Covalent Modification of p53 by (E)-1-(4-Methylpiperazin-1-yl)-3-(5-nitrofuran-2-yl)prop-2-en-1-one

Abstract

TP53 is commonly mutated in cancer, giving rise to loss of wild-type tumor suppressor function and increases in gain-of-function oncogenic roles. Thus, inhibition of mutant p53 and reactivation of wild-type function represents a potential means to target diverse tumor types. (E)-1-(4-Methylpiperazin-1-yl)-3-(5-nitrofuran-2-yl)prop-2-en-1-one (NSC59984), first identified from a high-throughput screen, induces wild-type p53 signaling and antiproliferative effects while inhibiting mutant p53 gain-of-function activities. Here, we investigate the specific mechanism of action of NSC59984 against p53. We found that NSC59984 reacts with thiols via an unusual Michael addition at the α-carbon. Covalent modification of p53 Cys124 and Cys229 was observed both following in vitro reaction and upon treatment of cells. Finally, we used a biotinylated form of NSC59984 and, separately, thermal proteome profiling to examine off-target effects, identifying several metabolic proteins involved in cellular metabolism as potential targets. These results demonstrate that covalent modification of p53 by NSC59984 leads to increased wild-type activity and suggest that potential reaction with metabolic enzymes may contribute to antiproliferative function.

The tumor suppressor protein p53 plays critical roles in the stress response by initiating cell cycle arrest and apoptosis. Consistent with the importance of these functions, TP53 is the most frequently mutated gene in cancer, with ∼50% of human tumors bearing changes in the TP53 alleles.¹ Oncogenic TP53 mutations are most commonly found in the DNA binding domain, especially at certain high-frequency hotspot positions which correlate with poor cancer-free survival of patients.² p53 mutations are classified as either contact mutants, which lead to structural changes that affect DNA binding, or conformational mutants that affect more general protein folding. In addition to inhibiting wild-type p53, mutations can also confer a gain-of-function oncogenic phenotype.³ For example, mutant p53 has been shown to bind to noncanonical DNA sites in several genes and regulate their expression.⁴ Additionally, mutant p53 can bind p63 and p73 to negatively regulate their transcription factor activity.⁵ Thus, the development of therapeutics that can inhibit the gain-of-function effects of mutant p53 and restore wild-type p53 tumor suppressor activity is a promising means to treat a number of tumor types.

Several small molecule inhibitors have been developed to target mutant p53 and restore wild-type p53 function.⁶⁻¹¹ Among the first inhibitors of mutant p53 was PRIMA-1.¹² Currently, a methylated PRIMA-1 derivative termed PRIMA-1^Met (APR246) is being evaluated in various clinical trials. PRIMA-1 induces apoptosis in a mutant p53-dependent manner, restores wild-type p53 conformation, and results in binding of p53 to canonical target promoters.¹² Both PRIMA-1 and PRIMA-1^Met are converted intracellularly to methylene quinuclidinone which then modifies Cys124 and Cys277 of p53 via a covalent reaction.^13,14 Modification of Cys277 increases the thermal stability of wild-type p53, as well as mutant p53 R175H and R273H; modification of both sites was found to be important for restoration of wild-type p53 activity in cells harboring p53 R175H.¹⁴ PK11007 is a 2-sulfonylpyrimidine molecule that was shown to stabilize the DNA-binding domain of both wild-type and mutant p53 by covalent modification of Cys182 and Cys277.¹⁵ Treatment of cells harboring specific p53 mutations with PK1107 resulted in reduced viability and increased transcription of p53 target genes.¹⁵ Additionally, small molecule covalent modifiers of the conformational p53 mutant Y220C have recently been reported to restore wild-type thermal stability.¹⁶

(E)-1-(4-methylpiperazin-1-yl)-3-(5-nitrofuran-2-yl)prop-2-en-1-one (NSC59984) was identified from a library screen as a molecule that induced wild-type p53 signaling in mutant p53 cells.¹⁷ It has been shown to induce cell death in cells from different tumor types with diverse p53 mutations, increasing p53 binding to the promoters of canonical target genes and inducing their transcription.¹⁷⁻¹⁹ Additionally, we have shown that NSC59984 treatment of esophageal adenocarcinoma cells harboring the p53 R248W mutation restored wild-type metabolic profiles.¹⁹ This activity was specific for p53 R248W, as esophageal adenocarcinoma cells harboring p53 R175H were not affected by NSC59984.¹⁹ However, the molecular mechanism by which NSC59984 reactivates mutant p53 remains unclear. Recently, NSC59984 was reported to induce mutant p53 degradation through the induction of ERK phosphorylation in a ROS-dependent manner.¹⁸ However, potential direct effects on p53 have not been explored.

Here, we have used biophysical and mass spectrometry approaches to investigate the mechanism of action of NSC59984. NSC59984 is unique among mutant p53 inhibitors in that it is readily amenable to modification with a biotin group, allowing the generation of a reagent that can be used to explore the mechanism of the inhibitor. Similar to other reactivators of mutant p53, we show that NSC59984 covalently modifies p53 on specific cysteine residues in the DNA binding domain. Molecular simulations predict that these modifications affect the conformation of the domain and stabilize conformations that mimic wild-type activity. Further, we examined possible off-target effects of NSC59984 and identified potential targets among metabolic enzymes.

Results

NSC59984 Reaction with Thiols

The molecule NSC59984 has been reported to inhibit mutant p53 activity in tumors.¹⁷ However, its mechanism of action has never been explicitly defined nor experimentally tested. NSC59984 is a simple α,β-unsaturated amide with a nitrofuran attached to the β-carbon and an N-methyl piperazine amide attached to the α-carbon. The expected reactivity for this molecule would be for nucleophiles to react at the β-carbon via a Michael addition reaction. In cells, the most likely nucleophiles that would react with NSC59984 are free thiols associated with glutathione, free cysteine, or cysteine side chains of proteins. Related molecules, such as 1 (Figure 1A), should also react in a similar pattern. Indeed, we found that 1 also has antiproliferative and pro-apoptotic activity comparable to NSC59984 (Figure S1). To directly test the reactivity of NSC59984, we combined the molecule with either glutathione, N-acetyl cysteine, or MESNA (2, a simple water-soluble thiol). When NSC59984 reacted with glutathione or N-acetyl cysteine in aqueous buffer at pH 7.4, two separable peaks were observable by HPLC in each case (Figure S2). Both peaks were isolated by HPLC, followed by NMR analysis, which showed, surprisingly, that the products (Figures 1A, 3, and 4) had the thiol attached to the α-carbon; the two peaks in the HPLC represent the diastereomers 3 and 4 (see Supporting Information, Compound Characterization). When NSC59984 reacted with MESNA, there was only one product peak observed by HPLC, which was expected since diastereomers are not formed in this reaction (Figure S2). The product from this reaction was characterized by NMR as 5 (Figure 1A), which again showed the sulfur attached to the α-carbon. This mode of reactivity was very surprising and contrary to the expected Michael addition reaction that should take place at the β-carbon.

