Chem. Res. Toxicol. 1991, 4 , 151-156

151

Articles Structure-Activity Study of Paracetamol Analogues: Inhibition of Replicative DNA Synthesis in V79 Chinese Hamster Cellst Ann M. Richard,*J Jan K. Hongslo,s Phillip F. Boone,li and Jerrn A. Holmes Carcinogenesis and Metabolism Branch (MD-68), U S . Environmental Protection Agency, Research Triangle Park, North Carolina 2771 1, Department of Environmental Medicine, National Institute of Public Health, Geitmyrsveien 75, N-0462 Oslo 4, Norway, and Environmental Health Research & Testing, Inc., Research Triangle Park, North Carolina 27709 Received October 1, 1990

Experimental and theoretical evidence pertaining to cytotoxic and genotoxic activity of paracetamol in biological systems was used to formulate a simple mechanistic hypothesis to explain the relative inhibition of replicative DNA synthesis by a series of 19 structurally similar paracetamol analogues, 5 of which were specifically analyzed for the current study. I t was hypothesized that the observed activity variation of the paracetamol analogues was based on the relative abilities of these compounds to undergo H atom loss a t the phenolic oxygen, and on the relative stabilities of the resulting free-radical species. Three calculated parameters were found to be relevant-the partial atomic charge on the ring carbon attached to the phenolic oxygen, the partial charge on the phenoxy radical oxygen, and the energy difference between the parent phenolic paracetamol analogue and the corresponding radical dissociation products. T h e variation in parameter values was significantly correlated with the relative inhibition of DNA synthesis and was easily rationalized in terms of the mechanistic hypothesis proposed. More specifically, competitive reaction with a tyrosyl radical species involving the transfer of a hydrogen atom a t the active site of ribonucleotide reductase was suggested as the underlying mechanistic basis for the observed activity variation of the paracetamol analogues. Comparison of calculated parameters for a model tyrosyl species and the paracetamol analogues was entirely consistent with this view.

Introduction Paracetamol (acetaminophen, 4-hydroxyacetanilide) is a widely used antipyretic analgesic which has been associated with a range of adverse biological effects. Overdosing may cause hepatic necrosis in man (I) and experimental animals ( 2 ) . Heavy alcohol consumption may predispose to liver damage caused by therapeutic doses of paracetamol (3). Chronic exposure of high doses of paracetamol have been reported to cause liver tumors in mice and bladder carcinomas in rats ( 4 , 5 ) . Furthermore, Sandler and co-workers (6) have found that long-term, daily use of paracetamol is associated independently with an increased risk of chronic renal disease. +Theresearch described in this paper has been reviewed by the Health Effecta Research Laboratory, U.S. Environmental Protection Agency, and approved for publication. Approval does not signify that the contents necessarily reflect the views and policies of the Agency, nor does mention of trade names or commercial products constitute endorsement or recommendation for use. * Author to whom correspondence should be addressed. 8 U.S. Environmental Protection Agency. INational Institute of Public Health. 11 Environmental Health Research & Testing, Inc.

0893-228x/91/2704-0151$02.50/0

In addition, a variety of genotoxic effects of paracetamol and related compounds have been demonstrated in several test systems (7-12). It has been shown in recent studies that paracetamol causes an immediate and highly specific inhibition of replicative DNA synthesis (13). Furthermore, studies with a cell line overproducing ribonucleotide reductase have demonstrated that paracetamol inhibits DNA synthesis by destroying a tyrosyl free radical on ribonucleotide reductase (141, possibly by means of a competitive H atom abstraction process. This is in accordance with a proposed mechanism for hydroxyurea inhibition of ribonucleotide reductase (15). Genotoxic effects, such as the induction of sister-chromatid exchanges and chromosomal aberrations, have been postulated to be a consequence of the subsequent deoxyribonucleotide pool imbalance or reduced DNA synthesis (14). Both theoretical and experimental investigations on the cytotoxic effects of paracetamol have provided evidence that the phenolic oxygen, rather than the amino nitrogen, is the energetically preferred site for H atom loss in paracetamol (16-18). In addition, much evidence has accumulated for the involvement of a free-radical intermediate in the conversion of paracetamol to the postulated ultimate reactive metabolite (N-acetyl-p-benzoquinone imine) for 0 1991 American Chemicd Societv

