Nitrogen Mustard Alkylates and Cross-Links p53 in Human Keratinocytes

Abstract

Cytotoxic blistering agents such as sulfur mustard and nitrogen mustard (HN2) were synthesized for chemical warfare. Toxicity is due to reactive chloroethyl side chains that modify and damage cellular macromolecules including DNA and proteins. In response to DNA damage, cells initiate a DNA damage response directed at the recruitment and activation of repair-related proteins. A central mediator of the DNA damage response is p53, a key protein that plays a critical role in regulating DNA repair. We found that HN2 causes cytosolic and nuclear accumulation of p53 in HaCaT keratinocytes; HN2 also induced post-translational modifications on p53 including S15 phosphorylation and K382 acetylation, which enhance p53 stability, promote DNA repair, and mediate cellular metabolic responses to stress. HN2 also cross-linked p53, forming dimers and high-molecular-weight protein complexes in the cells. Cross-linked multimers were also modified by K48-linked ubiquitination indicating that they are targets for proteasome degradation. HN2-induced modifications transiently suppressed the transcriptional activity of p53. Using recombinant human p53, HN2 alkylation was found to be concentration- and redox status-dependent. Dithiothreitol-reduced protein was more efficiently cross-linked indicating that p53 cysteine residues play a key role in protein modification. LC-MS/MS analysis revealed that HN2 directly modified p53 at C124, C135, C141, C176, C182, C275, C277, H115, H178, K132, and K139, forming both monoadducts and cross-links. The formation of intermolecular complexes was a consequence of HN2 cross-linked cysteine residues between two molecules of p53. Together, these data demonstrate that p53 is a molecular target for mustard vesicants. Modification of p53 likely mediates cellular responses to HN2 including DNA repair and cell survival contributing to vesicant-induced cytotoxicity.

Graphical Abstract

Introduction

Nitrogen mustard [bis(2-chloroethyl) methylamine, HN2] and related structural analogues including sulfur mustard [bis(2-chloroethyl) sulfide, SM] were first developed as chemical warfare agents.¹ HN2 was later developed for cancer chemotherapy and is currently used to treat cutaneous T-cell lymphomas.² HN2 and SM are potent vesicants causing damage to the skin, eyes, and lungs.^{1, 3} Both mustards contain two reactive chloroethyl side chains which can alkylate cellular macromolecules. Nucleic acids and proteins are major targets for HN2 and SM; alkylation of DNA results in the formation of diverse DNA lesions, including mono- and cross-linked adducts.^{1, 3} This can result in single- and double-strand DNA breaks, processes that activate DNA damage response pathways.^{3, 4} Protein modifications by mustards can alter their structure and function.^{3, 5} Both DNA and protein damage can lead to cell cycle arrest and cytotoxicity.

Cell fate after DNA damage is predominantly controlled by p53, a transcription factor regulating genes involved in processes such as cell cycle arrest, DNA repair, apoptosis, metabolism, autophagy, translational control and signaling feedback mechanisms.^{6, 7} Under homeostatic conditions, p53 is sequestered in an inactive form by its association with MDM2, an E3 ubiquitin ligase that targets p53 for degradation.^7–9 Activation of p53 by intracellular stress signals, particularly those involved in the DNA damage response, is known to occur via post-translational mechanisms such as site-specific phosphorylation and acetylation.⁸ Earlier studies showed that p53 protein levels are rapidly elevated following DNA damage induced by mustard vesicants both in cell culture and in animal models.^10–12 This is associated with upregulation of DNA damage signaling molecules including phosphorylated ATM (S1981), ATR (S426), DNA-PKcs (S2056), CHK1 (S317), CHK2 (T68), H2AX (S319), and p53 (S15).^10–13 Phosphorylation of p53 at S15 increases p53 stability and binding to target genes, which in turn, triggers activation of pathways that regulate vesicant-induced cell cycle arrest and apoptosis.^12–14

Covalent modification of cysteine residues in the p53 DNA binding domain by alkylating agents has been reported to regulate p53 function. For example, Kim et al.¹⁵ showed that 15-deoxy-Δ^12,14-prostaglandin J₂, an electrophilic cyclopentenone prostaglandin, selectively modifies p53 at C277 in MCF cells. This leads to accumulation of p53 in both cytosolic and nuclear fractions of cells and a reduction in its transcriptional activity.¹⁵ N-Ethylmaleimide (NEM), a thiol-derivatizing agent, has been reported to alter p53 activity in a concentration-dependent manner.^{16, 17} Whereas low concentrations of NEM increase DNA binding, higher concentrations reduce DNA binding.^{16, 17} Several thiol-alkylating agents, including APR-246 [2-(hydroxymethyl)-2-(methoxymethyl)quinuclidin-3-one] and the 2-sulfonylpyrimidine molecule PK11000 [5-chloro-2-(methylsulfonyl)-4-pyrimidinecarboxylic acid], have also been shown to reactivate mutant p53 by modification of cysteine residues in the core DNA binding domain.^18–20 Previously, we reported that HN2 cross-links thioredoxin and thioredoxin reductase forming several intermolecular complexes through alkylation between catalytic cysteines and selenocysteine residues; this leads to inhibition of the thioredoxin system and consequent oxidative stress.^{21, 22} In the present studies, we show that HN2 causes accumulation and nuclear retention of p53 and post-translational modified p53 in human keratinocytes, as a consequence of DNA damage response signaling. HN2 also directly targets p53 protein, resulting in the formation of several higher-molecular-weight p53 cross-linked proteins. This occurs primarily via HN2-modified cysteine residues in p53, a process that reduces its transcriptional activity. Taken together, these observations provide new insights into mechanisms of vesicant-induced alterations in keratinocyte growth and differentiation, which may be useful in the identification of countermeasures to mitigate dermal injury.

Materials and Methods

Caution:

HN2 is a highly toxic vesicant, and precautions were taken for its handling and preparation including the use of double gloves, safety glasses, masks, and other protective equipment to prevent exposures. The disposal of HN2 waste followed Rutgers University Environmental Health and Safety guidelines.

Chemicals and Reagents.

Dulbecco’s modified Eagle’s medium (DMEM; containing 4500 mg/L D-glucose, 110 mg/mL sodium pyruvate, and 584 mg/L L-glutamine; catalog no. 11995–065), fetal bovine serum, and penicillin/streptomycin were purchased from ThermoFisher Scientific/Invitrogen (Grand Island, NY). Recombinant wild-type human p53 protein was obtained from Enzo Life Sciences (Farmingdale, NY) and a p53 transcription factor assay kit from Cayman (Ann Arbor, MI). SuperSignal chemiluminescense substrates (West Dura, West Pico PLUS), Pierce™ ECL Western blotting substrate, Pierce™ BCA protein assay kit, NE-PER™ nuclear and cytoplasmic extraction reagents, and Halt™ protease and phosphatase inhibitor cocktails were obtained from Thermo Scientific (Rockford, IL). HN2 (mechlorethamine hydrochloride, catalog no. 122564), chlorambucil, melphalan, cisplatin, and all other chemicals were obtained from Sigma-Aldrich (St. Louis, MO) unless otherwise indicated.

