses of the indicated elution fractions, focusing on the Ub-sized bands (bottom). b SEC-based fractionation of HUWE1^HECT-driven ubiquitination reactions ± ATP in the presence of 20 μM BI8626 (top). SDS-PAGE analyses of the indicated elution fractions, focusing on the Ub-sized bands (bottom). Note that additional, delayed Ub-containing fractions are observed in the presence of ATP, which are not seen in inhibitor-free conditions. For the full SDS-PAGE, see Supplementary Fig. 9d. The lower benzyl moiety of BI8626, including the critical primary amino group is shown. Intact protein MS (c) and deconvoluted neutral MW spectrum (d) of the major BI8626-containing elution peak (orange; V ~ 1.8 mL) from (b). The calculated molecular weight (MW) of BI8626-Ub is specified. Intact protein MS (e) and deconvoluted neutral MW spectrum (f) of the minor BI8626-containing elution peak (orange; V ~ 1.4 mL) from (b). g SEC-based fractionation (top) and SDS-PAGE analysis (bottom) of a HUWE1^HECT-driven ubiquitination reaction, containing 20 μM derivative 1, analogous to (b). The lower benzyl moiety of derivative 1, including the critical amino group is shown. h SEC-based fractionation (top) and SDS-PAGE analysis (bottom) of a HUWE1^HECT-driven ubiquitination reaction, containing 20 μM derivative 2, analogous to (b). The lower benzyl moiety of derivative 2 that does not contain a primary amino group is shown. i SEC-based fractionation (top) and SDS-PAGE analysis (bottom) of a HUWE1^HECT-driven ubiquitination reaction, containing 20 μM BI8626, analogous to (b), but employing UbΔGG instead of WT Ub. j SEC-based fractionation of two independent HUWE1^HECT-driven ubiquitination reactions, containing 20 μM BI8626, analogous to (b), but devoid of the E1 or E2. k SEC-based fractionation (top) (bottom) of two independent ubiquitination reactions, containing 20 μM BI8626, analogous to (b), but devoid of HUWE1^HECT or employing a catalytically inactive HUWE1^HECT variant (C4341A) instead of the WT. Source data are provided as a Source Data file.

Fig. 3. BI8626 is an efficient and selective substrate of HUWE1^HECT.

a SEC-based analyses of HUWE1^HECT-driven ubiquitination reactions, containing the indicated BI8626 concentrations, monitoring the mono- and diubiquitinated BI8626 by absorbance at 340 nm. b Multi-turnover HUWE1^HECT-driven ubiquitination reactions ±20 μM purified mono-ubiquitinated BI8626 (BI8626-Ub), followed by a fluorescent Ub tracer. The bottom part of the gel is shown at a different exposure to allow for diUb and diUb-modified BI8626 to be distinguished. For protein input, see Supplementary Fig. 12b. c Quantification of assays as shown in (b). For each experiment, the amount of HUWE1^HECT autoubiquitination at 30 minutes was normalized to that under inhibitor-free (‘untreated’) conditions. Data are represented as mean ± SD (n = 3 biological replicates). Statistical analysis using un unpaired, two-sided t-test (ns (not significant): p $\ge$ 0.05). d SEC-based analyses of BI8626 ubiquitination, driven by the indicated HECT domains, monitoring the mono- and diubiquitinated BI8626 by absorbance at 340 nm. e Quantification of BI8626 monoubiquitination by the indicated HECT domains from SEC analyses as shown in (d). The signal produced by HUWE1^HECT was normalized to 1. Data are represented as mean ± SD (n = 2 biological replicates). Source data are provided as a Source Data file.