152 Chem. Res. Toxicol., Vol. 4, No. 2, 1991

the hepatotoxicity end p o i n t causing liver necrosis (17, 19-22). In accordance with these studies and on the basis of e x p e r i m e n t a l findings, i t has been suggested that the formation of a paracetamol p h e n o x y radical species accompanies, and m a y account for, the inactivation of ribonucleotide reductase (14). The present s t u d y will a t t e m p t to derive mechanistically relevant parameters to explain the activity variation of a series of paracetamol analogues on inhibition of replicative DNA synthesis. Modeling efforts will focus on the calculation of molecular electronic and energetic properties associated with relative propensities for the paracetamol analogues to u n d e r g o H atom loss, and t h e relative stabilities of the resulting free-radical p h e n o x y species.

Richard e t al. INHIBITION POTENCIES (IC50)

GROUP I

HIGH (.007 - . 0 2 2 M )

PARl *

G"", Q 0"

OH

0"

PAR(*

**)

0"

GROUP 11

MODERATE (.12-.5

M)

GROUP 111

Materials and Methods Chemicals. Caution: Skin contact with p-aminophenol and closely related analogues may cause dermatitis or skin sensitization, and inhalation can cause asthma a n d methemoglobin formation. Hence, latex gloves should be worn, a n d these chemicals should be handled in safety hoods. All test chemicals had stated purity ranging from 98% to 100%. 2,6-Dimethyl-4hydroxyacetanilide and 3,5-dimethyl-4-hydroxyacetanilide were generous gifts from Dr. Sidney D. Nelson, Department of Medicinal Chemistry, University of Washington, Seattle, WA. oAminophenol, m-aminophenol, o-hydroxyacetanilide, mhydroxyacetanilide, and phenetole were purchased from Aldrich Chemical Co., Milwaukee, WI. 0-and m-Hydroxyacetanilide were recrystallized from aqueous methanol. p-Ethoxyaniline was purchased from Eastman Kodak, Rochester, NY and the hydrochloride salt prepared in methanolic HCl. p-Hydroxyacetanilide (paracetamol), p-nitrophenol, and phenacetine were purchased from the Norwegian Medicinal Depot, Oslo. Other chemicals were obtained from the following sources: [methyl3H]thymidine (TdR; 45-49 Ci/mmol) and [14C]TdR (55 mCi/ mmol) from Amersham International, Amersham, U.K.; Eagle's MEM and fetal bovine serum from Gibco, Grand Island, NY; acetanilide, p-aminophenol, aniline, p-cresol, p-hydroxyphenol, 2,4-diaminophenol, p-(methylamino)phenol, and phenol from Sigma Chemical Co., St. Louis, MO. Cells. The V79 Chinese hamster cell line was kindly supplied by Dr.Lars Warngard, National Institute of Environmental Medicine, Stockholm, Sweden. The V79 cells were grown in Eagle's minimal essential medium (MEM), nonessential amino acids, 20 mM sodium pyruvate, 2 mM glutamine, 1000 IU of penicillin/mL, 0.1 mg of streptomycin/mL, and 5% fetal calf serum. The cells were grown in a humidified atmosphere of 5% C 0 2 in air. DNA Synthesis. V79 cells (2 X 106) were grown in [14C]TdR (0.01 pCi/mL) medium for 24 h and exposed to the test substance for 30 min. When necessary, the test substance was first dissolved in DMSO not exceeding a final concentration of 0.5%. The concentrations of test substance were based on preliminary studies over a wider concentration range (two concentrations per decade). After exposure to test substance, the incorporation of [3H]TdR (4 pCi/mL) during a 10-min pulse-labeling period was measured. T h e cells were washed with ice-cold PBS containing 50 mM unlabeled TdR, scraped off, and placed on Whatman GF/C filters. The filters were then successively washed with 4% perchloric acid, 70% ethanol, 95% ethanol, and 100% ethanol, air-dried, and dissolved in 0.4 N NaOH. The incorporation of [14C]TdR and [3H]TdR into DNA was determined by liquid scintillation counting. The [3H]TdR/[14C]TdRratios for exposed cells are expressed as percent of control cells. The values are means of a t least two experiments. Paracetamol was used as a positive control in all the experiments. Calculations. Initial geometries for each of the 19 paracetamol analogues in Figure 1 were graphically generated using the Macintosh CHEM3D modeling package (Cambridge Scientific Computing Inc., Cambridge, MA). Fully optimized geometries, implying variation of all bond distances, bond angles, and torsional angles, were subsequently obtained for each of the parent paracetamol analogues using the semiempirical AM1 method (23). These optimized geometries were used as starting points for the

PAN

PAT(?