Cell Culture and Treatments.

Immortalized adult human keratinocytes (HaCaT), CX-1, HEK293, and A431 cells were grown in DMEM supplemented with 10% fetal bovine serum, 100 units/mL penicillin, and 100 μg/mL streptomycin in a humidified atmosphere of 5% CO₂ at 37 °C. As components in serum including albumin are targets for covalent modifications by mustard vesicants, solutions of HN2 were freshly prepared in serum-free DMEM immediately before use. For the preparation of cell lysates, HaCaT cells (5 × 10⁶ cells) were incubated overnight in 15 cm culture dishes and then treated with HN2 (10–200 μM) or vehicle control. After 30 min to 6 h, cells were washed twice with ice-cold PBS (137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, 2 mM KH₂PO₄, pH 7.4), and removed from the plates with a cell scraper in 5 mL of PBS, and cell pellets were collected after centrifugation (800 g, 5 min). Lysates were prepared by sonication of the cells on ice with three 10 s pulses in five cell volumes of lysis buffer (PBS containing 0.1% Triton X-100 and 1% phosphatase and protease inhibitor cocktails) followed by centrifugation at 800 g for 5 min to remove cellular debris. Cytosolic and nuclear fractions were prepared from HaCaT cells using NE-PER reagents (Thermo Scientific) according to the manufacturer’s protocols. Protein concentrations were quantified with a BCA protein assay kit (Thermo Scientific) with bovine serum albumin as a standard.

Immunoprecipitation, Western Blotting and Slot Blotting.

For Western blotting, protein samples (50 μg) were subjected to reducing and denaturing SDS-PAGE (4–15% Criterion™ Tris-HCl gel, Bio-Rad, Hercules, CA) followed by electroblotting onto nitrocellulose membranes. For -sldot blotting, reduced and denatured proteins (15 μg) were applied to nitrocellulose membranes using a Bio-Dot SF microfiltration apparatus (Bio-Rad) according to the manufacturer’s instructions. Membranes were blocked in 5% nonfat dry milk in PBST (PBS containing 0.1% Tween 20) for 30 min at 37 °C. The blots were incubated with primary antibody overnight at 4°C followed by HRP-conjugated secondary antibody (Bio-Rad) for 30 min at 37 °C. Specific proteins were visualized using chemiluminescence, and the intensity of protein bands were quantified using NIH ImageJ software (https://imagej.nih.gov/ij/index.html). Antibodies used for this study were p53 (Santa Cruz sc-126, Dallas, TX or Cell Signaling #9282, Danvers, MA), phospho p53 (S15) (Cell Signaling #9284), acetyl p53 (K382) (Cell Signaling #2525), ubiquitin (Cell Signaling #3936), K48-linked polyubiquitin (Cell Signaling #12805), K63-linked polyubiquitin (Cell Signaling #12930), and β-actin (Santa Cruz sc-47778). For immunoprecipitation, whole cell lysates (500 μg protein) were incubated with a monoclonal antibody against p53 conjugated to agarose beads (50 μL, Santa Cruz sc-126 AC) overnight at 4 °C with gentle rotation. Immune-complexes were collected by centrifugation (3,000 g for 5 min) followed by 4 washes with ice-cold PBS. The immunoprecipitates were subjected to reducing and denaturing SDS-PAGE followed by Western blotting.

Transcriptional Activity Assay for p53.

Binding of p53 to a specific double-stranded DNA sequence containing the p53 response element was assayed with a p53 transcription factor assay kit (Cayman) according to the manufacturer’s protocols. Nuclear extracts (10 μg) were incubated overnight at 4 °C with immobilized oligonucleotides that contain the p53 response element. Bound protein was detected with a specific primary antibody against human p53 followed by incubation with an HRP-conjugated secondary antibody. The HRP signal was developed by a substrate provided by the manufacturer and absorbance was measured at 450 nm with a reference wavelength of 650 nm using a SpectraMax M3 microplate reader (Molecular Devices, San Jose, CA).

LC-MS/MS Analysis.

Methods for protein analysis by LC-MS/MS were previously described.^{22, 23} Briefly, recombinant human p53 (10 μg) was reduced with 10 mM dithiothreitol (DTT) at 37 °C for 20 min in a final volume of 50 μL in 50 mM potassium phosphate buffer (pH 7.4) and then incubated with HN2 (200 μM) or vehicle control at room temperature. After 1 h, reaction mixtures were separated by reducing and denaturing SDS−PAGE using a 4–15% Tris-HCl gel (Bio-Rad). After staining with Coomassie blue, bands containing p53 were cut from the gels. In-gel digestion with trypsin (Roche, Indianapolis, IN) and peptide extraction and reconstitution were performed as previously described.²² Peptide samples were analyzed by nano LC-MS/MS using a Dionex U3000 RSLC nanosystem (Dionex, Sunnyvale, CA) online with a Q Exactive HF mass spectrometer (Thermo Fisher Scientific, San Jose, CA), performed at the Biological Mass Spectrometry Facility at Rutgers University, as described previously.²³ Cross-linked peptides were identified using pLink2 (version 2.3.8, Chinese Academy of Sciences, Beijing, China) ²⁴ and searched against the UniProt human protein database with 20 ppm for MS and MS/MS tolerances. HN2-induced modification at cysteine, histidine, and/or lysine (+83.0735 and 101.0841 for HN2 cross-link and monoadduct, respectively), oxidation at methionine or tryptophan (+ 15.9950 Da), and carbamidomethylation at cysteine (+ 57.0214 Da) were set as dynamic modifications and included a maximum of 2 missed cleavages to identify spectra of adducted peptides. Results were annotated with b and y ions of peptide sequences and visualized by pLabel (Chinese Academy of Sciences); all cross-linked peptides and the location of cross-linking sites were manually validated.

Molecular Modeling.

The SAXS structure of full-length human p53²⁵ was kindly provided by Henning Tidow (University of Hamburg, Germany). Distributions of HN2 cross-links were visualized in the SAXS model of p53 using PyMOL (Schrödinger, New York, NY). Distances between nucleophilic side-chains of cross-linked residues were measured and compared to the maximum length of the HN2 molecule, which allowed for further confirmation of identified cross-links.

Data Analysis.

All data presented are representative of at least three independent experiments. A statistical analysis was performed using the Student’s t test, and a P value of p < 0.05 was considered statistically significant. Graphs were made using GraphPad Prism 5.0 software (La Jolla, CA), and error bars represented SE (n = 3–4).

