Mode of action of a DCAF16-recruiting targeted glue that can selectively degrade BRD9

Hughes, Scott J.; Stec, Wojciech J.; Davies, Colin T. R.; McGarry, David; Williams, Alicia; Carrara, Marta; Del Barco Barrantes, Ivan; Harris, Rebecca; Tasegian, Anna; Owens, Dominic D. G.; Fawcett, Alexander; Hellicar, John; Azevedo, Aleksandra; Meier, Gregor P.; Runcie, Andrew C.; Greger, Liliana; O’Rourke, Martin; Churcher, Ian; Pass, Martin; Brown, Giles A.; Ciulli, Alessio; Modis, Louise K.; Testa, Andrea

doi:10.1038/s41467-025-63594-w

Download PDF

Article
Open access
Published: 27 October 2025

Mode of action of a DCAF16-recruiting targeted glue that can selectively degrade BRD9

Nature Communications volume 16, Article number: 8516 (2025) Cite this article

6693 Accesses
31 Altmetric
Metrics details

Subjects

Abstract

Molecular glue degraders for therapeutic target proteins are emerging as a strategy in drug discovery. Here, we modify a BRD9 ligand with specific chemical fragments to create degrader compounds that we call Targeted Glues. When bound to the target protein, these create an altered protein-ligand interface that is recognised by a ligase. This interaction between the target and the E3 ligase leads to protein degradation and is stabilised by a reversible covalent interaction between our molecule and a specific cysteine in the ligase. By screening a library of BRD9 targeted compounds we discover AMPTX1, a potent selective and reversibly covalent BRD9 degrader. In cells, AMPTX-1 selectively recruits the E3 ligase, DCAF16, to BRD9 and drives BRD9 degradation, as demonstrated by co-immunoprecipitation-mass spectrometry. BRD9 degradation is primarily dependent on the engagement of the surface Cys58 of DCAF16; the formation of a covalent adduct to DCAF16 is facilitated by ternary complex formation with BRD9. BRD9 degradation is also achieved in vivo with AMPTX-1 in a mouse xenograft model after oral dosing due to the drug-like, orally bioavailable properties of the compound. This supports the concept that covalent recruitment of DCAF16 is a viable approach in the development of therapeutic degraders.

Template-assisted covalent modification underlies activity of covalent molecular glues

Article Open access 29 July 2024

Targeted protein degradation via intramolecular bivalent glues

Article Open access 21 February 2024

DCAF1-based PROTACs with activity against clinically validated targets overcoming intrinsic- and acquired-degrader resistance

Article Open access 04 January 2024

Introduction

Targeted protein degradation (TPD) with small molecule degraders, designed to remove a specific protein from the cell, is an emerging therapeutic modality in clinical trials for multiple indications, including solid and haematological malignancies and autoimmune diseases¹. Most small molecule degraders work by recruiting a specific protein of interest (POI) to an E3 ligase, by means of either heterobifunctional compounds (PROTACs) or monovalent compounds (molecular glues). Generally, a PROTAC shows measurable binary binding affinity for both the POI and E3 ligase and can be rationally designed for a specific POI. By contrast, a molecular glue degrader can show measurable binary binding affinity only to one of the two proteins, typically the E3 ligase (such as CRBN), with formation of the POI:glue:E3 complex made possible by ternary complex cooperativity. To date, rational design of molecular glues remains a largely unsolved challenge, and high-throughput screening of compound libraries continues to be the preferred strategy to discover monofunctional compounds able to induce protein degradation by direct recruitment of an E3 ligase^2,3.

While the advancement of clinical stage degraders has been made possible by the development of high-quality E3 ligase binders for VHL⁴ and CRBN^2,3, challenges remain in the discovery of drug-like ligands for different E3 ligases^5,6, limiting the protein degradation toolbox available to medicinal chemists and potentially limiting the range of degradable targets. Recent disclosures of KEAP1⁷, KLHDC2⁸ and DCAF1^9,10 ligands have demonstrated the feasibility of discovering substrate-competitive small molecules for different PROTAC modalities. However, even when high-quality ligands are available, the complex biology of E3 ligases makes it difficult to establish different E3 systems for general use in TPD, as the ubiquitination machinery may be differently regulated in diverse biological systems⁶.

A potential solution to this problem is ligase-agnostic cellular screening of compound libraries, with degradation of the POI as a phenotypic readout. Such compound libraries can be obtained by chemical modification of POI ligands at their solvent-exposed region: upon binding, the ligand modifies the POI surface to potentially enable de novo protein-protein interactions with an E3 ligase. If the interaction is functional, degradation of the POI would be observed. A proof of concept for this approach comes from serendipitously discovered glue degraders GNE-0011^11,12, CR8¹³ and UNC8153^14,15, simple derivatives of POI inhibitors, that present chemical fragments on the POI surface, allowing interactions with components of E3 ligase complexes (Fig. 1).

**Fig. 1: Chemical derivatisation of protein inhibitors to facilitate E3 ligase recruitment.**

We applied this ligase-agnostic approach via chemical modification of POI ligands at their solvent-exposed region to create chemical degraders of bromodomain-containing protein 9 (BRD9), defined hereafter as Targeted Glues^TM. BRD9 is a component of the non-canonical BRG-/BRM-associated factor (ncBAF) chromatin remodelling complex¹⁶, that has been shown to be degradable by both CRBN and VHL-based PROTACs^{17,18,19,20,21,22,23,24}. Two CRBN-based PROTAC degraders of BRD9, CFT8634 from C4 Therapeutics and FHD-609 from Foghorn Therapeutics, have shown encouraging efficacy in preclinical models of disease²⁵ and have been tested in Phase 1 clinical trials in oncology (NCT05355753 and NCT04965753)²⁶. However, no molecular glue degrader has been reported for BRD9 to date.

We report the discovery and elucidate the mode of action of a potent and selective BRD9-targeted glue degrader that covalently recruits DCAF16. The drug-like properties of the compound allowed a proof-of-concept degradation study in vivo, which demonstrates the potential for DCAF16 covalent recruitment by targeted glues to be a viable alternative to traditional CRBN or VHL-based PROTACs.

Results

AMPTX-1 is a potent and selective BRD9 degrader

A focused library of potential targeted glues was prepared by modification of high-affinity BRD9 ligands with a proprietary library of molecular scaffolds bearing diverse electrophilic warheads. We reasoned that diverse molecular scaffolds could form non-covalent interactions at the surface of BRD9, facilitating the formation of potential protein-protein interactions with an E3 ligase, aiding covalent E3 ligase engagement and BRD9 degradation. This library was screened in a HiBiT endpoint degradation assay, in which C-terminally tagged BRD9-HiBiT HEK293 cells were treated with compounds for 6 h, and BRD9 protein levels were quantified¹⁹. Compounds identified as having promising degradation activity were further optimised.

An optimised compound having a warhead with a cyanoacrylamide-bearing tetrahydroisoquinoline scaffold, AMPTX-1, was identified as a potent BRD9 degrader (DC₅₀ = 0.05 nM, D_max = 88%, upon 6 h treatment, Fig. 2). We reasoned that the cyanoacrylamide moiety may be important for the degradation activity. To test this hypothesis, a des-cyano analogue, compound 2, was prepared and tested in the same assay. As expected, compound 2 did not show protein degradation across the concentration range tested (Fig. 2a, b), suggesting that the cyanoacrylamide moiety was essential to the degradation activity of AMPTX-1. Both compounds were subjected to live cell kinetic profiling to better understand their activity over a 24 h period (Fig. 2c, Supplementary Fig. 1a). The results were comparable to those observed in the endpoint assay, with the D_max observed at concentrations above 1 nM and reached after 3 h of treatment. The degradation profile was sustained at higher concentrations throughout the duration of the experiment (24 h), and no effect on cell viability was observed (Supplementary Fig. 1e). As expected, the negative control compound 2 did not produce any degradation of the target (Supplementary Fig. 1a).

**Fig. 2: AMPTX-1 is a potent and selective degrader of BRD9.**

AMPTX-1 induced degradation of endogenous, untagged BRD9 in solid and liquid tumour cancer cell lines (MCF-7 and MV4-11, respectively) within 6 h (Fig. 2d; Supplementary Fig. 1b–d). Potent BRD9 degradation was maintained in both cell lines (MV4-11 - DC₅₀ = 0.5 nM, D_max = 93%; MCF-7 - DC₅₀ = 2 nM, D_max = 70%) with no effect on cell viability (Supplementary Fig. 1g, h), and a characteristic hook effect is noticeable at concentrations above 100 nM in both cell lines (Fig. 2d). To assess degradation selectivity across the proteome in a disease-relevant cell line (MV4-11), we conducted an unbiased quantitative multiplexed tandem mass tag (TMT) proteomics experiment. MV4-11 cells were treated in triplicate for 6 h with either AMPTX-1 or compound 2, each at a concentration of 100 nM (approximately 100-fold over the DC₅₀ of AMPTX-1). Out of 8350 quantified proteins, BRD9 was the only protein significantly degraded by AMPTX-1 (adjusted P-value = 1.84e-9) (Fig. 2e; Supplementary Data 1). Conversely, treatment with compound 2 resulted in no significant changes in any proteins, including all quantified BET family members (Supplementary Fig. 2a). Interestingly, BRD9 ligands containing a pyridone scaffold are known to bind BRD7 with similar affinity (BRD9 IC₅₀ = 37 nM; BRD7 IC₅₀ = 41 nM for the dimethyl pyridone analogue)²⁷, but no degradation of BRD7 was observed in this experiment, suggesting that selectivity of the compound AMPTX-1 may come from more effective and functional ternary complex formation and target ubiquitination, and not from selective target engagement. Consistent with the global proteomics, orthogonal immunofluorescence measurement of BRD4 and BRD7 protein levels in MV4-11 cells showed limited degradation of either protein following 24 h treatment with AMPTX-1 or compound 2 (Supplementary Fig. 2c, d).

AMPTX-1 covalently recruits DCAF16:DDB1 complex to BRD9

A close inspection of the degradation data in MV4-11 and MCF-7 cells (Fig. 2d) reveals the presence of a hook effect. This is a characteristic of bifunctional degraders and suggests saturation of two independent binding sites.