Figure 1.

NSC59984 reacts with free thiols via a Michael addition to the α-carbon. (A) Chemical structures of p53-reactivating molecules (NSC59984 and 1), MESNA (Sodium 2-mercaptoethanesulfonate (2)), adduct of Glutathione with NSC59984 (3), adduct of N-acetylcysteine with NSC59984 (4), and adduct of MESNA with NSC59984 (5). Potential sites for nucleophilic attack are labeled as α and β on the NSC59984 structure. The * label in 3, 4, and 5 represents the formation of a new stereogenic center after nucleophilic attack at the α carbon of NSC59984. Molecules 3 and 4 are a mixture of diastereomers, and 5 is a mixture of enantiomers. (B) Proposed scheme describing the reaction between 1 and 2. The * label represents the formation of a new stereogenic center after nucleophilic attack at the α carbon of 1. (C) ¹H NMR analysis of time-dependent reaction of model compounds 1 and 2, showing thiol modification. Amplitude was adjusted to reduce signal-to-noise ratio in the stack plot so that specific peaks are more visible. (D) Quantum mechanical calculations of the LUMO for 1 (upper) or NSC59984 (lower) explain the pattern of reactivity.

Figure 3.

Modeling of potential structural effects of NSC59984 modification of p53 R248W. (A) Interactions of the Cys124-adduct with the loop of the L1/S3 pocket. In both panels the protein is represented as gray ribbon; side chains of Lys120 and Cys124 are shown as sticks; nitrogen, oxygen, sulfur and hydrogen atoms are colored blue, red, yellow and white, respectively; the carbons of the Cys124-adduct are colored magenta, and the carbons of Lys120 are colored cyan. (B) Cys124 and Cys229 adducts at the parallel interface of p53DBD monomers. The figure was made by superimposing trajectory snapshots onto the C (blue) and A (gray) subunits of a crystal structure of the p53 tetramer bound to DNA.²² The Cys229 and Cys124 adducts, carbons colored purple and magenta, respectively, are part of the top protein monomer. Panel (B-b) shows the Cys229-adduct making two possible hydrogen bonds represented by yellow lines. Panel (B-c) shows the Cys124-adduct making three possible, stabilizing interactions with the loop immediately following S4. The formal +1 charged nitrogen and the partially negatively charged hydroxyl oxygen of Ser166 are indicated by plus and minus signs.

Figure 4.

Effects of biotinylated NSC59984. (A) Analysis of cellular proliferation was determined by fold change in CyQUANT measurement of cellular DNA following treatment with NSC59984-biotin (12 μM) for 72 h over carrier control in CP-A-WT or ESO26-R248W cells. Fold change in carrier-treated cells was normalized to 1. Statistical analysis was performed using a 2-way ANOVA test with Tukey correction. (B) Analysis of cellular proliferation was determined as in (A) following treatment with NSC59984 or NSC59984-biotin (12 μM) for 72 h over carrier control in ESO26-R248W cells. Fold change in carrier-treated cells was normalized to 1. Statistical analysis was performed using a 2-way ANOVA test with Tukey correction. (C) Western blot analysis of p53-biotin protein levels. CP-A-WT or ESO26 cells were treated with NSC59984-biotin (12 μM) for 72 h before cell lysis. Total protein was immunoprecipitated with Streptavidin Mag Sepharose beads and blotted for p53. (D) ESO26-R248W cells were treated with NSC59984-biotin (12 μM) for 2 h. Fixed cells were stained with streptavidin-AF488 (green), and counter stained for DNA (Hoechst, Blue). Images were captured on a Zeiss LSM710 confocal microscope. (E) ESO26-R248W cells were treated with NSC59984 (12 μM) for 12 h as described in panel (D).

To study this reaction in more detail, 1 and its reaction with 2 were used as a simpler system to facilitate analysis. Under the same reaction conditions, the only product isolated from the reaction was 6 (Figure 1B), where the sulfur is attached to the α-carbon. NMR characterization of this product clearly established the connectivity of the CH₂ to the nitrofuran. The reaction between 1 and 2 was also directly monitored by NMR (Figure 1C) by acquiring spectra over time as the reaction proceeds in deuterated aqueous, phosphate buffer, pH 7, at 37 °C. After 5 min, the starting material (1) is mostly consumed and there are some peaks that represent the formation of the product 6. However, there are numerous other peaks that are clearly visible. The chemical shifts of these peaks are consistent with resonances of alkene protons. After 1 h, there is no more starting material and the intermediate peaks are diminished. This trend continues until only the product 6 is clearly present, which is the only product isolated by HPLC. This interesting pattern of reactivity is probably due to the presence of the nitro group on the furan ring. The strong electron-withdrawing properties of the nitro group on the furan ring mostly likely play a role in directing nucleophilic attack to carbons other than the β-carbon in NSC59984 and 1.

Quantum mechanical calculations were performed with Gaussian to determine the LUMO of 1 as well as NSC59984 (Figure 1D). The molecular orbitals in the LUMOs clearly show density at the α-carbon, and smaller density at the β-carbon by comparison. This difference in molecular orbital density may explain why only the products with nucleophiles attached at the α-carbon are observed. The LUMOs also show substantial density at two carbons in the nitrofuran ring. Nucleophilic attack of 1 at either of these carbons by 2 would yield molecules with structures shown in 7 and 8. We speculate that these are the intermediates observed by NMR. Both 7 and 8 possess two sets of diastereotopic alkene protons, which are consistent with the resonances observed by NMR. While these intermediates may form initially, they likely revert to the starting materials, and ultimately the product of nucleophilic attack at the α-carbon, as observed by NMR. The initial formation of 7 and 8 likely reflects the kinetic products that form when a thiol nucleophile (such as 2) reacts with 1. These kinetic products may arise because nucleophilic attack at the furan ring is less sterically hindered compared to attack at the α-carbon. However, 7 and 8 lack a fully conjugated aromatic system which may result in a lack of stability leading to cleavage of the newly formed carbon–sulfur bonds and reformation of the original starting materials. The eventual formation of 6 likely represents the most thermodynamically stable product as it contains a fully aromatic furan ring. The unique LUMOs of 1 and NSC59984 appear to be directing the reaction of these molecules with 2 as well as controlling the reactivity with biological thiols. The ability of NSC59984 to readily react with biological thiols likely explains the reduction in glutathione levels previously observed in cells treated with this molecule.¹⁹