Results

Cross-Linking p53 in HaCaT Cells Treated with HN2

In initial experiments, we examined the effects of HN2 on p53 expression in HaCaT cells. Slot blotting revealed that there were no major alterations in total p53 protein 0–6 h after HN2 treatment (Figure 1A). In contrast, HN2 caused a concentration- and time-dependent increase in phosphorylation of p53 at S15 (Figure 1A). At 50 μM HN2, a significant increase in phospho p53 S15 was evident within 1 h; maximal increases were evident at 2 h with 50–200 μM (Figure 1A and not shown). HN2 also cross-linked p53 as evidenced by the formation of p53 dimers, trimers, and higher-molecular-weight oligomers, with molecular masses of around 100, 150, and >200 kDa, respectively, in SDS polyacrylamide gels (Figure 1B, 1C). Dimers of p53 appeared after 30 min with 50 μM HN2 (Fig. 1B); oligomers were evident at concentrations of HN2 greater than 50 μM (Figure 1C). HN2 also induced p53 S15 phosphorylation of monomers, dimers, trimers, and oligomers. These data indicate that HN2 induces a p53 DNA damage response in HaCaT cells.

Figure 1. HN2 cross-links p53 proteins in HaCaT cells.

Cells were treated with increasing concentrations of HN2 or vehicle control in serum-free medium. At the indicated times, cells were harvested, and total cellular lysates were prepared; protein expression was analyzed by (A) slot blots or (B, C) by SDS-polyacrylamide gel electrophoresis in reducing and denaturing gels followed by Western blotting. β-Actin was used as an example of a protein loading control. Representative Western blots and slot blots from one of three separate experiments are shown. Data are the mean ± SE (n = 3). *Significantly different (p < 0.05) from vehicle-treated controls. (B) Time-course of changes in p53 proteins in response to 50 μM HN2. (C) Effects of increasing concentrations of HN2 on the expression of p53 proteins. Protein expression was analyzed 2 h after HN2 treatment.

We next compared HN2 with other bifunctional alkylating agents used in cancer chemotherapy including chlorambucil, melphalan, and cisplatin. Each of these drugs was found to cross-link p53, although not as effectively as HN2 (Figure 2A). HaCaT cells are known to overexpress p53 mutant protein (containing H179Y and R282W mutations).^{26, 27} Cross-linking of p53 in cells containing wild-type p53 was next compared to cells with different p53 mutants. HN2-induced p53 cross-linking was evident in human embryonic kidney 293 (HEK-293) cells, which express wild-type p53, as well as cells expressing mutant p53 protein including CX-1, a cell line derived from a colorectal adenocarcinoma containing an R273H mutation, and A431, a cell line derived from an epidermoid carcinoma containing an R273H mutation (Figure 2B). Cross-linking of p53 was also observed in MLE-15 cells, an SV40 transformed murine lung epithelial cell line expressing wild-type p53.

Figure 2. Cellular cross-linking of p53 by bifunctional alkylating agents.

(A) HaCaT cells were incubated with 200 μM chlorambucil (CHL), melphalan (MEL), or cisplatin (CDDP) in serum-free medium. At the indicated times (0–6 h), cells were harvested, and cell lysates were prepared; p53 protein expression was analyzed by Western blotting. β-Actin was used as a protein loading control. Little to no cytotoxicity was observed after a 6 h exposure of cells to these agents (200 μM) as measured by the PrestoBlue assay. (B) Effects of HN2 on p53 cross-linking in cells with wild-type p53 and mutant p53. Cells were treated without and with HN2 (200 μM). After 2 h, lysates were prepared and analyzed for p53 cross-linking by Western blotting. Cell lines used include HaCaT (mutant p53 with H179Y and R282W mutations), HEK-293 (wild-type p53), CX-1 and A431 (mutant p53 with R273H mutation), and MLE-15 (wild-type p53). Data show representative Western blots from one of three separate experiments.

Nuclear Accumulation of p53 after HN2 Exposure

In further studies, we examined the effects of HN2 on expression and localization of p53 in HaCaT cells. Both cytoplasmic and nuclear fractions of the cells expressed p53 (Figure 3). HN2 caused a concentration-dependent increase in total and phosphorylated p53 in both of these fractions. In the cytoplasm, maximal increases in p53 were evident at 20–50 μM concentrations of HN2 after 2 h. Increases in total cytoplasmic p53 occurred at lower concentrations of HN2 (20–50 μM) when compared to phospho p53 (100–200 μM). In contrast, in the nucleus, total and phospho p53 increased with increasing concentrations of HN2 (10–200 μM). HN2 also cross-linked p53 in both cytoplasmic and nuclear fractions of the cells; greater amounts of cross-linked dimers and oligomers were evident in nuclear fractions. Protein monomers, as well as cross-linked dimers and oligomers, were phosphorylated following treatment with HN2 in both nuclear and cytoplasmic fractions of the cells.

Figure 3. Effects of HN2 on cytoplasmic and nuclear localization of p53 in HaCaT cells.

Cells were treated with increasing concentrations of HN2 (10–200 μM) or vehicle control in serum-free medium. After 2 h, cytosolic and nuclear fractions were prepared, and p53 and phospho p53 expression was analyzed by Western blotting and slot blotting. (A) HN2 cross-links p53 proteins in both cytosolic and nuclear fractions of HaCaT cells. Representative Western blots from three separate experiments are shown. β-Actin was used as a protein loading control. (B) Relative phospho p53 (S15) and p53 protein expression in HaCaT cells after HN2 treatment analyzed by slot blotting. Representative slot blots from three separate experiments are shown. Data are the mean ± SE (n = 3). *Significantly different (p < 0.05) from vehicle-treated controls.

Biphasic Response of p53 Activities in HaCaT Cells Treated with HN2

In further studies, we determined if HN2-induced modifications of p53 resulted in altered functional activity. Treatment of HaCaT cells with lower concentrations of HN2 (≤ 10 μM) for 2 h caused a concentration-dependent decrease in the DNA binding activity of p53, a response which persisted for at least 6 h (Figure 4A and not shown). The activity of p53 recovered at higher concentrations of HN2 (20–200 μM).

Figure 4. Effects of HN2 on transcriptional activity and post-translational modification of p53 in HaCaT cells.

Cells were treated with increasing concentrations of HN2 (10–200 μM) or vehicle control. After 2 h, whole cell lysates or nuclear fractions of cells were prepared. (A) Nuclear fractions were assayed for p53 transcriptional activity as described in the Materials and Methods section. Data are the mean ± SE (n = 3). *Significantly different (p < 0.05) from vehicle-treated controls. Cell lysates were analyzed by Western blotting for acetylation of p53 at K382 (B) and ubiquitination as indicated (C). β-Actin was used as an example of a loading control. (D) Ubiquitination assay of p53. Whole cell lysates were subjected to immunoprecipitation with a monoclonal antibody against p53 conjugated to agarose beads. K48-linked polyubiquitinated p53 and p53 in the immunoprecipitates were determined by Western blotting.

