Citrullination negatively regulates the functions of the p53 protein and opposes its ubiquitination and degradation

Significance

This study examines the effects of citrullination, a posttranslational modification catalyzed by peptidylarginine deiminase 4 (PAD4), on the tumor suppressor protein p53. The findings demonstrate that PAD4 catalyzes citrullination of p53, primarily targeting arginine residues within its DNA-binding domain. This modification markedly impairs p53’s ability to form its essential tetrameric structure and bind DNA, thereby compromising its transcriptional activity. Citrullination further disrupts p53’s regulation of cell cycle progression and apoptosis and interferes with its interaction with the E3 ubiquitin ligase MDM2, resulting in reduced ubiquitination and slower degradation of p53. Collectively, these results identify a previously unrecognized layer of p53 regulation and suggest that targeting PAD4 may represent a promising therapeutic approach in cancer.

Abstract

This study investigates the regulatory role of peptidylarginine deiminase 4 (PAD4)-mediated citrullination on the tumor suppressor protein p53. We demonstrate that p53 serves as a substrate for PAD4, undergoing citrullination at multiple arginine residues, including critical sites within its DNA-binding domain. Mass spectrometry identified eight citrullination sites, notably R158, R282, and R283, which were further validated in various cancer cell lines. Functional studies revealed that citrullination significantly impairs p53’s ability to form stable tetramers, essential for high-affinity DNA binding. Electrophoretic mobility shift assays and analytical ultracentrifugation confirmed reduced binding to consensus sequences in the p21 and MDM2 promoters. As a result, citrullination led to marked reductions in p21 and MDM2 transcriptional activation and altered regulation of ME2, as demonstrated by reporter assays and quantitative PCR. In addition, citrullination compromised p53’s roles in cell cycle control and apoptosis. Supporting these findings, citrulline-mimic mutants (arginine-to-glutamine substitutions) exhibited diminished transcriptional activity relative to wild-type p53. Furthermore, citrullination disrupted the interaction between p53 and its E3 ubiquitin ligase MDM2, reducing p53 ubiquitination and degradation, as shown by in vitro ubiquitination assays and cycloheximide chase experiments. Importantly, replacing glutamine with lysine at these key sites largely restored p53 activity, indicating that the loss of positive charge is central to the functional consequences of citrullination. Together, these findings identify PAD4-catalyzed citrullination as a regulatory mechanism that modulates p53 function and highlight PAD4 as a potential therapeutic target in cancer.

p53, often called the “guardian of the genome,” is a key tumor suppressor that regulates genes involved in cell cycle arrest, apoptosis, DNA repair, and senescence (1–4). Mutated in over half of human cancers, p53 plays a critical role in preventing malignant transformation (1, 5, 6). Its activity is tightly regulated by diverse posttranslational modifications (PTMs), including phosphorylation, acetylation, and ubiquitination, which affect its DNA binding, stability, localization, and protein interactions (7). In response to genotoxic stress, p53 undergoes rapid PTMs that enhance nuclear translocation and DNA binding (8, 9). Among these, phosphorylation is crucial for promoting tetramerization, stability, and transcriptional activity, while also fine-tuning cofactor interactions for precise control of tumor suppression (10, 11).

The mechanisms by which various kinases phosphorylate p53 have been well characterized (10). Around ten kinases can modify p53 at nine serine/threonine residues, modulating its apoptotic function (12–18). Among these, phosphorylation at Ser46 is especially important for triggering apoptosis in response to DNA damage (14, 17, 18). p53 is also extensively acetylated, a modification essential for its stress response (19, 20). Acetylation at Lys382 enhances DNA binding and activation of DNA damage response genes (21), while SIRT1, a NAD⁺-dependent deacetylase, reverses this modification, reducing p53 activity during aging and stress (22).

Peptidylarginine deiminase 4 (PAD4) catalyzes the conversion of peptidylarginine to peptidylcitrulline, a posttranslational modification known as citrullination. This process influences inflammation, differentiation, apoptosis, and histone gene regulation (23, 24). Although citrullination adds only 0.98 Da in mass, the loss of positive charge can significantly impact protein structure, function, and degradation (25, 26). Aberrant PAD4 activity is linked to diseases such as rheumatoid arthritis and cancer, where its expression is often elevated (27–29). Uniquely among PAD enzymes, PAD4 contains a nuclear localization signal, enabling it to regulate chromatin via histone citrullination (30). Elevated PAD4 and citrullination are commonly observed in tumors, and PAD4 is increasingly recognized as a promising therapeutic target in cancer (24, 31–33).

PAD4 regulates gene expression by citrullinating histone tails, altering chromatin structure to promote a repressive state in coordination with HDAC1 and HDAC2 (34). This contributes to the corepression of estrogen-responsive and p53 target gene promoters, including p21, GADD45, and PUMA (34, 35). PAD4 also citrullinates the histone chaperone nucleophosmin, promoting apoptosis via a p53-dependent pathway following DNA damage (36, 37). In addition, PAD4-mediated citrullination of the growth inhibitor ING4 disrupts its interaction with p53, reducing p53 acetylation and transcriptional activity (38). These findings highlight PAD4’s multifaceted role in modulating p53-dependent gene expression through both histone and nonhistone targets. Moreover, PAD4 and HDACs act as transcriptional corepressors of p53 via protein–protein interactions and histone deacetylation (34, 35).

This study reveals that PAD4 regulates p53 activity through direct citrullination. We show that p53 is a substrate of PAD4 in vitro and in cells and that citrullination impairs its tetramer formation, DNA binding, and target gene expression. This modification also disrupts p53-mediated cell cycle arrest and apoptosis. Moreover, citrullination interferes with p53 ubiquitination and degradation by weakening its interaction with MDM2. Together, these findings identify PAD4-mediated citrullination as a mechanism controlling p53 function.

