Chem -3iol. Interactions, 26 (1979) 1 L-25 0 EIsevier/North-HollandScientific Publishers Ltd.
MECHANISMS OF ACYMON OF C AN INVESTIGATION USING MUT.
:il
AROMATIC AMINES: BACTEBJA
JOHN D. SCRIBNER, SHARON R. FISK alrd NORMA K. SCRIBNER Pucifie ~~~~h~~~ Research Fo~~a~io~, I 102 Columbia Street, Seattle. WA 98104 (U.S.A.)
(Received October 3rd, 1978) (Revision received December 8th, 1978) (Accepted lkcember l!Xh, 1978)
SUMMARY
The ‘mutagenicities of groups of N-acetoxy-N-arylacetatnides, nitroarenes, arylamides and arylamines were determined in the $uZmoneEZa typhimurium tester strains TA93, TA1538, TA100, TA1535 and TA1537. Three broad classes of mutagenic activity were found, interpreted as follows: class A, including 2naphthyliumine, produced essentially only base-pair substitution wi~out induction of error-prone repair; class B, including ~-~ij~obip~enyl, caused considerable induction of error-prone repair, accompanied by a lower level of frame shifting; class C, including N-acetoxy-2-acetamidofluorene, produced high levels of frame shifting, with some induction of error-prone repair. Correlation of these results with known reactions of certain aromatic amine derivatives with nucleosides and nucleic acids, and with molecular orbital calculations, su.ggests that ‘i;he effect of class A k produced by small aromatic groups attached to extranucle~ he~roatoms in DNA bases, the effect of class B is caused by large aromatic groups attached to extranuclear heteroatoms or by arylamines attached to C-8 of guanine, while the effect of class C is caused by arylamides attached to C-S of guanine, probably rotating into the helix, as proposed by others. The data also suggest that the N-acetoxy-N-arylacetamides are generally useful models for ultimate metabelites derived in vivo, even if the in vnro metabolites do not carry an acetyl group. Finally, there is a roergh correlation between the sum of reversions induced in TA98 and TAlOO by the N-acetoxy-IV-arylacetamides and their previously determined local carcinogenicities. There is a poor correlation between mut~enic~ty in any one tester strain and carcinogenicity.
INTRODiJCTlON &tempts to establish a coherent genie effects found among aromatic
explanation for the variety of carcinoamines have generally met with frustra-
12 tion. Some of the problems observed, for example, are the observation that some amines are more active when acetylated [l] while others lose activity on acetylation [Z] and that several aromatic amides with comparable activity toward the rat mammary gland have widely disparate activities toward the rat liver [1,3,4]. Yet the hypothesis that N-hydroxyla”;ion followed by esterif’cation are the essential activating steps seems to hold true for most compounds considered separately [ 51. Work in this laboratory has been directed toward elucidating factors in the reactions of ultimate forms of aromatic amine carcinogens which may contribute to understanding the biological differences mentioned above. We have now learned enough about the chemistry of these compounds, both on an experimental and theoretical basis, to attempt to relate these chemical data to biological data. Although it is too soon to expect to obtain useful correlations with whole animal data, studies on bacterial mutagenesis may contribute to understanding the following questions: (1) Do N-acetoxy&arylacetamides serve as accurate models for the ultimate mutagens generated in vivo? (2) Are there qualitative differences among aromatic amines in their mutagenicity? (3) Can specific types of chemical lesions be associated with certain types of mutagenicity? (4) Is there a quantitative association between mutagenicity and tumor induction for aromatic amines or related compounds? In this paper we show that limited conclusions may be drawn in all of these areas, and that microbial mutagenesis may be a valuable aid in interpreting the significance of chemical events induced by carcinogens. It may even be possible 1.0 reverse the process and predict specific types of chemical lesions from knowledge of mutagenicity in SuZinonellatester strains. METHODS
Preparation of compounds. 1-Nitronaphthalene, 4-nitrobiphenyl, 2=nitroguorene and 2-nitronaphthalene are commercially available (Aldrich) and were purified by recrystallization. 2-Mitrophenanthrene and 4-nitrostilbene were prepared according to published procedures [ ,7 1. 2-Nitroanthracene was prepared by oxidation of 2-aminoanthracene, as follows: 3.7 g of 2-aminoanthracene suspended in 150 ml of @roleurn ether was treated with 3.3 g of mchloroperbenzoic acid and stirred 3 h at room temperature. The reaction mixture was then poured onto an alumina column packed in pet,roIleum ether and chromatographed in benzene. 111 mg of nitroanthracene was obtained, m.p., 179-181°C (lit. 180-181” [S]) with appropriate WV and IR spectra. N-Acetoxy-N-arylacetamides were prepared from the nitro arenes by reduction with ammonia and bydrogen sulfide, followed by peracetylation with acetic anhydride in ethyl acetate and trietbylamine. Amines not obtainable commercially were prepared by reduction of the corresponding nitro compounds with hydrazine and palladium/charcoal in ethanol. Amides weye prepared by acetylation with acetic anhydride in pyridine. N-Acetoxyamides were purified chromatography on silica gel in ethylacetate/benzene (7 : 3). Other compounds were recrystallized as necessary.
13 Mutagbzesis assuy. The assay was carried out as described by Ames, et al. [9], but using 35mm plates, rather than 100 mm, with proportionate reductions in components. S-9 fractions were obtained from male Fischer rats averaging 165 g and treated with Arochlor 1254 as described previously [ 91. A total of 25 g of liver was processed, and the S-9 fractions stored at -30°C in l-ml portions. Spontaneous mutation rates were consistent with those found by Ames and others, and are subtracted from the original data to give the figures reported. Plates for a single class of compounds were poured together and seeded together. Plates for dose-response curves on toxic compounds were also poured and seeded together. Toxic compounds were identified either by maxima in the initial dose-response curves or by reduction of the mutation level to below background (usually no colonies at all). All data are means of at least two plates and all background determinations were run in duplicate. RESULTS
In Tables I and II are summarized the relative potencies of the different compounds in the tester strains TASS, TA1538, TAlOO and TAX535 Testing was also carried out in TAY537, but this strain behaved in almost all
SOOk-
;
-1
200
Fig. 1. Mutagenesis by N-acetoxy-N-arylacetamides in Salmonella typhimurium tester strain TA1538. Aryl substituents are: P, 2-phenanthryl; F, 2-fluorenyl; X, 4-biphenyl; S, 4-stilbenyl; l-N, I-naphthyl; 2-N, 2-naphthyl,. 35 mm culture dishes were used. Fig, 2. Mutagenesis by nitroarenes iri Salmonella substituents as shown in Fig. 1. A, 2-anthracenyl.
typhimurium
t*~Y,erstrain TA1533.
Aryl
Fig. 3. ~uta~en~is by aryhmines in §~rn~~ella ~phjrn~ri~rn tester strain TAl!j3$. Aryl substituents as shown in Fig. 2. Curves shown are for optimum concentratioils of added S-9 protein: P, 0.5 mglplate; F, 1.0; X, 2.0; S, 0.5; 2-N, 2.0; A, 0.5. S-9 concentrations lower than 0.5 mg protein/plate were not tested. 1-Naphthylamine was inactive in the dose range sb own _ Fig. 4. Mutagenesis by N-aryiacetamides TAI538. Aryl substituents as shown in Fig. tions of added S-9 protein: P 0.5 mglplate; tions lower than 0.5 mg protein/plate were inactive in the dose range shown.
in Salmonella typhimurium tester strain 2. Curves shown are for optimum concentraF, 2.0; X, 2.0; S, 2.0; A, 0.5. S-9 coneentranot tested. I- and Z-~phthylace~mide were
cases as a less sensitive version of TA1538. A single exception was found, in which 4-aminobiphenyl was as active toward TA153’7 as toward TA98. Doseresponse curves are shown in Figs. l-4, for TA1538 only. Response varied with dose in much the same fashion for all four strains discussed herein. Without added S-9, the amines and amides produced no mutations above background. Because of the varying degrees of toxicity displayed, the data presented in the tables represent m~~rnurn rev./nmol values, based on at least 30 colonies/plate above background, or twice background if fewer colonies were observed at all doses. In Figs. 5 and 6 are shown dose-response curves for mutagenesis by nitro~~racene and nitrophenanthrene in all four strains. In Fig. 7 is a dose-response curve for mutagenesis by N-acetoxy4acetamidostilhene. Figs. 5-7 are included to indicate the high toxicity of 2%nitroanthracene and ~-nitrophen~threne, and the slightly lower toxicity of ~~acetoxy-~-ac~~~id~,~st~bene. It may also be noted from Figs. 3 and 4 that the opthmum concentration of S-9 protein varied for the different substrates. AS expected, the ~-ac~?toxy~ides and nit.roarenes were mutagenic without activation by liver homogenate. The patterns established in Table I are
2-Fluorene It-Biphenyl 2-Phenanthrene 4-Stilbene 1 -Naphthalene 2-Naphthalene 2-Anthracene
51.1 1.4 128 2.3 0.2 0.2 892
8.0 0.3 14.9 2.8 0.2 6.5
TA98
Strain
Mutagenesis by nitroarenes
2-Fluorene 4-Biphenyl 2-Phen~th~ne 4-Stilbene 1-Nanttbake 2-Naphthalene
Aryl substituent
(12X/0.5) (X74/25) (71.0/0.125) (5215) (22125) (33/25) (100/0.025)
(700/25) (25125) (250/5) (46l5) (4/5) (ll/si)
{colonies/pg
IN SALMONELLA
65.0 0.9 145 1.0 0.1 0.4 839
4.3 0.1 13.2 0.4 0.2 0.1
(154/0.5) (114125) (81/0.125) (23/5) (16/26) (58125) (94/0.0%5)
(7515) (12/25) (223151 (615) (4/S) (14ii;;z:
TAl538
on plateb
BY A’-ACETOXY-N-ARYLACETAMIDRS
No enzymes added. Revertants/nmol
MUTAGENESIS
TABLE I
7.9 4.2 62.4 11.7 0.8 4.3 1133
0.3 0.2 3.7 6.8 0.4 4.6
TAlOO
(188/5) (522f25) (35/0.125) (26/0.5) (111/25) (139.5) (254/0.05)
(23125) (18/25) (6215) (113f5) (B/5) (94/5)
TYPHIMURIUM
0.0 0.0 2.7 0.0 0.0 3.7 794
(O/25} (O/25) (3/0.25) (O/25) (2/25) (108/5) (8910.02S)
0.1(12/25) 0.1 (6125) 0.3 (5/5) 0.0 (015) 0.3 (6/5) 2.8 (5815)
TA1535
86.4 66.8 726 2.9 0.2 1.6 77
T~98
Strain
(coionie&g
(472/1/l) (40/1;2) (188/0.06/0.6) (751612) (6/S/2) (11/l/l) (400/1/l)
2Fiuorene 4-Biphenyl 2-Phenanthrene 4Stillbene 1-Naphthaiene 2L-Naphthalene 2-Anthracene
16.1 (360/5/2) 1.0 (24/6/2) 14.1(60/1/l) 2.5 (631512) 0.0 (O/5/1) 0.0 (O/S/l) 92 (39/0.1/0.5)
26.8 0.3 32.7 2.2 9.0 0.0 109
(600/5/Z) (7/5/2) (139/1/l) (471612) (O/5/1) (O/5/1) (46/0.1/0.5)
92.8 (613/1/l) 4.7 (28/l/2) 566 (146/0.05/0,5) 2.4 (61/6/l) 0.1(4/5/0.6) 0.6 (20/6/2) 96 (49/0.1/0.5)
TA1538
compound on pIate/mg S-Q protein added).
