"IIIII M“ III III .ddldfiualm [I ' a I‘m It”??? ”I“ III, II!!!” “III III III " "I II Ilflmfl ‘ MI 'I IIEIIIIIi I 'III .'I I I III II IIUI II“. ' .3”. II yr. ' IIII’ . ,, 3". Itl u. I-” III I 1‘ V I u-IIIIIISII III-J II IIIIII"IIII III ’ f‘uuflh 'I II 44m 1 II IIIIW‘dI “(WW WI“ W u Hth '— ‘W‘. _ _-===:'_:5, _‘L W?—~::—_— _ ‘ - ' -.w¢m-_ —_———;~ ‘ ——_~:— 'H—W— ~=~- I .3 F; . . _~ ‘ ¢ - v. _ . -_ - :- ~ ~ ,. d _’ -' . _. W... - *3 v- . -—-—-‘ _— N... p. - ~___ ' -- _.___"~——_~‘ . . - .__ .. , ..I 4r . _..- _, N. firm 3 ~r. : , ':‘m-:'\' -a . -<. _ _ - ‘ , w— _’- — 0 *— fi - 4— _ A - 2; - ‘ __ I I I I I ‘ II VI III,“ I“ II'IIIII I"‘II I III" I III” II" “II II III; {E} III III i“III I II Ih:)%hh I“ III“, M I f III” II; I”II III? I‘ "1| IIIII". WI I I I “I"“*II"“ I .. : w:w . II'MLIIIIIE: IIIIII“II*' «II 17".- W””‘ "' 4i" I 2‘ ‘ ,III I... II... .JW.“ 'l3umfl “IN“ ’|5 “I. :E I. 'I “I"I'“ILIIHH infiflI"“lfi {i ‘ ‘ I' I “,t‘. III : '_" II 'IIIIIII I . :II - III I m III. II‘ “II “In . “W W." I III II ' ‘1 I]; ”lly“ “wl Ifi .IIW lo ”I Id. 'Eé ’IIIII It“: I “”IIIIII“ ”I IIII IIIIIIIIIII IIII L ' II“? I": “II II'I 'IIfJ- I .' III ”I I IIII! “IM' ”._ II. IINII‘I’ II" I" H “I I“ 22E.'f‘2;v“:“‘1 I ”HIINIIU 333 j,. ‘”I III 'Imagii IIHIHIIII| ~5§ijfa. as; I.-.» WI ., II}.qIfIJIJI'f‘J‘i" ' II I ‘0 l I I. 5' .It'I-'.';I I", “I . 4 ( II I ‘0 IIT'III IWW II}; ”I. I’I'I' II I I ' THEE" This is to certify that the thesis entitled MOLECULAR CORRELATES OF 2-ACETYLAMINOFLUORENE CARCINOGENICITY presented by Edward Lee Schwartz has been accepted towards fulfillment of the requirements for Ph . D . dggree in Pharmacology & Toxicology / / // Major professor 0‘ (7 (5 Date L (I 7 , 0-7 639 OVERDUE FINES ARE 25¢ PER DAY mm mm < ,‘ Return to book drop to remove this checkout from your record. MOLECULAR CORRELATES OF Z-ACETYLAMINOFLUORENE CARCINOGENICITY By Edward Lee Schwartz A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Pharmacology & Toxicology 1979 ABSTRACT Molecular Correlates of 2-Acetylaminofluorene Carcinogenicity by Edward Lee Schwartz It has long been speculated that perturbations of the genetic material are causally related to the initiation of ne0plastic changes. The purpose of this investigation was to examine the nature of the covalent binding of the hepatic carcinogen 2-acetylaminofluorene (AAF) to hepatic DNA. The following parameters were evaluated: the locali— zation of carcinogen binding within the hepatic genome, the effect of continued carcinogen ingestion on further carcinogen binding, and the functional consequences of carcinogen-induced DNA damage on DNA tran- scriptional capacities. Ingestion of AAF (0.05% w/w) for l or 2 weeks resulted in an apparent reduction in capacity to metabolically activate the carcino- gen. An 85% decrease was observed after 2 weeks when rats were tested with [14C]AAF, but only a 60% reduction occurred after injection of [3HJN-hydroxy-AAF, demonstrating a reduction in both activation steps in the formation of an ultimate carcinogen from AAF. Analysis of the initial binding of N-OH-AAF to DNA indicated that the adducts were nonrandomly distributed 2 hr after injection. Approximately 85% of the bound carcinogen was located on less than 25% Edward Lee Schwartz of the total nuclear DNA. The distribution was analyzed by fractiona- ting chromatin into eu- and heterochromatin regions. Chromatin was fragmented by sonication or DNAse II digestion, and fractionated by glycerol gradient centrifugation or selective MgCl2 precipitation. Initially, more carcinogen was bound to the less dense chromatin-DNA regions. Binding to DNA was also greater on expressed regions of the genome, the in_vi!g.template-expressed DNA having 16 times the carcino- gen modification of the template repressed DNA. Rate of carcinogen loss from DNA was also not uniform among the chromatin fractions. Loss from the highly condensed pelleted hetero- chromatin was significantly slower than that of the less condensed chromatin regions. This may partially reflect the greater pr0portion of carcinogen binding to the N2 position of guanine as compared to the C8 position of guanine on the DNA of this fraction. Carcinogen-guanine adducts were analyzed by thin layer chromatography. The nonrandom nature of AAF-DNA interactions was also reflected in an apparent clustering of the 2 different carcinogen-guanine adducts within the genome. Fragments of DNA containing over 90% of the carci- nogen at the C8 position or over 40% of the carcinogen at the N2 position could be isolated. Furthermore, there was an inverse correla- tion between