LOW

PARl I

PAR10

PARI3

PAR12

(2.1-3.0 mM)

GROUP I V

INACTIVE 03.0

mM)

6 6 6"'

PARII'

OH

PAR15

PAR18

c€"iT*,

PAR11

PAR18 **COCH,

0"

6~3 PAR10

~

",I

CY,

on

Figure 1. Complete listing of PAR analogue structures grouped according to experimentally significant activity category, or inhibition potencies, where activity is inversely related to PAR concentration (ICa) which leads to a 50% reduction in rate of [3H]TdR incorporation in V79 cells; structures labeled with an asterisk (*) were specifically tested for the current study. full unrestricted AM1 (or UAM1) optimization of the phenoxy free radicals resulting from H atom loss from the corresponding parent phenols. All a b initio and semiempirical AM1 calculations were carried out using the GAUSSIAN86 program package (24) residing on an IBM 3090 located a t the EPA National Computer Center in Research Triangle Park, NC. Ab initio calculations were done a t the LCAO-MO-SCF level using the 3-21G basis set. Closed-shell restricted Hartree-Fock (RHF) calculations were performed on each of the AM1-optimized geometries of the 19 parent PAR analogues to obtain possible structure-activity parameters, e.g., energies and electronic properties. In addition, open-shell unrestricted Hartree-Fock (UHF) calculations were performed on each of the UAMl optimized geometries of the 14 unique phenoxy radicals. The UHF method allows for spin polarization and, hence, tends to give more realistic estimates of spin densities and charges (25). However, spin contamination for the UAMl//UHF/3-21G results was significant enough to lead to unreliable total energies, as estimated by the deviation of values from the pure doublet state value of 0.75. T o obtain more reliable total energies, restricted open-shell UAMl//ROHF/3-21G calculations were performed for the phenoxy radicals. For the two dimethylparacetamol phenoxy radical species, labeled PARl 1and PAR19 in Figure 1,the ROHF calculation would not converge. However, total ROHF energies and UHF energies for the remaining 12 PAR analogues were found to differ by a nearly constant scaling factm of 1,00009 f 0.00002. The UHF total energies were multiplied by this scaling factor to yield a reasonable estimate for the total ROHF energies of these three compounds. The energy difference between the parent PAR phenol and the higher energy phenoxy radical derivative was then calculated to provide an indication of' the relative stability of the radicals within the PAR series. The error associated with the energy difference due to the scaling of UHF energies is estimated to be less than 2%.

(s)

Results and Discussion The 19 paracetamol analogue structures analyzed i n the p r e s e n t s t u d y a r e listed i n Figure 1 and Tables I and I1 and grouped according to range of potency for inhibition of replicative DNA synthesis. ICso values represent the

Structure-Activity Study of Paracetamol Analogues

Chem. Res. Toxicol., Vol. 4, No. 2, 1991 153

Table I. Listing of Phenolic Paracetamol Analogues, Grouped According t o Concentration (IC,) T h a t Inhibits Rate of rSH]TdR Incorporation i n V79 Cells by 50%, a n d Calculated Properties PAR analogueso

ICm, mM Group I: High 0.007 0.018 0.019 0.022

PAR1* PAR2 PAR3 PAR4*

2,4-diaminophenol p-aminophenol o-aminophenol p-(methy1amino)phenol

PAR6* PAR7 PAR8 PARS*

p-cresol m-aminophenol p-hydroxyacetanilide (paracetamol) p-hydroxyphenol

PAR10 PARll PAR12

0 - hydroxyacetanilide

3,5-dimethyl-4-hydroxyacetanilide phenol

PAR14* PAR18 PAR19

p-nitrophenol m-hydroxyacetanilide 2,6-dimethyl-4-hydroxyacetanilide

QC (OH)'

QO (rad)d

AEe

(-1.0) -0.66 -0.66 -0.84

0.260 0.338 0.350 0.350

-0.530 -0.511 -0.519 -0.508

0.105 0.115 0.113 0.116

Group 11: Moderate 0.15 -0.17 0.19 -0.16 0.20 0.0 0.50 -0.37

0.368 0.404 0.369 0.355

-0.497 -0.498 -0.495 -0.497

0.126 0.137 0.126 0.122

0.368 0.368 0.383

-0.504 -0.507 -0.492

0.122 0.138 0.127

0.411 0.394 0.406

-0.474 -0.488 -0.469

0.137 0.141 0.142

'TOHb

Group 111: Low 2.1 2.2

(-0.48)