Mu&genesis by N-wykrce tamides
IL-Fluorene 4-Biphenyl 2-Pbenanthrene 4-Sti&iritne I-Naphthalene 2-Naphthalene 2-Anthracene
Aryi substituent
Reyertants/nmol
MUTAGENESIS BY ARYLAMINES IN SALMON.ELLA ~Y~~Z~U~~~~
TABLE 11
7.0 1.8 17.0 13.6 0.9 0.0 79
42.3 11.8 102 12.6 S.i 10.7 77
(166/S/2) (42/6/2) (72/l/2) (67/1/l) (O/W ). (O/S/l) (34/0.1/2)
(234/1/l) (?C,‘l/l) (26/9.06/0.5) (324/5/l) (36/l/l) (76/l/2) (400/l/l)
TAlOO
(2/l/0*5) (S/6/2) (16/0.2/l) (O/6/1) (20/1/l) (103/l/2) (71/1/l)
0.0 (l/6/2) 0.1 (S/S/Z) 2.7 (11/l/2) 0.6 (2/l/l) 0.1(2/6/l) 0.0 (O/5/1) 70 (34/0.1/2)
0.4 0.2 16.9 0.0 2.9 14.7 13.8
TA1635
3a In . .;
M k 0 75
3
v
” 20(
M
0.2
Fig. 5. Mutagenicity of 2-nitro~~ra~e~e represent different tester strains as shown.
in Salmonella
I c.4 0.6 pglplatfd
~p~~rnu~urn.
I (x6
The curves
Fig. 6. Mutagenicity of 2-nitrophenanthrene in Salmonella typhimurium. The curves represent different tester stsahts as shown. 2-Nit~~en~th~ne was inactive in TA1535 in the dose range shown.
generally carried over into Table IB, Tkhich shows compounds requiring enzymatic activation for mutagenicity. The data collected in these tables can be broadly s~rn~~ as follows. If one considers the sum of the reversions obtained in TA98 plus TAlOO as an indicator of total mutagenic activity, one finds an order of mutagenicity, phenanthryl > fluorenyl > stilbenyl > biphenyl > 2-~aphthyi > l-naphthyl, which holds within each of the four
0
;,pl.%
32
40
Fig. 7. Mutagenicity of N-acetoxy-4”acetamidostilbene in &lmondlu typhimurium. curves represent different tester strains as shown. ~-A~etoxy~-acet~id~tilbe~e inactive in TA3.538 and TA1535.
The was
classes of derivatives. Exceptions occurred with 2-naphthylacetamide, which was inactive, with N-acetoxy-2naphthylacetamide, which was more active than _~-a~etoxy~4-a~e~idobiphenyl and with 4-~inobiphenyl, which was slightly more active than 4-aminostilbene. In considering all compounds relative to all of the tester strains, we find three broad classes of behavior:.The fluorene and phen~th~ne derivatives comprise one class, in which mutagenicity toward the four tester strains is in the order 98 = 1538 > 100 > 1535. Stilbene and biphenyl derivatives form the second class, with the order 100 > 98 * 1538 % 1535. ZNapnthyl derivatives and l-naphthylamine form a third class, with the order 100 = 1535 > 98 * 1538. The remaining compounds were either inactive or, in the case of the anthracene derivatives, showed high activity toward all four strains. Some exceptions in detail to these broad cl~sifi~ations will be considered in the discussion. DISCUSSION
Until now, mutagenesisdata for aromatic mines in $~~~u~e~~~ have been reported simply as values obtained from the most sensitive strain [lo] or used to indicate the role of various activating systems in the mutagenic process [ 11-141. However, since the nature of the muta~ons is known in part, and the chemistry of the reactions of the ultimate carcinogens is known in part, it is now possible from a comparative study to make a preliminary judgement as to the roles of various types of chemical lesions in mutagenesis and to evaluate the significance of reactions of model ultimate carcinogens. The latter is particularly important, because it is such models studied in vitro which give us the most information about the detailed chemical events taking place between c~~inoge~ and nucleic acid. Such physi~ochemi~~ studies in turn serve as the basis for determining the validity of any theoretical approaches to predicting the reactivity of ultimate carcinogens. Since only mutagenesis by tne ~-acetoxy-~-a~~~etamides (in this study) can be presumed to be independent of activating enzymes, differences in potency between different amines, amides or nitroarenes in a given tester strainshould be viewed with caution, since they may be due in part to differences in me~bolism of the pro-mutagens. While such metabolic differences clearly may be relevant to carcinogenicity in whole animals, they can obscure differences in the activities of the ultimate mutagens. On the other hand, we will assume that different tester strains metabolize a given nitroarene in the same way, and (naturally) that S-9 metabolism is independent of t’hetester strainpresent. Before attempting to answer the questions posed in the introduction, we should review the nature of the mutationd involved, and consider the types of events which can take place as a result of chemical lesions. Histidine requirement in TAX35 is caused by a base substilxtion in the gene coding for the first enzyme of histidine biosynt~sis [ 14 1. This substitution can be reverted either by a direct mutation or by a variety of suppressor mutations. TA1538 contains a frameshift mutation in the histidinol dehydrogenase
19
gem?,resulting from deletion of a base. Essentizd to both of these strains is a lack of excision repair. TAlOO (derived from TA1535) and TA98 (derived from TM5381 in addition have been altered by addition of a plasmid which seems to produce a much-increased level of error-prone recombinational repair f16]. In order for recombination repair to take place, a gap must be formed in one strand of the DNA. ~though the details of the loss of excision repair are not known, it is likely that the excision endonuclease is missing. Thus, any gap formation would not be due to incomplete excision repair, but to incomplete replication. While deletions might also arise from depurination followed by action of an apurinic endonuclease, none of the known or predicted reactions of aromatic nit~nium ions with purines are of the type known to lead to depurination. In other cases, depurination may be a significant source of deletions. Incomplete replication could result either from blockage of base pairing or blockage of template-dependent DNA polymerase. Thus, we can now look for lesions which cause base-pair substitution without gap fo~ation, which cause substi~tion with gap fo~ation, which cause frame shifting with gap formation, and which cause frame shifting without gap formation. Unless a lesion is small, attack on N-l of adenine or N-3 of cytosine will likely prevent base pairing, rather than merely cause substitution, while attack on N-7 of guanine will cause depurination. Thus, any of these three common lesion types will likely either be lethal or result in recombination repair and should accordingly result in few mutations in TA1535, but with many more in TAlOO. Attack on extranuclear hetcroatoms by compounds which can be accommodated in the grooves of the helix should result in efficient base-pair substitution without gap formation, since there should be little interference with either base pairing or with polymerase. Such lesions should produce about equal numbers of mutations in TA1535 and TAlOO. Reversion of a frame shift mutation should be a highly unlikely event during gap-filling, since it would require a sequence of bases to be inserted correctly, rather than a single base. Thus, greater mutagenicity in TA98 than in TA1538 should be a rare obs~~ation. If it should occur, it would indicate extensive gap formation, with consequent inefficient frame shifting, and should be accompanied by a mutation level in TAlOO which is higher than that in TA98. This last prediction was developed out of consideration of the data presented here, but is fully confirmed by the results obtained previ~u~~ly by McCann et al. [16]. The statements made above should provide an adequate background for interpreting the result obtained in this experiment. Let us return, then, to question one. Do N-acetoxy-N-arylacetamides 8elve as accurate models for mutagenic metabolites generated in viva? The answer, of course, assumes that the metabolites generated for the tester strains are like those formed by reductases or oxidases in whole animals. With respect to the spectrum of mutagenic responses produced in the four tester strains, the answer appears to be yes; the aeetoxyamides arc useful models for in viva ultimate forms. That is, it appears ,justifiable to study ;the reactions of ~-acetoxyamid~s in vivo in order to elucidate the chemistry Qf arylamitws
