562
Chem. Res. Toxicol. 2002, 15, 562-569
Mass Spectrometric Characterization of Human Hemoglobin Adducts Formed in Vitro by Hexahydrophthalic Anhydride Monica H. Kristiansson,* Bo A. G. Jo¨nsson, and Christian H. Lindh Department of Occupational and Environmental Medicine, Institute of Laboratory Medicine, University Hospital, SE-221 85 Lund, Sweden Received November 29, 2001
Primary structural information of anhydride binding to endogenous proteins is of interest in order to determine the mechanism causing the type-I allergy seen in many anhydride-exposed workers. In addition, studies on specific protein adducts may generate new methods for biological monitoring. In this study, the binding of hexahydrophthalic anhydride (HHPA) to human hemoglobin (Hb) in vitro was investigated. The in vitro synthesized conjugates were analyzed using a hybrid quadrupole-time-of-flight mass spectrometer (Q-TOF) with electrospray ionization (ESI) to determine the number of HHPA adducts per Hb molecule. Structural information on the locations of the adducts was obtained through nanospray Q-TOF, liquid chromatography-ESI mass spectrometric analysis, and gas chromatography/mass spectrometric analysis of Pronase E and tryptic digests. Up to six adducts were found on the R-chain and five on the β-chain. The HHPA-adducts were localized to the N-terminal valine of the R- and β-chains of Hb and to lysine residues at positions 7, 11, 16, and 40 of the R-chain and 8, 17, 59, 66, and 144 of the β-chain. These results will constitute a basis for studies on structureactivity relationships as well as for development of methods for biological monitoring of acid anhydrides.
Introduction Allergy is a frequent and rapidly increasing disorder in the western world. However, there is still very little knowledge about the chemical structures of the allergens causing the development of the allergy. One reason for this may be that mainly large protein allergens have been studied and no common denominator of characteristic structures of these allergens has been determined so far (1). Studies of small molecules (haptens) that have the ability to form full antigens by reaction with endogenous protein may offer interesting possibilities for investigations of the chemical characteristics of antigens. However, there are only a few reports showing some characteristics of protein adducts of allergenic compounds (24). Organic acid anhydrides (OAAs)1 are low molecular weight, reactive chemicals with a wide use in the chemical industry today. It has been found that more than 60% of exposed workers may develop airway symptoms (5, 6). Furthermore, specific IgE antibodies against haptenserum albumin conjugates were found in a large fraction of the symptomatic workers (7). This indicates that type-1 allergy [according to the definition of Gell and Coombs (8)] is a major pathophysiological mechanism for the * To whom correspondence should be addressed at the Department of Occupational and Environmental Medicine, University Hospital, SE221 85 Lund, Sweden. Phone: +46 46 173148, Fax: +46 46 143702, E-mail:
[email protected]. 1 Abbreviations: ESI, electrospray ionization; Hb, hemoglobin; HHP acid, hexahydrophthalic acid; HHPA, hexahydrophthalic anhydride; HHPL, N-hexahydrophthaloyl-L-lysine; HHPV, NR-hexahydrophthaloyl-L-valine; OAA, organic acid anhydride; Q-TOF, quadrupole-timeof-flight; TFA, trifluoroacetic acid.
induction of the symptoms. Thus, the binding of the OAAs to endogenous proteins transforms the proteins into potent allergens. Some OAAs are active at exposure levels as low as only a few µg/m3, which make these highly sensitizing chemicals excellent model compounds for studies of allergenic structures (9, 10). The aim of this study was to investigate the chemical structure of the protein adducts in human hemoglobin (Hb) formed after in vitro reaction with a particularly strong sensitizer, hexahydrophthalic anhydride (HHPA) (11, 12, Rosqvist et al.).2
Materials and Methods Caution: Hexahydrophthalic anhydride (HHPA) and pentafluorobenzyl bromide (PFBBr) are hazardous chemicals and must be handled using proper safety measures. Materials. CsI, HHPA (99%) (cis-1,2-cyclohexanedicarboxylic anhydride), HCl, and trisodium phosphate dodecahydrate (Na3PO4‚12H2O) were purchased from Acros Organics (Geel, Belgium). NR-t-BOC-L-lysine and -L-valine, pentafluorobenzyl bromide (PFBBr), Pronase E, and trifluoroacetic acid (TFA) were obtained from Sigma Chemicals (St. Louis, MO). Trypsin (sequence grade) was obtained from Roche Diagnostics GmbH (Mannheim, Germany). ZipTipC18 pipet tips were obtained from Millipore (Bedford, MA). Acetic acid, acetone, disodium hydrogen phosphate dodecahydrate (Na2HPO4‚12H2O), NaH2PO4, NaCl, KCl, KH2PO4, ethylenedinitrilotetraacetic acid disodium salt dihydrate (EDTA-Na2‚2H2O), ethylenedinitrilotetraacetic acid tetrasodium salt tetrahydrate (EDTA-Na4‚4H2O), and calcium 2 Rosqvist, S., Nielsen, J., Welinder, H., Rylander, L., Lindh, C. H., and Jo¨nsson, B. A. G. Exposure-response relationships for hexahydrophthalic and methylhexahydrophthalic anhydrides using total plasma protein adducts as biomarkers (unpublished experiments).