0.0

2.4

0.0

Group I V Inactive >3.0 0.78 >3.0 0.21 >3.0 (0.0)

"Labels for PAR analogues analyzed specifically for the present study are followed by an asterisk (*). bRelative Hammett constant (27) computed in relation to OH position; greater uncertainty in estimates for multisubstituent compounds PARI, -11,and -19. Partial atomic charge on phenolic carbon in parent PAR compound obtained from AMl//RHF/3-21G calculation. Partial atomic charge on phenoxy radical oxygen obtained from UAMl//UHF/B-BlG calculation. e Difference in total energies of dissociation products (phenoxy radical and H atom) and parent PAR compound in atomic units; energies obtained from UAMl//ROHF/3-21G (UHF energies were scaled for PARll and -19; see text) and AMl//RHF/S-SlG calculations, respectively. Table 11. Listing of Nonphenolic Paracetamol Analogues, Grouped According t o Concentration (IC,) T h a t Inhibits Rate of [SH]TdR Incorporation i n V79 Cells by 50%, a n d Calculated ProDerties ICm, QC QO PAR analogueso mM (OCH2CH,)o (rad)b A E c Group 11: Moderate PAR5 p-ethoxyaniline 0.12 0.371 -0.511 0.115 PAR13 phenacetin

Group 111: Low 3.0 0.394

-0.495 0.126

PAR15 acetanilide PAR16 phenetole PAR17 aniline

Group IV: Inactive >3.0 >3.0 0.400 >3.0

-0.492 0.127

Partial atomic charge on phenyl carbon attached to ethoxy group in neutral species obtained from AMl//RHF/&BlG calculation. Partial atomic charge on phenoxy radical oxygen derivative obtained from UAMl//UHF/3-21G calculation (same values as for PARZ, -8, and -12, respectively, in Table I). 'Difference in total energies of dissociation products (phenoxy radical and H atom) and parent phenolic species in atomic units (same values as for PAR2, -8, and -12, respectively, in Table I); radical product energies obtained from UAM1/ /ROHF/3-21G calculations and parent energies from AMl//RHF/3-21G calculations.

concentration of PAR analogue that results in a 50% reduction in the rate of [3H]TdR incorporation in V79 cells. The potency groupings (Group I = high, etc.) represent categories of meaningful activity variation within experimental error and are intended to clarify and highlight significant trends in Tables I and 11. Experimental data for 14 of these analogues [not labeled with an asterisk (*) in Figure 1 and Tables I and 111 were published in an earlier study (26). To summarize the results of that study, the aminophenols were found to be significantly more potent inhibitors of replicative DNA synthesis than their acetylated forms and, in both series, inhibition activity decreased in order of para, ortho, and meta substitution. Alkylation of the phenol group was found to attenuate inhibition activity, while removal of the phenol group eliminated such activity. Finally, of the two dimethylated analogues of paracetamol, PARl 1, which

forms the more stable phenoxy radical (19), was an active inhibitor, while PAR19 was not. These results suggested that the stability of the phenolic free radical was mechanistically relevant to the inhibition of DNA synthesis in V79 cells. The activity variation expressed by the results of the earlier Holme et al. study (26) seemed consistent with crude estimates, based on Hammett u values ( 2 3 , of relative electron-withdrawing capabilities of the various substituents in relation to the phenolic oxygen. Hence, Hammett constants were used to suggest additional PAR analogues (PARl, -4, -6, -9, and -14) for which analysis might provide further insight into the mechanistic basis of the observed activity. On the basis of Hammett constants, PARl and PAR4 were both predicted to be more active than the most active analogue previously analyzed, Le., PAR2. PAR6 and PAR9 were chosen to test the importance of the amine nitrogen on activity; both were predicted to be more active than phenol (PAR12). Finally, PAR14 was predicted to be among the least active PAR analogues if reduction to the corresponding amine were not taking place in the V79 cells. A listing of the Hammett values used is provided in Table I for the phenolic PAR analogues. The experimental activities obtained for the five PAR analogues recommended for testing were consistent with the Hammett constant predictions, as indicated in Table I. The Hammett constants, however, provide only a crude estimate of the relative electronic-withdrawing effect on the phenolic oxygen and do not take into account the overall electronic environment produced by the interaction of multiple substituents within a molecule. Also, it was not clear whether these electronic effects reflected the relative stabilities and propensities for phenoxy radical formation. These uncertainties motivated more rigorous calculation of electronic and energetic parameters pertaining to phenoxy radical formation. Results of these calculations are summarized in Tables I and 11, where data for the phenolic PAR analogues, and corresponding phenoxy radical derivatives, have been listed separately from the ethoxy and nonphenolic PAR analogues. A va-

