Chem. Res. Toxicol. 1994, 7, 829-835
829
Enantiospecificity of Covalent Adduct Formation by Benzo[a]pyrene anti-Diol Epoxide with Human Serum Albumin Billy W. Day,*>+ Paul L. Skipper,$ Joseph Zaia,g Kuldip Singh," and Steven R. Tannenbaum$>§ Departments of Environmental & Occupational Health and Pharmaceutical Sciences, University of Pittsburgh, Pittsburgh, Pennsylvania 15238, and Department of Chemistry, Division of Toxicology, and George R. Harrison Spectroscopy Laboratory, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139 Received May 3, 1994@ Human serum albumin was reacted with the (+I- and (-)-enantiomers of r-7,t-8-dihydroxyt-9,t-10-epoxy-7,8,9,1O-tetrahydrobenzo[alpyrene to determine if the chiral nature of the protein influences adduct formation. The alkylated proteins were analyzed directly by fluorescence line narrowing spectroscopy, and their spectra were compared to those of the model synthetic adducts Nz-(7,8,9-trihydroxy-r-7,t-8,t-9,c-l0-tetrahydrobenzo[alpyren-lO-yl)histidine and 7,8,9trihydroxy-r-7,t-8,t-9,c-1O-tetrahydrobenzo[alpyren-lO-yl N-t-BOC-alaninate ester. The results from these analyses indicated that different adducts were formed by the enantiomers of the diol epoxide. The adducted proteins were also enzymatically digested, and the 8,9-cis-dihydroxy7,8,9,10-tetrahydrobenzo[alpyrene-containing adducts and hydrolysis products were isolated by boronate affinity chromatography. Diode array W, fast atom bombardment, and on-line atmospheric pressure ionization-mass spectral analysis of the HPLC purified products indicated that the more mutagenic and tumorigenic (+)-enantiomer forms carboxylic ester adducts with the protein at either Asp(187) or Glu(1881, while the (-)-enantiomer forms Nrhistidine adducts at His(146). This previously unrealized enantiospecificity of the reaction of benzo[alpyrene anti-diol epoxide with human serum albumin has important consequences for the application of the adducts as biomarkers of internal exposure.
Introduction
spectroscopy (5) or as partially or completely degraded adducts by fluorescence spectroscopy (6, 7)) 32P-postlaBenzo[alpyrene (BaPY is converted to several oxygenbeling procedures (8)) immunological assays (9)) and/or ated metabolites by enzymatic systems in vivo. The mass spectrometric methods (6, 7). metabolic path in humans whereby the 7 and 8 carbons The ultimate absolute configuration of anti-BaPDE is are oxidized by cytochrome P450 l A l , the resulting determined in the first oxidative step of BaP metabolism, epoxide is opened by epoxide hydrolase to yield the 7,8with subsequent hydrolysis and oxidation giving rise to dihydrodiol, which is further oxidized a t the 9 and 10 the various diastereomeric forms. Rat liver microsomes carbons by cytochrome P450 3A4 or the cytochromes produce the 7R, 8R-dihydrodiol of BaP in high enantioP450 1 (1-41, gives one of the most studied ultimate carcinogenic forms of the parent hydrocarbon: r-7,t-8meric excess, while human liver microsomes exhibit less dihydroxy-t-9,t-10-epoxy-7,8,9,10-tetrahydrobenzo[a]py- stereoselectivity in the formation of this dihydrodiol(28rene (anti-BaPDE). anti-BaPDE reacts with nucleophiles 56% ee) (10). Human skin produces both the (-)-R,Ron biological macromolecules to form covalent products, and (+I-S,S-enantiomers of the 7,8-dihydrodiol(11). (+Iwhich may be measured in biological samples from anti-BaPDE, which arises from the 7R,8R-dihydrodio19 environmentally exposed humans as intact DNA or is considered to be more mutagenic and tumorigenic than protein adducts by fluorescence line narrowing (FLN) the (-)-isomer in mammalian systems (12, 13). (-)-antiBaPDE, a minor product in the rat, arises from metabo* Author to whom correspondence should be addressed a t Center lism of the 7S,8S-dihydrodiol. The primary product of for Environmental and Occupational Health and Toxicology, University metabolism of the 7S,8S-dihydrodiol, again in the rat, is of Pittsburgh, 260 Kappa Dr., Pittsburgh, PA 15238;e-mail address:
[email protected]. the syn-diol epoxide. Methods to determine the origin +Universityof Pittsburgh. of the products arising from reaction of the enantiomers t Division of Toxicology, MIT. of anti-BaPDE with DNA adducts have been reported 8 Department of Chemistry, MIT. George R. Harrison Spectroscopy Laboratory, MIT. Present ad(14-18). There has been a report of a study of the dress: Department of Chemistry, University of Missouri-St. Louis. isolated enantiomers in protein-containing systems (13)) Abstract published in Advance ACS Abstracts, October 15, 1994. Abbreviations: anti-BaPDE, r-7,t-8-dihydroxy-t-9,t-lO-epoxy-7,8,9,- but the determination of enantioselectivity of a specific 10-tetrahydrobenzo[alpyene; M I , atmospheric pressure ionization; BaP, benzo[alpyrene; BaP triolyl, ~-7,t-8,t-9-trihydroxy-7,8,9,10-tet- protein toward diol epoxide adduct formation has not been examined. rahydrobenzo[alpyren-10-yl; cis and trans BaP tetrahydrotetrols, r-7,t8,t-9,t-10- and r-7,t-8,t-9,c-l0-tetrahydroxy-7,8,9,1O-tetrahydrobenzoRecent studies indicate that diol epoxide-serum al[alpyrenes; DAD, diode array detector; ee, enantiomeric excess; FAB, fast atom bombardmenc FLN, fluorescence line narrowing; ionspray, bumin adducts may be formed in humans exposed to pneumatically-assisted electrospray; PBS, 10 mM s o d i d p o t a s s i u m polycyclic aromatic hydrocarbons (19-21 ). As for invesphosphate buffered normal saline, pH 7.4;t-BOC, tert-butyloxycarbonyl. tigations which have identified the adducting xenobiotic @
0893-228x/94/2707-0829$04.50/00 1994 American Chemical Society
Day et al.