Earlier studies demonstrated that, in response to genotoxic stress, p53 activity is regulated through multiple processes including post-translational modifications.^{28, 29} We next analyzed the effects of HN2 on p53 acetylation and protein ubiquitination in HaCaT cells. HN2 was found to induce p53 acetylation at K382 and to trigger protein ubiquitination in a concentration-dependent manner (Figure 4B,4C); increases in ubiquitination of higher-molecular-weight proteins were more prevalent than those in lower-molecular-weight proteins (Figure 4C). Protein ubiquitination induced by HN2 was primarily mediated via K48-linked ubiquitination but not by K63-linkage (Figure 4C and not shown). Immunoprecipitation studies showed that K48-linked ubiquitin of p53 was more abundant in p53 dimers and oligomers than in p53 monomers (Figure 4D). These results suggest that intermolecular HN2 cross-links are targets for proteasome degradation.

Mechanism of HN2 Induced p53 Cross-Linking

Using human recombinant wild type p53 protein, mechanisms of HN2-induced p53 cross-linking were analyzed. Figure 5 shows that HN2 cross-links recombinant p53 in a concentration- and reduction status-dependent manner. DTT-reduced p53 protein exhibited a higher degree of cross-linking when compared to nonreduced p53, suggesting that cysteine residues play a key role in p53 cross-linking by HN2.

Figure 5. Effects of HN2 on recombinant human p53.

Recombinant wild-type human p53 (100 nM) was incubated without and with DTT (10 mM) at 37 °C. After 20 min, proteins were purified using Chroma Spin TE-10 columns to remove DTT and then incubated with increasing concentrations of HN2 (2–200 μM) or vehicle control in potassium phosphate buffer (100 mM, pH 7.4) at room temperature. After an additional 60 min, samples were analyzed by SDS-polyacrylamide gel electrophoresis under reducing and denaturing conditions followed by Western blotting.

We next used LC-MS/MS to identify modified residues in p53 monomers, dimers, trimers, and tetramers from HN2-treated p53 as well as monomers from untreated p53 control following SDS-polyacrylamide gel purification and in-gel trypsin digestion. Both HN2 monoalkylated (Table 1) and cross-linked (Tables 2–5) peptides were identified which were covalently bound to cysteine, histidine, and lysine residues on the p53 protein. Identified cross-links were further analyzed by their appearance in each protein band and validated by comparing the distance of two reactive sites on the p53 protein structure and the extended length of the HN2 molecule. For these studies, we used the p53 SAXS model that grouped these modifications into four classes: intrapeptide loop-links (Table 2), intramolecular cross-links (Table 3), intermolecular cross-links (Table 4), and mixed intramolecular and intermolecular cross-links (Table 5). Following HN2 treatment, both intrapeptide loop-links and intramolecular cross-links were predominantly detected in monomeric p53. These adducts were found to selectively cross-link amino acid residues in the protein that were in close proximity with a distance close to or less than the length of the HN2 molecule. In contrast, intermolecular and mixed-molecular cross-links were most abundant in the dimeric and multimeric p53 proteins; cross-linked residues were largely located in solvent accessible surface residues in the protein. Fragmentation spectra of HN2-alkylated peptides and their sequence assignments are shown in Figures 6 and 7, and in the Supporting Information in Figures S1–S5. Collectively, 7 cysteine residues on p53, including C124, C135, C141, C176, C182, C275, and C277, were primary targets for HN2 alkylation. HN2 also modified histidine and lysine residues, including H115, H178, K132, and K139, on p53 forming cross-links.

TABLE 1.

HN2 mono-links identified by LC-MS/MS

peptide sequence (amino acid position)	m/z	charge	tR (min)	[M+H]+ observed	[M+H]+ calculated	modification
LGFLHSGTAKSVTCTYSPALNK (111–132)	799.7603	3	25.88	2397.2660	2397.2720	C124 HN2 mono-link
	599.8214	4		2397.2663
SVTCTYSPALNK (121–132)	692.8604	2	19.45	1384.7136	1384.7140	C124 HN2 mono-link
	462.2427	3		1384.7134
MFCQLAK (133–139)	479.8797	2	18.26	957.4898	957.4896	M133 oxidation, C135 HN2 mono-link
	319.8349	3		957.4899
MFCQLAKTCPVQLWVDSTPPPGTR (133–156)	950.0862	3	39.67	2850.4045	2850.4147	M133 oxidation, C135 HN2 mono-link, C141 carbamidomethyl
MFCQLAKTCPVQLWVDSTPPPGTR (133–156)	723.8643	4	40.93	2891.4378	2891.4489	K139 HN2 mono-link, C135 carbamidomethyl, C141 carbamidomethyl
TCPVQLWVDSTPPPGTR (140–156)	978.0043	2	33.49	1956.0085	1956.0132	C141 HN2 mono-link
	652.3401	3		1956.0080
CPHHERCSDSDGLAPPQHLIR (176–196)	632.5577	4	20.40	2526.2071	2526.2088	C176 HN2 mono-link, C182 carbamidomethyl or C176 carbamidomethyl, C182 HN2 mono-link
	506.6755	5		2526.2077
	422.8295	6		2526.2075
VCACPGRDR (274–282)	567.7789	2	14.40	1135.5532	1135.5584	C275 carbamidomethyl, C277 HN2 mono-link
	378.8547	3		1135.5544

TABLE 2.

HN2 intrapeptide loop-links identified by LC-MS/MS

peptide sequence (amino acid position)	m/z	charge	tR (min)	[M+H]+ observed	[M+H]+ calculated	modification
SVTCTYSPALNKMFCQLAK (121–139)	736.3729	3	24.14	2206.1016	2206.1068	M133 oxidation, C124-C135 HN2 loop-link
	552.5330	4		2206.1065
	442.2273	5		2206.1042
SVTCTYSPALNKMFCQLAK (121–139)	754.0356	3	43.76	2261.0968	2261.1126	C124-K132 HN2 loop-link, M133 oxidation, C135 carbamidomethyl
MFCQLAKTCPVQLWVDSTPPPGTR (133–156)	926.8063	3	35.01	2776.4031	2776.3982	M133 oxidation, C135-C141 HN2 loop-link
	695.1061	4		2776.3980
	556.4861	5		2776.4024
MFCQLAKTCPVQLWVDSTPPPGTR (133–156)	944.4675	3	40.88	2833.4261	2833.4200	M133 oxidation, C135-K139 HN2 loop-link, C141 carbamidomethyl
RCPHHER (175–181)	339.8437	3	12.92	1017.5166	1017.5159	C176-H178 HN2 loop-link
RCPHHERCSDSDGLAPPQHLIR (175–196)	435.3856	6	23.54	2607.2788	2607.2778	C176-C182 HN2 loop-link
CPHHERCSDSDGLAPPQHLIR (176–196)	818.7325	3	23.45	2453.1842	2453.1924	C176-C182 HN2 loop-link
	614.3021	4		2453.1836
	491.6434	5		2453.1845
	410.3703	6		2453.1807
VCACPGRDR (274–282)	529.7868	2	37.29	1059.5257	1059.5186	C275-C277 HN2 loop-link
	353.8444	3		1059.5185

TABLE 5.