10.1021/tx0155911 CCC: $22.00 © 2002 American Chemical Society Published on Web 03/15/2002
Hemoglobin Adducts of Hexahydrophthalic Anhydride chloride dihydrate (CaCl2‚2H2O) were obtained from Merck (Darmstadt, Germany). NH4HCO3 came from BDH Laboratory Supplies (Poole, England). Acetonitrile, dichloromethane, methanol, and toluene were purchased from Lab-Scan (Dublin, Ireland). Tetra-n-butylammonium hydrogen sulfate (TBA) was obtained from Merck-Schuchardt (Hohenbrunn, Germany). Sex pheromone inhibitor iPD1 was purchased from Bachem (Bubendorf, Switzerland). [2H6]Hexahydrophthalic (HHP) acid was synthesized in the department’s laboratory together with Synthelec (Lund, Sweden). The dialysis was performed using Spectra/Por molecular porous membrane tubing, with a cutoff at 12 000-14 000 Da, from Spectrum (Gardena, CA). Buffers, Reagents, and Standards. Phosphate-buffered saline (PBS) containing EDTA (pH 7.4) was prepared by adding 8.0 g of NaCl, 0.2 g of KCl, 0.2 g of KH2PO4, 1.4 g of Na2HPO4‚ 12H2O, 0.45 g of EDTA-Na4‚4H2O, and 0.45 g of EDTA-Na2‚ 2H2O to 1 L of water. A solution of 0.13 M PFBBr was prepared prior to use by adding 0.2 mL of PFBBr to 10 mL of dichloromethane. A solution of 0.1 M TBA was prepared by adding 3.39 g of TBA, 1.79 g of Na2HPO4‚12H2O, and 1.90 g of Na3PO4‚12H2O to 100 mL of water. Standards of HHP acid were prepared by hydrolyzing HHPA in water and then further diluting with water. Synthesis of NE-Hexahydrophthaloyl-L-lysine (HHPL). R N -t-BOC-L-lysine (0.5 g) was dissolved in 30 mL of water. HHPA (0.6 g) was dissolved in 1 mL of acetonitrile which was then slowly added to the NR-t-BOC-L-lysine solution. The pH was maintained between 7 and 9 by addition of 5 M NaOH. After evaporation, the t-BOC protective group was removed using 20 mL of 1 M HCl in acetic acid at room temperature for 15 min. The solution was then evaporated, and the HHPL was purified with reversed-phase liquid chromatography (LC; Hewlett-Packard 1050 LC system; Palo Alto, CA) using a C18 Apex II column from Jones Chromatography (Lakewood, CO; 10 mm i.d. × 250 mm). The eluate was monitored through UV-detection at 220 nm. The samples were dissolved in 0.1% TFA in water, and 1.0 mL aliquots were injected through a loop. The mobile phase was 0.1% TFA for 1 min with a flow rate of 2 mL/min followed by a linear gradient to 32 min of 0.07% TFA in acetonitrile. HHPL was collected in 0.1 M NH4HCO3 solution and immediately evaporated. No HHP acid was found by LC-UV detection in the samples after the purification. Synthesis of Nr-Hexahydrophthaloyl-L-valine (HHPV). L-Valine (0.3 g) was dissolved in 10 mL of water. HHPA (0.4 g) was dissolved in 1 mL of acetonitrile which was then slowly added to the water solution. The pH was kept between 7 and 9 by addition of 5 M NaOH. The solution was evaporated, and the HHPV was purified using reversed-phase LC using a C18 Apex II column from Jones Chromatography (10 mm i.d. × 250 mm). The eluate was monitored through UV-detection at 220 nm. The samples were dissolved in 0.1% TFA in water, and 1.0 mL aliquots were injected through a loop. The mobile phase was 0.1% TFA for 1 min with a flow rate of 2 mL/min followed by a linear gradient to 32 min of 0.07% TFA in acetonitrile. The HHPV was collected in 0.1 M NH4HCO3 solution and immediately evaporated. No HHP acid was found by LC-UV detection in the samples after the purification. In Vitro Synthesis of Hb-HHPA Conjugates. Hb (R-chain of mass 15 126 amu and β-chain of mass 15 867 amu) was prepared from human erythrocytes obtained from one donor. These were washed and then lysed by addition of water and freezing. The cell debris was pelleted by centrifugation for 30 min at 18000g (J2-21M/E highspeed centrifuge, Beckman, Palo Alto, CA). The concentration of Hb was determined to 0.164 g/mL using Drabkin’s reagent from Sigma Diagnostics (St. Louis, MO). The conjugates were prepared by adding HHPA in acetonitrile to the Hb in 0.1 M phosphate buffer (pH 7.4) at room temperature. The pH was kept constant by addition of 0.1 M NaOH. The molar ratios of HHPA added to Hb were 40:1 (HbHHPA 1:40), 10:1 (Hb-HHPA 1:10), and 1:1 (Hb-HHPA 1:1). The solution was allowed to react for 24 h and was then dialyzed in PBS containing EDTA, for 2 days, and then in 50 mM NH4-
Chem. Res. Toxicol., Vol. 15, No. 4, 2002 563 HCO3, for 5 days. The buffer was changed daily, and 10 mL of each change was removed for quantification of HHP acid. The conjugates were stored at -20 °C until analysis. Digestion of Hb-HHPA with Pronase E. The conjugates were precipitated in 0.1 M HCl in acetone, washed 3 times with cold acetone, evaporated to dryness (SpeedVac AS 290; Savant, Farmingdale, NY), and dissolved in water. Pronase E was dissolved in 50 mM NH4HCO3 containing 5 mM CaCl2‚2H2O and added to the conjugates at a ratio of 1:10. The conjugates were incubated for 4 days at 37 °C during which one addition of Pronase E (Pronase E:Hb ratio of 1:10) was made daily, and finally the solutions were evaporated to dryness. Digestion of Hb-HHPA with Trypsin. The conjugates were precipitated in 0.1 M HCl in acetone, washed 3 times with acetone, evaporated to dryness, and dissolved in water. Trypsin was dissolved in 50 mM NH4HCO3 and added to the conjugates at a ratio of 1:50. The incubation was performed for 16-18 h at 37 °C after which the solutions were evaporated to dryness. Qualitative Analysis of Full-Length Hb-HHPA. Analyses of full-length Hb-HHPA conjugates were performed on a hybrid quadrupole-time-of-flight mass spectrometer (Q-TOF) (QSTAR; Applied Biosystems, Foster City, CA) with electrospray ionization (ESI) either on a nanospray source or on a turbo ionspray source coupled to a HPLC system (Hewlett-Packard 1050; Palo Alto, CA). The mass spectrometer was calibrated using CsI (m/z 132.9054) and the pentapeptide Sex pheromone inhibitor iPD1 (m/z 829.5398). Biospec Reconstruct from Applied Biosystems was used in order to reconstruct mass spectra of multiply charged ions to the zero-charge form. The algorithm was iterated 20 times. The conjugates were precipitated