20 and arylacetamides in vivo. We have already noted in the results that the sum of mutagenic responses in TA98 and TAlOO generahy falls in the same order for all aryl substituents within each of the classes of derivatives. More importantly, the relative potencies toward the d.ifferent tester strains generally falls in the same order for each derivative of a given arene. While N-acetoxy-4-acetamidostilbene is unique in being considerably more potent toward TA98 than TA1538, it is like the remaining stilbene derivatives in being less active toward TA98 than toward TA109. The conclusion reached in this paragraph is critical, for it has already been shown that the principal species bound to rat liver DNA in vivo after feeding of 2-acetamidofluorene lacks the acetyl group. While there is a shift toward greater sensitivity in TAlOO in going from N-acetoxy-2-acetamidofluorene to 2-aminofluorene or 2-acetamidofluorene, the response is greatest in TA98 for all three compounds, as well as for nitrofluorene. A lesser shift is seen for the phenanthryl derivatives (also more active toward TA98 than TA100) and the stilbenyl derivatives (more active in TAlOO for all derivatives).. Essentially no shift in activity is observed for the 2-naphthyl derivatives, which suggests that particularly for these compounds of specifically human interest it will be valid and useful to investigate the chemistry of nucleic acid interactions with the more easily handled hydroxamic acid esters. The second question posed is whether there are qualitative differences among aromatic amines in their mutagenicities. Since there has been a strong tendency to interpret the reactions of 2acetamidofluorene as representative of those of aromatic amine carcinogens generally,, this question is also critical. In the results section, also, we have noted that there are at least three different classes of mutagenicity spectra among the compounds which we have tested. One of these includes a known human carcinogen (2naphthylamine) which does not show any tendency to produce mutations by stimulation of error-prone repair, a mechanism of action considered important by many for the induction of tumors. Another known human carcinogen, 4-aminobiphenyl, does induce error-prone repair, and also shows a greater tendency to produce frame shift mutations than does 2-naphthylamine. The most toxic compounds in human cells in culture [ 171, the stilbene derivatives, are clearly most mu nit as a result of inducing errorprone repair. It would seem that, fcr anced view of the factors involved in aromatic amine carcinogenesis, detailed chemical and genetic analysis should be undertaken for at least one compound in each class of mutagens, with a view toward critical corn ‘son of the results. The third question proceeds nd phenomenology and inquires into the genetic significance of the di types of chemical attack resultin administration of aromatic amines to an organism. g summarized the types of mutagenicity spectra which might be obta d the biochemical history which they could reflect, we should now attempt to known chemistry of ultimate forms of aromatic amines to the m observed here. (It must be noted that no adducts have SaImoneZlaDNA after treatment under conditions of the
21 The most carefully studied interaction is that between N-acetoxy-2acetan+ dofluorene and DNA. This carcinogen attaches the fluorenylacetamido nitrogen to C-8 of guanine about 80% of the time, and an aromatic carbon to N2 of guanine for almost all of the remainder [ 18-201. There is strong evidence that the C-8 lesion produces local strand separat%on in DNA, as shown by physical studies and studies on sensitivity to endonuclease S1 [21231. These results are interpreted by an insertion-denaturation model, in
which the fluorene ring is rotated into the stacked bases and the gu&ne is rotated out of the helix [ 231. In contrast, space-fUing molecular models indicate that the N2 lesion can be accommodated in the small groove of a rigid DNA double helix without
causing any distortion.
Such a lesion could,
however, be expected to produce an alteration in the hydrogen-bonding capability of the nitrogen, as shown by the tremendoes reduction in basicity in g_oing from aniline to diphenylamine. Mutagenicity by .JU-acetoxy2-acetamidofiuorene appears to be exclusively frame shift mutagenicity, associated with the insertion-denaturation model. The amine, amide and nitro derivative, however, all show considerable induction of gap-filling. Recalling that in vivo data show most of the lesions from the amide to lack the acetyl group [24,25], we can consider all three compounds to lead to comparable C-8 lesions. While this apparently still results in insertion, it can also be seen from molecuhr models that the fluorenylamine residue can be accommodated much more easily than the acetylated residue, so that the frequency of insertion is reduced. However, the non-inserted aromatic residue can now block DNA polymerase and cause gap formation, as reflected in the increased mutagenicity in TAlOO. Extending our analysis to biphenyl derivatives, we recall first that 92% cf the biphenylacetamido residues bound to DNA in rat liver were found to ,pe N2 adducts, and that 65% of the residues from the reaction of the &fate ester of N-hydroxy-4-acetamidobiphenyl with DNA were of this type [19]. If’ the N2 adduct is unimportant, the mutagenicity spectrum of biphenyl derivatives should be comparable to that of fluorene derivatives. Instead, there is a pronounced tendency toward gap-filling, with TAlOO now more responsive than TA98. Thus, we should probably conclude that attack on N2 of guanine by an aromatic group blocks polymerization, without yet commenting on the mechanism of this blockage. In vitro, N-acetoxy-2-acetamidophenanthrene attacks N6 of adenine much more readily than it does C-8 of guanine [ 261 and other adducts are known ‘Dut not identified. Again, frame shifting is seen, possibly due to C-8 adduct, *and thus indicating the mutagenic efficiency of this particular lesion. Howc+vver,there is now a tendency toward base-pair substitution not seen with the fluorene and biphenyl derivatives (mutations in TA1535). Lacking other kn.owledge, we can then associate tbis substitution mechanism with the p attack. At the same time, because of the high mutagenicity in TAlClO, we may also conclude that this lesion also blocks polymerization. All of the known reactions of h7-acetoxy-4-acetamidostilbene (ob with nucleosides in vitro) involve base pairing sites, specifically N-9 of
22 cytosine, N-l and N6 of adenine and N-l and O6 of guanine [ 27-291. A large group attached to any of these sites should completely block base-pairing, yet would not be expected to produce a frame shift very often. Hence, N-acetoxy-4-acetamidostilbene sbould be a relatively weak mutagen in TA1538, but much stronger in TA98. We must still explain the increase in activity toward TA1538 seen for the remaining derivatives. Molecular orbital calculations (Ref. 26; J.D. Scribner and S.R. Fisk, unpublished) indicate that stilbenylhydroxylamine should react to some extent at C-8 of guanine, which could lead to a low level of frame shift mutations by insertion. Irl all cases, predominant reactions at the base-pairing atoms should result in gap formation and high mutagenicity in TAlOO. The mutagenicity spectrum of the naphthyl derivatives offers a chance to predict chemistry from mutagenicity, since adducts of 2naphthylamine derivatives with DNA are as yet unknown. The similarity between the mutagenicities of the 2naphthyl derivatives in TAlOO and TA1535 suggests that these compounds attack base-pairing atoms without blocking base-pairing. Such a result has been obtained with small alkylating agents [ 161, but even benzyl chloride is much more effective in TAlOO than in TA1535 [ 161. Available evidence, obtained in vitro with nucleosides, shows that benzyl chloride in water attacks predominantly N-3 of cytosine [ 301, N6 of adenine and N-7 of guanine, (A. Dipple and R. Moschel, unpublished) while molecular orbital calculations (Rei. 26; J.D. Scribner and S.R. Fisk, unpublished) suggest that the N-acetyl-2-naphthylnitrenium ion attacks N* of guan\ne and N6 of adenine. Consideration of the arguments applied to the mutagenicities of the biphenyl and phenanthrene derivatives leads to the tentative proposal that the 2naphthyl ultimate carcinogens attack N6 on adenine, and possibly 0” on guani:- ? or N4 on cytosine. It is already known that the principal adduct of l-naphthylhydroxylamine with DNA is a N-O adduct with O6 of guanine [ 311. This finding agrees well with the base-pair substitution behavior of 1-naphthylarnine in this study. The high mutagenicity of anthracene derivatives in all strains indicates (again from comparison of TA1538 and TA1535) both effective insertion lesions and (at least for the amide) non-blocking attack on extranuclear atoms. The high mutagenicity of X-anthrylacetamide in TAl535 suggests that lesions of the type predicted for 2naphthylamine derivatives should also be found for this compound, but accompanying attack on C-8 of guanine. To date, we have been unable to prepare N-acetoxy-2-anthrylacetamide, presumably because of the (predicted) high reactisity of this compound [ 321. We now arrive at the last of the four questions posed in the beginning: is there a quantitative association between mutagenicity and tumor induction for aromatic amines or related compounds? Here, success is mixed. All of the N-acetoxy-N-arylacetamides tested here hcve also been tested for local carcmogenicity. (Ref. 32; E.C. Miller, J.A. Miller and J.D. Scribner, unpublished) Of these, N-acetoxy-2-acetamidophenanthrene and N-acetoxy4-acetamidostilbene were comparably potent, while N-acetoxy-l-acetamido-
