Introduction

Proteins can self-assemble or hetero-assemble to form various higher-order complexes with diverse biological functions1,2. In response to environmental stimuli or mutations, they often oligomerize or aggregate, exhibiting a range of material properties from solid-like amyloid fibers to dynamic liquid droplets3,4,5,6. These aggregates can be nonfunctional and disease-causing. Historically, protein aggregates were primarily associated with diseases, such as neurodegenerative disorders7,8,9,10. However, growing evidence suggests that they can also play nonpathogenic and functional roles in regulating cell activities3,11. This has led to the intriguing concept of therapeutically exploiting protein oligomerization, whereby intentional modulation of protein aggregation could be leveraged for clinical benefit7. A compelling example supporting this perspective is the recent discovery of BI-3802, a small molecule that degrades B-cell lymphoma 6 (BCL6) protein. BI-3802 promotes the aggregation of BCL6, leading to its ubiquitination and degradation12. Despite several reports highlighting the therapeutic potential of small molecules in controlling protein oligomerization12,13,14,15, most such compounds have been identified serendipitously, often via phenotypic screening. Therefore, developing effective in vitro screening methods for targeted protein aggregation regulation remains a challenge, hindering rational drug design.

Nanopore sensing, a rapidly advancing technology, enables label-free examination of native proteins in a liquid-phase setting, making it ideal for studying protein response to changes in buffer composition or specific environmental stimuli16,17,18,19,20,21. When an external electric field is applied, analytes are driven electrophoretically or electroosmotically through nanoscale channels. Each translocation event, induced by a single analyte, results in a characteristic temporal alteration in the measured ionic current within the nanoscale channel. Analysis of these current signatures can reveal key properties of the analyte, including charge, size, conformational state, and interactions with other biomolecules18,22,23. These capabilities position nanopore sensing as a powerful platform for protein analysis22,24. In particular, quartz nanopores, known for their cost-efficiency and ease of fabrication22,25,26,27,28, overcome the cost barriers associated with conventional nanopore technologies, enabling high-throughput drug screening.

WD repeat domain 5 (WDR5) is an essential scaffolding protein that participates in multiple protein complexes and regulates a broad range of cellular processes29,30,31. Notably, it is a key component of the mixed-lineage leukemia (MLL/SET) complex and promotes Myc recruitment to chromatin at tumor-associated gene loci32,33. Intriguingly, both interaction events between WDR5 and MLL or Myc have been observed at the single-molecule level using engineered tFhuA nanopores34,35. Given WDR5’s central role in solid and hematologic malignancies and its ability to engage multiple partners to form higher-order complexes36,37, it raises the question of whether inducing its controlled oligomerization could represent a effective strategy for functional modulation.

In this study, we utilize quartz nanopores to screen our in-house library of symmetric small molecules for their capacity to selectively induce WDR5 oligomerization. This approach lead to the identification of WZ-1 as a potent and selective inducer of WDR5 aggregation. Mechanistic studies reveal that WZ-1 targets the WDR5 protein, promoting its oligomerization through a catalytic proximity thio-disulfide bond exchange (a process we term Catalytic Proximal Protein Oligomerization (CaPPO)), ultimately hindering the biological functions of WDR5 in cells. Our findings not only establish an efficient nanopore-based platform for discovering protein oligomerization inducers but also introduce CaPPO as a design paradigm for targeting WDR5 and other therapeutically relevant proteins.

Results

Identifying WZ-1 as a WDR5 oligomerization-inducing molecule using nanopore sensing

To facilitate the discovery of protein oligomerization inducers, we selected WDR5 as the model protein for several reasons. First, WDR5 is a well-characterized scaffold protein within the WD40 repeat (WDR) domain family, one of the most abundant protein interaction domains in the human proteome38. Second, several drug-like protein–protein interaction (PPI) inhibitors targeting WDR5 have been identified, providing a reference point to compare the effects of oligomerization inducers with those of conventional PPI inhibitors39. Third, prior studies have identified at least two functional binding pockets associated with WDR5: the WDR5 interaction WIN site40,41,42,43 and the WDR5-binding motif WBM site42,44, which provides an opportunity to explore whether oligomerization induction can disrupt WDR5 function.

To reliably detect WDR5 proteins in various oligomerization states, quartz nanopipettes were fabricated into nanopores with an average diameter of 22.2 ± 5.0 nm (n = 5), as confirmed by scanning electron microscopy (SEM) (Fig. S1). Linear I–V curves were observed in the protein buffer, with the nanopore conductance measured at 35.70 ± 4.87 nS (Fig. S2a). A thermal shift assay confirmed that WDR5 maintained its binding affinity for known inhibitors under high-salt conditions (Fig. S3), ruling out potential salt-induced alterations in protein functionality.

Our screening strategy leverages the sensitivity of quartz nanopores to different oligomeric states of proteins22. In all nanopore sensing experiments, the reference/ground Ag/AgCl electrode and the patch/working Ag/AgCl electrode were separately placed in the external reservoir (cis) and the pipette (trans), both filled with the same protein buffer. Analytes, including proteins and molecules, were introduced into the external reservoir (Fig. 1a). The consistent translocation characteristics of pure WDR5 across multiple nanopores, fabricated with the same parameters, were observed at an applied voltage of 250 mV, with a mean peak current of 76.33 ± 28.8 pA (Figs. 1b and S2b). This value represents the ion-excluded volume caused by the translocation of monomeric WDR5. Nonuniform translocation events of monomeric WDR5, along with atypical translocation events exhibiting significantly higher peak currents and longer dwell times, were ascribed to either impurities compromising protein purity or transient oligomerization, unfolding and surface adsorption during the translocation process22,45,46.

Fig. 1: Nanopore detection of monomeric and oligomeric WDR5 proteins.
Fig. 1: Nanopore detection of monomeric and oligomeric WDR5 proteins.The alternative text for this image may have been generated using AI.
Full size image

a Schematic illustration of the screening workflow for identifying WDR5 oligomerization inducers using nanopore detection. Briefly, 436 in-house compounds were grouped and incubated with WDR5. Nanopore sensing was then employed to record the translocation events of WDR5 from each mixture. After acquiring a sufficient number of events, the peak current amplitude and dwell time for each event were extracted and analyzed statistically. Compound groups showing significant increases in both parameters were either subdivided for subsequent screening rounds or identified as WDR5 oligomerization inducers in the final screening stage. Top right a: representative SEM image showing the diameter of the nanopore used in the screening experiments. b Representative ionic current traces and typical translocation events (filtered at 10 kHz for visualization) for monomeric and oligomeric WDR5 in protein buffer (1× PBS, 1 M KCl, pH = 8.6–9.4). Data were acquired at a sampling rate of 250 kHz. The diagram of translocation events are created in BioRender. Lab, C. (2026) https://BioRender.com/po6o2vy.