We first assessed whether BRD9 degradation is mediated via proteasome and Cullin RING E3 ligases (CRL), a class of NEDD8-dependent E3 ligases that are the most exploited for TPD approaches. Degradation of BRD9 induced by AMPTX-1 was blocked by the proteasome inhibitor Bortezomib and the NEDD8-activating E1 (NAE1) inhibitor MLN4924 (Fig. 3a), demonstrating that BRD9 degradation is proteasome and NAE1-dependent, and suggesting the ubiquitination step is CRL-mediated.

**Fig. 3: AMPTX-1-induced BRD9 degradation is dependent on the DCAF16 Cullin RING E3 ligase.**

We next performed a competition assay to assess whether degradation by AMPTX-1 could be impaired by an excess of either the BRD9 inhibitor BI-7273²⁷ or our warhead 3 (structure in Supplementary Fig. 3a). As expected, excess BI-7273 (10 µM, approximately 300-fold over IC₅₀ for BRD9) was able to block BRD9 degradation mediated by AMPTX-1, presumably through competition for BRD9 cellular engagement (Fig. 3a). In contrast, the same concentration of warhead 3 (10 µM) failed to prevent degradation (Fig. 3b). This suggests that the warhead alone could not efficiently engage with its putative E3 ligase target at a binary level, indicating that the mode of action of AMPTX-1 shows characteristics of a molecular glue but, with the presence of the hook effect, giving a hybrid mechanism of action.

To identify cellular CRL components engaged by AMPTX-1, we performed a ternary complex pulldown that utilises the HiBiT tag on BRD9 to enrich for proteins that form protein-protein interactions following treatment with AMPTX-1 or compound 2. BRD9-HiBiT HEK293 cells were treated for 1 h with DMSO or increasing concentrations of AMPTX-1 or compound 2 (1, 10, 300, and 1000 nM). To ensure maximum retention of the ternary complex, all samples were pre-treated for 1 h with 3 µM of MLN4924. Comparisons of protein recruitment to BRD9 by AMPTX-1 and compound 2 using proteomics showed significant enrichment of the substrate receptor DCAF16 and its adaptor DDB1, members of the CRL4^DCAF16 E3 ligase complex (Fig. 3c; Supplementary Data 2), in a dose-dependent manner, with increasing quantification observed across 1, 10, 300, and 1000 nM of AMPTX-1 (Supplementary Fig. 3b, Supplementary Data 2). DDB1 and DCAF16 were the only proteins significantly enriched at all concentrations, indicating excellent specificity of ternary complex formation by the compound.

To functionally validate the dependency on DCAF16, we generated DCAF16-knockout (KO) MCF-7 cells and assessed the impact on BRD9 degradation. AMPTX-1 showed significantly impaired activity in the KO cells compared to the WT parental control, consistent with a functional role for DCAF16 in the mode of action of AMPTX-1 (Fig. 3d).

The importance of the stereochemistry of the tetrahydroisoquinoline (THIQ) moiety for degradation potency and ternary complex formation was investigated. Chiral separation of a key intermediate in the synthesis of AMPTX-1 was performed (see “Methods” section) to obtain the separate enantiomers of AMPTX-1 (ent-1: 97.8% ee and ent-2: 91.0% ee). AMPTX-1-ent-1 and AMPTX-1-ent-2 were tested in MV4-11 cells using an immunofluorescence (IF)-based BRD9 degradation assay (Fig. 4a). Only one of the two enantiomers, AMPTX-1-ent-1, induced potent and deep degradation (DC₅₀ 0.2 nM, D_max 94%), while AMPTX-1-ent-2 showed significantly weaker activity (DC₅₀ 5 nM, D_max 43%). Cyanoacrylamides are known to reversibly form covalent bonds, and we confirmed reversible formation of an AMPTX-1 glutathione adduct in a dilution experiment (Supplementary Fig. 4)^28,29. Moreover, multiple cysteines on DCAF16 are known to be engaged by covalent degraders to induce targeted protein degradation^30,31,32,33. To assess whether a covalent reaction with DCAF16 occurred, we incubated recombinant DCAF16:DDB1:DDA1 (3.6 µM) complex with an excess of AMPTX-1-ent-1 or AMPTX-1-ent-2 (18 µM, 5 molar equivalents). Quantification of the modified/unmodified DCAF16 ratio was performed by ESI/MS analysis of the reaction mixtures (Fig. 4b, c). Despite the excess of warheads, the most abundant species was found to be the unmodified E3 ligase, with 71% and 82% unmodified DCAF16 with AMPTX-1-ent-1 and AMPTX-1-ent-2, respectively. The majority of modified DCAF16 showed mass shift consistent with a single adduct, but a small population (~4%) for both AMPTX-1-ent-1 and AMPTX-1-ent-2 showed mass shifts consistent with >1 modification, suggesting multiple cysteines can react with the warhead when present at 5-fold molar excess.

**Fig. 4: BRD9 and warhead stereochemistry facilitate covalent modification of Cysteine 58 on CRL4^DCAF16.**

The experiment was repeated in the presence of the bromodomain of BRD9 (residues 134-250; BRD9^BD) 3.6 µM, 1 molar equivalent: under these conditions, the main species measured was modified E3 ligase, with 72% and 64% adducted DCAF16 with AMPTX-1-ent-1 and AMPTX-1-ent-2, respectively. These data indicate that the presence of BRD9^BD shifts the equilibrium in favour of covalent adduct formation and suggest that engagement of BRD9 may be required to achieve sufficient labelling of DCAF16 in a cellular environment. While the molecular weight shift indicates only single adduct formation on DCAF16 by AMPTX-1-ent-1, a significant fraction (16%) of DCAF16 shows double or triple Cysteine modification by AMPTX-1-ent-2. This suggests that stereochemistry within the tetrahydroisoquinoline scaffold may guide the specificity of adduct formation, which may relate to how the different enantiomers are accommodated within the BRD9-DCAF16 ternary complex. Given the more potent degradation activity observed with AMPTX-1-ent-1 compared to AMPTX-1-ent-2, we postulated that a specific ternary complex, relying on engagement of a single cysteine in DCAF16, can yield much deeper degradation of BRD9 when compared to simultaneously engaging multiple cysteines sub-optimally.

Modified DCAF16 (AMPTX-1-ent-1) was subjected to trypsin digestion and peptide mapping. This experiment identified DCAF16-Cys58 as the cysteine primarily modified by AMPTX-1-ent-1 (Supplementary Fig. 5 and Supplementary Data 3). Intact MS of C58S mutant DCAF16 (DCAF16^C58S-DDB1) showed significantly less modification (18%) compared to wildtype DCAF16 (72%) when incubated with a 5-fold molar excess of AMPTX-1-ent-1 and BRD9^BD, confirming that Cys58 is the primary cysteine engaged by the compound (Fig. 4d). No adduct was observed on the C58S/C173S/C178S triple mutant DCAF16 (DCAF16^C58S/C173S/^C178S-DDB1), indicating that a minority of AMPTX-1-ent-1 adducts DCAF16-Cys173 and/or DCAF16-Cys178 recombinant proteins in the experimental conditions tested (Supplementary Fig. 5e).

To validate the cellular dependency on DCAF16-Cys58, we generated a homozygous C58S knock-in mutation of DCAF16 in MCF-7 cells using CRISPR-Cas9 genome editing and assessed the impact on BRD9 degradation. AMPTX-1-ent-1 showed significantly impaired activity compared to WT parental control, and similar to that observed in DCAF16 KO, which is consistent with a functional role for DCAF16-Cys58 in the degradation of BRD9 (Fig. 4e).

Oral administration of AMPTX-1 induces sustained BRD9 protein degradation in a xenograft model

To assess whether the BRD9-targeted glue degraders were suitable to demonstrate targeted protein degradation via DCAF16 in vivo, the mouse pharmacokinetic (PK) profiles of the single enantiomers were investigated upon intravenous (IV) and oral administration (PO). Oral bioavailability (F) was 20% and 30% for compound AMPTX-1-ent-1 and AMPTX-1-ent-2, respectively, with moderate elimination half-lives (Supplementary Table 1). Free fraction for AMPTX-1-ent-1 and AMPTX-1-ent-2 in plasma was found to be 0.6% and 0.5%, respectively. The free drug concentration profile over time, upon oral administration at 10 mg/kg is shown in Fig. 5a. At this dose, systemic exposure of free AMPTX-1-ent-1 was higher than the in vitro MV4-11 DC₅₀ (0.2 nM) for at least 4 h, suggesting that the PK properties of AMPTX-1-ent-1 were adequate for a pharmacodynamic (PD) proof-of-concept study in vivo.

**Fig. 5: In vivo Pharmacokinetic and Pharmacodynamic effects of AMPTX degraders.**

To ensure the free concentration of degrader in plasma was closer to DC₉₀ (1.4 nM), a dose of 50 mg/kg PO was chosen for the PD study. Mice subcutaneously implanted with MV4-11 xenografts received two oral doses of AMPTX-1 enantiomers, the first dose at t = 0 h, the second dose at t = 8 h. Tumour samples were collected at 10 and 24 h after the initial dose (2 and 16 h post last dose), and BRD9 protein levels were quantified by western blot analysis (Fig. 5b, c). At 2 h post last dose, a statistically significant 82% BRD9 degradation (p = 0.0091) was observed for the samples derived from the animals treated with AMPTX-1-ent-1, whereas only 31% degradation was achieved by the less active AMPTX-1-ent-2 at the same time point (Supplementary Fig. 6). Notably, degradation was maintained at 16 h post last dose, with 78% BRD9 degradation observed AMPTX-1-ent-1 treatment and 28% for AMPTX-1-ent-2 (Supplementary Fig. 6). At this time point the concentration of total drug in plasma being over 3 orders of magnitude lower compared to the levels detected at the 2 h post last dose, highlighting an extended pharmacodynamic effect.

