Targeting the p53 cancer mutants Y220C, Y220N, and Y220S with the small-molecule stabilizer rezatapopt

TL;DR

This study shows that the drug rezatapopt can stabilize and reactivate certain mutant forms of the p53 protein, which are linked to cancer, but with varying effectiveness across different mutations.

Contribution

The study demonstrates that rezatapopt binds to and stabilizes the Y220N and Y220S p53 mutants, expanding its potential therapeutic scope.

Findings

Rezatapopt binds to Y220N and Y220S p53 mutants with nanomolar affinity and restores wild-type-like stability.

High-resolution crystal structures reveal a conserved binding mode and key interactions with the protein backbone.

Rezatapopt reactivates p53 signaling in Y220C and Y220S mutant cells but shows limited effect on Y220N cells at tested concentrations.

Abstract

The cavity-creating p53 cancer mutation Y220C, which accounts for an estimated 125,000 new cancer cases per year, serves as an excellent paradigm for the development of mutant p53 reactivators. Several molecules that reactivate this thermolabile cancer mutant by targeting the mutation-induced crevice have been developed, and one of them, rezatapopt, is currently in clinical trials. The less frequently occurring Y220N and Y220S mutations are even more destabilizing than Y220C but create a similar surface crevice, raising the question of whether cancer patients with these mutations might also benefit from rezatapopt treatment. Here, we show that rezatapopt also binds to the Y220N and Y220S mutants, with nanomolar affinity, resulting in a full recovery of wild-type-like stability for the latter. High-resolution crystal structures of all three mutants bound to rezatapopt revealed a conserved binding mode, highlighting key interactions, including multipolar interactions of a fluorine substituent at a chiral center with the protein backbone. Consistent with the biophysical and structural data, rezatapopt reactivated p53 signaling in both Y220C and Y220S mutant cells by restoring the folded conformation and transcriptional activity, leading to anti-proliferative effects and apoptosis, albeit requiring higher compound concentrations in Y220S cells. The Y220N mutant, despite exhibiting high-nanomolar affinity for rezatapopt and substantial stabilization, did not show noticeable effects in cells at the concentrations tested, as rezatapopt binding resulted in only partial compensation for the mutation-induced loss of stability, for which we provide a structural explanation. Our data suggest that the development of clinical pan-Y220C/N/S reactivators, which could benefit an additional 10,000 patients per year, is challenging but not impossible.

Figures

Chemical structures of Y220C-reactivating molecules.The central scaffold of PMV Pharmaceuticals’ clinical compound rezatapopt merges two previously developed Y220C binders: the carbazole-based PK083 [[13](#CR13)] and the iodophenol-based PK5196, which features a subsite 2-targeting acetylene linker moiety [[14](#CR14)]. Rezatapopt analog 7 lacks the chiral center at the piperidine with fluorine substituent, whereas analog 8 contains a different moiety for targeting the subpocket around Pro153. Also shown are the chemical structures of the optimized iodophenol-based Y220C binder JC744, which ex

Affinity and stability effects of Y220C binders on the related Y220S and Y220N mutants.A Melting temperature of Y220C, Y220S, and Y220N mutant DBDs (stabilized quadruple-mutant variant, QM; M133L/V203A/N239Y/N268D) in their unbound state and with increasing concentrations of stabilizing compound, as measured by DSF. The dashed line indicates the melting temperature of the corresponding pseudo-wild-type DBD (QM), without Tyr220 mutation. B–E Representative ITC data and derived thermodynamic parameters of p53-Y220C/N/S mutant DBDs binding to rezatapopt and analogs; see also Table [2](#Tab2). B Y

Crystal structures of the p53 Y220C mutant DBD in complex with rezatapopt and close analogs.A Y220C-rezatapopt complex. The protein is shown as a surface representation colored according to electrostatic potential calculated using PyMOL (red: negative potential; blue: positive potential) and the ligand as a stick model. B Y220C-rezatapopt complex, ribbon diagram shown in a different orientation, with key interacting residues highlighted as stick models. Hydrogen bonds between the ligand and the mutant protein are shown as green dashed lines, and multipolar interactions of fluorine substituents

Structures of Y220S and Y220N mutant complexes with bound rezatapopt.A Superimposition of the binding modes of rezatapopt in the Y220C, Y220N, and Y220S mutants (based on chain A), showing perfect conservation of the overall binding mode. B. 2F_o_-F_c_ electron density map for the rezatapopt molecule in the Y220S complex (chain A), determined at a resolution of 1.4 Å, at a contour level of 1.8 sigma. C Structure of the Y220N mutant (chain B) highlighting the solvent-facing orientation of the Asn220 side chain in the unbound state, with Asn220-interacting structural water molecules highlighted

Rezatapopt induces target gene activation in Y220C and Y220S p53 mutants.A Immunoblot showing induction of p53 Y220C/S/N/H mutants in p53-null H1299 cells upon doxycycline treatment. B Fold change in the expression of p53 target genes *CDKN1A/p21*, *MDM2*, *BBC3*, and *BAX* after rezatapopt treatment, measured by RT-qPCR. Data are shown as mean ± SD (*n* = 3). C Immunoblot analysis of p53 and p21 protein levels following rezatapopt treatment. Images of the uncropped western blots are available in the Supporting Information.Fig. 6Rezatapopt inhibits cell growth by refolding Y220C and Y220S muta

Full text

Introduction

The transcription factor p53 is a key regulator of the cell cycle, safeguarding the integrity of our genome and preventing malignant transformation. It is inactivated by mutation in every second tumor, making the reactivation of p53 cancer mutants an attractive therapeutic strategy [1–5]. Most cancer-associated p53 mutations are missense mutations spread across the intrinsically unstable DNA-binding domain (DBD), with more than 2000 different missense variants reported in cancer [6]. Mechanistically, they can be classified as either DNA-contact mutants or conformationally unstable structural mutants that rapidly unfold and aggregate [7, 8]. A subgroup of destabilized structural mutants is temperature-sensitive, retaining an active, folded conformation at subphysiological temperature [9–11]. Such mutants can, in principle, be reactivated in cells at body temperature by binding to stabilizing small molecules that shift the folding-unfolding equilibrium toward a wild-type-like conformation capable of binding p53 DNA response elements and inducing transcription of p53 target genes.

