(+)-trans-anti-BPDE-N2-dG

• CAS: 65437-20-9
• Synonyms: (+)-trans-anti-Benzo[a]pyrene-dG; 10S(+)-trans-anti-[Py]-N2-dG; [7R-(7α,8β,9β,10α)]-2’-deoxy-N-(7,8,9,10-tetrahydro-7,8,9-trihydroxybenzo[a]pyren-10-yl)guanosine; Benzo[a]pyrene, guanosine deriv.; (+)-r-7,t-8-Dihydroxy-t-9,10-oxy-7,8,9,10-tetrahydro
• Molecular Formula: C30H27N5O7
• Molecular Weight: 569.6 g/mol
• SMILES: C1C(C(OC1N2C=NC3=C2NC(=NC3=O)NC4C(C(C(C5=C4C6=C7C(=C5)C=CC8=C7C(=CC=C8)C=C6)O)O)O)CO)O
• InChI: InChI=1S/C30H27N5O7/c36-10-18-17(37)9-19(42-18)35-11-31-24-28(35)33-30(34-29(24)41)32-23-22-15-7-6-13-3-1-2-12-4-5-14(21(15)20(12)13)8-16(22)25(38)27(40)26(23)39/h1-8,11,17-19,23,25-27,36-40H,9-10H2,(H2,32,33,34,41)/t17-,18+,19+,23-,25+,26+,27-/m0/s1
• InChIKey: WOKFWJGVLJVTLH-KCAXFQNZSA-N

As the predominant (~90%) DNA adduct from benzo[a]pyrene metabolic activation, (+)-trans-anti-BPDE-N2-dG is the indispensable stereoisomer-specific reference standard for carcinogenesis research. Unlike racemic mixtures or incorrect stereoisomers, this (+)-trans isomer exhibits a 3.5-fold higher G→T mutation frequency, 15-fold slower NER repair kinetics, and greater DNA bending (30–51°) than its (−)-trans counterpart. Procuring this authenticated stereoisomer ensures reproducible results in DNA repair inhibition assays, translesion synthesis studies, and biomarker quantification by LC-ESI-MS/MS. Only the exact (+)-trans configuration guarantees biologically relevant data.

(+)-trans-anti-BPDE-N2-dG (CAS 65437-20-9): Definitive Reference Standard for the Major Benzo[a]pyrene DNA Adduct in Carcinogenesis Research

(+)-trans-anti-BPDE-N2-dG (CAS 65437-20-9) is the dominant covalent DNA adduct formed from the metabolic activation of the environmental carcinogen benzo[a]pyrene, representing the 10S (+)-trans addition product of the ultimate carcinogen (+)-anti-benzo[a]pyrene-7,8-dihydrodiol-9,10-epoxide to the N² position of deoxyguanosine [1]. This bulky, stereochemically defined lesion is the most abundant BPDE-derived adduct in mammalian tissues following benzo[a]pyrene exposure, accounting for approximately 90% of all (+)-anti-BPDE DNA adducts and exhibiting a trans/cis formation ratio of 94:1 that starkly contrasts with the reduced regiospecificity observed for its (−)-anti-BPDE enantiomer [2]. As the principal mutagenic and tumor-initiating lesion underlying PAH-associated carcinogenesis, this compound is indispensable as a validated analytical reference standard and site-specific probe in DNA damage, repair, and translesion synthesis studies.

• [1] Zhao B, Wang J, Geacintov NE, Wang Z. Polη, Polζ and Rev1 together are required for G to T transversion mutations induced by the (+)- and (−)-trans-anti-BPDE-N2-dG DNA adducts in yeast cells. Nucleic Acids Res. 2006;34(2):417-425. PMID: 16415180. View Source
• [2] Suh M. Line narrowing spectroscopic studies of DNA-carcinogen adducts and DNA-dye complexes. Ph.D. Dissertation, Iowa State University; 1995. View Source

Why Stereochemistry Defines Function: The Irreplaceable Identity of (+)-trans-anti-BPDE-N2-dG in Adduct Biology

