Mapping the modification of histones by the myeloperoxidase-derived oxidant hypochlorous acid (HOCl)

Histones are critical for the packaging of nuclear DNA and chromatin assembly, which is facilitated by the high abundance of Lys and Arg residues within these proteins. These residues are also the site of a range of post-translational modifications, which influence the regulatory function of histones. Histones are also present in the extracellular environment, following release by various pathways, particularly neutrophil extracellular traps (NETs). NETs contain myeloperoxidase, which retains its enzymatic activity and produces hypochlorous acid (HOCl). This suggests that histones could be targets for HOCl under conditions where aberrant NET release is prevalent, such as chronic inflammation. In this study, we examine the reactivity of HOCl with a mixture of linker (H1) and core (H2A, H2B, H3 and H4) histones. HOCl modified the histones in a dose-and time-dependent manner, resulting in structural changes to the proteins and the formation of a range of post-translational modification products. N -Chloramines are major products following exposure of the histones to HOCl and decompose over 24 h forming Lys nitriles and carbonyls (aminoadipic semialdehydes). Chlorination and dichlorination of Tyr, but not Trp residues, is also observed. Met sulfoxide and Met sulfones are formed, though these oxidation products are also detected albeit at a lower extent, in the non-treated histones. Evidence for histone fragmentation and aggregation was also obtained. These results could have implications for the development of chronic inflammatory diseases, given the key role of Lys residues in regulating histone function.


Introduction
Histones are highly basic proteins that play an important role in packaging and stabilising DNA in cell nuclei. DNA (ca. 147 base pairs) wraps twice around an octamer of core histones, to form a nucleosome consisting of two H2A-H2B dimers and a tetramer of H3-H4 [1][2][3]. Nucleosome cores, unbound DNA of variable length, together with the linker histone H1, are then further packaged into chromatin [1]. Histones are rich in positively charged Lys and Arg residues, which facilitate the tight wrapping and packaging of the negatively charged DNA in chromatin [1]. These residues also play a key role in the regulatory function of these proteins [4]. Thus, histones can undergo multiple post-translational modifications (PTM), including methylation, acetylation, citrullination, phosphorylation, SUMOylation, ubiquination and ADP-ribosylation, which occurs particularly in the N-terminal region of the protein [5,6]. Combinations of these modifications to a particular genomic region constitutes a "histone-code", which alters the chromatin folding thereby regulating the accessibility of DNA to various binding and modification proteins, which can activate or repress gene expression [4,5,7].
Histones are a key component of neutrophil extracellular traps (NETs), and as such, can be released into the extracellular environment [8,9]. The release of NETs is a defensive pathway to trap and kill pathogens [10,11], but increasingly, these structures are implicated in the development of disease [12]. Decondensed chromatin and histones are released by a pathway termed NETosis, which is triggered by a range of stimuli [13]. The activation of peptidyl-arginine deiminase 4 (PAD4), which converts the Arg residues of histones into citrulline, promotes chromatin unfolding and is generally regarded as a driver of suicidal NET release [13,14]. In addition to histones, NETs also contain a range of defensive proteins, including myeloperoxidase (MPO), which retains its enzymatic activity to facilitate the extracellular elimination of pathogens [13,15]. Thus, MPO catalyses the reaction of hydrogen peroxide (H 2 O 2 ) with chloride ions to produce hypochlorous acid (HOCl), which is a strong oxidant and potent antimicrobial agent [16,17].
In addition to killing bacteria and invading pathogens, there is evidence that HOCl can also damage host tissue, which is linked to the development of inflammatory pathologies [16,17]. Proteins are a key target for HOCl, owing to their abundance and the presence of multiple reactive amino acid side chains, with Met > Cys ≫ cystine ~ His ~ α-amino > Trp > Lys > Tyr ~ Arg > Gln ~ Asn [18,19]. The HOCl-induced modification of amino acid side chains, together with the induction of fragmentation, crosslinking and aggregation alters the structure and function of proteins [20,21]. The reaction of HOCl with Tyr results in the formation of 3-chlorotyrosine (Cl-Tyr), which is commonly used as a biomarker to assess the involvement of MPO in disease [22]. Chromatin is also a target for HOCl, which can modify DNA to form chlorinated nucleobase products including 5-chlorocytidine, 5-chlorouracil and 8-chloroguanosine [23]; these are elevated in diseased tissues and fluids from patients with different inflammatory pathologies [24][25][26]. This, together with the knowledge that histones and active MPO are localised together on the NET backbone, suggests that histones could be a favourable target for HOCl in vivo. Previous studies have shown that exposure of the core histones H2A, H2B, H3 and H4 to HOCl results in the formation of Cl-Tyr and 3,5-dichlorotyrosine (diCl-Tyr), with the YXXK/KXXY sequence motif of histones H2A, H2B and H4 shown to be the preferred chlorination sites [27]. However, there is no information regarding the formation of other oxidation products, particularly those on the highly abundant Lys residues. This is significant given the reactivity of Lys with HOCl and the importance of post-translational modification of these residues in the histone-mediated regulation of DNA replication and transcription [2,4].