NSC59984 Reaction with p53

Having established that NSC59984 reacts with sulfur-containing molecules, we next investigated whether NSC59984 could react with cysteine-containing peptides from p53. As p53 mutation hot-spots tend to occur in the DNA-binding domain (DBD), we focused on five cysteine residues in that region: Cys124, Cys182, Cys229, Cys275, and Cys277. Previous studies have shown that Cys277 is the primary site of reaction with APR-246, with the reaction observed at Cys124 and the peptide containing Cys275/Cys277.¹⁴ PK11000 reacts with Cys182 and peptides containing Cys275/Cys277.¹⁵ Four synthetic peptides containing these sites were incubated with 10-fold or 100-fold molar excess of NSC59984 and mass spectrometry used to analyze any reaction products. When the Cys124 peptide was incubated with a 10-fold molar excess of NSC59984, a new mass was observed consistent with covalent modification of the peptide with NSC59984 (Figure S3A); the intensity of that new mass increased when reacted with 100-fold molar excess NSC59984 (Figure 2A). Reaction with the Cys182 peptide was observed with a shifted retention time at the higher concentration, but not at the lower concentration (Figure 2A). Covalent modification of Cys229 was observed at both NSC59984 concentrations, whereas the C275/C277 peptide did not react with NSC59984 at either concentration (Figure 2A, Figure S3A). Thus, in vitro, NSC59984 was able to react with peptides comprising Cys124 or Cys229, as well as a slight reaction with Cys182, but no reaction with Cys275/Cys277. To the best of our knowledge, this represents the first identification of a small molecule reacting with p53 Cys229.

Figure 2.

NSC59984 reacts with p53, covalently modifying cysteine residues. (A) Reconstructed mass spectra showing reaction of p53 peptides with 100-fold molar excess of NSC59984 for 3 h at 37 °C. Blue spectra are DMSO-treated controls and orange spectra are NSC59984 treated. In the panel for Cys182, the inset spectrum corresponds to the retention time of the control peptide without treatment. In the panel for Cys275, the dashed lines indicate the expected mass of singly- or doubly modified peptide. (B) MS/MS fragmentation spectra of the p53 R248W chymotryptic peptides 115–137 (upper) and 213–236 (lower) with masses consistent with addition of NSC59984 on Cys124 or Cys229, respectively. In both panels, the most intense b- and y-ions are colored with red and blue, respectively; precursor ion is shown in green. Fragment ions that contain the modified cysteine residue are indicated with yellow highlighting. In the peptide sequences, the ions detected are indicated.

Having identified in vitro sites of modification, we investigated whether NSC59984 also modified p53 in cells. We isolated p53 from NSC59984-treated ESO26 cells which have mutant p53-R248W¹⁹ to determine if covalent modification could be detected. Mass spectrometry analysis of the immunoprecipitated p53 identified peptides with adducts consistent with the addition of NSC59984 on Cys124, Cys229, and Trp146 (Figure 2B, Figure S3B). Estimation of the percent modified indicates that ∼20% of p53 was modified on either cysteine site. The cysteine adducts are consistent with our in vitro results using purified peptides. Since in vitro incubation of NSC59984 with tryptophan did not yield a reaction product (data not shown), we hypothesize that this modification results from an intramolecular transfer from a modified cysteine to Trp146 via a cycloaddition reaction (Figure S3C).²⁰

Molecular dynamics simulations were used to investigate how the adduct could affect the conformation and interactions of the p53 DBD. They revealed two main modes in which the adduct on Cys124 sterically affects the conformation of the L1 loop (Figure 3A). In Figure 3A-a, L1 wraps around the two branched lobes of the adduct. This keeps L1 in the vicinity of the “extended conformation” in which the side chain of Lys120 can bind a backbone phosphate of complexed DNA.^21,22 In Figure 3A-b, L1 straddles the two branched lobes of the adduct and adopts a range of intermediate conformations. These modifications influence the conformation of the druggable L1/S3 pocket, which can stabilize and reactivate the protein.²³ Using strict geometric criteria,²³ the Cys124 adduct increases the proportion of the active, open population of the L1/S3 pocket from 11 to 36%. As detailed in Table S1, the cysteine and tryptophan adducts were also found to form a number of long-lived hydrogen bonds within the DBD that could stabilize the conformation. Finally, the adducts could also affect the stability of the multimeric complex of the DBDs. As seen in Figure 3B-a, Cys229 and Cys124 occur at the interface of each parallel-aligned pair of p53 DBD subunits in the tetrameric complex bound to double-stranded DNA²² (only one pair of subunits is shown for clarity). Figure 3B-b details the interface around the Cys229-adduct, showing two possible hydrogen bonds with Gln100 and Lys101 of the N-terminus of the adjacent p53 DBD subunit. Figure 3B-c shows the Cys124-adduct making three possible stabilizing interactions with the loop immediately following S4: hydrogen bonds with Lys164 and Gln165, and an electrostatic attraction between the formal +1 charged nitrogen and the partially negatively charged hydroxyl oxygen of Ser166. Thus, the empirical increase in activity may be due to the adducts stabilizing native-like conformations of both the DBD monomer and the tetrameric complex that binds DNA.

Characterization of NSC59984 Off-Target Effects

Our observation of reaction of NSC59984 with glutathione and N-acetyl cysteine in vitro suggests that it is unlikely that p53 is the sole reaction target of this molecule in cells. We synthesized biotinylated NSC59984 (NSC59984-biotin, Figure S4A) to use in unbiased pulldown experiments to identify other potential targets. First, we established that biotinylation did not substantially affect the activity of NSC59984. Like NSC59984, NSC59984-biotin inhibited proliferation and promoted apoptosis only in ESO26 cells and not in CP-A-WT cells harboring wild-type p53 (Figure 4A, Figure S4B). Further, treatment of ESO26 cells with NSC59984-biotin showed similar decreased proliferation as observed for NSC59984 (Figure 4B). Western blots of streptavidin pulldowns from treated cells clearly showed that p53 was bound by NSC59984-biotin (Figure 4C). Incubation of NSC59984-biotin with the p53 cysteine-containing peptides resulted in covalent modification similar to that observed for NSC59984 (Figure S4C). Thus, the addition of the biotin moiety did not change the activity of NSC59984. Confocal microscopy was next used to characterize the uptake and distribution of NSC59984-biotin. After 2 h, NSC59984-biotin was observed intracellularly but dispersed, whereas by 12 h it was observed in the cytosol around the nucleus of the cells (Figure 4D,E). After 72 h, no further NSC59984-biotin was observed intracellularly (Figure S4D). We therefore focused on the identification of NSC59984-biotin targets 12 h after treatment.

We used NSC59984-biotin to identify potential reaction targets in two ways. First, ESO26 cells were treated with NSC59984-biotin or DMSO vehicle control and the reaction targets pulled down on streptavidin resin (Direct Method). Second, ESO26 cells were pretreated with either DMSO vehicle control or NSC59984 for 24 h, followed by treatment with NSC59984-biotin and streptavidin pulldown (Competition Method). In the Competition Method, reaction targets would be modified first by NSC59984, preventing the biotinylated form from reacting and causing a decrease in binding compared to the DMSO control. Proteins identified as 5-fold enriched by NSC59984-biotin over DMSO treatment in the Direct Method experiment and 2-fold enriched in the DMSO pretreatment over NSC59984 treatment in the Competition Method experiments were considered to be targets. Using these criteria, we identified four proteins in at least two replicate pulldown experiments (Table 1). p53 was not observed in these experiments, likely due to its low abundance in the cell. One of the proteins identified is PRDX4, a peroxiredoxin protein that catalyzes the reduction of hydrogen peroxide and organic hydroperoxides to water and alcohols, respectively.