830 Chem. Res. Toxicol., Vol. 7, No. 6, 1994
The instrumental arrangement and conditions were as previously described (24). Fast atom bombardment (FAB) mass spectra were obtained a t the MIT Mass Spectrometry Facility as previously described (23) on a JEOL HX110/HX110 mass spectrometer using a JEOL Csf gun operated a t 26 kV. Tandem mass spectrometry was carried out using all four sectors in ElBlEzBz configuration. Protonated peptide adducts were selected with the first mass spectrometer, and collision-induced dissociation (CID) of them took place in the field-free region after B1. The collision cell potential was held a t 3 kV and the ion collision energies a t 7 kV. He was used as the collision gas, introduced a t a pressure sufficient to reduce precursor ion signal to 3035%. The first mass spectrometer was operated so that only the species of the M H analyte was transmitted. Pneumatically-assisted electrospray mass spectra were obtained on a Perkin ElmerlSciex API I mass spectrometer equipped with an atmospheric pressure ionization source and a n ionspray interface, linked in tandem to the HPLCDAD W system with glass capillary tubing. The ionspray interface was maintained a t 5 kV. HighMaterials and Methods purity air was used as the nebulizing gas and was Chemicals. Caution: The (+)- and (-)-enantiomers maintained a t an operating pressure of 40 psi. Analytes of anti-BaPDE are probable human carcinogens. They were introduced directly from the HPLC system (at 0.2 should be handled in accordance with NIH guidelines mumin) without splitting. The orifice voltage was (25). generally set at 70 eV, and high-purity Nz was used as (+)-, (-)-, and (f)-r-7,t-8-Dihydroxy-t-9,t-lO-epoxy- the curtain gas, flowing a t 0.6 Umin. The quadrupole 7,8,9,10-tetrahydrobenzo[a]pyrene were purchased from was scanned from m l z 280 to 1000 in 8.07 slscan a t a the National Cancer Institute Chemical Carcinogen resolution of mlz 0.1. Repository, maintained by the Midwest Research InstiAmino acid composition and peptide sequence analyses tute, Kansas City, MO. NT-(7,8,9-Trihydroxy-r-7,t-8,t-9,c-were performed by Dr. John Hempel in the Department l0-tetrahydrobenzo[a]pyren-lO-yl)histidinylprolyltyof Molecular Genetics and Biochemistry a t the University rosine, N~-(7,8,9-tnhydroxy-r-7,t-8,t-9,c-lO-~trahydrobenzoof Pittsburgh. Composition was determined by 6 N HCl [alpyren-10-yl)histidine, 7,8,9-trihydroxy-r-7,t-8,t-9,~-10-hydrolysis, phenylthiohydantoin derivative formation, tetrahydrobenzo[a]pyren-10-ylN-t-BOC-alaninate ester, and HPLC analysis. Sequences were determined by and the cis- and trans-BaP tetrahydrotetrols were preEdman degradation and HPLC analysis. pared from (f)-r-7,t-8-dihydroxy-t-9,t-lO-epoxy-7,8,9,10- Adduct Synthesis and Isolation. Two silanized tetrahydrobenzo[a]pyrene as previously described (23, glass vials equipped with Teflon stir bars and caps were 26). each charged with human serum albumin (305 mg, 4.6 Instrumentation and Analytical Methods. HPLC pmol, fatty acid-free, Sigma Chemical Co., St. Louis, MO) separations were performed with adaptations of methods in 7.25 mL of 10 mM sodiundpotassium phosphate previously described (23) using similar columns. The buffered normal saline, pH 7.4 (PBS). The solutions were instrument was a Hewlett Packard 1090 Series I1 liquid treated with either (+I-anti-BaPDE or (-)-anti-BaPDE chromatograph equipped with a Hewlett Packard 1040 (100 pg, 331 nmol) in 0.1 mL of 19:l THF-Et3N. The diode array U V detector. All analyses used a primary mixtures were capped, stirred a t 37 "C for 2 h, and then diode array detector monitoring wavelength of 343 nm. extracted with ethyl acetate (4 x 10 mL). Residual ethyl Semipreparative separations were performed on either acetate was removed by bubbling Nz gas through the a Whatman Partisil M9 10125 ODS-3 7.8 x 300 mm C-18 solutions. Aliquots (0.25 mL) of the solutions were column a t a flow rate of 4.2 m u m i n or a Waters Prep removed for analysis by FLN spectroscopy. Pronase E Nova-Pak HR 60 A 6 pm 7.8 x 300 mm C-18 column a t (5 mg, Sigma) was added to each vial, and the mixtures a flow rate of 3.2 mumin. Analytical separations were were stirred a t 37 "C for 12 h. The solutions were diluted achieved on a Hewlett Packard ODS Hypersil 5 pm 2.1 with 10 mL of PBS, adjusted to pH 8 with 1 N NaOH, x 100 mm column a t a