in 0.1 M HCl in acetone and then washed 3 times with cold acetone and were either analyzed through flow injection in a 0.2 mL/min acetonitrile/water flow containing 0.05% TFA or analyzed with nanospray. In the latter case, the conjugates were dissolved in water, acidified with TFA, and purified using C18-zip tips prior to analysis. The samples were eluted with a solution of 75% methanol, 15% water, and 10% formic acid. The Q-TOF mass spectra were collected between 600 and 1700 amu. Qualitative Analysis of Pronase E Digests. Pronase E digests of Hb-HHPA were separated by HPLC (Hewlett-Packard 1050; Palo Alto, CA) on a C18 column (2.1 mm i.d. × 250 mm; Vydac, Hesperia, CA), collected in fractions, and analyzed for adducts with nanospray Q-TOF in the MS as well as the MS/ MS mode. The mobile phases consisted of water (A) and acetonitrile (B), both containing 0.05% TFA. The separation was performed with a gradient elution of from 1% to 10% B between 1 and 20 min, from 10% to 20% B between 20 and 30 min, from 20% to 100% B between 30 and 40 min, and finally isocratic at 100% B between 40 and 50 min. The flow rate was 0.3 mL/min. Fractions were collected every minute in Eppendorf tubes containing 0.1 mL of 50 mM NH4HCO3 using Fraction Collector Frac-100 from Pharmacia Biotech (Uppsala, Sweden). One-third of each sample was taken for quantification of HHP acid, and the remains of the fractions were evaporated to dryness. The dry samples were dissolved in 20-40 µL of water, acidified with TFA, and purified using C18-zip tips prior to analysis with nanospray. Samples were eluted with a mixture of 75% methanol, 15% water, and 10% formic acid. The generated peptides were named using the nomenclature suggested by Roepstorff and Fohlman (13). HHPL and HHPV in the fractions were identified by comparison with pure, synthesized standards [retention time (tR) of HHPL 21 min and tR of HHPV 37 and 43 min (double peak)]. The HHPL and HHPV standards were further characterized with Q-TOF, yielding a specific mass fragmentation. Fragmentation of HHPL ([M+H]+ ion at m/z 301.17) yielded m/z 283.17, 255.16, 155.07, 147.11, and 84.08. Fragmentation of HHPV ([M+H]+ ion at m/z 272.15) yielded m/z 254.14, 226.14, 155.07, 154.09, 118.09, and 72.08. Qualitative Analysis of Tryptic Digests. Tryptic digests of Hb-HHPA were separated by HPLC on a C18 column (2.1 mm i.d. × 250 mm, Vydac), collected in fractions, and analyzed for adducts with nanospray Q-TOF in the MS as well as the
564
Chem. Res. Toxicol., Vol. 15, No. 4, 2002
MS/MS mode. The mobile phases were water (A) and acetonitrile (B), both containing 0.05% TFA. The gradient elution was from 5% to 40% B in 50 min, from 40% to 100% B between 50 and 60 min, and finally isocratic at 100% B between 60 and 90 min. The flow rate was 0.3 mL/min. Two minute fractions were collected in Eppendorf tubes containing 0.2 mL of 50 mM NH4HCO3 using the Fraction Collector Frac-100. One-third of each sample was removed for quantification of HHP acid, and the remains of the fractions were evaporated to dryness. Prior to analysis with nanospray, 20-40 µL of water acidified with TFA was added to the dry samples, and these were further purified using C18-zip tips. The samples were eluted with 50 or 75% methanol with 0.1 or 10% formic acid. The theoretical sizes of tryptic digestion products of human Hb were obtained by tools in ProMac, a program supplied by Applied Biosystems. The sequence of human hemoglobin was obtained from SwissProt. Hydrolysis of Conjugates. The Hb-HHPA conjugates were hydrolyzed with 0.05 M HCl to release HHP acid. [2H6]HHP acid was added as an internal standard. After hydrolysis for 2 h at 100 °C, the samples were evaporated to dryness. Finally, the samples to be analyzed with liquid chromatography-electrospray ionization mass spectrometry (LC-ESI-MS/MS) were dissolved in 1.0 mL of water/methanol (60:40), and the residues to be analyzed with gas chromatography/mass spectrometry (GC/MS) were derivatized (see below). Quantification with LC-ESI-MS/MS. Quantification of HHP acid was performed on a triple quadrupole mass spectrometer with ESI (API 3000; Applied Biosystems) coupled to an HPLC system from Perkin-Elmer (Norwalk, CT). The HHP acid was separated from the Hb-matrix on a C18 column (2.1 mm i.d. × 50 mm, Genesis; Jones Chromatography, Lakewood, CO). The mobile phase consisted of water (A) and methanol (B), both containing 0.5% acetic acid. The separation was carried out using a linear gradient between 40% and 80% B in 4.5 min. The injection volume was 5 µL, and the flow rate 0.2 mL/min. The MS analysis was carried out using multiple reaction monitoring (MRM) in the negative ion mode at m/z 171.0/126.9 for HHP acid and at m/z 177.0/133.0 for [2H6]HHP acid. The temperature was set to 350 °C and the ion spray voltage to -2000 V. Pure nitrogen was used as curtain, turbo ionspray, collision and nebulizer gas. All samples were analyzed twice, and the mean value was calculated. Concentrations were determined by peak area ratios of HHP acid and [2H6]HHP acid. Quantification with GC/MS. Some samples, mainly with low levels of HHP acid, were derivatized as previously described (14) by adding 0.25 mL of 0.13 M PFBBr and 0.25 mL of 0.1 M TBA to the dry residues. The test tubes were placed in an ultrasonic bath for 90 min. Then 2 mL aliquots of hexane were added, and the samples were vortexed for 15 min and thereafter centrifuged for 10 min at 1500g (Sigma 3E-1, Deisenhofen, Germany). The samples were placed in a -20 °C freezer until the aqueous phase had frozen. The organic phases were transferred to new test tubes and evaporated to dryness in a flow of nitrogen. The samples were dissolved in 0.1 mL of toluene and transferred to injection vials. Analysis was performed on a VG Trio 1000 MS (Fisons, Manchester, U.K.) coupled to a Carlo-Erba 8065 GC equipped with an A200S autosampler (Carlo-Erba, Milan, Italy). The conditions of the GC/ MS analysis were as described by Lindh et al. (14). Concentrations were determined by peak area ratios of HHP acid and [2H6]HHP acid.