154 Chem. Res. Toxicol., Vol. 4, No. 2, 1991

riety of physical and electronic properties of potential relevance to the hypothesized mechanism were examined for each analogue. For the parent phenols, these included the following: octanol/ water partition coefficients; highest occupied and lowest unoccupied molecular orbital (HOMO and LUMO) energies; select bond densities; and Mulliken charges on a few atoms in relation to the phenol group. For the radicals, examined properties included the energy of formation and select spin densities and Mulliken charges. Properties found to have significant correlation with inhibition activity dose values are listed in Tables I and 11. QC values represent the partial atomic charge on the phenyl carbon attached to the OH or OCH2CH3group and indirectly reflect the relative electronic induction or withdrawing effects of substituents in relation to the phenolic or ethoxy oxygen. With exceptions mainly in the intermediate groupings I1 and 111, these positively charged values tend to increase with decreasing activity grouping, indicating that loss of electronic charge concentration in the vicinity of the phenolic oxygen is associated with reduced activity. This is consistent with the observation that decreasing negative charge concentration on the radical oxygen, QO, also tends to be associated with reduced activity. The energy difference listed in the last column of Table I most directly relates to the ease of formation and relative stability of phenoxy radicals within the PAR series and represents the stability of the dissociation products (phenoxy radical and hydrogen) with respect to the lower energy parent PAR structure. Comparison of hE values for the analogues listed indicates that the phenoxy radical is more stable with respect to the parent PAR analogue for the more active compounds (lower AE values) and, therefore, more likely to form and persist. The formation of a radical oxygen species from the parent phenol, involving loss of H, lowers the total electron density in the vicinity of the electronegative oxygen atom by about 0.2 electron unit, which could lead to the observed increase in energy. Hence, substituent replacements that increase this electron density could be expected to lower the energy difference, increase the likelihood of free-radical formation, and lend stability to the resulting radical. Each of these three calculated parameters are fundamental properties related to the underlying physical basis for the empirical Hammett constant. Hence, in addition to providing a more rigorous description of the electronic environment within the PAR analogues, these properties provide a theoretical rationale for the relevance of the Hammett constants. Qualitatively, the parameter values within the intermediate groupings I1 and I11 in Table I deviate somewhat from the associated activity trends. However, the Group I and IV activity categories, which represent the highest and lowest potency compounds, are clearly separated for all four parameter columns. Simple regression analysis of each of the four columns of data in Table I wm performed to gauge the statistical significance of the results. Correlation coefficients were obtained for log (ICmconcentration) versus each of the calculated parameters, where the inhibition potency for the Group IV "inactive" PAR analogues was set at 10 mM. The hE parameter, which has the clearest relationship with the proposed mechanistic hypothesis, gave the highest correlation coefficient "R" value of 0.86, while Hammett a's and QC and QO parameters had correlation coefficients of 0.83, 0.77, and 0.79, respectively. Clearly, these parameters bear a close physical relationship to one another. An analysis of these intercorrelations produced "R" values ranging from 0.74 for the (a,AE)pair to 0.92 for the (a,QO) pair. These statistical results are in accord with earlier discussion and