Next, we employed our established nanopore sensing platform to screen for molecules capable of inducing WDR5 oligomerization (Fig. 1a). Bivalent compounds, composed of two fragments that each bind the protein of interest (POI) and connected by a suitable linker, can simultaneously engage two POI molecules, thereby promoting dimerization47. Inspired by this principle, we screened an in-house library of 436 small-molecule compounds, most of which possess bivalent-like structures. To improve screening efficiency, we categorized the 436 compounds into nine groups (46–50 compounds per group, designated 1st-#1 to 1st-#9) based on structural similarity, and each group was evaluated using nanopore sensing (Fig. 2a, b). The mean peak current under identical conditions was determined using Gaussian fitting, and the rate of change was calculated by comparing the values in the mean before and after the introduction of the candidate compounds (Fig. S4). The rate of change served as the key metric for identifying groups containing potential oligomerization inducers. In the first screening round, group 1st-#9 produced the most pronounced effect, showing a ~45.1% increase in average peak current (Fig. 2a–d). These 46 compounds were then divided into five subgroups for a second-round screen, in which subgroup 2nd-#5 produced an even larger increase of 60.5% in mean peak current. Finally, the 10 candidate compounds from subgroup 2nd-#5 were tested individually, leading to the identification of WZ-1 (Fig. 2e) as the most potent inducer of WDR5 oligomerization.

Fig. 2: Statistical analysis of nanopore events for WDR5.
Fig. 2: Statistical analysis of nanopore events for WDR5.The alternative text for this image may have been generated using AI.
Full size image

a Representative ionic current traces depicting the translocation of WDR5 in the presence of mixture 1st-#9, mixture 2nd-#5, and WZ‑1 during screening rounds 1, 2, and 3, respectively. b Variations in the Gaussian fit mean values of WDR5 peak current amplitudes induced by incubation with all grouped compounds across the three screening rounds. c Violin plot showing the distribution of peak currents for WDR5 translocation before and after incubation with mixture 1st-#9(blue, n = 168; purple, n =  197), mixture 2nd-#5(blue, n = 230; purple, n = 143), and WZ‑1 (blue, n = 285; purple, n = 317). P values were calculated using two-sided, paired t-tests (without correction for multiple comparisons). Box plots show the mean (center line), the first and third quartiles (box bounds), and the whiskers extending to the most extreme data points within 1.5× interquartile range (IQR) from the quartiles. d Scatter plot of peak current amplitude versus dwell time for WDR5 in the absence and presence of different compound mixtures. e Molecular structure of WZ‑1. f Time-dependent ionic current traces monitoring WDR5 aggregation. WDR5 (500 nM) was incubated with 5 µM WZ‑1, and ionic currents were recorded for 5 min every 30 min at 250 mV. Traces acquired before incubation and at 60, 120, and 180 min post-incubation are shown. g Scatter plots of peak current amplitude versus dwell time for WDR5 at the indicated time points. h Column scatter plot showing peak current amplitude as a function of incubation time, illustrating the real-time aggregation kinetics of WDR5 in the presence of 5 µM WZ‑1. Source data are provided as a Source data file.

To exclude potential confounding factors, such as the change in translocation events, being solely caused by the binding of molecules to protein sites, we conducted real-time monitoring of the WDR5 translocation in the presence of high-affinity inhibitors OICR-942939,48 and SADS-149, which target WIN and WBM sites, respectively. In this experiment, we recorded translocation events in the presence of these two compounds and compared them to the behavior of the WDR5 monomer. As expected, neither OICR-9429 nor SADS-1 produced any discernible change in WDR5 translocation characteristics (Fig. S5a). In addition, control experiments confirmed that the observed signals were not due to micelles or precipitates formed by high concentrations of WZ-1 (Fig. S6). Together, these observations confirm that the changes in WDR5 translocation behavior are indeed induced by WZ-1 through its promotion of WDR5 oligomerization.

Next, we monitored the real-time oligomerization of WDR5 to assess its proteodynamics. To enable more stable and long-term detection of oligomerized WDR5 translocation events, we utilized a larger-diameter quartz nanopore. These pores generated smaller protein peak signals, reducing the risk of clogging and allowing efficient translocation of oligomeric species. Nanopore sensing was conducted over three hours at a WZ-1 concentration of 5 μM, with translocation events recorded in 5-min intervals every 30 min and visualized as scatter plots (Fig. 2f, g). Following the addition of WZ-1, the mean peak current gradually increased, reaching approximately twice its initial value. The aggregation behavior of WDR5 displayed a time-dependent pattern but did not follow a distinct sigmoidal trajectory (Fig. 2h). After the peak current plateauing at the 10× WZ-1 concentration, we increased the concentration of WZ-1 to 100×, which resulted in a broader peak-current distribution and a concurrent decrease in capture rate (Fig. S7), consistent with a concentration-dependent induction of WDR5 aggregation by WZ-1. Notably, this study represents the successful use of nanopore sensing for in vitro screening of protein oligomerization inducers, highlighting its potential as a platform for drug discovery.

The N-terminal domain and Cys248 residue of WDR5 underlie the specificity of WZ-1

To determine whether WZ-1 is indeed a WDR5-selective oligomerization-inducing small molecule, we conducted classical biochemical experiments. First, we confirmed the ability of WZ-1 to induce recombinant WDR5 protein oligomerization in vitro using immunoblotting. Interestingly, oligomer formation was observed only when the reducing agent dithiothreitol (DTT) was omitted from the loading buffer (Fig. 3a). Inspired by this clue, we hypothesized that nonreducing conditions are required for WZ-1-induced WDR5 oligomerization. To test this, we introduced the reducing agents DTT50 and glutathione (GSH)51 in our in vitro assays. As shown in Fig. 3b–d, both immunoblotting and nano sensing assays demonstrated that these reducing agents completely abolished the WZ-1-induced WDR5 oligomerization.

Fig. 3: Discovery of a thiol-dependent specific oligomerization inducer for WDR5.
Fig. 3: Discovery of a thiol-dependent specific oligomerization inducer for WDR5.The alternative text for this image may have been generated using AI.
Full size image

a Immunoblot analysis of 6 µM WDR5 treated with 100 µM WZ‑1 under reducing and nonreducing sample loading conditions. b Immunoblot analysis of 6 µM WDR5 treated with 100 µM WZ‑1 in the presence of 50 µM DTT or 50 µM GSH. c Representative ionic current traces of 500 nM WDR5 treated with 5 µM WZ‑1 in the presence of 1 mM GSH. d Column scatter plot of peak current amplitudes corresponding to the data in panel. P values were calculated using one-way ANOVA (without correction for multiple comparisons). c Box plots show the mean (center line), the first and third quartiles (box bounds), and the whiskers extending to the most extreme data points within 1.5× interquartile range (IQR) from the quartiles. (DMSO, n = 202; WZ-1, n =  318; WZ-1 + GSH, n = 132). e The mechanism of IAM blocking and immunoblot analysis of 6 µM WDR5 treated with 100 µM WZ‑1 together with either 50 µM IAM or DMSO (vehicle control). Diagram of IAM blocking is created in BioRender. Lab, C. (2026) https://BioRender.com/79yphxv. f Immunoblot analysis of 6 µM recombinant single-cysteine-mutant WDR5 proteins (C163S, C195S, C205S, C248S, C261S, C309S, and C334S) treated with 100 µM WZ‑1. g SEC-MALS analysis of 15 µM wild-type tag-free WDR5 and variants (C248S, 31–334, 31–334_C248S) treated with 200 µM WZ‑1. Calculated molecular weight of WDR5: 41.2 kDa. Source data are provided as a Source data file.