flow rate of 0.2 mUmin. The and treated with an additional 5-mg aliquot of Pronase solvents used were 0.1% CF3C0&UH~0(solvent A) and E a t 37 "C for 6 h with stirring. The digest solutions 0.1% CF&OZWCH&N (solvent B). The gradient generwere cooled to room temperature, diluted to 50 mL with ally used for adduct/BaP tetrahydrotetrols separations 50 mM N H 4 0 ~ C Hand , adjusted to pH 8.5 with NH40H. was as follows: time = 0, 85% solvent A, linear change The solutions were passed through boronate cis-diol to 10 min, 75% A; linear change to 34 min, 72% A affinity gels (3 g each column, Mi-Gel-601, Bio-Rad, (gradient 1). An additional system was used in some of Richmond, CA) twice. The columns were washed with the LC-MS experiments: time = 0, 100% A; linear 50 mL of 50 mM NH402CH. Bound materials were change to 60 min, 60% B (gradient 2). Additional W eluted with 150 mL of 0.9% HCOzH, pH 3. Each eluate spectra were determined on a Hewlett Packard 8452A was quickly adjusted to pH 7 with NH40H and then diode array spectrophotometer. stored a t -70 "C. Fractions of these solutions were lyophilized prior to HPLC analysis. Dried materials were Fluorescence line narrowing spectral analyses were dissolved in a mixture of HzO-CH&N-O.l% CF3COzH performed a t the MIT Laser Biomedical Resource Center.
on a specific protein, it has been previously shown that human hemoglobin can be modified by anti-BaPDE and chrysene diol epoxide in vivo (6, 7,22). The introduction of an additional marker of internal exposure, in which the exact responsible chemical species is identifiable, to the armament of existing methods for analysis of polycyclic aromatic hydrocarbon adducts will be a useful tool in the correlation of external dose to internal dose and health outcome for these xenobiotics. We recently began studies of the underlying chemistry of anti-BaPDEhuman serum albumin adducts in order to have the knowledge necessary to use them as a biomarkers for internal exposure. We reported that the racemic form of anti-BaPDE forms both carboxylic acid ester and NrHis(146) adducts with the protein in vitro (23,241. The intent of the present study was to determine which carboxylic acid residue(s) were alkylated by anti-BaPDE and if the chiral discrimination exhibited by human serum albumin toward other covalent and noncovalent ligands also extended to covalent adduct formation with anti-BaPDE.
+
Chem. Res. Toxicol., Vol. 7, No. 6, 1994 831
Enantiospecific Benzo[alpyrene-Albumin Adduction
E
6 m L
m
h
Ioi
His-Pro-Tyr
His-Pro
A
5
10
15
20
25
Time (min) Figure 2. Reversed-phase HPLC (gradient 1, see experimental section) of materials bound by boronate affinity medium after Pronase E digestion of human serum albumin adducted by (-)anti-BaPDE in vitro. The peaks at 10.7,12.2,and 14.2 min correspond, respectively, to the His, His-Pro, and His-Pro-Tyr Nr-imidazolyl adducts of the diol epoxide (see Figure 4). to immobilized phenylboronic acid cis-diol affinity columns, which recognize and bind small peptides containing the anti-BaPDE-derived moiety under mild basic conditions. The columns were washed to remove the majority of unadducted peptides, and then the adducts synthetic histidine adduct were eluted under acidic conditions, neutralized, and lyophilized. Boronate-bound fractions from each reaction were analyzed by semipreparative HPLC using diode array UV detection. In agreement with the FLN results, 1300 1500 1700 1900 the UV spectra of the adducts displayed red shifts in the S2 0-0 band of the pyrene chromophore corresponding A Wavenumber (cm") to the expected nucleophilic heteroatom on the protein Figure 1. FLN spectra of human serum albumin adducted in that was covalently bound to the C-10 atom of the r-7,tvitro with the enantiomers of anti-BaPDE and the synthetic 8,t-9-trihydroxy-7,8,9,lO-tetrahydrobenzo[alpyrene model ester and histidine adducts. Spectra were recorded at 4 moiety: a 2-nm red shift from the BaP tetrahydrotetrols K, and the laser excitation wavelength was 355 nm. Samples were dissolved in 5:41glycerol-H~O-C&CH~OH.The structureS2 0-0 ,Imm was noted in the adducts from the (+Idefining similarities and differences in the FLN spectra of enantiomer, while a 4-nm red shift was noted in the protein and synthetic adducts are the relative intensities of the adducts from the (-)-enantiomer. Determination of the