Structure-Based Identification of Kelch-like ECH-Associated Protein 1 as a Pharmacological Target of Electrophile-Containing Catechol-O-Methyltransferase Inhibitors

Abstract

Entacapone and nitecapone are electrophile-containing catechol-O-methyltransferase (COMT) inhibitors that are used to treat Parkinson’s disease in combination with L-DOPA. It is desirable to investigate whether they can covalently bind to cellular protein targets using their reactive electrophilic warheads. We identified Kelch-like ECH-associated protein 1 (KEAP1), a sensor for oxidative and electrophilic stress, as a potential pharmacological target of both drugs by performing covalent-based reverse docking. We confirmed that both drugs activate nuclear factor erythroid 2-related factor 2 (NRF2) by reversibly modifying C151 on KEAP1. Both drugs can enhance the expression of growth differentiation factor 15 (GDF15) and NRF2 downstream antioxidant response element (ARE) genes, both in vitro and in vivo. Furthermore, both drugs exhibit anti-inflammatory effects in an NRF2-dependent acute gout model. Our findings suggest that these two drugs could be repurposed for the treatment of NRF2-modulated inflammatory diseases, and the 3-methylene-acetylacetone group of nitecapone could serve as a new reversible covalent warhead.

Entacapone is a catechol-O-methyltransferase (COMT) inhibitor that has been used safely for over 20 years to treat Parkinson’s disease in combination with L-DOPA.¹ It also inhibits fat mass and obesity-associated protein (FTO), as identified recently by a structure-based virtual screening approach.² Entacapone binds to both targets noncovalently, as evidenced by crystallographic studies.²⁻⁴ However, entacapone has a well-known Michael acceptor, an electrophilic cyanoacrylamide group (Figure 1A), which can reversibly modify cysteine residues.⁵⁻⁷ No covalent binding targets of entacapone at the cellular level have been reported so far.

Figure 1.

Identification of KEAP1 as a reversible covalent binding target of entacapone. (A) The chemical structure of entacapone is shown, with the Michael acceptor group highlighted in red. (B) The covalent-based reverse docking process used to screen potential targets for entacapone, and KEAP1 as a representative potential target. (C) The binding interactions between entacapone and wild-type (WT) or C151 mutant (C151S) KEAP1-BTB domain measured using the Thermal Shift Assay. (D) Deconvoluted ESI-MS analysis of BTB WT and BTB-C151S mutant after incubation with entacapone. BTB exhibited a mass shift of 305.8 Da (left), while no adduct was detected for BTB-C151S under identical conditions (right). (E) Deconvoluted ESI-MS was employed to analyze the reversibility of entacapone and KEAP1-BTB. The protein/entacapone ratio was maintained at 1:1.5 for all three samples. In the middle and right samples, the reaction was preincubated and then exposed to a 2-fold or 10-fold concentration of DMF relative to entacapone. (F) Deconvoluted ESI-MS analysis was conducted to verify the formation of adducts between Cpd-1 and KEAP1-BTB WT after incubation. (G) The GSH reactivity of entacapone analogues was tested, and their proton affinity was calculated to evaluate their reactivity.

Reversible covalent drugs are a promising and innovative strategy in drug development. They may interact stably with the target and have lower immunogenicity and off-target effects than irreversible covalent inhibitors.⁸⁻¹¹ Cyanoacrylamide-based reversible covalent inhibitors have been developed for kinases including Fibroblast Growth Factor Receptor (FGFR), Ribosomal S6 Kinase 2 (RSK2), Epidermal Growth Factor Receptor (EGFR), and Bruton’s Tyrosine Kinase (BTK).^5,6,12,13 PRN1008, a BTK reversible covalent inhibitor, is in phase II/III clinical trials and shows potent and durable target inhibition.^6,14,15 In addition, the reversible covalent strategy has also been applied in Proteolysis Targeting Chimeras (PROTACs) technology to target specific proteins for degradation.^7,16

Kelch-like ECH-associated protein 1 (KEAP1), which contains several reactive cysteine residues, functions as an intracellular sensor for electrophiles and oxidants.¹⁷⁻¹⁹ Under normal conditions, KEAP1 binds to the transcription factor nuclear factor erythroid 2-related factor 2 (NRF2) through its Kelch domain, sequestering it in the cytoplasm and promoting its degradation through the ubiquitin-proteasome pathway. When cells are exposed to oxidative stress or electrophilic compounds, the cysteine residues of KEAP1 undergo covalent modifications, leading to a conformational change in the protein, and subsequent release of NRF2 from KEAP1, enabling NRF2 to translocate into the nucleus.²⁰⁻²² Once in the nucleus, NRF2 activates downstream genes containing antioxidant response element (AREs) to exert anti-inflammatory and antioxidant effects.²³⁻²⁵ Therefore, activating NRF2 by modulating KEAP1 could be beneficial for various diseases such as chronic lung and liver disorders, autoimmune diseases, neurodegenerative disorders, and metabolic disorders.²⁶⁻³¹ Small-molecule activators of NRF2, such as CDDO, DMF, and sulforaphane, have been studied extensively (Figure S1A).³²⁻³⁸ Only two drugs that target KEAP1 to activate NRF2, DMF and RTA-408, have been approved by the FDA for multiple sclerosis and Friedreich’s ataxia, respectively.³⁹⁻⁴² However, both drugs have shown some adverse effects in patients. DMF can cause flushing and gastrointestinal issues, decreased lymphocyte counts, and elevated transaminases. RTA-408 can also increase transaminases and cause headache and nausea.^40,41 Therefore, there is a need for safer and more drug-like NRF2 activators that can be applied to a broader spectrum of diseases.^31,35,42

Several covalent docking methodologies have been developed to expedite covalent ligands discovery.⁴³⁻⁴⁹ In this study, we employed a covalent-based reverse docking strategy to screen entacapone, which has an electrophilic cyanoacrylamide group, against protein targets with reactive cysteine residues. We found that entacapone can reversibly bind to C151 on KEAP1. We also found another electrophile-containing COMT inhibitor, nitecapone, which has a new 3-methylene-acetylacetone warhead and binds to KEAP1 similarly. Both drugs activated NRF2 in vitro and in vivo. They also showed anti-inflammatory activity in a mouse acute gout model, suggesting that they are a new class of oral KEAP1 reversible covalent binders.

