Article pubs.acs.org/crt
Hydrolytic Cleavage Products of Globin Adducts in Urine as Possible Biomarkers of Cumulative Dose: Proof of Concept Using Styrene Oxide as a Model Adduct-Forming Compound Jaroslav Mráz,† Iveta Hanzlíková,† Alena Moulisová,‡ Šárka Dušková,† Kamil Hejl,‡ Aneta Bednárǒ vá,‡ Ludmila Dabrowská,† and Igor Linhart*,‡ National Institute of Public Health, Prague, Šrobárova 48, CZ-10042 Prague, Czech Republic Department of Organic Chemistry, Faculty of Chemical Technology, University of Chemistry and Technology, Prague, Technická 5, CZ-166 28 Prague, Czech Republic
† ‡
S Supporting Information *
ABSTRACT: A new experimental model was designed to study the fate of globin adducts with styrene 7,8-oxide (SO), a metabolic intermediate of styrene and a model electrophilic compound. Rat erythrocytes were incubated with SO at 7 or 22 °C. Levels of specific amino acid adducts in globin were determined by LC/MS analysis of the globin hydrolysate, and erythrocytes with known adduct content were administered intravenously to recipient rats. The course of adduct elimination from the rat blood was measured over the following 50 days. In the erythrocytes incubated at 22 °C, a rapid decline in the adduct levels on the first day posttransfusion followed by a slow phase of elimination was observed. In contrast, the adduct elimination in erythrocytes incubated at 7 °C was nearly linear, copying elimination of intact erythrocytes. In the urine of recipient rats, regioisomeric SO adducts at cysteine, valine, lysine, and histidine in the form of amino acid adducts and/or their acetylated metabolites as well as SO-dipeptide adducts were identified by LC/MS supported by synthesized reference standards. S-(2-Hydroxy-1-phenylethyl)cysteine and S-(2-hydroxy-2-phenylethyl)cysteine, the most abundant globin adducts, were excreted predominantly in the form of the corresponding urinary mercapturic acids (HPEMAs). Massive elimination of HPEMAs via urine occurred within the first day from the erythrocytes incubated at both 7 and 22 °C. However, erythrocytes incubated at 7 °C also showed a slow second phase of elimination such that HPEMAs were detected in urine up to 50 days post-transfusion. These results indicate for the first time that globin adducts can be cleaved in vivo to modified amino acids and dipeptides. The cleavage products and/or their predictable metabolites are excreted in urine over the whole life span of erythrocytes. Some of the urinary adducts may represent a new type of noninvasive biomarker for exposure to adduct-forming chemicals.
■
humans.7 However, because globin adducts are determined mainly in healthy subjects, their wider use is impeded by the need for taking blood, which is invasive sampling. Although the formation and properties of various types of globin adducts have been described extensively, their subsequent fate following physiological removal of erythrocytes from circulation has been, to our knowledge, totally unexplored. To address this challenge, we hypothesize that the adducted globin undergoes in vivo proteolytic cleavage to free amino acids that re-enter endogenous metabolic pathways, whereas the amino acid adducts as nonphysiological polar compounds are rapidly excreted in urine unchanged and/or in the form of predictable metabolites. The structures of these
INTRODUCTION It has been known for more than five decades that electrophilic chemicals can bind covalently to the nucleophilic sites in proteins and nucleic acids in vivo and in vitro to form specific adducts. This can lead to a variety of undesirable effects, including DNA mutations and carcinogenesis.1−4 Adducts with the blood proteins globin and albumin have been used primarily as surrogate biomarkers of the DNA damage2,5 but nowadays are mainly employed as in vivo dosimeters in biomonitoring of occupational and environmental exposure to several groups of toxicologically relevant chemicals, such as alkylating agents, arylamines, or isocyanates.6 Globin adducts offer important advantages over DNA adducts because of their abundance and analytical accessibility. Compared to urinary metabolites, they have a well-defined and long lifetime, reflecting the life span of erythrocytes, which is ∼126 days in © XXXX American Chemical Society
Received: December 20, 2015