Richard et al. bolster confidence in the mechanistic interpretation and conclusions inferred from the data. To summarize the data in Table I, the Hammett a's and the three calculated parameters QC, QO, and AE display trends which consistently support the view that free-radical formation and stability play a central role in the activity of these PAR analogues. The AE parameter, which is most clearly associated with the proposed mechanism for activity and most closely correlated with activity values of the phenolic PAR analogues, is also the most difficult parameter to calculate. Hence, it is of interest from a QSAR standpoint to note that the simpler, more indirect measurements of electronic charge density do a reasonablejob of accounting for the activity variation. Some justification for this is provided by examination of the UHF Mulliken spin densities for the free-radical species. The unpaired electron is found to be primarily localized on the phenolic oxygen, in agreement with the findings of Koymans et al. (18). Hence, it is not unreasonable to find an association between electronic charge parameters for the phenolic oxygen and the relative ease of formation and resulting stability of the free-radical oxygen species. The phenolic PAR analogues in Table I have been segregated from the nonphenolic PAR analogues in Table I1 in order to facilitate a meaningful relative comparison of electronic and energetic parameters within each of the two subclasses. The inactivity of PAR15 and -17 in Table I1 was attributed to the requirement of a phenyl oxygen for activity in the earlier Holme et al. study (26). This is consistent with the mechanistic hypothesis in this study. It was also observed in that earlier study that the ethoxy PAR analogues (PAR5, -13, and -16) were consistently less active than their phenolic counterparts (PAR2, -8, and -12) but displayed the same activity ordering. The mechanism of activity of the ethoxy PAR analogues is unlikely to proceed through a radical dissociation involving an ethyl radical leaving group, and thus, other possible pathways must be considered. It is plausible to hypothesize the formation of hydroxyl intermediates which subsequently undergo H atom radical abstraction. However, such intermediates would be formed with less than 100% efficiency, accounting for the lower observed activity than that of the phenolic counterpart in Table I. Three calculated properties are listed in Table I1 for the ethoxy PAR analogues: QC values represent the partial atomic charge on the phenyl carbon attached to the ethoxy oxygen, and QO and AE values are the same as listed in Table I for the corresponding phenoxy radicals (PARB, -8, and -12,respectively). Each of these parameters displays variation which parallels the observed activity variation, consistent with the discussion pertaining to the corresponding phenols in Table I. As previously mentioned, recent experimental evidence has indicated that paracetamol destroys a tyrosyl radical species at the active site of the ribonucleotide reductase (14). This inhibition of normal ribonucleotide reductase activity could ultimately account for the experimentally observed inhibition of replicative DNA synthesis. If this were the case, the activities and calculated parameters in Table I should reflect the relative reactivities of PAR analogues with respect to the tyrosyl radical species and indicate whether the active PAR analogues would be likely to quench the tyrosyl radical species on the basis of a simple competitive reaction involving an H atom exchange. The tyrosyl portion of the ribonucleotide reductase active site is approximately modeled in this study with an unbound tyrosine molecule. While this simplified treatment neglects possible structural and topological requirements

Structure-Activity Study of Paracetamol Analogues Table 111. Listing of Molecular Structure and Calculated Properties for Tyrosine and Phenoxy Radical Derivative structure QC (OH)" QO (radlb hEc tyrosine

0

It

0.375

-0.501

0.128

,WH CH&H

Chem. Res. Toxicol., Vol. 4, No. 2, 1991 155

Registry NO.PAR1,95-86-3; PAR2,123-30-8; PAR3,95-55-6; PAR4,150-754; PAR5,156-43-4; PAR6,106-44-5; PAR7,591-27-5; PAR8, 103-90-2; PARS, 123-31-9; PAR10, 614-80-2; PAR11, 22900-79-4; PAR12,108-95-2; PAR13,62-44-2; PAR14, 100-02-7; PAR15, 103-84-4; PARl6, 103-73-1; PAR17, 62-53-3; PARl8, 621-42-1; PAR19, 6337-56-0.

References

Q'"' OH

Partial atomic charge on phenolic carbon in parent PAR compound obtained from AMl//RHF/3-21G calculation. b Partial atomic charge on phenoxy radical oxygen obtained from UAMl// UHF/3-21G calculation. e Difference in total energy of dissociation products (phenoxy radical and H atom) and parent tyrosine compound in atomic units; energies obtained from UAMl//ROHF/321G result and AMl//RHF/B-PIG calculation, respectively.

of the active site, it is assumed to be sufficient for modeling relevant electronic parameters. The chemical structure and values of the calculated parameters for the tyrosine parent phenol and corresponding free-radical species are listed in Table 111. The calculated parameters QC and AI3 place the tyrosyl analogue squarely in the Group I11 "low" activity category in Table I, i.e., at the separation juncture between active and inactive PAR analogues. This is entirely consistent with the above mechanistic hypothesis and suggests that PAR analogues in Groups I and I1 are more likely to undergo H atom loss than the tyrosyl phenol and, thus, effectively quench the tyrosyl radical species required for normal DNA replicative synthesis. In contrast, the Group IV PAR analogues are predicted to be less likely to undergo H atom loss and, hence, do not effectively quench the tyrosyl radical species or inhibit replicative DNA synthesis.