The four stereoisomeric anti-BPDE-N2-dG adducts—(+)-trans, (−)-trans, (+)-cis, and (−)-cis—are chemically identical in mass and formula yet display profoundly divergent biological properties that render generic substitution scientifically invalid. The (+)-trans isomer is derived from the highly tumorigenic (+)-anti-BPDE enantiomer, whereas the (−)-trans isomer originates from the nontumorigenic (−)-anti-BPDE enantiomer [1]. These adducts adopt distinct conformations within DNA duplexes (minor-groove vs. intercalated), induce different degrees of helical distortion, exhibit differential recognition and excision by human nucleotide excision repair (NER), and are bypassed with stereochemistry-dependent fidelity and efficiency by translesion synthesis polymerases [2][3]. Procurement of the precise (+)-trans stereoisomer, rather than a racemic mixture or an incorrect stereoisomer, is therefore essential for experimental reproducibility and biological relevance in carcinogenesis, DNA repair, and mutagenesis research.

• [1] Xu R, Mao B, Amin S, Geacintov NE. Bending and circularization of site-specific and stereoisomeric carcinogen-DNA adducts. Biochemistry. 1998;37(2):769-778. PMID: 9425101. View Source
• [2] Hess MT, Gunz D, Luneva N, Geacintov NE, Naegeli H. Base pair conformation-dependent excision of benzo[a]pyrene diol epoxide-guanine adducts by human nucleotide excision repair enzymes. Mol Cell Biol. 1997;17(12):7069-7076. PMID: 9372938. View Source
• [3] Zhang Y, Wu X, Guo D, Rechkoblit O, Wang Z. Two-step error-prone bypass of the (+)- and (−)-trans-anti-BPDE-N2-dG adducts by human DNA polymerases η and κ. Mutat Res. 2002;510(1-2):23-35. View Source

Quantitative Differentiation of (+)-trans-anti-BPDE-N2-dG Against Stereoisomeric and Functional Comparators

Adduct Formation Abundance and Regiospecificity: (+)-trans Dominates (94:1 trans/cis) vs. (−)-trans (63:22)

The (+)-anti-BPDE enantiomer, from which the target compound is derived, reacts with DNA to yield approximately 90% N2-dG adducts, with a trans-to-cis addition ratio of 94:1. In contrast, the (−)-anti-BPDE enantiomer yields only 50% N2-dG products alongside substantial non-specific adducts, with a reduced trans/cis ratio of 63:22 [1]. This means the target (+)-trans adduct is the overwhelmingly dominant lesion in biologically relevant exposures, while (−)-trans is a minor and less specific product. For analytical method development, the (+)-trans isomer must be the primary calibration standard because it represents the adduct species detected in human tissues at the highest abundance; use of a racemic mixture or the wrong stereoisomer introduces systematic quantification errors [1].

Evidence Dimension	Regiospecificity (% N2-dG adducts) and trans/cis addition ratio
Target Compound Data	~90% N2-dG adducts; trans/cis ratio = 94:1
Comparator Or Baseline	(−)-anti-BPDE: ~50% N2-dG adducts; trans/cis ratio = 63:22
Quantified Difference	1.8-fold higher N2-dG specificity; 47-fold greater trans selectivity (94:1 vs. 2.9:1 for (−)-anti when expressed as trans/cis odds ratio)
Conditions	In vitro reaction of (+)-anti-BPDE vs. (−)-anti-BPDE with DNA, analyzed by low-temperature fluorescence spectroscopy and ³²P-postlabeling

Why This Matters

For any quantitative DNA adduct assay (LC-MS/MS, ³²P-postlabeling, immunoassay), the absolute abundance dominance of the (+)-trans adduct means that a certified (+)-trans standard is non-negotiable for accurate calibration; substituting with a racemic mixture or an incorrect stereoisomer will introduce systematic bias and underestimate true exposure levels.

• [1] Suh M. Line narrowing spectroscopic studies of DNA-carcinogen adducts and DNA-dye complexes. Ph.D. Dissertation, Iowa State University; 1995. View Source

DNA Helical Distortion: (+)-trans Induces 30–51° Bending vs. 20–40° for (−)-trans, with Net Lesion Contributions of 21–26° vs. 5–19°