23 naphthalene was inactive. N-Acetoxy-2-acetamidofluorene was si@ficmt,ly less potent than the phenanthryl and stilbenyl compounds, while N-acetoxy4-acetamidobiphenyl and N-acetoxy-2-acetamidonaphthalene were about equally weak. The SUM of TA98 plus TAlOO is reasonably well correlated with carcinogenieity, with N-acetoxy-2-acetamidofluorene and N-acetoxy4-acetamidostilbene having changed places. Mutagenicity in any single strain is less well correlated, suggesting that no particular type of mutation can be particularly associated with carcinogenicity. In rat liver, which is effectively the activating system for the mutagenicity assay of Narylacetamides, 2-acetamidophenanthrene was inactive when fed at the same level as a dose of 2-acetamidofluorene which produced 60% liver tumors [I], yet it is a more powerful mutagen than 2-acetamidofluorene. While no comparison should be drawn between compounds activated by added microsomes and those which either do not require activation or vrhich are activated by the bacteria, it seems fair to compare the amines and the amides. Here, although X-fluorenamine is roughly five times more potent than 2-acetamidofluorene as a mutagen, it has been found to be a more slowly acting carcinogen [ 11. Thus, while bacterial mutagenesis appears to be a valuable prescreen for potential carcinogens, mutagenesis alone is inadequate as an explanation fnr carcinogenicity . Several predictions arise from the considerations advanced above. The first is that DNA treated with fluorenylhydroxylamine should be less sensitive to SI endonuclease than DNA bearing the same number of residues from Second, the lesions which are responsible N-acetoxy-2-acetamidofluorene. for aromatic amine-induced frame shift miltations should not block DNA polymerase in assays for copying fidelity. Third, naphthyl groups or smaller attached to extranuclear atoms will not block DNA polymerase, but larger groups will (beginning with xenyl). Fourth, N-hydroxy-Znaphthylarnine and esters of N-hydroxy-2naphthylacetamide will be found to react with extranuclear atoms of purines (and possibly pyrimidines). Confirmation of these predictions will add substance to the explanations advanced hem for the while correction will point the way to more observed mutagenicities, accurate explanations. ACKNOWLEDGEMENT This
work
was
sczpported
by United
States
Public
Health
Service Grant
No. CA18632. REFERENCES J.A. Miller, R.B. Sandin, E.C. Miller and H.P. Rusch; The carcinogenicity of cornpounds related to 2-acetylaminofluorene. II. Variations in the bridges and the 2-substituent, Cancer Res., 15 (1955) 188. G.M. Conzelman and L.E. Flanders, The metabolism and carcinogenicity of 2_acetamidonaphthalene, Proc. West. Pharmacol. Sot., 15 (1972) 96. E.C. Miller, R.B. Sandin, J.A. Miller and H.P. Rusch, The carcinogenicity of com-
24
4
5
6
7 8 9
13
11
12
13
14
15
16
17
18
I9 20
21
pounds related to 2-acetylaminofluorene. III. Aminobiphenyl and benzidine derivatives. Cancer Res., 16 (1956) 535. R.A. Andersen, M. Enomoto, E.C. Miller and J.A. Miller, Carcinogenesis and in‘ ? ‘J tion of the Walker 256 tumor in the rat by fruns-4-acetylaminostilbene, its N-hydi metabolite, and related compounds, Cancer Res., 24 (1964) 128. E.C. Miller and J.A. Miller, The metabolism of chemical carcinogens to reactive electrophiles and their possible mechanisms of action in carcinogenesis, in: C.E. Searle (Ed), Chemical Carcinogens, American Chemical Society, Washington DC, 1976, pp. 737-762. E.C. Miller, P.D. Lotlikar, H.C. Pitot, T;L. Fletcher and J.A. Miller, N-hydroxy metabelites of 2-acetylaminophenanthrene and 7-fluoro-2acetylaminofluorene as proximate carcinogens in the rat, Cancer Res., 26 (1966) 2239. M.Calvin and R.E. Buckles, The fr.eedom of rotation about the carbon-@arbon double bond. ~wtain substituted stilbenes, J. Am. Chem. Sot., 62 (1940) 3324. J.D. SC.-.bner and J.A. Miller, Synthesis of 2nitroanthracene and N-hydroxy-2-anthrylau+me, J. Chem. Sot., (1965) 5377. B.N. Ames, J. McCann and E. Yamasaki, Methods for detecting carcinogens and mutagens with the Salmonella/mammalian~microsc~me mutagenicity test, Mutat. Res., 31 (1975) 347. J. McCann, E. Choi, E. Yamasaki and B.N. Ames, Detection of carcinogens as mutagens in the Salmonella/microsome test: assay of 300 chemicals, Proc. Nat. Acad. Sci. (U.S.A.), 72 (1975) 5135. D. McGregor, The relationship of 2-acetamidofluorene mutagenicity in plate tests with its in vivo liver cell component distribution and its carcinogenic potential, Mutat. Res., 30 (1975) 305. W.T. Allaben, C.E. Weeks, S.C. Louie, E.J. Lazear and C.M. King, Effects of chemical structure of N-hydroxy_N-fluorenylacylamides on arylhydroxamic acid acyltransferase, sulfotransferase and deacylase activities, and on mutations in Salmonella typhimurium, Proc. Am. Assoc. Cancer Res., 19 (1978) 235. C.E. Weeks, W.T. Allaben, S.C. Louie, E.J. Lazear and C.M. King, Role of ,arylhydroxamic acid acyltransferase in the mutagenicity of N-hydroxy-N-2fluorenylacetamide in Salmonella typhimurium, Cancer Res., 38 (1978) 613. H.A.J. Schut, R.J. Wirth and S.S. Thorgeirsson, Mutagenic activation of N-hydroxy2acetylaminofluorene in the Salmonella test system: the role of deacetylation by liver and kidney fractions from mouse and rat, Mol. Pharmacol., 14 (1978) 682. B.N. Ames, F.D. Lee and W.E. Durston, An improved bacterial test system for the de&e&ion and classification of mutagens and carcinogens, Proc. Nat. Acad. Sci. (U.S.A.), 70 (1973) 782. J. McCann, N.E. Spingarn, J. Kobori and B.N. Ames, Detection of carcinogens as mutagens: bacterial tester strains with R factor plasmids, Proc. Nat. Acad. Sci. (U.S.A.), 82 (1975) 979. V.M. Maher, N. Birch, J.R.Ott and J.J. McCormick, Cytotoxicity of carcinogenic aromatic amides in normal and Xeroderma pigmentosum fibroblasts with different DNA repair capabilities, J. Natl. Cancer Inst., 54 (1975) 1287. E. Kriek, Persistent binding of a new reaction product of the carcinogen N-hydroxyN-2-acetylaminofluorene with guanine in rat liver DNA in vivo, Cancer Res., 32 (1972) 2042. E. Kriek and J.G. Westra, DNA bound forms of the carcinogen Lacetylaminofluorene, Abstr. Eur. Assoc. Cancer Res., (1973) 240. J.G. Westra, E. Kriek and H. Hittenhausen, Identification of the persistently bound form of the carcinogen N-acetyl-2-aminofluorene to rat liver DNA in vivo, Chem.Biol. Interact., 15 (1976) 149. J.F. Lf fevre, R.P.P. Fuchs and M.P. Daune, Comparative studies on the 7-iodo and ‘I-fluoro derivatives of N-acetoxy-N-2-acetylaminofluorene : binding sites on DNA and
25
22
23
24 25
26 27
28
29 30
31
32 33
conformation change of modified deoxytrinucleotides, Biochemistry, 17 (1978) 2561. H. Yamasaki, P. Pulk~bek, D. Grunberger and LB- Wai~~in, Differential excision from DNA of the C-8 and N2 guanosine adduets of N~c~l-Z~m~o~uorene by single strand-specific endonucleases, Cancer Res., 37 (1977) 3756. D. Grunberger and I.B. Weinstein, The base-displacement model: an explanation for the conformational and functional changes in nucleic acids mofidied by chemical carcinogens, in: J.M. Yuhas, R.W. Tennant and J.D. Regan (Eds.), Biology of Radiation Careinogenesis, Raven Press, New York, 1976, pp. 175-187. E. Kriek, Difference in binding of Z-ace~iaminofluorene to rat liver d~x~ibonuclei~ acid and ribosomal ribonucleic acid in u&o, Bioehem. Biophys. Acta, 161 (1968) 273. CC. Irving, R.A. Veazey and L.T. Russe!l, Po.ssible role of the glucuranide conjugate in the biochemical mechanism of binding of the carcinogen N-hydroxy-2-acetylaminofluorene to rat-liver deoxyribonucleic acid in vivo, Chem.-Bid. Interact., I (1969l 1970) 19. J-D. Scribner and N.K. Naimy, Adducts between the carcinogen Z~cet~idophen~threne and adenine and guanine of DNA, Cancer Res., 35 (1975) 1416. J.D. Scribner, D.L. Smith and J.A. McCloskey, Deamination of l-methylcytosine by the carcinogen N-acetoxy-4-acetamidostilbene: implications for hydrocarbon carcinogenesis, J. Org. Chem., 43 (1978) 2085. N.K. Scribner, D.L. Smith, K. Sehram, J.A. McCloskey and J.D. Scribner, Reactions of the carcinogen ~-acetoxy~~ce~mid~t~~ne wity nucleosides, Chem-Biol. Interact., 26 (1979) 27. N.K. Scribner and J.D. Scribner, Reactions of the carcinogen N-acetoxy-4-acetamidostilbene with polynucleotides in vitro, Chem.-Biol. Interact., 26 (1979) 47. R. Shapiro and S.J. Shiuey, Reactions of cytidine with 7-bromomethylbenz(a)anthracene, benzyl bromide, and p-methoxybenxyl bromide. Ratio of amino to 3 substitution, J. Org. Chem., 41(197 $1 1597. F.F. Kadlubar, J-A. Miller altd E.C. Miller, Guarql 06-arylamination and 06-arylation of DNA by the carcinogea N-hydroxy-1-naphthylamine, Cancer Res., 38 (1978) 3628. J.D. Scribner, J.A. Miller and E.C. Miller, Nucleophilic substitution on carcinogenic N-acetoxy-N-arylacetamides, Cancer Res., 30 (1970) 1570. E.C. Miller and J.A. Miller, Studies on the mechan~m of activation of aromatic amine and amide carcinogens to ultimate carcinogenic electrophilic reactants, Ann. NY. Acad. Sci., 163 (1969) 731.