In this study, we examine the reactivity of HOCl with a mixture of linker (H1) and core (H2A, H2B, H3 and H4) histones. We show that HOCl induces histone modification in a dose-and time-dependent manner and results in fragmentation, aggregation, loss of specific amino acids, and the formation different oxidation products. Under the conditions used in this study, histones H1 and H4 appeared to be the most susceptible to modification. Evidence was obtained for the loss of Lys and formation of unstable N-chloramines together with Lys nitriles and carbonyls (aminoadipic semialdehydes). A range of other oxidation products were formed, particularly on prolonged incubation with HOCl including Cl-Tyr, diCl-Tyr, Met sulfoxide, and Met sulfone. The formation of these products was observed on exposure of the histones to < 20fold molar excess HOCl, which is readily achievable under pathophysiological conditions and could be relevant to NET-derived host tissue damage at sites of inflammation.

Reagents and materials
All aqueous solutions were prepared using nanopure H 2 O from a MilliQ system (Millipore). All chemicals and reagents were of the highest purity available and purchased from Sigma-Aldrich / Merck unless stated otherwise. Experiments were performed with an unfractionated, commercial preparation of histones from calf thymus (Sigma-Aldrich, #H9250). A stock solution was prepared in PBS at a concentration of 2 mg/mL (133 μM assuming an average histone molecular mass of 15 kDa) for experiments. The concentration of HOCl was determined by UV absorbance at 292 nm at pH 11 using an extinction coefficient of 350 M − 1 cm − 1 [28].  (15 min) and H 2 O (3 × 5 min) and visualizing the protein bands using silver nitrate (0.2% w/v) and a developing solution containing sodium carbonate (30 g/L), 0.05% v/v formaldehyde and 2% w/v sodium thiosulfate. Colour development was stopped by the addition of EDTA (1.4% w/v), and the gel was imaged using a Sapphire Biomolecular Imager (Azure Biosystems). Change in the intensity of staining of protein bands was assessed using Image J.

Quantification of Lys by fluorescamine
Histones (5 μM) were treated with HOCl for 1 h before they were diluted (1:30) in 0.1 M borate buffer, pH 9. The protein samples (150 μL) were then mixed with ice-cold fluorescamine (50 μL, 0.3 mg/mL in acetone) in a black 96-well plate (Greiner). The plate was incubated at 20 • C for 15 min and fluorescence (λ ex 390 nm, λ em 475 nm) was subsequently recorded in a plate reader (Spectra Max i3x, Molecular Devices). The concentration of Lys was determined based on a standard curve prepared using 6-aminocaproic acid (0-250 μM).

LC-MS amino acid analysis and detection of oxidation products
Histones (20 μM) were treated with HOCl (0-400 μM), or MPO (100 nM) and H 2  The samples were diluted in 0.1% v/v formic acid before solid-phase extraction with activated mixed-mode strong cation exchange cartridges (Strata X-C, Phenomenex) [31]. The samples were washed with 0.1% v/v formic acid in water and 0.1% v/v formic acid in acetonitrile before elution from the cartridges with 20% (v/v) acetonitrile containing 1% (w/v) NH 4 OH. The eluted amino acids were dried in a vacuum concentrator for 4 h at 60 • C and the pellet reconstituted in 0.1% formic acid in water before the amino acids were quantified by electrospray ionisation-liquid chromatography-mass spectrometry (ESI LC-MS) using an Impact HD II Q-TOF mass spectrometer (Bruker Daltonics) in positive-ion-mode connected online to a Dionex Ultimate 3000 chromatography system (Thermo Fisher Scientific). The samples were separated using an Imtakt Intrada Amino Acid 100 × 3.0 mm column as described previously [31].  The peptide samples were acidified with 10% v/v TFA before adding them to the filter discs, followed by centrifugation at 1200g for 3 min, and washing twice with 50 μL 0.1% v/v TFA. Peptides were eluted by a solution containing acetonitrile (80% v/v) and formic acid (0.1% v/v) and centrifugation at 1200g into new tubes. Samples were then dried using a vacuum concentrator and reconstituted in 0.1% v/v formic acid (20 μL) for MS analysis. Histone peptide samples were analyzed by LC-MS on an Impact II ESI-QTOF (Bruker Daltonics) mass spectrometer in the positive ion mode with an Apollo ion source connected on-line to a Dionex Ultimate 3000 chromatography system (Thermo Fisher Scientific) with a CapLC pump. Analytes were separated using a 150 × 0.5 mm Luna C18 column (Phenomenex) at 20 • C with a flow rate of 20 μL min − 1 and gradient elution using 0.1% formic acid (Solvent A) and 80% acetonitrile/0.1% formic acid (Solvent B). The electrospray needle was held at 4500 V, with end plate offset of 500 V and temperature of 200 • C.
