Adduction of the Chloroform Metabolite Phosgene to Lysine Residues

The aim of this study was to assess whether phosgene is able to form irreversible adducts with this peptide and to investigate which residues are most...
1 downloads 0 Views 290KB Size
266

Chem. Res. Toxicol. 2003, 16, 266-275

Articles Adduction of the Chloroform Metabolite Phosgene to Lysine Residues of Human Histone H2B Laura Fabrizi,*,† Graham W. Taylor,‡ Benito Can˜as,‡ Alan R. Boobis, and Robert J. Edwards Sections on Clinical Pharmacology and Proteomics, Division of Medicine, Faculty of Medicine, Imperial College, Hammersmith Campus, Du Cane Road, London W12 0NN, U.K. Received June 6, 2002

Human exposure to trihalomethanes such as chloroform has been associated with both cancer and reproductive toxicity. While there is little evidence for chloroform mutagenicity or DNA adduct formation, in vivo studies in rats have demonstrated adduction to histones and other nuclear proteins. Histones play a key role in controlling DNA expression particularly through the acetylation of lysine residues in their N-termini. Therefore, we studied the reaction of phosgene, the major active metabolite of chloroform, with the N-terminus of human histone H2B (Hpep, Pro-Glu-Pro-Ala-Lys-Ser-Ala-Pro-Ala-Pro-Lys-Lys-Gly-Ser-Lys-Lys-Ala-Val-ThrLys-Ala-Gln-Lys) in a model chemical system. The aim of this study was to assess whether phosgene is able to form irreversible adducts with this peptide and to investigate which residues are most susceptible. Hpep was reacted with a range of phosgene concentrations (0.03-36 mM) at 37 °C, pH 7.4. The products of these reactions, analyzed by matrix-assisted laser desorption ionization MS, showed that up to three CO moieties could be adducted to the peptide. The singly and doubly adducted peptides were purified by HPLC and then hydrolyzed with trypsin to produce a series of fragments that were analyzed by HPLC-MS. The tryptic products showed that adduction occurred principally at lysine residues, and that all seven lysine residues of the peptide were subject to adduction. Collision-induced dissociation analysis using ion trap MS-MS of the tryptic fragment [Pro-Glu-Pro-Ala-Lys-Ser-Ala-Pro-Ala-Pro-Lys + CO] and of the full-length singly adducted peptide supported the role of lysine residues in adduction; the data also indicated that the N-terminal proline and the serine residues are susceptible. Addition of glutathione to the reaction mixture only partially attenuated adduct formation and allowed production of another adducted species, i.e., Hpep-CO-glutathione. The occurrence of such reactions to the N-termini of histones, if confirmed by in vivo studies, could help to explain the mechanism of chloroform carcinogenicity.

Introduction Halogenated byproducts, such as chloroform (CHCl3)1 and other trihalomethanes (THMs), are formed during water chlorination for human consumption (1). Their production raises public health concerns mainly related to potential human cancer risk (2) but also to effects on * To whom correspondence should be addressed: Laboratorio di Alimenti, Istituto Superiore di Sanita`, Viale Regina Elena, 299, 00161 Rome, Italy. Telephone: +39 06 49902713. Fax: +39 06 49387101. E-mail: [email protected]. † Current address: Laboratorio di Alimenti, Istituto Superiore di Sanita`, Viale Regina Elena 299, 00161 Rome, Italy. ‡ Section on Proteomics. 1 Abbreviations: b , series of peptide fragments derived from the n N-terminus; CHCl3, chloroform; COCl2, phosgene; ESI, electrospray ionization; GSH, glutathione; GS-CO-Cl, glutathionyl carbonyl chloride; GS-CO-SG, diglutathionyl dithiocarbonate; Hpep, N-terminus of human histone H2B, PEPAKSAPAPKKGSKKAVTKAQK, Pro-Glu-ProAla-Lys-Ser-Ala-Pro-Ala-Pro-Lys-Lys-Gly-Ser-Lys-Lys-Ala-Val-ThrLys-Ala-Gln-Lys; MALDI, matrix-assisted laser desorption ionisation; P450, cytochrome P450; TEA-CO2, triethylamine carbonate; TFA, trifluoroacetic acid; THMs, trihalomethanes; yn, series of peptide fragments derived from the C-terminus.

reproduction and development (3). Humans can be exposed to CHCl3 either by water ingestion or by skin absorption and inhalation during baths (4), showers (5), and particularly, in swimming pools (6). CHCl3 is frequently detected at relatively high concentrations (7), i.e., 0.1 µg/L to 1 mg/L in finished drinking water (8-11) and in chlorinated swimming pool water (6). CHCl3 toxicity is mediated through cytochrome P450-dependent (P450) metabolism, which leads to the formation of phosgene (COCl2) (12-14) and the dichloromethyl radical (•CHCl2) (15, 16). Both of these metabolites can bind irreversibly to endogenous nucleophiles (12-14, 17, 18). Under physiological oxygenation conditions, COCl2 is the major metabolite formed and is considered to be responsible for the acute toxicity of CHCl3 (12-14); however, its role in carcinogenicity remains to be elucidated (19). Although COCl2 is rapidly hydrolyzed in water, its reaction with amino-, thiol-, and hydroxyl-containing compounds in an aqueous environment has been demonstrated previously (14, 20-23). Most COCl2 produced in vivo is hydrolyzed to CO2 or scavenged by glutathione

10.1021/tx025565i CCC: $25.00 © 2003 American Chemical Society Published on Web 02/04/2003

Phosgene Adducts to Lysines of Human Histone H2B

(GSH), which generally protects against CHCl3 toxicity (24). The stable product diglutathionyl dithiocarbonate (GS-CO-SG) has been isolated and characterized in chemical systems and in microsomal incubations (14). However, recent studies have suggested that GSH may also play a role in CHCl3-mediated carcinogenicity (25), and we have shown previously that an intermediate metabolite, glutathionyl carbonyl chloride (GS-CO-Cl), can bind to nucleophiles present in the cell (23). This may provide an alternative route for the formation of COCl2 adducts and lead to toxicological consequences. CHCl3 is generally considered to be a nongenotoxic carcinogen (2, 26), although this view is not universally accepted (19, 27), as the precise mechanism of its carcinogenicity has still to be determined. Interestingly, in a study where rats were treated with [14C]CHCl3, no evidence of adduction to DNA was found, while covalent binding to histones (and other nuclear proteins) was detected (28). It has been known for some time that histones are an essential structural component of nucleosomes, which are in turn the building blocks of chromatin. However, more recently, it has come to light that the highly flexible N-terminal regions of histones, which contain a high proportion of nonhydrophobic, basic amino acids that extrude out of the nucleosome core, play an important role in gene transcription as well as in chromatin assembly (29, 30). If chemical modifications could occur in this region, then they would likely have important toxicological consequences, possibly including carcinogenicity. It has been shown previously that the N-terminal 23 residues of histone H2B are readily cleaved from nucleosome particles by trypsin, hence this region of the protein is particularly accessible in the assembled nucleosome structure (31). Therefore, we chose to investigate this region as a possible target for histone modification following exposure to CHCl3. Here, we describe the reaction between the CHCl3 metabolite COCl2 and the N-terminus of human histone H2B (Hpep) in a model chemical system, as well as the effect of GSH on this reaction.