The Y220C mutation, caused by an A-to-G transition at codon 220 of the TP53 gene, is an excellent paradigm for such a reactivation strategy. It is one of the most frequent p53 cancer mutations, accounting for approximately 1% of solid tumors and an estimated 125,000 new cancer cases per year worldwide (Table 1). In 2006, we showed that mutation of the tyrosine to a smaller cysteine creates a unique, extended surface crevice, and we suggested at the time that the Y220C mutant could be reactivated by molecules specifically targeting this site [12]. Since then, we have systematically mapped this binding site and developed several series of compounds that bind to the Y220C mutant, increase its thermostability and inhibit aggregation [13–18]. Some of these molecules were active in cancer cells with a homozygous Y220C mutation and reactivated a p53-dependent transcriptional program, providing cellular proof of concept [17, 19]. In addition, recent studies have exploited the thiol reactivity of Cys220 for the development of Y220C-specific covalent reactivators [20, 21]. Building on some of our earlier compound series (Fig. 1), PMV Pharmaceuticals have developed the Y220C reactivator rezatapopt [22], which induces full reactivation of the p53 transcriptional program in cells and inhibits tumor growth in vivo in xenograft and syngeneic mouse models [23]. Rezatapopt is currently being evaluated in a phase II clinical trial for patients with advanced solid tumors harboring the TP53 Y220C mutation [24]. In addition, several groups have developed bifunctional molecules, for example by linking Y220C binders to a zinc-chelating moiety [25] or by designing proximity-inducing drugs that recruit acetylases [26], deubiquitinases [27], or the transcriptional and epigenetic regulator BRD4 [28].Fig. 1. Chemical structures of Y220C-reactivating molecules.The central scaffold of PMV Pharmaceuticals’ clinical compound rezatapopt merges two previously developed Y220C binders: the carbazole-based PK083 [13] and the iodophenol-based PK5196, which features a subsite 2-targeting acetylene linker moiety [14]. Rezatapopt analog 7 lacks the chiral center at the piperidine with fluorine substituent, whereas analog 8 contains a different moiety for targeting the subpocket around Pro153. Also shown are the chemical structures of the optimized iodophenol-based Y220C binder JC744, which exhibits high-nanomolar affinity [18], and the carbazole-based compound PK9301 [29].Table 1. Frequency of missense p53 mutations at codon 220 in the GENIE cancer mutation database^a^.VariantUnique Patients^b^Percentage of Total (%)Extrapolated Global Cases p.a.^c^Y220C1222 (1159)0.675125,000Y220H69 (62)0.0357000Y220S64 (45)0.0336500Y220N50 (42)0.0255100Y220D28 (24)0.0142850^a^GENIE Cohort v17.0-public, January 2025; total of 196,244 unique patients [31].^b^The number of cases in which it is the only missense mutation in the TP53 gene is given in parentheses.^c^Calculated based on a cancer incidence of 20 million new cases per year worldwide (2022 GLOBOCAN report of the International Agency for Research on Cancer) [52].

Y220C is the most frequent, but not the only, cancer-associated mutation at codon 220. Three other mutations, Y220N, Y220S, and Y220H, also occur with an appreciable frequency, each accounting for an estimated 5000–7000 new cancer cases per year worldwide [29–31] (Table 1). We have shown that the Y220S mutant features a pronounced mutation-induced surface crevice, similar to that of the Y220C mutant, and that both the Y220S and Y220N variants are druggable, whereas the binding site is occluded in the crystal structure of the Y220H variant [29]. The Y220S and Y220N mutations are, however, significantly more destabilizing than the Y220C mutation [29], making the restoration of a wild-type-like stability and cellular potency more challenging. In principle, it should also be possible to develop clinical compounds for those two mutants; however, the low patient numbers make it economically unviable to initiate a full-scale industrial program targeting them individually.

To assess whether cancer patients with a Y220S or Y220N mutation might also benefit from treatment with PMV Pharmaceuticals’ current clinical candidate rezatapopt, we systematically evaluated its effects on these mutants. This included a comprehensive structural characterization of its binding mode in the Y220C, Y220S, and Y220N variants, as well as an assessment of its reactivation potential in the corresponding mutant cell lines.

Results and discussion

Effect of rezatapopt and analogs on Y220C/N/S mutant stability and their binding affinity