zero phonon lines at the excited state vibrational frequencies, S2 W ,I,,,= of the adducts from the (+)-enantiomer under which are reproducible. near neutral conditions (1:1 MeOH-H20) showed that the shift was not altered by a change in solvent/pH. corresponding to the initial composition of the HPLC Carboxylic acid esters from anti-BaPDE have a 2-nm red mobile phase, filtered, and analyzed by HPLC and/or shift in the S2 0-0 band from that seen with BaP LC-MS. Collected fractions were dried on a speedvac. tetrahydrotetrols in both acidic (CH3CN-H20-0.1% CF3Certain fractions were analyzed by DAD-UV in 1:l C02H) and near neutral (1:lMeOH-HzO) solvents, while MeOH-H20, and all were peracetylated as described alkylamine adducts of anti-BaPDE show a 2-nm red shift previously (23) for FAB-MS analysis. in the acidic solvent and no shift in the near neutral solvent (26). Histidine imidazolyl-anti-BaPDE adducts Results are unique by virtue of their 4-nm red shift (23). Additionally, only the enzymatically generated hydrolysate Human serum albumin was alkylated with the enanfrom the reaction of the (+)-enantiomer with human tiomers of anti-BaPDE in vitro under physiological serum albumin yielded BaP tetrahydrotetrols, products conditions a t a molar ratio of 14:l protein to epoxide. The which arise only from the enzymatic hydrolysis of ester protein solutions were extracted to remove BaP tetrahyadducts. The chromatograms are shown in Figures 2 and drotetrols which arise from hydrolysis of the epoxides and 3. from unstable adducts as well as any unreacted diol epoxide. An aliquot from each reaction mixture was HPLC fractions corresponding to adducts were analyzed by FLN spectroscopy. Comparison of spectra collected and analyzed further by FAB-MS. Since from each enantiomer-protein pair with those from the underivatized small peptide adducts of anti-BaPDE colsynthetic adducts Nr-(7,8,9-trihydroxy-r-7,t-8,t-9,c-lO-tet-lected in such a fashion do not yield easily interpretable rahydrobenzo[alpyren-10-y1)histidineand 7,8,9-trihydroxy- FAB mass spectra (23), each of the fractions was first r-7,t-8,t-9,c-10-tetrahydrobenzo[a]pyren-lO-y1 N-t-BOCperacetylated by treatment with acetic anhydridelpyrialaninate ester (Figure l ) , as well as those from BaP dine at room temperature. Only the chromatographic tetrahydrotetrols (data not shown; see ref 241, strongly fractions from the reaction of the (-)-enantiomer yielded suggested that (-)-anti-BaPDE formed only "-histidine results corresponding to anti-BaPDE peptide adducts. adducts, while the (+)-enantiomer formed only carboxylic The spectra and CID analysis of protonated molecular acid ester adducts. ions were identical to those we had previously noted (23) The remainder of the protein samples were separately from reaction of racemic anti-BaPDE, namely, BaP digested with Pronase E. The digests were then applied triolyl-His, BaP triolyl-His-Pro, and BaP triolyl-His-Pro-
/$
r/1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
Day et al.
832 Chem.Res.Toxicol.,Vol. 7,No. 6, 1994
I
+
1 Asp-Glu-Gly-Lys
HI+ ions were determined to originate from (+)-antiBaPDE-adducted peptides by virtue of the presence of a co-eluting m l z 303 ion, which corresponds to the BaP triolyl cation (also noted as the base ion for the BaP tetrahydrotetrols, i.e., the BaP tetrahydrotetrol [M H - Hz0]+ ion), a co-eluting ion corresponding to the [peptide portion of the adduct HI+, and a co-eluting adduct [M 2H12+ion. Attempts were made to perform amino acid composition analysis and peptide sequencing cis-BaP tetrahydrotetrol on the fractions corresponding to the adducts' elution w \ 01. , volumes, but contaminating unadducted peptides made 10 20 30 these analyses unrealizable. Hence, a different HPLC Time (min) solvent system ('gradient 2') with more polar starting Figure 3. Reversed-phase HPLC (gradient 1) of materials conditions was used in a n attempt to clear the peptide bound by boronate affinity medium after Pronase E digestion adducts from unadducted contaminants. By the time of of human serum albumin adducted by (+)-anti-BaPDE in vitro. The peaks a t 7.4,10.2,and 11.6 min are the diol epoxide-derived this latter analysis, the boronate-eluted material had carboxylic acid eater adducts of Asp-Glu-Gly-Lys-Ala-Ser, Aspbeen in solution (in a frozen state a t -70 "C) for several