Discussion

AMPTX-1 is a potent and selective covalent targeted glue that induces degradation of BRD9 by recruitment of DCAF16, a relatively uncharacterised E3 ligase. While DCAF16’s endogenous functional and biological role remains elusive, it has recently emerged as an E3 ligase that can be commandeered by small molecules. DCAF16-mediated degradation of neo-substrate BRD4 has been demonstrated by means of a variety of electrophilic bifunctional degraders^30,31,32,34. However, these compounds show deep target degradation only at relatively high (micromolar) concentrations and have been shown to react with a wide range of cysteine-containing proteins in the proteome³², limiting their potential for drug development. During the preparation of this manuscript, electrophilic derivatives of JQ1 were reported to recruit DCAF16 (and the off-target Cyclin G-associated kinase, GAK) to BRD4 by a template-assisted covalent modification mechanism^33,34. Consistent with our findings, DCAF16-Cys58 was shown to be the target of the BRD4-templated modification by the JQ1 derivative MMH2³³, and cryo-EM data revealed that BRD4^BD2 acts as a structural scaffold facilitating the covalent bond formation³³. Moreover, inspection of cryo-EM structures available for ternary complexes comprising BRD4 and DCAF16 reveals structural complementarity among the two proteins^33,35. As degradation of BRD4 by DCAF16 can be induced by diverse degrader chemotypes, it has been suggested that DCAF16 may play a role in BRD4 protein quality control³⁰.

Our work shows that BRD9 can also be efficiently glued to DCAF16 by a targeted glue bearing a reversible covalent warhead. The molecular complexity inherent within targeted glues may facilitate additional protein-ligand and protein-protein interactions, expanding the neo-substrate scope of DCAF16 beyond BRD4, as we have shown here. It has been noted that the structural plasticity of DCAF16 may confer high structural flexibility to the DDB1-DCAF16 E3 ligase complex³³, allowing for recognition of a variety of neo-substrates in addition to the proposed endogenous substrate, SPIN4³⁶.

Proteomics experiments indicate that AMPTX-1 selectively targets BRD9 for degradation by inducing the formation of a ternary complex with DCAF16. The absence of BRD7 degradation, which is engaged with similar potency by this class of bromodomain binders²⁷, also points to a model where specific protein-protein interactions between ligase and target protein, and/or preferential protein ubiquitinability³⁷, add a layer of selectivity beyond the simple target engagement. Selective degradation of BRD9 over BRD7 protein is also observed with CRBN-based PROTACs^17,20, but not VHL-based PROTACs¹⁸.

A unique, counterintuitive feature of the mode of action here described is the presence of a hook effect^38,39, albeit mild, at high compound concentrations in our protein degradation assays (Figs. 2 and 3). Interestingly, a similar hook effect was observed by Li et al.³³ in both ternary complex formation and degradation assays of related covalent glue molecules. Hook effects are often observed with heterobifunctional PROTACs because they can engage each protein individually. Typically, a hook effect results from saturation of the individual POI and E3 binding sites in 1:1 engagement, out-competing the 1:1:1 ternary complex formation at high degrader concentrations. However, despite the presence of a hook effect, no competition effects of AMPTX-1-induced BRD9 degradation were observed in the presence of a large excess of warhead 3, suggesting that the warhead alone is unable to effectively saturate the E3 ligase. Therefore, the hook effect observed with the targeted glue could be explained by an E3 ligase covalent loading process coupled to BRD9 binding site saturation, according to the following steps (Fig. 6), similar to what has been described for BRD4^33,34 and builds on our understanding of various degradation mechanisms.

1.
BRD9 engagement with AMPTX-1;
2.
ternary complex formation between DCAF16 and AMPTX-1-bound BRD9, leading to DCAF16 labelling;
3.
ubiquitin transfer and BRD9 dissociation from the complex;
4.
degradation cycle established: DCAF16 remains labelled and primed for engagement with additional BRD9 proteins; and
5.
hook effect: excess of targeted glue leads to saturation of both BRD9 and DCAF16.

**Fig. 6: Schematic representation of the E3 ligase loading process and hook effect.**

The hook effect suggests the persistence of a portion of modified DCAF16, essentially a neo E3 ligase³¹ bearing a covalently modified substrate receptor subunit (in this case a DCAF16:AMPTX-1 covalent adduct). This population of E3 ligase is primed for the target recruitment and is likely to persist inside the cell^40,41,42. The half-life of this species will depend on the half-life of DCAF16 (155.4 h, as reported from proteomics studies⁴³) and the reversibility of the covalent bond formed between DCAF16-Cys58 and AMPTX-1. Dissociation of the warhead, possible for this class of reversible covalent electrophiles, may be kinetically unfavoured because of protein-compound interactions that could stabilise the adduct and prevent the retro-Michael reaction by steric hindrance⁴⁴. Persistent modification of the long half-life E3 ligase DCAF16, together with the relatively slow resynthesis rate of BRD9 in MV4-11 (8.6 h as determined by cycloheximide chase assay, Supplementary Fig. 7), may underscore the high potency of this compound and the extended pharmacodynamic effect observed in our in vivo study.

Here, we also demonstrate the unprecedented potential of hijacking an E3 ligase via a covalent targeted glue to achieve robust and persistent target protein degradation in vivo in an animal model. Deep protein degradation was achieved in the tumour, and degradation levels were maintained when plasma concentrations of the drug dropped by over 1000-fold. This apparent disconnect between PK and PD may be a feature of the covalent mode of action, as previously demonstrated for reversibly covalent BTK inhibitors^44,45.

In summary, we have identified and characterised a potent and selective BRD9 degrader with a specific mode of action distinct from both PROTAC degraders and glues, which we term Targeted Glues^TM. Covalent modification of the E3 ligase DCAF16 is enabled by a templating effect based the BRD9 protein, which likely positions the covalent warhead in proximity of DCAF16-Cys58. Proteomics studies show that AMPTX-1 induces selective target degradation through the selective recruitment of DCAF16 to BRD9. Importantly, the bioavailability of AMPTX-1 supports using this chemistry for drug development. More generally, this work exemplifies a drug-like small molecule is able to selectively engage an undruggable protein surface, that of DCAF16, via a ligand to the target protein (BRD9), which enables its degradation via event-driven pharmacology. This approach could be used to recruit other E3 ligases and to expand the potential for TPD approaches and induced proximity pharmacology in the development of therapeutics.

Methods

Ethics

Animals used in this paper were female BALB/c nude mice, 6–8 weeks old, weighing 18–22 grams, sourced from Beijing Vital River Laboratory Animal Technology Co., Ltd.

Mice were housed in individually ventilated cages (IVCs) under controlled environmental conditions, with three animals per cage. The temperature was maintained at 20–26 °C, and the humidity was kept between 40% and 70%. Throughout the study, animals had ad libitum access to irradiated, sterilised dry pellet feed and sterile drinking water.

All procedures related to animal care, handling, and treatment were conducted in accordance with protocols approved by the Institutional Animal Care and Use Committee (IACUC) of WuXi AppTec, and in compliance with the guidelines of the Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC).

Routine monitoring was conducted daily to assess the impact of tumour growth and treatment on general health indicators such as mobility, food and water intake (observationally), body weight changes, grooming (eye/hair condition), and any other abnormal signs as defined in the study protocol. Mortality and clinical observations were documented by group and animal number.

Cell culture & models (incl. cell engineering)

HEK293 BRD9-HiBiT KI LgBit cells were obtained from Promega Corp. (CS3023412) MCF-7 (HTB-22) cells and MV4-11 (CRL-9591) cells were obtained from ATCC. The identity of the MCF-7 and MV4-11 cell lines was confirmed by STR by the vendor at purchase. MCF-7 DCAF16 KO and DCAF16^C58S (homozygous knock-in mutation) cells were generated using CRISPR-Cas9 genome editing (EditCo). Briefly, all cell lines were maintained in Dulbecco’s Modified Eagle’s Medium with glutaMAX (Gibco 31966-021) supplemented with 10% FBS (v/v) (Thermo Fisher A5256701). MV4-11 cells were obtained from ATCC (CRL-9591) and maintained in IMDM (Thermo Scientific 21056023) supplemented with 10% FBS. All lines were grown at 37 °C in a humidified 5% CO₂ atmosphere.

BRD9^BD (residues 134–250) protein production

The bromodomain of BRD9 (residues 134–250) was subcloned into a pRSET vector with a 6His-TEV affinity tag (pRSET_BRD9_134-250). 10 µl of competent cells of E. coli BL21 (DE3) were transformed using 1 µl of 50 ng/µl plasmid DNA following standard procedures. One colony was inoculated in 100 mL LB plus 2% glucose and grown overnight at 37 °C, with medium shaking (150 rpm) until OD₆₀₀ = 1.5. 50 mL was used to inoculate 1 Litre of Luria Broth (LB) medium plus 0.2% glucose or 1 L of Terrific Broth (TB, used readymade EZMix™ powder from Sigma plus 4% glycerol) medium. Two litres of each were set up. Starter OD₆₀₀ = 0.075. Cultures were grown to OD₆₀₀ = 1.1 and induced with 0.5 mM IPTG overnight at 17 °C. Cells were cooled in a cold cabinet at 4 °C for 30 min before induction. The next day, the cells were spun down at 2 × g (Beckman Avant J-26 XP rotor JLA8.1000) for 30 min, washed once with 25 mL cold PBS and stored at −80 °C. 1 mL of each cell pellet was lysed in 0.5 mL B-per in the presence of lysozyme at 1 mg/mL and DNAse I at 10 µg/mL. Post lysis, whole cell and soluble fractions were analysed on SDS-PAGE followed by Anti-His western blotting.

Cell pellets were thawed in the presence of Lysis Buffer (50 mM Tris pH 8.0, 300 mM NaCl, 10 mM Imidazole, 10% glycerol, 1 mM TCEP, 1 mg/ml lysozyme, (1:1000) DNase I stock for 10 µg/ml final, and 2 protease tablets (SIGMAFAST™ Protease Inhibitor Cocktail Tablets, EDTA-Free Cat# S8830)). Soluble proteins were released from cell pellets by using the cell disruptor (Constant Systems) at 25 PSI. Insoluble proteins were removed by high-speed centrifugation (Beckman Avant J-26 XP, Rotor JA25.50, 50,000 × g) for 30 min at 4 °C. For each litre of cell culture, 0.8 mL (1.6 mL of slurry) of Ni-NTA agarose beads (Ni-NTA Affinity Resin – Amintra, cat# ab270549) were employed in a 50 mL tube. Each 0.8 mL of resin was washed in 25 mL of water, followed by 25 mL of Buffer A (50 mM Tris, pH 8.0, 300 mM NaCl, 10 mM Imidazole, 10% glycerol, 1 mM TCEP). Binding was performed at 4 °C, gently rolling overnight.