A
DOI: 10.1021/acs.chemrestox.5b00518 Chem. Res. Toxicol. XXXX, XXX, XXX−XXX
Article
Chemical Research in Toxicology
Table 1. HPE Adduct Levels in Globin from Rat Erythrocytes Incubated with SO (IRE-SO) Determined after Acidic (AH) and Enzymatic (EH) Hydrolysis adduct level (nmol/g globin) HPEC (EH)
HPEV (AH)
HPEVL (EH)
HPEK (AH)
HPEK (EH)
concn of SO (mM)
temp. of incubation (°C)
batch of IRE-SO
21-
22-
21-
22-
21-
22-
21-
22-
21-
22-
2 5 5 10 10 20
22 7 22 7 22 7
2B 5A 5B 10A 10B 20A
190 320 460 578 611 608
81.4 128 190 229 257 263
18.2 21.6 75.4 74.8 257 229
13.1 16.4 67.2 54.9 185 149
7.7 8.3 32.5 27.7 106 80.8
9.4 9.3 38.3 31.1 124 85.3
3.7 3.4 13.8 12.8 49.0 41.8
4.4 3.5 15.7 15.6 59.9 49.6
0.6 0.6 2.3 2.2 7.1 5.0
0.7 0.7 3.6 2.8 11.7 7.6
ethyl)-DL-valine (22-HPEV), N-(2-hydroxy-1-phenylethyl)-DL-valyl-Lleucine (21-HPEVL), N-(2-hydroxy-2-phenylethyl)-DL-valyl-L-leucine (22-HPEVL), Nε-(2-hydroxy-1-phenylethyl)-DL-lysine (21-HPEK), Nε-(2-hydroxy-2-phenylethyl)-DL-lysine (22-HPEK), Nα-acetyl-Nε-(2hydroxy-1-phenylethyl)-DL-lysine (Ac-21-HPEK), Nα-acetyl-Nε-(2-hydroxy-2-phenylethyl)-DL-lysine (Ac-22-HPEK), and N-(2-hydroxy1,1,2,2-tetradeuteroethyl)-DL-valyl-L-leucine (HEVL-d4) are described in the Supporting Information. Mercapturic acids N-acetyl-S-(2hydroxy-1-phenylethyl)-L-cysteine (21-HPEMA), N-acetyl-S-(2-hydroxy-2-phenylethyl)-L-cysteine (22-HPEMA),21 and N-acetyl-S-(4fluorophenyl)-L-cysteine (FSPMA)22 were prepared as described previously in the literature. Racemic styrene 7,8-oxide (SO) was obtained from Sigma-Aldrich (Prague, Czech Republic). Methanol and acetonitrile (both LiChrosolv hypergrade for LC-MS), formic acid (98−100% Suprapur), and hydrochloric acid for hydrolyses (32%, GR for analysis) were obtained from Merck; HPLC/MS grade water was prepared using a Purite (Select Neptune) unit from Watrex (Prague, Czech Republic). Pronase from Streptomyces griseus (Cat. 53702) was from Calbiochem. SPE Phenomenex Strata X-C 33 μm columns (60 mg/3 mL), analytical HPLC columns (Phenomenex Aqua C18, 150 × 2 mm, 125 Å pore size, 5 μm particle size and Phenomenex Synergi 4 μ Polar RP, 150 × 3 mm), and semipreparative columns (Phenomenex Aqua C18, 250 × 10 mm, 125 Å pore size, 10 μm particle size) were purchased from Chromservis (Prague, Czech Republic). Study Outline. Rat erythrocytes were incubated with SO, and the levels of SO-specific adducts with globin were determined. Portions of the modified erythrocytes with known total amounts of the adducts were administered by transfusion to recipient rats. Levels of the globin adducts in the rat blood were determined immediately after the transfusion and thereafter over 50 days. Rat urine was analyzed to identify SO-specific cleavage products of the adducted globin to determine their recovery from globin adducts and to characterize the kinetics of elimination. Extensive work was undertaken to synthesize reference compounds of all individual SO-specific products to be determined in the rat globin and urine. Preparation of Rat Erythrocytes Modified with SO (IRE-SO). Four male Wistar rats, 300−350 g (Velaz, Czech Republic), were anaesthetized with pentobarbital, 60 mg/kg, and exsanguinated from the aorta. Individual blood samples were spun down at 3500 rpm for 15 min; plasma was removed, and the erythrocytes were washed twice with saline, pooled, and resuspended in an equal volume of saline. Six aliquots of resuspended pooled erythrocytes (4 mL each) were placed in 10 mL screw-cap tubes, and neat SO or SO solutions in dioxane (10% or 20%) were added to reach a final SO concentration of 2, 5, 10, or 20 mM (Table 1). The maximum concentration of dioxane in the erythrocyte suspension was 0.4% (v/v). The tubes were closed and placed on a roller mixer (1 rev/s) in a cabinet thermostat set at 7 or 22 °C for 16 h. The SO-modified erythrocytes (IRE-SO) were then spun down, washed, and again resuspended as described for blood samples. Two portions of the suspension of each batch of IRE-SO were administered by transfusion to recipient rats (2 mL/kg), and small aliquots (2 × 25 μL) were taken for two parallel spectrophotometric determinations of hemoglobin by Drabkin’s method.23 The remaining portion of IRE-SO was used to isolate globin for determining the SO-specific adducts (nmol/g globin).