To explore the selectivity of WZ-1 for WDR5, we obtained all 10 commercially available WDR proteins (WDR16, WDR21B, WDR21C, WDR23, WDR24, WDR30, WDR31, WDR40A, WDR77, and WDR86) to assess whether WZ-1 could induce their oligomerization. As shown in Fig. S5b, WZ-1 induced oligomerization exclusively in WDR5, with no detectable oligomerization in any of the other tested WDR proteins. These findings indicate that WZ-1 selectively induces aggregation of WDR5.

Given that WZ-1 contains a disulfide motif, we hypothesized that its disulfide bond might function as a latent covalent warhead, reacting with cysteine residues in WDR5. To test this, we pretreated WDR5 with iodoacetamide (IAM) to block free cysteine residues before incubation with WZ-152 (Fig. 3e). As shown in Fig. 3e, cysteine blocking completely abolished WDR5 oligomerization, supporting the involvement of cysteine residues in the WZ-1-mediated process.

To determine which cysteine residue is involved in WZ-1-triggered oligomerization, we individually mutated each of WDR5’s ten cysteines to serine (CS mutation). Seven single-mutant recombinant WDR5 proteins (C163S, C195S, C205S, C248S, C261S, C309S, and C334S) were successfully purified. Following WZ-1 treatment, most mutants still exhibited oligomer formation. However, several mutations, particularly C205S, modestly reduced oligomerization, whereas the C248S mutant—despite displaying minimal background oligomerization—completely lost the ability to form WZ-1-induced oligomers (Fig. 3f). This result suggests that cys248 is the key residue involved in WZ-1-induced WDR5 oligomerization, and other cysteine residues facilitate the progress of oligomerization. Similar results were observed using the SEC-MALS assay. As shown in Fig. 3g, wildtype WDR5 protein appeared mostly as a monomer in solution, and WZ-1 resulted in the formation of dimers, while the C248S mutation remained as monomers.

It has been observed that the N-terminal residues of WDR5 can bind to the WIN site of a neighboring molecule in the crystal lattices38,53, suggesting a potential role in WDR5 oligomerization. Consistently, AlphaFold structure prediction of a WDR5 dimer produced a symmetric homodimer stabilized by WIN-site interactions (Fig. 4a). Notably, the Cys248 residues of each monomer are positioned at the dimer interface, approximately 7.6 Å apart (Fig. 4b), providing a structural rationale for why Cys248 is the preferred site for disulfide bond formation. Supporting this model, SEM-MALS analysis, immunoblotting analysis and Cryo-EM analysis demonstrate that the N-terminal tail markedly enhances WZ-1-induced aggregation (Fig. 3g, Figs. S5d and S8), consistent with the transient interactions mediated by WIN sites and the N-terminal tail bring cys248 or other cysteine residues in close proximity for the WZ-1 catalyzed reaction. Cryo-EM analysis of WDR5 treated with WZ-1 further demonstrated the coexistence of monomers, dimers, and higher-order oligomers (Fig. 4c).

Fig. 4: Structural and morphological characterization of WDR5 oligomerization.
Fig. 4: Structural and morphological characterization of WDR5 oligomerization.The alternative text for this image may have been generated using AI.
Full size image

a Crystal structure of monomeric WDR5 (light blue; PDB ID: 2H68) showing the positions of cysteine residues. b Predicted dimeric WDR5 structure (pink and teal) generated by AlphaFold, with a spatial separation of 7.6 Å between the two cys248 residues. c Representative 2D classification images illustrating distinct tag-free WDR5 aggregation states induced by WZ‑1. d Schematic diagram of the CaPPO process, showing the formation of dimeric and oligomeric WDR5.

WZ-1 induces WDR5 oligomerization through a disulfide bond switch mechanism

Having identified the N-terminal tail and Cys248 as key elements in regulating WDR5 dimerization, we next investigated the binding mechanism between WZ-1 and WDR5. We focused on whether the 2-aminothiazole core and the disulfide motif were essential for WZ-1 to function as an oligomerization inducer54. To this end, we conducted a structure–activity relationship (SAR) study WZ-1: WZ-2, WZ-3, and WZ-4, in which the 2-aminothiazole core was simplified, and WZ-5 and WZ-6, which lacked the disulfide motif (Fig. 5a). We then evaluated their binding affinity to WDR5 alongside their ability to induce oligomerization. As shown in Figs. 5b and S9, only symmetric compounds containing the intact 2-aminothiazole core (WZ-1, WZ-4, and WZ-6) exhibited low-micromolar dissociation constants (Kd) compared with WZ-2, WZ-3, and WZ-5, underscoring the importance of the 2-aminothiazole core in WDR5 binding. Interestingly, replacing the disulfide motif in WZ-1 with a hydrazine group (WZ-6) completely abolished oligomerization, even though the compound retained both symmetry and the 2-aminothiazole core. These findings outline a strategy for triggering protein oligomerization that differs from conventional bivalent compound design (Figs. 5c, d, and S5c). Our preliminary structure-activity relationship analysis suggests that the 2-aminothiazole core mediates stable binding to WDR5, while the disulfide motif acts as a scout element, initiating WDR5 oligomerization.

Fig. 5: Specific chemical catalytic mechanism of WZ‑1.
Fig. 5: Specific chemical catalytic mechanism of WZ‑1.The alternative text for this image may have been generated using AI.
Full size image

a The preliminary SAR study of WZ-1. b Affinity dissociation constants (Kd) of WZ‑1 derivatives measured by MicroScale thermophoresis (MST) assay. c, d Coomassie blue staining analysis of WDR5 following treatment with WZ‑1 and its derivatives. e HPLC‑HRMS analysis of 1 mM WZ‑1, 100 mM NAC, or a mixture containing both 1 mM WZ‑1 and 100 mM NAC. f Schematic diagram of the thiol-disulfide exchange reaction induced by WZ‑1, resulting in the formation of dimeric NAC.

Unlike many other covalent warheads, a disulfide bond can be highly dynamic under physiological conditions due to a series of reversible reactions collectively known as thio-disulfide exchanges55. We therefore hypothesized that the disulfide bond in WZ-1 could exchange with Cys248 of WDR5, initiating the formation of a new intermolecular disulfide bond between WDR5 molecules and thereby promoting oligomerization. We further speculated that WZ-1 might function catalytically-being released after the exchange to trigger subsequent thio-disulfide reactions. To test this hypothesis, we carried out a model reaction between WZ-1 and N-acetylcysteine (NAC). HPLC-HRMS analysis revealed that the thiol group of NAC initially reacted with WZ-1 via a thio-disulfide exchange, generating a disulfide intermediate (A2) and releasing intermediate A1. Intermediate A1 was subsequently trapped by ethyl bromide to form A3. Meanwhile, A2 could undergo a second reaction with NAC, yielding the corresponding NAC dimer (dNAC) (Figs. 5e, f, and S10).