MECHANISMS OF ACYMON OF C AN INVESTIGATION USING MUT.
:il
AROMATIC AMINES: BACTEBJA
JOHN D. SCRIBNER, SHARON R. FISK alrd NORMA K. SCRIBNER Pucifie ~~~~h~~~ Research Fo~~a~io~, I 102 Columbia Street, Seattle. WA 98104 (U.S.A.)
(Received October 3rd, 1978) (Revision received December 8th, 1978) (Accepted lkcember l!Xh, 1978)
SUMMARY
The ‘mutagenicities of groups of N-acetoxy-N-arylacetatnides, nitroarenes, arylamides and arylamines were determined in the $uZmoneEZa typhimurium tester strains TA93, TA1538, TA100, TA1535 and TA1537. Three broad classes of mutagenic activity were found, interpreted as follows: class A, including 2naphthyliumine, produced essentially only base-pair substitution wi~out induction of error-prone repair; class B, including ~-~ij~obip~enyl, caused considerable induction of error-prone repair, accompanied by a lower level of frame shifting; class C, including N-acetoxy-2-acetamidofluorene, produced high levels of frame shifting, with some induction of error-prone repair. Correlation of these results with known reactions of certain aromatic amine derivatives with nucleosides and nucleic acids, and with molecular orbital calculations, su.ggests that ‘i;he effect of class A k produced by small aromatic groups attached to extranucle~ he~roatoms in DNA bases, the effect of class B is caused by large aromatic groups attached to extranuclear heteroatoms or by arylamines attached to C-8 of guanine, while the effect of class C is caused by arylamides attached to C-S of guanine, probably rotating into the helix, as proposed by others. The data also suggest that the N-acetoxy-N-arylacetamides are generally useful models for ultimate metabelites derived in vivo, even if the in vnro metabolites do not carry an acetyl group. Finally, there is a roergh correlation between the sum of reversions induced in TA98 and TAlOO by the N-acetoxy-IV-arylacetamides and their previously determined local carcinogenicities. There is a poor correlation between mut~enic~ty in any one tester strain and carcinogenicity.
INTRODiJCTlON &tempts to establish a coherent genie effects found among aromatic
explanation for the variety of carcinoamines have generally met with frustra-
12 tion. Some of the problems observed, for example, are the observation that some amines are more active when acetylated [l] while others lose activity on acetylation [Z] and that several aromatic amides with comparable activity toward the rat mammary gland have widely disparate activities toward the rat liver [1,3,4]. Yet the hypothesis that N-hydroxyla”;ion followed by esterif’cation are the essential activating steps seems to hold true for most compounds considered separately [ 51. Work in this laboratory has been directed toward elucidating factors in the reactions of ultimate forms of aromatic amine carcinogens which may contribute to understanding the biological differences mentioned above. We have now learned enough about the chemistry of these compounds, both on an experimental and theoretical basis, to attempt to relate these chemical data to biological data. Although it is too soon to expect to obtain useful correlations with whole animal data, studies on bacterial mutagenesis may contribute to understanding the following questions: (1) Do N-acetoxy&arylacetamides serve as accurate models for the ultimate mutagens generated in vivo? (2) Are there qualitative differences among aromatic amines in their mutagenicity? (3) Can specific types of chemical lesions be associated with certain types of mutagenicity? (4) Is there a quantitative association between mutagenicity and tumor induction for aromatic amines or related compounds? In this paper we show that limited conclusions may be drawn in all of these areas, and that microbial mutagenesis may be a valuable aid in interpreting the significance of chemical events induced by carcinogens. It may even be possible 1.0 reverse the process and predict specific types of chemical lesions from knowledge of mutagenicity in SuZinonellatester strains. METHODS
Preparation of compounds. 1-Nitronaphthalene, 4-nitrobiphenyl, 2=nitroguorene and 2-nitronaphthalene are commercially available (Aldrich) and were purified by recrystallization. 2-Mitrophenanthrene and 4-nitrostilbene were prepared according to published procedures [ ,7 1. 2-Nitroanthracene was prepared by oxidation of 2-aminoanthracene, as follows: 3.7 g of 2-aminoanthracene suspended in 150 ml of @roleurn ether was treated with 3.3 g of mchloroperbenzoic acid and stirred 3 h at room temperature. The reaction mixture was then poured onto an alumina column packed in pet,roIleum ether and chromatographed in benzene. 111 mg of nitroanthracene was obtained, m.p., 179-181°C (lit. 180-181” [S]) with appropriate WV and IR spectra. N-Acetoxy-N-arylacetamides were prepared from the nitro arenes by reduction with ammonia and bydrogen sulfide, followed by peracetylation with acetic anhydride in ethyl acetate and trietbylamine. Amines not obtainable commercially were prepared by reduction of the corresponding nitro compounds with hydrazine and palladium/charcoal in ethanol. Amides weye prepared by acetylation with acetic anhydride in pyridine. N-Acetoxyamides were purified chromatography on silica gel in ethylacetate/benzene (7 : 3). Other compounds were recrystallized as necessary.
13 Mutagbzesis assuy. The assay was carried out as described by Ames, et al. [9], but using 35mm plates, rather than 100 mm, with proportionate reductions in components. S-9 fractions were obtained from male Fischer rats averaging 165 g and treated with Arochlor 1254 as described previously [ 91. A total of 25 g of liver was processed, and the S-9 fractions stored at -30°C in l-ml portions. Spontaneous mutation rates were consistent with those found by Ames and others, and are subtracted from the original data to give the figures reported. Plates for a single class of compounds were poured together and seeded together. Plates for dose-response curves on toxic compounds were also poured and seeded together. Toxic compounds were identified either by maxima in the initial dose-response curves or by reduction of the mutation level to below background (usually no colonies at all). All data are means of at least two plates and all background determinations were run in duplicate. RESULTS
In Tables I and II are summarized the relative potencies of the different compounds in the tester strains TASS, TA1538, TAlOO and TAX535 Testing was also carried out in TAY537, but this strain behaved in almost all
SOOk-
;
-1
200
Fig. 1. Mutagenesis by N-acetoxy-N-arylacetamides in Salmonella typhimurium tester strain TA1538. Aryl substituents are: P, 2-phenanthryl; F, 2-fluorenyl; X, 4-biphenyl; S, 4-stilbenyl; l-N, I-naphthyl; 2-N, 2-naphthyl,. 35 mm culture dishes were used. Fig, 2. Mutagenesis by nitroarenes iri Salmonella substituents as shown in Fig. 1. A, 2-anthracenyl.
typhimurium
t*~Y,erstrain TA1533.
Aryl
Fig. 3. ~uta~en~is by aryhmines in §~rn~~ella ~phjrn~ri~rn tester strain TAl!j3$. Aryl substituents as shown in Fig. 2. Curves shown are for optimum concentratioils of added S-9 protein: P, 0.5 mglplate; F, 1.0; X, 2.0; S, 0.5; 2-N, 2.0; A, 0.5. S-9 concentrations lower than 0.5 mg protein/plate were not tested. 1-Naphthylamine was inactive in the dose range sb own _ Fig. 4. Mutagenesis by N-aryiacetamides TAI538. Aryl substituents as shown in Fig. tions of added S-9 protein: P 0.5 mglplate; tions lower than 0.5 mg protein/plate were inactive in the dose range shown.
in Salmonella typhimurium tester strain 2. Curves shown are for optimum concentraF, 2.0; X, 2.0; S, 2.0; A, 0.5. S-9 coneentranot tested. I- and Z-~phthylace~mide were
cases as a less sensitive version of TA1538. A single exception was found, in which 4-aminobiphenyl was as active toward TA153’7 as toward TA98. Doseresponse curves are shown in Figs. l-4, for TA1538 only. Response varied with dose in much the same fashion for all four strains discussed herein. Without added S-9, the amines and amides produced no mutations above background. Because of the varying degrees of toxicity displayed, the data presented in the tables represent m~~rnurn rev./nmol values, based on at least 30 colonies/plate above background, or twice background if fewer colonies were observed at all doses. In Figs. 5 and 6 are shown dose-response curves for mutagenesis by nitro~~racene and nitrophenanthrene in all four strains. In Fig. 7 is a dose-response curve for mutagenesis by N-acetoxy4acetamidostilhene. Figs. 5-7 are included to indicate the high toxicity of 2%nitroanthracene and ~-nitrophen~threne, and the slightly lower toxicity of ~~acetoxy-~-ac~~~id~,~st~bene. It may also be noted from Figs. 3 and 4 that the opthmum concentration of S-9 protein varied for the different substrates. AS expected, the ~-ac~?toxy~ides and nit.roarenes were mutagenic without activation by liver homogenate. The patterns established in Table I are
2-Fluorene It-Biphenyl 2-Phenanthrene 4-Stilbene 1 -Naphthalene 2-Naphthalene 2-Anthracene
51.1 1.4 128 2.3 0.2 0.2 892
8.0 0.3 14.9 2.8 0.2 6.5
TA98
Strain
Mutagenesis by nitroarenes
2-Fluorene 4-Biphenyl 2-Phen~th~ne 4-Stilbene 1-Nanttbake 2-Naphthalene
Aryl substituent
(12X/0.5) (X74/25) (71.0/0.125) (5215) (22125) (33/25) (100/0.025)
(700/25) (25125) (250/5) (46l5) (4/5) (ll/si)
{colonies/pg
IN SALMONELLA
65.0 0.9 145 1.0 0.1 0.4 839
4.3 0.1 13.2 0.4 0.2 0.1
(154/0.5) (114125) (81/0.125) (23/5) (16/26) (58125) (94/0.0%5)
(7515) (12/25) (223151 (615) (4/S) (14ii;;z:
TAl538
on plateb
BY A’-ACETOXY-N-ARYLACETAMIDRS
No enzymes added. Revertants/nmol
MUTAGENESIS
TABLE I
7.9 4.2 62.4 11.7 0.8 4.3 1133
0.3 0.2 3.7 6.8 0.4 4.6
TAlOO
(188/5) (522f25) (35/0.125) (26/0.5) (111/25) (139.5) (254/0.05)
(23125) (18/25) (6215) (113f5) (B/5) (94/5)
TYPHIMURIUM
0.0 0.0 2.7 0.0 0.0 3.7 794
(O/25} (O/25) (3/0.25) (O/25) (2/25) (108/5) (8910.02S)
0.1(12/25) 0.1 (6125) 0.3 (5/5) 0.0 (015) 0.3 (6/5) 2.8 (5815)
TA1535
86.4 66.8 726 2.9 0.2 1.6 77
T~98
Strain
(coionie&g
(472/1/l) (40/1;2) (188/0.06/0.6) (751612) (6/S/2) (11/l/l) (400/1/l)
2Fiuorene 4-Biphenyl 2-Phenanthrene 4Stillbene 1-Naphthaiene 2L-Naphthalene 2-Anthracene
16.1 (360/5/2) 1.0 (24/6/2) 14.1(60/1/l) 2.5 (631512) 0.0 (O/5/1) 0.0 (O/S/l) 92 (39/0.1/0.5)
26.8 0.3 32.7 2.2 9.0 0.0 109
(600/5/Z) (7/5/2) (139/1/l) (471612) (O/5/1) (O/5/1) (46/0.1/0.5)
92.8 (613/1/l) 4.7 (28/l/2) 566 (146/0.05/0,5) 2.4 (61/6/l) 0.1(4/5/0.6) 0.6 (20/6/2) 96 (49/0.1/0.5)
TA1538
compound on pIate/mg S-Q protein added).