Results

Covalent-Based Reverse Docking

We collected 125 unique protein targets with active cysteines known to be modified by covalent ligands from the public databases.⁵⁰⁻⁵³ We used a covalently based reverse docking strategy to screen entacapone and nitecapone against these targets (Figure 1B). Four stereoisomers of covalent adduct of entacapone with the cysteine residue and two stereoisomers of covalent adduct of nitecapone with the cysteine residue were generated, individually. After combining the docking scores and the number of hydrogen bonds formed with the protein binding pocket for all stereoisomers, we identified a few potential targets for entacapone and nitecapone (Table S1). We chose KEAP1, a known drug target modified by covalent binders, for further studies.

Reversible Covalent Interaction between Entacapone and KEAP1

We used a thermal shift assay (TSA) to verify the direct interaction between entacapone and the purified BTB domain of KEAP1. The shift in the melting temperature (ΔT_m) of the KEAP1-BTB protein increased with the dose of entacapone compared to the DMSO control. This effect was reduced when we mutated the C151 of KEAP1-BTB (C151S), suggesting a covalent interaction (Figure 1C). We also used mass spectrometry to analyze the compound-protein adduct. We found a product with an extra mass of 305.71 Da in the entacapone-treated KEAP1-BTB protein. This matched the molecular weight of entacapone (305.29 Da). No adduct formation was observed in entacapone-treated C151S protein under the same experimental conditions, indicating that C151 on KEAP1-BTB is essential for covalent bond formation (Figure 1D).

We used an ultraviolet–visible (UV–vis) spectroscopy-based method to test the reversibility of the interaction of entacapone with GSH.⁵ As the dose of GSH increased, the intensity of entacapone’s main UV–visible band at 380 nm decreased. The K_d value is 3.2 mM (Figure S1B). Furthermore, the 380 nm peak reappeared after diluting the mixture with PBS, showing the reversibility of the cyanoacrylamide-containing entacapone with the thiol group of cysteine, consistent with the previous result (Figure S1C).⁵ Additionally, we employed the KEAP1-BTB protein to validate this characteristic. Upon incubation with entacapone, a mass shift of 305.7 Da was observed at a ratio of 56.6%. Subsequent incubation with a high concentration of DMF (M_w: 144 Da), a traditional irreversible covalent KEAP1 modifier, resulted in mass shifts of 305.4 (reduced to 8.9%) and 144.3 Da (75%). Importantly, higher concentrations of DMF completely abolished entacapone modification due to competitive binding with C151, leading to a mass shift of 144.2 Da (87.4%). These findings further support the notion of reversibility of cyanoacrylamide-containing entacapone with the KEAP1-BTB protein (Figure 1E).

The nitrile group is a strong electron-withdrawing group, which increases the electropositivity and reactivity of the β-carbon atom of the Michael acceptor. Thus, we made three tool compounds by removing the nitrile group (Cpd-1), replacing it with acetyl (Cpd-2) or fluorine (Cpd-3) to study its role in the reactivity of entacapone. The UV–visible spectroscopy assays showed that none of the three compounds reacted with GSH (Figure S1D). We also saturated the double bond of the cyanoacrylamide group and made Cpd-4. Unlike entacapone, the TSA assay showed that all tool compounds failed to increase the Tm of KEAP1-BTB protein such as entacapone (Figure S1E). The MS analysis confirmed this, as no adduct was found in the Cpd-1-treated KEAP1-BTB protein (Figure 1F). We used the quantum mechanics (QM) approach to calculate the proton affinity (PA) of entacapone and its analogues. We found that all nitrile-free compounds had PA values higher than those of entacapone, which is consistent with the GSH reactivity results (Figure 1G). Our results show that entacapone reversibly covalently binds to C151 of KEAP1 and that the cyanoacrylamide of entacapone is essential for this binding. Thus, KEAP1 is a potential covalent binding target of entacapone.

KEAP1-Entacapone Binding Mode Analysis

Except for the covalent bond formed between the cyanoacrylamide and C151, the noncovalent interactions are also critical for entacapone binding with KEAP1 (Figure 2). The docking pose reveals that the hydroxyl group at the meta position on the nitrocatechol ring formed a hydrogen bond with S146. The model also showed that the deprotonated ortho hydroxyl group of entacapone formed a charge–charge interaction with R135, and the nitro group interacted with K131.

Figure 2.

Binding model of entacapone with KEAP1 (docking pose based on structure PDB ID: 5DAD), COMT (PDB ID: 4XUD), and FTO (PDB ID: 6AK4). The hydrogen-bond interactions between the ligand and the protein are color-coded in pink, while the gray dashed lines represent ion chelation interactions. The graphics were generated using UCSF Chimera.⁵⁴

It is interesting to compare how entacapone binds to KEAP1, COMT, and FTO. We used the COMT structure (4XUD) as a reference for modeling the COMT-entacapone binding complex (Figure 2). The nitrocatechol moiety of entacapone chelated with metal ion similar to the crystal ligand in the COMT complex structure. In striking contrast, in the crystal complex structure of entacapone bound with FTO, the nitrile group chelated with the binding-site metal ion instead and the hydroxyl group at the meta position of the nitrocatechol ring formed hydrogen bonds with R322 and Y106 (Figure 2). This is rare as a drug molecule binds with different targets in very different binding modes, illustrating interesting binding characteristics employed by a multitarget ligand.