Based on these results, and considering that Cys248 is positioned next to the WBM site while the 2-aminothiazole core of WZ-1 is critical for binding and aggregation, we propose a two-step mechanism for WZ-1-induced WDR5 oligomerization (Fig. 4d). In the first step, the 2-aminothiazole core of WZ-1 binds the WBM site of WDR5, positioning its disulfide bond for reaction with the surface-exposed Cys248. In the second step, the N-terminal tail promotes further proximity between WDR5 molecules, enabling protein-to-protein disulfide exchange between the WZ-1-Cys248 adduct and cysteine residues on adjacent WDR5 molecules, thereby driving oligomer growth. Whether the cysteine residue engaged in this exchange is Cys248 itself or another surface cysteine likely determines whether the reaction terminates or continues, thus controlling the extent of oligomerization.

WZ-1 inhibits cancer cell proliferation by disrupting the interaction between WDR5 and its partner proteins

Next, we investigated the cellular effects of WZ-1. Using the sulforhodamine B (SRB) assay, we assessed the anti-cancer activity of WZ-1 against four colon cancer cell lines: CT26, RKO, DLD1, and SW620. As shown in Fig. 6a, WZ-1 effectively inhibited proliferation in all four cell lines, with IC50 values ranging from 1.1 to 6.2 μM, proving more potent than the reported WDR5 inhibitor SADS-1 (IC50 > 30 μM). In contrast, WZ-1 did not substantially inhibit proliferation in normal colon cell lines NCM460 and immortalized human colon fibroblasts, indicating a degree of selectivity toward cancer cells (Fig. S11a). Although we didn’t detect endogenous WDR5 oligomerization in CT26 and RKO cells, we observed dimerization of exogenously expressed WDR5 only after treatment with WZ-1 (Fig. 6b), suggesting that WZ-1 can also induce WDR5 oligomerization in cells.

Fig. 6: Functional inhibition of WDR5 triggered by protein oligomerization inducing strategy.
Fig. 6: Functional inhibition of WDR5 triggered by protein oligomerization inducing strategy.The alternative text for this image may have been generated using AI.
Full size image

a CT26, RKO, DLD1, and SW620 cells were treated with WZ‑1 or SADS‑1 at varying concentrations for 72 h, and cell viability was assessed by SRB assay. b CT26 and RKO cells were cultured in Opti‑MEM supplemented with 100 µM BSO for 8 h, followed by immunoblot analysis of exogenous WDR5 after treatment with 10 µM WZ‑1. c, d Schematic diagram of GSEA analysis workflow and bar plot of the top 30 enriched GSEA terms for RKO cells treated with WZ‑1. The software GSEA (v4.1.0) and MSigDB were used to perform gene set enrichment analysis. The gene expression matrix and rank genes were input by Signal2Noisenormalization method. Enrichment scores and P value was calculated in default parameters. |NES| > 1, NOM P < 0.05 were considered to be different in two groups. e GSEA enrichment plots for WBM site— and WIN site-related pathways in RKO cells (WIN site: E2F targets, G2M checkpoint; WBM site: Myc targets v1, Myc targets v2). f, g 293T cells overexpressing tagged target plasmids were cultured in Opti‑MEM supplemented with 100 µM BSO for 8 h, then treated with varying concentrations of WZ‑1 (0, 5, and 10 µM) for 3 h. Co‑immunoprecipitation assays using WDR5 as bait protein demonstrated its interaction with MLL1 (f) and with c‑Myc (g). Source data are provided as a Source data file.

To understand the link between WZ-1-induced WDR5 oligomerization and cell proliferation inhibition, we performed RNA-Seq on CT26 and RKO cells after WZ-1 treatment and conducted GSEA analysis on the significantly changed genes (Fig. 6c). Interestingly, we observed a significant negative enrichment of E2F and G2M checkpoint signaling pathways in WZ-1 treated cells compared to controls (Figs. 6d, e, and S11d, e). Consistent with these findings, we also discovered that WZ-1 caused G0/G1 phase cell cycle arrest (Fig. S11b). These results closely parallel previous findings that WDR5 WIN-site inhibitors cause a decrease in gene expression associated with the E2F and G2M checkpoint signaling pathways42,56.

Notably, we noticed that WZ-1 treatment also inhibits the expression of Myc target genes (Fig. 6d, e). Given that WDR5 directly interacts with c-Myc via its WBM site33, and that Cys248 is positioned adjacent to this site (Fig. S11c), these results further suggest that WZ-1-induced WDR5 oligomerization disrupts its full functionality in both the WIN and WBM interactions. To test this hypothesis, we examined whether WZ-1-induced WDR5 oligomerization affects its interaction with MLL1 or c-Myc. As shown in Fig. 6f, g, WZ-1 treatment markedly disrupted WDR5 binding to both MLL1 and c-Myc. Collectively, these results suggest that WZ-1 inhibits cancer cell proliferation by impairing both WIN- and WBM-mediated functions of WDR5, thereby compromising its role in these critical oncogenic pathways.

Discussion

In this study, we used nanopore sensing technology to screen for WDR5 oligomerization inducers. Through this approach, we efficiently identified WZ-1 as a specific inducer of WDR5 oligomerization after only three rounds of screening from a library of 436 compounds. The discovery of WZ-1 underscores the success of nanopore-based screening. Nanopore sensing technology, characterized by single-molecule sensitivity, is particularly effective for screening oligomerization inducers due to several key factors: (i) It allows for throughput screening at relatively low sample concentrations (100–500 nM), (ii) rapid, label-free detection within minutes minimizes errors arising from protein modification and denaturation, (iii) the cost-effective and accessible quartz nanopores enhance the applicability and generalizability of this method, (iv) the adaptability of this technique to diverse proteins eliminates the need for costly antibodies etc17,18,24,27,28,57. Thus, nanopore-based screening offers great promise in identifying lead compounds and advancing protein oligomerization inducers.

Of interest is the fact that, upon the administration of WZ-1 alone, the peak currents and capture rate dramatically diminish after a rapid surge of peak currents (Fig. S12). This unusual phenomenon has also been seen in nanopore detection of α-Syn oligomerization24. To capture this transient event, samples incubated for less than 30 min were used for recording and statistical analysis, yielding a rate of change of 74.3%, slightly higher than the positive controls from the second screening round. This effect was consistently reproduced across independent experiments, confirming its robustness and nonrandom nature. Notably, increasing the WZ-1 concentration 50-fold markedly amplified the effect. This phenomenon is not influenced by alterations in charge preference following aggregation (Table S1). SEM imaging of nanopores following WDR5 aggregation detection (Fig. S13a) revealed irregular agglomerates near the pore surface, which SEM-EDS analysis confirmed to be protein aggregates (Fig. S13b). These findings suggest that WDR5 molecules, concentrated near the nanopore by electrophoretic forces, undergo hyper-polymerization in the presence of WZ-1, thereby hindering the translocation of unbound WDR5. Given that these aggregates are stabilized by disulfide bonds, it is reasoned that the nanopore test is unlikely to produce false oligomerization signals caused by agglomerates dissociation. The absence of over high blocking peaks supports this speculation. Collectively, these results highlight the strong influence of WZ-1 binding on WDR5’s oligomeric state and structural dynamics.