Glu-Gly-Lys,and Arg-Asp-Glu-Gly-Lys,respectively. The peaks weeks. Re-analysis by LC-MS of this material showed a t 15.5 and 21.9 min are the 9,lO-trans and 9,lO-czs BaP the presence of the three original adducted peptides noted tetrahydrotetrols, respectively, which presumably arose from hydrolysis of ester adducts. in the first analysis plus evidence of three new adducts (Figure 5). A typical API mass spectrum, showing Tyr, with the latter being the most abundant (Figure 4; overlapping signals from different adducts, is shown in BaP triolyl = r-7,t-8,t-9-trihydroxy-7,8,9,10-tetrahydroben- Figure 6. Figure 7 displays the co-elution of [M HI+, zo[aIpyren-lO-yl). Analysis by FAB-MS of the acetylated [peptide HI+, and [M 2H12+ions for the major adduct adducts from the reaction of the (+)-enantiomer showed (750 m 12). The presence of co-eluting unadducted peponly acetylated BaP tetrahydrotetrols and unadducted tides was dramatically reduced when gradient 2 was peptides (data not shown), Le., the adducts from the (+)employed. The LC fractions were analyzed for amino enantiomer were not stable to derivatization. acid composition, peptides were sequenced by the Edman Since we could not be certain that these unadducted method, and the results were compared to the LC-MS peptides originated from peptide adducts of (+)-antiexperiments (Table 1). BaPDE or were merely co-eluting unadducted peptides, By taking into consideration both the API-LC-MS we undertook a different analytical approach to elucidate and peptide sequencing data, it was apparent that the the (+)-enantiomer adduct structures. The peptide adthree peptide adducts which were present in both the ducts eluted from the boronate cis-diol affinity column original and later LC-MS experiments ([M HI+ m l z formed by the (+)-enantiomer were analyzed by LC-MS 908, 750, and 906) had a common origin, the portion of using a pneumatically-assisted electrospray (ionspray) the protein spanning residues 186-192. The major atmospheric pressure ionization (API) source on a single adduct, which yielded [M + HI+ m l z 750 and [M Nal+ quadrupole instrument. The use of 'gradient 1' (see m I z 772, contains the peptide Asp-Glu-Gly-Lys, which experimental section) showed the presence of three uniquely corresponds to residues 187-190 in human adducts with m l z 908, 750, and 906, respectively, plus serum albumin. Since this adducted peptide had W the trans and cis BaP tetrahydrotetrols. The adduct [M 5oj
+
q /i;tJ\ ,
,
,
\',
,
+
+
,
+
+
+
+
+
r X 1.3
.
R=
r
D
F
284.8
1326.8
I
100 200 300
400
500 600
760
800
900
100
m/Z
200 300
400
synthetic
500
600 7QO 800
i
9h
Figure 4. Fast atom bombardment collision-induced dissociation mass spectra of peracetylated (-)-anti-BaPDE adducts noted in Figure 2. (A) His adduct; (B) His-Pro adduct; ( C ) His-Pro-Tyr adduct; (D) synthetic His-Pro-Tyr adduct prepared from the tripeptide (23). These are the same adducts noted from reaction of the racemic diol epoxide with human serum albumin and analyzed by FAB-MS in ref 23, where discussion of the fragmentation can be found.
Enantiospecific Benzo[a]pyrene-Albumin
Chem. Res. Toxicol., Vol. 7, No. 6, 1994 833
Adduction
750
I
I-,
--
A
733
, - , / , k , - y , - , y
-
$
-
,
_ I _
449
Q
d
II I
750
,
,
,
,
n
,J,\
fl
375'5 ,
,
,
,
,
,
-
A
906
18
14
22
Time (min) Figure 7. Selected ion chromatograms showing co-elution of the [M + HI+, [peptide + HI+, and [M + 2H12+ions from the major carboxylic acid ester adduct (BaPtriolyl-Asp-Glu-Gly-Lys) of (+)-anti-BaPDE. carboxylic acid ester adduct peptides noted were AspGlu-Gly-Ly-Ala-Ser(Asp(187)-Ser(192), adduct [M HI+ m / z 908) and Arg-Asp-Glu-Gly-Lys(Arg(186)-Lys(190), adduct [M HI+ mlz 906). Again, the carboxylate side chain of either Asp(187) or Glu(188) is the likely nucleophile that reacted with anti-BaPDE. The three peptide adducts which appeared only after prolonged storage of the boronate-bound material, and thus are most likely artifacts arising from the unstable nature of carboxylic acid esters of anti-BaPDE (27),correspond to Asn-LeuGly-Lys (Asn(429)-Lys(4321, [M HI+ m l z 7331, LysPro-Lys (Lys(536)-Lys(538), [M HI+ mlz 6741, and Lys ([M HI+ m l z 449) adducts. It was not possible to determine whether it was the N-terminal amino or the Lys-+amino adducts that the carboxylic acid esters had rearranged to during the prolonged storage period.