Nitrogen gas was used for both the nebuliser (0.7 Bar) and as the drying gas (6.0 L min − 1 ). MS1 precursor scans (150-2200 m/z) were followed by data dependent MS/MS fragmentation of the three most intense precursors with sampling rates of 2 Hz.
The MS data was searched against the bovine reference Uniprot proteome (UP000009136) using MaxQuant software (version 1.6.0.1). Modifications included in the analysis of the histones were oxidation on Met (M) (+ 16 and + 32 Da), chlorination (+ 34 Da) and dichlorination (+ 68 Da) on Tyr (Y) and Trp (W), nitrile (− 4 Da), aminoadipic semialdehyde (− 1 Da) and 2-aminoadipic acid (+ 15 Da) on Lys (K), and oxindolyalanine (+ 16 Da), dioxindolyalanine (+ 32 Da), N-formylkynurenine (+ 32 Da) and kynurenine (+ 4 Da) on Trp (W). Quantification of the occupancy of modified peptides from the histone samples were assessed using Compass QuantAnalysis (version 4.3). The occupancy is defined as the peak area for a particular precursor ion with a particular charge state matching a specific modified peptide divided by the sum of the peak areas for all identified peptides (modified and nonmodified) with the same charge state matching the same peptide sequence. Missed cleavage as well as other charge states were accounted for when calculating the occupancy. Histone sequence alignments were made in MEGA X (version 10.1.8).

Quantification of protein carbonyls with 2,4-dinitrophenylhydrazine
The formation of carbonyls on the histones after treatment with HOCl was quantified using a commercial protein carbonyl content assay kit (Sigma-Aldrich) in accord with the manufacturer's instructions. Briefly, histones were mixed with an equal volume of 2,4-dinitrophenylhydrazine solution (DNPH, from kit) and incubated for 10 min at 21 • C. The proteins were precipitated by mixing with ice-cold trichloroacetic acid (TCA, 20% w/v), and pelleted by centrifugation at 13,000 g for 2 min and washed twice with ice-cold acetone. The pellet was then dissolved in guanidine (6 M) and the carbonyl concentration determined by measuring the absorbance at 375 nm using a plate reader (Spectra Max i3x, Molecular Devices) with an extinction coefficient of 22,000 M − 1 cm − 1 . The protein concentration was determined using the Pierce™ BCA Protein Assay Kit (Thermo Fisher) according to the manufacturer's instructions. Data are expressed as nmol carbonyl/mg protein.
The formation of carbonyls was also visualised with the Oxyblot™ Protein Oxidation Detection Kit (Millipore) following separation by SDS-PAGE and Western blotting with an anti-DNPH antibody as described previously [29]. Histones were treated with HOCl before being denatured by the addition of an equal volume of SDS (12%) and derivatisation of the carbonyl groups with DNPH (supplied in kit) for 15 min at 21 • C. The supplied Neutralising reagent was added before reducing the samples by treatment with DTT (50 mM). Protein (25 μg) was then separated by SDS-PAGE using Novex 16% Tricine gels (Thermo Fisher) at 200 V for 35 min, and transferred onto a polyvinylidene fluoride (PVDF) membrane (Thermo Fisher) at 20 V for 7 min. Membranes were blocked with 5% (w/v) BSA in PBS-T (0.1% Tween-20 in PBS) for 2 h at 21 • C, washed three times in PBS-T for 5 min, and then incubated with the primary antibody (rabbit anti-DNP, dilution 1:150) overnight at 4 • C. The membranes were washed three times in PBST and incubated in HRP-conjugated secondary antibody (1:300) for 1 h at 21 • C. The membrane was washed a further three times in PBS-T and imaged with ECL enhanced chemiluminescence substrate (PerkinElmer) using a Sapphire Biomolecular Imager (Azure Biosystems). To ensure equal loading in the wells, membranes were stained with Ponceau-S (0.1% w/v in 5% v/v acetic acid) for 10 s and washed briefly with PBS-T before imaging.

Statistical analyses
Statistical analyses were performed using GraphPad Prism (version 9; GraphPad Software) using 1-way or 2-way ANOVA with P < 0.05 taken as significant. Data represent mean ± S.E.M. from at least 3 independent experiments in each case, with the details of the specific multiple comparison tests used outlined in the Figure Legends.