Experimental Procedures Materials. The peptide, Hpep (Pro-Glu-Pro-Ala-Lys-Ser-AlaPro-Ala-Pro-Lys-Lys-Gly-Ser-Lys-Lys-Ala-Val-Thr-Lys-Ala-GlnLys, i.e., PEPAKSAPAPKKGSKKAVTKAQK), which represents residues 1-23 of human histone H2B was synthesized using Boc solid-phase chemistry by Alta Bioscience (Birmingham, UK). Its composition was confirmed by amino acid analysis. COCl2 [20% (v/v) solution in toluene] was obtained from Fluka (Gillingham, U.K.). L-(Tosylamido-2-phenyl) ethyl chloromethyl ketone (TPCK)-treated bovine pancreas trypsin attached to agarose, triethylamine, R-cyano-4-hydroxycinnamic acid, GSH, formic acid, and trifluoroacetic acid (TFA) were purchased from Sigma (Poole, U.K.). Acetonitrile (Hypersolv, far UV grade) and methanol (Hypersolv) were purchased from Merck-BDH (Lutterworth, U.K.). Caution: phosgene is highly toxic and should be handled using appropriate safety measures (i.e. in fully functioning fumehood and wearing protective gloves). Reaction Conditions. Reactions between COCl2 (0.03-36 mM) and Hpep (0.25 mM) were carried out for 30 min at 37 °C in 50 mM triethylamine carbonate (TEA-CO2) buffer, pH 7.4 (final volume 200 µL); control reactions, where COCl2 was omitted, were also carried out. In some cases, GSH (0.03-3 mM) was also added to the reaction mixture. The TEA-CO2 buffer was freshly prepared by dissolving 50 mM TEA in water and adjusting the pH by addition of solid carbon dioxide. This buffer was chosen as it is volatile. Although no formal assessment was made of the reaction kinetics, qualitatively there was no obvious

Chem. Res. Toxicol., Vol. 16, No. 3, 2003 267 difference in the amount of adducted product produced with this compared with other buffers. Reactions were initiated by addition of COCl2 as a 20% (v/v) solution in toluene, or diluted in acetonitrile as necessary. After 30 min, samples were immediately cooled on ice, snap frozen at -20 °C and freezedried to simultaneously remove water and any residual COCl2. The dried residues were resuspended in 10-50 µL of acetonitrile:water (1:1) containing 0.1% (v/v) TFA, and analyzed by matrix-assisted laser desorption ionization (MALDI)-MS on a Micromass MALDI Instrument, using R-cyano-4-hydroxycinnamic acid as a matrix. This instrument is highly resolving with each peak appearing as an isotopic distribution representing the relative natural abundance of 13C in the peptides. Here, product masses were quantified using peaks corresponding to 12C ions. HPLC Purification. Reaction products were purified by HPLC. The system consisted of a Hewlett-Packard HP 1090 Liquid Chromatograph, in combination with a µBondapack C18 Waters semipreparative column (0.8 × 30 cm). The column was equilibriated in 7% (v/v) acetonitrile containing 0.1% TFA, and after sample loading, a linear gradient of 7-27% acetonitrile was applied over 24 min at a flow rate 2 mL/min. The eluate was monitored at 210 nm, and 2 mL fractions were collected throughout the chromatographic separation. Fractions containing [Hpep + CO] or [Hpep + 2 CO] were identified by MALDIMS and pooled separately. Tryptic Digestion. Peptides purified by HPLC (approximately 15 nmol) were digested with 0.4 units of immobilized trypsin (0.02 mL of agarose-trypsin; activity 20 units/mL) for 30 min at 37 °C in 50 mM TEA-CO2 buffer, pH 7.4 (final volume 1 mL). Under these conditions, trypsin was expected to hydrolyze peptide bonds on the C-terminal side of each lysine in Hpep, to produce a series of fragments that could be used to identify the position(s) of COCl2 adduction. At the end of the reaction, samples were centrifuged at 13000g for 1 min and the supernatants freeze-dried. The dried residues were resuspended in 7% (v/v) acetonitrile containing 0.04% (v/v) TFA (10-50 µL) and analyzed by HPLC-electrospray ionization (ESI)-MS using a Phenomenex Aqua C18 column (2 mm × 15 cm, packing size 5 µm, pore size 125 Å) on a Jasco modular HPLC. The column was equilibriated in water containing 0.04% (v/v) TFA and, after sample loading (5-25 µL), a linear gradient of 0-60% acetonitrile containing 0.04% TFA was applied over 60 min at a flow rate of 0.1 mL/min. The eluate was monitored at 210 nm using a MD 1510 diode array detector. ESI-MS analysis was carried out on-line using a VG (Micromass) Quattro II triple quadrupole MS. Where needed for further analysis, the eluate was collected in 0.2 mL fractions throughout the chromatographic separation. Analysis of Adducted Peptides by MS-MS. Collisioninduced dissociation analysis of HPLC purified [Hpep + CO] and HPLC purified tryptic fragments of [Hpep + CO] was performed by MS-MS using a Finnigan LCQ Deca Ion Trap MS. For this analysis, samples were freeze-dried and resuspended in 10 µL methanol:water (4:1) containing 0.05% (v/v) formic acid and applied directly into the instrument as a nanospray.

Results Formation of COCl2 Adducts with Hpep. A series of reactions was performed in which 0.25 mM Hpep was reacted with a range of concentrations of COCl2 (0.0336 mM). The products of these reactions were analyzed by MALDI-MS. In the absence of COCl2, the major [M + H]+ peak detected with an m/z of 2348 corresponded to Hpep (Figure 1a). Concentrations of less that 0.1 mM COCl2 produced no detectable effect on the spectrum of the peptide. However, with 0.1 mM COCl2 an additional peak with an m/z corresponding to Hpep plus 26 was detected (Figure 1b), and with 0.3 mM COCl2, ions corresponding to Hpep plus 26, 52, and 78 were found (Figure 1c). These ions indicate adduction to the peptide

268

Chem. Res. Toxicol., Vol. 16, No. 3, 2003

Fabrizi et al.