Results

p53 Is a Direct Substrate of PAD4-Mediated Citrullination.

Previous studies have shown that PAD4 regulates p53 target gene expression by citrullinating histones and interacting with p53 and histone-modifying enzymes such as HDAC1, HDAC2, and PRMT1 (34, 35). Protein–protein interaction assays identified direct binding between the N-terminal domain of PAD4 (residues 1 to 300) and the C-terminal domain of p53 (residues 301 to 393) (34, 35). Unexpectedly, we found that p53 is not only a binding partner but also a direct substrate of PAD4 enzymatic activity. To test this, we performed kinetic assays comparing wild-type PAD4 (PAD4_WT) with a catalytically inactive mutant (PAD4_C645S), using a histone H4 peptide as a positive control (SI Appendix, Fig. S1A). PAD4_WT showed a dose-dependent increase in activity with H4 peptide, while PAD4_C645S had minimal activity. We then used recombinant full-length p53 as a substrate. PAD4_WT activity increased markedly with p53 concentration, whereas PAD4_C645S remained inactive (Fig. 1A). The catalytic constant (k_cat) of PAD4_WT was approximately five times higher than that of the inactive mutant (Fig. 1A), confirming that p53 is a bona fide substrate of PAD4.

Fig. 1.

The process of in vitro and cellular citrullination of the p53 protein is facilitated by PAD4. (A) The activity of PAD4 was evaluated using purified recombinant p53 protein as the in vitro substrate, alongside either the wild-type PAD4 (PAD4_WT) or the catalytically inactive mutant (PAD4_C645S). The kinetic parameters, K_m and k_cat, are indicated in the lower panel. (B) The diagram highlights the citrullination sites on the p53 protein, as determined through LC–MS/MS analysis, which is also presented in Table 1. (C) In vitro citrullination of p53 was performed using recombinant PAD4 with 10 mM MgCl₂ or CaCl₂ at 25 °C, followed by immunoblotting with anti-PAD4, anti-p53, and anti-cit-p53 antibodies. Citrullination was also assessed in the presence of 100 μM PAD4 inhibitors GSK484 (D) or BMS-P5 (E). (F) MCF-7 cells were treated with 1 μM ionomycin for 24 h. (G and H) HCT116 and A375 cells were treated with varying ionomycin concentrations for 24 h. (I–K) MCF-7, HCT116, and A375 cells were exposed to 10 mM CaCl₂ for 2 h, then treated with GSK484 (I and K) or BMS-P5 (J) for 1 h. Cit-p53/p53 ratios were shown and normalized to GAPDH.

Identification and Validation of p53 Citrullination Sites.

We used LC–MS/MS to identify PAD4-mediated citrullination sites on p53, analyzing both recombinant p53 and p53 isolated from MCF7 cells. Eleven citrullinated arginine residues were identified in recombinant p53, including R110, R158, R174, R181, R209, R280, R282, R283, R290, R306, and R363 (Table 1 and SI Appendix, Fig. S1B). Among these, eight sites including R158, R181, R209, R280, R282, R283, R306, and R363 were also detected in p53 isolated from MCF7 cells (Table 1 and SI Appendix, Fig. S1C). Six of these residues are located in the DNA-binding domain (Fig. 1B), suggesting potential effects on p53 function. To validate these sites, we generated polyclonal antibodies against citrullinated p53 peptides at R110, R158, R282/283, R306, and R363. Immunoblotting following in vitro citrullination confirmed their specificity (SI Appendix, Fig. S1 D–H). Recombinant p53 was incubated with PAD4 under four conditions: untreated (Set 1), CaCl₂ (Set 2), GSK484 (Set 3), and CaCl₂ plus GSK484 (Set 4). Citrullinated p53 was detected only in Set 2, consistent with PAD4’s calcium dependence. Among the antibodies, only the anti-cit-R363 antibody specifically recognized citrullinated p53 and not the R363Q or R363K mutants (SI Appendix, Fig. S1 H–I) and was selected for further use. Consistent with prior findings that PAD4 is not activated by Mg²⁺ (39), citrullinated p53 was detected only with CaCl₂, showing a time-dependent increase (Fig. 1C). PAD4 inhibitors GSK484, BMS-P5, and BB-Cl-amidine effectively suppressed p53 citrullination (Fig. 1 D–E and SI Appendix, Fig. S1J).

Table 1.

The citrullination sites of p53 protein identified by LC–MS/MS

Modified peptide sequence	Site	G2 (in vitro)	Velos (in cells)
K.TYQGSYGFRLGFLHSGTAK.S	R110	V†	ND†
R.VRAMAIYK.Q	R158	V	V
K.QSQHMTEVVRR.C	R174	V	ND
R.CPHHERCSDSDGLAPPQHLIR.V	R181	V	V
R.VEYLDDRNTFR.H	R209	V	V
R.VCACPGRDRR.T	R280/R282	V	V
R.DRRTEEENLR.K	R282/R283	V	V
R.RTEEENLRK.K	R283/R290	V	ND
K.RALPNNTSSSPQPK.K	R306	V	V
K.DAQAGKEPGGSRAHSSHLK.S	R363	V	V
Sequence coverage (%) *		76	56

^*The best sequence coverage among all experiments is reported.

^†“V”—valid; “ND”—not detected. All evaluations were made by manual inspection according to spectrum quality and relative sequence-determining fragment ions and neutral loss ions in MS2 spectra.