urinary products would be indicative of specific reactive xenobiotics, and their concentrations would reflect the extent of binding to the protein. Therefore, they might be considered as alternative biomarkers of cumulative exposure to adductforming xenobiotics. For this hypothesis to be tested, a new experimental model was designed using rat erythrocytes incubated with the adductforming chemical in vitro. The level of adducts with globin is determined, and a dose of the modified erythrocytes is administered by transfusion to unexposed recipient rats. Urine of the recipient rats is then analyzed for the ultimate products resulting from hydrolytic cleavage of the adducted globin. This concept offers several advantages over direct administration of adduct-forming chemicals. Whereas the applied dose of the parent chemical can be limited by its toxicity, it was shown that intravenous administration of erythrocytes with even considerably high adduct levels was well-tolerated.8 Further, a known initial amount of the globin adducts transferred to the body allows for accurately determining a quantitative relationship between these adducts and their cleavage products in the urine. In contrast, direct administration of the parent compound would inevitably produce adducts with nucleophilic amino acids in a range of proteins that would be excreted in addition to cleavage products from adducted globin. In the current study, styrene 7,8-oxide (SO), an electrophilic metabolite of styrene known to bind to DNA and proteins, was used as a model adduct-forming xenobiotic. SO has been classified by IARC as a category IIA carcinogen9 and together with styrene is one of the most extensively studied occupational and environmental toxicants.4,10−17 Previously, we characterized regio- and stereoisomeric patterns of Cys, Lys, and His adducts in human globin incubated in vitro with SO using a method consisting of total protein hydrolysis followed by GC/ MS and HPLC/MS.18 The same procedure was employed in this study to determine Cys, Val, and Lys adducts with SO in the rat erythrocytes used for transfusion. Using our new experimental model, ultimate urinary products resulting from degradation of specific SO adducts in globin were identified, and quantitative relationships between the SO adducts in globin and the corresponding cleavage products in urine are described.