Figure 2. HPLC purification of adducted and nonadducted Hpep. A typical HPLC chromatogram showing the separation of adducted and nonadducted Hpep. In this case, 36 mM COCl2 was reacted with 0.25 mM Hpep and subjected to preparative HPLC under the conditions described in the Experimental Procedures. Peaks labeled 0, 1, and 2 indicate the elution of unadducted Hpep, [Hpep + CO], and [Hpep + 2CO], respectively.

Figure 1. MALDI-MS spectra showing the production of Hpep adducts at different COCl2 concentrations. A series of reactions were performed in which varying amounts of COCl2 were reacted with 0.25 mM Hpep and the products analyzed by MALDI-MS. The labels for peaks 0, 1, 2 and 3 indicate the number of CO adducts on Hpep, i.e., unadducted Hpep, [Hpep + CO], [Hpep + 2CO] and [Hpep + 3CO], respectively. Details of the reaction conditions, sample preparation and analysis are described in the Experimental Procedures. In the control reaction (a), no COCl2 was added to the mixture. In the reactions performed with (b) 0.1 mM and (c) 0.3 mM COCl2, COCl2 was diluted in acetonitrile before addition, and this solvent comprised 5% (v/v) of the final reaction mixture. In the reaction with (d) 36 mM COCl2, COCl2 was added to the mixture in toluene [final concentration 2% (v/v)].

of 1 × CO, 2 × CO, and 3 × CO groups, respectively. It should be noted that in MALDI-MS relative ion intensities cannot be compared between ions, hence this techniques does not reveal which adduct is the major product in each of the reactions. In reactions with higher concentrations of COCl2, no further products were clearly identified (Figure 1d). Although there was a suggestion of a small peak at around Hpep plus 104 mass units, which could have been due to [Hpep + 4 CO], this could not be resolved from the background noise (Figure 1, panels c and d). Several small additional peaks were also seen in the spectra derived from control and reaction mixtures. A peak with an m/z corresponding to Hpep + 56 was present in the Hpep preparation. This is probably a minor impurity formed as a byproduct during peptide synthesis, possibly due to addition of a tert-butyl group, which is used to protect Ser and Thr residues. In further studies described below, reaction mixtures were purified by HPLC and this also removed the impurity. The other minor peaks present in the spectra were either [M + Na]+ or [M + K]+ species of the major products. No products

corresponding to cross-linked dimeric Hpep or larger Hpep complexes were observed. Localization of COCl2 Adduction on Hpep. The products of the reaction between 36 mM COCl2 and 0.25 mM Hpep were analyzed and purified by HPLC (Figure 2). The analysis showed that under these conditions both [Hpep + CO] and [Hpep + 2CO] were produced and that the [Hpep + CO] adduct appeared to be the major product. No peak corresponding to [Hpep + 3CO] was identified. Purified [Hpep + CO] and [Hpep + 2CO] as well as unadducted Hpep were then digested with trypsin. The digests of the three products were analyzed by HPLC-MS (Figure 3, Table 1) in order to identify the fragments that had been adducted with COCl2. The chromatogram of the Hpep digest (Figure 3a) comprised six UV-absorbing peaks, and based on the expected trypsin cleavage pattern (occurring C-terminal to each lysine residue), these contained ions corresponding to all of the predicted fragments of Hpep (Table 1). In addition, KGSK and KAVTK were found. These fragments occurred as a result of trypsin hydrolysis between the two adjacent lysine residues in Hpep at positions 12 and 13 and 15 and 16. They were also present in 2 h digests, suggesting that the further hydrolysis of peptides by trypsin to GSK and AVTK occurred only very slowly. This is typical for peptides with an N-terminal lysine and a free R-amino group (32). The chromatograms of the tryptic digests of [Hpep + CO] (Figure 3b) and [Hpep + 2CO] (Figure 3c) showed several additional peaks with increased retention times. MS analysis of the peaks produced from [Hpep + CO] showed that all of the fragments identified in the Hpep digest were also present in this digest and, in addition, a number of adducted peptides were also detected (Table 1). Most of the adducted peptides were larger than the nonadducted peptides due to incomplete trypsin digestion of the peptide, presumably as a result of CO adduction to various lysine residues and subsequent inhibition of hydrolysis by trypsin at these sites (Table 1). Digestion of [Hpep + 2CO] also produced most of the fragments obtained in the Hpep digest, except for the peptides GSK, KGSK, KAVTK, which were absent (Table 1). A number of larger singly adducted peptides were detected. All of these adducted peptides were also detected in the [Hpep

Phosgene Adducts to Lysines of Human Histone H2B

Chem. Res. Toxicol., Vol. 16, No. 3, 2003 269

Table 1. Identification of Peptide Fragments Resulting from Trypsin Digestion of Hpep, [Hpep + CO], and [Hpep + 2CO]a [Hpep + CO]

[Hpep + 2 CO]

peak

Hpep

1 2 3 4 5

K (146.8b), GSK (291.3b), KGSK (418.8,b 210.0c)d AQK (345.9,b 173.5c) AVTK (418.0,b 209.5c) KAVTK (546.0,b 273.6c) PEPAK (540.9,b 270.9c)

K (147b), GSK (290.7b), KGSK (419.0,b 209.9c)d, [PEPAK + CO] (566.5b) AQK (345.9b) AVTK (418.0,b 209.4c) KAVTK (546.0,b 273.6c) PEPAK (540.9,b 271.1c), KAVTK + CO (571.6b)

6 7 8

SAPAPK (569.9,b 285.6c) no peak no peak

SAPAPK (570.0b, 285.7c) [AQK + CO] (371.7)b [GSKKAVTK + CO] (422.2c), [KAVTKAQK + CO] (450.2),c [KGSKKAVTK + CO] (486.8c) [PEPAKSAPAPKK + CO] (623.7c) [PEPAKSAPAPK + CO] (559.8c)

9 10

no peak no peak

K (147.1b), [PEPAK + CO] (566.7b) AQK (346.9b) AVTK (418.0b; 209.3c) none identified PEPAK (541.0,b 271.1c), [KAVTK + CO] (572.8,b 287.1c) SAPAPK (570.1,b 285.6c) none detected [GSKKAVTK + CO] (422.3c), [KAVTKAQK + CO] (450.1c) none identified [PEPAKSAPAPK + CO] (560.0c)

a Preparations of adducted and nonadducted Hpep were subjected to trypsin digestion and then separation by HPLC. The peak numbers correspond to those indicated in Figure 3. Each fragment was identified by its m/z ratio as determined by ESI-MS. b The mass of each tryptic peptide is shown in parentheses and corresponds to [M + H]+ ion. c The mass of each tryptic peptide is shown in parentheses and corresponds to [M + 2H]2+ ions. d Both KGSK and GSKK have the same mass, therefore, ion trap MS-MS was used to confirm the structure of this fragment as KGSK.