The weak WDR5 dimers observed as background in our Western blot experiments are likely due to nonspecific covalent bond formation under nonreducing conditions. During incubation, dynamic oligomers mediated by N-terminal interactions bring two WDR5 molecules into close proximity, facilitating unintended covalent linkage. This dimer signal is more apparent in Western blot analysis than in SEC-MALS. In Western blot, denaturation of WDR5 into peptides exposes cysteine residues, which can further promote aggregation. In contrast, in SEC-MALS, these dynamic oligomers dissociate during chromatographic separation, as the proteins partition into the pores of the column matrix and elute in the mobile phase. Such behavior is typical of weak protein complexes, which often appear as separate peaks during SEC.

WZ-1-induced oligomerization of WDR5 appears to be driven by two synergistic mechanisms: the binding of its molecular core to WDR5, followed by thio-disulfide exchange, and WDR5’s inherent tendency to dimerize, which may offset the comparatively low micromolar affinity of the molecular core. Another important point is that thio-disulfide exchange-based protein oligomerization might require an appropriate cysteine residue near the target protein’s binding site. This CaPPO strategy of protein intervention relies solely on chemical catalysis and shows high specificity, eliminating the need for endogenous cellular processing machinery and thus reducing the risk of decreased drug efficacy due to intercellular variability58,59,60.

It is important to acknowledge that this thio-disulfide exchange based protein oligomerization inducer may face several challenges: (i) so far, it has only been tested on WDR5, and further exploration is needed to determine its applicability to other proteins, (ii) the disulfide bonds induced by WZ-1 are susceptible to cleavage under strong reducing conditions, so we need to develop strategies to ensure stable linkages that remain unaffected by intracellular environments, (iii) challenges persist in translating oligomerization inducers into viable drugs, primarily stemming from unforeseen oligomerization scenarios, such as excessive oligomerization and off-target proteins interactions. Considering the principle of disulfide bond exchange, focusing on the reaction kinetics of cysteine residues may be more effective for improving protein specificity.

Our results suggest that the anti-tumor effect of WZ-1 is due to the induction of WDR5 aggregation, which blocks the interaction between WDR5 and c-Myc, as well as the transcription of downstream genes (Fig. 7). The CaPPO mechanism by which the degradation of oligomerized WDR5 occurs remains elusive, requiring further experimental investigation. The aggregation of WDR5 seems to hinder its interaction with c-Myc, possibly because the Cys248 site is close to the WBM binding domain. Interestingly, our data indicate that WZ-1 also affects the interaction between WDR5 and MLL1. This suggests that this protein intervention approach, based on inducing aggregation, has the potential for multifunctional modulation. Another point worth mentioning is that the aggregated protein exhibits similarities to prion-like proteins, such as those involved in neurodegeneration, which is known for its strong intercellular transfer and transmission capacities9,61,62,63,64. This suggests that if compound-induced oligomerization can mimic this process, the protein oligomerization inducer might enhance its therapeutic potential.

Fig. 7: The mechanism schematic of functional inhibition of WDR5 by WZ-1.
Fig. 7: The mechanism schematic of functional inhibition of WDR5 by WZ-1.The alternative text for this image may have been generated using AI.
Full size image

Created in BioRender. Lab, C. (2026) https://BioRender.com/s6wmfas.

Methods

Reagents

Primary antibodies against WDR5 (db4107), MLL1 (db12766), c-Myc (db1667), HA (db2603), FLAG (db1435), and GAPDH (db106) were purchased from Diagnostic Biosystems (Hangzhou, China). The antibody against ASH2L (#5019) was obtained from Cell Signaling Technology (Boston, MA, USA).

OICR-9429 (HY-16993), DTT (HY-15917), GSH (HY-D0187), IAM (HY-34477), and buthionine sulfoximine (BSO, HY-106376) were obtained from MedChemExpress (MCE, New Jersey, USA). N-acetylcysteine (NAC, A7250) was purchased from Sigma-Aldrich (St. Louis, MO, USA). The compound SADS-1 was synthesized in Dr. Chen’s laboratory.

Synthesis of compounds

All solvents were purified according to standard methods prior to use. NMR spectra were recorded for 1H NMR at 500 MHz and for 13C NMR at 125 MHz. For 1H NMR, tetramethylsilane (TMS) (δ = 0), DMSO-d6 (δ = 2.500) or CDCl3 (δ = 7.260) served as internal standard, and data were reported as follows: chemical shift, integration, multiplicity (s = singlet, d = doublet, t = triplet, q = quartet, m = multiplet), and coupling constant (s) in Hz. For 13C NMR, TMS (δ = 0), DMSO-d6 (δ = 39.52) or CDCl3 (δ = 77.16) was used as internal standard and spectra were obtained with complete proton decoupling. LC-MS of all final products were confirmed on an Agilent 1290 HPLC-6224 time of fight mass spectrometer using PhenomenexLuna 5 μ C18, 100 Å, 150 × 4.60 mm 5-micron column at a flow rate of 0.5 mL/min using liner gradients buffer B in A (B: CH3OH containing 0.1% formic acid, A: H2O containing 0.1% formic acid). Mobile phase B was increased linearly from 5 to 95% over 7 min and 95% over the next 2 min, after which the column was equilibrated to 5% for 1 min. The optical rotation was conducted at 25 °C using JASCO P-1030/1010 polarimeter, with a wavelength of 589 nm. The WZ-2 was commercially obtained.

5,5′-disulfanediylbis(4-(4-chlorophenyl) thiazol-2-amine) (WZ-1)

4-(4-chlorophenyl)-5-thiocyanatothiazol-2-amine (200 mg, 0.75 mmol, 1.0 equiv) and glutathione (300 mg, 0.98 mmol, 1.3 equiv) were dissolved in a mixture of DMF, MeCN and distilled water (V/V/V = 4/4/1). Then the reaction mixture was stirred at room temperature. Upon the completion of the reaction via TLC (PET: EA, 1/1), the mixture was concentrated in vacuo and diluted with H2O (3 mL). The resulting precipitate was filtered and purified via silica gel chromatography to give the yellow compound. (227 mg, 63%) 1H NMR (500 MHz, DMSO-d6) δ 7.63 (s, 4H), 7.59 (d, J = 8.5 Hz, 4H), 7.34 (d, J = 8.5 Hz, 4H). 13C NMR (125 MHz, DMSO-d6) δ 171.0, 157.0, 132.8, 132.3, 130.3, 127.7, 107.4. HRMS (ESI): m/z calcd for C18H13Cl2N4S4 [M + H]+: 482.9395, found: 482.9394.