HOCl treatment induces fragmentation and aggregation of histones
The preparation of isolated calf thymus histones containing H1, H2A, H2B, H3 and H4 (1 mg/mL) was treated with increasing concentrations of HOCl (0-500 μM) for 15 min and 24 h before separation by SDS-PAGE and visualisation of the protein bands with silver staining. The protein bands were assigned to histone H1, H2A, H2B, H3 and H4 on the basis of their molecular weights [3]. Changes in the relative intensity of the individual histone proteins bands was apparent with increasing concentrations of HOCl, which were more striking following incubation for 24 h (Fig. 1A-B). A significant decrease in the intensity of the bands assigned to H1 and H4 together with an increase in high molecular mass aggregates was observed with > 50 μM HOCl and 24 h treatment ( Fig. 1C-F). There were also changes seen in the density of the bands assigned to H2A, H2B and H3, but these were more variable between replicate experiments ( Figure S1 and S2). A similar pattern of reactivity was observed in analogous experiments where the histones were treated for 24 h with an MPO/H 2 O 2 /Cl − system to generate HOCl enzymatically ( Figure S3). The loss of histone H1, formation of high molecular mass aggregates and change in H4 staining was not seen in control experiments with only MPO or H 2 O 2 , consistent with HOCl-induced structural changes ( Figure S3).

Modification of amino acids and formation of oxidation products
Histones are rich in Lys residues, which together with His, are favourable targets for HOCl, resulting in the formation of N-chloramines, which can undergo secondary reactions to promote further modifications of the protein [20,32]. Therefore, studies were performed to examine the formation and stability of histone N-chloramines using the TNB assay [29]. Exposure of histones (50 μM) to HOCl (200 μM) resulted in the formation of unstable N-chloramines, as evidenced by the consumption of TNB ( Fig. 2A). Under these conditions, the formation of N-chloramines accounted for ca. 55% of the initial HOCl added to the histones. The histone-derived N-chloramines decomposed in a time-dependent manner on incubation over 24 h (at 21 • C). The modification of Lys residues on histones after reaction with HOCl was also evaluated by monitoring changes in binding of the fluorescent probe fluorescamine [29]. Exposure of the histones (5 μM) to HOCl (25-1000 μM) resulted in a significant decrease in fluorescamine fluorescence with ≥ 50 μM HOCl, consistent with the modification of Lys residues (Fig. 2B).
The loss of Lys and other amino acid residues known to be reactive with HOCl, including Met, Trp, His, Arg and Tyr, was examined further using LC-MS amino acid analysis following protein hydrolysis [31]. The concentration of each parent amino acid was normalised to the concentration of Phe, which was not affected by the HOCl treatment (data not shown), to account for differences in the efficiency of protein hydrolysis between experiments. Evidence was obtained for a dose-dependent decrease in Lys, His, Tyr and Arg on treating the histones (20 μM) with increasing concentrations of HOCl (0-400 μM) for 24 h (Fig. 3A-D). The loss of these amino acids was generally greater on exposure of the histones to HOCl for 24 h compared to 15 min ( Fig. 3A-D). In contrast, the data for Met and Trp were found to be highly variable between experiments (data not shown). This may be associated with a poor, inconsistent, recovery of these amino acids in the non-treated histones, which was not seen with the other amino acids.
The loss of Arg on exposure of histones to > 10-fold molar excess HOCl was confirmed in separate experiments with PTQ, which displays altered fluorescence on binding to Arg ( Figure S4). In this case, the change in histone Arg concentration measured by PTQ was independent of the incubation time ( Figure S4). A higher molar ratio of oxidant: protein was required to see a significant loss of Arg compared to the modification of Lys, or loss of His, and Tyr (Fig. 3). This is consistent with the lower reactivity of HOCl with the Arg guanidino group compared to the Lys amino group [18].
In addition to the loss of parent amino acids, a dose-dependent increase in the formation of Cl-Tyr was observed on reaction of the histones with HOCl (Fig. 3E) and the MPO/H 2 O 2 /Cl − system (Fig. 3F). In experiments with HOCl, the concentration of Cl-Tyr formed was similar on treating the histones for 15 min and 24 h, with the exception of the 20-fold molar excess condition, where a decrease relative to that seen on 15 min incubation was apparent (Fig. 3E). This may reflect the formation of diCl-Tyr on longer incubation of the histones with HOCl, which was not quantified in these experiments. In contrast, a greater formation of Cl-Tyr was observed with the MPO/H 2 O 2 /Cl − system on comparing 15 min and 24 h incubation (Fig. 3F), though the overall levels of this oxidation product were somewhat lower than that observed with HOCl (Fig. 3E). This may be related to exposure of the histones to a flux of HOCl rather than adding a bolus of the oxidant. However, there was no significant change in the concentration of any of the parent amino acids seen in experiments with MPO/H 2 O 2 /Cl − (data not shown). This could reflect the incomplete conversion of H 2 O 2 to HOCl by MPO, such that the histones are exposed to a lower concentration of oxidant compared to the experiments with reagent HOCl.