Results Detection Limit and Precision of LC-ESI-MS/MS Method. HHPA was hydrolyzed to HHP acid in water and further diluted to prepare a standard curve. The same amount of [2H6]HHP acid and HCl (final concentration of 0.05 M) was added to each sample. The samples were evaporated to dryness and dissolved in 1 mL of methanol/water (40:60). The detection limit was determined as the concentration giving a peak height of 3
Kristiansson et al.
Figure 1. Electrospray quadrupole-time-of-flight mass spectrometric analysis of the full-length conjugate of human HbHHPA. Reconstructed mass spectrum of (a) Hb-HHPA at a molar ratio of 1:40 and (b) Hb-HHPA at a molar ratio of 1:1.
times the noise and was found to be 0.5 ng/mL. The within-day precision was determined through preparation of 10 samples of Hb-HHPA tryptic digests at 2 different concentrations of HHP acid (4 and 40 ng/mL). The samples were hydrolyzed as described above. Analysis of each sample was performed twice, and the mean was determined. The standard deviation of 10 mean values was used in order to calculate the coefficient of variation (CV). The CV for the higher concentration was 5% and for the lower concentration 3%. Analysis of Full-Length Hb-HHPA. Conjugates were synthesized by adding HHPA in acetonitrile to Hb at pH 7.4, at room temperature, and allowed to react for approximately 24 h. The samples were dialyzed to remove hydrolyzed HHP acid. Quantitative analysis of HHP acid in the dialysis buffer showed that 91, 76, and 76% of the added HHPA was found in the dialysis buffer as HHP acid for Hb-HHPA 1:40, 1:10, and 1:1, respectively, whereas, when performing quantitative analysis of hydrolyzed HHP acid from synthesized conjugates, 11, 18, and 15% of added HHPA was bound to Hb for Hb-HHPA 1:40, 1:10, and 1:1, respectively. Analyses of full-length Hb-HHPA conjugates with flow injection and nanospray Q-TOF were performed in order to determine the number of adducts formed for each conjugate. Spectra with multiply charged ions were obtained and reconstructed using the program Biospec Reconstruct obtained from Applied Biosystems. The mass spectra of Hb-HHPA 1:40 and 1:1 are shown in Figure 1. The masses of 15 126 and 15 867 amu corresponded to the R- and the β-chain of Hb, respectively. Additional masses of 154 amu were obtained from formation of HHPA-adducts. The reconstructed mass spectra showed up to six adducts of HHPA on the R-chain
Hemoglobin Adducts of Hexahydrophthalic Anhydride
Chem. Res. Toxicol., Vol. 15, No. 4, 2002 565 Table 1. Summary of Modified Amino Acids Found in Hb-HHPA 1:40 Pronase E Digests Analyzed with Q-TOF-MS
Figure 2. LC-ESI-MS/MS quantification of HHP acid in fractions of Hb-HHPA 1:40 Pronase E digests separated with HPLC on a C18 column. Fractions were collected, and parts of each sample were hydrolyzed with HCl to obtain HHP acid. LCESI-MS/MS settings: mobile phase consisted of water (A) and methanol (B), both containing 0.5% acetic acid. The separation was carried out using a linear gradient between 40% and 80% B in 4.5 min at a flow rate of 0.2 mL/min. The MS analysis was carried out with multiple reaction monitoring (MRM) in the negative ion mode at m/z 171.0/126.9 (HHP acid) and at m/z 177.0/133.0 ([2H6]HHP acid).
Figure 3. GC/MS quantification of HHP acid in fractions of Hb-HHPA 1:1 Pronase E digests separated with HPLC on a C18 column. Fractions were collected, and parts of each sample were hydrolyzed with HCl to obtain HHP acids which were derivatized using PFBBr. GC/MS settings: selected ion monitoring of dipentafluorobenzyl hexahydrophthalate (PFB-HHP) at m/z 153 and 351 and of PFB-[2H6]HHP at m/z 159 and 357.
and five on the β-chain for conjugate Hb-HHPA 1:40, four adducts on the R-chain and four on the β-chain for conjugate Hb-HHPA 1:10 (data not shown), and two adducts on the R-chain and two on the β-chain for conjugate Hb-HHPA 1:1. Analysis of Hb-HHPA Pronase E Digests. The HbHHPA conjugates were digested with Pronase E at 37 °C for 4 days and then separated using HPLC. Approximately the same amount of Hb (0.1 mg) was used for each conjugate. Fifty fractions were collected and analyzed quantitatively and qualitatively for adducts. Quantitative analysis of Hb-HHPA 1:40 with LC-ESIMS/MS (Figure 2) and Hb-HHPA 1:10 and Hb-HHPA 1:1 with GC/MS (Figure 3) showed three distinct fractions containing HHP acid. Although the retention times of the digests varied slightly for the three conjugates, the pattern of three major fractions containing HHP acid after hydrolysis remained the same. The recoveries of the analyses were 101, 99, and 107% for conjugates HbHHPA 1:40, 1:10, and 1:1, respectively. Fractions with quantifiable amounts of HHP acid after hydrolysis were further analyzed with Q-TOF. The adducts found in the different fractions are shown in
fraction no.