2,2′-disulfanediyldianiline (WZ-3)

2-mercaptoaniline (250 mg) was added to a 25 mL round-bottom flask, followed by the addition of 5 mL of anhydrous ethanol to dissolve the compound. Subsequently, 508 mg of iodine was introduced, and the mixture was stirred at room temperature. After an overnight reaction, the solution turned colorless, and TLC indicated the complete disappearance of the starting material. Then, 100 mL of EA was added, and the mixture was washed three times with 50 mL of saturated saline solution. The organic phase was collected and dried over anhydrous sodium sulfate. The residue was purified via silica gel chromatography to yield a white solid (190 mg, 76%). 1H NMR (500 MHz, DMSO-d6) δ 7.08 (ddd, J = 8.0, 7.1, 1.5 Hz, 2H), 7.00 (dd, J = 7.5, 1.5 Hz, 2H), 6.73 (dd, J = 8.0, 1.5 Hz, 2H), 6.43 (td, J = 7.5, 1.5 Hz, 2H), 5.45 (s, 4H). HRMS (ESI) m/z calcd for C12H13N2S2 [M + H]+: 249.0515. Found: 249.0510.

5,5′-disulfanediylbis(4-methylthiazol-2-amine) (WZ-4)

4-methyl-5-thiocyanatothiazol-2-amine (68 mg, 0.4 mmol, 1.0 equiv) and glutathione (152 mg, 0.52 mmol, 1.3 equiv) were dissolved in a mixture of DMF, MeCN and distilled water (V/V/V = 4/4/1). Then the reaction mixture was stirred at room temperature. Upon the completion of the reaction via TLC (PET: EA, 1/1), the mixture was concentrated in vacuo and diluted with H2O (3 mL). The resulting precipitate was filtered and purified via silica gel chromatography to give the white compound (69 mg, 60%). 1H NMR (500 MHz, DMSO-d6) δ 7.70 (s, 4H), 2.23 (s, 6H). HRMS (ESI) m/z calcd for C8H11N4S4 [M + H]+: 290.9861. Found: 290.9862.

4-(4-chlorophenyl)-5-(methylthio) thiazol-2-amine (WZ-5)

To a solution of 4′-chloroacetophenone (6.1 g, 39.5 mmol) in a mixture of 48% HBr (20 mL) and dimethylsulfoxide (20 mL) in a sealed flask. This mixture was heated at 60 °C for 12 h and then cooled. After the addition of isopropanol (15 mL) and ethyl ether (15 mL), the solution was stirred for another 0.5 h and allowed to stand overnight in the ice box. The precipitate was filtered and washed with ethyl ether to afford the desired sulfonium salt as white crystals. To a solution of the crude product (3.7 g, 10 mmol) in ethyl alcohol (20 mL), thiourea (912 mg, 12 mmol, 1.2 equiv) was added at room temperature and then refluxed for 2 h. The precipitated product was collected by filtration, washed with water (3 × 10 mL) and dichloromethane (3 × 10 mL). The resulting residue was purified via silica gel chromatography (PET/EA = 10:1) to afford a yellow solid (1.8 g, 75% yield). 1H NMR (500 MHz, DMSO-d6) δ 7.91–7.88 (m, 2H), 7.48–7.44 (m, 2H), 7.33 (s, 2H), 2.32 (s, 3H). 13C NMR (125 MHz, DMSO-d6) δ 168.1, 150.8, 133.3, 132.2, 130.0, 128.0, 111.0, 21.7. HRMS (ESI) m/z: [M + H]+ calcd for C10H10ClN2S2:256.9968; found: 256.9964.

5,5′-(hydrazine-1,2-diyl)bis(4-(4-chlorophenyl)thiazol-2-amine) (WZ-6)

Two hundred eighty-seven milligrams of 2-amino-4-(4-chlorophenyl)-5-bromothiazole was added to a 50 mL sealed tube, followed by the addition of 5 mL acetonitrile and 5 mL hydrazine hydrate. The mixture was heated to 80 °C and allowed to react overnight. Afterward, 50 mL water was added, and the mixture was extracted three times with 50 mL ethyl acetate. The ethyl acetate layers were combined, dried over anhydrous sodium sulfate, and concentrated. The residue was purified via silica gel chromatography to yield a white solid (85 mg, 30%). 1H NMR (500 MHz, DMSO-d6) δ 7.87–7.86 (m, 2H), 7.84–7.80 (m, 2H), 7.47–7.43 (m, 2H), 7.28–7.23 (m, 2H), 7.17 (s, 2H), 5.14 (s, 2H). HRMS (ESI) m/z calcd for C18H15Cl2N6S2 [M + H]+: 449.0171. Found: 449.0169.

Expression and purification of WDR5

WDR5 and its corresponding mutants were cloned into the pET-28a vector (#69864-3, Addgene) to produce an N-terminal 6×His-tagged fusion protein and expressed in Escherichia coli Rosetta 2 (DE3) cells (the primers are listed in Table S2). A single transformed colony was inoculated into LB medium containing kanamycin and grown at 37 °C, 220 rpm for 8 h. The culture was expanded in fresh LB medium with kanamycin and grown for an additional 12 h. This secondary culture was then inoculated at a 1:100 ratio into 400 mL LB medium containing kanamycin and grown at 37 °C, 220 rpm until the optical density at 600 nm (OD600) reached 0.4–0.6. Protein expression was induced by adding 500 µM isopropyl β-D-thiogalactopyranoside (IPTG) and continuing incubation under the same conditions.

Cells were harvested by centrifugation at 4000 × g for 5 min at 4 °C, resuspended in 80–100 mL lysis buffer (20  mM HEPES, 300 mM NaCl, 10% glycerol, 0.1 mM PMSF, pH 7.5), and passed through a 70-mesh strainer for dispersion. Cell disruption was performed using a pre-chilled high-pressure homogenizer at 750 bar. The lysate was clarified by centrifugation at 8000 × g for 30 min at 4 °C, followed by filtration through a 0.45 µm membrane.

The supernatant was loaded onto a Ni-NTA affinity column (AKTA Purifier system) pre-equilibrated with lysis buffer. Bound protein was eluted with a gradient of increasing imidazole concentration, and protein elution was monitored at 280 nm. Fractions containing WDR5 were pooled, aliquoted into 1.5 mL tubes, and further purified by size-exclusion chromatography (SEC) on a GE column using buffer containing 20 mM HEPES, 200 mM NaCl, and 0.3 mM TCEP, pH 7.5.