The formation of other Tyr-derived oxidation products, including di-Tyr and 3,4-dihydroxyphenylalanine (DOPA), Met sulfoxide, and the Trp-derived products 5-hydroxy-tryptophan and kynurenine, which have been reported in studies with other proteins [20] was also assessed. While there was no consistent evidence for an increase in the formation of DOPA, Met sulfoxide, 5-hydroxy-tryptophan and kynurenine on the histones following treatment with HOCl, there was a dose-dependent increase in the formation of di-Tyr ( Figure S5). The formation of di-Tyr was also quite variable between experiments but was significant in experiments where the histones were incubated with a 20-fold molar excess of HOCl for 24 h ( Figure S5).
The experiments were then extended to examine the formation of protein carbonyls, including aminoadipic semialdehydes, which are a decomposition product of Lys-derived, N-chloramines, and could contribute to the significant loss of Lys amine groups observed following incubation of the histones and HOCl for 24 h (Fig. 3A). A timedependent increase in the formation of protein carbonyls, as evidenced by derivatisation of these groups with DNPH was observed on reaction of the histones (50 μM) with HOCl (200 μM) (Fig. 4A). The localisation of the carbonyl groups to specific histone proteins was then examined by separation of the proteins by SDS-PAGE and Western blotting with an anti-DNPH antibody to visualise the histones containing carbonyls. There was a time-dependent increase in the formation of carbonyls on incubation of the histones with a 4-fold molar excess of HOCl over 24 h, indicated by the greater intensity of the multiple bands in the DNPH Western blot (Fig. 4B). This is consistent with carbonyl formation on different histone proteins present in the commercial preparation used, particularly histones H1, H2A, H2B and H3. However, it is difficult to precisely assign the bands to individual histones given the similarity in the molecular weight of histones H2A, H2B and H3. Equal protein loading was confirmed by Ponceau red staining of the membrane ( Figure S6). In contrast, there was no change in the staining of the anti-DNP antibody over time on incubation of the histones in the absence of HOCl (data not shown).

LC-MS peptide mass mapping studies on HOCl-modified histones
To gain a greater understanding of the extent and nature of HOClinduced modifications to specific histones, tryptic digests were prepared after exposure of the histone preparation (20 μM) to HOCl (0-200 μM) for 15 min and 24 h, and the samples were analyzed by LC-MS. The most abundant proteins present in the calf thymus histone preparation were confirmed to be histones following a database search against the bovine proteome from Uniprot. The distribution of different identified histone proteins in the samples was determined, with H2B found to be the most abundant histone (44%), followed by H4 (27%), H2A (19%), H1 (6%) and H3 (4%), based on the number of peptide spectrum matches (PSMs). This is somewhat comparable to the distribution of histones seen on separation by SDS-PAGE, though H4 appears to be less abundant than H3, on using silver staining to visualise the protein bands (Fig. 1).
The experiments were extended to examine the formation of PTMs on different amino acid residues, including Tyr, Lys, Met and Trp, following exposure of the histones to HOCl. Evidence was obtained for the chlorination and dichlorination of Tyr (Y) on each of the histones, with a total of 16 chlorination and 9 dichlorination sites detected.  Table 1 shows the local sequences containing the modified Tyr residues on each type of histone, whereas Table S1 shows a complete list of all detected tryptic peptides, including those with missed cleavage sites, which yielded longer peptides, where trypsin failed to cleave at the predicted Lys or Arg residue. In addition, in some cases, single amino acid substitutions were apparent, owing to the presence of multiple isoforms of the histone proteins (Table S1). Most Tyr residues are well conserved across the individual core histones (H2A, H2B, H3 and H4) based on sequence alignments ( Figure S7-S10). However, histone H1 has many isoforms and lacks the sequence conservation seen with the core histones ( Figure S11). Therefore, peptides from histone H1 isoforms have been identified more specifically (Table 1 and S1).