Possible intra-molecular and inter-molecular HN2 cross-links identified by LC-MS/MS

peptide sequence (amino acid position)	m/z	charge	tR(min)	[M+H]+ observed	[M+H]+ calculated	modification
α peptide: SVTCTYSPALNK (121–132)	715.3506	3	16.92	2144.0318	2144.0375	αC124-βC176 HN2 cross-link
β peptide: CPHHER (176–181)	537.0154	4		2144.0379
	429.8129	5		2144.0370
α peptide: SVTCTYSPALNK (121–132)	532.7640	4	18.74	2128.0347	2128.0347	αC124-βC275 HN2 cross-link, βC277 carbamidomethyl or αC124-βC277 HN2 cross-link, βC275 carbamidomethyl
β peptide: VCACPGR (274–280)	426.4125	5		2128.0328
α peptide: TCPVQLWVDSTPPPGTR (140–156)	1074.2120	3	32.82	3219.6170	3219.6175	αC141-βC124 HN2 cross-link
β peptide: SVTCTYSPALNK (121–132)	805.9109	4		3219.6178
	644.9294	5		3219.6186
α peptide: TCPVQLWVDSTPPPGTR (140–156)	809.6602	4	32.80	3235.6139	3235.6124	αC141-βC124 HN2 cross-link, αW146 oxidation
β peptide: SVTCTYSPALNK (121–132)	648.1285	5		3235.6126
α peptide: TCPVQLWVDSTPPPGTR (140–156)	699.3549	4	33.27	2792.3922	2792.3931	αC141-βC135 HN2 cross-link, βM133 oxidation
β peptide: MFCQLAK (133–139)	559.3165	5		2792.3939
α peptide: TCPVQLWVDSTPPPGTR (140–156)	887.4486	4	36.02	3544.7687	3544.7674	αC141-βC182 HN2 cross-link
β peptide: CSDSDGLAPPQHLIR (182–196)	710.3595	5		3544.7672
α peptide: TCPVQLWVDSTPPPGTR (140–156)	900.7817	3	28.35	2698.3257	2698.3261	αC141-βC275 HN2 cross-link, βC277 carbamidomethyl or αC141-βC277 HN2 cross-link, βC275 carbamidomethyl
β peptide: VCACPGR (274–280)	675.5875	4		2698.3262
	540.7811	5		2698.3264
α peptide: CSDSDGLAPPQHLIR (182–196)	993.1661	3	27.57	2974.4791	2974.4760	αC182-βC124 HN2 cross-link
β peptide: SVTCTYSPALNK (121–132)	744.6254	4		2974.4763
	595.9469	5		2974.4748
	496.5865	6		2974.4784
α peptide: CSDSDGLAPPQHLIR (182–196)	633.8214	4	29.07	2531.2558	2531.2566	αC182-βC135 HN2 cross-link
β peptide: MFCQLAK (133–139)	507.2580	5		2531.2567
α peptide: CSDSDGLAPPQHLIR (182–196)	818.7325	3	23.39	2453.1842	2453.1845	αC182-βC275 HN2 cross-link, αC277 carbamidomethyl
β peptide: VCACPGR (274–280)	614.3016	4		2453.1838
	491.6435	5		2453.1827
	410.3703	6		2453.1807
α peptide: CSDSDGLAPPQHLIR (182–196)	818.7349	3	23.08	2453.1864	2453.1845	αC182-βC277 HN2 cross-link, αC275 carbamidomethyl
β peptide: VCACPGR (274–280)	614.5527	4		2453.1848
	491.6431	5		2453.1848
α peptide: VCACPGR (274–280)	541.5868	3	14.62	1622.7463	1622.7460	αC277-βC176 HN2 cross-link, αC275 carbamidomethyl
β peptide: CPHHER (176–181)	406.4420	4		1622.7464
	325.5556	5		1622.7461

TABLE 3.

Intra-molecular HN2 cross-links identified by LC-MS/MS

peptide sequence (amino acid position)	m/z	charge	tR(min)	[M+H]+ observed	[M+H]+ calculated	modification
α peptide: SVTCTYSPALNK (121–132)	800.0592	3	25.68	2396.2624	2396.2641	αC124-βH115 HN2 cross-link
β peptide: LGFLHSGTAK (111–120)	600.7850	4		2396.2638
α peptide: SVTCTYSPALNK (121–132)	736.7086	3	24.17	2206.1027	2206.1068	αC124-βC135 HN2 cross-link
β peptide: MFCQLAK (133–139)	552.5326	4		2206.1064
	442.4279	5		2206.1061
α peptide: SVTCTYSPALNK (121–132)	742.3738	3	22.02	2222.1010	2222.1017	αC124-βC135 HN2 cross-link, βM133 oxidation
β peptide: MFCQLAK (133–139)	556.5313	4		2222.1011
	445.1200	5		2222.1018
α peptide: TCPVQLWVDSTPPPGTR (140–156)	695.1093	4	34.46	2776.3987	2776.3982	αC141-βC135 HN2 cross-link
β peptide: MFCQLAK (133–139)	556.4859	5		2776.3975
α peptide: CSDSDGLAPPQHLIR (182–196)	876.4341	3	19.36	2625.2879	2625.2884	αC182-βC176 HN2 cross-link
β peptide: RCPHHER (175–181)	526.0649	5		2625.2881
	438.3872	6		2625.2879
α peptide: CSDSDGLAPPQHLIR (182–196)	824.0706	3	19.83	2469.1870	2469.1873	αC182-βC176 HN2 cross-link
β peptide: CPHHER (176–181)	618.3029	4		2469.1874
	494.8440	5		2469.1884
	412.5096	6		2469.1864

TABLE 4.