We tested the effect of rezatapopt and two analogs from the patent literature, compounds 7 and 8 (see Fig. 1; refs. [32, 33]), on the thermostability of the Y220C, Y220N, and Y220S mutant DBDs by monitoring thermal unfolding using differential scanning fluorimetry (DSF) (Fig. 2A and Supplementary Table S1). All measurements with recombinant protein were performed using a stabilized quadruple mutant DBD variant, QM (residues 94-312; M133L/V203A/N239Y/N268D), which we have routinely used in the past for biophysical and structural studies of conformationally unstable p53 mutants [12, 34, 35]. The Y220C, Y220N, and Y220S mutations reduced the melting temperature, Tm, of the DBD by 9.0, 12.4, and 12.7 °C, respectively. Rezatapopt and its analogs increased the thermal stability of all three mutants in a concentration-dependent manner. For the Y220C and Y220S mutants, this resulted in a Tm increase of up to 11.4 and 13.0 °C, respectively, at the highest compound concentration tested (250 μM), thereby in both cases compensating for the mutation-induced stability loss. The maximal Tm increase upon rezatapopt addition for the Y220N mutant was 7.7 °C. This is a substantial increase in stability but does not fully compensate for the mutation-induced stability loss (ΔTm = −12.4 °C), leaving a net loss of ~5 °C compared with the wild-type DBD. Analogs 7 and 8 had a similar stabilizing effect as rezatapopt on all mutants but were less potent overall.Fig. 2. Affinity and stability effects of Y220C binders on the related Y220S and Y220N mutants.A Melting temperature of Y220C, Y220S, and Y220N mutant DBDs (stabilized quadruple-mutant variant, QM; M133L/V203A/N239Y/N268D) in their unbound state and with increasing concentrations of stabilizing compound, as measured by DSF. The dashed line indicates the melting temperature of the corresponding pseudo-wild-type DBD (QM), without Tyr220 mutation. B–E Representative ITC data and derived thermodynamic parameters of p53-Y220C/N/S mutant DBDs binding to rezatapopt and analogs; see also Table 2. B Y220C mutant with compound 7. C Y220C mutant with rezatapopt. D Y220S mutant with rezatapopt. E Y220N mutant with rezatapopt.

We next measured the dissociation constant, Kd, of rezatapopt by isothermal titration calorimetry (ITC) (Table 2, Fig. 2, and Supplementary Fig. S1). Rezatapopt bound to the Y220C mutant with a Kd of 22.3 ± 4.2 nM and to the Y220S mutant with a Kd of 96.6 ± 7.5 nM. The Y220N mutant exhibited a significantly reduced affinity for rezatapopt compared with Y220C or Y220S, though it remained in the high-nanomolar range (Kd = 526 ± 56 nM). Consistent with published data, analog 7, lacking the fluorine substituent at the piperidine moiety, had an almost six-fold reduced affinity for the Y220C mutant (Kd = 119 ± 10 nM). Analog 8, which features the fluoro-piperidyl substituent but incorporates an oxadiazole moiety in place of the alkyne linker and a different subsite 2-targeting aromatic moiety, bound with a Kd of 65 ± 18 nM. We note that the ITC Kd for the Y220C-rezatapopt complex is approximately 10 times higher than the reported value measured by surface plasmon resonance [23].Table 2. Dissociation constants of rezatapopt and its analogs measured by isothermal titration calorimetry (ITC).p53 mutantCompoundKd (nM)^a^Y220Crezatapopt22.3 ± 4.2 (6)Y220Srezatapopt96.6 ± 7.5 (2)Y220Nrezatapopt526 ± 56 (6)Y220C7119 ± 10 (3)Y220C864.7 ± 18.4 (2)^a^Average of n (in parentheses) independent measurements ± SD.

Structural basis of rezatapopt binding to Y220C/N/S mutants and its energetic implications

At the time of writing, no crystal structure of the Y220C DBD with the clinical compound rezatapopt was available in the Protein Data Bank; only structures with two closely related analogs lacking the fluorine substituent on the piperidine moiety have been deposited. Here, we determined crystal structures of the Y220C mutant in complex with rezatapopt and three related compounds (compounds 5, 7, and 8; cf. Fig. 1). In addition, we determined cocrystal structures of the Y220S and the Y220N mutants with rezatapopt, plus the structure of the ligand-free Y220N mutant. All structures were determined at high resolution, ranging from 1.38 to 1.87 Å (Supplementary Table S2).

In the structure of the Y220C-rezatapopt complex (1.49 Å resolution), the central indole scaffold sits at the center of the mutation-induced cavity and is sandwiched between the S7/S8 loop residues Pro222 and Pro223 on one side, and Val147, Thr150, and Pro151 on the other side (Fig. 3A, B). The CF_3_ anchor is buried at the deepest part of the pocket, and two of the fluorine substituents form favorable interactions with the Cys220 sulfur atom (S-F distances = 3.4 and 3.5 Å). Compared with the parent carbazole PK083 [13] and its analogs bearing a fluorinated ethyl anchor [36], the central, N-alkylated indole scaffold is shifted by about 1.3 Å toward the Cys220 C_α_ atom, enabled by a flip of the Cys220 side chain (Fig. 3C). The 4-aminopiperidine moiety points toward the solvent and forms a hydrogen bond with the hydroxyl group of Thr150. The fluorine substituent interacts directly with the protein through multipolar interactions with the backbone carbonyl groups of Asp148 and Ser149 (C-F distances = 3.4 and 3.3 Å, respectively). Such a stabilization would not be possible with the R enantiomer, explaining the reported superior potency of rezatapopt over its enantiomer and its more than threefold higher affinity compared with the parent compound without fluorination of the piperidine moiety [22]. The phenylamine moiety binds in subsite 2, where the aromatic ring packs against Pro153 via a CH-π stacking interaction, and the amine group forms a hydrogen bond with the backbone carbonyl oxygen of Cys220, as observed for the corresponding compound PK5196 from the iodophenol series [14] (Fig. 3B, C). Both ring substituents point toward the solvent. Preferential packing of aromatic rings at this site has also been observed for several fragment hits [16, 37].Fig. 3. Crystal structures of the p53 Y220C mutant DBD in complex with rezatapopt and close analogs.A Y220C-rezatapopt complex. The protein is shown as a surface representation colored according to electrostatic potential calculated using PyMOL (red: negative potential; blue: positive potential) and the ligand as a stick model. B Y220C-rezatapopt complex, ribbon diagram shown in a different orientation, with key interacting residues highlighted as stick models. Hydrogen bonds between the ligand and the mutant protein are shown as green dashed lines, and multipolar interactions of fluorine substituents as magenta dashed lines. C Superimposition of the binding modes of the core PMV scaffold 5 and the parent molecules PK083 (PDB: 2VUK) and PK5196 (PDB: 4AGQ). D Complex of Y220C with compound 7. E, F Complex of the Y220C mutant with compound 8, shown in the same way as the Y220C-rezatapopt complex in panels (A) and (B), respectively.