the extent of carcinogen modification of the DNA of a particular chromatin fraction, and the percent of that modification occurring at the N2 position of guanine. A "cooperative base displacement" model has been proposed as a potential explanation of the clustering effect observed. Binding of the carcinogen to the C8 position of guanine has been demonstrated to alter the double stranded conformation of the DNA molecule and cause Edward Lee Schwartz local regions of denaturation. These regions are more susceptible to further carcinogen attack at the C8 position, but are relatively resis- tant to attack at the N2 position. A reduction in the amount of newly bound carcinogen located at the N2 position compared to the C8 position after 2 weeks of AAF ingestion provided further support for this model. Structural alterations of DNA and potential functional template damage was also suggested by studies of the transcription of DNA iso- lated from rats fed AAF (0.03%). E, ggli_RNA polymerase was used to transcribe the DNA under conditions in which re-initiation of RNA synthesis by the enzyme was blocked. A reduction of 40-50% in template capacity was observed after 4 days of AAF ingestion. This reduction was due to premature termination of RNA synthesis without a change in RNA synthesis initiation. Premature termination was observed on DNA purified from eu- and heterochromatin fractions obtained by sonication and glycerol gradient centrifugation. No reduction occurred on DNA of the pelleted heterochromatin fraction. This template did show a re- duction in the rate of polymerization of adenosine and uridine nucleo- tides into the growing RNA chain. All parameters returned to control values when animals were fed a basal diet for 7 days after 4 days of AAF ingestion. The major finding of this study was that steric factors, including the conformation of DNA in chromatin, influence the initial binding of carcinogen to DNA as well as the subsequent binding that occurs with continued carcinogen exposure. Interaction of carcinogen with DNA.EB yiyg_is neither a random nor a static phenomenon. With constant car- cinogen exposure, aspects of metabolic activation, DNA binding, and possibly repair are subject to continued alterations. ACKNOWLEDGEMENTS I am indebted to Denise Forbes, Joan Seger, and Sandra Moore for their excellent technical assistance in the work described in this dissertation. Thanks also to Dr. Emmett Braselton for the mass spectral analysis, to Dr. Fred Beland of the National Center for Toxicological Research for providing me with chemical samples and to Dr. Miriam Poirier of the National Cancer Institute for making available her unpublished findings. I would also like to express my gratitude to Diane Hummel for her routinely excellent help in pre- paring this, as well as many other manuscripts. I would like to convey Special thanks to my many friends here at MSU, and in particular to Dr. Jay Goodman, for providing support, encouragement and intellectual stimulation during the past four years. ii TABLE OF CONTENTS Page ACKNOWLEDGEMENTS ii LIST OF TABLES v LIST OF FIGURES vi LIST OF ABBREVIATIONS viii INTRODUCTION 1 1. Experimental Rationale 2. Carcinogenicity of 2-Acetylaminof1uorene (AAF)- ------- 9 3. Effects of 2-Acetylaminofluorene on DNA Structure ------ 12 4. Biochemical Effects of 2-Acetylaminofluorene Binding to Macromolecules 17 5. Chromatin Structure and Fractionation 20 6. Intragenomal Distribution of DNA Damage and Repair----- 23 7. Experimental Objectives 27 MATERIALS AND METHODS 29 1. Animals; Carcinogen Administration 29 2. Isolation of Chromatin 29 3. Chromatin Fractionation 30 4. Synthesis of Nascent RNA 33 5. Analysis of Carcinogen Binding to Chromatin Components- 34 6. Chromatography of Carcinogen-DNA Base Adducts -------- 35 7. Mass Spectral Analysis 36 8. Preparation of Purified DNA 36 9. Template Activity of DNA and Chromatin 37 10. Other Methods 38 RESULTS 39 1. Characterization of Chromatin Fractionation Procedures- 39 2. Assessment of the Purity of radiolabelled N-OH-AAF---- 49 3. Analysis of Carcinogen Binding and Loss 54 4. Analysis of Carcinogen Binding to DNA by Site 64 5. DNAse II: Mg012 Fractionation of Chromatin and Analysis of Carcinogen-DNA Binding 73 6. Effect of Continued AAF Ingestion on Carcinogen Binding to DNA 82 7. Effect of AAF Ingestion on DNA Transcriptional Capacity 85 iii TABLE OF CONTENTS (continued) Page DISCUSSI N 107 CONCLUSIONS 135 BIBLIOGRAPHY 139 iv Table 10 11 12 13 14 LIST OF TABLES Page Evidence for gene localization after DNAse II-MgCl2 chromatin