Inter-molecular HN2 cross-links identified by LC-MS/MS

peptide sequence (amino acid position)	m/z	charge	tR (min)	[M+H]+ observed	[M+H]+ calculated	modification
α peptide: SVTCTYSPALNK (121–132)	884.1159	3	23.12	2649.3275	2649.3261	αC124-βC124 HN2 cross-link
β peptide: SVTCTYSPALNK (121–132)	663.3376	4		2649.3265
	530.8719	5		2649.3271
α peptide: MFCQLAK (133–139)	425.9598	4	18.03	1700.8173	1700.8181	αC135-βC176 HN2 cross-link
β peptide: CPHHER (176–181)	341.5016	5		1700.8169
α peptide: MFCQLAK (133–139)	429.9569	4	17.94	1716.8059	1716.1759	αC135-βC176 HN2 cross-link, αM133 oxidation
β peptide: CPHHER (176–181)	343.8763	5		1716.8130
α peptide: MFCQLAK (133–139)	567.6086	3	18.23	1700.8113	1700.8102	αM133 oxidation, αC135-βC275 HN2 cross-link, βC277 carbamidomethyl or αM133 oxidation, αC135-βC277 HN2 cross-link, βC275 carbamidomethyl
β peptide: VCACPGR (274–280)	425.9577	4		1700.8106
	340.9681	5		1700.8101
α peptide: TCPVQLWVDSTPPPGTR (140–156)	948.7345	4	41.02	3789.9090	3789.9089	αC141-βC141 HN2 cross-link
β peptide: TCPVQLWVDSTPPPGTR (140–156)	758.9879	5		3789.9091
α peptide: TCPVQLWVDSTPPPGTR (140–156)	952.9829	4	41.03	3805.9038	3805.9039	αC141-βC141 HN2 cross-link, αW146 oxidation
β peptide: TCPVQLWVDSTPPPGTR (140–156)	762.1875	5		3805.9031
α peptide: TCPVQLWVDSTPPPGTR (140–156)	680.0890	4	25.33	2714.3283	2714.3289	αC141-βC176 HN2 cross-link
β peptide: CPHHER (176–181)	543.8723	5		2714.3283
α peptide: TCPVQLWVDSTPPPGTR (140–156)	891.1974	4	35.97	3560.7567	3560.7623	αC141-βC182 HN2 cross-link, αW146 oxidation
β peptide: CSDSDGLAPPQHLIR (182–196)	713.1581	5		3560.7603
α peptide: CSDSDGLAPPQHLIR (182–196)	637.8192	4	26.98	2547.2512	2547.2515	αC182-βC135 HN2 cross-link, βM133 oxidation
β peptide: MFCQLAK (133–139)	510.4568	5		2547.2526
α peptide: CSDSDGLAPPQHLIR (182–196)	825.9133	4	31.41	3299.6278	3299.6258	αC182-βC182 HN2 cross-link
β peptide: CSDSDGLAPPQHLIR (182–196)	660.9315	5		3299.6266
	550.2440	6		3299.6234
α peptide: VCACPGR (274–280)	541.5871	3	15.76	1622.7467	1622.7460	αC275-βC176 HN2 cross-link, αC277 carbamidomethyl
β peptide: CPHHER (176–181)	406.4420	4		1622.7461
α peptide: VCACPGR (274–280)	536.2525	3	15.90	1606.7437	1606.7432	αC275-βC277 HN2 cross-link, αC277 carbamidomethyl, βC275 carbamidomethyl or αC277-βC275 HN2 cross-link, αC275 carbamidomethyl, βC277 carbamidomethyl
β peptide: VCACPGR (274–280)	402.4417	4		1606.7433
	322.1545	5		1606.7428

Figure 6. Mass spectrum of HN2 cross-linked peptides at m/z 494.8439.

(A) Mass spectrum of the quintuply charged cross-linked peptides with monoisotopic m/z 494.8439. (B) Extracted-ion chromatogram of the parent ion in different p53 protein bands. (C) Higher-energy collisional dissociation (HCD) fragmentation of the [M + 5H]⁵⁺ precursor ion at m/z 494.8439. The sequences of cross-linked peptides were identified as CSDSDGLAPPQHLIR (residues 182–196 on p53; α peptide) and CPHHER (residues 176–181 on p53; β peptide). Identified fragments are indicated with α or β to denote the peptide from which they originated, together with their charge states. This cross-link involves residues C182 and C176 from α and β peptide, respectively. The inset shows the cross-linked fragments with the identified b and y ions labeled. (D) Stereo view of the structure of full-length wild-type human p53 tetramer. The molecular model was based on the SAXS model of human p53 in the absence of DNA. The subdomains of the DNA binding domain (DBD) are shown in cyan, green, magenta, and blue, respectively, with ribbon representation. Oligomerization domains (ODs) are shown in red. (E) Close-up of the interface of two p53 DBD subunits. Distances (Å) between thiol groups of C176 and C182 are labeled.

Figure 7. Mass spectrum of HN2 cross-linked peptides at m/z 532.7640.

(A) Mass spectrum of the quadruply charged cross-linked peptides with monoisotopic m/z 532.7640. (B) Extracted-ion chromatogram of the precursor ion at m/z 532.7640 in different p53 protein bands. (C, D) HCD fragmentation of the [M + 4H]⁴⁺ m/z 532.7640 precursor ion. The sequences of cross-linked peptides were identified as SVTCTYSPALNK (residues 121–132 on p53; α peptide) and VCACPGR (residues 274–280 on p53; β peptide) with one HN2 cross-link and one carbamidomethylation (cam) on cysteine. Identified fragments are indicated with α or β to denote the peptide from which they originated, together with their charge states. This cross-link involves residues C124 and C275 (C) or C124 and C277 (D) from the α and β peptide, respectively. The inset shows the cross-linked fragments with the identified b and y ions labeled. (E) Cross-linking sites on the SAXS model of human p53. Cross-linked and/or carbamidomethylated residues C124, C275, and C277 (colored in red) are located at the surface of the p53 structure. The distances (Å) between thiol groups of cysteine residues are labeled.

Figure 6 shows an example of HN2 cross-linking two peptides on p53 forming an intramolecular cross-link. The isotopic spectrum revealed that it was a quintuply charged ion, as indicated by the mass difference of 0.2 Da between neighboring peaks, with a monoisotopic mass of 494.8439 Da ([M + H]⁺ 2469.1884 Da) (Figure 6A). The mass of this precursor ion matched the mass of peptides CPHHER (residues 176–181 on p53; monoisotopic [M] 777.3341 Da) and CSDSDGLAPPQHLIR (residues 182–196 on p53; monoisotopic [M] 1607.7726 Da) plus one HN2 molecule (theoretical mass of protonated cross-link: [M + H]⁺ 2469.1873 Da). The selective ion chromatogram showed that it appeared with a retention time at 20.01 ± 0.11 min (mean ± ER, n = 4) and was most abundant in monomeric p53 from HN2-treated samples (Figure 6B). It was also detected in the dimeric, trimeric, and tetrameric p53 bands from HN2-treated protein but not in the monomeric p53 controls.