■
EXPERIMENTAL SECTION
Materials. Cysteine adducts, S-(2-hydroxy-1-phenylethyl)-L-cysteine (21-HPEC), and S-(2-hydroxy-2-phenylethyl)-L-cysteine (22HPEC)19 were prepared as previously described. Syntheses of the other amino acid adducts, namely, 21-HPEC-d8 and 22-HPEC-d8, N(2-hydroxy-1,1,2,2-tetradeuteroethyl)-DL-valine (HEV-d4),20 N-(2-hydroxy-1-phenylethyl)-DL-valine (21-HPEV), N-(2-hydroxy-2-phenylB
DOI: 10.1021/acs.chemrestox.5b00518 Chem. Res. Toxicol. XXXX, XXX, XXX−XXX
Article
Chemical Research in Toxicology
These were washed with diluted phosphoric acid (pH 3.2, 2 × 1 mL) and eluted with 80% aqueous MeOH (methanolic fraction, 1 mL) and then with a 5:95 mixture of 28−30% aqueous NH4OH and MeOH (ammonia fraction, 1 mL). Both the methanolic and ammonia fractions were evaporated in a vacuum rotary evaporator to dryness, and the residues were reconstituted with water (1 mL) before analysis. The methanolic fraction was analyzed for HPEMAs, whereas all other analytes were eluted in the ammonia fraction. HPLC-MS/MS Analysis. All analyses were performed on an HPLC-MS system consisting of an UltiMate 3000 HPLC instrument coupled to an Orbitrap Q Exactive high resolution mass spectrometer (Thermo). Globin hydrolysates and ammonia fractions of SPE-cleaned urine samples were analyzed on a Phenomenex Synergi 4 μ Polar RP column heated at 40 °C. Mobile phase A (0.1% formic acid) was pumped at a flow rate of 0.3 mL/min (0−5 min); then, mobile phase B (MeOH) was added with the concentration increasing linearly from 0 to 50% (5−15 min) and held constant for an additional 3 min. Thereafter, the mobile phase returned to 100% A for 2 min, and the column was equilibrated for another 2 min. Simultaneous determination of the analytes was performed in the positive electrospray ionization (+ESI) mode using a parallel reaction monitoring (PRM) experiment to monitor MS2 spectra obtained by collisionally induced dissociation of quasimolecular ions (M + H)+ at a uniform collision energy of 25 auxiliary units. The capillary voltage was set to 3.5 kV and temperature to 320 °C. Sheath gas, auxiliary gas, and sweep gas flow rates were 48, 11, and 2 L/min. Transitions used for quantitation of the individual analytes and internal standards were as follows (m/z): 21-HPEV, 238.1438 → 118.0863; 22-HPEV, 238.1438 → 220.1332; 21-HPEVL, 351.2278 → 72.0808; 22-HPEVL, 351.2278 → 174.1277; 21-HPEC, 242.0845 → 121.0648; 22-HPEC, 242.0845 → 135.0263; 21- and 22-HPEK, 267.1703 → 130.0863; Ac-21-HPEK, 309.1809 → 189.1234; Ac-22-HPEK, 309.1809 → 291.1703; HPEH (V), 276.1343 → 156.0768; HEVL-d4, 279.2216 → 120.1321; 21-HPEC-d8, 250.1347 → 127.1025, and HEV-d4, 166.1376 → 120.1321. Analyses for 21- and 22-HPEMA in the methanolic fraction of SPEcleaned urine samples were conducted on a Phenomenex Aqua C18 column heated at 40 °C. Mobile phases A (0.1% formic acid) and B (0.1% formic acid in acetonitrile) were pumped at a total flow rate of 0.3 mL/min using the gradient described above. Determination was performed in the negative electrospray ionization (−ESI) mode using a PRM experiment to monitor MS2 spectra obtained by collisionally induced dissociation of quasimolecular ions (M − H)− at a collision energy of 25 auxiliary units. The settings at the electrospray probe were as described above. The transition used for quantitation of both 21- and 22-HPEMA was m/z 282.0806 → 153.0376; the internal standard FSPMA was monitored at m/z 256.0449 → 127.0023. For calibration of the adduct determination in the globin hydrolysates, blank globin samples were spiked with appropriate amounts of synthetic reference compounds and with internal standard(s) before the hydrolytic step and then treated in the same way as authentic samples. HEV-d4 was used as an internal standard to quantify HPEV and HPEK in acidic hydrolysates and 21-HPEC-d8 (to quantify HPEC) and HEVL-d4 (to quantify HPEV, HPEVL, and HPEK) in enzymatic hydrolysates. For calibration of urine analyses, blank urine samples were spiked with appropriate amounts of synthetic reference compounds and with internal standard(s), transferred onto SPE columns, and worked up as authentic samples. FSPMA was an internal standard for quantitation of HPEMAs in the methanolic fraction, whereas 21-HPEC-d 8 and HEVL-d 4 were used for quantitation of adducts eluted in the ammonia fraction (HPEC, HPEV, HPEVL, and Ac-HPEK). Whenever an analyte or an internal standard was eluted in the form of two peaks of diastereomers, the sum of both peak areas was calculated. Concentration ranges of the calibration samples and the amounts of the internal standard(s) were adapted to the actual adduct levels in authentic globin or urine samples and, for individual analytes, varied within 3 orders of magnitude from 0−5 ng/sample to 0−5 μg/ sample. All calibration curves were linear, typically with R2 > 0.99.