m/z
23 24 25 26 27 28 33 37 38 43
301.18 301.17 301.18 301.18, 402.22 358.19 358.19, 372.21 569.28 254.13, 272.15 254.13, 272.15 254.14, 272.15
position of adduct
β K59
sequence K+HHPA K+HHPA K+HHPA K+HHPA, KT+HHPA KG/GK+HHPA KG/GK+HHPA, KA+HHPA GNPK+HHPA V+HHPA V+HHPA V+HHPA
Table 1. The mass spectrum of fraction 24 of the HbHHPA 1:40 Pronase E digest showed an [M+H]+ ion at m/z 301.17, and MS/MS yielded fragments of m/z at 283.17, 255.16 (immonium ion of lysine with HHPA), 155.07, 147.11, and 84.08 (immonium ion of lysine), consistent with the fragmentation of the HHPL standard. Mass spectra of fractions 37, 38, and 43 all contained [M+H]+ ions of m/z 272.15 corresponding in mass to HHPV. In addition, ions were found at m/z 254.14 corresponding to loss of water from the [M+H]+ ion. MS/ MS of the ion at m/z 272.15 yielded fragments of m/z 254.14, 226.14 (immonium ion of valine with HHPA), 155.07, 154.09, 118.09, and 72.08 (immonium ion of valine). The mass spectra were in accordance with the HHPV standard. Several fractions (23, 25, 26, 27, 28) with lower content of HHP acid also contained adducted lysine and/or adducted dipeptides of lysine and glycine, lysine and alanine, and lysine and threonine. Fraction 33 contained a singly charged HHPA-adducted tetrapeptide (m/z at 569.28) which was identified as β 56-59 through the y-series ions y1”* (m/z 301.17) to y4”* (m/z 569.28), all modified by 154 Da, and the b-series ions b2 (m/z 172.07) to b3 (m/z 269.12) and the 154 Da modified b4* (m/z 551.28), suggesting that HHPA was bound to K59. For Hb-HHPA 1:40, the lysine adducts made up 57% and the valine adducts 32% of the total amount of quantified HHP acid. The ratio of lysine and valine adducts was approximately constant for all conjugates, independent of the molar ratio of HHPA and Hb. Analysis of Hb-HHPA Tryptic Digests. To identify the positions of the HHPA-adducts in the globin sequence, the Hb-HHPA conjugates were digested with trypsin at 37 °C for 16-18 h. The tryptic digests were separated by HPLC and collected in 45 fractions that were analyzed qualitatively with nanospray Q-TOF and quantitatively with GC/MS after hydrolysis of the fractions. The quantifications of HHP acid for Hb-HHPA 1:40 and 1:1 are shown in Figure 4. The quantification of HbHHPA 1:10 showed the same pattern as Hb-HHPA 1:40 and 1:1 (data not shown). The recoveries of HHP acid were 103, 93, and 83% for Hb-HHPA 1:40, 1:10, and 1:1, respectively. The fractions containing hydrolyzable HHP acid were analyzed with Q-TOF. The results of these analyses are summarized in Tables 2-4. The quantification of HHP acid showed that adducted peptides eluted more or less in fractions 19-33 and all HHPA-adducts found in conjugates Hb-HHPA 1:10 and 1:1 were also found in conjugate Hb-HHPA 1:40. Since the most relevant conjugate for in vivo comparison would be the one with the lowest molar ratio of added HHPA, only
566
Chem. Res. Toxicol., Vol. 15, No. 4, 2002
Kristiansson et al. Table 3. Adducted Peptides from Hb-HHPA 1:10 Tryptic Digests Analyzed with Nanospray Q-TOF-MS peptide
position of adduct
fraction no.
no. of charges
R 1-11+HHPA R 1-11+2HHPA R 1-16+HHPA R 1-16+2HHPA
V1 V1, K7 V1 V1, K7/K11
19, 20 24 20 25
2 2 3 2, 3
R 12-31+HHPA R 32-56+HHPA
K16 K40
20, 21, 22 27
3 3,4
β 1-8+HHPA β 1-17+HHPA β 9-30+HHPA
V1 V1 K17
19, 20 25 26
2 3 2,3
24 25 21, 22
3 3 2
β 41-61+HHPA K59 β 66-82+HHPA K66 β 133-146+HHPA K144
m/z 663.36 740.40 613.65 997.03, 665.02 733.02 1014.15, 760.87 553.78 674.03 1191.60, 794.74 814.06 651.35 802.42
Table 4. Adducted Peptides from Hb-HHPA 1:40 Tryptic Digests Analyzed with Nanospray Q-TOF-MS peptide
Figure 4. GC/MS quantification of HHP acid in fractions of Hb-HHPA tryptic digests separated with HPLC on a C18 column. (a) Hb-HHPA 1:40 tryptic digests and (b) Hb-HHPA 1:1 tryptic digests. Fractions were collected, and parts of each sample were hydrolyzed with HCl to obtain HHP acids which were derivatized using PFBBr. GC/MS settings: selected ion monitoring of dipentafluorobenzyl hexahydrophthalate (PFBHHP) at m/z 153 and 351 and PFB-[2H6]HHP at m/z 159 and 357. Table 2. Adducted Peptides from Hb-HHPA 1:1 Tryptic Digests Analyzed with Nanospray Q-TOF-MS peptide
position of adduct
fraction no.
no. of charges
m/z
R 1-11+HHPA R 1-11+2HHPA R 12-31+HHPA β 1-8+HHPA β 9-30+HHPA β 41-61+HHPA β 133-146+HHPA
V1 V1, K7 K16 V1 K17 K59 K144
19, 20 24 20 20 26 24 21
2 2 3 2 3 3 2
663.36 740.39 733.02 553.78 794.75 814.05 802.41
mass spectra of Hb-HHPA 1:40 adducts also found in HbHHPA 1:1 are described. Fraction 19 contained the highest level of hydrolyzable HHP acid for all three conjugates, and analysis with nanospray Q-TOF showed a doubly charged ion of m/z 663.35, suggesting HHPA bound to peptide R 1-11. Product ion scan of the [M+2H]2+ ion of m/z 663.35 generated strong singly charged b-series ions of b1* (m/z 254.13), b2* (m/z 367.22), and b3* (m/z 454.24) all modified by 154 Da, giving evidence of HHPA binding to the N-terminal valine of the R-chain (Figure 5). A complete, unmodified y”-series except for the last y”-ion confirmed the peptide sequence. A weak [M-154+2H]2+ ion at m/z 586.33 was detected showing the loss of HHPA from the peptide. The formation of [M-154+nH]n+ ions was a characteristic feature for all adducted peptides analyzed in the MS/MS mode. In fraction 20, molecular ions were found at m/z 553.78 and 733.02. Product ion scan of the [M+2H]2+ ion of m/z
position of adduct
fraction no.