The following day, Ni-NTA beads were collected through centrifugation at a low speed (Eppendorf 5810 R, Rotor 1730 × g for 10 min) and collected into one 50 mL Falcon tube and washed twice with 50 mL of buffer A. The beads were transferred to a 10 mL OmniFit column. The column was washed with 20 mL of Buffer A and then 20 mL with 15% (45 mM Imidazole) Buffer B (50 mM Tris, pH 8.0, 300 mM NaCl, 300 mM imidazole, 1 mM TCEP, 10% glycerol). Protein elution was then achieved using a step elution at 100% Buffer B. For His-tag cleavage, samples were incubated with TEV protease (1:10) at 20 °C overnight.

For final purification, samples were loaded onto a size exclusion S75 26/60 column pre-equilibrated in 25 mM Hepes pH 7.5, 200 mM NaCl, 10% glycerol, 1 mM TCEP. Fractions were collected and analysed by SDS-PAGE. Highest purity fractions were concentrated, aliquoted and frozen at −80 °C.

DCAF16:DDB1ΔB and DCAF16^C58S:DDB1ΔB protein production

WT, DCAF16^C58S(1–216), and DCAF16^{C58S/C173S/C178S} (1–216) with N-terminal 6His-TEV affinity tags, were subcloned into individual pOET1 vectors. DDB1(1–1140, Δ396-705, linker) and full length DDA1(1–102) were both subcloned into a single pOET5.1 duet vector for co-expression. SF21 insect cells in suspension were grown in SF-900TM II SFM media (Gibco®) at 28 °C and 0.8 × g in shake flasks with vented caps on a shaking platform. Cells were passaged every 2–3 days to a density within the log phase of the growth to maintain optimal growth conditions. The SF21 cell line used was routinely passaged at least 4 times before use in an expression growth medium and had a viability of above 95%.

On the day of infection, the cell number was adjusted to 2 × 106 cells/mL by the addition of fresh SF900II media prior to co-infection with the two baculovirus constructs at a 1:1 ratio and MOI of 2.5. Approximately 72 h after infection, the cell pellet was harvested by centrifugation at 7000 × g for 20 min at 4 °C before freezing the cell pellet at −80 °C prior to purification.

Cells were counted using a Beckman Coulter Vi-Cell BLU Cell Viability Analyzer, which automates the Trypan Blue Dye Exclusion method for cell viability analysis.

Harvested cells were resuspended in 50 mM Tris-HCl, pH 8.0, 300 mM NaCl, 20 mM Imidazole, 10% Glycerol, 2 mM TCEP (lysis buffer), supplemented with protease inhibitors (AbCam) and Nuclease A (GenScript). The cells were disrupted by sonication (Q500 QSonica), 5 × 20 s pulse, 40 s rest, and then clarified by centrifugation at 185,000 × g for 1 h (Ti45 rotor, Optima XPN-80 Ultracentrifuge). The soluble supernatant was filtered using a 0.45 μm membrane prior to nickel affinity chromatography.

The filtered supernatant was applied directly to 10 mL Ni2+ HisTrap Crude FF resin (Cytiva).

The resin was washed using 5 CV lysis buffer and eluted using a 10 CV 0–100% lysis buffer:IMAC elution buffer gradient (50 mM Tris-HCl, pH 8.0, 300 mM NaCl, 600 mM Imidazole, 10% Glycerol, 2 mM TCEP). Elution fractions containing the DCAF16:DDB1:DDA1 complex were pooled and then treated with TEV protease in a 1:15 ratio. The sample was dialysed overnight at 4 °C into 50 mM Tris-HCl pH 8.0, 300 mM NaCl, 20 mM Imidazole, 10% Glycerol, 1 mM TCEP.

Following dialysis, the cleaved DCAF16:DDB1:DDA1 sample was passed through 10 mL Ni2+ HisTrap Crude FF resin (Cytiva) and the flow through collected. The subsequent sample was diluted to reduce the NaCl concentration to <50 mM and applied to a HiTrap Q FF column (Cytiva), which had been equilibrated into 25 mM Tris-HCl, pH 8.2, 50 mM NaCl, 10 2% Glycerol, 1 mM TCEP (IEX buffer). A 40 CV 0–0.5 M NaCl IEX buffer gradient, followed by 5 CV of 1 M NaCl IEX buffer, was used to elute the protein.

The protein was then concentrated by ultrafiltration (Amicon) and loaded onto a Superdex 200 column (Cytiva) equilibrated in 25 mM Tris-HCl, pH 8.0, 300 mM NaCl, 10% Glycerol, 1 mM TCEP. The protein eluted as a single peak representing a 1:1:1 DCAF16:DDB1:DDA1 complex. The protein was snap frozen in liquid N2 and stored at −80 °C prior to use.

BRD9-HiBiT endpoint and kinetic assays

HEK293 BRD9-HiBiT KI (LgBiT) Cells (Promega CS3023412) were seeded at 8 × 10³ cells per well in 36 μL of media in sterile white bottom 384-well plates (Greiner 781080) and incubated overnight at 37 °C with 5% CO₂. The following day, compounds were prepared at 1000× final concentration in DMSO and diluted 1:100 in maintenance media. In case of competition of mechanism of action studies (NEDDylation- or proteasome-dependence), plates were pre-treated with DMSO, 3 µM NAE1 inhibitor MLN4924, 10 µM proteasome inhibitor Bortezomib, 10 µM BI-7273 (BRD9 POI ligand) or 10 μM of warhead 3 with a Multidrop Pico8 digital dispenser and incubated for 1 h. Cells were treated with AMPTX-1 and incubated for 6 h. BRD9 expression was quantified using the NanoGlo Lytic Endpoint assay (Promega N3050). The lytic buffer was equilibrated to room temperature for 10–15 min before the lytic reagent was prepared by adding substrate (1:50) and LgBit protein (1:100) in the lytic buffer. 40 µl lytic reagent mix was added to each well of the cell plate, and the plates were centrifuged briefly at ~50 G. Plates were incubated with shaking (350 rpm/0.05 × g) on a shaking platform for 12–24 min before reading using the LUM plus module of a BMG Pherastar FSX. The percentage of BRD9 remaining was calculated by normalisation to average data from high and low control wells (DMSO-treated cells and no cells, respectively).

For live cell kinetic 8 × 10³, Hek293 BRD9-HiBiT KI LgBit cells per well were seeded in 80 μL of phenol red-free OptiMEM media (Gibco 11058021) supplemented with 4% FBS (ATCC 302025) in white opaque 384-welled plates (GBO 781080). Following overnight incubation, 20 μL of assay media supplemented with 1:40 of Nano-Glo Endurazine (Promega N2571) was added to each well and incubated at 37 °C for 3 h. Compounds of interest were dispensed into plates using the FlexDropIQ non-contact dispenser and normalised for DMSO concentration to 0.1% (vol/vol). Plates were placed immediately in the ClarioSTAR microplate reader equipped with ACU (BMG) that was pre-normalised to 5% CO₂ and 37 °C. Recording was taken every 5 min with 0.8 s integration time.

Resulting measurements were analysed by determination of fractional luminescence and normalisation to the first time point. Four-parameter variable slope dose response, as well as one-phase decay fitting, was performed in GraphPad Prism.

BRD9 Immunofluorescence assay

A suspension of MV4-11 cells (ATCC, CRL-9591) was prepared in phenol red-free assay media supplemented with 10% FBS, and cells were seeded at 20,000 cells per well (45 μL) in sterile black poly-D-lysine-coated 384-well plates (Greiner 781948). Compounds were prepared at 1000× final concentration in DMSO, diluted 1:100 in assay media, and 5 μL compound was added to each well of the cell plate. Cells were incubated for 24 h at 37 °C with 5% CO₂. All the following incubations for immunofluorescence staining were at room temperature. 15 μL of 16% PFA was added to each well (3.7% final concentration) and the cells were fixed for 15 min, then washed twice with DPBS. Cells were permeabilised with 0.1% Triton X-100 for 10 min, Triton X-100 was removed, then blocked with 1% BSA in DPBS for 1 h. Cells were stained with 25 μL anti-BRD9 (clone E4Q3F) antibody (CST 48306) diluted 1:25,600 in 1% BSA in DPBS for 2–3 h. Wells were washed twice with DPBS, then incubated with 25 μL of 1% BSA containing a 1:1000 dilution of Anti-rabbit Alexa Fluor™ 647 secondary antibody (Thermo Scientific A21244; Lot no.: 2836809) and 1 μg/mL Hoechst nuclear counter stain (Abcam ab228551) for 1 h. For selectivity assays, cells were stained with 25 μL anti-BRD4 BL-149-2H5 antibody (Bethyl Lab A700-004, Clone: BL-149-2H5, Lot no.: 4) diluted 1:2000 or anti-BRD7 D9K2T (CST 15125S, Clone: D9K2T Lot no.: 1) diluted 1:1000 in 1% BSA/ 0.5% Triton X-100 in DPBS overnight at 4 °C. Wells were washed twice with DPBS prior to imaging on a Perkin Elmer Operetta CLS with a 10× air lens. Images were processed using Harmony High-Content Imaging and Analysis Software (Perkin Elmer), and the mean contrast ratio of Alexa Fluor™ 647 in central nuclei was used to quantify BRD9 protein levels. Results were normalised using DMSO and a no-primary antibody control, and a four-parameter variable slope dose response was fitted in GraphPad Prism. To determine the effect of compounds on the viability of the cells at the required time point, the intensity measured from the Hoeschst nuclear counterstain in each well of treated cells was normalised to the DMSO control, and a four-parameter variable slope dose response was fitted in GraphPad Prism.