PAD4-Dependent Citrullination of p53 in Cancer Cells.

We next examined citrullinated p53 (cit-p53) in cells and detected its expression across multiple cancer cell lines (Fig. 1 F–K). In MCF-7 cells, ionomycin treatment, which enhances PAD4 activity, increased the cit-p53 to total p53 ratio (cit-p53/p53) to 1.79 (Fig. 1F). Similar increases were observed in HCT116 and A375 cells (Fig. 1 G and H and SI Appendix, Fig. S1K), as well as in Jurkat cells (SI Appendix, Fig. S1 L and M). To assess PAD4 specificity, we treated cells with PAD4 inhibitors. GSK484 reduced cit-p53 levels in MCF-7 and A375 cells (Fig. 1 I and K and SI Appendix, Fig. S1N), while BMS-P5 and BB-Cl-amidine suppressed cit-p53 in HCT116 cells (Fig. 1J and SI Appendix, Fig. S1O). GSK484 also decreased cit-p53 in Jurkat cells (SI Appendix, Fig. S1 P and Q). These results confirm that PAD4 catalyzes p53 citrullination in both in vitro and cellular settings. PAD4 activation increases citrullinated p53, while its inhibition suppresses this modification, indicating a positive correlation between PAD4 activity and cit-p53 levels.

PAD4 Activity and Expression Directly Regulate the Levels of Citrullinated p53 Protein.

To determine whether PAD4 expression directly influences p53 citrullination, we transfected cells with either wild-type PAD4 (PAD4_WT) or the catalytically inactive mutant PAD4_C645S. In HCT116 cells, PAD4_WT overexpression significantly increased citrullinated p53 levels, with a cit-p53/p53 ratio of 1.46, whereas PAD4_C645S had minimal effect (ratio 1.13) (Fig. 2 A and B). A similar pattern was observed in MCF-7 cells, where PAD4_WT raised the ratio to 2.02, compared to 1.12 for PAD4_C645S (Fig. 2 C and D). This phenomenon was also observed in A375 cells (SI Appendix, Fig. S2 A–D). These findings confirm that both PAD4 expression and its catalytic activity are required for p53 citrullination.

Fig. 2.

Citrullination of the p53 protein, facilitated by the enzyme PAD4, has been observed in multiple cell lines. (A and C) HCT116 and MCF-7 cells were transfected with PAD4_WT or PAD4_C645S plasmids for 24 h, followed by immunoblotting for PAD4, p53, and cit-p53. (B and D) Quantified cit-p53/p53 ratios are shown. (E and G) PAD4 was silenced in HCT116 and A375 cells using siRNA for 48 h; corresponding cit-p53/p53 ratios are shown in (F and H). (I and K) PAD4-overexpressing HCT116 and MCF-7 cells were treated with GSK484; corresponding cit-p53/p53 ratios are shown in (J and L). All ratios were normalized to GAPDH and presented as mean ± SD from three independent experiments. A Student’s t test was conducted for statistical analysis. Significance levels were set at *P < 0.05, **P < 0.01, and ***P < 0.001.

We next assessed the effect of PAD4 knockdown on p53 citrullination. In HCT116 cells, silencing endogenous PAD4 with two distinct siRNAs significantly reduced citrullinated p53 levels, lowering the cit-p53/p53 ratios to 0.51 and 0.32, respectively (Fig. 2 E and F). Similar reductions were observed in A375 cells, with ratios decreasing to 0.38 and 0.26 following siRNA treatment (Fig. 2 G and H). In stable PAD4-overexpressing HCT116 and A375 cells, subsequent PAD4 knockdown also led to reduced citrullinated p53 levels (SI Appendix, Fig. S2 E–H), further confirming that PAD4 is required for p53 citrullination.

We further confirmed that PAD4 inhibition reduces citrullinated p53 levels in PAD4-overexpressing cells. In HCT116 cells, GSK484 treatment led to a dose-dependent decrease in citrullinated p53, lowering the cit-p53/p53 ratio to 0.40 (Fig. 2 I and J). Similar effects were observed in PAD4-overexpressing MCF-7 cells, where GSK484 also reduced the ratio to 0.40 (Fig. 2 K and L). A similar reduction was observed in PAD4-overexpressing A375 cells treated with GSK484 (SI Appendix, Fig. S2 I and J), and comparable effects were seen with BMS-P5 (SI Appendix, Fig. S2 K and L). These findings demonstrate that PAD4-induced citrullination of p53 is effectively suppressed by PAD4 inhibitors. Together, our data show that p53 citrullination is directly regulated by PAD4 expression and catalytic activity in multiple cancer cell lines.

Citrullination of the p53 Protein Negatively Impacts Its Ability to Bind DNA and Its Transcriptional Activity.

We next investigated how citrullination affects p53 function by analyzing its quaternary structure in the presence or absence of DNA using analytical ultracentrifugation (AUC). Native p53 bound to DNA containing consensus binding sites and formed p53–DNA complexes. To assess the impact of citrullination, we examined p53 binding to promoter sequences of its target genes, p21 and MDM2 (referred to as DNA_p21 and DNA_mdm2; Fig. 3 A and B). Without DNA_p21, p53 existed in a monomer–dimer equilibrium (purple dashed line, Fig. 3 A, Upper panel). Upon DNA_p21 binding, p53 formed tetramers (green line, Fig. 3 A, Upper panel). In contrast, citrullinated p53 remained monomeric (purple dashed line, Fig. 3 A, Lower panel) and failed to bind DNA_p21 (green line, Fig. 3 A, Lower panel). Similar results were observed with DNA_mdm2 (Fig. 3B). These findings show that native p53 forms functional DNA-bound tetramers, while citrullinated p53 fails to bind DNA and cannot assemble into tetrameric complexes.