The closely related rezatapopt analog 7, which lacks the piperidine fluorination and has a different substitution pattern at the Pro153-targeting phenyl moiety, exhibited essentially the same binding mode (Fig. 3D). In the structure of the Y220C mutant bound to compound 8 (Fig. 3E, F), in which the acetylene linker is replaced by an oxadiazole moiety, the position of the indole scaffold is slightly shifted, and the side chain of Cys220 is flipped toward Val157 due to the larger spatial requirements of the oxadiazole linker. As a result, the Cys220 sulfur atom is further away from the CF_3_ group (closest S-F distances = 4.0 and 3.9 Å in chain A and B, respectively). Unlike rezatapopt, this compound does not form a hydrogen bond with the Cys220 carbonyl group but instead forms an alternative hydrogen bond with the carbonyl group of Pro151 (Fig. 3F). Interaction with Pro153 occurs via the pyrazole moiety through hydrophobic and CH-π interactions.

The high-resolution crystal structures of the Y220S and Y220N mutants in complex with rezatapopt revealed the same overall binding mode, with only marginal shifts compared to that observed in the Y220C mutant (Fig. 4A). There was excellent electron density in both cases, allowing unambiguous modeling of the ligand, as shown for the Y220S mutant in Fig. 4B. Almost all key interactions are conserved in the Y220N and Y220S complexes, including the targeting of Pro153, the hydrogen bond with the backbone carbonyl of Cys220, and the multipolar interactions of the fluoropiperidine moiety (Fig. 4D, F). The differences explaining the reduced binding affinity in both cases are the altered coordination of the CF_3_ anchor by the mutated side chain at the bottom of the binding site and binding-induced reorientations of the mutated side chain. Similarly to the Y220C mutant (Fig. 3B, C), there is a flip of the side chain of residue 220 to accommodate the CF_3_ moiety, which is particularly pronounced for the Y220N mutant. In the ligand-free structure of the Y220N mutant, the side-chain amide points toward the solvent and interacts with structural water molecules (Fig. 4C). Upon ligand binding, the Asn220 side chain is pushed toward Val157 in the hydrophobic core of the protein, which also flips in return, to accommodate rezatapopt (Fig. 4D). As a result, the side-chain amide group of Asn220 is buried in an energetically unfavorable hydrophobic environment and is no longer saturated with hydrogen bonds. Combined with the loss of stabilizing interactions with the CF_3_ group, this accounts for the approximately 25-fold reduced potency of rezatapopt compared with the Y220C mutant, as well as the lack of full recovery of wild-type stability. In the ligand-free Y220S mutant structure, the Ser220 hydroxyl also forms hydrogen bonds with structural water molecules that are displaced upon rezatapopt binding (Fig. 4E, F). Again, the side chain reorientates upon rezatapopt binding, but, being smaller, it can be accommodated in an energetically more favorable way than an asparagine, resulting in a hydrogen bond with the backbone carbonyl group of Thr155. In this case, there is therefore no additional penalty from binding, and the relatively modest affinity difference compared with the Y220C mutant can be largely attributed to the missing interaction between residue 220 and the CF_3_ anchor.Fig. 4. Structures of Y220S and Y220N mutant complexes with bound rezatapopt.A Superimposition of the binding modes of rezatapopt in the Y220C, Y220N, and Y220S mutants (based on chain A), showing perfect conservation of the overall binding mode. B. 2F_o_-F_c_ electron density map for the rezatapopt molecule in the Y220S complex (chain A), determined at a resolution of 1.4 Å, at a contour level of 1.8 sigma. C Structure of the Y220N mutant (chain B) highlighting the solvent-facing orientation of the Asn220 side chain in the unbound state, with Asn220-interacting structural water molecules highlighted as red spheres and water-mediated hydrogen bonds shown as orange dashed lines. D Structure of the Y220N-rezatapopt complex in which the mutated side chain is pushed into an unfavorable position facing the hydrophobic core of the protein. E Structure of the Y220S mutant in its free state (PDB entry 6SI2), with selected structural water molecules in the immediate environment of the mutated side chain shown as red spheres. F Structure of the Y220S-rezatapopt complex. Hydrogen bonds between the ligand and the mutant protein are shown as green dashed lines, and multipolar interactions between fluorine substituents and the protein as magenta dashed lines. The hydrogen bond between the Ser220 side chain and the backbone carbonyl of Thr155 is highlighted as an orange dashed line.

Rezatapopt induces a strong transcriptional response in Y220C and Y220S mutant cells

