Rapid Covalent-Probe Discovery by Electrophile-Fragment Screening

Abstract

Covalent probes can display unmatched potency, selectivity, and duration of action; however, their discovery is challenging. In principle, fragments that can irreversibly bind their target can overcome the low affinity that limits reversible fragment screening, but such electrophilic fragments were considered nonselective and were rarely screened. We hypothesized that mild electrophiles might overcome the selectivity challenge and constructed a library of 993 mildly electrophilic fragments. We characterized this library by a new high-throughput thiol-reactivity assay and screened them against 10 cysteine-containing proteins. Highly reactive and promiscuous fragments were rare and could be easily eliminated. In contrast, we found hits for most targets. Combining our approach with high-throughput crystallography allowed rapid progression to potent and selective probes for two enzymes, the deubiquitinase OTUB2 and the pyrophosphatase NUDT7. No inhibitors were previously known for either. This study highlights the potential of electrophile-fragment screening as a practical and efficient tool for covalent-ligand discovery.

Figure 1

Electrophile-fragment library adheres to the “rule-of-three”. Distribution of (A) chloroacetamides and acrylamides in the library (B) molecular weights of the fragments, including the electrophile moiety, (C) number of heavy atoms, (D) c log P values, (E) number of rotatable bonds, (F) number of hydrogen-bond acceptors, and (G) number of hydrogen-bond donors. The library largely adheres to the “rule of three” for fragment libraries.

Figure 2

High-throughput thiol-reactivity measurements across electrophile fragments. (A) Schematic description of the high-throughput thiol-reactivity assay. In the presence of TCEP, DTNB is reduced to TNB^2–, which has a strong absorbance at 412 nm and is yellow under natural light. Alkylation of TNB^2– by an electrophile fragment reduces the observed absorbance. (B) Example of the reactivity measurement for PCM-0102854. (C) Example of a second-order kinetic rate calculation. The data are fitted to a second-order reaction. [A] is the concentration of the electrophile, and [B] is the concentration of TNB^2–. The rate is determined by a linear regression of the data across 4 h of measurement. (D) Distribution of the rates of all the electrophile fragments in the library. (E) Thiol-reactivity-rate distributions for the fragments containing prevalent substructures. (F) Chemical structures of chloroacetamide substructures in the library that are represented by at least 10 fragments (see Figure S2 in the Supporting Information for the same analysis for acrylamides).

Figure 3

Intact protein LC/MS screen identifies hits for most targets (A) (top) Outline of our screening pipeline. Compounds are pooled, five in each well, incubated with a target protein for 24 h and read via LC/MS. (bottom) Example of the MS deconvoluted spectrum for NUDT7 with no compound (blue) and after 24 h incubation with five compounds (green). Note that the shift in the mass of the protein corresponds to 100% labeling of PCM-0102951. (B) Summary of the quantified labeling of 10 proteins by the electrophile library. Blue represents 100% binding and white no labeling or data not available (see labeling assignment in Methods).

Figure 4

Discovery of a selective OTUB2 inhibitor by fragment growing. (A) Cocrystal structures of OTUB2 in complex with (from top left) PCM-0102998, PCM-0103080, PCM-0102660, PCM-0103011, PCM-0103007, PCM-0102954, PCM-0103050, PCM-0102153, PCM-0102305, PCM-0102821, and PCM-0102500 (see Figure S13 in the Supporting Information). Structures with compounds containing the chloroacethydrazide motif are boxed. (B) Percent covalent labeling of OTUB2 with compounds containing the identified motif, PCM-0102300, PCM-0103009, PCM-0102142, PCM-0102998, PCM-0102954, and PCM-0102355 (Figure S13 in the Supporting Information), at 200 μM (blue) and 100 μM (green). Compounds boxed in (A) are marked with asterisks. (C) Percent covalent labeling of OTUB2 with selected next-generation compounds at 100 μM (see Figure S16 in the Supporting Information for all analogues and Table S4 in the Supporting Information for percent labeling). (D) Dose–response measurement of percent labeling by next-generation OTUB2 binders. All labeling in (B)–(D) was measured via LC/MS after 24 h incubation at 4 °C. (E) Inhibition of OTUB2 in an enzymatic assay (2.5 h preincubation in the presence of 2 mM free cysteine). (F) Gel-based ABPP selectivity assessment using a fluorescent activity-based DUB probe in HEK293 cells overexpressing OTUB2 (see Figure S21 in the Supporting Information for the full gel and additional controls). Cells were incubated with DMSO or increasing concentrations of OTUB2-COV-1 and then lysed, and the DUBs were fluorescently labeled with an alkyne ABPP probe: the higher the level of labeling by an inhibitor, the lower the level of labeling by the ABPP probe. The only bands that are reduced by OTUB2-COV-1 up to 30 μM correspond to OTUB2-GFP (at 65 kDa) and a degradation product (∼55 kDa). This demonstrates both OUTB2’s cellular engagement with OTUB2-COV-1 in cells and its high selectivity over other DUBs (see Figure S20 in the Supporting Information for the corresponding experiment in lysates). (G) Chemical structures of selected next-generation OTUB2 binders. The chloroacethydrazide motif is highlighted in red.

Figure 5

Discovery of a potent NUDT7 inhibitor by fragment merging. (A) Chemical structures of similar hit compounds that exhibited 68% (PCM-0102716), 88% (PCM-0102512), and 100% (PCM-0102558, PCM-0102298, PCM-0102951, and PCM-0102938) labeling of NUDT7 in the primary screen. Compound NUDT7-REV-1 is a noncovalent fragment (purple) that was identified as a NUDT7 binder in a crystallography soaking screen (see (E)). NUDT7-COV-1 (blue) is a merged compound based on PCM-0102716 (magenta) and NUDT7-REV-1. (B) The six hits identified in the primary screen stabilize NUDT7 by 4.5–8.1 °C in a T_m shift assay. (C) Labeling percentage of compounds PCM-0102558, PCM-0102951, PCM-0102298, PCM-0102716, PCM-0102512, and PCM-0102938 at 5–200 μM, 24 h, 4 °C. (D) Cocrystal structures of NUDT7 with compounds PCM-0102951, PCM-0102558, and PCM-0102716. (E) Overlay of the crystal structures of NUDT7 with compound PCM-0102716 (pink) and with the noncovalent fragment NUDT7-REV-1 (purple). (F) The cocrystal structure of NUDT7 with the merged compound NUDT7-COV-1 adopts the exact same binding mode as the two separate fragments. (G) Enzymatic inhibition of NUDT7 by NUDT7-COV-1 and NUDT7-REV-1. The data shown include results with and without a 30 min protein preincubation in the presence of the compounds. (H) Intracellular target engagement is demonstrated by thermal stabilization of FLAG-NUDT7 by NUDT7-COV-1 in intact HEK293 cells. After transfection, cells were treated with 20 μM NUDT7-COV-1 or DMSO for 30 min before being heated to the indicated temperatures.