From the MS/MS fragmentation spectrum, the detection of a series of unmodified αy₁-αy₁₂ and βy₁-βy₅ ions confirmed a cross-link with the peptide sequence of CSDSDGLAPPQHLIR (α peptide) and CPHHER (β peptide), in which HN2 cross-linked residue C182 on the α peptide and C176 on the β peptide (Figure 6C). This is supported by a series of modified αb₂-αb₈ doubly charged ions with mass increases corresponding to the addition of one molecule HN2 and the CPHHER peptide. The distances of two thiol groups between C176 and C182 were found to be approximately 7.5 Å and 11.7 Å for two residues on the same monomer or two residues on different monomers, respectively (Figure 6E). The distance of C176 and C182 on the same molecule, which is close to the extended length of HN2 (~7.5 Å),³⁰ is highly suggestive of an intramolecular cross-link.

Another example of mixed intra- and intermolecular HN2 cross-links is shown in Figure 7. It is a quadruply charged ion, as evidenced by a mass difference of 0.25 Da between neighboring peaks, with monoisotopic m/z 532.7640 (Figure 7A). This precursor ion was detected predominantly in HN2-modified dimers, followed by trimers, tetramers, and monomers; it was not detected in the p53 control monomers (Figure 7B). An MS/MS analysis of m/z 532.7640 identified two distinct fragmentation spectra for this cross-linked ion.

Figure 7C shows HN2 cross-linked C124 on α peptide SVTCTYSPALNK (residues 121–132 on p53) and C275 on β peptide VCACPGR (residues 274–280 on p53) with one carbamidomethyl modification on C277 of β peptide. The sequence assignment of this modified peptide was based on the following observations: (1) unmodified αb₂, αb_3, and αy₁-αy₈ fragment ions on SVTCTYSPALNK peptide and unmodified βy₁-βy₃ on VCACPGR peptide; (2) βy₄ and βy₅ displaying a mass shift of 57.0214 Da corresponding to the addition of one carbamidomethyl group on C277 of β peptide; and (3) αb₅ and αb₆ displaying a mass shift corresponding to the addition of one HN2 and β peptide and βb₃ displaying a mass shift corresponding to the addition of one HN2 and α peptide. This fragment spectrum was detected in HN2-modified monomers, dimers, trimers, and tetramers. The detection in monomeric p53 suggests an intramolecular linking of two cysteines. The SAXS model of human p53 shows that the distance between two thiol groups of cross-linked cysteines is 10.1 Å (Figure 7E), which is greater than the extended length of HN2, indicating a flexible p53 structure in this region in an aqueous environment.

Figure 7D shows a different sequence assignment for the precursor ion at m/z 532.7640; HN2 cross-linking occurred between C124 on α peptide SVTCTYSPALNK and C277 on β peptide VCACPGR. This fragment spectrum was detected in HN2-modified dimers, trimers, and tetramers but not in HN2-treated monomers, suggesting intermolecular cross-linking. This is supported by the fact that both C124 and C277 are located on the surface of the p53 structure and the distance of sulfur atoms between two cysteines is approximately 14.5 Å (Figure 7E), which is significantly greater than the extended length of HN2.

Discussion

Functional human p53 is a tetramer composed of four identical 393 amino acid polypeptides.³¹ Each subunit consists of several functional domains including an N-terminal transactivation domain, a DNA binding core domain, a tetramerization domain, and a C-terminal domain (see model in Figure 6D). The regulation of p53 in response to DNA damage is tightly controlled by site-specific post-translational modifications, interactions with cofactors, and/or modulation of DNA or chromatin binding.^{29, 31} In the present studies, we identified p53 as a molecular target for HN2 in keratinocytes. HN2 treatment caused a time- and concentration-dependent increase in expression of p53 in cytoplasmic and nuclear fractions of the cells. This was associated with site-specific post-translational modifications of p53 including phosphorylation at S15 and acetylation at K382, as well as K48-linked ubiquitination. Phosphorylation of p53 at S15 is mediated by protein kinases including ATM/ATR/DNA PK and CHK1/CHK2. This causes conformational changes in the p53 structure,⁸ which can reduce affinity for MDM2, an important negative regulator of p53.^{7, 8} In turn, this can lead to accumulation of p53 in the nucleus and transcriptional activation.^{7, 8} Our findings that p53 is phosphorylated at S15 after HN2 exposure are consistent with previous reports of rapid activation of ATM/ATR/DNA PKcs pathways following exposure to DNA damaging agents including ultraviolet light, ionizing radiation, and mustard vesicants.^{10, 12, 32} It should be noted that genotoxic insult could also induce phosphorylation of p53 on other sites important in regulating p53 activity including S9, S20, S46, and T18.⁸ S15 phosphorylation also enhances the association of p53 with histone acetyltransferases such as CBP/p300 leading to acetylation of histones in the neighboring vicinity and in the C-terminal domain of p53 including K382, promoting gene transcription.^{29, 33} Acetylation of p53 can suppress ubiquitin binding increasing p53 stability and its transcriptional functions.^{29, 33}

Cytosolic p53 is known to interfere with pathways important in maintaining cell homeostasis including the regulation of oxidative stress.^{34, 35} It has also been reported to regulate glucose, lipid, amino acid, and nucleotide metabolism.³⁵ In addition, in response to genotoxic stress, p53 can modulate cell death pathways including apoptosis via protein-protein interactions with members of the Bcl-2 family at cytosolic and mitochondrial sites, and ferroptosis via alterations in iron metabolism and reactive oxygen species.^{35, 36} Thus, our findings of increased p53 in the cytoplasm of HaCaT cells in response to HN2 may contribute to mechanisms of vesicant toxicity that are transcription-independent.⁵

In previous studies, we reported that, in A549 lung epithelial cells, HN2 cross-links thioredoxin and thioredoxin reductase, forming intermolecular protein complexes and, as a consequence, inhibition of the thioredoxin system.^{21, 22} Similarly, the present studies demonstrate that HN2 can cross-link p53, forming several high-molecular-weight protein aggregates, including dimers, trimers, tetramers, and oligomers in intact cells and in vitro using recombinant p53 protein. The formation of p53 aggregates by HN2 is p53 status independent. Thus, p53 cross-links were detected in cells expressing wild-type p53 (e.g., HEK-293 and MLE-15 cells) and in cells expressing mutant p53 including HaCaT, A431, and CX-1 cells. All of these cells are known to overexpress p53.²⁷ In contrast, no cross-links were evident in cells expressing low levels of p53 including A549, HeLa, and MCF-7 cells (data not shown). These findings indicate that the extent of HN2 modification depends on expression levels of p53. This may lead to differential damage responses in different cell types following HN2 exposure. Oligomerization and aggregation of p53 were also observed after treatment of HaCaT cells with other bifunctional alkylating agents including melphalan, chlorambucil, and cisplatin. These observations suggest that p53 is a preferential cellular target for alkylating agents. This is supported by an earlier study showing that 2-phenylethynesulfonamide caused p53 aggregation in MEF cells.²⁰ Formation of p53 stable cross-links has also been described in cellular extracts of MEF cells or immunoaffinity-purified wild-type murine p53 protein after glutaraldehyde treatment.^{20, 37} We also found that HN2 significantly increased K48-linked, but not K63-linked, ubiquitination of p53; this was more pronounced in p53 cross-linked oligomers than in monomers. These data suggest that cross-linking promotes proteasomal degradation of p53 resulting in decreases of p53.