In some cases, the chlorination (and dichlorination) of the peptide Tyr-containing peptides was observed together with other modifications, including Met (M) sulfoxide and Lys (K) nitrile (Table 1 and S1). The extent and nature of histone Tyr chlorination was dependent on the concentration of HOCl (Table 1). The extent of chlorination increases in a concentration-and time-dependent manner on exposure of histones to HOCl, as indicated by the number of PSMs from chlorinated and The amino acid and Cl-Tyr concentrations were normalised Phe, which is unchanged by HOCl treatment under these conditions. Data represent the mean ± S.E.M. of n ≥ 6 experiments. Statistically significant change compared to the untreated (or PBS) control with *, **, *** and **** showing P < 0.05, 0.01, 0.001 and 0.0001 respectively, determined by 2-way ANOVA with a Dunnett's post hoc test. #### shows a significant change between each H 2 O 2 concentration by 2-way ANOVA with a Tukey's post-hoc test. dechlorinated peptides (Fig. 6A). Treating the histone with ≥ 5-fold molar excess HOCl for 24 h resulted in a greater number of chlorinated peptide PSMs (Fig. 6A) and was more likely to result in the detection of peptides that also contained Lys nitriles and/or Met sulfoxide (Table 1 and S2). No evidence was obtained in this study for chlorination of Trp (W) residues, which has been reported on exposure of other proteins to HOCl [20]. Similarly, there was no reproducible formation of other Trp-derived oxidation products, including oxindolyalanine, dioxindolyalanine, N-formylkynurenine and kynurenine. This may reflect the low abundance of Trp in histone proteins.
To further investigate whether specific Tyr residues are more or less susceptible to HOCl-induced chlorination, the occupancy of different chlorinated Tyr residues was examined by analysing the extracted ion chromatograms for the chlorinated peptides m/z values and defining the peak areas at the corresponding retention time. The area under the peak defines the intensity of the modified peptide in the specific sample. The overall occupancy was calculated as the intensity of the modified peptide(s) divided by the total intensity of unmodified peptide(s) and modified peptide(s), that is, the total of the intensities for all peptides containing the specific Tyr residue. Fig. 6B-D shows the calculated occupancy for selected Tyr residues in histone H1 (Tyr-71) and H4 (Tyr-51 and Tyr 72), which appear to be particularly sensitive to chlorination by HOCl ( Table 1). As expected on the basis of the above data, the extent of chlorination and dichlorination of histone H1 (peptide ALAAGYDVEKNNSR from isoforms H1.1, H1.2, H1.3, H1.4 and H1.12) and histone H4 (peptides ISGLIYEETR and DAVTYTEHAK) was dependent on the concentration of HOCl and the incubation time (Fig. 6). In each case, a greater extent of chlorination and particularly dichlorination, was observed on treating the histones with HOCl for 24 h compared to 15 min. These data provide support for a role for histone-derived N-chloramines (on Lys or His residues) in promoting the transfer of chlorine to Tyr residues.
As indicated above, there was also evidence for the formation of Lys oxidation products with a total of 25 and 12 sites containing Lys nitriles and aminoadipic semialdehydes, respectively, following exposure to HOCl (Table 2 and Table S2-3). The MS/MS fragmentation spectrum for peptide ALAAAGYDVEKNNSR on histone H1 containing a nitrile at Lys-75 is shown in Fig. 5D. The formation of 2-aminoadipic acid was also observed, but the number of PSMs detected were very limited in this case (data not shown). Missed cleavages were observed in most peptides containing modified Lys side chains (Tables S2 and S3), consistent with the failure of trypsin to recognize modified Lys residues. The formation of nitrile and aminoadipic semialdehydes on Lys increased in a concentration-dependent manner on exposure of histones to HOCl based on the number of PSMs (Fig. 7A-B). There was no appreciable timedependent change in nitrile formation on comparing the PSMs on treating the histones with HOCl for 15 min and 24 h (Fig. 7A). However, the aminoadipic semialdehyde PSMs were greater at 15 min compared to 24 h (Fig. 7B). This contrasts with the quantification of carbonyls by DNPH on HOCl-treated histones, which increased with incubation time (Fig. 4B).
Met is a key target for HOCl [19] but it is also readily oxidised during sample preparation and storage. Thus, relatively high yields of Met sulfoxide were observed in the peptides from the non-treated histones, and only a small increase in the formation of this PTM was apparent with HOCl treatment (Fig. 7C). The extent of Met sulfoxide formation was similar on incubation of the histones with < 10-fold molar excesses of HOCl on comparing 15 min and 24 h incubation. A significant decrease in the presence of Met sulfoxide was seen at 24 h when the histone was exposed to a 10-fold excess of HOCl (Fig. 7C). This could reflect over-oxidation of the Met, or the formation of other PTMs on the Met-containing peptides. However, although there was a small HOCl-dependent increase in the formation of Met sulfone on the histones, again the number of PSMs were greater following 15 min exposure, rather than 24 h (Fig. 7D, Table 3 and S4). Sulfone formation was generally seen at similar positions to sulfoxide formation, and some of the peptides containing the sulfone PTM also contain Cl-Tyr (Table 3 and S4).

Discussion
Reaction of a mixture of calf thymus histones with HOCl resulted in the alteration of protein structure and the formation of N-chloramines, which was accompanied by a loss in Lys, His, Tyr, Met and Arg. Modifications were observed on both the linker histone (H1) and the core histones (H2A, H2B, H3 and H4), and occurred in a time-and dosedependent manner. There was evidence for the formation of Lysderived nitriles and carbonyls (aminoadipic semialdehydes), and chlorination and dichlorination of Tyr at multiple sites in each type of histone. Met sulfoxide and Met sulfone were also observed, though these products were abundant in the non-modified histones. A similar pattern of modification was seen in analogous experiments where HOCl was generated enzymatically by MPO in the presence of H 2 O 2 and Cl − .