Next, we tested whether the binding of rezatapopt to the Y220S or Y220N mutant, and the resulting stabilization, translates into a restoration of p53 function in cells. For this purpose, we generated doxycyline-inducible H1299 cell lines of Y220C, Y220S, Y220N, and Y220H (Fig. 5A). We first measured the mRNA levels of p53 target genes CDKN1A/p21 (cell-cycle arrest), MDM2 (p53 negative-feedback loop), BBC3/PUMA (proapoptotic), and BAX (proapoptotic) after 24 h rezatapopt treatment (Fig. 5B). Consistent with PMV’s recently published data on other cell lines [23], rezatapopt resulted in a strong upregulation of all four target genes, with a fold change for CDKN1A > 100. In the Y220S mutant cell line, a similarly strong upregulation of the target genes was observed, except for BBC3/PUMA, for which the increase was significant but less pronounced. The rezatapopt-induced upregulation of CDKN1A/p21 expression in the Y220C and Y220S mutant cell lines was also confirmed at the protein level by western blotting (Fig. 5C). However, the onset of transcriptional reactivation in the Y220S cells was consistently shifted to an approximately 16-fold higher compound concentration. This shift can be readily explained by the roughly 4-fold lower binding affinity combined with the greater mutation-induced stability loss in the Y220S mutant. In contrast, no induction of p53 target gene expression was observed for the Y220N mutant at the rezatapopt concentrations tested, consistent with the biophysical data showing that rezatapopt binding does not fully restore wild-type-like stability (Fig. 2A). This was confirmed in cells by monitoring the folding state of the mutants upon rezatapopt treatment using conformation-specific antibodies (Fig. 6A). PAb1620 binds to the folded state of the p53 DBD, whereas PAb240 recognizes a cryptic epitope on β-strand S7 that becomes exposed upon unfolding of the DBD [38]. The antibody DO-1 binds to the intrinsically disordered N-terminal region of p53 and thus detects total p53 protein levels, independent of the folding state of the DBD. Rezatapopt reduced the amount of unfolded Y220C mutant protein and increased the amount of folded protein at low micromolar concentrations. Again, the same trend was observed for the Y220S mutant, although higher compound concentrations were required. For the Y220N and Y220H mutants, no refolding was observed, and they remained predominantly unfolded even at the highest rezatapopt concentration tested (64 μM). The rezatapopt-induced increase in folded protein levels and restoration of transcriptional activity of the Y220C and Y220S mutants also resulted in reduced proliferation (Fig. 6B, C) and cell viability (Fig. 6D, E). Rezatapopt reduced the viability of the Y220C cell line, with an IC_50_ of 130 nM. A significant viability reduction was also observed for the Y220S mutant (IC_50_ = 3.5 μM), but not for the Y220N or Y220H mutants. Interestingly, a reduction in cell proliferation was also seen for the Y220N mutant at high rezatapopt concentrations (10 μM), but not for the Y220H variant. This could indicate a weak, partial reactivation for the Y220N mutant that manifests only at later time points. By contrast, target gene induction was measured at 24 h, when such subtle effects may remain below the limit of detection. Supporting this interpretation, proliferation curves for treated Y220N cells begin to diverge only after 48 h.Fig. 5. Rezatapopt induces target gene activation in Y220C and Y220S p53 mutants.A Immunoblot showing induction of p53 Y220C/S/N/H mutants in p53-null H1299 cells upon doxycycline treatment. B Fold change in the expression of p53 target genes CDKN1A/p21, MDM2, BBC3, and BAX after rezatapopt treatment, measured by RT-qPCR. Data are shown as mean ± SD (n = 3). C Immunoblot analysis of p53 and p21 protein levels following rezatapopt treatment. Images of the uncropped western blots are available in the Supporting Information.Fig. 6. Rezatapopt inhibits cell growth by refolding Y220C and Y220S mutant p53 into its wild-type conformation.A Immunoblot showing detection of unfolded and wild-type-like folded p53 mutant protein following immunoprecipitation with conformation-specific antibodies after rezatapopt treatment. B Growth curves of p53 Y220C/S/N/H-expressing H1299 cells following rezatapopt treatment. Data are shown as mean ± SD (n = 3). C Relative area under the curve (AUC) derived from the growth curves in (B). Data are shown as mean ± SEM (n = 3); p-values determined in comparison to respective control treatment by two-way ANOVA with Dunnett’s multiple comparisons test. D Dose-response curve showing relative cell viability following rezatapopt treatment, assessed by CellTiterGlo assay. Data are shown as mean ± SD (n = 3). E IC_50_ values calculated from (D). Data are shown as mean and 95% CI (n = 3). F Colony formation assay showing clonogenic growth under rezatapopt treatment. Representative images from n = 2 replicates are shown. G Cell-cycle analysis assessed by EdU/PI staining and flow cytometry, showing the percentage of cells in S-phase after rezatapopt treatment normalized to the respective control treatment. H Percentage of apoptotic cells assessed by Annexin V staining and flow cytometry following rezatapopt treatment.

Consistent with the upregulation of p53 target genes, rezatapopt reduced clonogenic growth of the Y220C- and Y220S-expressing cells in long-term colony formation assays (Fig. 6F and Supplementary Fig. S2). In addition, FACS analysis showed induction of cell-cycle arrest (Fig. 6G), and Annexin V staining revealed rezatapopt-induced apoptosis for both mutants (Fig. 6H), again requiring higher doses for the Y220S mutant.

Implications for the development of clinical pan-Y220C/N/S reactivators

Our data show that rezatapopt induces the same response in Y220S cells as in Y220C cells but that its cellular potency is reduced by more than one order of magnitude. Cancer patients with a Y220S mutation may therefore also benefit from rezatapopt treatment, but the higher concentrations needed to elicit a p53 reactivation may be problematic, with potential off-target activity and side effects due to dose-limiting toxicity; especially given the already high dosage of rezatapopt (2000 mg once-daily) in the current clinical trials for patients with Y220C mutation. In contrast, the thermal stabilization of the Y220N mutant upon rezatapopt treatment is not sufficient to fully compensate for the mutation-induced stability loss to a level that restores a wild-type-like folding state in cells at 37 °C and reactivation of the p53 signaling program at pharmacologically relevant concentrations.