fractionation 24 Composition of glycerol gradient chromatin fractions--— 40 Chromatin fractionation by selective MgCl2 precipita- tion 44 Comparison of glycerol gradient and MgCl2 fractionation 46 Analysis of initial carcinogen binding to cellular components 55 Initial binding of 3H—N-OH-AAF to chromatin DNA-------- 59 Rate of carcinogen loss from chromatin DNA and chroma— tin proteins 63 Thin layer chromatography of carcinogen-base adducts--- 74 Analysis of [3HJ-N-OH-AAF binding to DNA. DNAseII- MgCl2 fractionation of chromatin 79 Effect of continuous carcinogen ingestion on metabolic activation capacities 83 Effect of carcinogen ingestion of the sites of newly- bound carcinogen 84 Synthesis of RNA from control DNA samples isolated from chromatin fractions 89 RNA and protein contamination of purified DNA samples—- 93 Mathematical analysis of the rate of ribonucleotide polymerization 103 Figure 10 ll 12 13 14 15 16 17 18 19 LIST OF FIGURES Page Carcinogen-guanine adducts after AAF administration---- 13 Sedimentation profile of hepatic chromatin 32 Physical association of nascent RNA with chromatin ----- 42 Effect of resonication on the solubility of pelleted chromatin 47 Mass spectral analysis of [3H]-N-OH-AAF 50 Molecular ion distribution of [3H]-N-OH-AAF 52 Binding of carcinogen to chromatin components-—------—- 56 Distribution of DNA-bound carcinogen in glycerol gra- dient chromatin fractions 60 Loss of carcinogen from chromatin DNA (control diet)-—- 65 Loss of carcinogen from chromatin DNA (AAF diet)—- ----- 67 Thin layer chromatography of carcinogen-base adducts--- 69 Mass spectral analysis of N-(guanin—8-y1)-aminofluorene 71 Diagramatic representation of chromatin fractionation by DNAse II-MgCl2 75 Time course of chromatin fractionation by DNAseII-MgCl2 76 Correlation between extent of carcinogen binding to DNA and percent as N2 adduct 80 Analysis of DNA template capacity 86 Rate of RNA chain growth 90 Effect of AAF ingestion on DNA template capacity ------- 94 Correlation between RNA chain size and RNA polymerase turnover 96 vi LIST OF FIGURES (continued) Figure 20 21 22 23 Page Effect of protein contamination of DNA on RNA synthesis 99 Effect of RNA contamination of DNA on RNA synthesis---- 101 Kinetics of nucleotide triphosphate polyermization ----- 105 Extent of formation of dG-8-AAF, in vitro, in relation to the concentration of N-Ac-AAF 124 vii LIST OF ABBREVIATIONS 2-acetylaminofluorene N-OH-Z-acetylaminofluorene N-(guanosin-C8-y1)—2-aminof1uorene 3-(guanosin-Nz-yl)-2-acetylaminofluorene N-acetoxy-2-acetylaminof1uorene dimethylbenz[a]anthracene dimethylnitrosamine Ndmethyl-N-nitrosourea benzo[a]pyrene methylmethanesulfonate Adenine guanine cytidine uridine viii INTRODUCTION 1. Rationale The realization that some chemicals can be causally implicated in the induction of certain human cancers (146) has substantially shaped the state of cancer research today. Furthermore, rapid advances in the field of molecular biology in the last two decades have led to a multitude of speculations regarding potential mechanism(s) of these chemical carcinogens at the cellular and molecular levels. However, despite this enormous amount of research, in a recent review of the subject Heidelberger concludes that there are "very few final and definitive answers to questions about mechanisms [of chemical car- cinogenesis]" (79). In fact, a series of experiments performed by Berenblum and associates (13) over 30 years ago remain the corner- stone for much of the theoretical discussions of carcinogenesis under consideration at the present time. These experiments demonstrated that carcinogenesis could be divided into two stages, initiation and promotion. This pessimism reflects, to a large extent, the apparent com- plexity of the cancer problem. It has been suggested that cancer is not a single disease entity, but rather should be thought of as a heterogeneous classification of neoplastic growths (50). Likewise carcinogenesis probably results from alterations at several points in the regulation of cell proliferation and differentiation resulting in 2 more than one possible mechanism of carcinogen action (168,169). Different mechanisms could yield a variety of tumor phenotypes. This conclusion as much reflects a lack of complete understanding of the control of normal cell growth and differentiation as it does any definitive findings regarding the neoplastic cell. A question of primary concern in the study of the molecular biology of cancer is the nature of the critical macromolecular tar- get(s) for chemical