Fig. 3.

Citrullination of the p53 protein results in functional impairment, negatively impacting its ability to form quaternary structures, interact with DNA, and preserve transcriptional activity. (A and B) Size distribution plots of p53 (Upper) and citrullinated p53 (Lower) bound to p21 promoter DNA (DNA_p21) or MDM2 promoter DNA (DNA_mdm2). The MDM2 probe includes two binding sites, yielding a peak corresponding to two p53 tetramers bound to one promoter (2x p53 + DNA_mdm2). (C and D) EMSA showing p53 or cit-p53 binding to DNA_p21 and DNA_mdm2 using biotin-labeled probes detected by streptavidin-HRP. (E and F) Reporter assays in HEK293FT cells cotransfected with PAD4_WT, p53_WT, and either p21 or ME2_RE3 promoter reporters. Luciferase activity was measured after 48 h. (G and H) qPCR analysis of p21 and ME2 mRNA in cells co-overexpressing PAD4 and p53. Data represent mean ± SD from three independent experiments. A Student’s t test was conducted for statistical analysis. Significance levels were set at *P < 0.05, **P < 0.01, and ***P < 0.001.

We further assessed p53–DNA interactions using electrophoretic mobility shift assays (EMSA) with biotin-labeled DNA probes derived from the p21 and MDM2 promoters. Complexes were detected using streptavidin–HRP. As shown in Fig. 3C, native p53 bound to DNA_p21, producing a distinct band shift (lanes 2 to 4) compared to DNA alone (lane 1). In contrast, no shift was observed with citrullinated p53 (lanes 5 to 7). Similar results were obtained using DNA_mdm2 (Fig. 3D). These findings, consistent with the AUC data (Fig. 3 A and B), confirm that citrullination impairs p53’s DNA-binding ability.

We further evaluated p53 transcriptional activity using a reporter assay containing the promoters of p21 (a positively regulated target) and ME2 (a negatively regulated target). p53 is known to repress ME2 expression by binding to two response elements, RE1 and RE3, in its promoter (39). Overexpression of p53 increased p21 promoter-driven luciferase activity, while coexpression with PAD4 modestly reduced this activity (Fig. 3E). In contrast, PAD4 overexpression significantly enhanced luciferase activity from the ME2 promoter (RE1 and RE3), indicating that citrullination impairs p53’s repressive function on ME2 (Fig. 3F and SI Appendix, Fig. S3A).

We also examined the effect of PAD4 overexpression on endogenous p21 and ME2 expression using qPCR. Consistent with reporter assay results, p53 overexpression increased p21 mRNA levels, while coexpression with PAD4 reduced them (Fig. 3G). Conversely, ME2 mRNA levels were suppressed by p53 overexpression but restored upon PAD4 coexpression (Fig. 3H). Protein levels followed the same trend: p21 decreased and ME2 increased with PAD4 overexpression (SI Appendix, Fig. S3 B and C). Together, these findings indicate that citrullination impairs p53 function by disrupting its tetramerization, DNA binding, and transcriptional activity.

Citrullination at Residues R158, R282, and R283 Impairs the Tetramerization and DNA-Binding Ability of the p53 Protein.

Mass spectrometry identified eight arginine residues as potential citrullination sites in both in vitro and cellular samples. To evaluate their functional relevance, we generated citrulline-mimic mutants by substituting glutamine for arginine at R158, R181, R209, R280, R282, R283, R306, and R363. This approach enabled us to assess how citrullination affects p53 tetramerization and DNA binding at the p21 and MDM2 promoters.

Among the citrulline-mimic mutants, p53_R158Q showed the most severe functional impairment. Unlike wild-type p53 (Fig. 3 A and B), p53_R158Q remained monomeric even in the presence of DNA_p21 or DNA_mdm2, resulting in a marked loss of DNA binding (Fig. 4 A–C and E). In contrast, p53_R363Q behaved like p53_WT, forming tetramers and effectively binding both promoters (Fig. 4 A, B, D, and E). Additional mutants revealed varying degrees of disruption. p53_R280Q remained dimeric with p21 DNA (SI Appendix, Fig. S4A), and p53_R283Q showed weak tetramer formation (SI Appendix, Fig. S4B). Similar patterns were seen with p53_R280Q, R282Q, and R283Q at the MDM2 promoter (SI Appendix, Fig. S4 C and D). EMSA confirmed these observations (SI Appendix, Fig. S4 E–M). p53_R280Q retained DNA_p21 binding (SI Appendix, Fig. S4F), consistent with its dimeric form. In contrast, p53_R282Q and R283Q exhibited reduced binding to both p21 and MDM2 promoters (Fig. 4A and SI Appendix, Fig. S4 G, K, and L). These results highlight R158 citrullination as a major disruptor of p53 tetramerization and DNA binding, with R282 and R283 contributing to partial functional loss.

Fig. 4.

The quaternary structure, DNA-binding capacity, and transcriptional activity of the citrulline-mimic p53 mutants. (A and B) Size distribution plots of p53_R158Q (Upper) and p53_R363Q (Lower) bound to p21 (DNA_p21) or MDM2 (DNA_mdm2) promoter DNA. (C–E) EMSA showing DNA binding of p53_R158Q and p53_R363Q to DNA_p21 and DNA_mdm2 at molar ratios of 2:1, 4:1, and 6:1. (F and G) Reporter assays in HCT116_p53^−/− (p21 promoter; n = 4) and HEK293FT (ME2_RE1 promoter; n = 3) cells cotransfected with citrulline-mimic p53 mutants. Luciferase activity was measured at 24 h and normalized to p53 mutant protein levels (Fig. S4 O and P). Statistical significance was assessed by Student’s t test. Significance levels were set at *P < 0.05, **P < 0.01, and ***P < 0.001. All black stars represent comparisons with p53_WT, whereas red stars signify comparisons among the designated groups.