Table 1. Proteins Identified to Pulled Down Upon Treatment with NSC59984-biotin.

accession	gene	description	ESO NSC/DMSO direct rep 1	ESO DMSO/NSC competition rep 1	ESO NSC/DMSO direct rep 2	ESO DMSO/NSC competition rep 2
P00352	ALDH1A1	retinal dehydrogenase 1	16.01	3.68	18.76	4.27
O60610	DIAPH1	protein diaphanous homologue 1	100.00	100.00	100.00	100.00
Q13162	PRDX4	peroxiredoxin-4	6.10	40.23	100.00	2.67
Q01650	SLC7A5	large neutral amino acids transporter small subunit 1	100.00	100.00	100.00	100.00

We observed a high background of nonspecific interactors in NSC59984-biotin-untreated samples in these experiments, likely due to endogenous biotinylation. Thus, we performed thermal proteome profiling (TPP) experiments to look for additional targets of NSC59984 as an orthogonal approach. In these experiments, ESO26 cells were treated with DMSO or NSC59984, then changes in the thermal stability of proteins were measured by quantitative mass spectrometry (Figure S5A). From two biological replicate experiments, 2177 proteins were identified and quantified in both. As observed from the distribution of Euclidian distances (Figure S5B), most proteins showed a similar response to DMSO and NSC59984 treatment. Ten proteins were found to reproducibly have shifted melting temperatures upon NSC59984 treatment with adjusted p-value < 0.05 (Figure 5). Of those, five have functions related to cell death and/or metabolism and mitochondrial function – FAM136A, OPA1, LCN2, ANKHD1, and TIGAR. These were of particular interest as increased apoptosis and metabolic alterations have both been shown to occur in cells treated with NSC59984.¹⁹ We have recently shown that NSC59984 treatment in ESO26 cells leads to decreased activity of phosphofructokinase 1, which is inhibited by TIGAR; further, knockdown of TIGAR decreased the antiproliferative effects of NSC59984, consistent with a role of this protein in NSC59984 activity.¹⁹ These effects are consistent with a specific effect of NSC59984 on TIGAR in addition to its effects on mutant p53.

Figure 5.

Thermal proteome profiling of NSC59984 targets. Thermal melting curves for selected proteins found to be significantly (adjp < 0.05) affected by NSC59984 treatment of ESO cells. Proteins are grouped by molecular function. Curves in light/dark blue correspond to the DMSO-treated samples and curves in red/brown the NSC59984-treated samples. p-Values were calculated from z-scores generated from the distribution of Euclidean distances output by the mineCETSA R package using the Robust Z-Score method; a Benjamini–Hochberg multiple hypothesis correction was applied.

We next investigated whether NSC59984 modulated TIGAR stability via direct reactivity or whether its altered thermal stability was due to indirect effects. Purified recombinant TIGAR was incubated with NSC59984 in vitro for 30 min and then potential reaction analyzed by mass spectrometry. We identified four sites of NSC59984 reaction on cysteine residues of TIGAR: Cys114, Cys161, Cys254, and Cys256 (Figure S6). Quantitation of the amount of modified peptide for these sites showed that ∼5–10% of TIGAR protein was modified on a given site (Table 2). Thus, it is possible for NSC59984 to react with TIGAR, suggesting potential effects of this inhibitor beyond mutant p53.

Table 2. TIGAR Peptides Modified by NSC59984 In Vitro.

peptide	residue	% modified
AAREECPVFTPPGGETLDQVK	Cys114	9.8%
EECPVFTPPGGETLDQVK	Cys114	4.7%
EADQKEQFSQGSPSNCLETSLAEIFPLGK	Cys161	9.3%
EQFSQGSPSNCLETSLAEIFPLGK	Cys161	5.1%
EVKPTVQCICMNLQDHLNGLTETR	Cys254	6.8%
EVKPTVQCICMNLQDHLNGLTETR	Cys256	2.9%

Discussion

The p53 tumor suppressor is an essential protein for the protection of genomic stability. As such, it is frequently mutated or inactivated in tumors. Reactivation of mutant p53, therefore, represents a potential strategy for the development of new chemotherapeutics. NSC59984 was identified in a screen as a reactivator of mutant p53, and studies into its effects have demonstrated that it leads to increased transcription of p53 targets, increased p63 transcriptional activity, and apoptosis. Further, we recently found that NSC59984 treatment of esophageal adenocarcinoma cells bearing p53 R248W resulted in a reversion of altered cellular metabolism to a more normal state.¹⁹ However, despite this understanding of the effects of NSC59984, its specific mechanism of action with regard to p53 had not been previously studied. We further observed that the structure of NSC59984 is amenable to biotinylation without affecting the activity, allowing for development of a tool to study the mechanism of action. Thus, we set out to investigate the reaction of NSC59984 with p53 and potential off-target effects.

We show here that NSC59984 covalently modifies p53 on Cys124 and Cys229 in treated cells. Cys124 is commonly observed to be targeted by reactivating compounds,¹⁵ but Cys229 has not previously been observed to be a targeted site to the best of our knowledge. Analysis of the reaction of NSC59984 with thiol analogs in vitro indicates that reaction with cysteine residues in p53 occurs through an unusual Michael addition at the α-carbon. This interesting pattern of reactivity is likely due to the presence of the nitro group on the furan ring, with the strong electron-withdrawing properties of the nitro group directing nucleophilic attack to carbons other than the β-carbon in NSC59984. In the crystal structure of the p53 core domain (residues 94–289) bound to DNA, Cys124 and Cys229 are surface exposed and therefore accessible for reaction.^24,25 Intriguingly, NSC59984 appears to exhibit some reaction specificity with particular cysteine residues. Cys182 and Cys277 have been reported to be the most solvent accessible of the thiols in the p53 core domain,²⁵ yet Cys182 is only slowly modified by NSC59984 in vitro and Cys277 is nonreactive. This suggests that NSC59984 is not simply reacting with the most accessible sites and that some element of preferential reaction exists. Our findings suggest that tuning the cysteine-specificity of NSC59984 may be a viable avenue for future drug design efforts.

Mechanism studies on other reactivators also found that reaction generally occurs on cysteine residues of p53 through a canonical Michael addition, with PK11007 reacting by nucleophilic aromatic substitution.²⁶ Modification of wild-type or mutant p53 with methylene quinuclidinone increases the thermostability of the p53 core domain by 2–3 degrees;¹⁴ likewise, PK11007 also thermally stabilizes p53.¹⁵ Recently reported p53 Y220C-specific covalent compounds have been shown to thermally stabilize p53 Y220C, a mutation that particularly results in a loss in thermal stability of mutant p53.¹⁶ Our modeling results indicate that modification of these cysteine residues would stabilize the DNA-binding domain of p53 in both the monomeric and tetrameric conformations, enhancing p53 transcriptional activity.