Notably, emerging preclinical evidence challenges the longstanding notion that full restoration of canonical p53 signaling is essential for therapeutic success. Partial reactivation, even at levels insufficient to trigger classical apoptosis, can nonetheless drive tumor regression through alternative pathways. In breast cancer, the efficacy of CDK4/6 inhibitors correlates with p53-DREAM-mediated senescence, where geroconversion (the transition to irreversible senescence) depends on the activation of this axis [39]. This concept is further illustrated in mouse models of Y220C mutant tumors treated with rezatapopt: although activation of canonical p53 target genes is transient, lasting mere hours post-administration, the p53-DREAM axis facilitates sustained repression of proliferation genes. This suggests that senescence induction, augmented by immune-mediated clearance of senescent cells in immunocompetent environments, is the primary therapeutic mechanism [23]. These results align with our prior research on modeling partial p53 recovery in hematologic malignancies, where even incomplete functional restoration led to enduring anti-proliferative effects and immune activation, culminating in disease regression [40]. Consequently, even modest stabilization of Y220 mutants may suffice to engage senescence or immune rejection pathways, offering clinical advantages. This possibility appears promising, particularly when considered alongside adjuvant strategies such as immune checkpoint inhibitors, which may further amplify these mechanisms.

Our data suggest that the development of potent pan-Y220C/N/S reactivators is challenging but not impossible, and provide a potential roadmap for achieving this goal. Key for the design of pan-Y220C/N/S reactivators is to start from a basic scaffold that positively engages with the mutated side chain of each variant, possibly via a CF_3_ moiety, and does not induce a flip as in the case of rezatapopt that pushed the asparagine side chain of the Y220N mutant into an energetically unfavorable hydrophobic environment. In this context, it is noteworthy that the CF_3_ moiety of the carbazole-based PK9301 (compound 3 in Fig. 1), for which the position of the CF_3_ anchor in the binding pocket is slightly shifted relative to that in the rezatapopt complex, directly interacts with the side chain of Ser220 (PDB entry 6SI3), resulting in nearly equipotent binding to the Y220C and Y220S mutants [29]. Future design efforts on the Y220C mutant that also profile the binding to the Y220S and Y220N mutants early in the development of the basic scaffold have a good chance of yielding potent pan-Y220C/N/S reactivators. Our structural data also provide a framework for the rational design of novel bifunctional compound modalities that selectively target Y220X mutants, including mutant-specific molecular glue compounds [41] and proximity-inducing drugs that enable targeted delivery of cytotoxic effector proteins [42].

Materials and methods

Chemical compounds

Rezatapopt and analogs 7 and 8 were purchased from MedChemExpress (Cat. No.: HY-156633, HY-148416, and HY-145759, respectively). Rezatapopt precursor 5 was synthesized following a published protocol [43]. Detailed synthesis and analytical data for this compound are given in the Supporting Information.

Protein expression and purification

Y220C/N/S mutant p53 DBDs (residues 94-312, containing the stabilizing mutations M133L, V203A, N239Y, and N268D; collectively referred to as QM), as well as the corresponding pseudo-wild-type QM variant, were expressed and purified as described previously [29]. Briefly, for recombinant protein expression in E. coli C41, a pET24a-based vector was used encoding for a fusion protein containing an N-terminal hexahistidine tag, followed by the lipoyl domain of Bacillus stearothermophilus dihydrolipoamide acetyltransferase (Uniprot entry P11961, residues 2-85) for improved solubility, a TEV protease cleavage site, and the human p53 DBD variant of interest (residues 94-312). The DBD mutants were purified at 4 °C by Ni-NTA chromatography, followed by TEV protease cleavage overnight, affinity chromatography using a heparin column, and a final size-exclusion chromatography step on a Superdex 75 column (final buffer 25 mM HEPES, pH 7.5, 150 mM NaCl, 0.5 mM TCEP).

X-ray crystallography

Crystals of Y220C/N/S mutants of the human p53 DBD (with the stabilizing mutations M133L, V203A, N239Y, and N268D; [34]) were grown by vapor diffusion using the sitting drop technique. Y220C mutant crystals were grown at 20 °C by microseeding. Protein buffer: 6 mg/mL in 25 mM HEPES, pH 7.5, 150 mM NaCl, 0.5 mM TCEP. Reservoir buffer: 100 mM HEPES, pH 7.0, 19% (w/v) polyethylene glycol 4000. They were soaked for 2-3 h with 17-19 mM compound in 100 mM HEPES, pH 7.2, 10 mM sodium phosphate, pH 7.2, 19% (w/v) polyethylene glycol 4000, 20% (v/v) glycerol, 150 mM KCl.

Complexes of the Y220S and Y220N mutants with rezatapopt were obtained by cocrystallization at 20 °C, mixing protein solution (6 mg/mL in 25 mM HEPES, pH 7.5, 150 mM NaCl, 0.5 mM TCEP, 1 mM rezatapopt, 2% DMSO) with the following reservoir buffers: Y220S, 0.2 M ammonium nitrate, 20% PEG 3350 (2:1 ratio); Y220N, 10% PEG 8000, 8% ethylene glycol, 0.1 M HEPES pH 7.5 (1:1 ratio). Crystals of the ligand-free Y220N mutant were grown at 4 °C with 22.5% (w/v) BMW PEG smear broad, 0.1 M potassium sodium tartrate, 10% (w/v) ethylene glycol, 0.1 M sodium cacodylate, pH 5.3-5.8. The Y220S and Y220N crystals were cryoprotected with reservoir buffer supplemented with 23% ethylene glycol. Crystals were flash frozen in liquid nitrogen, and X-ray datasets were collected at 100 K at beamline X06SA of the Swiss Light Source, Switzerland and beamlines I03 and I04 of the Diamond Light Source, UK. Diffraction data were integrated with XDS [44] and scaled with AIMLESS [45], which is part of the CCP4 package [46]. The structures were solved by difference Fourier methods in PHENIX [47] with PDB entry 6SHZ as a starting model and initial rigid body refinement. The structures were then refined using iterative cycles of manual model building in COOT [48] and refinement in PHENIX [47]. Restraint dictionaries for the ligands were generated using the Grade Web Server (http://grade.globalphasing.org). The geometry of the final models was validated using MolProbity [49]. Data collection and refinement statistics are listed in Supplementary Table S2. Structural figures were prepared using PyMOL (www.pymol.org).