In a direct head-to-head T4 ligase-mediated cyclization assay using site-specifically modified 11-mer oligonucleotide duplexes, the (+)-trans-anti-BPDE-N2-dG adduct induced inferred bending angles of 30–51° for 90–92% of circular ligation products, with the most abundant circular product exhibiting a bending angle of 40°. In the same experimental system, the stereoisomeric (−)-trans-anti-BPDE-N2-dG adduct produced bending angles of only 20–40°, with the most abundant circular product at 26 ± 2° [1]. The net contribution per individual lesion to overall bending was 21–26° for (+)-trans vs. 5–19° for (−)-trans. Furthermore, the ratio of circular to linear DNA multimers reached 8–9 for the (+)-trans-modified duplex (circle contour sizes 99–121 bp) but only ~1 for (−)-trans (154–176 bp), indicating dramatically more efficient DNA cyclization driven by (+)-trans-induced distortion [1]. Unmodified control DNA exhibited circular product formation at least 10-fold less efficiently, with bending angles in the 14–19° range [1].

Evidence Dimension	DNA bending angle (inferred from circular ligation product distribution)
Target Compound Data	30–51° bending range; most abundant product 40°; net lesion contribution 21–26°; circular/linear ratio 8–9 at 99–121 bp
Comparator Or Baseline	(−)-trans-anti-BPDE-N2-dG: 20–40° bending range; most abundant product 26 ± 2°; net lesion contribution 5–19°; circular/linear ratio ~1 at 154–176 bp. Unmodified DNA: 14–19°; circular formation ≥10-fold less efficient
Quantified Difference	1.5- to 1.8-fold greater bending for (+)-trans in most abundant product (40° vs. 26°); 4- to 5-fold greater net lesion bending contribution; ~8-9 fold higher cyclization efficiency
Conditions	T4 DNA ligase cyclization of site-specific (+)-trans or (−)-trans modified 11-mer duplex 5'-d(CACAT[G*]TACAC)·d(TGTACATGTGG) at 16°C; electrophoretic separation and circular/linear product quantification

Why This Matters

The markedly greater DNA distortion induced by the (+)-trans adduct is a direct structural correlate of its unique recognition by NER, replication proteins, and damage-signaling kinases; researchers studying DNA damage recognition or architectural protein binding must use the correct stereoisomer to recapitulate in vivo conformational signals.

• [1] Xu R, Mao B, Amin S, Geacintov NE. Bending and circularization of site-specific and stereoisomeric carcinogen-DNA adducts. Biochemistry. 1998;37(2):769-778. PMID: 9425101. View Source

Nucleotide Excision Repair Efficiency: (+)-trans Repaired 15-Fold Slower than AAF-C8-dG; Cis Adducts Repaired 10-Fold Faster than Trans

In a cell-free human NER assay using site-specifically modified DNA substrates, human nucleotide excision repair excised the predominant (+)-trans-BPDE-N2-dG adduct 15 times less efficiently than a standard N-acetylaminofluorene-C8-dG (AAF-C8-dG) lesion situated in the same sequence context [1]. No significant difference in repair rate was observed between the (+)-trans and (−)-trans stereoisomers; however, changing the adduct configuration from trans to cis accelerated NER excision by approximately 10-fold for both (+)-cis and (−)-cis adducts [1]. Conversely, when the complementary cytosine opposite the lesion was replaced with adenine, excision of the cis adducts (but not the trans adducts) was reduced by an additional ~10-fold. The full dynamic range across all four stereoisomeric configurations and base-pairing contexts exceeded 100-fold in repair rates [1]. This demonstrates that stereochemistry—not merely chemical identity—governs repair protein recognition efficiency.

Evidence Dimension	Human NER dual incision efficiency (relative rate)
Target Compound Data	(+)-trans-BPDE-N2-dG: repair rate 15-fold lower than AAF-C8-dG standard; no difference vs. (−)-trans
Comparator Or Baseline	AAF-C8-dG (reference lesion): repair rate set as 15-fold faster baseline. (+)-cis-BPDE-N2-dG: ~10-fold faster repair than (+)-trans. (−)-cis-BPDE-N2-dG: ~10-fold faster than (−)-trans
Quantified Difference	≥100-fold range in NER rates across stereoisomeric and base-pairing combinations; trans isomers uniformly ~10-fold more resistant to NER than their cis counterparts
Conditions	Cell-free human NER assay in HeLa cell extracts; site-specific adducts in identical 135-mer DNA duplex sequence; dual incision product quantification

Why This Matters

The extreme resistance of the (+)-trans adduct to NER explains its in vivo persistence and its dominance as the primary mutagenic lesion in human tissues; studies of DNA repair inhibitor efficacy or damage persistence require the authentic (+)-trans stereoisomer, as the easily-repaired cis isomers do not model the biologically relevant repair-resistant lesion.