As covalent drugs irreversibly modify their targets, it is critical to understand potential off-target effects of these molecules. We have used two orthogonal approaches to investigate other proteins that could be reaction targets of NSC59984–pulldowns using a biotinylated form of the inhibitor and thermal proteome profiling. Although the two methods indicated different targets, both suggest that NSC59984 has effects on metabolic enzymes. A similar thermal proteome profiling method used to investigate potential targets of methylene quinuclidinone identified asparagine synthetase (ASNS) as significantly affected by treatment.²⁷ ASNS is a metabolic enzyme that catalyzes the synthesis of asparagine from aspartate and glutamine.²⁸ PRIMA-1^Met has further been shown to inhibit thioredoxin reductase 1 (TXNRD1) by targeting the selenocysteine in the catalytic domain, thereby depleting cells of reduced glutathione and increasing reactive oxygen species to induce p53-independent cell death.^28,29 Thus, PRIMA-1^Met also targets proteins involved in cellular metabolism, as we observe for NSC59984.

NSC59984 is being developed as a reactivator of mutant p53 for cancer treatment. Our results indicate that this molecule exerts its effects on p53 through covalent modification of specific cysteine residues which leads to stabilization of the DNA-binding domain. Using two distinct methods to look for potential off-target effects, we observed only a small number of affected proteins, suggesting specificity in the reaction with p53. Among those, we found that NSC59984 can covalently modify TIGAR in vitro, suggesting that this is a potential direct effect. Our thermal stability analysis showed that NSC59984 stabilizes TIGAR, consistent with our previous findings that TIGAR levels and activity were increased by NSC59984 treatment.¹⁹ Thus, effects on TIGAR may cooperate with effects on p53 to direct tumor cells to a more “normal” metabolic environment while restoring more normal p53 tumor suppressor activity.

Materials and Methods

Chemical Synthesis

The syntheses of NSC59984 and compound 1 are outlined in Scheme 1. Using HATU as the coupling agent, amide formation between 5-nitrofuran-2-acrylic acid and N-methyl-piperazine afforded NSC59984. Compound 1 was synthesized by the reaction of 5-nitrofuran-2-acrylic acid with a dimethylamine using carbonyl diimidazole (CDI) as the coupling agent.

Scheme 1. (i) HATU, DIPEA, N-Methyl-piperazine, Anhydrous DMF, RT; (ii). CDI, DIPEA, 2.0 M Dimethylamine Solution in THF, Anhydrous THF, RT.

Reactions were conducted under a nitrogen atmosphere. Commercial solvents and reagents were purchased from Millipore Sigma, except for N-methyl-piperazine (Oakwood Products, Inc.) and Mesna (Combi-Blocks, Inc.). HPLC separation was performed on an Agilent (Santa Clara, CA) 1260 Infinity II Series RP-HPLC with automatic fraction collection using ultraviolet detection at 260 nm. Waters (Milford, MA, USA) XBridge BEH C18 OBD Prep columns (130 Å, 5 μm, 10 × 250 mm) were used in conjunction with solvents A and B for purification at 45 °C. Solvent A was 0.1% TFA in water and solvent B consisted of 90% acetonitrile in water.

Synthesis of 1-(4-Methylpiperazin-1-yl)-3-(5-nitrofuran-2-yl)prop-2-en-1-one (NSC59984)

In a flame-dried flask, 5-nitrofuran-2-acrylic acid (355 mg, 1.94 mmol) was dissolved in anhydrous DMF (5.0 mL) under an argon atmosphere. HATU (737 mg, 1.94 mmol), DIPEA (0.68 mL, 3.88 mmol), and N-methyl-piperazine (0.22 mL, 1.94 mmol) were added sequentially to the reaction mixture, which was stirred at room temperature for 20 h. After solvent removal by evaporation, the crude mixture was diluted with H₂O (20.0 mL), transferred to a separation funnel, and extracted with dichloromethane (40 mL × 8). The organic phase was dried over Na₂SO₄, filtered, and concentrated. The crude material was then purified by flash chromatography using a gradient 0 to 10% MeOH in DCM to obtain the desired product as an orange solid (236 mg, 0.89 mmol, 46%).

¹H NMR (400 MHz, DMSO-d₆) δ 7.77 (d, 1H, J = 3.9 Hz), 7.38 (d, 1H, J = 15.4 Hz), 7.30 (d, 1H, J = 15.4 Hz), 7.26 (d, 1H, J = 3.9 Hz), 3.65 (bs, 2H), 3.57 (bs, 2H), 2.33 (bs, 4H), 2.20 (s, 3H). ¹³C NMR (125 MHz, DMSO-d₆) δ 162.94, 153.73, 151.57, 126.99, 122.30, 115.63, 114.90, 54.96, 54.19, 45.47, 45.05, 41.64. HRMS–ES (m/z): [M + H]⁺ calcd. for C₁₂H₁₅N₃O₄ 266.1141, found 266.1136.

Synthesis of N,N-Dimethyl-3-(5-nitrofuran-2-yl)acrylamide 1

In a flame-dried flask, 5-nitrofuran-2-acrylic acid (100 mg, 0.55 mmol) was dissolved in anhydrous THF (2.0 mL) under an argon atmosphere. Carbonyl diimidazole (88.5 mg, 0.55 mmol) was added at room temperature, and the mixture was stirred for 10 min before adding dropwise a 2.0 M solution of dimethylamine in THF (0.27 mL, 0.55 mmol). The resulting mixture was stirred at room temperature for 20 h. After solvent removal by evaporation, the crude material was diluted with dichloromethane (20 mL), transferred to a separation funnel, and washed with aq. sat. NaHCO₃ (20 mL × 1), H₂O (20.0 mL × 1), and brine (20.0 mL × 1). The organic phase was dried over Na₂SO₄, filtered, and concentrated. The crude material was then purified by flash chromatography using a gradient of 0 to 5% MeOH in DCM to obtain the desired product as a yellowish solid (80.3 mg, 0.38 mmol, 70%).

¹H NMR (400 MHz, DMSO-d₆) δ 7.77 (d, 1H, J = 3.9 Hz), 7.35 (d, 1H, J = 15.3 Hz), 7.25 (d, 1H, J = 3.8 Hz), 7.23 (d, 1H, J = 15.4 Hz), 3.15 (s, 3H), 2.94 (s, 3H). ¹³C NMR (125 MHz, DMSO-d₆) δ 162.94, 153.73, 151.57, 126.99, 122.30, 115.63, 114.90, 54.96, 54.19, 45.47, 45.05, 41.64. HRMS–ES (m/z): [M + H]⁺ calcd. for C₉H₁₀N₂O₄ 266.1141, found 266.1136.