Transfusion of Modified Erythrocytes to Recipient Rats. Rats were anesthetized with pentobarbital (60 mg/kg), and their vena jugularis were cannulated. The IRE-SO suspension (2 mL/kg body wt), given to 2 animals per each IRE-SO batch, was administered by a syringe; the cannula was left in the vein, and its inner space was rinsed by the injection of saline (0.2 mL). An initial blood sample (0.5 mL) was withdrawn through the cannula 10 min after transfusion. The cannula was then removed; the vena jugularis was tied off, and the wound was sewed up. Rats were placed in glass metabolic bowls to collect the total urine samples over 24 h on days 1 and 2 and then on days 8, 14, 20, 27, 42, and 49. Water (50 mL/day) but no food was provided in the metabolic bowls. After each collection of urine, the bowls were repeatedly rinsed to wash out urine quantitatively, thereby diluting it to ∼80−100 mL/24 h. Urine samples were filtered through a paper filter and stored at −20 °C before analysis. Blood samples (0.5−1 mL) were taken under diethyl ether anesthesia from the plexus ophthalmicus at the end of each urine collection period. The total amount (nmol) of each adduct introduced by transfusion to the recipient rat was calculated as a multiple of administered volume of the IRE-SO suspension (mL), concentration of hemoglobin in the IRE-SO suspension (g/mL), and the adduct level in globin (nmol/g globin). Animal experiments were approved by the Committee for Animal Protection of the Czech Ministry of Health. Intraperitoneal Administration of HPEV and HPEC. Synthetic valine derivatives 21- and 22-HPEV were used as mixtures of two diastereomeric racemates each, whereas cysteine derivatives 21- and 22-HPEC were mixtures of two diastereomers, each with natural Lconfiguration at cysteine. Solutions of 21-HPEV, 22-HPEV, 21-HPEC, and 22-HPEC in saline (1 mg/2 mL) were administered intraperitoneally to male Wistar rats, 300−350 g, at a dose of 1 mg/kg body wt (4 animals/compound). Rats were then placed in glass metabolic bowls to collect total urine samples over 24 h on days 1 and 2 as described above. Isolation of Globin. Blood samples (∼1 mL) were spun down at 3500 rpm for 15 min, and plasma was removed. Erythrocytes were washed twice with saline and hemolyzed by the addition of distilled water (0.3 mL) and sonication (5 min). Globin was precipitated from the hemolyzate by a 98:2 (v/v) ice-cold mixture of acetone and concentrated hydrochloric acid (6 mL), washed at least three times with ice-cold acetone until the solvent was colorless, then once with ice-cold diethyl ether, and dried in a rotary vacuum concentrator at 50 °C for 5 min. Globin samples were stored at 4 °C before analysis. Sample Preparation for Globin Analyses. Acidic Hydrolysis (AH). Globin samples (5 mg) were dissolved in 12 M HCl (0.5 mL); the solutions were spiked with an internal standard solution and heated at 100 °C for 16 h. Depending on the target compounds and their concentration range in the analyzed globins, 1−1000 ng of the internal standard HEV-d4 in 100 μL of solution was added to the samples. Acidic hydrolysates were evaporated in a Teflon-coated rotary vacuum concentrator to dryness; the residues were reconstituted in 10% aqueous ethanol (1 mL) and filtered through a 0.45 μm PVDF Durapore membrane filter (Ultrafree, Merck) before analysis. Enzymatic Hydrolysis (EH). Globin samples (5 mg) were dissolved in distilled water (0.4 mL), and the solutions were spiked with internal standard aqueous solutions (HEVL-d4 and/or 21-HPEC-d8, 100 μL, 1−1000 ng). Phosphate buffer (0.1 M, pH 7.4, 0.45 mL) and a 1% suspension of Pronase in the same buffer (50 μL) were added. Vials were closed and gently shaken in a thermostat at 37 °C for 16 h. Enzymatic hydrolysates were spun