carcinogenesis. Although the recent statement by the Millers that "there appears to be no firm basis at this time to decide the nature of the critical molecular change in any instance of chemical carcinogenesis" (148) echoes that of Heidelberger, altera- tions of DNA have received the most consideration. The interest in DNA as a molecular target has evolved directly from one of the oldest thoeries of carcinogenesis, the somatic mutation hypothesis, and has paralleled an increased understanding of the process of mutagenesis of single cells in culture. The somatic mutation theory of carcinogenesis is usually attri- buted to Boveri, who in 1914 suggested that the origin of a malignant tumor is "a certain abnormal chromatin constitution, the way in which it originates having no significance (22)." Similar hypotheses were presented by Tyzzer in 1916 (198) and Bauer in 1928 (11). The concept was mostly clearly elaborated by a committee consisting of S. Bayne- Jones, R.G. Harrison, C.C. Little, J. Northop and J.B. Murphy, who stated in 1938 that "The new property of the cell appears to develop suddenly, becomes a fixed character, and is transmitted to its de- scendants. It gives evidence of being a somatic muatation" (quoted in 3 ref. 168). During this time, the first experiments demonstrating the ability of purified chemicals to induce cancer in experimental animals were performed. The first synthetic carcinogen, dibenz(a,h)anthra- cene, was demonstrated by Kenaway and Hieger in 1930 (99), and shortly thereafter the carcinogenic hydrocarbon benzo(a)pyrene was isolated from coal tar and identified (34). However, there was little appre- ciation at the time for the relationship between chemicals in the environment and human cancer (45). Important experimental evidence and further conceptual support for the somatic mutation theory was provided by Berenblum (12,13) who, based on the work of Rous (180), proposed a two stage model of carcino- genesis. In these studies, animal skin was painted for a brief time with a subcarcinogenic dose of a chemical carcinogen (initiation) followed a variable time later with a relatively longer exposure to a second noncarcinogenic chemical (promoter). Exposure to either the initiating or promoting regimen alone did not produce tumors; nor were tumors seen when the sequence of initiation-promotion were reversed. A working hypothesis for two-stage carcinogenesis has been pr0posed by Boutwell (21): "the initiator results in the formation of permanent and heritable but unexpressed changes in the cell genome, and the promoter causes the phenotypic expression of these changes in geno- " The idea that neoplastic transformation was a multistage type C O 0 process was a departure from the concept dominant at the time, which held that cancer was a rapid, irreversible, single step change (49). The potential role of a somatic mutation in the initiation phase of carcinogenesis gained new life in the late 19503 mainly due to the 4 work of E.C. and J.A. Miller who demonstrated that many apparently nonreactive carcinogens could be metabolically activated to a reactive species capable of covalently binding to macromolecules (145). Alternatively, the possibility of a permanent and heritable change in the genome resulting from the interaction of a carcinogen with a repressor regulatory protein was pr0posed by Pitot and Heidelberger (166) based on the work of Monod and Jacob (151). The phenotypic changes which result in cancer can also be envisioned as occurring without any chemical modifications of DNA as the normal program of cellular differentiation has been accepted to be a process that alters the availability, but not the information content of the total DNA complement (169). The unifying concept of chemical carcinogenesis, therefore, is that the aberration that produces cancer can be traced to a crucial "malfunction" of the genetic material, a paradigm that encompasses all of the models described above. This malfunction thus can be the result of chemical, viral or epigenetic changes. A common property of virtually all ultimate chemical carcinogens is that they are strong electrophilic reactants (146). Thus, chemical carcinogens have the potential to react with nucleophilic sites on cellular macromolecules. These nucleophilic sites are abundant in RNA, DNA and protein and include nitrogen, oxygen and carbon atoms (146). Attempts have been made to correlate a selective binding to DNA, but not RNA or protein, and carcinogenicity; these studies have (24,87,132) or have not (111) been successful in