• [1] Hess MT, Gunz D, Luneva N, Geacintov NE, Naegeli H. Base pair conformation-dependent excision of benzo[a]pyrene diol epoxide-guanine adducts by human nucleotide excision repair enzymes. Mol Cell Biol. 1997;17(12):7069-7076. PMID: 9372938. View Source

Translesion Synthesis Fidelity: (+)-trans Directs 76% Accurate C Insertion vs. 89% for (−)-trans, with Elevated A (14%) and G (7%) Misinsertions

In a site-specific in vivo translesion synthesis assay in wild-type yeast cells, the nucleotide insertion spectrum opposite a single (+)-trans-anti-BPDE-N2-dG adduct was 76% C (accurate), 14% A (misinsertion), and 7% G (misinsertion). By contrast, the stereoisomeric (−)-trans-anti-BPDE-N2-dG adduct produced 89% C, 4% A, and 4% G [1]. The (+)-trans adduct thus exhibits a 3.5-fold higher frequency of A misinsertion (14% vs. 4%) and a 1.75-fold higher frequency of G misinsertion than the (−)-trans adduct, directly corresponding to a substantially greater propensity for G→T and G→C transversion mutations—the signature mutations observed in human lung cancers associated with PAH exposure. Translesion synthesis was reduced by 26–37% in rad30Δ (Polη-deficient) cells, while rev3Δ (Polζ-deficient) cells exhibited near-total ablation of bypass for both adducts [1].

Evidence Dimension	Nucleotide insertion fidelity opposite site-specific adduct in vivo
Target Compound Data	(+)-trans-anti-BPDE-N2-dG: 76% C, 14% A, 7% G; TLS reduced 26–37% in Polη-deficient cells
Comparator Or Baseline	(−)-trans-anti-BPDE-N2-dG: 89% C, 4% A, 4% G; TLS reduced similarly in Polη-deficient cells; near-total TLS abolition in Polζ-deficient cells for both adducts
Quantified Difference	3.5-fold higher A misinsertion (14% vs. 4%); 1.75-fold higher G misinsertion (7% vs. 4%); 13 percentage-point lower accurate C insertion (76% vs. 89%)
Conditions	Site-specific adduct in single-stranded vector transformed into S. cerevisiae wild-type and polymerase mutant strains; progeny plasmid sequencing for nucleotide insertion identity at lesion site

Why This Matters

The (+)-trans adduct is intrinsically more mutagenic than its (−)-trans stereoisomer, generating the G→T transversion signature characteristic of smoking-associated lung cancer; accurate mutagenesis studies and polymerase inhibitor screening require the authentic (+)-trans substrate to model clinically relevant mutation patterns.

• [1] Zhao B, Wang J, Geacintov NE, Wang Z. Polη, Polζ and Rev1 together are required for G to T transversion mutations induced by the (+)- and (−)-trans-anti-BPDE-N2-dG DNA adducts in yeast cells. Nucleic Acids Res. 2006;34(2):417-425. PMID: 16415180. View Source

Tumorigenic Potency of Precursor Diol Epoxide: (+)-anti-BPDE Carries 60–70% of Benzo[a]pyrene Tumor-Initiating Activity vs. Nontumorigenic (−)-anti-BPDE

Mouse skin tumor initiation studies demonstrate that the (+)-anti-BPDE enantiomer—the metabolic precursor of the target compound—possesses 60–70% of the total tumor-initiating activity of the parent benzo[a]pyrene. In contrast, the (−)-anti-BPDE enantiomer, which gives rise to the (−)-trans adduct, is essentially nontumorigenic in the same assay [1]. This differential tumorigenicity is not attributable to metabolic activation differences, as both enantiomers are the reactive diol epoxide forms directly administered to tissue; rather, it reflects the fundamentally different biological processing of the resulting N2-dG adducts. The (+)-trans adduct is thus the lesion that translates PAH exposure into tumor initiation in vivo, whereas the (−)-trans adduct is a biologically silent bystander [1][2].