Conclusion The present theoretical study has provided support for the mechanistic hypothesis proposed and has determined molecular parameters that could be potentially useful for activity prediction of paracetamol analogues. Clearly, the present analysis, in common with most SAR analyses, has modeled a tremendously simplified version of the actual biological problem due to lack of detailed knowledge and practical considerations. For instance, PAR energy calculations assume isolated gas-phase molecules, not molecules in an aqueous biological medium. Possible metabolic intermediates such as further hydroxylated forms of the paracetamol analogues and transesterification products were not explicitly considered. Also, interaction with the biological receptor, which would mediate and affect the relative energetics of radical formation within the PAR series, was neglected. In addition, the calculations fail to take into account the ability of some of the resonancestabilized PAR radicals to undergo rapid dimerization and polymerization reactions (1s21). This could account for the seeming discrepancy between the AI3 value for PAR11 and the experimental observation that this radical, which is unable to form such coupling products due to the two ortho methyl groups, forms one of the most stable phenol radicals (19, 21). In spite of the simplifications of the present study, however, the results of the current analysis do seem to support a plausible mechanistic hypothesis and can provide a departure point for further studies.

Acknowledgment. We thank Ms. Christine Bjerrge for technical assistance and Dr. Sidney D. Nelson, Dr. James R. Rabinowitz, Dr. Stephen Nesnow, and Dr. Gilda Loew for valuable comments and suggestions.

(1) Prescott, L. F., Wright, N., Roscoe, P., and Brown, S. S. (1971) Plasma paracetamol half-life and hepatic necrosis in patients with paracetamol overdosage. Lancet i, 519-522. (2) Mitchell, J. R., Jollow, D. J., Potter, W. Z., Gillette, J. R., and Brodie, B. B. (1973) Acetaminophen-induced hepatic necrosis. IV. Protective role of glutathione. J. Pharmacol. Exp. Ther. 187, 211-217. (3) Floren, C.-H., Thesleff, P., and Nilsson, A. (1987) Severe liver damage caused by therapeutic doses of acetaminophen. Acta Med. Scand. 222, 285-288. (4) Flaks, A., and Flaks, B. (1983) Induction of liver cell tumors in IF mice by paracetamol. Carcinogenesis 4, 363-368. (5) Flaks, A., Flaks, B., and Shaw, A. (1985) Induction by paracetamol of bladder and liver tumors in the rat. Acta Pathol. Microbiol. Immunol. Scand., Sect. A 93, 367-377. (6) Sandler, D. P., Smith, J. C., Weinberg, C. R., Buckalew, Y. M., Dennis, V. W., Blythe, W. B., and Burgess, W. P. (1989) Analgesic use and chronic renal disease. N. Engl. J. Med. 320,1238-1243. (7) Sakaki, M., Yoshida, S., and Hiraga, K. (1983) Additional effect of acetaminophen on the mutagenicity and clastogenicity of Nmethyl-N-nitro-N-nitrosoguanidine in cultured Chinese hamster CHO-KI cells. Mutat. Res. 122, 367-372. (8) Dybing, E., Holme, J. A., Gordon, W. P., Saderlund, E. J., Dahlin, D. C., and Nelson, S. D. (1984) Genotoxicity studies with paracetamol. Mutat. Res. 138, 21-32. (9) Hongslo, J. K., Christensen, T., Brunborg, G., Bjarnstad, C., and Holme, J. A. (1988) Genotoxic effects of paracetamol in V79 Chinese hamster cells. Mutat. Res. 204, 333-343. (10) Kocisova, J., Rossner, P., Binkova, B., Bavorova, H., and Sram, R. J. (1988) Mutagenicity studies on paracetamol in human volunteers. I. Cytogenetic analysis of peripheral lymphocytes and lipid peroxidation in plasma. Mutat. Res. 209, 161-165. (11) Corbett, M. D., Corbett, B. R., Hannothiaux, M.-H., and Quintana, S. J. (1989) Metabolic activation and nucleic acid binding of acetaminophen and related arylamine substrates by the respiratory burst of human granulocytes. Chem. Res. Toxicol. 2, 260-266. (12) Eiche, A., Bexell, G., and Sandelin, K. (1990) Genotoxicity of p-aminophenol in somatic and germ line cells of Drosophila melanogaster. Mutat. Res. 240,87-92. (13) Hongslo, J. K., Bjarnstad, C., Schwarze, P. E., and Holme, J. A. (1989) Inhibition of replicative DNA synthesis by paracetamol in V79 Chinese hamster cells. Toxicol. in Vitro 3, 13-20. (14) Hongslo, J. K., Bjorge, C., Schwarze, P. E., Brogger, A., Mann, G., Thelander, T., and Holme, J. A. (1990) Paracetamol inhibits replicative DNA synthesis and induces sister-chromatid exchange and chromosomal aberrations by inhibition of ribonucleotide reductase. Mutagenesis 5, 475-480. (15) Kjaller Larsen, I., Sjoberg, B.-M., and Thelander, L. (1982) Characterization of the active site of ribonucleotide reductase of Escherichia coli, bacteriophage T4 and mammalian cells by inhibition studies with hydroxyurea analogues. Eur. J. Biochem. 125, 75-81. (16) Loew, G . H., and Goldblum, A. (1985) Metabolic activation and toxicity of acetaminophen and related analogs: A theoretical study. Mol. Pharmacol. 27, 375-386. (17) Van de Straat, R., de Vries, J., de Boer, H. J. R., Vromans, R. M., and Vermeulen, N. P. E. (1987) Relationship between paracetamol binding to and its oxidation by two cytochromes P-450 isozymes-a proton nuclear magnetic resonance and spectrophotometric study. Xenobiotica 17, 1-9. (18) Koymans, L., Van Lenthe, J. H., Van de Straat, R., Donn6-Op den Kelder, G . M., and Vermeulen, N. P. E. (1989) A theoretical study on the metabolic activation of paracetamol by cytochrome P-450 Indications for a uniform oxidation mechanism. Chem. Res. Toxicol. 2, 60-66. (19) Fischer, V., and Mason, R. P. (1984) Stable free radical and benzoquinone imine metabolites of an acetaminophen analogue. J. Riol. Chem. 259, 10284-10288. (20) West, P. R., Harman, L. S., Josephy, P. D., and Mason, R. P. (1984) Acetaminophen: enzymatic formation of a transient phe-