demonstrating a relationship between the extent of binding to DNA and the degree of carcinogenicity of a particular chemical. Such correlations are 5 probably not useful for demonstrating that DNA is the critical cellu- lar target since the relative binding most likely reflects only the metabolic capacities of the cell; furthermore only some of the inter- actions of carcinogens with nucleic acids appear to be important in the initiation of carcinogenesis (185). Other studies have shown a good correlation between in_yiyg_ carcinogenicity and the mutagenicity of chemicals in bacterial systems (133). Similar but less extensive results have been reported for mammalian cell culture systems (86). Correlations between mutagenesis and carcinogenesis, however, do not clarify whether mutations are the underlying cause of cancer or merely reflect the nearly universal electrophilic nature of carcinogens and the possibility that critical targets for mutagenesis and carcinogenesis might be nucleOphilic centers, though not necessarily identical. More convincing evidence for the role of mutations in cancer come from studies of cells from patients having the rare disease xeroderma pigmentosa, which is characterized by an extremely high incidence of skin cancer as a consequence of exposure to sunlight. Cells from these patients have a greatly reduced capacity to repair UV or chemical- induced damage to DNA (32). In a similar manner, when cells from Poecilia formosa were exposed to UV light in_yi££2_and injected into suitable hosts, a high incidence of thyroid tumors resulted. The cells of this species contain a photoreactivating enzyme that cleaves the UV induced pyrimidine dimers. Exposure of the cells to visible light after UV irradiation substantially prevented the development of thyroid tumors upon subsequent innoculation of the cells (77), demon— strating that repair of DNA damage reverses malignant transformation. 6 Using a series of 5 criteria, it has been demonstrated that a somatic mutation, but not an epigenetic change, is the basis for malignant transformation of BHK cells by chemical carcinogens in yi££9_(20). Finally, neoplastic transformation in_yit£g_has also been obtained (8) after a direct perturbation of the DNA template by 5-bromodeoxyuridine and near ultraviolet light treatment, without any concomitant altera- tion of any other cellular macromolecules. The experiments cited above all provide evidence for the premise that alterations in DNA and subsequent somatic mutations are important in carcinogenesis. Interaction of carcinogens with DNA during the initiation stage is usually described as a multi-step process: 1) metabolic activation of the chemical to an electrophile; 2) covalent binding to DNA; 3) possibility of DNA repair (error free or error prone); and 4) heritably altered DNA sequence. In the skin tumor two stage system, the requirements for initiation can be satisfied with a single carcinogen exposure(154), and only a relatively short exposure is required for initiation in the liver (163). Based in part on an analogy to mutagenesis studies, it is often assumed that initiation can be described as a random single or multi "hit" phenomena. Such an analysis has been done for human retinoblastoma by Knudson §£_§1, (101), who suggest that the "mutation theory assumes that carcino- genesis is related to discrete changes occurring at random and at a constant average rate." The concept of initiation as a random mutational event has also been extended to suggest a random DNA interaction. In a hypothesis suggesting that oncogeny may be adaptive ontogeny, Nery (158) states 7 that "in general, chemical carcinogens and ionizing radiation randomly alter one or more bases in DNA." On the other hand, Burnet (25) reasons that there is a nonrandom component in the somatic-mutational process. Likewise, Huberman 25 a1. (85) reported that the frequency of chemically induced transformation in_yi££2_was 20-fold greater than that of a specific mutational event (development of ouabain resis- tance). As one possible explanation, they suggest that the genes for transformation may be located at hot spots within the genome, which have a higher frequency of mutation. Alternatively, the difference in frequency between mutagenesis and transformation ig_yi££g_may reflect the relative size or number of target genes for transformation. Recently, it has been reported (113) that the ratio of transformed/ ouabain resistant mutants was 21 for AcAAF and only 12 for benzo(a)- pyrene. The fact that these two carcinogens had differing