Table 2. Suggested Structures of Adduction within [Hpep + CO]

Figure 3. HPLC separation of tryptic digests of adducted and nonadducted Hpep. The HPLC purified peptides (a) Hpep(PEPAK∧SAPAPK∧K∧GSK∧K∧AVTK∧AQK), (b) [Hpep + CO], and (c) [Hpep + 2CO], derived from the reaction of 36 mM COCl2 with 0.25 mM Hpep, were digested with trypsin (∧ ) trypsin cleavage sites) and then analyzed by on-line ESI-MS to determine the identity of peptides in each of the peaks. For each, the UV trace (210 nm) is shown. Numbers refer to the 10 major peaks detected in the three chromatograms. The identity of the peptides found in each of the peaks are indicated in Table 1.

+ CO] digest, although some that were detected in the [Hpep + CO] digest were absent in the [Hpep + 2CO] digest (Table 1). Presumably, this was due to further modification of Hpep by additional adduction. The UV HPLC trace indicates an increase in late-running peaks, such as peak 9 and other minor peaks occurring between peaks 9 and 10 and after peak 10; however, no adducted species could be identified in these peaks, possibly due to the complexity of the species formed. Altogether, these data are consistent with adduction to lysine residues at positions 5, 11, 12, 15, 16, and 20. In addition, in the digestion mixture of [Hpep + CO], the fragment [AQK + CO] was found, suggesting adduction to the C-terminal residue, also a lysine (position 23). Furthermore, in both the digestion mixtures of [Hpep + CO] and of [Hpep +

2CO], adduction to the N-terminal fragment [PEPAK + CO] was found (Table 1). As trypsin was able to cleave this peptide from Hpep, adduction does not appear to involve a lysine residue in this case. On the basis of the structures of the adducted and nonadducted peptides, identified in the tryptic digests of Hpep compared with that of [Hpep + CO], it is possible to postulate where COCl2 adduction occurred (Table 2).

270

Chem. Res. Toxicol., Vol. 16, No. 3, 2003

Fabrizi et al.

Table 3. Analysis of the Singly-Charged Ion [PEPAKSAPAPK + CO + H]+ by Collision-Induced Dissociation MSa

a [PEPAKSAPAPK + CO], resulting from the tryptic digestion of [Hpep + CO], was purified by HPLC (Figure 3) and then subjected to ion trap MS-MS where the [PEPAKSAPAPK + CO + H]+ ion was fragmented. Adducted and nonadducted peptide fragments were identified from their m/z ratios. Peptides indicated by lowercase letters refer to putative fragments that were not detected in the spectrum (in particular, peptides of low mass, in this case b1-b3 and y1-y3, are especially difficult to discern from the background). Peptides shown in uppercase letters correspond to fragments identified in the spectrum and those that are underlined correspond to peptides adducted by CO. The suggested site of CO adduction is indicated in bold. b Some products were detected only as dehydrated species.

It can be seen that most of the products can be explained by a bridge formed by COCl2 between two nucleophiles (structures 1-5). All of these products involve one or more lysine residues and/or the N-terminal proline. However, in three cases, where again adduction to lysine residues and the N-terminal proline is implicated (structures 6-8), the information obtained did not allow identification of the second putative nucleophile. Collision-Induced Dissociation of the Tryptic Digestion Fragment [PEPAKSAPAPK + CO]. Among the adducted peptide fragments resulting from digestion of [Hpep + CO] and [Hpep + 2CO], the peptide [PEPAKSAPAPK + CO] was prominent and apparently relatively pure (peak 10, Figure 3, Table 1). Therefore, an HPLC-purified preparation of [PEPAKSAPAPK + CO] was analyzed further using ion trap MS-MS. Atypically for a peptide of its mass and composition, which would normally be expected to produce doubly charged ions only, [PEPAKSAPAPK + CO] was detected as both singly and doubly charged ions, with the singly charged ion being more intense. Consequently, these two ionic species were analyzed separately by ion trap MS-MS following helium collision-induced dissociation. Under the conditions used, the induced dissociation was limited so that cleavage occurred principally at a single amide bond in the peptide. However, it was evident from examples where there was a CO bridge across the position of peptide dissociation that the energy introduced was sufficient to simultaneously cleave one of the two bonds linking CO to the peptide. Consequently, it appears that in such cases the CO group is left attached to either of the corresponding bn or yn ions (e.g., the appearance of the unmodified b3 and b4 ions, and of the adducted y7 and y9 ions as shown in Table 3, and of both the unmodified and the adducted y6 ions as shown in Table 4). Analysis of the resultant mixture of cleaved products allowed identification of series of peptide fragments derived from both the N-terminus (bn series) and the C-terminus (yn series) of [PEPAKSAPAPK + CO] and

from this analysis, attempts were made to deduce the position of CO adduction. The dissociation spectrum of the [PEPAKSAPAPK + CO + H]+ ion revealed that in the bn series, b3 and b4 were the only nonadducted peptide fragments found (Table 3). The other larger fragments in this series (b5 b11) were all found to be adducted to CO, with the exception of b10, which was not found in the spectrum either as adducted or as nonadducted species. These data suggest that adduction occurred on the central lysine of PEPAKSAPAPK (i.e., at position 5). In the yn series, y4 to y6 were not adducted, while y7, y9, and y11 were adducted (Table 3). These data also indicate that adduction occurred on the lysine at position 5. The evidence that the fragment [PEPAKSAPAPK + CO] is in this case singly ionized suggests that its N-terminal amine is adducted. Therefore, [PEPAKSAPAPK + CO + H]+ is likely to contain a bridge between the lysine residue at position 5 and the N-terminal proline. The dissociation spectrum of the [PEPAKSAPAPK + CO + 2H]2+ ion showed that in the bn series, both b3 and b4 were found as nonadducted species, while all fragments from b6 to b11 were adducted, with the exceptions of b5 and b10 which were not found in the spectrum either as adducted or as nonadducted species (Table 4). These data suggest that adduction occurred either at the central lysine residue at position 5 and/or at the adjacent serine residue at position 6. In the yn series, y2, y5, and y6 were found as nonadducted species, while all fragments from y6 to y11 were found adducted to CO, with the exception of y8, which was not detected. Notably, y6 was found as both adducted and nonadducted species. These data also indicate the involvement of the lysine and serine residues in CO adduction. It seems probable that [PEPAKSAPAPK + CO + 2H]2+ contains a bridge where CO links the lysine residue at position 5 with the serine residue. Collision-Induced Dissociation of Adducted Hpep. Ion trap MS-MS was also applied to the analysis of the singly adducted full-length peptide [Hpep + CO], pro-