To further assess how citrullination at R158, R282, and R283 affects p53 transcriptional activity, we performed luciferase reporter assays using the p21 and ME2 promoters (Fig. 4 F and G and SI Appendix, Fig. S4N). Luciferase signals were normalized to p53 protein levels (SI Appendix, Fig. S4 O and P). Both p53_R158Q and p53_R282Q/R283Q mutants showed markedly reduced transcriptional activity at the p21 promoter (Fig. 4F). Replacing glutamine with lysine in these mutants (p53_R158K and p53_R282K/R283K) largely restored activity, suggesting that positive charge at these sites is essential. Similarly, these mutants failed to repress ME2 promoter activity at RE1 and RE3 elements (Fig. 4G and SI Appendix, Fig. S4N), while the lysine-substituted forms restored inhibitory function. These results indicate that citrullination impairs p53-mediated transcription by disrupting key charge-dependent interactions.

Among the eight citrullination sites, R306 and R363 lie outside p53’s core DNA-binding domain. Reporter assays showed that p53_R363Q retained transcriptional activity comparable to p53_WT at both the p21 and ME2 promoters (Fig. 4 F and G and SI Appendix, Fig. S4N). In contrast, p53_R306Q exhibited enhanced p21 promoter activity but reduced repression at the ME2 promoter, indicating a partial loss of inhibitory function (Fig. 4 F and G and SI Appendix, Fig. S4N). Overall, these results highlight that citrullination at R158, R282, and R283 disrupts p53 tetramerization, DNA binding, and transcriptional regulation, with R158 being especially critical for maintaining p53 activity.

Citrullination at Residues R158, R282, and R283 Results in a Loss of p53’s Ability to Enhance the Expression of Downstream Genes p21 and MDM2.

Building on the finding that citrullination impairs p53 transcriptional activity, we examined how citrulline-mimic and non-citrulline-mimic mutants affect expression of downstream targets p21 and MDM2. As expected, p53_R158Q and p53_R282Q/R283Q showed significantly reduced induction of p21 and MDM2 mRNA (Fig. 5 A and B), whereas the corresponding lysine-substituted mutants, p53_R158K and p53_R282K/R283K, largely restored target gene expression (Fig. 5 A and B). p53_R363Q maintained activity comparable to p53_WT, and p53_R306Q slightly increased p21 and MDM2 expression relative to wild-type p53 (Fig. 5 A and B).

Fig. 5.

The regulation of downstream genes targeted by p53 is affected by citrulline-mimic p53 mutants. (A and B) Relative mRNA levels of p21 and MDM2 in HCT116_p53^−/−cells transfected with citrulline-mimic or non-citrulline-mimic p53 mutants for 24 h. (C and D) Protein levels of p21 and MDM2 measured after 48 h of transfection. All protein ratios were normalized to GAPDH, p21/p53 and MDM2/p53 ratios are shown in green. A Student’s t test was conducted for statistical analysis. Significance levels were set at *P < 0.05, **P < 0.01, and ***P < 0.001. All black stars represent comparisons with p53_WT, whereas red stars signify comparisons among the designated groups.

We next evaluated p21 and MDM2 protein expression in cells expressing different p53 mutants. Both p53_R158Q and p53_R282Q/R283Q showed markedly reduced p21 and MDM2 levels, with significantly lower p21/p53 and MDM2/p53 ratios compared to p53_WT (Fig. 5C and SI Appendix, Fig. S5 A and B). The non-citrulline-mimic mutant p53_R158K fully restored target protein expression, while p53_R282K/R283K showed partial recovery (Fig. 5C and SI Appendix, Fig. S5 A and B). p53_R363Q behaved similarly to p53_WT across all measures (Fig. 5D and SI Appendix, Fig. S5 A and B). Although p53_R306Q displayed higher total p53 levels (Fig. 5D), its p21/p53 and MDM2/p53 ratios remained at 0.5 to 0.6, indicating partial functional impairment (SI Appendix, Fig. S5 A and B).

We further evaluated how PAD4 overexpression affects p53-mediated regulation of downstream targets p21 and ME2. PAD4 overexpression significantly reduced p53_WT’s ability to induce p21 mRNA and to repress ME2 mRNA (SI Appendix, Fig. S5 C and D). In the absence of PAD4 overexpression, the non-citrulline-mimic mutants p53_R158K and p53_R282K/R283K regulated p21 and ME2 expression similarly to p53_WT, enhancing p21 and repressing ME2. However, PAD4 overexpression impaired their transcriptional activity, mirroring the effect on p53_WT (SI Appendix, Fig. S5 C and D). At the protein level, PAD4 overexpression similarly reduced p21 induction and ME2 repression across p53_WT and the non-citrulline-mimic mutants (SI Appendix, Fig. S5 E–G). These results confirm that citrullination at R158, R282, and R283 significantly impairs p53’s ability to regulate its target genes at both mRNA and protein levels.

Citrullination at Residues R158, R282, and R283 Compromises p53’s Regulatory Roles in Cell Cycle and Cell Apoptosis.

Given that citrullination impairs p53’s DNA-binding and transcriptional activity, leading to reduced expression of target genes, we next examined whether this modification affects p53’s ability to regulate cell cycle progression and apoptosis.