Our findings that HN2-induced alkylation of p53 is more effective when the protein is reduced by DTT suggest that cysteine residues are key for cross-linking. This is supported by our LC-MS/MS analysis showing that HN2 predominantly alkylated cysteine residues in p53 and, to a lesser extent, lysine and histidine residues. This is in agreement with studies on the reactivity of nitrogen mustards, including HN2 and tris(2-chloroethyl) amine (HN3) with model peptides containing selective nucleophilic amino acids. In these studies, nitrogen mustard-induced adduct formation was shown to occur more readily with cysteine than with lysine or histidine.³⁸

Earlier studies demonstrated that cysteine modifications in the p53 core domain, including reversible oxidation, nitrosylation, and glutathionylation, as well as nonreversible thiol-alkylation, alter the activity of p53.^{15, 17, 18, 39–41} Our HPLC-MS/MS analysis revealed that, of the ten cysteine residues in the p53 core domain, seven were modified by HN2 including C124, C135, C141, C176, C182, C275, and C277, forming both monoadducts and cross-links. Figure 8 shows the solvent accessibility and distance between cysteine sulfur atoms in a subunit of DNA binding domain based on the SAXS model of human p53.²⁵ Cysteine 124, 182, and 277 are exposed on the protein surface (Figure 8B) and their thiol groups most accessible for alkylation (accessibility values are 6.03, 26.52, and 15.96, respectively); thus, these residues are likely preferential sites for HN2 modification. These findings are consistent with previous reports that C124, C182, and C277 are the most reactive nucleophiles on the p53 DBD for reactions with N-ethylmaleimide, 2-sulfonylpyrimidines, PK11000, and APR-246.^{17, 19, 41} Mutagenesis studies have also shown that substitution of serine for cysteine at position 121 or 272 (corresponding to C124 and C277 in human p53) partially blocked transactivation.⁴² In line with these studies, Kim et al.¹⁵ reported that 15-deoxy-Δ^12,14-prostaglandin J₂ abrogated DNA binding as a result of directly targeting C277 on p53 in MCF-7 cells, which expresses wild-type p53 protein. Moreover, mass spectrometry demonstrated that C124 and C141 in human p53 can be glutathionylated and that S-glutathionylation inhibits p53 DNA binding.⁴⁰ We also found that C275 is partially exposed on the protein surface (Figure 8B), and it may be accessible for HN2 alkylation after side chain rotation. Moreover, HN2 modified three buried cysteines, C135, C141, and C176, all of which have solvent accessibility values of zero (Figure 8B). These residues are located either at the beginning or at the end of β-strands. p53 contains intrinsically disordered regions that undergo conformational changes in response to stress signals, redox modification, and chemical alkylation.⁴³ It is possible that initial HN2 modifications trigger conformational changes in p53, which uncover buried residues making them accessible to HN2 alkylation. This may also bring distant reactive thiols in closed proximity, which would facilitate the formation of intramolecular cross-links. Modification of C135, C141, and C176 has been reported previously by a reaction of T-p53C-Y220C with 3-benzoylacrylic acid.⁴⁴ Cysteine mutants of murine p53 at C132, C138, and C272 (corresponding to C135, C141, and C275 in human p53) were found to partially block DNA binding activity while mutation of murine p53 at C173 (corresponding to C176 in human p53) completely abolished DNA binding and transactivation activity.⁴² In addition, HN2 cross-linked cysteine with several lysine and histidine residues including H115, H178, K132 and K139 in the p53 core domain, modifications preventing DNA binding due to steric hindrance. This may result in the reduction of active p53 protein and decreases in DNA binding activity, which can lead to the suppression of p53 transcription.

Figure 8. Molecular model of human p53 core DNA binding domain.

(A) Location of cysteine residues in the p53 DNA binding domain (top) and relative distances between the thiol groups of cysteine residues in the p53 model (bottom). Cysteine residues from one subunit are shown in a stick representation in red. (B) Surface of the p53 core domain (top) and solvent accessibility surface of cysteine residues in the p53 model (bottom). Solvent-exposed cysteine residues are shown in red.

It is of interest to note that HN2 suppressed p53 DNA binding in a concentration-dependent manner in HaCaT cells; inhibition was more pronounced at low concentrations of HN2 (10 μM) when compared to higher concentrations (100–200 μM). These data suggest that HN2 modulates p53 and DNA binding via distinct mechanisms at different HN2 concentrations and that these are likely dependent on many factors including levels of p53 and post-translationally modified p53, proteins that regulate p53, DNA binding cofactors, chromatin remodeling proteins, and the extent of HN2 modifications/cross-links on p53 and possibly other proteins.⁴⁵ In addition, we cannot rule out the possibility that HN2 alkylates side chains of other amino acids besides cysteine, lysine, and histidine, processes that may also alter p53 activity. Further studies are needed to better define mechanisms by which HN2 modulates p53-DNA binding.

In summary, at least two distinct mechanisms are likely involved in the regulation of p53 by HN2; one is activation of DNA damage response and the second is p53 protein covalent modifications. The former is likely to be mediated by nuclear and cytosolic accumulation of total p53 and HN2-induced post-translational modifications, processes that are important in mediating p53 transactivation and regulating cell homeostasis. The latter is due to site-specific binding of HN2 to p53, a process that may contribute to the inactivation of p53. HN2/p53 cross-links are targets for proteasomal degradation leading to decreases in levels of p53 and its functional activity. HN2 cross-linking of p53 forming dimers and tetramers is consistent with the homodimeric structure of the protein. The formation of higher-molecular multimers indicates that multiple p53 subunits are in close proximity. Differential expression of p53 subunits in the cytoplasm and nucleus may be responsible for the increased formation of HN2 cross-linked multimers. There remains a need to better define mechanisms by which HN2 cross-linking of p53 leads to the formation of dimers and multimers in the cytoplasm and nucleus of cells and the role of these modifications on the functional activity of p53. One can speculate that the formation of p53 modifications may have a dominant negative effect or a gain/loss of function of p53 (reviewed in Aubrey et al.³⁶). Any of these activities may play a role in abetting/hindering HN2-induced toxicity. Drug candidates that can counter these effects may be effective mustard countermeasures.

Abbreviations

HN2

nitrogen mustard or mechlorethamine

SM

sulfur mustard

DTT

dithiothreitol