Phosgene Adducts to Lysines of Human Histone H2B

Chem. Res. Toxicol., Vol. 16, No. 3, 2003 271

Table 4. Analysis of the Doubly-Charged Ion [PEPAKSAPAPK + CO + 2H]2+ by Collision-Induced Dissociation MSa

a [PEPAKSAPAPK + CO], resulting from the tryptic digestion of [Hpep + CO], was purified by HPLC (Figure 3) and then subjected to ion trap MS-MS where the [PEPAKSAPAPK + CO + 2H]2+ ion was fragmented. Details of the presentation of the data are given in the legend to Table 3. b Some products were detected only as [M + CO + 2H]2+ ions. c Some products were detected only as dehydrated species.

Table 5. Analysis of the Doubly-Charged Ion [Hpep + CO + 2H]2+ by Collision-Induced Dissociation MSa

a The adducted peptide [Hpep + CO] resulting from the reaction between 0.1 mM COCl and 0.25 mM Hpep was purified by HPLC 2 (Figure 2) and then subjected to ion trap MS-MS where the [Hpep + CO + 2H]2+ ion was analyzed. Details of the presentation of the data b 2+ are given in the legend to Table 3. Some products were detected only as [M + CO + 2H] ions. The suggested predominant region of adduction is indicated in bold.

duced in the reaction of Hpep with 0.1 mM COCl2. Its dissociation spectrum showed a complex pattern with many of the cleaved peptides occurring as both adducted

and nonadducted species (Table 5). This suggests that there are multiple adduction sites on the peptide. Nevertheless, in the bn series, a fairly complete set of

272

Chem. Res. Toxicol., Vol. 16, No. 3, 2003

Fabrizi et al.

Figure 4. MALDI-MS spectra showing the effect of GSH on COCl2 adduction to Hpep. A series of reactions were performed in which varying concentrations of GSH were added to reaction mixtures containing 0.3 mM COCl2 and 0.25 mM Hpep. Details of the reaction conditions, sample preparation and analysis are described in the Experimental Procedures. The products of (a) the control reaction mixture (without GSH) and of reaction mixtures containing (b) 0.3 mM and (c) 3 mM GSH were analyzed by MALDI-MS. Peaks labeled 0, 1, 2, 3, 0 + COSG, 1 + COSG, and 2 + COSG indicate Hpep, [Hpep + CO], [Hpep + 2CO], [Hpep + 3CO], [Hpep + COSG], [Hpep + CO + COSG], and [Hpep + 2CO + COSG], respectively.

adducted peptide fragments was found from b12 to b22 (except for b13) and in the yn series, a complete set of adducted fragments from y7 to y21 was found. These data suggest that the region common to both series, KGSKKA, is a predominant site of CO adduction. Effect of GSH on the Reaction between COCl2 and Hpep. The effect of GSH on adduct formation was assessed in a series of reactions using increasing concentrations of GSH. Reaction of 0.3 mM COCl2 with 0.25 mM Hpep produced mono-, di-, and tri-adducted peptide species (Figure 4a) and addition of 0.03 mM GSH (data not shown) or 0.3 mM GSH (Figure 4b) had little or no effect on adduct formation. However, with 3 mM GSH, adduction was clearly attenuated, as neither di- or triadducted Hpep was detectable, although the monoadducted peptide was still found (Figure 4c). Interestingly, in the presence of 0.3 mM GSH, two new species were found, i.e., [Hpep + CO + CO-SG] and [Hpep + 2CO + CO-SG] (Figure 4b). With 3 mM GSH, the species [Hpep + CO + CO-SG] was also found as well as another species, [Hpep + CO-SG] (Figure 4c). Therefore, GSH was only able to partially attenuate CO adduction to Hpep when at a concentration in excess of that of COCl2 and its presence also produced additional adduct species. However, the ion intensities recorded for these species were extremely low (Figure 4).

Discussion We have investigated the reaction between the CHCl3/ THM metabolite COCl2 and the peptide Hpep, which represents the N-terminus of histone H2B, in a model chemical system. The results of the study indicate that COCl2 readily reacts with Hpep causing the addition of up to three CO moieties to the peptide. Further, it was found that lysine residues are the most susceptible sites of adduction within this peptide. This inference is supported by several findings. First, the majority of the fragments found in the tryptic digests of [Hpep + CO] and [Hpep + 2CO] were larger than those produced from the digestion of unadducted Hpep. The masses of the fragments indicate not only adduction of CO but also

inhibition of cleavage by trypsin at lysine residues shown to be cleavable in unadducted Hpep. This suggests that the lysine residues were modified by CO adduction. Second, the collision-induced dissociation spectra of both the singly and doubly charged ions of the tryptic digest fragment [PEPAKSAPAPK + CO] showed that CO was adducted to the lysine residue at position 5. Third, the collision-induced dissociation spectrum of the whole singly adducted peptide [Hpep + CO] indicated that the region of Hpep comprising amino acids 11-17 (KKGSKKA), which is rich in lysine residues, was a predominant area of adduction. It would appear that all seven lysine residues of Hpep are able to react with COCl2. Indeed, from an examination of the various tryptic fragments produced in the digest mixtures of [Hpep + CO] and [Hpep + 2CO], it can be seen that each of the larger fragments is the result of inhibition of trypsin digestion at one or more sites involving lysine residues. Therefore, all six of the internal lysine residues of Hpep bind CO. In addition, the Cterminal lysine at position 23 also appears to be adducted, as the digest fragment [AQK + CO] was found. Even when using the lowest amount of COCl2 possible to produce detectable adduction, several lysine residues on Hpep were still found to be modified. This suggests that CO binding occurs randomly on lysine residues and, therefore, that the HPLC-purified preparations of [Hpep + CO] and [Hpep + 2CO] were actually mixtures of different singly and doubly adducted regioisomers, respectively. These data are consistent with the fact that the side chains of lysine residues contain nucleophilic amines and, except for its N-terminal proline, Hpep has no other strong nucleophiles. It is known that the N-terminal regions of histone proteins contain a high proportion of nonhydrophobic, basic amino acids. Consequently, these regions are highly flexible and are able to extrude away from the histone core (29, 33). This property is likely to afford a high level of accessibility to all of the lysine residues in the N-termini and, thus, allow ready reactivity with small electrophiles such as COCl2. Here, the