down; the supernatants were filtered through a 0.45 μm PVDF Durapore filter and diluted with ethanol (0.1 mL) before analysis. Sample Preparation for Urine Analyses. A single solid-phase extraction (SPE) protocol was developed to isolate all SO-specific degradation products from rat urine. The capacity of the sorbent was tested for 1 mL of rat urine as obtained after washing metabolic bowls, i.e., diluted to ∼100 mL/24 h. Strata X-C 33 μm columns (60 mg/3 mL, Phenomenex) were preconditioned with MeOH (1 mL) and equilibrated with phosphoric acid solution, pH 3.2 (1 mL). Rat urine samples acidified with 1 M phosphoric acid to pH 3.2 (1 mL) and spiked with internal standards were transferred onto the columns. C
DOI: 10.1021/acs.chemrestox.5b00518 Chem. Res. Toxicol. XXXX, XXX, XXX−XXX
Article
Chemical Research in Toxicology
■
RESULTS AND DISCUSSION Synthesis of Reference Standards. When nucleophiles, such as thiol groups of cysteine or amino groups in proteins and amino acids, react with SO, two regioisomeric products are formed. Attack at the α- and β-carbon of the oxirane ring leads to 2-hydroxy-1-phenylethyl (21-HPE) and 2-hydroxy-2-phenylethyl (22-HPE) derivatives, respectively (Scheme 1). There-
Scheme 2. Synthesis of Valine and Valyl-Leucine Adducts with SO
Scheme 1. Regio- and Stereoselectivity of the Reaction of SO with SH and NH2 Groups in Proteins
Scheme 3. Synthesis of Lysine Adducts with SO fore, direct reaction of SO with unprotected or N-protected amino acids leads to a mixture of regioisomers. The adducts of SO with cysteine, lysine, and histidine have been previously prepared by direct reactions of SO with cysteine, Nα-protected lysine, and Nα-protected histidine, respectively, followed by HPLC separation.18 However, only small amounts of pure products could be obtained by this procedure. For sufficient amounts of the individual regioisomers to be obtained, new synthetic procedures have been developed. Cysteine adducts 21- and 22-HPEC were prepared by a regioselective ring opening reaction of SO with protected cysteine as described recently.19 Valine adducts 21- and 22-HPEV were synthesized by reductive amination of 3-methyl-2-oxobutanoic acid (MOBA) with 2-amino-2-phenylethanol (22-APE) and 2-amino-1phenylethanol (21-APE), respectively (Scheme 2). This reaction led to the formation of a new stereogenic center at the α-carbon of valine. As a result, two diastereomeric racemates of both 21- and 22-HPEV were formed when racemic 22- and 21-APE, respectively, were used. Individual diastereomeric racemates were isolated by semipreparative HPLC. Valyl-leucine adducts 21- and 22-HPEVL were prepared analogously by reductive amination of N-(3-methyl-2-oxobutanoyl)-L-leucine ethyl ester (MOB-L-Et) as shown in the Scheme 2. Two major diastereomers of each HPEVL regioisomer were distinguished by HPLC as well as in NMR spectra. Lysine adducts 21- and 22-HPEK were synthesized by the reaction of 22- and 21-APE, respectively, with protected 2amino-5-bromohexanoic acid (Scheme 3). Nα-Acetyl derivatives Ac-21- and Ac-22-HPEK were prepared in an analogous way from 2-acetamido-5-bromohexanoic acid (Scheme 3). Formation and Analysis of Globin Adducts with SO in Rat Erythrocytes in Vitro. Suspensions of rat erythrocytes were incubated with SO at four concentration levels (2, 5, 10, and 20 mM) and two temperatures (7 and 22 °C) for 16 h. Globin samples isolated from these erythrocytes were hydro-