Differential scanning fluorimetry (DSF)

Melting temperatures, Tm values, of p53 DBD variants were determined by DSF using an Agilent MX3005P real-time qPCR instrument (excitation/emission filters = 492/610 nm). Measurements were performed in a 96-well plate with an assay buffer consisting of 5 μM mutant DBD in 25 mM HEPES, pH 7.5, 150 mM NaCl, 0.5 mM TCEP, 2.5% (v/v) DMSO, and varying concentrations of small-molecule stabilizers. The fluorescent dye SYPRO Orange (5000×, Invitrogen) was added at a dilution of 1:1000 (total volume of 20 μL per well). The fluorescence signal was then monitored upon temperature increase from 25 to 95 °C, at a heating rate of 3 °C/min, and Tm values were calculated after fitting the fluorescence curves to the Boltzmann function. DSF measurements were performed in at least three technical replicates, and Tm values are given as the mean with standard deviation.

Isothermal titration calorimetry (ITC)

ITC experiments were performed using an Affinity ITC LV (low volume) instrument (TA Instruments) at 20 °C with a stirring rate of 75 rpm. The cell unit contained 30 μM Y220 mutant DBD in a 25 mM HEPES (pH 7.5), 150 mM NaCl, 0.5 mM TCEP and 2% (v/v) DMSO assay buffer. The syringe contained 240 μM compound in the same buffer, using injection steps of 2 μL (first injection 0.5 μL) and an injection interval of 200 s; 25 injections in total. Data analysis was performed using the NanoAnalyze Data Analysis software (Version 3.10.0) supplied with the instrument.

Generation of cell lines

For the generation of plasmids for conditional expression of p53 mutants, WT p53 cDNA was PCR-amplified using the primers 5‘-ATG GAG GAG CCG CAG TCA GAT-3‘ and 5‘-TCA GTC TGA GTC AGG CCC TTC-3‘ and cloned into a Gateway™ entry vector using the pENTR™/D-TOPO™ Cloning Kit (Thermo Scientific, K240020SP) according to the manufacturer’s protocol. Y220C (chr17:7,674,872T>C), Y220S (chr17:7,674,872T>G), Y220N (chr17:7,674,873A>T, and Y220H (chr17:7,674,873A>G) mutations were introduced using the QuickChange Multi Site Directed Mutagenesis Kit (Agilent, 210513) according to the manufacturer’s protocol using the following primers. Y220C: 5‘-AGT GTG GTG GTG CCC TGT GAG CCG CCT GAG GT-3‘, Y220S: 5‘-AGT GTG GTG GTG CCC TCT GAG CCG CCT GAG GT-3‘, Y220N: 5‘-AGT GTG GTG GTG CCC AAT GAG CCG CCT GAG GT-3‘, Y220H: 5‘-AGT GTG GTG GTG CCC CAT GAG CCG CCT GAG GT-3‘. Mutated cDNAs were subsequently transferred into pINDUCER20 [50] plasmid using Gateway™ LR Clonase™ II Enzyme-Mix (Invitrogen, 11791020) according to the manufacturer’s protocol. For lentivirus production, HEK-293T cells (ATCC, CRL-3216) were transfected with lentiviral plasmids and the helper plasmids pMD2.g and psPAX2 [51] using PEI transfection reagent (MedChemExpress, HY-K2014) according to manufacturer’s protocol. Supernatants containing lentivirus were collected 48 and 72 h after transfection, filtered through 0.45 μm filters (Sarstedt, 83.1826), supplemented with 8 μg/mL hexadimethrine bromide (Sigma, 107689) and used immediately for spin-infection (1 h, 1500 rpm at 37 °C) of H1299 cells (ATCC, CRL-5803). Transduced cells were selected with 500 µg/mL neomycin (Gibco, 10131027). Expression of p53 mutants was induced with 500–1000 ng/mL doxycycline (Sigma, D9891) for 48 h and maintained during rezatapopt treatment.

Immunoblot

Cells were treated with 0.1–2 µM rezatapopt or DMSO as solvent control for 48 h. Cells were lysed in NP-40 Buffer (50 mM Tris-HCl, 150 mM NaCl, 5 mM EDTA, 2% NP-40, pH 8.0) supplemented with cOmplete ULTRA protease inhibitor cocktail (Roche, 4693124001) and sonicated using a Bioruptor (Diagenode) for 5 × 30 s. 20 μg protein samples were prepared using NuPAGE™ LDS Sample Buffer (Invitrogen, NP0007) and NuPAGE™ Sample Reducing Agent (Invitrogen, NP0009) and separated on NuPAGE 4–12% Bis-Tris polyacrylamide gels (Invitrogen, WG1402) using MOPS buffer (Invitrogen, NP0001). Following transfer to Immun-Blot PVDF Membrane (BioRad, 1620177) using NuPAGE™ Transfer Buffer (Invitrogen, NP00061), antigens were detected using the antibodies: p53 (Santa Cruz Biotechnology, sc-126; antibody DO-1; 1:1000), p21 (Santa Cruz Biotechnology, sc-6246; 1:200), and β-actin (Abcam, ab6276; 1:2500). Detection was performed using ImageLab (v6.0.1) with secondary goat anti-mouse IgG Fc HRP antibody (Invitrogen, A16084; 1:2500) and WesternBright Sirius chemiluminescent HRP conjugate (Advansta, K-12043). β-actin was detected using goat anti-mouse Alexa-488 conjugate (Invitrogen, A-11029; 1:2500).