transfor— mation/mutation ratios in the same cell line suggests that the rela- tive number of target genes cannot completely explain the frequency of transformation in 23532, In a similar view, in their analysis of the two stage model of carcinogenesis, Scribner and Suss (182) suggest that "poor initiators act either through an excess number of random hits or a limited number of selective hits, whereas good initiators can achieve a high number of selective hits." The question, then, is whether the carcinogenic process involves the expression of an aberrant phenotype superimposed on a background of random DNA damage, or whether the critical DNA lesion is more Specific to carcinogens than would be predicted due to random DNA damage alone. Proceeding from the Scribner and Suss (182) model, one can ask what the nature of the selectivity of good initiators is. A 8 conclusion of random DNA damage would predict that initiation and promotion are independently occurring events, while a conclusion of non-random damage would suggest some chemicals are highly carcinogenic because of the inherent nature or site of the particular DNA damage they induce. In the former case, the promotional aspects of carcino- genesis would be the major determinant of frequency of tumor forma- tion, while the latter case allows for initiation events to enter into determining potency of carcinogenesis. Presently, potential inter- actions between events in initiation and promotion are only poorly understood. It is clear that gaps still exist in our understanding of the initiation process. Much work has been done on the chemical nature of carcinogen DNA interactions as well as the mutational properties of chemical carcinogens. As pointed out by Weinstein §£_al, (203), however, there has been a tendency to think of initiation as a single random point mutation resulting from errors in replicating the damaged DNA, a hypothesis that may not be consistent with the apparently high efficiency of initiation of some chemicals. It is important, then, to investigate further the nature of initiation, the first step in carcinogenesis. Questions which need to be answered and which can only be partially addressed in this thesis relate to the nature, location and persistence of chemical modification of DNA, the funC* tional effects of these lesions, the specificity (if any) of these lesions in determining carcinogenic potency, and the relationship between the initiating and promoting phases of carcinogenesis. 9 2. Carcinogenicity of 2-Acetylaminofluorene The aromatic amine 2-acetylaminof1uorene (AAF) is a highly active experimental carcinogen that was first investigated by Wilson g£_al, in 1941 (201). As a class, the aromatic amines have been demonstrated to be carcinogenic in a variety of species, including rodents, rabbits, dogs and humans (107). Indications of human cancer risk was first reported in 1895 by Rehn (176), who observed a high incidence of bladder tumors in German aniline dye workers who were occupationally exposed to a variety of aromatic amines. Until World War I, more than 80% of the worldwide production of aniline and other aromatic amines was manufactured by the German chemical industry, and other investi— gators soon confirmed Rehn's observation. Commercial production of aniline dyes in the U.S. began during World War I; this was followed some years later by reports of increased bladder cancer among these American workers and also those in the rubber industry, where various aromatic amines were used as antioxidants (107). It has been sug- gested that these human bladder tumors were due to contamination of aniline with other aromatic amines, such as 2-aminonaptha1ene, 4- aminobiphenyl, and benzidine (201). Besides causing bladder cancer in man, dog, rabbit and rats, AAF induces liver, mammary gland and ear duct gland carcinomas in rats (107,206). Typically, hepatocellular carcinomas are induced in rats by feeding 0.03% (w/w) AAF in the diet for approximately 6-9 months, at which time 100% of the animals will have tumors. During the initial weeks of carcinogen exposure, animals gain weight at a slower rate than controls, however, no hepatic cell death occurs for at least 10 7 weeks of carcinogen ingestion (measured by loss of prelabelled DNA) (3,207). Within 2 to 3 weeks of AAF ingestion, a proliferation of oval (bile duct?) cells in the liver begins (48); this is followed after 7 weeks by hepatic parenchymal cell proliferation (3). The change of the parenchymal cells from hyperplastic to ne0plastic takes place slowly, with