NMR Experiments

NSC59984 and l-Glutathione

NSC59984 (5.6 mg, 0.02 mmol) and l-glutathione (32.4 mg, 0.10 mmol) were dissolved in PBS buffer 10X (5.0 mL) at pH = 7.4 in a 20 mL vial with stirring. The reaction was heated at 36.4 °C for 24 h. The formation of adducts was monitored by mass spectrometry. After cooling, the reaction was put in dry ice and lyophilized. The two diastereomers 3 were separated by HPLC.

¹H NMR (400 MHz, D₂O) δ 7.50 (d, 0.5H, J = 8.6 Hz), 7.49 (d, 0.5H, J = 8.6 Hz), 6.59 (d, 0.5H, J = 9.9 Hz), 6.58 (d, 0.5H, J = 9.9 Hz), 4.64–4.55 (m, 2H), 4.45 (app q, 1H, J = 7.5 Hz), 4.36–4.32 (m, 1H), 3.93 (s, 2H), 3.78 (t, 1H, J = 6.3 Hz), 3.68–3.58 (m, 2H), 3.56–3.19 (m, 4H), 3.16–3.05 (m, 3H), 2.96 (s, 3H), 2.93–2.80 (m, 1H), 2.56–2.47 (m, 2H), 2.18–2.11 (m, 2H). ¹³C NMR (125 MHz, D₂O) δ 174.64, 173.75, 173.60, 173.55, 171.90, 170.19, 170.13, 156.89, 156.75, 151.31, 114.84, 114.67, 112.29, 112.22, 53.69, 53.02, 52.90, 52.78, 42.99, 42.91, 42.84, 42.76, 41.64, 41.12, 40.49, 39.49, 39.39, 31.18, 31.11, 30.65, 30.60, 30.54, 30.36, 26.00, 25.95. HPLC RT: 10.68 min.

¹H NMR (400 MHz, D₂O) δ 7.50 (d, 0.5H, J = 7.2 Hz), 7.49 (d, 0.5H, J = 7.1 Hz), 6.61–6.59 (m, 1H), 4.63–4.60 (m, 1H), 4.53–4.32 (m, 3H), 3.92 (s, 2H), 3.78 (t, 1H, J = 6.2 Hz), 3.65–3.05 (m, 9H), 2.96 (s, 3H), 2.96–2.88 (m, 1H), 2.54–2.50 (m, 2H), 2.18–2.12 (m, 2H). ¹³C NMR (125 MHz, D₂O) δ 174.68, 174.62, 173.89, 173.62, 173.58, 171.88, 171.80, 170.21, 163.13, 162.85, 157.00, 156.82, 151.31, 117.41, 115.09, 114.83, 114.66, 112.34, 112.23, 53.72, 53.17, 52.84, 52.81, 52.66, 43.02, 42.79, 41.74, 40.59, 39.84, 39.50, 39.34, 31.22, 30.06, 29.97, 26.02, 25.99. HPLC RT: 11.24 min.

NSC59984 and N-Acetyl-l-cysteine

NSC59984 (5.3 mg, 0.02 mmol) and N-acetyl-l-cysteine (16.3 mg, 0.10 mmol) were dissolved in PBS buffer 10X (5.0 mL) at pH = 7.4 in a 20 mL vial with stirring. The reaction was heated at 36.4 °C for 24 h. Formation of adducts was monitored by mass spectrometry. After cooling, the reaction was put in dry ice and lyophilized. The two diastereomers were separated by HPLC.

¹H NMR (500 MHz, D₂O) δ 7.48 (d, 0.5H, J = 11.2 Hz), 7.47 (d, 0.5H, J = 11.3 Hz), 6.57 (d, 1H, J = 12.0 Hz), 6.56 (d, 1H, J = 12.0 Hz), 4.61–4.55 (m, 1H), 4.48–4.36 (m, 2H), 4.33–4.30 (m, 1H), 3.67–3.03 (m, 9H), 2.95 (s, 1.7H), 2.94 (s, 1.3H), 2.89–2.83 (m, 1H), 2.01 (s, 3H). ¹³C NMR (125 MHz, D₂O) δ 174.83, 173.74, 173.66, 170.17, 170.04, 163.41, 163.13, 162.84, 157.00, 151.29, 117.42, 115.09, 114.86, 114.68, 112.25, 112.19, 53.42, 52.85, 52.78, 42.95, 42.88, 42.69, 40.64, 40.37, 39.41, 30.99, 30.68, 30.50, 30.11, 21.70. HPLC RT: 14.29 min.

¹H NMR (500 MHz, D₂O) δ 7.48 (d, 0.5H, J = 9.4 Hz), 7.47 (d, 0.5H, J = 9.2 Hz), 6.59–6.56 (m, 1H), 4.67–4.57 (m, 1H), 4.38–4.33 (m, 2H), 3.65–3.56 (m, 2H), 3.55–3.49 (m, 1H), 3.46–3.35 (m, 1H), 3.29–3.04 (m, 5H), 2.94 (s, 3H), 2.01 (s, 3H). ¹³C NMR (125 MHz, D₂O) δ 175.45, 173.62, 170.34, 157.13, 157.00, 151.28, 114.87, 114.70, 112.33, 112.21, 54.01, 53.73, 52.88, 52.82, 52.72, 43.01, 42.92, 42.83, 40.75, 40.11, 39.48, 39.37, 31.24, 31.17, 30.32, 30.20, 21.76. HPLC RT: 14.86 min.

NSC59984 and Mesna

NSC59984 (20.0 mg, 0.08 mmol) and Mesna (61.9 mg, 0.10 mmol) were dissolved in PBS buffer 10× (5.0 mL) at pH = 7.4 in a 20 mL vial with stirring. The reaction was heated at 36.4 °C for 24 h. The formation of adducts was monitored by mass spectrometry. After cooling, the reaction was put in dry ice and lyophilized. The crude material was purified by HPLC.

¹H NMR (600 MHz, D₂O) δ 7.48 (d, 0.5H, J = 14.0 Hz), 7.47 (d, 0.5H, J = 14.2 Hz), 6.59 (d, 0.5H, J = 11.2 Hz), 6.59 (d, 0.5H, J = 11.0 Hz), 4.64–4.58 (m, 1H), 4.45 (t, 0.5H, J = 7.4 Hz), 4.41 (t, 0.5H, J = 7.3 Hz), 4.37–4.34 (m, 1H), 3.71–3.66 (m, 1H), 3.66–3.59 (m, 2H), 3.55–3.50 (m, 1H), 3.48–3.38 (m, 1H), 3.36–3.20 (m, 2H), 3.17–3.04 (m, 4H), 3.01–2.96 (m, 1H), 2.95 (s, 3H), 2.91–2.85 (m, 1H). ¹³C NMR (150 MHz, D₂O) δ 170.43, 170.29, 157.19, 157.12, 151.31, 114.97, 114.79, 112.32, 112.24, 52.92, 52.82, 50.67, 50.34, 43.01, 42.88, 42.76, 40.94, 40.44, 39.47, 30.36, 30.03, 23.80, 23.62. HPLC RT: 10.86 min.