KEAP1/NRF2 Pathway Activation by Entacapone

KEAP1 regulates NRF2 and protects cells from oxidative stress.^20,22 We tested if the reversible covalent binding of entacapone to KEAP1 can stabilize the NRF2 levels. We co-expressed Flag-tagged KEAP1 and Flag-tagged NRF2 in HEK293T cells and found that entacapone stabilized NRF2 in a dose-dependent way (Figures 3A and S2A). We also co-expressed Flag-tagged KEAP1-C151S and Flag-tagged NRF2 in HEK293T cells. We found that entacapone, like CDDO-Me, could not stabilize NRF2 with the C151S mutation (Figure 3A). Consistently, Cpd-1, which did not bind to KEAP1, also failed to stabilize NRF2 (Figure S2B). We used the PathHunter U2OS KEAP1–NRF2 functional assay (DiscoverX, catalog no. 93–0821C3) to check the nuclear translocation of NRF2 by entacapone. The results showed that entacapone promoted the nuclear translocation of NRF2 reporter with an EC₅₀ of 9.9 μM (Figure 3B).

Figure 3.

Activation of KEAP1/NRF2 pathway by entacapone. (A) Co-transfection of wild-type (WT) or C151 mutant (C151S) KEAP1 and NRF2 with Flag tag in HEK293T cells treated with DMSO, entacapone, or CDDO-Me as positive control for assessing NRF2 stability via Western blot. (B) Determination of NRF2 nuclear translocation EC₅₀ values of entacapone in U2OS cell line with reporter system using multiple doses. (C) Proteomic profiling reveals the effect of 100 μM entacapone treatment on HT-1080 cells. The upregulated targets are shown in red, and the downregulated targets are shown in blue. Notably, arrows indicate various downstream targets of the KEAP1/NRF2 pathway, including HO-1, GLCM, and NQO1, which exhibit differential expression levels. (D) Further analysis of protein variations and assessment of pathway enrichment affected by entacapone treatment. The data were analyzed using DAVID.⁵⁵ (E) Immunoblot analysis of protein expression levels including NRF2, NQO1, and GCLM in HT-1080 cells treated with entacapone at 50 and 100 μM, with DMSO as a control.

We further performed proteomics analysis in HT-1080 cells to investigate the effects of entacapone on the KEAP1-NRF2 pathway. The results showed a significant increase in the protein levels of NRF2 targets such as HO-1, GCLM, and NQO1 (Figure 3C). The NRF2 pathway was strongly enriched in the changed proteins after entacapone treatment (Figure 3D). These proteomic data were further validated by Western blot analysis, where both NRF2 and its downstream targets exhibited higher expression levels after entacapone treatment (Figure 3E). We also assessed the transcriptional activity of NRF2-driven genes, and the results were consistent with the observed protein changes (Figure S2C). In summary, our results provide strong evidence that entacapone can stabilize NRF2 protein and activate the NRF2 pathway by binding to C151 of KEAP1.

Nitecapone as a New and Efficient Electrophilic Modifier of KEAP1

Given the pivotal role of the Michael acceptor in entacapone’s interaction with KEAP1, we tested if nitecapone, another COMT inhibitor with a reactive 3-methylene-acetylacetone group, could also covalently bind to KEAP1 (Figure 4A). Similar to the experiments performed with entacapone, we used the UV–visible spectroscopy assay to measure the reactivity of nitecapone with GSH. We found that nitecapone reacted with GSH in a dose-dependent manner with a K_d of 2.5 mM (Figure S3A). We also used the thermal shift assay to confirm that nitecapone increased the ΔT_m of KEAP1-BTB-WT but not of KEAP1-BTB-C151S (Figure 4B). Additionally, the saturated compound (Cpd-5) of nitecapone did not increase the ΔT_m of KEAP1-BTB-WT, showing the importance of the 3-methylene-acetylacetone group and C151 for binding of nitecapone to KEAP1 (Figure S3B,C). We further confirmed this interaction by mass spectrometry. Nitecapone-treated KEAP1-BTB protein, but not KEAP1-BTB-C151S, had an extra adduct with an extra mass of 265.89 Da, which matched the molecular weight of nitecapone (265.22 Da) (Figure 4C). To check the reversibility of the 3-methylene-acetylacetone warhead of nitecapone, we performed a dilution assay between nitecapone and GSH. We saw that the absorbance peak of nitecapone at 380 nm recovered in a no GSH buffer after dilution (Figure S3D). We also incubated KEAP1-BTB protein with nitecapone, with a mass shift of 265.8 Da observed at a ratio of 64.1%. Higher concentrations of DMF completely abolished the modification of nitecapone, resulting in a mass shift of 144.3 Da (86.5%). These findings further support the reversible covalent bond formed by the 3-methylene-acetylacetone warhead (Figure 4D).

Figure 4.

Electrophilic COMT inhibitor nitecapone exhibits KEAP1/NRF2 activation. (A) Model of compound nitecapone with a 3-methylene-acetylacetone warhead in forming reversible covalent adducts with the thiol group of cysteine residue in the protein. (B) TSA confirms the binding interactions between nitecapone and wild-type (WT) or C151 mutant (C151S) KEAP1 at different doses, and the shift in melting temperature (ΔT_m) was presented for WT or C151S mutant. (C) Deconvoluted ESI-MS analysis of BTB WT and BTB-C151S mutant after incubation with nitecapone. BTB WT exhibited a mass shift of 265 Da (left), while no adduct was detected for BTB-C151S under identical conditions (right). (D) Deconvoluted ESI-MS was employed to analyze the reversibility of nitecapone and KEAP1-BTB. The protein:nitecapone ratio was maintained at 1:0.5. In the right sample, the reaction was preincubated and then exposed to a 30-fold concentration of DMF relative to nitecapone. (E) NRF2 nuclear translocation EC₅₀ values of nitecapone determined in U2OS cells using a reporter system at multiple doses. (F) Western blot experiment to assess the stability of NRF2 protein co-expressed with WT or C151 mutant KEAP1 and treated with DMSO, entacapone, nitecapone, or tolcapone.