N-chloramines are a key product of the reaction of HOCl with proteins and are formed on the side chain of Lys and His residues, together with the N-terminal amino groups [20,32]. Histones are rich in Lys residues, and the formation of chloramines accounted for a significant proportion of the HOCl added (55% with 4-fold molar excess HOCl). It is well established that protein chloramines can undergo a variety of secondary reactions, resulting in the formation of different oxidation products and the transfer of chlorine to other sites within the protein (reviewed [20,21]). Decomposition of the histone chloramines occurred over 24 h and was associated with the formation of Lys-derived oxidation products, together with Cl-Tyr and diCl-Tyr, particularly in experiments with higher molar ratios (> 5-fold) of HOCl. Evidence for the formation of histone-derived carbonyls following treatment with HOCl was obtained in experiments using DNPH and confirmed in peptide mass-mapping studies. Carbonyl formation in this case, is attributed to the elimination of HCl from protein monochloramines (RNHCl) resulting in intermediate imines, which are hydrolysed to form ammonium ions and aminoadipic semialdehydes [20,32,33].
Based on the peptide mass-mapping studies, aminoadipic semialdehydes were most prevalent on the histone H1 isoforms, which are particularly Lys-rich, but these products were also formed on each of the core histones. Interestingly, a higher number of peptides containing aminoadipic semialdehydes were observed on treating the histones with HOCl for 15 min compared to 24 h (based on PSMs). This may reflect

Table 1
Chlorination of specific Tyr (Y) residues seen on exposure of histones to different molar ratios of HOCl for 24 h unless indicated otherwise.
additional modification of the peptides containing this PTM, and the formation of other products with mass changes that were not included in the MaxQuant database search. It is also possible that DNPH is derivatising other reactive moieties formed on the histones following reaction with HOCl. For example, DNPH has been reported to react with protein sulfenic acids in addition to aldehyde/ketone groups [34]. However, although HOCl can form protein-bound sulfenic acids [21], histones lack free Cys residues, so it is less likely that this reaction will be applicable.
Lys nitriles were also formed at multiple sites on each of the histones, which is attributed to the decomposition of protein dichloramines (RNCl 2 ) [35,36]. In general, Lys nitrile formation was only observed on treating the histones with > 5-fold excess of HOCl, which may favour the generation of protein dichloramines compared to monochloramines. Nitrile formation did not appear to increase over time, on the basis of the number of PSMs. However, again this may be related to the formation of other PTMs on the nitrile-containing peptides on prolonged incubation of the histones with HOCl. In support of this, peptides containing Lys nitriles and Cl-Tyr/diCl-Tyr were observed on histones H1, H2B and H3 following 24 h treatment with HOCl (Table 1). In this study, there was no evidence for the formation of protein-bound, 2-aminoadipic acid, on reaction of protein Lys residues with HOCl, in contrast to other studies where these products were more abundant than Lys nitriles [35]. The reason for this is not certain but may reflect differences in the nature of the proteins examined, or the reactivity of HOCl, used here, compared to the MPO/H 2 O 2 /Cl − system, used in previous work [35]. Exposure of histones to HOCl resulted in the chlorination and dichlorination of Tyr but not Trp residues. This could reflect the relatively low abundance of Trp residues within the sequence of the different histones. Similarly, no evidence was obtained for the consistent formation of other Trp oxidation products, including oxindolyalanine (5hydroxytryptophan), dioxindolyalanine, N-formylkynurenine and kynurenine, which are formed on exposure of other proteins to HOCl [37][38][39]. The formation of Cl-Tyr and diCl-Tyr was seen at multiple sites on the histone H1 isoforms and each of the core histones H2A, H2B, H3 and H4, in accord with previous experiments with isolated histones [27]. In general, the extent of Tyr chlorination was significantly greater on treating the histones with HOCl for 24 h rather than 15 min. Some Tyr residues, including Tyr-71 on some H1 isoforms, Tyr-40 on H2B, Tyr-51 and Tyr-72 on H4, appeared to be particularly susceptible to chlorination, as evidenced by the detection of Cl-Tyr in experiments with low, equimolar, concentrations of HOCl.