LUMO Quantum Mechanical Calculations

The molecular geometry was optimized with Gaussian 16-A03 (Gaussian, Inc.) using a DFT B3LYP hybrid functional³⁰ with an aug-cc-pVDZ basis set and a CPCM/H2O solvent model.³¹ The molecular orbitals were then visualized with GaussView.

Molecular Dynamics Simulations and Analysis

Molecular dynamics trajectories were produced for the p53DBD with and without the adduct on each of Cys124, Cys229, and Trp146 separately. These protein models were based on the A-chain in the PDB: 2XWR crystal structure.³² Simulations were done with the NAMD 2.13b2 software package³³ using the CHARMM36 all-atom force field.³⁴ The adducts were modeled with CHARMM-GUI.³⁵ Weak potentials (10 kcal/mol) were used to constrain the zinc ion to its four ligand side chains. The systems were hydrated with TIP3P waters in a square, periodic-boundary-condition cell that started at 70.0 × 66.0 × 59.0 Å. Sodium and chloride ions were added to electrically neutralize and provide an ionic strength of 150 mM using Visual Molecular Dynamics.³⁶ The minimum distance between the protein and the cell wall was 12.0 Å. Long-range electrostatic and van der Waals interactions were truncated with a CUTOFF of 12.0 Å and a SWITCHDIST of 8.0 Å. The Langevin method was used to maintain the temperature at 310.0 K, and the Langevin–Piston method was used to maintain the pressure at 1 ATM. A 2.0 fs integration time-step was enabled using the SETTLE mechanism to restrain covalent bond lengths to equilibrium values. Trajectories were calculated for a minimum of 1000.0 ns each. Conformational clusters and pictures were produced with UCSF Chimera.³⁷ Analyses of hydrogen bonds and all other structural parameters were done with CHARMM.³⁸

Mass Spectrometry

Synthetic p53 peptides were incubated with NSC59984 at 10:1 or 100:1 inhibitor:peptide molar ratio; incubation with an equal volume of DMSO was also performed as a vehicle control. Reactions were performed in 20 mM Tris pH 7.4 and incubated at 37 °C for 3 h. Following reaction, peptides were injected onto a 50 × 2 mm Synergi Fusion-RP C18 column (Phenomenex) using an Exion HPLC (SCIEX) and then analyzed on an X500B Q-TOF mass spectrometer (SCIEX). Data were analyzed and mass spectra reconstructed in SCIEX-OS.

Immunoprecipitated p53 was separated by gel electrophoresis and the p53 band cut from the gel. The protein was in-gel digested with trypsin or chymotrypsin at 37 °C for 16 h, as described;³⁹ reduction was not performed to retain potential cysteine modification by NSC59984, only alkylation of free sites. Dried peptides were separated on a 75 μm × 15 cm, 2 μm Acclaim PepMap reverse phase column (Thermo Scientific) at 300 nL/min using an UltiMate 3000 RSLCnano HPLC (Thermo Scientific). Peptides were eluted into a Thermo Orbitrap Fusion mass spectrometer with parent full-scan mass spectra collected in the Orbitrap mass analyzer, fragmentation in the HCD cell (HCD normalized energy 32%, stepped ±3%), and product ions analyzed in the ion trap. Proteome Discoverer 2.4 (Thermo) was used to search the data against the subset of human proteins from the Uniprot database using SequestHT. The search was limited to tryptic or chymotryptic peptides, with maximally two missed cleavages allowed. Cysteine carbamidomethylation and methionine oxidation were set as variable modifications. Modification of cysteine, tryptophan, and histidine by NSC59984 was also included as a variable modification. For all searches, the precursor mass tolerance was 10 ppm, and the fragment mass tolerance was 0.6 Da. Modified spectra were manually validated.

Recombinant TIGAR (OriGene Technologies) was incubated with a 10-fold molar excess of NSC59984 in PBS buffer (pH 7.2) at 25 °C for 1 h. The reacted TIGAR was alkylated with iodoacetamide without reduction to retain potential cysteine modification by NSC59984 and then in-solution digested with trypsin at 37 °C. Peptides were desalted and dried. Dried peptides were separated on a 75 μm × 15 cm, 2 μm Acclaim PepMap reverse phase column (Thermo Scientific) at 300 nL/min using an UltiMate 3000 RSLCnano HPLC (Thermo Scientific). Peptides were eluted into a Thermo Orbitrap Exploris 480 mass spectrometer with parent full-scan mass spectra collected in the Orbitrap mass analyzer, fragmentation in the HCD cell (HCD normalized energy 32%, stepped ±3%), and product ions analyzed in the Orbitrap. Proteome Discoverer 3.0 (Thermo) was used to search the data against the subset of human proteins from the Uniprot database using SequestHT. The search was limited to tryptic peptides, with maximally two missed cleavages allowed. Cysteine carbamidomethylation and methionine oxidation were set as variable modifications. Modification of cysteine by NSC59984 was also included as a variable modification. For all searches, the precursor mass tolerance was 10 ppm, and the fragment mass tolerance was 0.02 Da. Modified spectra were manually validated.

Cell Culture

Treatment of ESO26-R248W cells with NSC59984 ((E)-1-(4-methylpiperazin-1-yl)-3-(5-nitrofuran-2-yl)prop-2-en-1-one) was carried out in complete RPMI media at 12 μM for 72 h unless otherwise stated. DMSO was used as a carrier control for all treatments.

Cell Proliferation

The CyQuant Cell Proliferation Assay Kit (Life Technologies) was used according to the manufacturer’s instructions. Cells were seeded onto poly-d-lysine coated plates and treated with NSC59984 for 72 h, after which the medium was removed from the cells and replaced with fresh medium without drug. On five consecutive days, the medium was removed from the cells and the plate was stored at −80 °C at least overnight to lyse the cells. Once plates from all 5 days had been lysed CyQuant reagents were added for 15 min before samples were analyzed on the Synergy HI microplate reader (BioTek). Results were calculated from the standard curve and are represented as fold change over control. A 2-way ANOVA test with Tukey correction was carried out for statistical analysis. The results shown were obtained from at least three independent experiments. Results were plotted and analyzed with GraphPad Prism. Error bars show the standard deviation. p-Values smaller than 0.05 were indicated as significant.

Apoptosis Assay

The ApoAlert Annexin V-FITC Apoptosis Kit (Takara) was used according to the manufacturer’s instructions. Cells were seeded and stimulated for 72 h. Analysis was carried out on a FACS Calibur (BD Biosciences). Quadrants were defined by single stain controls and the percentages of apoptotic cells were calculated from both early and late apoptotic fractions. A 2-way ANOVA test with Tukey correction was carried out for statistical analysis. The results shown were obtained from at least three independent experiments. Results were plotted and analyzed with GraphPad Prism. Error bars show the standard deviation p-values smaller than 0.05 were indicated as significant.