We used the co-expression assay to confirm that nitecapone stabilized NRF2 protein in a dose-dependent manner (Figure S3E). In the PathHunter U2OS KEAP1-NRF2 functional assay, nitecapone had an EC₅₀ of 4.5 μM (Figure 4E). Again, nitecapone did not stabilize NRF2 with the KEAP1-C151S protein, showing that binding of nitecapone to KEAP1 depended on C151 (Figure 4F). While we treated HEK293T cells with tolcapone, a COMT inhibitor without an electrophilic group, NRF2 was not activated, and ΔT_m of KEAP1-BTB-WT was significantly decreased in the TSA assay (Figures 4F and S3F,G). Therefore, nitecapone used a new 3-methylene-acetylacetone warhead to form a reversible covalent bond with C151 of KEAP1, and thus activate NRF2 in cells.

Both Entacapone and Nitecapone Alleviate MSU-Induced Acute Gout Model

NRF2 is needed for GDF15 induction in mouse bone marrow-derived macrophages (BMDMs), and GDF15 induction can reduce inflammation in mice.^21,56,57 We tested if entacapone and nitecapone could induce the GDF15 and NRF2-GDF15 signaling pathway in BMDMs. Gdf15 expression in BMDMs increased significantly after treatment with the compounds, with nitecapone being more active than entacapone (Figure 5A). The activation of NRF2 downstream targets was also confirmed in BMDMs treated with entacapone and nitecapone. Nitecapone increased the expression of NRF2 targets Gdf15, Gclm, and Hmox1 more than entacapone (Figure 5B), consistent with the stabilization and nuclear translocation results of NRF2 (Figures 3B and 4E,F).

Figure 5.

Entacapone and nitecapone activate downstream targets of KEAP1/NRF2 in mouse bone marrow-derived macrophages (BMDMs) independent of COMT or FTO mediation. (A) Primary mouse bone marrow-derived macrophages were obtained and induced, then they were treated with 100 μM entacapone or nitecapone, and the activation of mRNA levels of Gdf15 was identified in a time-course analysis range from 1, 3, 6 to 12, 24 h (n = 3). (B) Primary mouse bone marrow-derived macrophages were treated with entacapone or nitecapone for 6 h. mRNA expression levels of KEAP1/NRF2 downstream targets (Gdf15, Csr, Tal, Cat, Gclm, and Hmox1) were detected (n = 3). (C, D) Relative mRNA expression levels of Gdf15 in BMDMs from Comt^+/+ or Comt^–/– (C) or Fto^+/+ or Fto^–/– (D) mice were determined after treatment with DMSO control, entacapone or nitecapone at various doses, and CDDO-Me (n = 3 per group).

Entacapone and nitecapone are known as COMT inhibitors, and entacapone has an additional target FTO.^1,2 We tested if COMT or FTO inhibition could activate NRF2. We used COMT- or FTO-knockout (KO) BMDMs treated with entacapone and nitecapone. There were no significant differences in the mRNA expression of Gdf15 after treatment with either compound (Figure 5C,D). These results suggest that the COMT or FTO does not affect NRF2 activation. We note that our previous study on entacapone inhibiting FTO required a much longer treatment time.² This is different from the fast stabilization of the NRF2 levels after a few hours of entacapone treatment.

Next, we used the MSU-induced acute gout model, which relies on NRF2 activation, to test the pharmacological effects of entacapone and nitecapone.²¹ Both drugs exhibited strong anti-inflammatory activity and effectively alleviated MSU-induced acute gout. Those in vivo results were consistent with the higher activation of GDF15 in serum and NRF2 targets with nitecapone than entacapone (Figure 6A–C).

Figure 6.

Entacapone and nitecapone effectively reduce inflammation in an MSU-induced gouty arthritis model by activating the KEAP1/NRF2 pathway. (A) Experimental mice at 6–8 weeks were divided into different groups and were given 100 mpk doses of entacapone or nitecapone to treat different time points (n = 3 per group). Changes in the expression levels of downstream target protein GDF15 were detected in mice blood by the ELISA assay kit methods. (B) Statistics of foot diameter of mice in different groups, saline as the control group and MSU as the experimental group, were measured at 48 h. * P < 0.05, ** P < 0.01, ***P < 0.001, entacapone or nitecapone vs MSU gouty model group. Data processed with Graph Prism. (C) Representative results of mouse feet after treatment with vehicle, entacapone, and nitecapone, respectively. Each image represents the feet from at least 6 mice. (D) The relative mRNA expression levels of the NRF2 pathway in the liver were assessed after 2 h treatment with DMSO control, entacapone, or nitecapone. (E) The relative mRNA expression levels of the NRF2 pathway in the kidney were assessed after 2 h treatment with DMSO control, entacapone, or nitecapone. (F) Drug distribution in liver and kidney after treatment with entacapone (n = 3). (G) Drug distribution in liver and kidney after treatment with nitecapone (n = 3).

To further examine the in vivo activation of the NRF2 pathway, we used quantitative PCR to measure the transcription levels of the NRF2 targets in different tissues. The results showed that liver and kidney had higher transcription levels of Gdf15 and Hmox1 among the 11 tissues tested (Figure S4A,B). Interestingly, entacapone activated Gdf15, Hmox1, Csr, Cat, Tadol, and Gclm more in the liver (Figure 6D), while nitecapone activated Gdf15, Hmox1, Csr, Gclm, and Nqo1 more in the kidney (Figure 6E). Furthermore, our Western blot results confirmed that entacapone and nitecapone both stabilized the NRF2 protein and activated its downstream targets in the kidney (Figure S4C). We also investigated the tissue distribution of entacapone and nitecapone through pharmacokinetic (PK) analysis and found more favorable distributions of entacapone in the liver and nitecapone in the kidney (Figure 6F,G). This finding may help explain the tissue-specific fashion of activating NRF2 downstream signaling by these two drug molecules. In summary, entacapone and nitecapone activated GDF15 through the KEAP1/NRF2 pathway and reduced inflammation in a gout mouse model.