Mu&genesis by N-wykrce tamides
IL-Fluorene 4-Biphenyl 2-Pbenanthrene 4-Sti&iritne I-Naphthalene 2-Naphthalene 2-Anthracene
Aryi substituent
Reyertants/nmol
MUTAGENESIS BY ARYLAMINES IN SALMON.ELLA ~Y~~Z~U~~~~
TABLE 11
7.0 1.8 17.0 13.6 0.9 0.0 79
42.3 11.8 102 12.6 S.i 10.7 77
(166/S/2) (42/6/2) (72/l/2) (67/1/l) (O/W ). (O/S/l) (34/0.1/2)
(234/1/l) (?C,‘l/l) (26/9.06/0.5) (324/5/l) (36/l/l) (76/l/2) (400/l/l)
TAlOO
(2/l/0*5) (S/6/2) (16/0.2/l) (O/6/1) (20/1/l) (103/l/2) (71/1/l)
0.0 (l/6/2) 0.1 (S/S/Z) 2.7 (11/l/2) 0.6 (2/l/l) 0.1(2/6/l) 0.0 (O/5/1) 70 (34/0.1/2)
0.4 0.2 16.9 0.0 2.9 14.7 13.8
TA1635
3a In . .;
M k 0 75
3
v
” 20(
M
0.2
Fig. 5. Mutagenicity of 2-nitro~~ra~e~e represent different tester strains as shown.
in Salmonella
I c.4 0.6 pglplatfd
~p~~rnu~urn.
I (x6
The curves
Fig. 6. Mutagenicity of 2-nitrophenanthrene in Salmonella typhimurium. The curves represent different tester stsahts as shown. 2-Nit~~en~th~ne was inactive in TA1535 in the dose range shown.
generally carried over into Table IB, Tkhich shows compounds requiring enzymatic activation for mutagenicity. The data collected in these tables can be broadly s~rn~~ as follows. If one considers the sum of the reversions obtained in TA98 plus TAlOO as an indicator of total mutagenic activity, one finds an order of mutagenicity, phenanthryl > fluorenyl > stilbenyl > biphenyl > 2-~aphthyi > l-naphthyl, which holds within each of the four
0
;,pl.%
32
40
Fig. 7. Mutagenicity of N-acetoxy-4”acetamidostilbene in &lmondlu typhimurium. curves represent different tester strains as shown. ~-A~etoxy~-acet~id~tilbe~e inactive in TA3.538 and TA1535.
The was
classes of derivatives. Exceptions occurred with 2-naphthylacetamide, which was inactive, with N-acetoxy-2naphthylacetamide, which was more active than _~-a~etoxy~4-a~e~idobiphenyl and with 4-~inobiphenyl, which was slightly more active than 4-aminostilbene. In considering all compounds relative to all of the tester strains, we find three broad classes of behavior:.The fluorene and phen~th~ne derivatives comprise one class, in which mutagenicity toward the four tester strains is in the order 98 = 1538 > 100 > 1535. Stilbene and biphenyl derivatives form the second class, with the order 100 > 98 * 1538 % 1535. ZNapnthyl derivatives and l-naphthylamine form a third class, with the order 100 = 1535 > 98 * 1538. The remaining compounds were either inactive or, in the case of the anthracene derivatives, showed high activity toward all four strains. Some exceptions in detail to these broad cl~sifi~ations will be considered in the discussion. DISCUSSION
Until now, mutagenesisdata for aromatic mines in $~~~u~e~~~ have been reported simply as values obtained from the most sensitive strain [lo] or used to indicate the role of various activating systems in the mutagenic process [ 11-141. However, since the nature of the muta~ons is known in part, and the chemistry of the reactions of the ultimate carcinogens is known in part, it is now possible from a comparative study to make a preliminary judgement as to the roles of various types of chemical lesions in mutagenesis and to evaluate the significance of reactions of model ultimate carcinogens. The latter is particularly important, because it is such models studied in vitro which give us the most information about the detailed chemical events taking place between c~~inoge~ and nucleic acid. Such physi~ochemi~~ studies in turn serve as the basis for determining the validity of any theoretical approaches to predicting the reactivity of ultimate carcinogens. Since only mutagenesis by tne ~-acetoxy-~-a~~~etamides (in this study) can be presumed to be independent of activating enzymes, differences in potency between different amines, amides or nitroarenes in a given tester strainshould be viewed with caution, since they may be due in part to differences in me~bolism of the pro-mutagens. While such metabolic differences clearly may be relevant to carcinogenicity in whole animals, they can obscure differences in the activities of the ultimate mutagens. On the other hand, we will assume that different tester strains metabolize a given nitroarene in the same way, and (naturally) that S-9 metabolism is independent of t’hetester strainpresent. Before attempting to answer the questions posed in the introduction, we should review the nature of the mutationd involved, and consider the types of events which can take place as a result of chemical lesions. Histidine requirement in TAX35 is caused by a base substilxtion in the gene coding for the first enzyme of histidine biosynt~sis [ 14 1. This substitution can be reverted either by a direct mutation or by a variety of suppressor mutations. TA1538 contains a frameshift mutation in the histidinol dehydrogenase
19
gem?,resulting from deletion of a base. Essentizd to both of these strains is a lack of excision repair. TAlOO (derived from TA1535) and TA98 (derived from TM5381 in addition have been altered by addition of a plasmid which seems to produce a much-increased level of error-prone recombinational repair f16]. In order for recombination repair to take place, a gap must be formed in one strand of the DNA. ~though the details of the loss of excision repair are not known, it is likely that the excision endonuclease is missing. Thus, any gap formation would not be due to incomplete excision repair, but to incomplete replication. While deletions might also arise from depurination followed by action of an apurinic endonuclease, none of the known or predicted reactions of aromatic nit~nium ions with purines are of the type known to lead to depurination. In other cases, depurination may be a significant source of deletions. Incomplete replication could result either from blockage of base pairing or blockage of template-dependent DNA polymerase. Thus, we can now look for lesions which cause base-pair substitution without gap fo~ation, which cause substi~tion with gap fo~ation, which cause frame shifting with gap formation, and which cause frame shifting without gap formation. Unless a lesion is small, attack on N-l of adenine or N-3 of cytosine will likely prevent base pairing, rather than merely cause substitution, while attack on N-7 of guanine will cause depurination. Thus, any of these three common lesion types will likely either be lethal or result in recombination repair and should accordingly result in few mutations in TA1535, but with many more in TAlOO. Attack on extranuclear hetcroatoms by compounds which can be accommodated in the grooves of the helix should result in efficient base-pair substitution without gap formation, since there should be little interference with either base pairing or with polymerase. Such lesions should produce about equal numbers of mutations in TA1535 and TAlOO. Reversion of a frame shift mutation should be a highly unlikely event during gap-filling, since it would require a sequence of bases to be inserted correctly, rather than a single base. Thus, greater mutagenicity in TA98 than in TA1538 should be a rare obs~~ation. If it should occur, it would indicate extensive gap formation, with consequent inefficient frame shifting, and should be accompanied by a mutation level in TAlOO which is higher than that in TA98. This last prediction was developed out of consideration of the data presented here, but is fully confirmed by the results obtained previ~u~~ly by McCann et al. [16]. The statements made above should provide an adequate background for interpreting the result obtained in this experiment. Let us return, then, to question one. Do N-acetoxy-N-arylacetamides 8elve as accurate models for mutagenic metabolites generated in viva? The answer, of course, assumes that the metabolites generated for the tester strains are like those formed by reductases or oxidases in whole animals. With respect to the spectrum of mutagenic responses produced in the four tester strains, the answer appears to be yes; the aeetoxyamides arc useful models for in viva ultimate forms. That is, it appears ,justifiable to study ;the reactions of ~-acetoxyamid~s in vivo in order to elucidate the chemistry Qf arylamitws