Evidence Dimension	Mouse skin tumor-initiating activity of precursor diol epoxide enantiomers
Target Compound Data	(+)-anti-BPDE (precursor of target adduct): 60–70% of benzo[a]pyrene tumor-initiating activity; highly tumorigenic
Comparator Or Baseline	(−)-anti-BPDE (precursor of comparator adduct): essentially nontumorigenic; benzo[a]pyrene (parent carcinogen): 100% reference activity
Quantified Difference	Qualitative to semi-quantitative difference: (+)-anti-BPDE is tumorigenic; (−)-anti-BPDE is nontumorigenic; (+)-anti-BPDE accounts for majority of BaP's carcinogenic potency
Conditions	Two-stage mouse skin carcinogenesis model; topical application of enantiomerically pure diol epoxides followed by TPA promotion

Why This Matters

Only the (+)-trans adduct formed from the tumorigenic (+)-anti-BPDE enantiomer is pathobiologically relevant for studying the initiation step of chemical carcinogenesis; using the wrong stereoisomer or a racemic precursor in cell transformation or animal studies models a lesion with no demonstrated tumor-initiating capacity, invalidating carcinogenicity conclusions.

• [1] Suh M. Line narrowing spectroscopic studies of DNA-carcinogen adducts and DNA-dye complexes. Ph.D. Dissertation, Iowa State University; 1995. View Source
• [2] Xu R, Mao B, Amin S, Geacintov NE. Bending and circularization of site-specific and stereoisomeric carcinogen-DNA adducts. Biochemistry. 1998;37(2):769-778. PMID: 9425101. View Source

Analytical Detection Sensitivity: LC-ESI-MS/MS Achieves 1 amol LOD, Enabling Quantification at 1 Adduct per 10¹¹ Nucleotides in Human Lung DNA

A validated liquid chromatography-electrospray ionization tandem mass spectrometry (LC-ESI-MS/MS) method employing high-resolution/accurate mass analysis achieved a limit of detection of 1 attomol (amol) of BPDE-N2-dG on-column, corresponding to 1 adduct per 10¹¹ nucleotides (approximately 1 adduct per 10 human lung cells) using 40 μg of human lung DNA [1]. This sensitivity exceeds that of the conventional ³²P-postlabeling assay (~1 adduct per 10¹⁰ nucleotides) by roughly an order of magnitude [1]. Application of this method to 29 human lung DNA samples yielded quantifiable BPDE-N2-dG adducts in 20 samples, with smoker lung DNA containing a mean of 3.1 adducts per 10¹¹ nucleotides versus 1.3 adducts per 10¹¹ nucleotides in nonsmoker lung DNA—a 2.4-fold difference attributable to tobacco smoke exposure [1]. The assay's ability to detect and quantify this specific adduct in human tissues at ultra-trace levels is critically dependent on the availability of authenticated (+)-trans-anti-BPDE-N2-dG as the primary calibration and internal standard.

Evidence Dimension	Analytical sensitivity for BPDE-N2-dG detection in human tissue DNA
Target Compound Data	LOD: 1 amol on-column (1 adduct/10¹¹ nucleotides); smoker lung: 3.1 adducts/10¹¹ nt; nonsmoker lung: 1.3 adducts/10¹¹ nt
Comparator Or Baseline	³²P-postlabeling LOD: ~1 adduct/10¹⁰ nucleotides (approximately 10-fold less sensitive); no direct stereoisomer comparison in this analytical method
Quantified Difference	~10-fold sensitivity improvement over ³²P-postlabeling; 2.4-fold adduct level difference between smoker and nonsmoker lung DNA
Conditions	LC-ESI-MS/MS on Orbitrap mass spectrometer; N=29 human lung DNA samples; quantification against authentic (+)-trans-anti-BPDE-N2-dG standard

Why This Matters

For any human biomonitoring or molecular epidemiology study aiming to link PAH exposure to cancer risk, the availability of a highly pure, stereochemically defined (+)-trans-anti-BPDE-N2-dG reference standard is the rate-limiting factor for achieving and validating the sub-femtomolar sensitivity required to detect this lesion in clinical specimens.