Conclusions and Discussion

Covalent inhibitors are a promising class of small molecules for new drug discovery.^13,58,59 We used a covalent-based reverse docking approach to discover the potential targets of entacapone and nitecapone, noncovalent COMT inhibitors with reactive electrophile groups. Our study indicates that KEAP1 is a likely target of entacapone and nitecapone in vivo. We also describe the different binding modes of entacapone with different targets. It is interesting that a compound adopts very different binding modes to bind to different protein targets. Notably, the 3-methylene-acetylacetone warhead could expand the toolbox of covalent inhibitor discovery.

Both entacapone and nitecapone activated the KEAP1/NRF2 pathway in vitro and in vivo. They reduced inflammation in an MSU-induced foot gouty arthritis model, suggesting that they could treat NRF2-related diseases. Entacapone has been reported to protect against acute kidney injury by inhibiting ferroptosis.⁶⁰ In our study, we found that the ferroptosis pathway was also enriched in the proteomic analysis in addition to the NRF2 pathway, implying that entacapone may have multiple mechanisms (Figure 3D). Nitecapone has been reported to protect against diabetic nephropathy.^61,62 This disease is highly associated with KEAP1/NRF2 pathway.⁶³⁻⁶⁶ Nitecapone had strong tissue distribution and NRF2 activation in the kidneys, implying that nitecapone might be useful for inflammatory kidney diseases. More studies are needed to fully explore their therapeutic potential.

Experimental Methods

Covalent-Based Reverse Docking

The protein–ligand complexes with covalent cysteine modes were retrieved from the public databases and relevant literature on covalent inhibitors. Receptor structures were preprocessed using the Protein Local Optimization Program (PLOP).⁶⁷ We collected 125 unique receptors from Homo sapiens, each with at least one complex structure for every covalent compound containing a cysteine residue in the binding site, for covalent docking using AutoDock4. Entacapone ligands were prepared in four stereoisomers (R,R; R,S; S,S; S,R), and nitecapone ligands were prepared in two stereoisomers (R; S) following covalent docking guidelines.⁴⁶ The chain and atom number of the cysteine residue used for covalent docking of each target were defined based on the protein–ligand complex structure. The results of each chiral format were evaluated based on docking scores and hydrogen-bond numbers. For each receptor, 10 docking poses were generated and evaluated based on AutoDock4 scores and hydrogen-bond numbers (with a threshold of at least 2). After filtering, the top candidates based on docking scores were further assessed through visualization, and the 11 final candidates are presented in Table S1.

Expression and Purification of Recombinant KEAP1

Human KEAP1-BTB domain (48–180 amino acids) purification was performed as previously described.³² The cDNA was subcloned into the pET28b vector to incorporate an N-terminal hexahistidine (His) tag. The S172A point mutation was introduced into the KEAP1-BTB cDNA to enhance the protein stability. Wild-type (WT) KEAP1-BTB and KEAP1-BTB-C151S mutant were expressed in BL21(DE3) cells. The cells were grown in LB medium and then induced with 0.25 mM IPTG after reaching OD₆₀₀ = 0.5–0.8, and incubated overnight at 18 °C with shaking at 200 rpm. Cells were harvested by centrifugation at 4000 rpm and 4 °C for 20 min and lysed by sonication with a 25 mM Tris-HCl (pH 8.0), 150 mM NaCl buffer. The lysate was then centrifuged at 12,000 rpm and 4 °C for 1 h. The crude protein was purified by Ni-affinity chromatography and eluted with 25 mM Tris-HCl (pH 8.0), 150 mM NaCl, and 250 mM imidazole buffer. After overnight Thrombin cleavage to remove the His tag, the purified protein was further purified using a Superdex 75 increase column with 25 mM Tris-HCl (pH 8.0) and 150 mM NaCl buffer. The final purified protein was confirmed by SDS-PAGE and stored at −80 °C.

Fluorescence-Based Thermal Shift Assay

The fluorescence-based thermal shift assay was performed as previously described.³² WT KEAP1-BTB and KEAP1-BTB-C151S mutant proteins were incubated with indicated concentrations of compounds and 5 × SYPRO Orange solution in a 20 μL reaction system. A Bio-Rad 96-well clear plate was used, and DMSO served as the control with a content of 0.1% in all groups. The reaction was detected using a Bio-Rad PCR system, with an initial temperature of 25 °C for 3 min followed by a 0.5 °C increase every 8 s until reaching 95 °C. Snapshots were taken at each 0.5 °C interval. After completion of the program, data was outputted and analyzed using the Bio-Rad software.

UV–Vis Spectrometry Assay

In the reactivity assessment, entacapone and nitecapone were incubated at a concentration of 200 μM with GSH of varying millimolar concentrations. The incubation was performed at room temperature using PBS as the dilution buffer. Absorbance measurements were then carried out in the wavelength range of 250–600 nm. In the investigation of the compounds’ reversible properties, the reaction products were subjected to 10-fold dilution in either plain PBS buffer or PBS buffer containing different GSH concentrations. The change in absorbance was quantified by using a TECAN scanning model machine, followed by subsequent data analysis.

Proton Affinity Calculation

Computational methods were used to study the reactivity of entacapone and its analogues. All geometric optimizations were performed with the Gaussian16 package. The minimum energy structure and the corresponding anion structure were optimized using the DFT method, and the proton affinity of anions in water was also calculated under the same level of the calculation method. The b3lyp functional and 6-311+G* basis set were used for all DFT calculations. The Grimme dispersion scheme (D3) was applied to describe the weak dispersion interaction, and the polarizable continuum model (PCM) was used to implicitly consider the solvent effect.