Fig. 4. HUWE1^FL promotes compound ubiquitination in a manner that competes with protein substrate modification.

a Multi-turnover HUWE1^FL-driven ubiquitination reactions in the presence of the indicated BI8626 concentrations, without protein substrate, followed by a fluorescent Ub tracer; for protein input, see Supplementary Fig. 13a. b Multi-turnover HUWE1^FL-driven MCL1 ubiquitination reactions in the presence of the indicated BI8626 concentrations, followed by a fluorescent Ub tracer; for protein input, see Supplementary Fig. 13f. c Multi-turnover HUWE1^FL-driven ubiquitination reactions in the presence of the indicated BI8626 derivatives (20 μM), without protein substrate, followed by a fluorescent Ub tracer; for protein input, see Supplementary Fig. 14e. d Multi-turnover HUWE1^FL-driven MCL1 ubiquitination reactions in the presence of the indicated BI8626 derivatives, followed by a fluorescent Ub tracer; for protein input, see Supplementary Fig. 14f. e Multi-turnover HUWE1^FL ubiquitination reactions in the presence of 20 μM BI8626 or purified BI8626-Ub, without protein substrate, followed by a fluorescent Ub tracer; for protein input, see Supplementary Fig. 18a. f Multi-turnover HUWE1^FL-driven MCL1 ubiquitination reactions in the presence of 20 μM BI8626 or purified BI8626-Ub, followed by a fluorescent Ub tracer; for protein input, see Supplementary Fig. 18b. g Cartoon of our click chemistry-based strategy for detecting inhibitor (derivative 6; Table 1) ubiquitination in vitro. An azide-linked fluorophore is symbolized by a star. h Proof-of-principle experiment for the detection of HUWE1^FL-driven derivative 6 ubiquitination, according to (g). Note that at high exposure (top), a small amount of Ub chain formation on the compound is observed. i Multi-turnover HUWE1^FL-driven ubiquitination reactions in the presence of varying concentrations of derivative 6 and MCL1, followed by fluorescence (for compound ubiquitination) and IB (for MCL1 ubiquitination), respectively. The Ub signal is shown by Coomassie staining. Source data are provided as a Source Data file.

Fig. 5. Ubiquitination of BI8626 derivatives occurs in cells.

a Representative targeted MS² spectrum of a BI8626-modified C-terminal Ub peptide (precursor m/z 278.1506, z = 2), generated by tryptic digestion of cell lysates, following Ub-IP. Annotated fragment ions at m/z 249.6405 (z = 2) and 498.2745 (z = 1) correspond to a glycine–BI8626 conjugate. Additional fragment ions derived from BI8626 itself are observed at m/z 91.0545 (z = 1), 212.6164 (z = 2), 221.1296 (z = 2), 334.1778 (z = 1), 424.2258 (z = 1), and 441.2523 (z = 1), corresponding to characteristic cleavages of the compound. Structural representations of key fragmentation events are indicated, with boxed parts corresponding to the respective fragment ions. b Cartoon of our click chemistry-based strategy for detecting inhibitor ubiquitination in cell lysates. Elements of this figure were created in BioRender. Pohl, P. (2025) https://BioRender.com/sygo5n6. c Proof-of-principle experiment for the click chemistry-based detection of cell-based derivative 6 ubiquitination according to (b). Additionally, the total protein is shown by Coomassie staining (bottom). Compound 7 is clickable, but devoid of a primary amino group, providing a negative control. d Click-chemistry-based detection of cellular derivative 8 ubiquitination in the absence and presence of the UBA1 inhibitor TAK243, analogous to (c). Note that compound 8 is used here instead of 6, for it provided enhanced signal (see Methods). The two sections are part of the same gel, shown with different exposures. e Click-chemistry-based detection of cellular derivative 8 ubiquitination in HEK293T WT and HUWE1 KO cells, analogous to (c). Additionally, the HUWE1 level and an actin-loading control are shown based on IB. f Quantification of the ubiquitinated compound (across all species, including 8-Ub, 8-Ub₂ and 8-Ub_n) from (e), normalized to WT cells. Data are represented as mean ± SD (n = 4 biological replicates). Statistical analysis with an unpaired, two-sided t-test **p = 0.0036. g Click-chemistry-based detection of cellular derivative 8 ubiquitination in HEK293T HUWE1 KO cells, analogous to (c), upon overexpression of the indicated HUWE1 variants. C4341S catalytically inactive. The HUWE1 level and an actin-loading control are shown based on IB. Source data are provided as a Source Data file.