lyzed under acidic conditions or enzymatically by Pronase. Adducts with cysteine, lysine, histidine, and N-terminal valine arising from the attack by nucleophilic thiol or amino groups or imidazole ring nitrogens at both α- and β-carbons of the SO oxirane ring were detected in the globin digests. Because both the HPE moiety and corresponding amino acid contain an asymmetric (stereogenic) carbon, each regioisomer is composed of two diastereomers with retained L-configuration of the amino acid. In this study, adducts identified in acidic hydrolysates by comparison of their chromatographic retention times and MS2 spectra with those of reference standards included 21 and 22 regioisomers of HPEV and HPEK. Further, five peaks with MS2 spectra characteristic for HPE adducts of histidine (HPEH I−V) were detected. However, as authentic standards of the HPEH adducts that form a complex regio- and diastereomer mixture (theoretically, 8 isomeric species) were not available, the individual isomers observed in our sample could not be identified unequivocally. In agreement with a previous report,18 cysteine adducts 21- and 22-HPEC were not found in the acidic hydrolysate due to their instability under the harsh acidic conditions used. D
DOI: 10.1021/acs.chemrestox.5b00518 Chem. Res. Toxicol. XXXX, XXX, XXX−XXX
Article
Chemical Research in Toxicology
Table 2. Detection of HPE Adducts in the Acidic (AH) and Enzymatic (EH) Hydrolysates of Globin and in Urine (U) presence in matrix adducts
a
number of HPLC peaks
21-HPEV 22-HPEV 21-HPEVL 22-HPEVL
2 2 2 2
21-HPEC 22-HPEC 21-HPEMA 22-HPEMA
2 2 1 1
21-HPEK 22-HPEK Ac-21-HPEK Ac-22-HPEK
1 1 1 1
HPEHa Ac-HPEHa
5 6
retention time (min) Val Adducts 5.02, 7.60 6.07, 6.69 13.4, 13.6 14.3, 14.5 Cys Adducts 3.40, 3.95 3.54, 3.74 11.0 11.2 Lys Adducts 1.76 1.84 8.22 7.98 His Adducts 1.55, 1.70, 1.79, 2.12, 2.51 5.31, 5.44, 6.08, 6.40, 6.60, 8.12
AH * *
EH
U
* * *
* * * *
* *
* *
* * * *
* * * *
*
* *
Reference standards were not available (identification based on mass spectra only).
peptides composed of natural amino acids. Consequently, peptidic bonds adjacent to adducted amino acids may resist Pronase hydrolysis. We reported examples of similar dipeptidic adducts formed by Pronase hydrolysis of adducted globin previously.24,25 The HPE adducts with cysteine (21- and 22-HPEC), valine (21- and 22-HPEV), lysine (21- and 22-HPEK), and valylleucine (21- and 22-HPEVL) in six batches of IRE-SO were determined quantitatively by HPLC/MS analysis of acid or Pronase-hydrolyzed adducted globin. Their levels in globin from six batches of rat erythrocytes incubated in vitro with SO (IRE-SO) are shown in (Table 1). Very good correlations (R2 > 0.99) were obtained between levels of the corresponding adducts as measured after acidic and after enzymatic hydrolysis, i.e., 21-HPEV vs 21-HPEVL, 22-HPEV vs 22-HPEVL, 21HPEK (AH) vs 21-HPEK (EH), and 22-HPEK (AH) vs 22HPEK (EH). More importantly, however, considerable differences were noted between absolute values of adduct levels in acidic and enzymatic hydrolysates, demonstrating that neither acid nor Pronase hydrolysis is universally applicable for all types of the adducts. Valine is the N-terminal amino acid in all four globin chains, of which two α-chains N-terminate with Val-Leu and two βchains with Val-His. HPE adducts at both α- and β-chain termini provide HPEV on acidic hydrolysis. Assuming that reactivity of SO with N-terminal Val in both α- and β-chains is equal and that yields of HPEV and HPEVL upon acidic and enzymatic hydrolysis are quantitative, the ratio of HPEVL (EH) to HPEV (AH) should be 0.50. The mean values found for 21HPEVL (EH) to 21-HPEV (AH) and 22-HPEVL (EH) to 22HPEV (AH) were 0.38 and 0.59, respectively, confirming that reactivity of the N-termini of rat globin α- and β-chains with SO is indeed similar. In the Pronase hydrolysate, a small but clearly detectable amount of 22-HPEV was found beside 22-HPEVL (molar ratio of 1:80), whereas no 21-HPEV was detected, thereby demonstrating a quantitatively negligible yet interesting example of regioselectivity of Pronase action.