Cell Culture and Treatment

HT-1080 (ATCC) and HEK293T (ATCC) were maintained in Dulbecco’s modified Eagle’s medium (Gibco) containing 10% fetal bovine serum (Sigma) and incubated at 37 °C under humidified conditions and 5% CO₂. To obtain BMDMs, bone marrow cells from WT and transgenic mice were cultured for 5 to 7 days in Dulbecco’s modified Eagle’s medium (DMEM, Gibco) containing 10% FBS and conditioned media from L929 cell culture.

Cell Transfection

HEK293T cells were plated on 24-well or 12-well plates at a density of 0.7 × 10⁵ or 1.2 × 10⁵ cells per well. Flag-NRF2 and Flag-KEAP1 WT/C151S plasmids were transfected by Lipo8000 (Beyotime) according to the manufacturer’s instructions. 24 h after the transfection, the cells were treated with the indicated concentration of compounds for another 6 h. The cells were then washed three times with PBS and collected for Western blot.

Electrospray Ionization Mass Spectrometer (ESI MS)

Purified KEAP1-BTB (WT or C151S) was incubated with compounds for 2 h at 37 °C in 25 mM Tris and 150 mM NaCl, pH 8.0 buffer. Covalent adduct detection between compounds and BTB proteins was performed by using liquid chromatography-mass spectrometry (LC-MS). The LC-MS analysis was conducted by using a Thermo Scientific Vanquish UHPLC coupled with a Thermo Q Exactive Plus Biopharm system. Chromatographic separation was achieved using an ACQUITY UPLC Protein BEH C4 column (300 Å, 1.7 μm, 2.1 mm × 100 mm) at a column temperature of 60 °C. The mobile phases consisted of 0.1% formic acid in water and acetonitrile containing 0.1% formic acid. The gradient elution program was as follows: 0–5 min, 5% B; 5–10 min, 5% to 90% B; 10–12 min, 90% B; 12–12.1 min, 5% B; and 12.1 to 15 min, 5% B. The sample injection volume was 2 μL, and the flow rate was maintained at 0.3 mL/min. Mass spectrometry acquisition parameters were set as follows: Sheath gas flow rate (45), Aux gas flow rate (10), Spray voltage (3.5 kV), Capillary temperature (320 °C), S-Lens RF Level (55), Aux gas heater temperature (310 °C), Scan Type (Full MS), Microscans (5), Resolution (17500), AGC target (3e6), Maximum IT (200 ms), Number of scan ranges (1), Scan range Spectrum (1000–2500 m/z), Data type (Profile). Data analysis involved the usage of the Intact Mass Analysis module in BioPharma Finder 4.1.

Mice Gout Model

The mice gout model was performed as previously described.²¹ The eight-week-old male mice were intraperitoneal injection with Vehicle or Entacapone (100 mg/kg) or Nitecapone (100 mg/kg) twice with an interval of 12 h. One hour after the last compound injection the left hind foot pad of mice was injected intradermally with 40 μL of saline or MSU (Invivogen). The foot was measured by calipers 24 or 48 h after injection, and all animal experiments were approved by the Ethics Committee.

Chemistry

All final compounds were characterized using liquid chromatography–mass spectrometry (LC-MS), high-resolution mass spectrometry (HRMS), and nuclear magnetic resonance (NMR) experiments (refer to the Supporting Information). The purity of each compound evaluated in biochemical assays exceeded 95%.

Synthesis of Cpd-1

Cpd-1 was prepared in two synthetic steps from 3,4-dihydroxy-5-nitrobenzaldehyde, according to the following procedure:

Step 1: Synthesis of (E)-3-(3,4-dihydroxy-5-nitrophenyl)acrylic Acid (2)

A mixture of 3,4-dihydroxy-5-nitrobenzaldehyde (500 mg, 2.73 mmol) and malonic acid (570 mg, 5.46 mmol) in anhydrous pyridine (15 mL) and piperidine (0.2 mL) was stirred at 60 °C under an argon atmosphere for 48 h. The reaction mixture was then cooled to room temperature (rt) and diluted with H₂O (30 mL). Then, the mixture was neutralized with 1M HCl and extracted with ethyl acetate (100 mL × 3). The combined organic phases were washed with brine and dried by anhydrous Na₂SO₄. Concentrated under vacuum, purified by column chromatography (dichloromethane/methanol = 9:1) to yield the product (490 mg, 81%). MS [M + H]⁺ calcd for C₉H₇NO₆ 226.1, found 226.1.

Step 2: Synthesis of (E)-3-(3,4-dihydroxy-5-nitrophenyl)-N,N-diethylacrylamide (Cpd-1)

A solution of compound 2 (100 mg, 0.44 mmol) in SOCl₂ (10 mL) was stirred at 80 °C for 2 h. Then, the reaction mixture was concentrated under vacuum to give the crude intermediate which was added to a solution of diethylamine (33 mg, 0.44 mmol) in THF (10 mL) at 0 °C, and the mixture was stirred at rt for 4 h, and then quenched by 1 M HCl and extracted with DCM (50 mL × 3). The combined organic phases were dried over Na₂SO₄ and concentrated under vacuum. The residue was purified by Prep-HPLC (H₂O/MeCN = 44:56, 0.1%TFA, twice) to obtain the product (9.6 mg, 0.6%). MS [M – H]⁻ calcd for C₁₃H₁₆N₂O₅ 279.1, found 279.1. ¹H NMR (400 MHz, DMSO-d₆) 10.38 (s, 1H), 7.67 (d, J = 2.0 Hz, 1H), 7.33 (dd, J = 8.7, 6.6 Hz, 2H), 6.92 (d, J = 15.3 Hz, 1H), 3.47 (m, 2H), 3.34 (m, 2H), 1.11 (t, J = 7.0 Hz, 3H), 1.02 (t, J = 7.0 Hz, 3H). ¹³C NMR (101 MHz, DMSO-d₆) δ 164.98, 148.38, 143.45, 140.11, 138.29, 126.22, 118.48, 118.33, 115.34, 29.56, 15.77, 13.75. HRMS (ESI): calcd for C₁₃H₁₅N₂O₅ [M – H]⁻: 279.0981; found 279.0986.