The susceptibility of individual Tyr residues to chlorination seen here differs to previous studies where the extent of Tyr chlorination was quantitatively measured using LC-MS in experiments with the isolated core histones H2A, H2B, H3 and H4 [27]. Previously, preferential chlorination of Tyr residues was observed in sequence motifs where the Tyr was within 3 residues of a Lys or His residue, with Tyr-39 of H2A, Tyr-43, 83 and 121 of H2B, Tyr-41 of H3 and Tyr-88 of H4 found to be the most efficiently chlorinated residues [27]. This was postulated to be due to the initial reaction of HOCl with the more kinetically favoured Lys and His residues and the resulting chloramines facilitating chlorine transfer to Tyr [40,41]. Similarly, Tyr chlorination on reaction of HOCl with apolipoprotein A-I was favoured in the sequence motif YXXK or KXXY [42]. However, not all Tyr residues with the YXXK/KXXY motif were as readily chlorinated on the core histones, suggesting that the protein secondary structure also plays a role in determining selectivity [27]. In this study, of the readily modified residues, only Tyr-40 on H2B is located in a YXXK sequence motif, though we have evidence for chlorination of all the previously characterised histone chlorination sites at higher HOCl concentrations (5 and 10-fold molar excess). The reason for the difference in selectivity could be related to the use of a heterogeneous preparation of histones, which are not present in the same molar ratios, compared to experiments with each histone in isolation. It is possible that here, some Tyr chlorination could be facilitated by intermolecular chlorine transfer reactions [40].
Protein chloramines can also transfer chlorine and promote damage to other substrates, including DNA [43]. This could be relevant in the case of histones, given their close association with DNA both within the nucleus and in the extracellular environment, particularly on NETs and related structures. Whether histone chloramines can promote protein-DNA or protein-protein cross-links in NETs is unknown. However, previous studies have revealed that chlorinated polyamines play an important role in mediating the cross-linking of different NET proteins, including MPO and histones, which is believed to help stabilise the NET structure [44].
The reaction of histones with HOCl also resulted in protein fragmentation and aggregation, which was more extensive on 24 h compared to 15 min incubation, as evidenced by the loss in staining intensity of the some of the parent histone bands and detection of highmolecular-mass protein bands. Protein aggregation was only observed on exposure of the histones to > 5-fold HOCl for 24 h. Again, this may reflect secondary reactions mediated by the decomposition of protein chloramines. The aggregates are seen under reducing conditions, and some evidence was obtained for the formation of diTyr on exposure of Table 2 Nitrile and aminoadipidic semialdehyde formation on specific Lys residues seen on exposure of histones to different molar ratios of HOCl for 24 h unless indicated otherwise.  Figure S11). e Sulfoxide formation on Met (M) with nitrile and aminoadipic semialdehyde. f Cl 1 -Tyr formation on Tyr (Y) with nitrile only.
the histones to HOCl particularly at higher oxidant concentrations, which could contribute to the observed cross-linking. However, further experiments will be required to determine the precise nature of the cross-links and determine whether the aggregates are formed as a result of intra-or inter-molecular reactions. Extensive fragmentation of the histones was not observed under the conditions used here (≤ 10-fold HOCl). This is consistent with studies with other proteins, where higher molar excesses of HOCl are required to induce significant backbone cleavage [20,45,46]. The biological significance of the oxidative PTMs formed on the histones following exposure to HOCl has yet to be established. However, post-translational modification of Lys and Arg residues plays a critical role in regulating DNA replication, transcription and repair, so it could be postulated that exposure of histones to HOCl will alter function [4,5,7]. In particular, evidence was obtained for different Lys-derived PTMs, including nitriles and aminoadipic semialdehydes, which would alter the overall charge of the histones, particularly as these products are formed concurrently with the loss of Arg residues. In addition, the modification of Lys, Arg, Tyr and other residues was observed on exposure of the histones to low molar excesses of HOCl, which are readily achievable in vivo, particularly during chronic inflammation [47]. Studies are in progress to assess the extent and nature of HOCl-derived histone modifications on NETs released from human neutrophils. In addition, further studies will be necessary to fully assess the functional implications of HOCl-induced histone modification, including in the extracellular environment, such as following NETosis.
The mechanisms by which NETs promote disease are complex and incompletely understood [12]. Extracellular histones, which comprise ca. 70% of NET-associated proteins [8], are important in driving NET-derived host tissue damage [12,48,49]. Histones are highly cytotoxic, owing to their ability to disrupt cell membranes and are known to induce endothelial damage and dysfunction [2,48,49]. Recent studies have demonstrated that the damaging effects of histones can be modulated by PTMs, including citrullination, owing to the alteration of protein charge, which modulates electrostatic interactions [50]. Taken together, this suggests that the reaction of histones with HOCl and the resulting modifications of particularly Lys and Arg residues could have broad biological implications. This might be especially significant in the context of chronic inflammatory diseases, such as atherosclerosis, but could also be relevant for other acute NET-related pathologies, such as sepsis and COVID-19.

Declaration of competing interest
None.

Data availability
Data will be made available on request.