Plasmid construction and lentivirus transduction

WDR5 and c-Myc were subcloned into pCDH-CMV-MCS-EF1-Puro plasmids (#CD510B-1, System Biosciences) with an indicated tag for transient expression using the primers listed in Table S2). Lentivirus was produced by HEK293 cells with pCMV-dR8.91 (packaging vector, P0259, miaolingbio.Inc), pMD2.G-VSVG (envelope vector, #12259, Addgene), and targeted plasmids co-transfected using Lipofectamine 2000 (#11668019, Invitrogen). Medium containing virus was harvested 48 h after transfection and filtered through a 0.45-μm Millipore filter. For infection, cells were grown to 20–30% confluency in six-well plate, and 1 mL of each virus was added with 2 μL polybrene (6 mg/mL). After the cells were infected for 12–16 h, the medium was changed for fresh medium.

MicroScale thermophoresis (MST) assay

Recombinant His-tagged WDR5 protein and fluorescently labeled RED-tris-NTA dye (NanoTemper Technologies, MO-L018) were diluted to final concentrations of 800 and 100 nM, respectively, in 1× PBS-T buffer. A total of 90 µL of diluted WDR5 protein was mixed with 90 µL of RED-tris-NTA dye and incubated at room temperature in the dark for 30 min to allow labeling.

Small-molecule compounds were prepared as a 50 µM stock solution in 1× PBS-T buffer containing 10% DMSO. Sixteen consecutive two-fold serial dilutions of the compound were prepared in the same buffer, with a final volume of 10 µL per dilution. For interaction measurements, 10 µL of labeled WDR5 protein (400 nM in 1× PBS-T buffer) was added to each 10 µL aliquot of serially diluted compound.

The resulting mixtures were loaded into MST capillaries (NanoTemper Technologies, MOK022) and analyzed using a Monolith NT.115 instrument controlled by MO. Control software (NanoTemper Technologies). Binding affinities (Kd) were determined using MO. Affinity Analysis software.

Nanopore fabrication

All the nanopores were fabricated using a laser-assisted pipette puller (Sutter Instrument, P-2000, USA) by pulling quartz capillaries (Sutter Instrument, QF100-50-7.5). One capillary was then positioned onto the holder of the puller by aligning and securing it in the groove. The capillary was subsequently pulled to generate two similar nanopipettes using an optimized two-line pulling parameter1: HEAT = 775, FIL = 4, VEL = 30, DEL = 170, PUL = 802; HEAT = 805, FIL = 3, VEL = 20, DEL = 145, PUL = 180. It is worth noting that the pore geometry is significantly influenced by the instrument’s state and ambient conditions, including laser intensity, temperature and humidity. Under our experimental conditions (20 °C and 34% humidity), the average size of our nanopores with this protocol is 22.2 ± 5.0 nm (n = 5) unless stated otherwise, based on the SEM estimation.

Conductance measurement of nanopores

Before each experiment, nanopore conductance measurement was conducted in a solution of 1× PBS and 1 M KCl (pH = 8.6–9.4) by measuring the current ramping from −400 to +400 mV at a rate of 50 mV. Subsequently, the current–voltage plots were linearly fitted, and the conductance was estimated by subtracted the slope of the line to ensure that the consistency of nanopore sizes used in each experiment.

Translocation experiments and drug screening

All translocation experiments through nanopores were conducted in a solution of 1× PBS and 1 M KCl (pH = 8) at room temperature. A nanopipette filled with electrolyte (≈10 µl) was immersed into the aforementioned solution (≈200 µl). Two freshly prepared Ag/AgCl electrodes were inserted into the nanopipette and the bath, respectively, serving as working and reference electrodes. Measurements were performed using an Axopatch 700B patch clamp amplifier (Molecular Devices, USA). The signal was digitized with a Digidata 1440B at a sampling rate of 250 kHz. For protein measurement experiments, WDR5 proteins at a final concentration of 500 nM were added to the bath. A bias, typically 250 mV in this study unless stated otherwise, was applied to drive the translocation of WDR5 in an outside-to-inside direction. Peak currents of pure WDR5 were detected and recorded during each translocation experiment, ensuring consistency in nanopore geometries by using freshly pulled nanopores. For oligomerization inducer detection from our in-house compounds library, molecular mixture at selected concentrations (single molecule: protein = 2.5: 1 or 10: 1) were added to the bath and incubated for half an hour. After incubation, all relevant peaks were detected and recorded for subsequent data analysis.

SEC-MALS analysis

Size-exclusion chromatography with multi-angle light scattering (SEC-MALS) were performed using a Superdex 200 increase column (Cytiva) combined with multi-angle laser light scattering using a Wyatt HELEOS-II 18-angle photometer coupled to a Wyatt Optilab rEX differential refractometer (Wyatt Technology Corp). Experiments were conducted at room temperature using tag-free WDR5 wild-type, C248S mutant, and N-terminal truncation mutant proteins at a concentration of 1.0 mg/mL (27 µM), and a flow rate of 0.5 mL/min in 20 mM Hepes pH 7.5, 200 mM NaCl. Before analysis, the proteins were pre-incubated with 200 µM WZ-1 or DMSO control for 10 min and then centrifugated at 4 °C, 12,000 rpm for 2 min. The data were analyzed using the ASTRA 6.1 software (Wyatt) and GraphPad Prism 8.0. The molecular mass was determined across the protein elution peak65.

Nano-sensing data acquisition and analysis

A custom-written program in MATLAB (R2022b) was employed to analyze the translocation events of WDR5. In summary, the following steps were performed: (1) The I-t signal containing the peak segments was exported using the Clamfit and loaded by MATLAB for subsequent analysis. (2) The trace was filtered using a 10 kHz low-pass filter and subtracted to compensate for any baseline fluctuations. (3) Peak signals were distinguished from background noise using a threshold method, allowing for the identification of the start and end points of each peak. Peaks amplitude >45 pA; dwell time >10 μs were selected, and the current values within the pulse plateau were averaged to obtain mean values as peak currents66. (4) Each peak was idealized to extract relevant features. Dwell time is calculated from the width of the peak, and Peak current is determined by the average points within the signal of each peak, respectively. (5) Finally, peak statistics are then extracted (e.g., number of peaks, dwell time, and peak current)

Defining the rate of change

All Peak currents and dwell times from a single experiment were Gaussianized and averaged. The change in the mean value of the WDR5 current signal before and after each addition of compound mixtures was compared, and the rate of change was defined as follows:

Rate of change = (mean value of peak current in experimental group − mean value of peak current in control group)/mean value of peak current in control group.

A change of more than 50 percent was considered positive. The Gaussian fitting and figure plotting were performed using Originlab (2023) and Excel (Microsoft Office 365), and the results were further imported into Adobe Illustrator (Adobe CC) for visualization.

Cell culture

The CT26, RKO, DLD1, and SW620, 293T, NCM460 cell lines were purchased from the Chinese Academy of Sciences National Collection of Authenticated Cell Cultures (Shanghai, China). Immortalized human colon fibroblasts were obtained from Hunan Fenghui Biotechnology Co., Ltd (Changsha, China). All cell lines were examined by STR profiling and were monitored for mycoplasma contamination semiannually. CT26, DLD1 and NCM460 cells were cultured in RPMI-1640 medium supplemented with 10% fetal bovine serum (Cytiva, Cat#SH30396.03). RKO cells were cultured in Eagle’s minimum essential medium supplemented with 10% FBS. SW620 cells were cultured in L-15 media supplemented with 10% FBS. 293T cells were cultured in DMEM medium supplemented with 10% FBS. Immortalized human colon fibroblasts were cultured with special culture medium (Fenghui, FH1118-21). All cell lines were cultured in a humidified 5% CO2 incubator at 37 °C.