We first used qPCR to evaluate the mRNA expression of cell cycle arrest–related genes, including CDKN1B (p27) and GADD45A. In cells overexpressing p53_WT, both genes were significantly upregulated. In contrast, cells expressing the p53_R158Q or p53_R282Q/R283Q mutants showed no increase in CDKN1B or GADD45A expression (Fig. 6 A and B). Notably, the p53_R158K mutant retained the ability to induce these target genes, and the p53_R363Q mutant exhibited transcriptional activity comparable to p53_WT, with a marked elevation in CDKN1B and GADD45A mRNA levels (Fig. 6 A and B). To assess the functional impact on cell cycle regulation, we performed flow cytometry analysis. Consistent with the qPCR results, p53_WT overexpression induced a strong G1 phase arrest, whereas the p53_R158Q and p53_R282Q/R283Q mutants failed to trigger significant arrest (SI Appendix, Fig. S6 A and B). In contrast, the p53_R158K and p53_R363Q mutants preserved their ability to induce G1 phase arrest, mirroring the effect of p53_WT (SI Appendix, Fig. S6 A and B).

Fig. 6.

Citrulline-mimic p53 mutants impair p53 regulation of cell cycle arrest and apoptosis. (A and B) HCT116_p53^−/− cells were transfected with citrulline-mimic p53 mutants for 18 h, then cultured in 1% FBS for 6 h; mRNA levels of CDKN1B and GADD45A were measured. (C and D) Cells were transfected for 24 h, followed by 12 h in 32.5 mM glucose; BAX and Bcl-xL mRNA levels were assessed. (E) Cells were transfected for 24 h, then cultured in 32.5 mM glucose for 24 h; apoptosis was analyzed by Annexin V/PI staining and flow cytometry. (F) Quantification of apoptotic cells. Data represent mean ± SD from three independent experiments. Statistical significance was determined using Student’s t test. Significance levels were set at *P < 0.05, **P < 0.01, and ***P < 0.001. Black asterisks indicate comparisons with p53_WT, while red asterisks indicate comparisons among the designated groups.

We next evaluated whether citrullination affects p53-mediated apoptosis. Overexpression of p53_WT significantly upregulated the proapoptotic gene BAX, while p53_R158Q and p53_R282Q/R283Q mutants showed markedly reduced BAX expression (Fig. 6C). The p53_R158K and p53_R363Q mutants restored BAX expression to levels comparable to p53_WT. Similarly, p53_WT repressed the antiapoptotic gene BCL2L1 (Bcl-xL), whereas this effect was lost in p53_R158Q and p53_R282Q/R283Q, but retained in p53_R158K and p53_R363Q (Fig. 6D). Annexin V/PI staining followed by flow cytometry showed that p53_WT increased apoptosis, while the p53_R158Q and p53_R282Q/R283Q mutants failed to do so (Fig. 6 E and F). In contrast, p53_R158K and p53_R363Q induced apoptosis effectively. These results demonstrate that citrullination at R158, R282, and R283 compromises p53-mediated cell cycle arrest and proapoptotic functions, underscoring the importance of this modification in modulating its tumor suppressor activity.

Citrullination Disrupts the Ubiquitination Process of p53, Thereby Impeding Its Degradation.

Citrullination may affect p53 ubiquitination and stability, as p53 is primarily degraded via the ubiquitin–proteasome pathway. To investigate this, we performed in vitro ubiquitination assays. Native recombinant p53 showed efficient ubiquitination in the presence of ubiquitin and the E3 ligase MDM2, as indicated by characteristic ubiquitin ladders (Fig. 7A). In contrast, citrullinated p53 failed to form these ladders (Fig. 7B), suggesting that citrullination impairs p53 ubiquitination and may block its proteasomal degradation.

Fig. 7.

Ubiquitination and protein degradation of citrulline-mimic p53 proteins. (A and B) In vitro ubiquitination assays were performed with recombinant p53 or cit-p53 proteins incubated with Ub and MDM2 at 30 °C for 90 min; Ub-p53 was detected using anti-p53. (C–F) Citrulline-mimic p53 mutants (R158Q, R283Q, R306Q, R363Q) were similarly assessed. Ub-p53/p53 ratios were quantified using ImageJ. (G–L) HCT116_p53^−/− cells were transfected with p53_WT or citrulline-mimic p53 mutants for 24 h, followed by CHX treatment (150 μg/ml, 3 h) to block protein synthesis. p53 degradation was evaluated by immunoblotting and quantified. All protein ratios were normalized to GAPDH. Half-lives were estimated using one-phase decay fitting in Prism7. Data are presented as mean ± SD from three independent experiments. Statistical significance was assessed by Student’s t test. Significance levels were defined as *P < 0.05, **P < 0.01, and ***P < 0.001.

To identify specific citrullination sites that affect p53 ubiquitination, we performed in vitro assays using citrulline-mimic mutants. Most mutants showed impaired ubiquitination compared to p53_WT (Fig. 7 C–F). Notably, p53_R158Q, p53_R283Q, and p53_R306Q displayed markedly reduced ubiquitination ladders (Fig. 7 C–E), while p53_R363Q showed only a modest reduction (Fig. 7F). Other mutants exhibited variable effects (SI Appendix, Fig. S7). These results suggest that citrullination at specific sites can significantly disrupt p53 ubiquitination.