no. of charges
R 1-7+HHPA R 1-11+HHPA R 1-11+2HHPA R 1-16+2HHPA
V1 V1 V1, K7 V1, K7/K11
R 1-16+3HHPA
V1, K7, K11 25
2,3
R 1-31+3HHPA
V1, K7, K16 24
3,4
R 12-31+HHPA R 32-56+HHPA
K16 K40
20 23
3 3,4
β 1-8+HHPA β 1-17+HHPA
V1 V1
19, 20 21, 22
2 2,3
β 1-17+2HHPA
V1, K8
24, 25
2,3
β 1-30+HHPA
V1
23
3,4
β 9-30+HHPA
K17
23, 24, 25
2,3
β 41-61+HHPA
K59
22
2,3
22 20
3 2
β 66-82+HHPA K66 β 133-146+HHPA K144
20 19, 20 22 22, 23
1 2 2 2,3
m/z 883.44 663.35 740.40 997.04, 665.03 1074.06, 716.37 1219.94, 915.20 733.02 1014.16, 760.87 553.78 1010.53, 674.03 1087.56, 725.37 1105.90, 829.68 1191.61, 794.74 1220.60, 814.06 651.36 802.42
Figure 5. Product ion spectrum of molecular ion [M+2H]2+ (m/z 663.35) of peptide R 1-11 with one HHPA-adduct on the N-terminal valine (fraction 20 of C18 separated Hb-HHPA 1:40 tryptic digests). Analysis with nanospray quadrupole- time-offlight mass spectrometry.
553.78 (suggesting HHPA bound to peptide β 1-8) showed binding of HHPA to the N-terminal valine of the β-chain due to the presence of b-series ions from b1* at
Hemoglobin Adducts of Hexahydrophthalic Anhydride
Figure 6. Product ion spectrum of molecular ion [M+3H]3+ (m/z 814.06) of peptide β 41-61 with one HHPA-adduct on K59 (fraction 22 of C18 separated Hb-HHPA 1:40 tryptic digests). Analysis with nanospray quadrupole-time-of-flight mass spectrometry.
m/z 254.13 to b4* at m/z 605.31, all modified by 154 Da, corresponding to HHPA bound to valine. An [M-154+2H]2+ ion at m/z 476.75 was also detected. The triply charged molecular ion of m/z 733.02 was sequenced to peptide R 12-31 with one HHPA-adduct by MS/MS. The b-series ions b2 (m/z 143.08) to b4 (m/z 386.18) were unmodified, and the rest of the b-series ions, b5* (m/z 668.33) to b13* (m/z 1452.69), were modified by 154 Da. This information as well as data from the doubly charged y”-series localized the adduct to R K16. These y”-series ions were unmodified from y6” (m/z 344.68) to y15” (m/z 765.37) and modified from y16”* (m/z 906.44) to y19”* (m/z 1063.51). A strong [M-154+3H]3+ ion at m/z 681.67 was found. Product ion scan of the [M+2H]2+ ion of m/z 802.42 in fraction 21 yielded the sequence of β 133-146 with one adduct of HHPA bound to residue K144. The sequence was obtained from the b-series ions of unmodified b2 (m/z 199.14) to b11 (m/z 1003.55; b9 missing) and b12* (m/z 1285.70) to b13* (m/z 1448.75) modified by 154 Da. The y”-series ions from y1” (m/z 156.07) to y2” (m/z 319.13) were unmodified whereas y3”* (m/z 601.28) to y12”* (m/z 1405.71) were modified by 154 Da. An [M-154+2H]2+ ion at m/z 725.39 was found. The Q-TOF spectrum of fraction 24 showed two ions at m/z ratios corresponding to peptides with matching molecular weights of adducted peptides. The first was a triply charged ion at m/z 814.06, suggesting β 41-61 with one HHPA-adduct. Product ion scan of this ion gave a spectrum of an [M-154+3H]3+ ion at m/z 762.71 and doubly charged y”-series ions from y4”* (m/z 313.20) to y19”* (m/z 1073.53) (y10”* missing), agreeing with the amino acid sequence of β 43-58 plus an additional mass of 154 Da (Figure 6). This indicated that the adduct was located on either of the residues in the sequence β 5861. The y”-series ions y1” (m/z 147.11) to y11”* (m/z 1309.70) confirmed the sequence of β 51-61 with an additional mass of 154 Da from y3”* (m/z 528.33) to y11”*, making it reasonably to assume that HHPA was bound to K59. Also, the b-series ions b2 (m/z 295.14) to b17 (m/z 1815.81) and the doubly charged b-series b10 (m/z 566.26) to b17 (m/z 908.40) and b19* (m/z 1098.01) to b20* (m/z 1147.54) ions confirmed the predicted sequence with an HHPA-adduct on β K59. The other molecular ion in fraction 24 was a doubly charged ion of m/z 740.40 which corresponded to R 1-11 modified by two HHPA molecules. Product ion scan yielded an [M-154+2H]2+ ion at m/z 663.37, and the 154
Chem. Res. Toxicol., Vol. 15, No. 4, 2002 567
Da modified b-series ions b1* (m/z 254.14) to b6* (m/z 737.37) and the 308 Da modified b7** (m/z 1019.54) to b8** (m/z 1120.59) determined the sites of modification to V1 and K7. An almost complete y”-series supported this determination by unmodified y1” (m/z 147.11) to y4” (m/z 461.27), and ions modified by 154 Da from y5”* (m/z 743.43) to y10”* (m/z 1226.66). MS/MS of the molecular ion [M+3H]3+ of m/z 794.74 found in fraction 26 yielded the sequence of β 9-30 with one HHPA bound to β K17. The information was obtained from the b-series ions which were unmodified from b2 (m/z 159.08) to b8 (m/z 786.41) and modified by 154 Da from b9* (m/z 1068.58) to b14* (m/z 1624.83). The y”-series ions were unmodified from y1 (m/z 175.12) to y13 (m/z 1314.67). The doubly charged b-series ions ranged from b9* (m/z 534.80) to b20* (m/z 1076.05), all modified by 154 Da. In addition, the doubly charged y”-series ions were obtained from y8” (m/z 379.71) to y19”* (m/z 1063.06), and were modified by 154 Da from y14”* (m/z 798.93). Also, an [M-154+3H]3+ ion at m/z 743.38 was found.