Co-immunoprecipitation

Twenty-four hours after transfection with pCDH or tagged target plasmids, the culture medium was replaced with Opti-MEM supplemented with 100 µM BSO and incubated for 8 h. Cells were then treated with WZ-1 for an additional 3 h, harvested, and lysed in ice-cold RIPA buffer (50 mM Tris-HCl, 150 mM NaCl, 0.1% SDS, 1% NP-40, 0.5% sodium deoxycholate, pH 7.4, 0.1 mM sodium orthovanadate, 5 µg/mL leupeptin, and 0.1 mM PMSF) at 4 °C for 30 min. Lysates were clarified by centrifugation at 15,000 × g for 30 min at 4 °C. An aliquot of the supernatant was mixed with 6× loading buffer, heated at 95 °C for 30 min, and used as the input sample. For immunoprecipitation, FLAG beads (Smart-Lifesciences, Cat# SA042005) were washed twice with 1% NP-40 buffer (25 mM Tris-HCl, 300 mM NaCl, 10% glycerol, 1% NP-40, pH 7.4) before incubation with the remaining supernatant in a 1.5 mL microcentrifuge tube overnight at 4 °C on a 360° rotator. After incubation, beads were pelleted by centrifugation and washed 3–5 times with 0.2% NP-40 buffer (25 mM Tris-HCl, 500 mM NaCl, 10% glycerol, 0.2% NP-40, pH 7.4). The final bead pellet was resuspended in 2× loading buffer, heated at 95 °C for 30 min, and subjected to downstream analysis.

Western blot analysis

An equal amount of protein samples (20–40 µg) was pipetted and loaded into 10% Tris-glycine gels filled with Tris-glycine SDS running buffer. After the proteins were resolved, they were transferred onto methanol-activated PVDF membrane using a trans-blot electrophoretic transfer cell. Upon completion of membrane transfer, the membrane was blocked with the blocking buffer (1× TBS-T containing 5% skim milk) and incubated for 30 min at room temperature on a horizontal shaker. Subsequently, the membranes were washed three times for 5 min each with 1× Tris-buffered saline Tween 20 (TBS-T) for subsequent blue staining or immunoblotting. The solution containing primary antibodies was then added and placed on a horizontal shaker at 4 °C for overnight incubation. Following this, the PVDF membranes were washed with 1× TBST on a shaking shaker, repeating the wash three times for durations of 15, 5, and 5 min, respectively. Next, the membranes were incubated in the solution containing HRP-labeled secondary antibodies corresponding to the species of the primary antibodies. This solution was diluted to 1:5,000 in 1× TBS-T buffer containing 5% skimmed milk and incubated on a horizontal shaker at room temperature for 1 h. After the completion of the incubation period, the membranes were washed three times with 1× TBS-T. Finally, the PVDF membrane was uniformly covered with an ECL exposure solution and placed into the AI680 imager for exposure and imaging. For all Western blot experiments mentioned, each experiment was independently repeated at least twice with similar results.

SRB assay

All cells were cultured in Opti-MEM medium supplemented with 100 µM BSO and maintained in a humidified incubator at 37 °C with 5% CO₂. Cell lines were tested for mycoplasma contamination every 3 months. For the SRB assay, cells seeded in 96-well plates were fixed with 10% (w/v) trichloroacetic acid at 4 °C for 2 h. The cells were then rinsed thoroughly with deionized water and dried at 65 °C. Next, cells were stained with SRB solution at room temperature for 25 min, washed with 1% glacial acetic acid to remove unbound dye, and dried again at 65 °C. Finally, bound dye was solubilized in 10 mM Tris-base, and absorbance was measured at 515 nm using a microplate reader.

RNA seq analysis

After treatment with WZ-1 for 12 h in Opti-MEM medium, RKO and CT26 cells were harvested and immediately subjected to rapid freezing in liquid nitrogen before storage at −80 °C. Subsequently, the samples were sent to LC-Bio Technologies (China, Hangzhou) Co for detection and analysis.

Cell cycle analysis

Cell cycle distribution analysis was conducted using the Cell Cycle Analysis Kit (Beyotime Biotechnology, cat#C1052). After seeding cells in 6-well plates, they were treated with indicated compounds. The cells were then harvested and fixed in 70% ethanol at 4 °C overnight. Then cells were washed twice with PBS and incubated in the dark with propidium iodide (PI) and RNase A for 30 min at 37 °C. Finally, the cell cycle phases were analyzed using a flow cytometer (BD FACscanta™ II).

Cryo-EM data acquisition and analysis

Initially, 1 mg/mL tag-free WDR5 was incubated with 200 µM WZ-1 on ice for 30 min before vitrification. Aliquotes of 3 µl of the reaction mixture were applied to glow-discharged holey carbon grids (Quantifoil Au R1.2/1.3). The grids were blotted for 3.5 s and flash-frozen in liquid ethane with Vitrobot (Mark IV, Thermo Fisher Scientific). The prepared grids were transferred to a Titan Krios operating at 300 kV equipped with a Falcon 4 detector. Movie stacks were automatically collected using EPU software (Thermo Fisher Scientific) with a defocus range from −0.8 to −1.6 µm at a nominal magnification of ×130,000, yielding a pixel size of 0.93 Å on images. Movies were recorded with a 3.42 s exposure to give each stack a total dose of 52 e/Å2. A total of 929 movie stacks were imported and processed with cryoSPARC version 4.3.1. 1,150,237 particles were auto picked with the Blob picker program and subjected to 2 rounds of 2D classification to remove nonprotein particles. Later, four additional datasets were collected with a protein concentration of 0.1 mg/mL to compare the oligomerization states of WT WDR5 (561 movies), WT WDR5 with WZ-1 (588 movies), WDR5 C248S with WZ-1 (667 movies) and WDR5 31-334 with WZ-1 (537 movies). All datasets were subjected to the same data processing protocol. After 2 rounds of 2D classification, the number of particles representing monomer, dimer and oligomer WDR5 classes were used for analysis.

Statistical analysis

Statistical analysis was performed using GraphPad Prism 7.0 software. The number of biological replicates for each experiment is indicated in the figure legends. Differences between means were determined using one-way analysis of variance (ANOVA) and were considered significant at P < 0.05.

Chemical probe consideration

WZ-1 was evaluated against community guidelines for chemical probes67 and shows reproducible, target-relevant cellular activity consistent with WDR5 biology. As it does not yet meet all probe validation criteria, including extensive proteome-wide selectivity profiling, WZ-1 is used here as a target-directed small-molecule research tool.

Reporting summary

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