distinct stages of hyperplastic nodules, nodules with atypical cells, and small hepatocellular carcinomas observed (177). The hyperplastic nodules are thought to be the precursors of the neoplastic cells, however, most of the nodules can revert into normal hepatic structures if carcinogen exposure is stopped before a critical point is reached (193). AAF has also been reported to induce hepatic tumors after a single injection when administered to newborn (1-2 day-old) mice (60). The metabolic activation of AAF in the liver has been worked out in some detail. Rats fed AAF convert it to a metabolite, N-hydroxy-Z- acetylaminofluorene (N-OH-AAF), which is then excreted in the urine as a glucuronide conjugate (36). Unlike the ring hydroxy metabolites of AAF, the N-hydroxy metabolite is a more active and versatile carcino- gen than the parent compound (144,147). The N—hydroxylation reaction is catalyzed by the endoplasmic reticulum cytochrome P-450 mixed function oxidase enzyme system (194). Although administration of N- OH—AAF to rats resulted in DNA and protein bound derivatives, the compound had little or no reactivity in 31532, implicating a second metabolic activation step in ziyg (146). It has been proposed (see 107) that this second step is an esterification of N-OH-AAF to a highly reactive sulfate ester. A 11 cytosolic enzyme which catalyzes this reaction utilizing 3'-phos- phoadenosine 5'—phosphosu1fate as a sulfate donor has been studied in_ yi££2_(38). A good correlation has been demonstrated between the activity of the hepatic sulfotransferase and the carcinogenicity of N— OH-AAF in different rodent species and sexes. Depletion of sulfate pools by p—hydroxyacetanilide reduced the protein binding of N-OH-AAF; this effect was reversed by administration of sulfate anions. Because of its high reactivity, N-acetoxy-AAF (AcAAF) is used in zi££g_in place of the sulfate ester; it yields carcinogen residues identical to those obtained after in_yigg administration of AAF (107). Other enzymatic pathways for the formation of electrophiles from AAF have been described. One of these is a peroxidase-catalyzed one electron oxidation of N-OH-AAF to yield a free nitroxide radical, which can then dismutate to form AcAAF (10). Alternatively, a cyto- solic acetyltransferase has been reported to form the strong electro— phile N-acetoxy-Z—aminofluorene in the rat liver (91). N-OH-AAF can also be converted in the liver to the weakly electrOphilic o—glucuro- nide (142), a reaction which may be important in extrahepatic car- cinogenesis (89). Finally, it has been proposed that AAF-sulfate esters formed in the cytoplasm would be too unstable to result in the binding of AAF residues to nuclear DNA (107). However, recently it has been reported that activation of AAF and DNA binding can occur in isolated nuclei apparently due to nuclear envelope monoxoygenase (98) and sulfotransferase (190) enzymes. 12 3. Effects of 2-Acetylaminofluorene on DNA Structure and Tran- scription Progress has been made in the identification of covalent adducts of carcinogen residues with DNA after AAF exposure. The identifica— tion of a reaction product of guanosine and AcAAF in_yi££9_(108) was soon followed by several studies dealing with adduct formation in_yiyg_ (46,91,104,106,192,205). Three carcinogen-DNA adducts have been identified, all representing binding to the nucleotide guanosine. The major adduct involves binding of the amine group of the carcinogen to the C8 position of guanine. This reaction occurs either with the loss of the acetyl group from AAF (65% of total DNA binding) or retention of the acetyl group (25% of total DNA binding). The remaining 5-15% of carcinogen bound to DNA involves the binding of the ring 3 position of AAF to the nitrogen on the 2 position of guanine, with retention of the acetyl group of AAF (90,91,106,205). Binding to the C8 and N2 positions of guanine is illustrated in Figure 1; no studies have analyzed the potential functional differences acetylated vs. deacety- lated carcinogen CS—guanine modification. Differences in carcinogen binding between RNA and DNA have been reported. Thus, while 65% of carcinogen-DNA residues have lost the acetyl group of AAF, approximately 75% of the fluorene residues bound to cytoplasmic RNA retained the acetyl group (91,92,104,105). This finding might reflect relative differences in reaction pathways generating electrophilic species capable of binding to macromolecules in the cytoplasm or in the nucleus (91). Differences in the rate of loss of carcinogen from RNA and DNA after a single injection of AAF or N-OH-AAF have also been reported. Loss of carcinogen from RNA occurs 13 Figure l. Carcinogen-guanine adducts after AAF administration. Structures of DNA-bound fluorene residues after in_vivo admini- stration of AAF (205). Upper adduct is N-(guanin—8—yl)-2- aminofluorene, lower adduct is 3-(guanin-Nz—yl)—2-acetylamino- fluorene. 