Discussion This study shows that HHPA, when added to Hb in vitro, forms adducts with N-terminal valine and several lysine residues of both the R- and the β-chain. The characterization of HHPA-adducted proteins is of importance for two reasons. The main reason for such studies is to obtain information on allergenic chemical structures but may also be used for generation of methods for biological monitoring. Natural protein allergens have been studied by, e.g., Aalberse (1) in an attempt to establish some structural characteristics responsible for the allergenicity. However, their conclusion was that allergens are structurally heterogeneous, and therefore they were unable to deduce any specific allergenic structures. Our study was based on conjugates between an endogenous protein and HHPA. Such protein-OAA conjugates have been shown to be immunologically active (10). Studies of the changes when endogenous, nonallergenic proteins react with a low molecular weight compound and thereby form allergens will generate important information on allergenic chemical structures. In this study, we have developed a quantitative LCESI-MS/MS method of measuring HHP acid using deuterium-labeled HHP acid as an internal standard. Such a quantitative method offers the advantage of actually knowing the amount of adducts formed. Protein adducts of a number of different compounds have been structurally characterized in several other studies. However, only in a few cases have the adducts been analyzed in a quantifiable manner such as measurement of Hb and albumin adducts of benzo[a]pyrene diolepoxide (15) and detection of adducts to human Hb and albumin after incubation with phosgene (16). In those cases, the quantification was achieved through UV detection of chromophore compounds at a specific wavelength and radiometric detection of 14C-labeled phosgene, respectively. The total amount of HHP acid found in the dialysate and bound to the Hb was close to 100% of the total HHPA used in the synthesis of the conjugates. Thus, it is evident that almost all HHPA adducts are hydrolyzed from the protein using 0.05 M HCl for 2 h at 100 °C. This is in accordance with the results from Lindh and Jo¨nsson (4).
568
Chem. Res. Toxicol., Vol. 15, No. 4, 2002
In the fractions containing hydrolyzable HHPA, adducts only with valine and lysine were found. This is in accordance with results from earlier studies on cyclic acid anhydrides using photospectrometric methods (17). Other amino acid residues known to be reactive to acid anhydrides are cysteine, tyrosine, histidine, serine, and threonine. We did not find conjugation to any of these amino acids in our study. A probable explanation is that the conjugates of cysteine, tyrosine, and histidine are unstable whereas the reactivity to serine and threonine is too low in comparison with lysine and the N-terminal valine. In the analysis of full-length Hb-HHPA, we found six adducts on the R-chain and five on the β-chain. We also found a total of 11 specific adducts in the study of the tryptic peptides. Thus, it seems reasonable to assume that the majority of the adducts formed have been mapped. The amount of HHPA binding to the lysine residues was only about twice as high as that for the N-terminal valine despite the fact that there is 1 N-terminal valine but 11 lysine residues on each globin chain. Thus, the reaction rate for HHPA with the N-terminal valine is higher than for the lysine residues. This is expected since the pKa for the R-amino group on valine is about 2 units lower than for the -amino group of lysine. There was no major difference in the binding to the amino acids regardless of the molar ratio of HHPA added. This indicates that the binding of HHPA, even at lower molar ratios, may be similar to the results described here. The HPLC separation of Pronase E digests of HbHHPA as well as valine-HHPA standards resulted in two peaks of HHPA-adducted valine. This was probably because either of the two acyl carbons of HHPA could undergo nucleophilic substitution by the valine R-amine group or the lysine -amine group, thus generating two diastereomers with different properties. The HHPA-adducted peptides were monitored, and the positions of the adducts were determined through the band y”-series ions. Besides this, evidence supporting the presence of an HHPA-adduct was also obtained by the loss of the mass of HHPA from the parent ion when performing MS/MS. The size of the peak of the peptide ion losing HHPA varied depending on the peptide analyzed, possibly depending on the stability of the HHPAadduct in the amino acid sequence. The quite tedious and time-consuming way of detecting adducted peptides through nanospray analysis of collected fractions instead of LC/MS is compensated by several advantages. The quantification makes it possible to limit the analyses to only those fractions containing adducts which can then be thoroughly analyzed with MS and MS/MS. Since nanospray has a very low flow rate, on the order of nanoliters per minute, one can utilize the sample to its fullest, performing several MS/MS experiments. If LC/MS was to be used instead, the flow would have to be splited in order to quantify the content of HHP acid. Since the split ratio seldom is stable, this would lead to a loss of accuracy regarding the recovery of HHP acid. The trypsin was not able to cleave the protein at lysine residues adducted with HHPA. All peptides modified on the lysine residues therefore consisted of at least two theoretically calculated tryptic peptide sequences, which was of help in identifying the adducts. Similar results were also observed by others (18).
Kristiansson et al.