+
+
-ae
453 6 489 3 75
44f
I
I
+
520 1
+
+
Discussion
400
600
800 mlz
Figure 6. Pneumatically-assistedelectrospraymass spectrum of the peak corresponding to the major human serum albumin carboxylic acid ester adduct (BaP triolyl-Asp-Glu-Gly-Lys, [M + HI+ m l z 750) from (+I-anti-BaPDE.m l z 778 = [BaP triolylAsp-Glu-Gly-Lys+Nal+,m l z 448 = [Asp-Glu-Gly-Lys+ HI+, m l z 375.6 = [BaP triolyl Asp-Glu-Gly-Lys + 2H12+,m l z 303 = [BaPtriolyll+.The contaminatingsignals include those from the later-eluting (+)-anti-BaPDEester adduct BaP triolyl-Arg-AspGlu-Gly-Lys,[M + HI+ m l z 906, m l z 604 = [Arg-Asp-Glu-Gly+ Lys + HI+, m l z 453.6 = [BaP triolyl-Arg-Asp-Glu-Gly-Lys 2H12+.The signals at m l z 489 and 520 are from contaminating, unadducted peptides. spectral properties characteristic of a carboxylic acid ester rather than alkylamine adduct, the likely adducting nucleophilic atom is a n oxygen of one of the carboxylate side chains, either Asp(187) or Glu(188). The other two
We have shown in previous papers (23, 24) that racemic anti-BaPDE, when reacted with human serum albumin, forms both Nr-imidazolyl-His(146) adducts and carboxylic ester adducts with then unidentified acidic residueb). The racemically adducted protein yielded complex, incompletely resolvable FLN spectra, which were indicative of both carboxylic acid ester and Nrimidazolyl adducts. The FLN spectra simplified upon extensive dialysis of the protein to spectra which were equivalent to synthetic Nr-His-BaPDE adducts (24). The histidine adducts from digestion of the protein were amenable to isolation, peracetylation, and identification by FAJ3-MS analysis. The presence of carboxylic acid ester adducts was proven by study of the incorporation of l80from Hz180 into BaP tetrahydrotetrols released from the protein by hydrolysis under different pH conditions, where ester hydrolysis could be shifted between unimolecular and bimolecular mechanisms (23). The
834 Chem. Res. Toxicol., Vol. 7, No. 6, 1994
Day et al.
Table 1. Peptide Adducts frcsm Reaction of (+)-anti-BaPDEand Human Serum Albumin Characterized by API-LC-MS and Edman Seauencing elution order (gradient 2)
+
adduct IM H1+ peptide-[M HI+ aa sequence position
+
1
2
3
4
5
6
733 431 Asn-Leu-Gly-Lys 429-432
908 606 Asp-Glu-Gly-Lys-Ala-Ser 187-192
674 372 Lys-Pro-Lys 536-538
449 147 Lys
750 448 Asp-Glu-Gly-Lys 187-190
906 604 Arg-Asp-Glu-Gly-Lys 186-190
identity of the carboxylic acid side chain(s) adducted by racemic anti-BaPDE was undetermined in previous studies due to the somewhat unstable nature of carboxylic esters of this type. In this study, we used a different analytical approach for determining structure of the carboxylic acid ester adducts. The use of ionspray LC-MS, which has recently been shown to be effective in determination of sites of xenobiotic-protein adduct formation (281, allowed for the detection and elucidation of the structure of peptides bearing carboxylic acid ester adducts of anti-BaPDE. Even under the mild conditions required for ionization and detection of [M HI+ ions of adducts by this method, enough energy was imparted to the analytes to cause partial fragmentation. The adducts were identified in the LC-MS experiments by their [M HI+ and [M 2H12+ions and co-eluting fragments corresponding to the [peptide HI+ ion of the adducted peptide and the [BaP triolyl]+ ion. The present study shows that human serum albumin exhibits remarkable enantiospecificity in its formation of covalent adducts upon reaction with the enantiomers of anti-BaPDE. NT-His(146)of human serum albumin is alkylated by the (-)-enantiomer of anti-BaPDE, while the more mutagenic and tumorigenic (+)-enantiomer alkylates the carboxylic acid side chain of either Asp(187) o r Glu(188). Numerous studies have examined the enantiospecificity and site specificity of small moleculebinding to serum albumins, including both non-covalent and covalent bond-forming ligands (29-31). In fact, immobilized serum albumins are quite efficient chiral stationary phases for the separation of the enantiomers of many compounds (32). In the cases of covalent bondforming ligands examined to date, the protein seems t o be nearly equally reactive to both enantiomers of a given ligand but discriminative in the sites of reaction. To the best of our knowledge, this is the first study to actually identify the exact sites of covalent bond formation by electrophilic enantiomers with human serum albumin a t the amino acid level. Examination of published partial molecular models of human serum albumin indicates a possible explanation for the enantioselective reaction of the protein with antiBaPDE. His(146) is in the hairpin turn connecting helix 8 to helix 9 in subdomain IB, and Asp(187) and Glu(188) lie approximately two a-helical turns back from the C-terminal end of helix 10 (33). These three helices form a region roughly shaped liked the combination of the letters VI', where His(146) is near the vertex of the V and Asp(187) and Glu(188) are near the bottom of the I. Even though the articles describing the crystallographic structure of human serum albumin have been in print for 2 or more years, the atomic coordinates of its structure have not been made publicly available. Hence, we could not perform computer graphics and computational modeling of these binding sites to further study the origins of the enantioselectivity.
+
+
+
+
This previously unrealized enantiospecificity of the reaction of anti-BaPDE with human serum albumin has important consequences for the application of the adducts as biomarkers of internal exposure. GC-MS and other methods that measure BaP tetrahydrotetrols released from the protein by hydrolysis will be indicative of adduction by (+)-anti-BaPDE only. Methods that do not partition the Hid1461 and carboxylic acid ester adducts such as immunoassays or broad-band fluorimetric assays will give a measure of both (+)- and (-)-anti-BaPDE adducts combined. Selective determination of either the (-1-anti-BaPDE adduct or both will require the development and application of analytical methods which do not depend on sample volatility, such as the HPLC-MS approaches described in this paper.