Cell viability – CellTiter-Glo

A suspension of MV4-11 cells was prepared in phenol red-free assay media supplemented with 10% FBS, and cells were seeded at 5000 cells per well (90 μl) into a white 96-well plate (Greiner 655083). Compounds were prepared at 1000× final concentration in DMSO, diluted 1:100 in assay media, and 10 μL of compound was added to each well of the cell plate. 100 μl of CellTiter-Glo (Promega) was immediately added to cells in the Day 0 plate, the plate was shaken for two minutes and incubated at room temperature in the dark for an additional eight minutes, before luminescence was measured using a Pherastar plate reader (BMG). Treated plates were incubated for five days at 37 °C, 5% CO₂. On completion of incubation, CellTiter-Glo was added to each well and the luminescence measured. Cell viability was calculated using the day 0 and day 5 luminescence values, and a four-parameter variable slope dose response was fitted in GraphPad Prism.

Western blot protocol

MV4-11 cells were treated with AMPTX-1 and compound 2 at the indicated concentrations for 24 h, washed in PBS and lysed in AML lysis buffer (50 mM Tris, pH 7.6, 150 mM NaCl, 2% SDS, 10% Glycerol). Protein lysates were mixed with 2× LDS sample buffer, SDS loading buffer (Invitrogen, NP0007), with 10 mM DTT. Samples were boiled, and approximately 20 µg was loaded in an SDS-PAGE gel. Proteins were transferred to 0.45 µM nitrocellulose membrane. After blocking in 5% milk, membranes were incubated with primary antibodies BRD9 (CST, 58906, Clone: E9R2I Lot no.: 3), diluted 1:1000 and Rhodamine Anti-Tubulin. StarBright Blue 700 goat anti-rabbit IgG (BIO-RAD, 12004161, Lot no.: 64433886) secondary antibody was used, dilution 1:10,000, and blots were visualised with a ChemiDoc MP (BIO-RAD) and imaged processed in Image Lab software (BIO-RAD).

Cycloheximide chase experiment

A suspension of MV4-11 cells was prepared in phenol red-free assay media supplemented with 10% FBS, and cells were seeded at 20,000 cells per well (45 μL) in sterile black poly-D-lysine-coated 384-well plates (Greiner 781948). Cells were treated with 20 μg/mL of cycloheximide, and samples were removed at various time points from 30 min to 48 h. Samples were fixed. 15 μL of 16% PFA was added to each well (3.7% final concentration), and the cells were fixed for 15 min, then washed twice with DPBS. Cells were permeabilised with 0.1% Triton X-100 for 10 min, Triton X-100 was removed, then blocked with 1% BSA in DPBS for 1 h. Cells were stained with 25 μL anti-BRD9 E4Q3F antibody (CST 48306, Lot no.: 1) diluted 1:25,600 in 1% BSA in DPBS for 2–3 h. Wells were washed twice with DPBS, then incubated with 25 μL of 1% BSA containing a 1:1000 dilution of Anti-rabbit Alexa Fluor™ 647 secondary antibody (Thermo Scientific A21244, Lot no.: 2836809) and 1 μg/mL Hoechst nuclear counter stain (Abcam ab228551) for 1 h. Wells were washed twice with DPBS prior to imaging on a Perkin Elmer Operetta CLS with a 10× air lens. Images were processed using Harmony High-Content Imaging and Analysis Software (Perkin Elmer), and the mean contrast ratio of Alexa Fluor™ 647 in central nuclei was used to quantify BRD9 protein levels. Results were normalised using DMSO and a no-primary antibody control, and data fitted using non-linear regression and one-phase exponential decay.

Immunoprecipitation mass spectrometry (IP-MS) sample preparation

HEK293 cells expressing BRD9-HiBiT were seeded in 100 mm plates at 7 × 10⁶ cells/dish. The following day, all plates were pre-treated with MLN4924 at 3 µM for 30 min before the addition of DMSO, AMPTX-1 or compound 2 at the indicated concentrations for 2 h at 37 °C. Cell monolayers were washed 1× in ice-cold PBS, scraped into PBS, pelleted at 300 × g for 5 min and washed 1× in PBS. Cell pellets were resuspended in 500 µL NP-40 lysis buffer (Thermo, J60766.AK) containing protease inhibitors + 250 U benzonase (Thermo Fisher, 88701). Cells were lysed on ice for 30 min and cleared through centrifugation at 16,000 × g for 15 min. Supernatants were directly transferred to HiBiT antibody (Promega N7200, Lot no.: 0000562061, diluted 1:1000), bound dynabeads protein G (Thermo Fisher, 10765583) and incubated on an Eppendorf rotator for 30 min. Beads were isolated with a magnetic rack and washed 2× in PBS and 1× in H2O and resuspended in 25 μL of 50 mM TEAB. BRD9-enriched samples were reduced (10 mM TCEP, 55 °C for 1 h), alkylated (18.75 mM iodoacetamide, room temperature for 30 min) and then digested from the beads with trypsin (1.25 µg trypsin; 37 °C, overnight). Beads were placed on a magnetic rack, and digested peptides were removed and subjected to TMT labelling.

TMT LC-MS global proteomics sample preparation

MV4:11 cells were seeded at 8 × 10⁵ for 6 h. After seeding, cells were immediately treated with 100 nM of the indicated compounds in triplicate. At indicated time points, cells were washed twice with PBS, pelleted, and resuspended in 100 µL of lysis buffer (50 mM Tris pH 7.6, 150 mM NaCl, 2% SDS, 10% glycerol, protease inhibitors (Pierce A32961), 1 mM PMSF, 250 U benzonase (Thermo Fisher, 88701) and sonicated for 1 min at 100% power with a probe sonicator). Cell debris was removed by centrifugation at 21,000 × g for 10 min.

FASP-based tryptic digestion

246 µg of each sample was used for FASP as follows: Dithiothreitol was added to a final concentration of 83.3 mM, incubated at 99 °C for 5 min, then cooled to room temperature. VIVACON 500 filter units, 30,000 MWCO (Sartorius Stedim, Biotech GmbH, 37079 Goettingen, Germany) were washed with 8 M urea in 100 mM Tris/HCl pH 8.5, centrifuged at 14,000 × g, 20 °C for 15 min. Samples were loaded in 50 µL aliquots plus 200 µL 8 M urea, centrifuged at 14,000 × g, 20 °C for 15 min. The filter units were washed with 8 M urea, centrifuged at 14,000 × g, 20 °C for 15 min. 100 μL 50 mM Iodoacetamide in 8 M urea was added to the filter units and incubated for 30 min in the dark. Filter units were centrifuged at 14,000 × g, 20 °C for 10 min, then washed with 100 μL 8 M urea, centrifuged at 14,000 × g, 20 °C for 15 min. This wash step was repeated two times, for a total of three washes. Filters were further washed with 100 μL 50 mM Triethylammonium Bicarbonate Buffer pH 8.5 (TEAB, Thermo Fisher Scientific), centrifuged at 14,000 × g, 20 °C for 10 min. This wash step was repeated two times, for a total of three washes. Filter units were transferred to fresh tubes, and 1.25% (w/w) trypsin (Pierce MS grade, Thermo Fisher Scientific, Loughborough, LE11 5RG, UK) in 60 µL 50 mM TEAB was added. The Filter units were sealed with Parafilm, and digestion was performed overnight at 37 °C with shaking. Following overnight digestion, the filter units were centrifuged at 14,000 × g, 20 °C for 20 min, and the flow through containing the peptides was retained. The filter units were washed with 40 μL 50 mM TEAB, centrifuged at 14,000 × g, 20 °C for 10 min and then 50 μL 0.5 M NaCl, centrifuged at 14,000 × g, 20 °C for 20 min, and these washes were added to the flow through. Peptide samples were desalted and cleaned up using SepPak cartridges according to the manufacturer’s instructions (Waters, Milford, Massachusetts, USA). Eluate from the SepPak cartridge was evaporated to dryness and resuspended in 100 µl 50 mM TEAB, and 10 µl was used for a peptide assay. All chemicals from Merck Life Science UK Limited, Dorset, SP8 4XT, unless otherwise stated.

TMT labelling, high pH reversed-phase chromatography

For global proteomics samples, approximately 60 µg of each sample was labelled with Tandem Mass Tag (TMTpro) 18-plex reagents according to the manufacturer’s protocol (Thermo Fisher Scientific, Loughborough, LE11 5RG, UK). In addition, 5 µg of all 34 samples was combined to make a bridging sample, and two 60 µg aliquots of this bridging sample were labelled with the TMTpro 126 C tag. The samples were labelled as follows: TMT-127N = DMSO 6 h Repeat 1, TMT-127C = compound 2 6 h Repeat 1, TMT-128N = AMPTX-1 6 h Repeat 1, Experiment 2 layout; TMT-127N = DMSO 6 h Repeat 2, TMT-127C = compound 2 6 h Repeat 2, TMT-128N = AMPTX-1 6 h Repeat 2, TMT-131N = DMSO 6 h Repeat 3, TMT-131C = compound 2 6 h Repeat 3, TMT-132N = AMPTX-1 6 h Repeat 3.

For IP-MS, the entire digest was TMT labelled according to the manufacturer’s protocol (Thermo Fisher Scientific, Loughborough, LE11 5RG). The labelling was as follows: TMT-126C = DMSO Repeat 1, TMT-127N = 10 nM compound 2 Repeat 1, TMT-127C = DMSO Repeat 2, TMT-128N = 10 nM compound 2 Repeat 2, TMT-128C = DMSO Repeat 3, TMT-129N = 10 nM compound 2 Repeat 3, TMT-129C = 1 nM AMPTX-1 Repeat 1, TMT-130N = 300 nM compound 2 Repeat 1, TMT-130C = 10 nM AMPTX-1 - Repeat 1, TMT-131N = 300 nM compound 2 Repeat 2, TMT-131C = 10 nM AMPTX-1 Repeat 2, TMT-132N = 300 nM compound 2 Repeat 3, TMT-132C = 10 nM AMPTX-1 Repeat 3, TMT-133N = 300 nM AMPTX-1 Repeat 1, TMT-133C = 300 nM AMPTX-1 Repeat 2, TMT-134N = 300 nM AMPTX-1 Repeat 3, TMT-135N = 1000 nM AMPTX-1 Repeat 1, TMT-134C = 1000 nM compound 2 repeat 1.