Synthesis of Cpd-2

Cpd-2 was prepared in one synthetic step from 3,4-dihydroxy-5-nitrobenzaldehyde, according to the following procedure:

Synthesis of (E)-2-(3,4-Dihydroxy-5-nitrobenzylidene)-N,N-diethyl-3-oxobutanamide (Cpd-2)

To a solution of 3,4-dihydroxy-5-nitrobenzaldehyde (500 mg, 2.73 mmol) and N,N-diethyl-3-oxobutanamide (557 mg, 3.55 mmol) in toluene (10 mL) were added piperidine (23 mg, 0.27 mmol), acetic acid (16 mg, 0.27 mmol), and molecular sieves (600 mg). The reaction mixture was stirred at 110 °C for 16 h. The mixture was then quenched by water, extracted by EtOAc (50 mL × 3), dried over anhydrous Na₂SO₄, concentrated under vacuum, and purified by column chromatography and Prep-HPLC to obtain a white solid as the target product (14.5 mg, 1.6%). MS [M + H]⁺ 323, ¹H NMR (400 MHz, DMSO-d₆) δ 7.67 (d, J = 2.1 Hz, 1H), 7.56 (s, 1H), 7.31 (d, J = 2.1 Hz, 1H), 3.43 (m, 2H), 3.05 (m, 2H), 2.40 (s, 3H), 1.10 (t, J = 7.1 Hz, 3H), 0.80 (t, J = 7.1 Hz, 3H). ¹³C NMR (101 MHz, DMSO-d₆) δ 196.33, 166.33, 147.78, 143.66, 137.62, 137.18, 136.16, 123.88, 119.59, 117.38, 42.24, 38.25, 26.09, 13.67, 12.18. HRMS (ESI): calcd for C₁₅H₁₇N₂O₆ [M – H]⁻: 321.1087; found 321.1091.

Synthesis of Cpd-3

Cpd-3 was prepared in five synthetic steps from 3,4-dimethoxybenzaldehyde, according to the following procedure:

Step 1: Synthesis of 3,4-Dimethoxy-5-nitrobenzaldehyde(2)

To a solution of 3,4-dimethoxybenzaldehyde (2 g, 12 mmol) in acetic acid (20 mL) was added nitric acid (12 mL, 120 mmol). The reaction mixture was stirred at rt for 16 h. The mixture was quenched by water, extracted by DCM (50 mL × 3), dried over anhydrous Na₂SO₄, concentrated under vacuum, and purified by column chromatography to obtain a white solid as the target product (1.4 g, 55.1%). MS [M + H]⁺ 212.

Step 2: Synthesis of Ethyl (Z)-3-(3,4-dimethoxy-5-nitrophenyl)-2-fluoroacrylate(3)

To a solution of ethyl 2-(diethoxy phosphoryl)-2-fluoroacetate (1.6 g, 6.6 mmol) in dry THF (30 mL) was added Et₃N (1.34 g,13.2 mmol), followed by magnesium bromide (1.23 g, 6.6 mmol). The reaction mixture was heated to 50 °C, and 3,4-dimethoxy-5-nitrobenzaldehyde (1.4 g, 6.6 mmol) was added. The reaction mixture was stirred at 50 °C for 16 h. The mixture was concentrated under vacuum and purified by column chromatography to obtain the target product (1.6 g, 80.6%). MS [M + H]⁺ 300.

Step 3: Synthesis of (Z)-3-(3,4-Dimethoxy-5-nitrophenyl)-2-fluoroacrylic Acid(4)

To a solution of ethyl (Z)-3-(3,4-dimethoxy-5-nitrophenyl)-2-fluoroacrylate (1.6 g, 5.5 mmol) in EtOH/H₂O (50:10 mL) was added lithium hydroxide (0.7 g, 16.5 mmol). The reaction was stirred at rt for 16 h. The mixture was quenched with water, adjusted pH to 3, extracted by EtOAc (50 mL × 3), dried over anhydrous Na₂SO₄, concentrated under vacuum, and purified by column chromatography to obtain the target product (1 g, 66.6%). MS [M + H]⁺ 272.

Step 4: Synthesis of (Z)-3-(3,4-Dimethoxy-5-nitrophenyl)-N,N-diethyl-2-fluoroacrylamide(5)

A solution of (Z)-3-(3,4-dimethoxy-5-nitrophenyl)-2-fluoroacrylic acid (1 g, 3.67 mmol), Et₃N (0.4 g, 5.51 mmol), and DIEA (1.43 g, 11.01 mmol) in DMF (15 mL) was stirred at 0 °C for 3 h. Then, HATU (2.1 g, 5.51 mmol) was added. The reaction was stirred at rt for 16 h. The mixture was concentrated under vacuum and purified by column chromatography to obtain the target product (1 g, 83.4%). MS [M + H]⁺ 327.

Step 5: Synthesis of (Z)-3-(3,4-Dihydroxy-5-nitrophenyl)-N,N-diethyl-2-fluoroacrylamide Cpd-3

To a solution of (Z)-3-(3,4-dimethoxy-5-nitrophenyl)-N,N-diethyl-2-fluoroacrylamide (200 mg, 0.61 mmol) in DCM (2 mL) was added BBr3 (0.6 mL, 6.1 mmol) slowly at 0 °C. The reaction mixture was stirred at rt for 16 h. Then, the mixture was quenched with water, adjusted pH to 7, extracted by EtOAc (50 mL × 3), dried over anhydrous Na₂SO₄, concentrated under vacuum, and purified by prep-HPLC to obtain the target product (42.7 mg, 23.3%). MS [M + H]⁺ 299, ¹H NMR (400 MHz, DMSO-d₆) δ 10.62 (s, 1H), 10.28 (s, 1H), 8.02 (s, 1H), 7.52 (s, 1H), 6.77(d, J = 37.3 Hz, 1H), 3.38 (d, J = 4.4 Hz, 4H), 1.11 (s, 6H). ¹³C NMR (101 MHz, DMSO-d₆) δ 161.17, 160.88, 152.33, 151.05, 149.59, 145.85, 139.16, 118.76, 116.97, 116.89, 112.28, 109.48, 109.44, 42.60, 14.46, 12.51. HRMS (ESI): calcd for C₁₃H₁₄FN₂O₅ [M – H]⁻: 297.0887; found 297.0892.