20 and arylacetamides in vivo. We have already noted in the results that the sum of mutagenic responses in TA98 and TAlOO generahy falls in the same order for all aryl substituents within each of the classes of derivatives. More importantly, the relative potencies toward the d.ifferent tester strains generally falls in the same order for each derivative of a given arene. While N-acetoxy-4-acetamidostilbene is unique in being considerably more potent toward TA98 than TA1538, it is like the remaining stilbene derivatives in being less active toward TA98 than toward TA109. The conclusion reached in this paragraph is critical, for it has already been shown that the principal species bound to rat liver DNA in vivo after feeding of 2-acetamidofluorene lacks the acetyl group. While there is a shift toward greater sensitivity in TAlOO in going from N-acetoxy-2-acetamidofluorene to 2-aminofluorene or 2-acetamidofluorene, the response is greatest in TA98 for all three compounds, as well as for nitrofluorene. A lesser shift is seen for the phenanthryl derivatives (also more active toward TA98 than TA100) and the stilbenyl derivatives (more active in TAlOO for all derivatives).. Essentially no shift in activity is observed for the 2-naphthyl derivatives, which suggests that particularly for these compounds of specifically human interest it will be valid and useful to investigate the chemistry of nucleic acid interactions with the more easily handled hydroxamic acid esters. The second question posed is whether there are qualitative differences among aromatic amines in their mutagenicities. Since there has been a strong tendency to interpret the reactions of 2acetamidofluorene as representative of those of aromatic amine carcinogens generally,, this question is also critical. In the results section, also, we have noted that there are at least three different classes of mutagenicity spectra among the compounds which we have tested. One of these includes a known human carcinogen (2naphthylamine) which does not show any tendency to produce mutations by stimulation of error-prone repair, a mechanism of action considered important by many for the induction of tumors. Another known human carcinogen, 4-aminobiphenyl, does induce error-prone repair, and also shows a greater tendency to produce frame shift mutations than does 2-naphthylamine. The most toxic compounds in human cells in culture [ 171, the stilbene derivatives, are clearly most mu nit as a result of inducing errorprone repair. It would seem that, fcr anced view of the factors involved in aromatic amine carcinogenesis, detailed chemical and genetic analysis should be undertaken for at least one compound in each class of mutagens, with a view toward critical corn ‘son of the results. The third question proceeds nd phenomenology and inquires into the genetic significance of the di types of chemical attack resultin administration of aromatic amines to an organism. g summarized the types of mutagenicity spectra which might be obta d the biochemical history which they could reflect, we should now attempt to known chemistry of ultimate forms of aromatic amines to the m observed here. (It must be noted that no adducts have SaImoneZlaDNA after treatment under conditions of the
21 The most carefully studied interaction is that between N-acetoxy-2acetan+ dofluorene and DNA. This carcinogen attaches the fluorenylacetamido nitrogen to C-8 of guanine about 80% of the time, and an aromatic carbon to N2 of guanine for almost all of the remainder [ 18-201. There is strong evidence that the C-8 lesion produces local strand separat%on in DNA, as shown by physical studies and studies on sensitivity to endonuclease S1 [21231. These results are interpreted by an insertion-denaturation model, in
which the fluorene ring is rotated into the stacked bases and the gu&ne is rotated out of the helix [ 231. In contrast, space-fUing molecular models indicate that the N2 lesion can be accommodated in the small groove of a rigid DNA double helix without
causing any distortion.
Such a lesion could,
however, be expected to produce an alteration in the hydrogen-bonding capability of the nitrogen, as shown by the tremendoes reduction in basicity in g_oing from aniline to diphenylamine. Mutagenicity by .JU-acetoxy2-acetamidofiuorene appears to be exclusively frame shift mutagenicity, associated with the insertion-denaturation model. The amine, amide and nitro derivative, however, all show considerable induction of gap-filling. Recalling that in vivo data show most of the lesions from the amide to lack the acetyl group [24,25], we can consider all three compounds to lead to comparable C-8 lesions. While this apparently still results in insertion, it can also be seen from molecuhr models that the fluorenylamine residue can be accommodated much more easily than the acetylated residue, so that the frequency of insertion is reduced. However, the non-inserted aromatic residue can now block DNA polymerase and cause gap formation, as reflected in the increased mutagenicity in TAlOO. Extending our analysis to biphenyl derivatives, we recall first that 92% cf the biphenylacetamido residues bound to DNA in rat liver were found to ,pe N2 adducts, and that 65% of the residues from the reaction of the &fate ester of N-hydroxy-4-acetamidobiphenyl with DNA were of this type [19]. If’ the N2 adduct is unimportant, the mutagenicity spectrum of biphenyl derivatives should be comparable to that of fluorene derivatives. Instead, there is a pronounced tendency toward gap-filling, with TAlOO now more responsive than TA98. Thus, we should probably conclude that attack on N2 of guanine by an aromatic group blocks polymerization, without yet commenting on the mechanism of this blockage. In vitro, N-acetoxy-2-acetamidophenanthrene attacks N6 of adenine much more readily than it does C-8 of guanine [ 261 and other adducts are known ‘Dut not identified. Again, frame shifting is seen, possibly due to C-8 adduct, *and thus indicating the mutagenic efficiency of this particular lesion. Howc+vver,there is now a tendency toward base-pair substitution not seen with the fluorene and biphenyl derivatives (mutations in TA1535). Lacking other kn.owledge, we can then associate tbis substitution mechanism with the p attack. At the same time, because of the high mutagenicity in TAlClO, we may also conclude that this lesion also blocks polymerization. All of the known reactions of h7-acetoxy-4-acetamidostilbene (ob with nucleosides in vitro) involve base pairing sites, specifically N-9 of
22 cytosine, N-l and N6 of adenine and N-l and O6 of guanine [ 27-291. A large group attached to any of these sites should completely block base-pairing, yet would not be expected to produce a frame shift very often. Hence, N-acetoxy-4-acetamidostilbene sbould be a relatively weak mutagen in TA1538, but much stronger in TA98. We must still explain the increase in activity toward TA1538 seen for the remaining derivatives. Molecular orbital calculations (Ref. 26; J.D. Scribner and S.R. Fisk, unpublished) indicate that stilbenylhydroxylamine should react to some extent at C-8 of guanine, which could lead to a low level of frame shift mutations by insertion. Irl all cases, predominant reactions at the base-pairing atoms should result in gap formation and high mutagenicity in TAlOO. The mutagenicity spectrum of the naphthyl derivatives offers a chance to predict chemistry from mutagenicity, since adducts of 2naphthylamine derivatives with DNA are as yet unknown. The similarity between the mutagenicities of the 2naphthyl derivatives in TAlOO and TA1535 suggests that these compounds attack base-pairing atoms without blocking base-pairing. Such a result has been obtained with small alkylating agents [ 161, but even benzyl chloride is much more effective in TAlOO than in TA1535 [ 161. Available evidence, obtained in vitro with nucleosides, shows that benzyl chloride in water attacks predominantly N-3 of cytosine [ 301, N6 of adenine and N-7 of guanine, (A. Dipple and R. Moschel, unpublished) while molecular orbital calculations (Rei. 26; J.D. Scribner and S.R. Fisk, unpublished) suggest that the N-acetyl-2-naphthylnitrenium ion attacks N* of guan\ne and N6 of adenine. Consideration of the arguments applied to the mutagenicities of the biphenyl and phenanthrene derivatives leads to the tentative proposal that the 2naphthyl ultimate carcinogens attack N6 on adenine, and possibly 0” on guani:- ? or N4 on cytosine. It is already known that the principal adduct of l-naphthylhydroxylamine with DNA is a N-O adduct with O6 of guanine [ 311. This finding agrees well with the base-pair substitution behavior of 1-naphthylarnine in this study. The high mutagenicity of anthracene derivatives in all strains indicates (again from comparison of TA1538 and TA1535) both effective insertion lesions and (at least for the amide) non-blocking attack on extranuclear atoms. The high mutagenicity of X-anthrylacetamide in TAl535 suggests that lesions of the type predicted for 2naphthylamine derivatives should also be found for this compound, but accompanying attack on C-8 of guanine. To date, we have been unable to prepare N-acetoxy-2-anthrylacetamide, presumably because of the (predicted) high reactisity of this compound [ 321. We now arrive at the last of the four questions posed in the beginning: is there a quantitative association between mutagenicity and tumor induction for aromatic amines or related compounds? Here, success is mixed. All of the N-acetoxy-N-arylacetamides tested here hcve also been tested for local carcmogenicity. (Ref. 32; E.C. Miller, J.A. Miller and J.D. Scribner, unpublished) Of these, N-acetoxy-2-acetamidophenanthrene and N-acetoxy4-acetamidostilbene were comparably potent, while N-acetoxy-l-acetamido-