Offline HpRP fractionation

For global proteomics, samples were combined to a total of 100 µg each TMT plex and desalted using a SepPak cartridge according to the manufacturer’s instructions (Waters, Milford, Massachusetts, USA). Eluate from the SepPak cartridge was evaporated to dryness and resuspended in buffer A (20 mM ammonium hydroxide, pH 10) prior to fractionation by high pH reversed-phase chromatography using an Ultimate 3000 liquid chromatography system (Thermo Fisher Scientific). In brief, the sample was loaded onto an XBridge BEH C18 Column (130 Å, 3.5 µm, 2.1 mm × 150 mm, Waters, UK) in buffer A, and peptides were eluted with an increasing gradient of buffer B (20 mM Ammonium Hydroxide in acetonitrile, pH 10) from 0 to 95% over 60 min. The resulting fractions (20 in total) were evaporated to dryness and resuspended in 1% formic acid prior to analysis by nano-LC MSMS using an Orbitrap Fusion Lumos mass spectrometer (Thermo Scientific).

For IP-MS, the entire labelled material was combined prior to SepPak desalting as described above.

Nano-LC MS3

High pH RP fractions were further fractionated using an Ultimate 3000 nano-LC system in line with an Orbitrap Fusion Lumos mass spectrometer (Thermo Scientific). In brief, peptides in 1% (vol/vol) formic acid were injected onto an Acclaim PepMap C18 nano-trap column (Thermo Scientific). After washing with 0.5% (vol/vol) acetonitrile 0.1% (vol/vol) formic acid peptides were resolved on a 250 mm × 75 μm Acclaim PepMap C18 reverse-phase analytical column (Thermo Scientific) over a 150 min organic gradient, using 7 gradient segments in solvent B (1–6% for 1 min, 6–15% for 58 min, 15–32% for 58 min, 32–40% for 5 min, 40–90% for 1 min, held at 90% for 6 min and then reduced to 1% over 1 min) with a flow rate of 300 nl min⁻¹. Solvent A was 0.1% formic acid, and Solvent B was aqueous 80% acetonitrile in 0.1% formic acid. Peptides were ionised by nano-electrospray ionisation at 2.0 kV using a stainless-steel emitter with an internal diameter of 30 μm (Thermo Scientific) and a capillary temperature of 300 °C. For global proteomics, 20 fractions were collected for SPS-MS3 analysis. For IP-MS, six fractions were collected.

All spectra were acquired using an Orbitrap Fusion Lumos mass spectrometer controlled by Xcalibur 3.0 software (Thermo Scientific) and operated in data-dependent acquisition mode using an SPS-MS3 workflow. FTMS1 spectra were collected at a resolution of 120,000, with an automatic gain control (AGC) target of 200,000 and a max injection time of 50 ms. Precursors were filtered with an intensity threshold of 5000, according to charge state (to include charge states 2–7) and with monoisotopic peak determination set to Peptide. Previously interrogated precursors were excluded using a dynamic window (60 s ± 10 ppm). The MS2 precursors were isolated with a quadrupole isolation window of 0.7 m/z. ITMS2 spectra were collected with an AGC target of 10,000, max injection time of 70 ms and CID collision energy of 35%.

For FTMS3 analysis, the Orbitrap was operated at 50,000 resolution with an AGC target of 50,000 and a max injection time of 105 ms. Precursors were fragmented by high-energy collision dissociation (HCD) at a normalised collision energy of 60% to ensure maximal TMT reporter ion yield. Synchronous Precursor Selection (SPS) was enabled to include up to 10 MS2 fragment ions in the FTMS3 scan.

TMT LS-MS proteomics data analysis

Raw data files were converted to mzML format using msconvert proteowizard (version 3.0.22167). Database searches were performed using MSFragger⁴⁶ (version 4.0) within FragPipe (version 21.1). A precursor mass tolerance of 20 ppm and fragment mass tolerance of 0.6 Da was used. Two missed cleavages were allowed with a minimum peptide length of six and a maximum length of fifty. Searches were performed against the reference human proteome from Uniprot (release 2023_05), including only reviewed accessions.

Carbamidomethylation of cysteine (+57.02146 Da) and TMTpro labelling of Lysine (+304.20715 Da) were set as fixed modifications. Oxidation of methionine (+15.9949 Da), N-terminal acetylation (+42.0106 Da), and N-terminal TMTpro labelling (+304.20715 Da) were set as variable modifications. False-discovery rate filtering was set to 1% at the PSM, peptide, and protein levels. Each channel intensity was normalised to the median intensity of all channels to account for protein loading differences, and a log₂ transformation was applied. For global proteomics, internal reference scaling normalisation⁴⁷ was performed to correct for batch effect between plexes and proteins with less than two unique peptides were excluded. Statistical analysis was performed with the moderated t-statistics implemented in the limma package⁴⁸ (v 3.56.2) within the R framework (v 4.3.1). A single linear model was constructed to test the degrader vs an equal concentration of the negative control compound. The Benjamini & Hochberg method for multiple comparisons was used to adjust the p-values. Proteins were considered significantly altered if they met the criteria of an adjusted p-value < 0.01 and an absolute log₂ fold change >1.

Intact MS

DCAF16-DDB1 or DCAF16^C58S-DDB1 or DCAF16^{C58S/C173S/C178S}-DDB1 (3.6 µM) was incubated with 18 µM compound in the presence or absence of 3.6 µM BRD9^BD (1:1:5) at RT for 2 h (n = 1). The samples were diluted to approximately 0.1 mg/ml with 0.1% formic acid 5% acetonitrile. 10 µl of sample was loaded onto the Sciex Exion LC, and a 5 min reverse-phase gradient was used.

Buffer A was 0.1% formic acid, and Buffer B was 0.1% formic acid in 100% acetonitrile. The flow was set to 500 µl starting at 5% B, leading to 45% B over 3 minutes before a 95% B wash and equilibration at 5% B. A Phenomenex bioZen 3.6 µm, XB-C8, 50 × 2.1 mm column was used. The flow from the column was passed into the Sciex X500B mass spectrometer, collecting data in positive ion mode. To enable ionisation of the eluate, the source was set to 400 °C, 5500 V with gas at 50 psi. A TOF mass window of 500 to 3000 Da was collected, scanning at 0.5 s. The X500B was calibrated with a positive calibration mix; the error for this experiment was estimated at <1 Da. The resultant TIC was deconvoluted using BioToolKit. The relative abundances were calculated using the peak area for DCAF16 species 1 (24,919.7) in the intact MS spectra as a reference m/z. The relative abundances were then calculated using 24,919.7 + 692 Da, and +1384 Da, respectively, for 1:1 and 1:2 ratios.

Peptide mapping mass spectrometry

For sample preparation, 5 µM DCAF16-DDB1 and 5 µM compound, in the presence or absence of 5 µM BRD9^BD, were incubated at RT for 20 h in 50 mM ammonium bicarbonate. Samples were added to a 2 mL Lo-Bind plate and incubated with 12.5 mM TCEP in 8 M Urea for 45 min at 60 °C while mixing at 500 rpm (0.07 × g) on a shaking plate holder. Each sample was then incubated with 20 mM Iodoacetamide in 50 mM Ammonium Bicarbonate for 45 min at RT while mixing +. For the digest, samples were incubated overnight at 37 °C in 100 µL of 1 µg/mL trypsin. Digests were quenched using 30% Formic Acid (aq) for 5 min prior to injection on the LC-MS system.

The column used was an Acquity UPLC CSH C18 column, 1.7 µ m, 2.1 mm × 100 mm, which was kept at 60 °C during analysis. Buffer A was 0.1% formic acid in water, and Buffer B was 0.1% formic acid in Acetonitrile. The flow was set to 500 µl/min starting at 3% A, leading to 90% A over 57.5 min, before a 97% B wash and equilibration at 3% A. Data was acquired in data-independent acquisition mode (MSe) across the m/z range of 300–1200. A collision energy ramp of 30–60 V was used to generate MS/MS data. Data was processed within the peptide mapping module for Waters UNIFI software. The data within UNIFI was filtered to display only peptides that met the following criteria: (i) the peptide had a binder attached; (ii) the observed mass of the peptide was within ±10 ppm of its theoretical mass, and (iii) the peptide had an MS response of >50,000.

Animal studies and immunoblotting procedures

All animal handling, care, and treatments for this study were conducted according to protocols approved by the Institutional Animal Care and Use Committee (IACUC) of WuXi AppTec, in compliance with the standards of the Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC). Female Balb/c nude mice were inoculated subcutaneously in the right flank with MV4-11 tumour cells (10 × 106) in 0.2 mL PBS mixed with Matrigel (50:50) to promote tumour growth. Once the tumours reached an average volume of 379 mm³ on day 29 post-inoculation, animals were randomised into treatment groups. The compounds were administered orally twice daily, formulated in a solution containing 5% DMSO, 5% Solutol, and 90% HPβCD (15% v/v), with an 8-h interval between doses. Plasma and tumour samples were collected at 10 and 24 h following the first dose.

For immunoblotting, protein lysates were prepared in RIPA buffer and separated using NuPAGE® Novex 4–12% Bis-Tris gels, followed by transfer onto a nitrocellulose membrane with the iBlot®2 Gel Transfer Device. Membranes were blocked for 2 h at room temperature in Odyssey blocking buffer (LI-COR, 927-60001). Primary antibodies were diluted in Odyssey blocking buffer with 0.1% Tween 20 and incubated with the membrane overnight. BRD9 antibody (Cell Signaling, 58906, Clone: E9R2I, Lot no.: 3) was used at 1:1000, with Vinculin (Sigma, SAB4200080, Clone: V284) at 1:10,000 as a loading control. Secondary antibodies, IRDye® 680RD Goat anti-Mouse IgG (LI-COR, 926-68070, Clone: Polyclonal, Lot no.: C91023-06) and IRDye® 800CW Goat anti-Rabbit IgG (LI-COR, 926-32211, Clone: Polyclonal, Lot no.: C91112-09), were diluted 1:10,000 in Odyssey blocking buffer with 0.1% Tween20 and incubated for 1 h at room temperature. Band intensities were quantified using Image Studio software. Statistical significance was calculated using Dunnett’s T3 multiple comparison.