In the enzymatically hydrolyzed globin, 21-HPEC, 22-HPEC, 21-HPEK, 22-HPEK, and HPEH isomers were detected by HPLC-MS as expected. Rather surprisingly, only a small amount of a single diastereomer of 22-HPEV and no 21-HPEV was detected in the globin hydrolysate even from erythrocytes incubated with 20 mM SO. In the HPLC system used, individual diastereomers of 21-HPEC, 22-HPEC, 21-HPEV, and 22-HPEV were clearly resolved. In contrast, diastereomers of 21- and 22-HPEK coeluted, probably because the stereogenic carbons in these compounds are separated by as many as 6 or 7 bonds such that physicochemical properties do not differ significantly. In addition to the adducts with single amino acids, four peaks at m/z 351.2278 corresponding to HPE derivatives of valyl-leucine (an N-terminal dipeptide of rat globin α-chain) (HPEVL) were found. In the collisionally activated MS2 spectra, the earlier eluting pair of isomers gave identical fragment ions at m/z 231.1704 corresponding to (MH − SO)+ and m/z 72.0808 corresponding to (CH3)2−CH−CHNH2+, whereas the later eluting pair gave MS2 fragment ions at m/z 333.2174 (MH − H2O)+ and m/z 174.1277 (MH − SO − CH2CH(CH3)2)+. On the basis of these fragmentation patterns, the first and second pairs of compounds could be identified as 21- and 22-HPEVL, respectively, with each consisting of 2 diastereomers. This assignment was confirmed by comparison with synthetic reference standards of both HPEVL regioisomers, although the patterns of diastereomers in synthetic and “natural” 21-HPEVL were not identical due to differences in the absolute configuration at valine between “natural” HPEVLs released from globin and the synthetic reference standards. Nevertheless, MS2 spectra of all diastereomers of the same regioisomer were virtually identical. An overview of the adducts detected in both AH and EH globin samples is given in Table 2. The adducts with amino acids Cys, His, and Lys have been reported in the Pronase-hydrolyzed globin samples previously,18 whereas dipeptide HPEVL adducts have not, to the best of our knowledge, been described as of yet. Formation of HPEVL due to incomplete Pronase hydrolysis is not surprising because physiological substrates of Pronase are proteins and E
DOI: 10.1021/acs.chemrestox.5b00518 Chem. Res. Toxicol. XXXX, XXX, XXX−XXX
Article
Chemical Research in Toxicology
identical for both adducts (HPEC and HPEV), were observed. In rats given erythrocytes incubated with 5 mM SO at 7 °C (IRE-SO 5A), the adduct levels declined slowly in a nearly linear fashion (Panels 5A in Figures 2 and 3). The linear decline
The levels of HPEK adducts in acidic hydrolysates were approximately 1 order of magnitude higher than those in enzymatic hydrolysates. This discrepancy can be explained by hindered proteolytic action of Pronase at adducted lysine. Thus, only acidic hydrolysis provides true HPEK levels in globin. Unlike in HPEV and HPEK, direct comparison of HPEC adduct levels in enzymatic and acidic hydrolysates is not possible due to HPEC decomposition in the strongly acidic medium.18 In analogy with HPEK, we assume that the level of HPEC in globin is also greatly underestimated if determined after hydrolysis by Pronase. When the levels of HPE adducts with Val, Lys, His, and Cys in globin are plotted against the SO concentration in the incubation medium, the first three curves are convex, whereas the curve for Cys adducts is concave (Figure 1). The concave curve may reflect a saturation of
Figure 2. Elimination of 21- and 22-HPEC in globin of rats that received transfusion of erythrocytes incubated with 5 and 10 mM SO at 7 °C (batches A) and 22 °C (IRE-SO batches B). Each point represents the mean of values measured in two animals.
Figure 1. Incubation of rat erythrocytes with SO. Relationships between HPE adduct levels in globin and SO concentration in the incubation mixture at 7 and 22 °C. Regioisomers of HPEV, HPEK, and HPEH were determined in acidic and those of HPEC in enzymatic hydrolysates.
cysteine binding sites at SO concentrations between 10 and 20 mM. Under the conditions of apparent saturation, 21- plus 22HPEC levels were