Acknowledgment. This work was supported by NIH Grants ES04675 and ES07020 and by American Cancer Society Grants SIG-11-1 and IRG-58-31. The MIT Laser Biomedical Resource Center is supported by NIH Grant RR02594 (to M. Feld). The MIT Mass Spectrometry Facility is supported by NIH Grant RR00317 (to K. Biemann).
References (1) Hall, M., and Grover, P. L. (1988) Stereoselective aspects of the metabolic activation of benzo[a]pyrene by human skin in vitro. Chem.-Biol. Interact. 64,281-296. (2) Bjelogrlic, N., Peng, R. X., Park, S. S., Gelboin, H. V., Honkakoski, P., Pelkonen, O., and Vahakangas, K. (1993) Involvement of P450 1Al in benzo(a)pyrene but not in benzo(a)pyrene-7,8-dihydrodiol activation by 3-methylcholanthrene-induced mouse-liver microsomes. Pharmacol. Toxicol. 3, 319-324. (3) Schmalix, W. A., Maser, H., Kiefer, F., Reen, R., Wiebel, F. J., Gonzalez, F., Seidel, A., Glatt, H., Greim, H., and Doehmer, J. (1993) Stable expression of human cytochrome-P450 1Al cDNA in V79 Chinese-hamster cells and metabolic-activation of benzo[alpyrene. Eur. J . Pharmaco1:Environ. Toxicol. Pharmacol. 248, 251-261. (4) States, J . C., Quan, T. H., Hines, R. N., Novak, R. F., and Rungemorris, M. (1993) Expression of human cytochrome-P450 1Al in DNA-repair deficient and proficient human fibroblasts stably transformed with an inducible expression vector. Carcinogenesis 14,1643-1649. (5) Jankowiak, R., Day, B. W., Lu, P., Doxtader, M. M., Skipper, P. L., Tannenbaum, S. R., and Small, G. J. (1990) Fluorescence line narrowing spectral analysis of in vivo hemoglobin adducts of benzo[alpyrene diol epoxide: comparison to synthetic analogues. J. Am. Chem. SOC.112,5866-5869. (6) Day, B. W., Naylor, S., Gan, L.-S., Sahali, Y., Nguyen, T. T., Skipper, P. L., Wishnok, J. S., and Tannenbaum, S. R. (1990) Molecular dosimetry of polycyclic aromatic hydrocarbon epoxides and diol epoxides via hemoglobin adducts. Cancer Res. 50,46114618. (7) Weston, A., Rowe, M. L., Manchester, D. K., Farmer, P. B., Mann, D. L., and Harris, C. C. (1989) Fluorescence and mass spectral evidence for the formation of benzo[alpyrene anti-diol epoxideDNA and -hemoglobin adducts in humans. Carcinogenesis 10, 251-257. ( 8 ) Shields, P. G., Kato, S., Bowman, E. D., Petruzzelli, S., Cooper, D. P., Povey, A. C., and Weston, A. (1993) Combined micropreparative techniques with synchronous fluorescence spectroscopy or 32P-postlabellingassay for carcinogen-DNA adduct determination. IARC Sci. Publ. No. 124,243-254. (9)Mumford, J . L., Lee, X., Lewtas, J., Young, T. L., and Santella, R. (1993) DNA adducts as biomarkers for assessing exposure to
Enantiospecific Benzo[alpyrene-Albumin Adduction polycyclic aromatic hydrocarbons in tissues from Xuan Wei women with high exposure to coal combustion emissions and high lung cancer mortality. Enuiron. Health. Perspect. 99, 83-87. (10)Hall, M., Forrester, L. M., Parker, D. K, Grover, P. L. and Wolf, C. R. (1989)Relative contribution of various forms of cytochrome P450 to the metabolism of benzo[a]pyrene by human liver microsomes. Carcinogenesis 10,1815-1821. (11)Hall, M.,and Grover, P. L. (1988)Stereoselective aspects of the metabolic activation of benzo[alpyrene by human skin in vitro. Chem.-Biol. Interact. 64,281-296. (12)Brookes, P., and Osborne, M. R. (1982)Mutation in mammalian cells by stereoisomers of anti-benzo[a]pyrene-diol-epoxidein relation to the extent and nature of the DNA reaction products. Carcinogenesis 3, 1223-1226. (13)Pelling, J. C.,Slaga T. J., and DiGiovanni, J. (1984)Formation and persistence of DNA, RNA, and protein adducts in mouse skin exposed to pure optical enantiomers of 7j3,8a-dihydroxy-9a,lOaepoxy-7,8,9,10-tetrahydrobenzo[alpyrene in vivo. Cancer Res. 44, 1081-1086. (14)Kim, S. K.,Geacintov, N. E., Brenner, H. C., and Harvey, R. G. (1989)Identification of conformationally different binding sites in benzo[alpyrene diol epoxide-DNA adducts by low temperature fluorescence spectroscopy. Carcinogenesis 10,1333-1335. (15)Jankowiak, R. (1992)Heterogeneous distributions and dispersive photodissociation