GSH reversibility study

A solution containing 0.5 mM AMPTX-1 and 5.0 mM L-glutathione (4, reduced form) in a 9:1 mixture of PBS buffer and DMSO was allowed to sit at ambient temperature in a sealed well of a 100 µL v-bottom 384-well plate for 2 h (performed in duplicate).

After this time, LC−MS of each well revealed that 38.4% (average of two experiments) of AMPTX-1 ([M + H⁺]⁺ ion with m/z 692.6) had reacted with glutathione (reduced form) to form the corresponding adduct GSH adduct 5 ([M + H⁺]⁺ m/z 999.8 and [M + H⁺]²⁺ m/z 500.5). This amount of adduct formation was formed in ~2 h and was maintained up to at least 24 h in a sealed well.

The wells were then diluted by 10× by removing 5 µL of the original solution and transferring it into a separate well containing 45 µL of a 9:1 mixture of PBS buffer and DMSO. This produced a solution containing 0.05 mM AMPTX-1 and 0.5 mM L-glutathione (reduced form) in a 9:1 mixture of PBS buffer and DMSO, which was allowed to sit at ambient temperature in a sealed well of a 100 µL v-bottom 384-well plate for 4 h (performed in duplicate).

After this time, LC−MS of each well revealed that <2% (average of two experiments) of AMPTX-1 ([M + H⁺]⁺ ion with m/z 692.6) was now a GSH adduct 5 with glutathione ([M + H⁺]⁺ m/z 999.8 and [M + H⁺]²⁺ m/z 500.5).

These data support the dynamic equilibrium and reversible nature of the reaction of AMPTX-1 with glutathione. The amount of adduct 5 formation peaked at ~2 h and was stable up to 24 h, and then after dilution, the amount of adduct decreased and reversed to reform AMPTX-1 (see Fig. 7 and Supplementary Fig. 3).

**Fig. 7: Schematic representation of the GSH reversibility study described in the “Methods” section.**

Plasma protein binding assay

Plasma protein binding of the test compound was determined in vitro by ultracentrifugation using standard assay conditions. In brief, CD-1 mouse plasma was spiked with test compound to a final concentration of 2 µM and pre-incubated for 45 minutes at 37 °C, prior to centrifugation at 627,000 x g for 180 min. Following centrifugation, ultrafiltrate was collected via pipette from the middle portion of the aqueous layer, mixed with control plasma 1-1 (v/v) and quenched with a 5-fold excess of acetonitrile containing an appropriate internal standard. A 0-min control was similarly prepared by spiking CD-1 mouse plasma with test compound at 2 µM followed by immediate 1-1 (v/v) dilution with control ultrafiltrate and quenched with acetonitrile containing internal standard. All processed samples were centrifuged at 6000 x g for 10 minutes, the supernatant was collected and diluted 1-1 (v/v) with water prior to bioanalysis using a specific LC-MS/MS method. Fraction unbound (fup) was determined by the peak area ratio of ultrafiltrate/peak area ratio of 0-min control.

Synthetic chemistry

See supplementary method file. AMPTX-1 qualifies as a molecular glue degrader per the following guidelines. https://www.chemicalprobes.org/info/molecular-glues.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

The mass spectrometry global proteomics dataset and IP-MS dataset used in this study have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository under the accession codes PXD062354 (global proteomics in MV4-11) and PXD062355 (IP-MS in HEK293 cells expressing BRD9-HiBiT), respectively. Unprocessed data used to create all the figures is available in the source data file. Raw data for the intact mass spectrometry and peptide mapping have been deposited to the ProteomeXchange Consortium via the massIVE partner repository under the accession numbers MSV000098521 [https://massive.ucsd.edu/ProteoSAFe/dataset.jsp?task=6ca1c2f30d5e4311bf0b3d60fb0de202] (intact MS) and MSV000098613 [https://massive.ucsd.edu/ProteoSAFe/dataset.jsp?task=f3386d66bb93493781f416e574f16492] (peptide mapping), respectively. Source data are provided with this paper.

References

Békés, M., Langley, D. R. & Crews, C. M. PROTAC targeted protein degraders: the past is prologue. Nat. Rev. Drug Discov. 21, 181–200 (2022).
Article PubMed PubMed Central Google Scholar
Domostegui, A., Nieto-Barrado, L., Perez-Lopez, C. & Mayor-Ruiz, C. Chasing molecular glue degraders: screening approaches. Chem. Soc. Rev. 51, 5498–5517 (2022).
Article CAS PubMed Google Scholar
Dewey, J. A. et al. Molecular glue discovery: current and future approaches. J. Med. Chem. 66, 9278–9296 (2023).
Article CAS PubMed PubMed Central Google Scholar
Diehl, C. J. & Ciulli, A. Discovery of small molecule ligands for the von Hippel-Lindau (VHL) E3 ligase and their use as inhibitors and PROTAC degraders. Chem. Soc. Rev. 51, 8216–8257 (2022).
Article CAS PubMed PubMed Central Google Scholar
Ishida, T. & Ciulli, A. E3 ligase ligands for PROTACs: how they were found and how to discover new ones. SLAS Discov. 26, 484–502 (2021).
Article CAS PubMed Google Scholar
Shah Zaib Saleem, R., Schwalm, M. P. & Knapp, S. Expanding the ligand spaces for E3 ligases for the design of protein degraders. Bioorg. Med. Chem. 105, 117718 (2024).
Article CAS PubMed Google Scholar
Du, G. et al. Exploring the target scope of KEAP1 E3 ligase-based PROTACs. Cell Chem. Biol. 29, 1470–1481.e31 (2022).
Article CAS PubMed PubMed Central Google Scholar
Hickey, C. M. et al. Co-opting the E3 ligase KLHDC2 for targeted protein degradation by small molecules. Nat. Struct. Mol. Biol. 31, 311–322 (2024).
Article CAS PubMed Google Scholar
Schröder, M. et al. DCAF1-based PROTACs with activity against clinically validated targets overcoming intrinsic- and acquired-degrader resistance. Nat. Commun. 15, 275 (2024).
Article ADS PubMed PubMed Central Google Scholar
Li, A. S. M. et al. Discovery of nanomolar DCAF1 small molecule ligands. J. Med. Chem. 66, 5041–5060 (2023).
Article ADS CAS PubMed PubMed Central Google Scholar
Blake, R. A. Abstract 4452: GNE-0011, a novel monovalent BRD4 degrader. Cancer Res. 79, 4452–4452 (2019).
Article Google Scholar
Shergalis, A. G. et al. CRISPR screen reveals BRD2/4 molecular glue-like degrader via recruitment of DCAF16. ACS Chem. Biol. 18, 331–339 (2023).
Article CAS PubMed Google Scholar
Słabicki, M. et al. The CDK inhibitor CR8 acts as a molecular glue degrader that depletes cyclin K. Nature 585, 293–297 (2020).
Article PubMed PubMed Central Google Scholar
Hanley, R. P. et al. Discovery of a potent and selective targeted NSD2 degrader for the reduction of H3K36me2. J. Am. Chem. Soc. 145, 8176–8188 (2023).
Article ADS CAS PubMed PubMed Central Google Scholar
Nie, D. Y. et al. Recruitment of FBXO22 for targeted degradation of NSD2. Nat. Chem. Biol. 20, 1597–1607 (2024).
Article CAS PubMed Google Scholar
Kadoch, C. et al. Proteomic and bioinformatic analysis of mammalian SWI/SNF complexes identifies extensive roles in human malignancy. Nat. Genet. 45, 592–601 (2013).
Article CAS PubMed PubMed Central Google Scholar
Remillard, D. et al. Degradation of the BAF Complex Factor BRD9 by Heterobifunctional Ligands. Angew. Chem. Int. Ed. 56, 5738–5743 (2017).
Article CAS Google Scholar
Zoppi, V. et al. Iterative design and optimization of initially inactive Proteolysis Targeting Chimeras (PROTACs) identify VZ185 as a potent, fast, and selective von Hippel–Lindau (VHL) based dual degrader probe of BRD9 and BRD7. J. Med. Chem. 62, 699–726 (2019).
Article CAS PubMed Google Scholar
Riching, K. M. et al. Translating PROTAC chemical series optimization into functional outcomes underlying BRD7 and BRD9 protein degradation. Curr. Res. Chem. Biol. 1, 100009 (2021).
Article CAS Google Scholar
Wang, C. et al. Discovery of CW-3308 as a potent, selective, and orally efficacious PROTAC degrader of BRD9. J. Med. Chem. 67, 14125–14154 (2024).
Article CAS PubMed Google Scholar
Weisberg, E. et al. BRD9 degraders as chemosensitizers in acute leukemia and multiple myeloma. Blood Cancer J. 12, 1–10 (2022).
Article Google Scholar
Duan, H. et al. Discovery of a highly potent and selective BRD9 PROTAC degrader based on E3 binder investigation for the treatment of hematological tumors. J. Med. Chem. 67, 11326–11353 (2024).
Article CAS PubMed Google Scholar
Sabnis, R. W. Novel compounds for targeted degradation of BRD9 and their use for treating cancer. ACS Med. Chem. Lett. 13, 17–18 (2022).
Article CAS PubMed Google Scholar
Sabnis, R. W. BRD9 bifunctional degraders for treating cancer. ACS Med. Chem. Lett. 12, 1879–1880 (2021).
Article CAS PubMed PubMed Central Google Scholar
Vetma, V., O’Connor, S. & Ciulli, A. Development of PROTAC degrader drugs for cancer. Annu. Rev. Cancer Biol. 9, 119–140 (2025).
Article Google Scholar
Livingston, J. A. et al. A Phase I study of FHD-609, a heterobifunctional degrader of bromodomain-containing protein 9, in patients with advanced synovial sarcoma or SMARCB1-deficient tumors. Clin. Cancer Res. 31, 628–638 (2025).
Article CAS PubMed Google Scholar
Martin, L. J. et al. Structure-Based design of an in vivo active selective BRD9 inhibitor. J. Med. Chem. 59, 4462–4475 (2016).
Article ADS CAS PubMed PubMed Central Google Scholar
Serafimova, I. M. et al. Reversible targeting of noncatalytic cysteines with chemically tuned electrophiles. Nat. Chem. Biol. 8, 471–476 (2012).
Article CAS PubMed PubMed Central Google Scholar
Senkane, K. et al. The proteome-wide potential for reversible covalency at cysteine. Angew. Chem. Int. Ed. 58, 11385–11389 (2019).
Article CAS Google Scholar
Sarott, R. C. et al. Chemical specification of E3 ubiquitin ligase engagement by cysteine-reactive chemistry. J. Am. Chem. Soc. 145, 21937–21944 (2023).
Article ADS CAS PubMed Google Scholar
Zhang, X., Crowley, V. M., Wucherpfennig, T. G., Dix, M. M. & Cravatt, B. F. Electrophilic PROTACs that degrade nuclear proteins by engaging DCAF16. Nat. Chem. Biol. 15, 737–746 (2019).
Article CAS PubMed PubMed Central Google Scholar
Lim, M. et al. DCAF16-based covalent handle for the rational design of monovalent degraders. ACS Cent. Sci. 10, 1318–1331 (2024).
Article CAS PubMed PubMed Central Google Scholar
Li, Y.-D. et al. Template-assisted covalent modification underlies activity of covalent molecular glues. Nat. Chem. Biol. 20, 1640–1649 (2024).
Article CAS PubMed PubMed Central Google Scholar
Hassan, M. M. et al. Exploration of the tunability of BRD4 degradation by DCAF16 trans-labelling covalent glues. Eur. J. Med. Chem. 279, 116904 (2024).
Article CAS PubMed PubMed Central Google Scholar
Hsia, O. et al. Targeted protein degradation via intramolecular bivalent glues. Nature 627, 204–211 (2024).
Article ADS CAS PubMed PubMed Central Google Scholar
Zhang, X., Thielert, M., Li, H. & Cravatt, B. F. SPIN4 is a principal endogenous substrate of the E3 ubiquitin ligase DCAF16. Biochemistry 60, 637–642 (2021).
Article CAS PubMed Google Scholar
Crowe, C. et al. Mechanism of degrader-targeted protein ubiquitinability. Sci. Adv. 10, eado6492 (2024).
Article CAS PubMed PubMed Central Google Scholar
Riching, K. M., Caine, E. A., Urh, M. & Daniels, D. L. The importance of cellular degradation kinetics for understanding mechanisms in targeted protein degradation. Chem. Soc. Rev. 51, 6210–6221 (2022).
Article CAS PubMed Google Scholar
Casement, R., Bond, A., Craigon, C. & Ciulli, A. Mechanistic and structural features of PROTAC ternary complexes. In Targeted Protein Degradation Vol. 2365 (eds Cacace, A. M., Hickey, C. M. & Békés, M.) 79–113 (Springer US, New York, NY, 2021).
Chaudhry, C. Mathematical model for covalent proteolysis targeting chimeras: thermodynamics and kinetics underlying catalytic efficiency. J. Med. Chem. 66, 6239–6250 (2023).
Article CAS PubMed Google Scholar
Gabizon, R. et al. Efficient targeted degradation via reversible and irreversible covalent PROTACs. J. Am. Chem. Soc. 142, 11734–11742 (2020).
Article ADS CAS PubMed PubMed Central Google Scholar
Grimster, N. P. Covalent PROTACs: the best of both worlds?. RSC Med. Chem. 12, 1452–1458 (2021).
Article CAS PubMed PubMed Central Google Scholar
Zecha, J. et al. Peptide level turnover measurements enable the study of proteoform dynamics. Mol. Cell. Proteom. 17, 974–992 (2018).
Article CAS Google Scholar
Bradshaw, J. M. et al. Prolonged and tunable residence time using reversible covalent kinase inhibitors. Nat. Chem. Biol. 11, 525–531 (2015).
Article CAS PubMed PubMed Central Google Scholar
Langrish, C. L. et al. Preclinical efficacy and anti-inflammatory mechanisms of action of the bruton tyrosine kinase inhibitor rilzabrutinib for immune-mediated disease. J. Immunol. 206, 1454–1468 (2021).
Article CAS PubMed PubMed Central Google Scholar
Kong, A. T., Leprevost, F. V., Avtonomov, D. M., Mellacheruvu, D. & Nesvizhskii, A. I. MSFragger: ultrafast and comprehensive peptide identification in mass spectrometry-based proteomics. Nat. Methods 14, 513–520 (2017).
Article CAS PubMed PubMed Central Google Scholar
Plubell, D. L. et al. Extended multiplexing of tandem mass tags (TMT) labeling reveals age and high fat diet specific proteome changes in mouse epididymal adipose tissue. Mol. Cell Proteom. 16, 873–890 (2017).
Article CAS Google Scholar
Ritchie, M. E. et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res. 43, e47 (2015).
Article PubMed PubMed Central Google Scholar