Phosgene Adducts to Lysines of Human Histone H2B

Chem. Res. Toxicol., Vol. 16, No. 3, 2003 273

Scheme 1. Reaction Pathways from CHCl3 to COCl2 Adducts Formed with Hpepa

a All of the reactants and products were either identified here or have been found in previous studies, except for those in parentheses, which are theoretical intermediates, which have not been isolated. The two adducted species of Hpep shown arise independently. Therefore, it is possible to have mixtures of these adducts occurring on a single Hpep molecule.

absence of any specificity for particular lysine residues indicates that they all have similar nucleophilicity and none appear hindered. COCl2 is a bifunctional reactive electrophile. In the presence of nucleophiles, it generally loses both its chlorine atoms, producing a CO bridge between the two nucleophiles. If the nucleophile is a primary amine, COCl2 can also form an isocyanate, which is rather unstable in water (34), although it can further react with other nucleophiles present giving rise to CO bridges. Our results show that the adducted masses corresponding to [Hpep + CO], [Hpep + 2CO] and [Hpep + 3CO] are consistent with intramolecular bridges formed between two nucleophiles within Hpep. No evidence of intermolecular bridges was observed, although this was not surprising given the relatively dilute solutions of Hpep used. Therefore, it is envisaged that the structure most likely to be formed is adduction of CO to form an intramolecular bridge between two lysine residues. Indeed, in a previous study (22), where [14C]phosgene was incubated with human blood, it was found that COCl2 can form a cyclic adduct within albumin, bridging two lysine residues which occurred at positions 195 and 199. In Hpep, a similar type of adduction was found between lysine residues at positions 5-11, 12-15, 15-16, and 16-20. These data also indicate that, as well as lysine residues, other amino acids participate in bridge formation. Evidence for the involvement of the N-terminal proline in CO adduction comes from the collision-induced dissociation of [PEPAKSAPAPK + CO + H]+, where it appears that the adduct forms a bridge between the N-terminal proline and the lysine residue at position 5. The serine residue at position 6 may also be subject to CO adduction, as is suggested by the collision-induced dissociation spectrum of [PEPAKSAPAPK + CO + 2H]2+, where it appears that the serine residue is bridged by CO to the neighboring lysine residue. Moreover, it is also possible that bridges between lysine residues and peptide

bonds (amidic nitrogens) may be formed, although no direct evidence was produced for them in this study. Certainly, in a previous study (35), where CHCl3 was incubated with microsomes and hemoglobin, COCl2 was shown to form a cyclic adduct bridging a cysteine and the adjacent peptide bond in the sequence. When [14C]phosgene was incubated with human blood, a cyclic adduct between the N-terminus of hemoglobin and the adjacent peptide bond in the sequence was identified (22). Hpep does not contain any cysteine residues and so no evidence for adduction via this amino acid was produced here. In fact, histone proteins are completely devoid of cysteine residues and so, unlike other proteins (22, 23), are not subject to adduction in this way. For Hpep, in one case, adduction appeared to occur independently of lysines. This is suggested by the appearance of the N-terminal fragment [PEPAK + CO] derived from the tryptic digests of both [Hpep + CO] and [Hpep + 2CO]. Adduction of CO to this fragment appears to occur at residues other than lysine, as it is the product of cleavage between the apparently unmodified lysine at position 5 and the serine residue at position 6. If lysine is not involved in the adduction, then it would appear likely that the N-terminal proline, being the other strong nucleophile in PEPAK, may be the site of CO adduction. However, no direct evidence for this was produced. To consider the effect of possible scavenging on the reaction between Hpep and COCl2 that might occur under more physiological conditions, we added GSH to the reaction mixture. Although GSH was able to partially attenuate the extent of the reaction, even levels of GSH comparable to those found in cell were insufficient to completely inhibit CO adduction to Hpep. Moreover, additional larger moieties with masses corresponding to addition of CO-SG groups to the peptide were found. It would be of considerable interest to more fully characterize these species, although the low ion intensities observed in the spectra means that the current techniques would be difficult to apply and rather suggests that the

274

Chem. Res. Toxicol., Vol. 16, No. 3, 2003

products may require some purification and concentration before such studies can be performed. Nevertheless, their appearance suggests that COCl2 is probably able to react with GSH to produce the reactive intermediate metabolite GS-CO-Cl, which may then bind to Hpep, unless it first reacts with a second GSH molecule to form GS-CO-SG, as has been shown previously in reactions between COCl2 and GSH (14, 23). Alternatively, an intermediate isocyanate product formed between COCl2 and Hpep could react with GSH, leading to an irreversible adduct to Hpep, which includes the -SG moiety (Scheme 1). Among the known posttranslational modifications of histones (36), it is the acetylation and deacetylation of lysine residues contained in the N-terminal regions and their effects on normal gene transcription and chromatin assembly that has been studied most extensively (29, 30). When these lysines are acetylated, the overall positive charge of the histones is reduced and, consequently, so is their interaction with negatively charged DNA. This permits a partial dissociation of DNA from the nucleosome and allows an increase in normal gene transcription and, subsequently, cell growth to occur. Conversely, deacetylation of lysine residues allows reformation of the nucleosome to a closed conformation, reducing gene transcription and avoiding uncontrolled DNA expression. Histone acetylation is essential to cell function (37-40) and the level is established and maintained by multiple histone acetyltransferases and deacetylases (38). Accumulating evidence suggests that deregulation of the activities of histone acetylases and deacetylases plays a causative role in the generation of cancer (41-44). Modification of histone acetylation or the lysine targets for this activity has also been shown to result in adverse developmental outcome (45-48). As our data indicate adduction of carbonyl groups onto Hpep, it is tempting to suggest that COCl2 might mimic the effect of acetylation, in so much as it eliminates the positive charge of the lysine residues. Such an effect might be expected to cause an increase in gene expression, which may eventually lead to proliferation, a known effect of CHCl3 exposure (49-51). Although we have been able to demonstrate the potential of COCl2 to adduct to functionally critical lysine residues, this now need to be demonstrated in more biologically relevant models, such as hepatocytes, liver slices or indeed whole animals. Certainly, it has been shown in the liver of rats treated in vivo with [14C]CHCl3 that histones and other nuclear proteins become radiolabeled (28), and this suggests that COCl2 (produced in the endoplasmic reticulum) is able to cross the nuclear membrane and reach the histones in the nucleus. In conclusion, our results show that, in a model chemical system, COCl2 can bind to the N-terminus of histone H2B, principally through lysine residues. GSH can only partially attenuate this reaction, and at the same time its presence allows the production of other minor adducted species. Our data suggest that the simple reaction system employed produced a multitude of regioisomer products, and consequently, it was not possible to separate them completely and to characterize all of their structures. However, regardless of their precise structure, the involvement of functionally essential lysine residues, if confirmed by in vivo studies, could have major implications for explaining the mechanism of CHCl3 carcinogenicity and developmental toxicity.