We assessed the stability and half-life of citrulline-mimic p53 mutants using cycloheximide (CHX) chase assays. Mutants p53_R158Q, p53_R282Q, p53_R283Q, and p53_R306Q exhibited significantly slower degradation than p53_WT (Fig. 7 G–I and K), consistent with their impaired ubiquitination (Fig. 7 C–E and SI Appendix, Fig. S7F). While p53_WT had a half-life of approximately 6 to 12 h, these mutants showed minimal degradation (approximately 10%) over 9 h, with most protein remaining (Fig. 7 G–K). In contrast, p53_R363Q degraded at a rate similar to p53_WT, reflecting normal ubiquitination and turnover (Fig. 7L). Non-citrulline-mimic mutants (R158K, R282K, R283K, R306K, and R363K) also showed degradation kinetics comparable to p53_WT (SI Appendix, Fig. S8).

In summary, citrullination at residues R158, R282, R283, and R306 significantly impairs p53 ubiquitination and degradation, while preserving positive charge at these sites restores normal turnover comparable to p53_WT.

Citrullination Interferes the Interaction Between p53 and MDM2 Protein.

Since MDM2 is the E3 ligase responsible for p53 ubiquitination and degradation, we investigated whether citrullination affects their interaction. Given that citrulline-mimic mutants exhibit impaired ubiquitination and stability (Fig. 7 and SI Appendix, Figs. S7 and S8), we hypothesized that citrullination weakens the p53–MDM2 interaction. Pull-down assay confirmed that ionomycin treatment and PAD4 overexpression significantly weakened p53–MDM2 binding (Fig. 8A and SI Appendix, Fig. S9A). Similarly, coimmunoprecipitation (Co-IP) assay in HCT116 cells, ionomycin treatment markedly decreased the interaction between endogenous p53 and MDM2 (Fig. 8B), with comparable results in HEK293FT cells (SI Appendix, Fig. S9B).

Fig. 8.

Protein–protein interactions between citrulline-mimic p53 mutants and MDM2 proteins. (A) HEK293FT cells were transfected with His-p53 and MDM2, followed by ionomycin treatment for 24 h or cotransfected with PAD4 for 48 h. His-tag pull-down was performed to assess p53–MDM2 interaction by immunoblotting. (B) HCT116 cells treated with ionomycin for 24 h were subjected to p53 immunoprecipitation to examine p53–MDM2 binding. (C and D) Alpha assays were conducted in HEK293FT cells expressing citrulline-mimic p53 mutants, incubated with GST-MDM2 at 25 °C for 1 h. Titration curves were generated to assess binding. (E) B_max values from titration curves indicate MDM2 binding to citrulline-mimic p53. Data are mean ± SD from three independent experiments. Statistical analysis was performed using Student’s t test. Significance levels were set at *P < 0.05, **P < 0.01, and ***P < 0.001. Black stars represent comparisons with p53_WT, and red stars indicate comparisons among designated groups.

We further employed an amplified luminescent proximity homogeneous assay (Alpha) to quantify p53–MDM2 interactions. p53_WT showed a robust, concentration-dependent increase in Alpha signal, following a hyperbolic kinetic profile (black line, Fig. 8C). In contrast, p53_R158Q displayed a significantly reduced interaction, with only about half the maximal signal (B_max) of p53_WT (orange line, Fig. 8 C and E), consistent with impaired ubiquitination. Both p53_R158K and p53_R363Q showed B_max values similar to p53_WT, indicating intact MDM2 binding (Fig. 8 C and E). In contrast, p53_R283Q, p53_R282Q/R283Q, and p53_R306Q showed reduced MDM2 binding, as indicated by lower B_max values compared to p53_WT (Fig. 8 D and E). We assessed the interactions of non-citrulline-mimic mutants p53_R282K, p53_R283K, and p53_R306K with MDM2 (SI Appendix, Fig. S9 C and D). Unlike their glutamine-substituted counterparts, these mutants showed MDM2 binding similar to p53_WT, with comparable B_max values. Although p53_R363Q exhibited slightly elevated Alpha signals, the B_max values did not differ significantly from p53_WT (Fig. 8 C and E).

To further investigate the impact of citrullination, we overexpressed PAD4 to enhance p53 modification and examined its interaction with MDM2. PAD4 overexpression significantly reduced MDM2 binding to p53_WT, as reflected by a marked decrease in B_max (SI Appendix, Fig. S9 E and J). Similarly, under PAD4 overexpression, p53_R158K, p53_R283K, and p53_R306K showed reduced MDM2 interactions and decreased B_max values (SI Appendix, Fig. S9 F–H and J). However, these reductions were milder than those observed for p53_WT, likely because lysine substitution prevents PAD4-mediated citrullination at these sites. In contrast, the p53_R363K mutant behaved similarly to p53_WT, exhibiting a comparable loss of MDM2 binding affinity and a similar reduction in B_max under PAD4 overexpression (SI Appendix, Fig. S9 I and J).

Collectively, these findings indicate that citrullination disrupts the protein–protein interactions between p53 and MDM2, particularly at residues R158, R283, and R306. This observation is consistent with previous in vitro ubiquitination and cycloheximide (CHX) assay results (Fig. 7 and SI Appendix, Figs. S7 and S8), suggesting that impaired interactions at these sites contribute to reduced p53 ubiquitination and, consequently, hinder its degradation.