• [1] Villalta PW, Hochalter JB, Hecht SS. Ultrasensitive high-resolution mass spectrometric analysis of a DNA adduct of the carcinogen benzo[a]pyrene in human lung. Anal Chem. 2017;89(23):12735-12742. PMID: 29111692. View Source

High-Impact Application Scenarios for (+)-trans-anti-BPDE-N2-dG: Where Stereochemical Authenticity Determines Experimental Validity

Site-Specific Mutagenesis and Carcinogenesis Mechanism Studies

Construction of site-specifically modified oligonucleotides or plasmid vectors containing the (+)-trans-anti-BPDE-N2-dG adduct enables precise determination of mutation spectra and frequencies in mammalian or yeast systems. As demonstrated in yeast, the (+)-trans adduct directs 14% A misinsertion (G→T transversions) versus only 4% for the (−)-trans stereoisomer, a 3.5-fold difference critical for modeling the G→T mutational signature found in TP53 from smoking-associated lung cancers [1]. Use of this authenticated stereoisomer is essential because even minor contamination with the (−)-trans adduct will dilute the mutagenic potency and produce artefactually reduced G→T mutation frequencies, undermining genotype-phenotype correlations in cancer driver gene studies.

Human Nucleotide Excision Repair Kinetics and Drug Sensitivity Assays

The (+)-trans-anti-BPDE-N2-dG adduct is repaired 15-fold less efficiently than the standard AAF-C8-dG lesion and approximately 10-fold less efficiently than its cis stereoisomers in cell-free human NER assays [2]. This repair resistance makes it the ideal substrate for testing small-molecule NER inhibitors, chemosensitizers, or evaluating NER-deficient cancer cell lines (e.g., xeroderma pigmentosum models). The 10-fold repair rate differential between trans and cis configurations means that experiments using a trans-adduct substrate are probing a fundamentally different repair kinetic regime than those using cis adducts; procurement of the pure (+)-trans isomer is mandatory to model clinically persistent lesions.

Human Biomonitoring and Molecular Epidemiology of PAH Exposure

The validated LC-ESI-MS/MS method achieving 1 amol LOD for BPDE-N2-dG has enabled the first direct quantification of this adduct in human lung DNA, revealing 3.1 vs. 1.3 adducts per 10¹¹ nucleotides in smokers versus nonsmokers, respectively [3]. This 2.4-fold differential provides a quantitative biomarker of PAH exposure that is directly linked to the tumorigenic (+)-trans adduct species. Accurate calibration requires the (+)-trans-anti-BPDE-N2-dG reference standard because the three other stereoisomers co-elute or partially co-elute under most chromatographic conditions; without an authentic stereoisomer-specific standard, the measured adduct levels cannot be confidently assigned to the biologically relevant (+)-trans form.

DNA Damage-Signaling and Chromatin Remodeling Studies

The (+)-trans-anti-BPDE-N2-dG adduct induces significantly greater DNA bending (30–51°, net lesion contribution 21–26°) than the (−)-trans isomer (20–40°, net 5–19°) as quantified by T4 ligase cyclization [4]. This enhanced structural distortion in the minor groove is the molecular signal that determines recruitment of damage-recognition proteins such as XPC-RAD23B and UV-DDB. Studies of ATM/ATR kinase activation, γ-H2AX foci formation, or chromatin immunoprecipitation of repair factors at damaged sites must use the (+)-trans isomer to faithfully replicate the architectural distortion that initiates the DNA damage response in vivo.

• [1] Zhao B, Wang J, Geacintov NE, Wang Z. Polη, Polζ and Rev1 together are required for G to T transversion mutations induced by the (+)- and (−)-trans-anti-BPDE-N2-dG DNA adducts in yeast cells. Nucleic Acids Res. 2006;34(2):417-425. PMID: 16415180. View Source
• [2] Hess MT, Gunz D, Luneva N, Geacintov NE, Naegeli H. Base pair conformation-dependent excision of benzo[a]pyrene diol epoxide-guanine adducts by human nucleotide excision repair enzymes. Mol Cell Biol. 1997;17(12):7069-7076. PMID: 9372938. View Source
• [3] Villalta PW, Hochalter JB, Hecht SS. Ultrasensitive high-resolution mass spectrometric analysis of a DNA adduct of the carcinogen benzo[a]pyrene in human lung. Anal Chem. 2017;89(23):12735-12742. PMID: 29111692. View Source
• [4] Xu R, Mao B, Amin S, Geacintov NE. Bending and circularization of site-specific and stereoisomeric carcinogen-DNA adducts. Biochemistry. 1998;37(2):769-778. PMID: 9425101. View Source