Fabrizi et al.

Acknowledgment. The research was supported in part by the Italian Foundation for the Research on Cancer (F.I.R.C.) and by the Department of Health and Food Standards Agency. The conclusions/views stated represent those of the authors and not the policy of the Department of Health or the Food Standards Agency.

References (1) Bunn, W. W., Haas, B. B., Deane, E. R., and Kleopfer, R. D. (1975) Formation of trihalomethanes by chlorination of surface water. Environ. Lett. 10, 205-213. (2) Fawell, J. (2000) Risk assessment case studyschloroform and related substances. Food Chem. Toxicol. 38 (1 Suppl.), S91-S95. (3) Nieuwenhuijsen M. J., Toledano M. B., Eaton N. E., Fawell J., and Elliot P. (2000) Chlorination disinfection byproducts in water and their association with adverse reproductive outcomes: a review. Occup. Environ. Med. 57, 73-85. (4) Corley, R. A., Gordon, S. M., and Wallace, L. A. (2000) Physiologically based pharmacokinetic modeling of the temperaturedependent dermal absorption of chloroform by humans following bath water exposure. Toxicol. Sci. 53, 13-23. (5) Kuo, H. K., Chiang, T. F., Lo, I. I., Lai, J. S., Chan, C. C., and Wang, J. D. (1998) Estimates of cancer risk from chloroform exposure during showering in Taiwan. Sci. Tot. Environ. 218, 1-7. (6) Lahl, U., Batjer, K., Duszeln, J. V., Gabel, B., and Thiermann, W. (1981) Distribution and balance of volatile halogenated hydrocarbons in the water and air of covered swimming pools using chlorine for water disinfection. Water Res. 15, 803-814. (7) IARC (1991) Chlorinated drinking water; chlorination byproducts; some other halogenated compounds; cobalt and cobalt compounds. IARC monographs on the evaluation of carcinogenic risks to humans, Lyon, France. (8) Uden, P. C., and Miller, J. W. (1983) Chlorinated acids and chloral in drinking water. J. Am. Water Works Assoc. 75, 524-526. (9) Symons J. M., Bellar, T. A., Carswell, J. K., De Marco, J., Kropp, K. L., Robeck, G. G., Seeger, D. R. Slocum, C. J., Smith, B. L., and Stevens, A. A. (1975) National organics reconnaissance survey for halogenated organics. J. Am. Water Works Assoc. 67, 634647. (10) Bellar, T. A., Lichtenberg, J. J., and Kroner, R. C. (1974) The occurrence of organohalides in chlorinated drinking water. J. Am. Water Works Assoc. 66, 703-706. (11) Rook, J. J. (1974) Formation of haloforms during chlorination of natural water. Water Treat. Exam. 23, 234-243. (12) Pohl L. R., Bhooshan, B., Whittaker, N. F., and Krishna, G. (1977) Phosgene: a metabolite of chloroform. Biochem. Biophys. Res. Commun. 79, 648-691. (13) Pohl L. R., Martin, J. L., and George, J. W. (1980) Mechanism of metabolic activation of chloroform by rat liver microsomes. Biochem. Pharmacol. 29, 3271-3276. (14) Pohl, L. R., Branchflower, R. V., Highet, R. J., Martin, J. L., Nunn, D. S., Monks, T. J., George, J. W., and Hinson, J. A. (1981) The formation of diglutathionyl dithiocarbonate as a metabolite of chloroform, bromotrichloromethane, and carbon tetrachloride. Drug Metab. Dispos. 9, 334-339. (15) Tomasi, A., Albano, E., Biasi, F., Slater, T. F., Vannini, V., and Dianzani, M. U. (1985) Activation of chloroform and related trihalomethanes to free radical intermediates in isolated hepatocites and in the rat in vivo as detected by the ESR-spin trapping technique. Chem. Biol. Interact. 55, 303-316. (16) Testai, E., Di Marzio, S., Di Domenico, A., and Piccardi, A. (1995) An in vitro investigation of the reductive metabolism of chloroform. Arch. Toxicol. 70, 83-88. (17) De Biasi, A., Sbraccia, M., Keizer, J., Testai, E., and Vittozzi, V. (1992) The regioselective binding of CHCl3 reactive intermediates to microsomal pospholipids. Chem. Biol. Interact. 85, 229-242. (18) Gemma, S., Ade, P., Sbraccia, M., Testai, E., and Vittozzi, L. (1996) In vitro quantitative determination of phospholipid adducts of chloroform intermediates in hepatic and renal microsomes from different rodent strains. Environm. Toxicol. Pharm. 2, 233242. (19) Vittozzi, L., Gemma, S., Sbraccia, M., and Testai, E. (2000) Comparative characterization of CHCl3 metabolism and toxicokinetics in rodent strains differently susceptible to chloroforminduced carcinogenicity. Environ. Toxicol. Pharmacol. 8, 103110. (20) Babad, H., and Zeiler, A. G. (1973). The chemistry of phosgene. Chem. Rev. 73, 75-91.