23 naphthalene was inactive. N-Acetoxy-2-acetamidofluorene was si@ficmt,ly less potent than the phenanthryl and stilbenyl compounds, while N-acetoxy4-acetamidobiphenyl and N-acetoxy-2-acetamidonaphthalene were about equally weak. The SUM of TA98 plus TAlOO is reasonably well correlated with carcinogenieity, with N-acetoxy-2-acetamidofluorene and N-acetoxy4-acetamidostilbene having changed places. Mutagenicity in any single strain is less well correlated, suggesting that no particular type of mutation can be particularly associated with carcinogenicity. In rat liver, which is effectively the activating system for the mutagenicity assay of Narylacetamides, 2-acetamidophenanthrene was inactive when fed at the same level as a dose of 2-acetamidofluorene which produced 60% liver tumors [I], yet it is a more powerful mutagen than 2-acetamidofluorene. While no comparison should be drawn between compounds activated by added microsomes and those which either do not require activation or vrhich are activated by the bacteria, it seems fair to compare the amines and the amides. Here, although X-fluorenamine is roughly five times more potent than 2-acetamidofluorene as a mutagen, it has been found to be a more slowly acting carcinogen [ 11. Thus, while bacterial mutagenesis appears to be a valuable prescreen for potential carcinogens, mutagenesis alone is inadequate as an explanation fnr carcinogenicity . Several predictions arise from the considerations advanced above. The first is that DNA treated with fluorenylhydroxylamine should be less sensitive to SI endonuclease than DNA bearing the same number of residues from Second, the lesions which are responsible N-acetoxy-2-acetamidofluorene. for aromatic amine-induced frame shift miltations should not block DNA polymerase in assays for copying fidelity. Third, naphthyl groups or smaller attached to extranuclear atoms will not block DNA polymerase, but larger groups will (beginning with xenyl). Fourth, N-hydroxy-Znaphthylarnine and esters of N-hydroxy-2naphthylacetamide will be found to react with extranuclear atoms of purines (and possibly pyrimidines). Confirmation of these predictions will add substance to the explanations advanced hem for the while correction will point the way to more observed mutagenicities, accurate explanations. ACKNOWLEDGEMENT This
work
was
sczpported
by United
States
Public
Health
Service Grant
No. CA18632. REFERENCES J.A. Miller, R.B. Sandin, E.C. Miller and H.P. Rusch; The carcinogenicity of cornpounds related to 2-acetylaminofluorene. II. Variations in the bridges and the 2-substituent, Cancer Res., 15 (1955) 188. G.M. Conzelman and L.E. Flanders, The metabolism and carcinogenicity of 2_acetamidonaphthalene, Proc. West. Pharmacol. Sot., 15 (1972) 96. E.C. Miller, R.B. Sandin, J.A. Miller and H.P. Rusch, The carcinogenicity of com-
24
4
5
6
7 8 9
13
11
12
13
14
15
16
17
18
I9 20
21
pounds related to 2-acetylaminofluorene. III. Aminobiphenyl and benzidine derivatives. Cancer Res., 16 (1956) 535. R.A. Andersen, M. Enomoto, E.C. Miller and J.A. Miller, Carcinogenesis and in‘ ? ‘J tion of the Walker 256 tumor in the rat by fruns-4-acetylaminostilbene, its N-hydi metabolite, and related compounds, Cancer Res., 24 (1964) 128. E.C. Miller and J.A. Miller, The metabolism of chemical carcinogens to reactive electrophiles and their possible mechanisms of action in carcinogenesis, in: C.E. Searle (Ed), Chemical Carcinogens, American Chemical Society, Washington DC, 1976, pp. 737-762. E.C. Miller, P.D. Lotlikar, H.C. Pitot, T;L. Fletcher and J.A. Miller, N-hydroxy metabelites of 2-acetylaminophenanthrene and 7-fluoro-2acetylaminofluorene as proximate carcinogens in the rat, Cancer Res., 26 (1966) 2239. M.Calvin and R.E. Buckles, The fr.eedom of rotation about the carbon-@arbon double bond. ~wtain substituted stilbenes, J. Am. Chem. Sot., 62 (1940) 3324. J.D. SC.-.bner and J.A. Miller, Synthesis of 2nitroanthracene and N-hydroxy-2-anthrylau+me, J. Chem. Sot., (1965) 5377. B.N. Ames, J. McCann and E. Yamasaki, Methods for detecting carcinogens and mutagens with the Salmonella/mammalian~microsc~me mutagenicity test, Mutat. Res., 31 (1975) 347. J. McCann, E. Choi, E. Yamasaki and B.N. Ames, Detection of carcinogens as mutagens in the Salmonella/microsome test: assay of 300 chemicals, Proc. Nat. Acad. Sci. (U.S.A.), 72 (1975) 5135. D. McGregor, The relationship of 2-acetamidofluorene mutagenicity in plate tests with its in vivo liver cell component distribution and its carcinogenic potential, Mutat. Res., 30 (1975) 305. W.T. Allaben, C.E. Weeks, S.C. Louie, E.J. Lazear and C.M. King, Effects of chemical structure of N-hydroxy_N-fluorenylacylamides on arylhydroxamic acid acyltransferase, sulfotransferase and deacylase activities, and on mutations in Salmonella typhimurium, Proc. Am. Assoc. Cancer Res., 19 (1978) 235. C.E. Weeks, W.T. Allaben, S.C. Louie, E.J. Lazear and C.M. King, Role of ,arylhydroxamic acid acyltransferase in the mutagenicity of N-hydroxy-N-2fluorenylacetamide in Salmonella typhimurium, Cancer Res., 38 (1978) 613. H.A.J. Schut, R.J. Wirth and S.S. Thorgeirsson, Mutagenic activation of N-hydroxy2acetylaminofluorene in the Salmonella test system: the role of deacetylation by liver and kidney fractions from mouse and rat, Mol. Pharmacol., 14 (1978) 682. B.N. Ames, F.D. Lee and W.E. Durston, An improved bacterial test system for the de&e&ion and classification of mutagens and carcinogens, Proc. Nat. Acad. Sci. (U.S.A.), 70 (1973) 782. J. McCann, N.E. Spingarn, J. Kobori and B.N. Ames, Detection of carcinogens as mutagens: bacterial tester strains with R factor plasmids, Proc. Nat. Acad. Sci. (U.S.A.), 82 (1975) 979. V.M. Maher, N. Birch, J.R.Ott and J.J. McCormick, Cytotoxicity of carcinogenic aromatic amides in normal and Xeroderma pigmentosum fibroblasts with different DNA repair capabilities, J. Natl. Cancer Inst., 54 (1975) 1287. E. Kriek, Persistent binding of a new reaction product of the carcinogen N-hydroxyN-2-acetylaminofluorene with guanine in rat liver DNA in vivo, Cancer Res., 32 (1972) 2042. E. Kriek and J.G. Westra, DNA bound forms of the carcinogen Lacetylaminofluorene, Abstr. Eur. Assoc. Cancer Res., (1973) 240. J.G. Westra, E. Kriek and H. Hittenhausen, Identification of the persistently bound form of the carcinogen N-acetyl-2-aminofluorene to rat liver DNA in vivo, Chem.Biol. Interact., 15 (1976) 149. J.F. Lf fevre, R.P.P. Fuchs and M.P. Daune, Comparative studies on the 7-iodo and ‘I-fluoro derivatives of N-acetoxy-N-2-acetylaminofluorene : binding sites on DNA and
25
22
23
24 25
26 27
28
29 30
31
32 33
conformation change of modified deoxytrinucleotides, Biochemistry, 17 (1978) 2561. H. Yamasaki, P. Pulk~bek, D. Grunberger and LB- Wai~~in, Differential excision from DNA of the C-8 and N2 guanosine adduets of N~c~l-Z~m~o~uorene by single strand-specific endonucleases, Cancer Res., 37 (1977) 3756. D. Grunberger and I.B. Weinstein, The base-displacement model: an explanation for the conformational and functional changes in nucleic acids mofidied by chemical carcinogens, in: J.M. Yuhas, R.W. Tennant and J.D. Regan (Eds.), Biology of Radiation Careinogenesis, Raven Press, New York, 1976, pp. 175-187. E. Kriek, Difference in binding of Z-ace~iaminofluorene to rat liver d~x~ibonuclei~ acid and ribosomal ribonucleic acid in u&o, Bioehem. Biophys. Acta, 161 (1968) 273. CC. Irving, R.A. Veazey and L.T. Russe!l, Po.ssible role of the glucuranide conjugate in the biochemical mechanism of binding of the carcinogen N-hydroxy-2-acetylaminofluorene to rat-liver deoxyribonucleic acid in vivo, Chem.-Bid. Interact., I (1969l 1970) 19. J-D. Scribner and N.K. Naimy, Adducts between the carcinogen Z~cet~idophen~threne and adenine and guanine of DNA, Cancer Res., 35 (1975) 1416. J.D. Scribner, D.L. Smith and J.A. McCloskey, Deamination of l-methylcytosine by the carcinogen N-acetoxy-4-acetamidostilbene: implications for hydrocarbon carcinogenesis, J. Org. Chem., 43 (1978) 2085. N.K. Scribner, D.L. Smith, K. Sehram, J.A. McCloskey and J.D. Scribner, Reactions of the carcinogen ~-acetoxy~~ce~mid~t~~ne wity nucleosides, Chem-Biol. Interact., 26 (1979) 27. N.K. Scribner and J.D. Scribner, Reactions of the carcinogen N-acetoxy-4-acetamidostilbene with polynucleotides in vitro, Chem.-Biol. Interact., 26 (1979) 47. R. Shapiro and S.J. Shiuey, Reactions of cytidine with 7-bromomethylbenz(a)anthracene, benzyl bromide, and p-methoxybenxyl bromide. Ratio of amino to 3 substitution, J. Org. Chem., 41(197 $1 1597. F.F. Kadlubar, J-A. Miller altd E.C. Miller, Guarql 06-arylamination and 06-arylation of DNA by the carcinogea N-hydroxy-1-naphthylamine, Cancer Res., 38 (1978) 3628. J.D. Scribner, J.A. Miller and E.C. Miller, Nucleophilic substitution on carcinogenic N-acetoxy-N-arylacetamides, Cancer Res., 30 (1970) 1570. E.C. Miller and J.A. Miller, Studies on the mechan~m of activation of aromatic amine and amide carcinogens to ultimate carcinogenic electrophilic reactants, Ann. NY. Acad. Sci., 163 (1969) 731.