rates of benzo[a]pyrene diol-epoxide enantiomer-DNA and -poly(dG-dC)poly(dG-dC). Biophys. Chem. 42, 133-146. (16)Geacintov, N. E., Cosman, M., Mao, B., Alfano, A., Ibanez, V., and Harvey, R. G. (1991)Spectroscopic characteristics and site Vsite I1 classification of cis and trans benzo[a]pyrene diolepoxide enantiomer-guanosine adducts in oligonucleotides and polynucleotides. Carcinogenesis 12,2099-2108. (17)Ruche, C. J., Jeffrey, A. M., Mao, B., Alfano, A., Kim, S. K, Ibanez, V., and Geacintov, N. E. (1991)Dependence of conformations of benzo[alpyrene diol epoxide-DNA adducts derived from stereoisomers of different tumorigenicities on base sequence. Chem. Res. Toxicol. 4,311-317. (18)Jankowiak, R., Lu, P.-Q., Small, G. J., and Geacintov N. E. (1990) Laser spectroscopic studies of DNA adduct structure types from enantiomeric diol epoxides of benzo[a]pyrene. Chem. Res. Toxicol. 3,39-46. (19)Lee, B. M.,Yin, B. Y., Herbert, R., Hemminki, K., Perera, F. P., and Santella, R. M. (1991)Immunologic measurement of polycyclic aromatic hydrocarbon-albumin adducts in foundry workers and roofers. Scand. J. Work Enuiron. Health 17,190-194. (20)Dellomo M., and Lauwerys, R. R. (1993)Adducts to macromolecules in the biological monitoring of workers exposed to polycyclic aromatic-hydrocarbons. Crit. Rev. Toxicol. 23,111-126.
Chem. Res. Toxicol., Vol. 7, No. 6, 1994 835 (21) Sherson, D., Sabro, P., Sigsgaard, T., Johansen, F., and Autrup, H. (1990)Biological monitoring of foundry workers exposed to polycyclic aromatic hydrocarbons. Brit. J.Ind. Med. 47,448-453. (22)Day, B.W., Skipper, P. L., Wishnok, J. S., Coghlin, J.,Hammond, S. K., Gann,P., and Tannenbaum, S. R. (1990)Identification of an in uiuo chrysene diol epoxide adduct in human hemoglobin. Chem. Res. Toxicol. 3,340-343. (23)Day, B. W., Skipper, P. L., Zaia, J., and Tannenbaum, S. R. (1991) Benzo[alpyrene anti-diol epoxide covalently modifies human serum albumin carboxylate side chains and imidazole side chain of histidine(l46). J . Am. Chem. SOC.113,8505-8509. (24)Day, B. W., Doxtader, M. M., Rich, R. H., Skipper, P. L., Singh, K, Dasari, R. R., and Tannenbaum, S. R. (1992)Human serum albumin-benzo[alpyrene anti-diol epoxide adduct structure elucidation by fluorescence line narrowing spectroscopy. Chem. Res. Toxicol. 5, 71-76. (25)NIH Guidelines for the Laboratory Use of Chemical Carcinogens (1981)NIH Publication No. 81-2385,US.Government Printing Office, Washington, DC. (26)Naylor, S., Gan, L.-S., Day, B. W., Pastorelli, R., Skipper, P. L., and Tannenbaum, S. R. (1990)Benzo[alpyrene diol epoxide adduct formation in mouse and human hemoglobin: physicochemical basis for dosimetry. Chem. Res. Toxicol. 3,111-117. (27)Skipper, P.L., Naylor, S., Gan, L.-S., Day, B. W., Pastorelli, R., and Tannenbaum, S. R. (1989)Origin of tetrahydrotetrols derived from human hemoglobin adducts of benzo[alpyrene. Chem. Res. Toxicol. 2,280-281. (28)Springer, D. L., Bull, R. J., Goheen, S. C., Sylvester, D. M., and Edmonds, C. G. (1993)Electrospray-ionization mass spectrometric characterization of acrylamide adducts to hemoglobin. J. Toxicol. Enuiron. Health 40,161-176. (29)Mayer, S.,Mutschler, E., Benet, L. Z., and Spahn-Langguth, H. (1993)In vitro and in vivo irreversible plasma protein binding of beclobric acid enantiomers. Chirality 5,120-125. (30)Dubois, N., Lapicque, F., Abiteboul, M., and Netter, P. (1993) Stereoselective protein-binding of ketoprofen-effect of albumin concentration and of the biological system. Chirality 5,126-134. (31)Ding, A., Ojingwa, J. C., McDonagh, A. F., Burlingame, A. L., and Benet, L. Z. (1993)Evidence for covalent binding of acyl glucuronides to serum-albumin via a n imine mechanism as revealed by tandem mass-spectrometry. Proc. Natl. Acad. Sci. U S A . 90, 3797-3801. (32)AUenmark, S. (1993)Chiral discrimination by albumin: a mechanistic study of liquid chromatographic optical resolution of nonaromatic carboxylic acids. Chirality 5,295-299. (33)He, X. M., and Carter, D. C. (1992)Atomic structure and chemistry of human serum albumin. Nature 358,209-215.