It is questionable whether HHPA-adducted Hb takes part in the pathophysiological mechanism since hemoglobin is shielded by the erythrocyte walls. Hemoglobin has been used mainly as a model protein for future studies of more relevant endogenous proteins. However, in a study of lung injuries in rats exposed to trimellitic anhydride, alveolar spaces filled with modified proteins, red blood cells, and macrophages ingesting crystal-like hemoglobin were found (19). This suggests that Hb can possibly be involved in the immune response in some cases. Once the primary structure modifications due to HHPA are elucidated, future work would be to evaluate known adducted peptides in cell systems with T-cells; e.g., the amount or expression of cytokines produced could be a measure of the potential ability of each peptide to trigger immune responses. If some adducted peptides would be more relevant than others, this or these peptides could be used as biomarkers as an alternative to hydrolysis of the total amount of adducted proteins. Thus, in this way we would gain both theoretical structural information but also obtain practical use of our results.
Acknowledgment. This work was supported by the Swedish Medical Research Council (Project 11336), the Swedish Council for Work Life Research, AMF Insurance, the Vårdal Foundation, the European Union (Project BMH4-CT-98-3951), the Swedish Council for Planning and Coordination of Research, and the Medical Faculty at Lund University.
References (1) Aalberse, R. C. (2000) Structural biology of allergens. J. Allergy Clin. Immunol. 106, 228-238. (2) Jo¨nsson, B. A. G., Wishnok, J. S., Skipper, P. L., Stillwell, W. G., and Tannenbaum, S. R. (1995) Lysine adducts between methyltetrahydrophthalic anhydride and collagen in guinea pig lung. Toxicol. Appl. Pharmacol. 135, 156-162. (3) Day, B. W., Jin, R., and Karol, M. H. (1996) In vivo and in vitro reactions of toluene diisocyanate isomers with guinea pig hemoglobin. Chem. Res. Toxicol. 9, 568-573. (4) Lindh, C. H., and Jo¨nsson, B. A. G. (1998) Human hemoglobin adducts following exposure to hexahydrophthalic anhydride and methylhexahydrophthalic anhydride. Toxicol. Appl. Pharmacol. 153, 152-160. (5) Venables, K. M. (1989) Low molecular weight chemicals, hypersensitivity, and direct toxicity: the acid anhydrides. Brit. J. Ind. Med. 46, 222-232. (6) Nielsen, J., Welinder, H., Horstmann, V., and Skerfving, S. (1992) Allergy to methyltetrahydrophthalic anhydride in epoxy resin workers. Brit. J. Ind. Med. 49, 769-775. (7) Yokota, K., Johyama, Y., Yamaguchi, K., Fujiki, Y., Takeshita, T., and Morimoto, K. (1997) Specific antibodies against methyltetrahydrophthalic anhydride and risk factors for sensitization in occupationally exposed subjects. Scand. J. Work, Environ. Health 23, 214-220. (8) Gell, P. G. H., and Coombs, R. R. A. (1963) The classification of allergic reactions underlying disease. In Clinical aspects of immunology (Gell, P. G. H., and Coombs, R. R. A., Eds.) pp 317337, Blackwell, Oxford, U.K. (9) Welinder, H., Zhang, X. D., Gustavsson, C., Bjo¨rk, B., and Skerfving, S. (1995) Structure-activity relationships of organic acid anhydrides as antigens in an animal model. Toxicology 103, 127-136. (10) Zhang, X. D., Welinder, H., Jo¨nsson, B. A. G., and Skerfving, S. (1998) Antibody responses of rats after immunization with organic acid anhydrides as a model of predictive testing. Scand. J. Work, Environ. Health 24, 220-227. (11) Welinder, H. E., Jo¨nsson, B. A. G., Nielsen, J. E., Ottosson, H. E., and Gustavsson, C. A. (1994) Exposure-response relationships in the formation of specific antibodies to hexahydrophthalic anhydride in exposed workers. Scand. J. Work, Environ. Health 20, 459-465.
Hemoglobin Adducts of Hexahydrophthalic Anhydride (12) Nielsen, J., Welinder, H., Jo¨nsson, B. A. G., Axmon, A., Rylander, L., and Skerfving, S. (2001) Exposure to hexahydrophthalic and methylhexahydrophthalic anhydridessdose-response for sensitization and airways effects. Scand. J. Work, Environ. Health 27, 327-334. (13) Roepstorff, P., and Fohlman, J. (1984) Proposal for a common nomenclature for sequence ions in mass spectra of peptides. Biomed. Mass Spectrom. 11, 601. (14) Lindh, C. H., and Jo¨nsson, B. A. G. (1997) Determination of hexahydrophthalic acid and methylhexahydrophthalic acid in plasma after derivatisation with pentafluorobenzyl bromide using gas chromatography and mass spectrometric detection. J. Chromatogr. 691, 331-339. (15) Helleberg, H., and To¨rnqvist, M. (2000) A new approach for measuring protein adducts from benzo[a]pyrene diolepoxide by high performance liquid chromatography/tandem mass spectrometry. Rapid Commun. Mass Spectrom. 14, 1644-1653.
Chem. Res. Toxicol., Vol. 15, No. 4, 2002 569 (16) Noort, D., Hulst, A. G., Fidder, A., van Gurp, R. A., de Jong, L. P. A., and Benschop, H. P. (2000) In vitro adduct formation of phosgene with albumin and hemoglobin in human blood. Chem. Res. Toxicol. 13, 719-726. (17) Palacia´n, E., Gonza´lez, P. J., Pin˜eiro, M., and Herna´ndez, F. (1990) Dicarboxylic acid anhydrides as dissociating agents of protein-containing structures. Mol. Cell. Biochem. 97, 101-111. (18) Moll, T. S., Harms, A. C., and Elfarra, A. A. (2000) A comprehensive structural analysis of hemoglobin adducts formed after in vitro exposure of erythrocytes to butadiene monoxide. Chem. Res. Toxicol. 13, 1103-1113. (19) Zeiss, C. R., Leach, C. L., Smith, L. J., Levitz, D., Hatoum, N. S., Garvin, P. J., and Patterson, R. (1988) A serial immunologic and histopathologic study of lung injury induced by trimellic anhydride. Am. Rev. Respir. Dis. 137, 191-196.
TX0155911