Synthesis of Cpd-4

Cpd-4 was prepared in two synthetic steps from 3,4-dihydroxy-5-nitrobenzaldehyde, according to the following procedure:

Step 1: Synthesis of (E)-2-Cyano-3-(3,4-dihydroxy-5-nitrophenyl)-N,N-diethylacrylamide(3)

To a solution of 3,4-dihydroxy-5-nitrobenzaldehyde (1 g, 5.5 mmol) in isopropyl alcohol (10 mL) was added 2-cyano-N,N-diethylacetamide (0.9 g, 6.5 mmol), followed by piperidine (0.23 g, 2.7 mmol). The reaction mixture was stirred at 85 °C for 16 h. The mixture was then concentrated under vacuum and purified by column chromatography to obtain the target product (1g, 62.5%). MS [M + H]⁺ 306.

Step 2: Synthesis of 2-Cyano-3-(3,4-dihydroxy-5-nitrophenyl)-N,N-diethylpropanamide (Cpd-4)

To a solution of 3 (300 mg, 0.98 mmol) in MeOH (5 mL) was added NaBH₄ (37 mg, 0.98 mmol). The reaction mixture was stirred at rt for 16 h. The mixture was concentrated under vacuum and purified by prep-HPLC to obtain the target product (101. 2 mg, 33.5%). MS [M + H]⁺ 308. ¹H NMR (400 MHz, DMSO-d₆) δ 10.12 (d, J = 37.5 Hz, 2H), 7.26 (d, J = 2.0 Hz, 1H), 6.99 (d, J = 2.1 Hz, 1H), 4.40 (t, J = 7.6 Hz, 1H), 3.30–3.23 (m, 2H), 3.19 - 3.08 (m, 2H), 2.99 (dd, J = 12.5, 7.6 Hz, 2H), 0.99 (t, J = 7.1 Hz, 3H), 0.93 (t, J = 7.1 Hz, 3H). ¹³C NMR (101 MHz, DMSO-d₆) δ 163.50, 147.45, 140.85, 136.86, 127.14, 121.20, 118.42, 115.14, 41.90, 40.35, 35.33, 34.65, 14.17, 12.57. HRMS (ESI): calcd for C₁₄H₁₆N₃O₅ [M – H]⁻: 306.1090; found 306.1091.

Synthesis of Cpd-5

Cpd-5 was prepared in two synthetic steps from 3,4-dihydroxy-5-nitrobenzaldehyde, according to the following procedure:

Step 1: Synthesis of 3-(3,4-Dihydroxy-5-nitrobenzylidene)pentane-2,4-dione(3)

To a solution of 3,4-dihydroxy-5-nitrobenzaldehyde (1 g, 5.5 mmol) in isopropyl alcohol (10 mL) was added acetylacetone (0.65 g, 6.5 mmol), followed by piperidine (0.23 g, 2.7 mmol). The reaction mixture was stirred at 85 °C for 16 h. The mixture was concentrated under vacuum and purified by column chromatography to obtain the target product (400 mg, 28.5%). MS [M + H]⁺ 266.

Step 2: Synthesis of 3-(3,4-Dihydroxy-5-nitrobenzyl)pentane-2,4-dione (Cpd-5)

To a solution of 3 (300 mg, 1.13 mmol) in ethanol (5 mL) was added NaBH₃CN (71.13 mg, 1.13 mmol). The reaction was stirred at rt for 16 h. The mixture was concentrated under vacuum and purified by prep-HPLC to obtain the target product (10.95 mg, 3.6%). MS [M – H]⁻ 266. ¹H NMR (400 MHz, DMSO-d₆) δ 10.03 (s, 2H), 7.17 (d, J = 2.0 Hz, 1H), 6.88 (d, J = 2.0 Hz, 1H), 4.21 (t, J = 7.4 Hz, 1H), 2.87 (d, J = 7.4 Hz, 2H), 2.10 (s, 6H). ¹³C NMR (101 MHz, DMSO-d₆) δ 203.97, 147.51, 140.53, 136.77, 129.25, 120.55, 114.28, 67.54, 31.91, 30.13. HRMS (ESI): calcd for C₁₂H₁₂NO₆ [M – H]⁻: 266.0665; found 266.0666.

Data Availability Statement

The docking poses of 11 candidates for ENT and NIT (Table S1) are available at https://www.huanglab.org.cn/for_submission_wp.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsptsci.3c00281.

• List of entacapone and nitecapone covalent docking candidates; features and performance of entacapone or its analogues in the UV–vis spectrometry assay and TSA; stabilization of NRF2 protein in HEK293T cells by entacapone; performance of nitecapone in the UV–vis spectrometry assay and its effect on stabilizing the NRF2 protein in HEK293T cells; activation of KEAP/NRF2 pathway after treatment with entacapone or nitecapone and the drug distribution in tissues; Western blot original figures; and HRMS, ¹H NMR, ¹³C NMR, and LC-MS spectra of representative compounds (PDF)

Author Contributions

^○ P.W. and Y.L. contributed equally to this work. The manuscript was written through the contributions of all authors. All authors have given approval to the final version of the manuscript.

Financial support from Beijing Municipal Science & Technology Commission (Z201100005320012 to N.H.) and Tsinghua University.

The authors declare no competing financial interest.