156 Chem. Res. Toricol., Vol. 4, No. 2, 1991 noxyl free radical. Biochem. Pharmacol. 33, 2933-2936. (21) Fischer, V., West, P. R., Nelson, S. D., Harvison, P. J., and Mason, R. P. (1985) Formation of 4-aminophenoxyl free radical from the acetaminophen metabolite N-acetyl-p-benzoquinone imine. J. Biol. Chem. 260, 11446-11450. (22) Van de Straat, R., Vromans, R. M., Bosman, P., de Vries, J., and Vermeulen, N. P. E. (1988) Cytochrome P-450 mediated oxidation of substrates by electron-transfer; Role of oxygen radicals and of 1- and 2-electron oxidation of paracetamol. Chem.-Biol. Interact. 64, 267-280. (23) Dewar, M. J. S., Zeobisch, E. G., and Healy, E. F. (1985) AM1: A new general purpose quantum mechanical molecular model. J. Am. Chem. SOC.107,3902-3909. (24) Frisch, M. J., Binkley, J. S., Schlegel, H. B., Raghavachari, K., Melius, C. F., Martin, R. L., Stewart, J. J. P., Bobrowicz, F. W.,

Richard et al. Rohlfing, C. M., Kahn, L. R., Defrees, D. J., Seeger, R., Whiteside, R. A., Fox, D. J., Fleuder, E. M.,and Pople, J. A., (1984) GAUSSIAN 86, Carnegie-Mellon Quantum Chemistry Publishing Unit, Pittsburgh, PA. (25) Clark, T. (1985) A Handbook of Computational Chemistry: A practical guide to chemical structure and energy calculations, Wiley-Interscience, New York. (26) Holme, J. A., Hongslo, J. K., Bj~rnstad,C., Harvison, P. J., and Nelson, S. D. (1988) Toxic effects of paracetamol and related structures in V79 Chinese hamster cells. Mutagenesis 3, 51-56. (27) Hansch, C., and Leo, A. (1979) Substituent Constants for Correlation Analysis in Chemistry and Biology, Wiley-Interscience, New York. Hammett, L. P. (1935) Effect of structure on the reactions of organic compounds. Chem. Rev. 17, 125-139.