Discussion

This study provides strong correlative evidence that citrullination at specific arginine residues impairs p53 function. We demonstrate that p53 not only interacts with PAD4 but also serves as a substrate for its enzymatic activity (Fig. 9). Mass spectrometry identified eight citrullinated arginine residues, primarily within the DNA-binding domain (Table 1), suggesting functional consequences. PAD4-dependent citrullination was confirmed across various cancer cell lines, with increased cit-p53 levels upon PAD4 overexpression and decreased levels upon PAD4 inhibition or silencing (Figs. 1 and 2). Citrullination converts positively charged arginines into neutral citrullines, disrupting electrostatic interactions essential for p53 tetramer formation and DNA binding, particularly at R158, R282, and R283 (Figs. 3, 4, and 9 A and B). This modification impairs p53’s transcriptional regulation of downstream targets, including activation of p21 and MDM2, and repression of ME2 (Figs. 4, 5, and 9C). Quantitative PCR and immunoblotting confirmed that mutants p53_R158Q and p53_R282Q/R283Q had diminished target gene activation, which was restored by lysine substitutions (Fig. 5). Citrullination also reduces p53’s ability to induce cell cycle arrest and apoptosis (Fig. 6). Citrullinated p53 fails to undergo normal MDM2-mediated ubiquitination and degradation, as shown by in vitro assays and citrulline-mimic mutants (Figs. 7 and 9 D–F). Furthermore, citrullination impairs the interaction between p53 and MDM2, further interfering with its degradation (Figs. 8 and 9D). Together, these results indicate that citrullination disrupts DNA binding, gene regulation, protein stability, and MDM2 interaction, thereby undermining p53’s tumor suppressor function.

Fig. 9.

The functional implications of p53 citrullination facilitated by PAD4. This model illustrates that under specific conditions, a substantial increase in cellular Ca²⁺ concentration markedly activates PAD4. As a result, dimeric or monomeric forms of p53 undergo citrullination by PAD4. The citrullinated p53, whether dimeric or monomeric, is unable to form functional tetramers (A) or effectively bind to DNA (B), leading to impaired transcriptional activity (C). In addition, citrullinated p53 fails to interact with MDM2 (D), resulting in reduced ubiquitination (E) and protection from degradation (F).

p53 is a central transcription factor that regulates key cellular processes, and its function is tightly controlled by various posttranslational modifications (PTMs) (7). Here, we identify citrullination as a PTM that compromises p53 tetramerization, DNA binding, and transcriptional activity. While previous studies demonstrated that PAD4 influences p53 target gene expression through histone citrullination (34, 35), our findings further reveal that PAD4 also directly citrullinates p53, thereby modulating its function. Most citrullination sites lie within p53’s core DNA-binding domain, affecting its structural integrity and function. Moreover, citrullination disrupts p53–MDM2 interaction and downstream degradation. Although MDM2 binding and oligomerization sites are located at the p53 C-terminus (40, 41), citrullination-induced conformational changes in the DNA-binding domain likely interfere with this interaction. These results underscore the structural complexity of p53 regulation by citrullination and highlight the need for further structural studies.

Beyond p53 itself, citrullination can modulate p53 activity by altering interactions with regulatory proteins. For instance, citrullination of ING4 weakens its interaction with p53, reducing p53 acetylation and transcriptional activity (38). Citrullination also intersects with other PTMs, such as methylation and acetylation. For example, H3R26 citrullination antagonizes H3K27 methylation (42, 43). Our findings suggest that citrullination may similarly impact ubiquitination of p53. Under normal conditions, ubiquitination promotes p53 degradation to regulate stress responses and maintain homeostasis(44). During cellular stress, lysine acetylation blocks ubiquitination, stabilizing p53 and enhancing its transcriptional activity (45). This balance is crucial for p53’s tumor suppressor function. In contrast, citrullination impairs p53 activity, and its accumulation does not enhance tumor suppression.

Previous studies also showed that p53 activates PAD4 under cellular stress (36, 46), identifying PAD4 as a downstream effector. Our results now reveal that PAD4, in turn, regulates p53 stability and activity through citrullination, affecting both transcription and degradation. This suggests a possible negative feedback loop between p53 and PAD4, similar to the p53 and MDM2 axis (47), in which p53 upregulates PAD4 and PAD4 restricts p53 activation. Such feedback may fine-tune the stress response and maintain homeostasis. Notably, PAD4-mediated citrullination may also affect mutant p53, which often gains oncogenic functions that promote tumor progression and chemoresistance (48, 49). Citrullination may alter mutant p53’s stability, conformation, and cofactor interactions. Given PAD4’s frequent overexpression in cancer (50, 51), aberrant citrullination likely contributes to tumor progression and represents a potential therapeutic target. PAD4 is also active in neutrophils, where it promotes tumor development (52). In our study, PAD4 inhibitors reduced citrullinated p53 levels. Several PAD4-specific inhibitors show antitumor effects and may enhance immune checkpoint therapy (33, 53). Targeting PAD4 could restore p53 stability and function by blocking its citrullination.

In conclusion, PAD4-mediated citrullination disrupts p53’s DNA binding and tetramerization, impairs transcriptional regulation, and interferes with MDM2-mediated degradation. This identifies citrullination as a critical regulatory mechanism of p53 in cancer. Targeting PAD4 may offer a strategy to restore p53 function, warranting further investigation into its therapeutic potential.

Materials and Methods

Comprehensive details of all materials and experimental procedures employed in this study are provided in the SI Appendix, Materials and Methods. These encompass cell culture conditions, plasmid construction, protein expression and purification, PAD4 enzymatic assays, liquid chromatography–tandem mass spectrometry (LC–MS/MS), antibody generation, in vitro ubiquitination assays, analytical ultracentrifugation (AUC), electrophoretic mobility shift assays (EMSA), reporter gene assays, cycloheximide (CHX) chase assays, flow cytometric analysis of cell cycle progression and apoptosis, immunoprecipitation, pull-down and AlphaScreen assays, quantitative PCR (qPCR) primer sequences, and other pertinent methodologies.

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

All study data are included in the article and/or SI Appendix.