Download references

Acknowledgements

This work was funded by Amphista Therapeutics. No additional funding or grants were given to any author for this work. The authors would like to thank Dr Victoria Lovatt (Amphista Therapeutics Ltd) for helpful discussions: Dr Kate Heesom (Director of the Bristol University Proteomic Facility, UK), Dr Joshodeep Boruwa and Dr Kamalkishor Landge (Syngene International Ltd, India), Dr Samuel Bloor, Camila Torres (Sygnature Discovery, UK) and Marieke Lamers, Jason Pembroke (Domainex, UK) for their assistance with this work.

Author information

Authors and Affiliations

Amphista Therapeutics Ltd, Cambridge, UK
Scott J. Hughes, Wojciech J. Stec, Colin T. R. Davies, David McGarry, Alicia Williams, Marta Carrara, Ivan Del Barco Barrantes, Rebecca Harris, Anna Tasegian, Dominic D. G. Owens, Alexander Fawcett, John Hellicar, Aleksandra Azevedo, Gregor P. Meier, Andrew C. Runcie, Liliana Greger, Martin O’Rourke, Ian Churcher, Martin Pass, Giles A. Brown, Louise K. Modis & Andrea Testa
Centre for Targeted Protein Degradation, School of Life Sciences, University of Dundee, Dundee, UK
Alessio Ciulli
Division of Biological Chemistry and Drug Discovery, School of Life Sciences, James Black Centre, University of Dundee, Dundee, UK
Alessio Ciulli

Authors

Scott J. Hughes
View author publications
Search author on:PubMed Google Scholar
Wojciech J. Stec
View author publications
Search author on:PubMed Google Scholar
Colin T. R. Davies
View author publications
Search author on:PubMed Google Scholar
David McGarry
View author publications
Search author on:PubMed Google Scholar
Alicia Williams
View author publications
Search author on:PubMed Google Scholar
Marta Carrara
View author publications
Search author on:PubMed Google Scholar
Ivan Del Barco Barrantes
View author publications
Search author on:PubMed Google Scholar
Rebecca Harris
View author publications
Search author on:PubMed Google Scholar
Anna Tasegian
View author publications
Search author on:PubMed Google Scholar
Dominic D. G. Owens
View author publications
Search author on:PubMed Google Scholar
Alexander Fawcett
View author publications
Search author on:PubMed Google Scholar
John Hellicar
View author publications
Search author on:PubMed Google Scholar
Aleksandra Azevedo
View author publications
Search author on:PubMed Google Scholar
Gregor P. Meier
View author publications
Search author on:PubMed Google Scholar
Andrew C. Runcie
View author publications
Search author on:PubMed Google Scholar
Liliana Greger
View author publications
Search author on:PubMed Google Scholar
Martin O’Rourke
View author publications
Search author on:PubMed Google Scholar
Ian Churcher
View author publications
Search author on:PubMed Google Scholar
Martin Pass
View author publications
Search author on:PubMed Google Scholar
Giles A. Brown
View author publications
Search author on:PubMed Google Scholar
Alessio Ciulli
View author publications
Search author on:PubMed Google Scholar
Louise K. Modis
View author publications
Search author on:PubMed Google Scholar
Andrea Testa
View author publications
Search author on:PubMed Google Scholar

Contributions

Conception: S.H., W.S., D.M., M.O., A.C., I.C., A.Te. (Andrea Testa). Experimental design and execution: S.H., W.S., M.C., L.G., C.D., A.W., I.D., R.H., A.Ta. (Anna Tasegian), D.O., A.F., J.H., A.A., G.M., A.R. Data analysis, interpretation, and generation of figures: S.H., W.S., D.M., M.C., L.G., M.O., I.C., M.P., G.B., A.C., L.M., A.Te. Authored, edited and substantively revised the manuscript: S.H., W.S., M.C., L.G., M.O., M.P., G.B., L.M., A.Te.

Corresponding authors

Correspondence to Louise K. Modis or Andrea Testa.

Ethics declarations

Competing interests

All co-authors are/were employees and/or shareholders of Amphista Therapeutics Ltd. The Ciulli laboratory at the University of Dundee receives or has received sponsored research support from Almirall, Amgen, Amphista Therapeutics, Boehringer Ingelheim, Eisai, Merck KaaG, Nurix Therapeutics, Ono Pharmaceutical and Tocris-Biotechne.

Peer review

Peer review information

Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Figs. 1-7

Description of Additional Supplementary Files

Supplementary Data 1

Supplementary Data 2

Supplementary Data 3

Reporting Summary

Transparent Peer Review file

Source data

Source Data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Hughes, S.J., Stec, W.J., Davies, C.T.R. et al. Mode of action of a DCAF16-recruiting targeted glue that can selectively degrade BRD9. Nat Commun 16, 8516 (2025). https://doi.org/10.1038/s41467-025-63594-w

Download citation

Received: 31 December 2024
Accepted: 20 August 2025
Published: 27 October 2025
Version of record: 27 October 2025
DOI: https://doi.org/10.1038/s41467-025-63594-w