Western Blot

Cells were lysed with freshly prepared RIPA buffer containing 50 mM Tris–HCl, 10 mM MgCl2, 20% glycerol, and 1% Triton X-100 with protease and phosphatase inhibitors (Roche) and centrifuged at 10,000 × g for 10 min at 4 °C. Protein concentration was determined by BCA assay (Pierce). Samples for Western blot analysis were combined with NuPAGE LDS sample buffer (Life Technologies) containing 0.05% β-mercaptoethanol, and incubated at 95 °C for 5 min. Twenty μg of protein was loaded onto a 4–12% SDS-polyacrylamide gel. The resolved proteins were blotted onto the nitrocellulose membrane and detected using TP53 D01 (Santa Cruz) antibody. Results were imaged using ECL and the Molecular Imager ChemiDoc Touch (BioRad).

NSC-Biotin Pulldowns

Control and treated ESO26-R248W cells were lysed with freshly prepared RIPA buffer containing 50 mM Tris–HCl, 10 mM MgCl₂, 20% Glycerol and 1% Triton X-100 with protease and phosphatase inhibitors (Roche) and centrifuged at 10,000g for 10 min at 4 °C. Protein concentration was determined by BCA assay (Pierce) and 1.5 mg of total protein was analyzed for each pulldown. Samples were precleared on Protein G beads (Bio-Rad) for 1 h before incubation with Streptavidin Mag Sepharose beads (GE Healthcare) for 1 h on a rotating shaker. Following incubation, the unbound sample was removed, and the pulldown samples were washed twice in TBS + 8 M urea before resuspension in PBS. Pulldown samples were then in-solution digested using S-traps (Protifi) following the manufacturer’s protocols. Dried samples were analyzed on a Thermo Orbitrap Fusion mass spectrometer. Proteome Discoverer 2.4 (Thermo) was used to search the data against the subset of human proteins from the Uniprot database using SequestHT. Label-free quantitation of proteins was performed in Proteome Discoverer and used to calculate enrichment ratios. All experiments were performed in replicate.

Thermal Proteome Profiling

Untreated ESO26-R248W cells were lysed by freeze–thaw method in freshly prepared buffer containing 25 mM Tris pH 7.5, 10 mM MgCl₂, 1 mM DTT, and 4% NP-40 and centrifuged at 20,000 × g for 20 min at 4 °C. The lysate was split evenly and incubated with either DMSO or NSC59984 for 1 h at 25 °C. Samples were divided equally into 5 microfuge tubes and incubated in a thermocycler for 3 min at either 37, 44.5, 52, 59.5, or 67 °C followed by 3 min at 25 °C. Samples were centrifuged at 20,000xg for 20 min at 4 °C to pellet insoluble proteins. Following thermal cycling and fractionation, the protein concentration in each sample was determined using a BCA colorimetric assay (Thermo); proteins were normalized at the 37 °C condition and equal volumes of higher temperature for each treatment were used. Proteins were reduced using 10 mM dithiothreitol at 56 °C for 1 h and alkylated with 20 mM iodoacetamide in the dark for 30 min. Following alkylation, protein digestion was performed overnight at 37 °C at a 1:50 ratio of trypsin to protein concentration. The samples were desalted using C18 desalting columns (Thermo) and peptide concentration was assessed using a BCA peptide colorimetric assay (Thermo). Tandem Mass Tag (TMT) labeling was performed using TMT or TMTpro reagents (Thermo) following the manufacturer’s instructions. The label peptides were pooled and then dried. High pH fractionation was then performed using a spin-column kit (Thermo). Fractionated peptides were analyzed on a Thermo Orbitrap Exploris 480 mass spectrometer similar to the TIGAR reactions with the exception of 0.7 Da isolation window being used. Data were processed and output using Proteome Discoverer 3.0 and melting curves were generated using the mineCETSA R package. Data were scaled and curves were fitted and plotted using the functions within the mineCETSA package. p-Values were calculated from z-scores generated from the distribution of Euclidean distances output by the mineCETSA package^40,41 using the Robust Z-Score method; a Benjamini-Hochberg multiple hypothesis correction was applied. Mass spectrometry data have been deposited to the massIVE repository with the data set identifier MSV000095372.⁴²

Immunofluorescence (IF)

Cells were grown and stimulated on PDL-coated 18 mm glass coverslips for 72 h. Cells were then fixed using 5% paraformaldahyde, permeabilized using 0.5% Triton X-100, counter-stained with Hoechst 33342 at 10 nM. TP53 antibody (D01) (Santa Cruz) was diluted 1:50 before incubation, and a secondary antimouse AF647 antibody was used at 5 μg/mL before counter-staining. Streptavidin-488 (Abcam) was diluted 1:800. Images were captured on a Zeiss LSM710 confocal microscope capturing images at optimum slice depth for each sample. Gain settings and exposure time were kept constant between samples.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsptsci.4c00447.

• Characterization of novel compounds; figures indicating the cellular effects of compound 1 (Figure S1); HPLC chromatograms of reaction of NSC59984 with glutathione and N-acetyl cysteine (Figure S2); mass spectra showing reaction of NSC59984 with p53 (Figure S3); structure, cellular distribution, and peptide reaction of NSC-biotin (Figure S4); plots of data normalization and Euclidean distances of melting curves for thermal proteome profiling experiments (Figure S5); plots of MS/MS spectra for modified TIGAR peptides (Figure S6)(PDF)

• Tables containing mass spectrometry results and statistical analyses (Table S1) (XLSX)

Author Present Address

^∥ Department of Chemistry, University of Wisconsin-Madison, Madison, Wisconsin 53706, United States (A.J.P.)

Author Present Address

^⊥ The Green Center for Systems Biology, The Lyda Hill Department of Bioinformatics, and The Department of Biophysics, University of Texas Southwestern Medical Center, Dallas, TX 75390, United States (J.C.D.).

Author Present Address

^# The Pennsylvania State University College of Medicine, Hershey, PA 17033, United States (G.C.L.).

Author Present Address

^∇ Institute for Bioscience and Biotechnology Research, University of Maryland, Rockville, MD 20850, United States (H.D.T.).

Author Present Address

^○ Protein Research Center, Agropur, Le Sueur, MN 56058, United States (D.S.).

Author Present Address

^◆ University of San Francisco, San Francisco, CA 94117, United States (H.N.).

Author Contributions

^§ K.B. and M.R. contributed equally.

Author Contributions

L.M.J., D.H.A., and E.A. conceived of experiments; K.B., M.R., A.J.P., J.C.D., T.K.M., G.C.L., J.P.K., S.R.D., H.D.T., D.S., H.N., and R.O. performed experiments; K.B., M.R., A.J.P., J.C.D., T.K.M., G.C.L., D.H.A., and L.M.J. analyzed the data; K.B., M.R., A.J.P., D.H.A., and L.M.J. wrote the paper. All authors have read and approved the final version of the paper.

The authors declare no competing financial interest.