Phosgene Adducts to Lysines of Human Histone H2B (21) Gyllenhaal, O., and Vessman, J. (1988) Phosgene as a derivatizing reagent prior to gas and liquid chromatography. J. Chromatogr. 435, 259-269. (22) Noort, D., Hulst, A. G., Fidder, A., van Gurp, R. A., de Jong, L. P., and Benschop, H. P. (2000) In vitro adduct formation of phosgene with albumin and hemoglobin in human blood. Chem. Res. Toxicol. 13, 719-726. (23) Fabrizi, L., Edwards, R. J., Taylor, G. W., and Boobis, A. R. (2001) Adducts of the Chloroform Metabolite Phosgene. In Advances in Experimental Medicine and Biology, volume 500. Biological Reactive Intermediates VI. Chemical and Biological Mechanisms in Susceptibility to and Prevention of Environmental Diseases (Dansette, P. M., Snyder, R., Delaforge, M., Gibson, G. G., Greim, H., Jollow, D. J., Monks, T. J., and Sipes, I. G., Eds.) pp 129132, Kluwer Academi/Plenum Publishers, New York. (24) Kim, S. K., Kim, Y. C., and Kim, Y. C. (1998) Effects of Singly Administered Betaine on Hepatotoxicity of Chloroform in Mice. Food Chem. Toxicol. 36, 655-661. (25) Rossi, S., Gemma, S., Fabrizi, L., Testai, E., and Vittozzi, L. (1999) Time-dependence of chloroform-induced metabolic alterations in the liver and kidney of B6C3F1 mice. Arch. Toxicol. 73, 387393. (26) Butterworth: B. E., Templin, M. V., Borghoff, S. J., Conolly, R. B., and Kedderis, G. L. (1995). The role of regenerative cell proliferation in chloroform-induced cancer. Toxicol. Lett. 82-83, 23-26. (27) Robbiano, L., Mereto, E., Migliazzi Morando, A., Pastore, P., and Brambilla, G. (1998) Increased frequency of micronucleated kidney cells in rats exposed to halogenated anaesthetics. Mutat. Res./Genet. Toxicol. Envir. Mutagen. 413, 1-6. (28) Diaz Gomez, M. I., and Castro, J. A. (1980) Covalent binding of chloroform metabolites to nuclear proteins - No evidence for binding to nucleic acids. Cancer Lett. 9, 213-218. (29) Grunstein, M. (1997) Histone acetylation in chromatin structure and transcription. Nature 389, 349-352. (30) Struhl, K. (1998) Histone deacetylation and trascriptional regulatory mechanism. Genes Dev. 12, 599-606. (31) Bohm, L., and Crane-Robinson, C. (1984) Proteases as structural probes for chromatin: the domain structure of histones. Biosci. Rep. 4, 365-386. (32) Croft, L. R. (1980) Introduction to Protein Sequence Analysis, Chapter 1, p 11, John Wiley & Sons Ltd., Page Bros Ltd., Norwich, U.K. (33) Wu, J., and Grunstein, M. (2000) 25 years after the nucleosome model: chromatin modifications. Trends Biochem. Sci. 25, 619623. (34) Barton, D., and Ollis, W. D. (1979) Nitrogen Compounds, Carboxylic Acids, Phosphorus Compounds. In Comprehensive Organic Chemistry. The synthesis and Reactions of Organic Compounds. (Sutherland, I. O., Ed.) Vol. 2, pp 521-526, Pergamon Press, Oxford. (35) Pereira, M. A., Chang, L. W., Ferguson, J. L., and Couri, D. (1984) Binding of chloroform to the cysteine of hemoglobin. Chem.-Biol. Interact. 51, 115-124.

Chem. Res. Toxicol., Vol. 16, No. 3, 2003 275 (36) Strahl, B. D., and Allis, C. D. (2000) The language of covalent histone modifications. Nature 403 (6765), 41-45. (37) Zhang, W., Bone, J. R., Edmondson, D. G., Turner, B. M., and Roth, S. Y. (1998) Essential and redundant functions of histone acetylation revealed by mutation of target lysines and loss of the Gcn5p acetyltransferase. EMBO J. 17, 3155-3167. (38) Johnson, C. A., and Turner, B. M. (1999) Histone deacetylases: complex transducers of nuclear signals. Semin. Cell Dev. Biol. 10, 179-188. (39) Magnaghi-Jaulin, L., Ait-Si-Ali, S., and Harel-Bellan A. (1999) Histone deacetylase in signal transduction by growth regulatory signals. Semin. Cell Dev. Biol. 10, 197-203. (40) Marks, P. A., Richon, V. M., and Rifkind, R. A. (2000) Histone deacetylase inhibitors: inducers of differentiation or apoptosis of trasformed cells. J. Natl. Cancer Inst. 92, 1210-1216. (41) Cress, W. D., and Seto, E. (2000) Histone deacetylases, transcriptional control, and cancer. J. Cell. Physiol. 184, 1-16. (42) Fenrick, R., and Hiebert, S. W. (1998) Role of histone deacetylases in acute leukemia. J. Cell. Biochem. 30-31 (Suppl.), 194-202. (43) Kouzarides, T. (1999) Histone acetylases and deacetylases in cell proliferation. Curr. Opin. Genet. Dev. 9, 40-48. (44) Chung, D. (2002) Histone modification: the ‘next wave’ in cancer therapeutics. Trends Mol. Med. 8 (4) Suppl., S10-S11. (45) Timmermann, S., Lehrmann, H., Polesskaya, A., and HarelBellan, A. (2001) Histone acetylation and disease. Cell Mol. Life Sci. 58, 728-736. (46) Yu, Y. E., Zhang, Y., Unni, E., Shirley, C. R., Deng, J. M., Russell, L. D., Weil, M. M., Behringer, R. R., and Meistrich, M. L. (2000) Abnormal spermatogenesis and reduced fertility in transition nuclear protein 1-deficient mice. Proc. Natl. Acad. Sci. U.S.A. 97, 4683-4698. (47) Xin, Z., Allis, C. D., and Wagstaff, J. (2001) Parent-specific complementary patterns of histone H3 lysine 9 and H3 lysine 4 methylation at the Prader-Willi syndrome imprinting center. Am. J. Hum. Genet. 69, 1389-1394. (48) Phiel, C. J., Zhang, F., Huang, E. Y., Guenther, M. G., Lazar, M. A., and Klein, P. S. (2001) Histone deacetylase is a direct target of valproic acid, a potent anticonvulsant, mood stabilizer, and teratogen. J. Biol. Chem. 276, 36734-36741. (49) Larson, J. L., Wolf, D. C., and Butterworth, B. E. (1995) Induced regenerative cell proliferation in livers and kidneys of male F-344 rats given chloroform in corn oil by gavage or ad libitum in drinking water. Toxicol. 95, 73-86. (50) Larson, J. L., Wolf, D. C., Me´ry, S., Morgan, K. T., and Butterworth, B. E. (1995) Toxicity and cell proliferation in the liver, kidneys and nasal passages of female F-344 rats, induced by chloroform administered by gavage. Food Chem. Toxicol. 33, 443456. (51) Templin, M. V., Jamison, K. C., Wolf, D. C., Morgan, K. T., and Butterworth, B. E. (1996) Comparison of chloroform-induced toxicity in the kidneys, liver, and nasal passages of male OsborneMendel and F-344 rats. Cancer Lett. 104, 71-78.

TX025565I