Proliferation/IC50 assay

2 ×10^4^ cells were seeded into 96-well plates (Sarstedt, 83.3925). After 24 h, cells were treated with 102.4 nM–20 µM rezatapopt or DMSO as solvent control. Cells were imaged every 4 h using the IncuCyte S3 live-cell analysis system (Sartorius). Confluence data were collected using IncuCyte S3 software (v2018A) and analyzed using the area under the curve (AUC) to calculate IC_50_ values. Alternatively, cells were lysed and quantified after 72 h treatment using CellTiter-Glo® Luminescent Cell Viability Assay (Promega, G7570) according to the manufacturer’s protocol.

Reverse transcription quantitative PCR

Cells were treated with 0.25–64 µM rezatapopt or DMSO as solvent control for 24 h. After 24 h, total RNA was isolated using RNeasy Mini Kit and reverse transcribed using SuperScript™ VILO™ cDNA Synthesis Kit (Invitrogen, 11754050) according to the manufacturer’s protocol. Resulting cDNA was used for qPCR using ABsolute qPCR Mix SYBR Green (Thermo Scientific, AB1158B) using the following primers: CDKN1A fw 5’-TGG AGA CTC TCA GGG TCG AAA-3’, CDKN1A rev 5’-CCG GCG TTT GGA GTG GTA-3’, MDM2 fw 5’- CAT GCA ATG AAA TGA ATC CC-3’, MDM2 rev 5’-GGA AGC CAA TTC TCA CGA AG-3’, BAX fw 5’-GGG TTG TCG CCC TTT TCT ACT T, BAX rev 5’-AGC CCA TGA TGG TTC TGA TCA G, BBC3 fw 5’-ACC TCA ACG CAC AGT ACG AG-3’, BBC3 rev 5’-GAG ATT GTA CAG GAC CCT CCA-3’, GAPDH fw 5’-CTA TAA ATT GAG CCC GCA GCC-3’, GAPDH rev 5’-ACC AAA TCC GTT GAC TCC GA-3’. Data were analyzed using the ΔΔC_t_ method.

Immunoprecipitation

Cells were treated with 1–64 µM rezatapopt or DMSO as solvent control for 4 h. Cells were harvested, lysed in cold IP buffer (50 mM Tris-HCl, 150 mM NaCl, 1 mM EDTA, 1% NP-40, pH 8.0) and sonicated using a Bioruptor (Diagenode) for 3 × 10 s. Lysates were diluted to 1 mg/mL with cold IP buffer. 350 µL lysate were incubated with 50 µL Protein G Sepharose® 4 Fast Flow (Merck, GE17-0618-01) and 3 µg antibody (folded p53: PAb1620, Merck, MABE339; or unfolded p53: PAb240, Merck, OP29L) rotating for 4 h at 4 °C. Beads were washed twice with 1 mL IP buffer (RT for PAb1620, 4 °C for PAb240) before boiling the beads in 20 µL SDS loading buffer consisting of NuPAGE™ LDS Sample Buffer (Invitrogen, NP0007) and NuPAGE™ Sample Reducing Agent (Invitrogen, NP0009). Input samples were prepared with 17.5 µg protein (5% of final IP). Immunoblot was performed as described above, using Pierce™ peroxidase-conjugated recombinant Protein A/G (Thermo Scientific, 32490, 1:5000) instead of secondary antibody.

Colony formation assay

5 × 10⁴ cells were seeded per 6 cm plate (Sarstedt, 83.3901). After 24 h, cells were treated with 0.1–20 µM rezatapopt or DMSO as a solvent control. Treatments were refreshed every 3 days until control plates reached confluency. Plates were then fixed with 70% ethanol for 30 min at RT, before staining for 15 min with 0.2% crystal violet in 10% ethanol (Sigma-Aldrich, HT90132). Plates were scanned using Epson Scan (v3.24 G), and images were processed with Adobe Photoshop CS6 (v13.0.1). For quantification, the stain was eluted with 20% acetic acid, and absorbance was measured at 590 nm using a CYTATION3 imaging plate reader with Gen5 software (v3.08).

Apoptosis assay

Cells were treated with 0.1–20 µM rezatapopt or DMSO as solvent control for 48 h. Cells and media supernatants were collected, pelleted, and resuspended in Annexin V-APC conjugate (MabTag, AnxA100) diluted in Annexin V binding buffer (BD Biosciences, 556454) according to the manufacturer’s protocol. The suspension was incubated in the dark for 20 min at RT, washed in Annexin V binding buffer and analyzed by flow cytometry (Beckman-Coulter Cytoflex LX Cytometer using CytExpert (v2.6) Flow cytometry data were analyzed using FlowJo (v10.8.1).

Cell-cycle analysis

Cells were treated with 0.1–20 µM rezatapopt or DMSO as solvent control for 48 h. For EdU incorporation, cells were incubated with 10 µM 5-ethynyl-2′-deoxyuridine (EdU; baseclick, BCN-001) for 2 h prior to collection of cells, media and supernatants, and fixation with ice-cold 70% ethanol over night at −20 °C. Following fixation, cells were washed with blocking buffer (1% BSA in PBS), then permeabilized with 0.5% Triton X-100 in PBS for 30 min at RT. The EdU Click-iT reaction was performed by incubating cells for 30 min at RT in the dark in 250 µL Click-iT reaction mix (4 mM CuSO₄, 1 µM Eterneon-Red 645 Azide (baseclick, BCFA-201), 100 mM ascorbic acid in PBS). After the reaction, cells were washed once with blocking buffer and once with 0.1 M glycine in PBS, followed by staining with 20 μg/mL propidium iodide (Invitrogen, P3566) in the presence of 100 μg/mL RNaseA (AppliChem, A3832). Cells were analyzed by flow cytometry (Beckman-Coulter Cytoflex LX Cytometer using CytExpert (v2.6). Flow cytometry data were analyzed using FlowJo (v10.8.1).

Supplementary information

Supporting Information uncropped western blots