15 with a half—life of approximately 3 days, with complete loss by 4 weeks (91). Loss of carcinogen from DNA occurs with a half-life of l- 7 days, however, 10-15% of the maximal DNA-bound carcinogen remains stably bound up to 10 weeks after a single injection (90,91,106,192). The persistently bound carcinogen-DNA residue has been identified as the NZ-guanine adduct (205). Binding of AAF to the N2 position of guanine does not occur with RNA (105), however, possibly reflecting structural differences between the 2 molecules. The potential importance of the persistent carcinogen—DNA residue was suggested by studies of Epstein_g£.al. (46). Rats were fed AAF for 10 weeks to induce hyperplastic nodules and were then returned to the control diet for 4 weeks. DNA isolated from cells in the hyper— plastic nodules, but not from the surrounding normal hepatocytes, was found to contain distortions possibly due to persistent carcinogen-DNA moieties. Extensive studies on the structural alterations due to modifica— tion of DNA in 21339 with AcAAF have been carried out by Weinstein and associates (116, 208,209), Fuchs and Daune (56, 58,59) and Kapuler and Michelson (97). Circular dichroism (56,59) and melting curve analysis (56,59,116) suggested that carcinogen modified bases were shifted outside the double helix; formaldehyde unwinding studies (58) indi- cated that each bound fluorene residue gave rise to weak points in the DNA; and viscosity and light scattering findings (56) indicate that AAF introduces hinge points in the DNA molecule. Based in part on these results, a "base displacement model" has been proposed (59,116) to describe the structural effect of AAF interaction with double stranded nucleic acids: 1) AcAAF attacks the 08 position of guanine 16 when it is made accessible by normal DNA "breathing", 2) a disruption of the normal G-C hydrogen bond base pairing occurs, 3) there is a rotation of the guanine base around the glycosidic bond (between guanine N9 and deoxyribose Cl) from the normal anti to a syn confor- mation, 4) the modified guanine is displaced from its normal coplanar position relative to the adjacent bases, 5) a stacking of the fluorene residue parallel to adjacent bases then occurs, 6) the large size of the fluorene molecule gives rise to distortion of the DNA helix and localized regions of denaturation. The conformational changes resulting from either 08 or N2 guanine modification was further examined by digestion of the DNA with a single strand-specific Sl endonuclease (55,208). Digestion of only the C8 guanine adduct, but not the N2 adduct, occurred. This suggests that the former causes major conformational changes, while the latter does not cause major distortions of native DNA structure. Model building studies indicate that the fluorene residue of the N2 guanine adduct could occupy the minor groove of the DNA helix with less disturbance of helical structure (209). Other studies (208) demon- strated that 5-35 base pairs of DNA were released for every gua-AAF residue released with $1 nuclease digestion, giving an approximation of the extent of denaturation due to each bound AAF residue. Using a different technique, Fuchs and Daune (58) estimated a denaturation of 7 to 23 base pairs, while exposure of human cells in culture to AcAAF induces an excision repair of approximately 140 nucleotides per AAF modification (175). The finding that the N2 guanine adduct is not susceptible to digestion by $1 nuclease suggests that an important l7 factor governing DNA repair in_yizg_is the ability of repair enzymes to recognize a structurally altered template. The base displacement model predicts that binding of AAF to C8- guanine would be sterically inhibited when the nucleotide is strictly held in a static double helical structure (55). When DNA was reacted with AcAAF in a high ionic environment, which is known to stabilize nucleic acid structure, the rate and extent of DNA modification was reduced (116). Similarly, the rate and extent of the reaction of AcAAF with denatured DNA was 2-3 times greater than that of native DNA (116). Finally reaction of AcAAF with E, ggli_tRNA, which contains both double stranded and "looped" single stranded regions, occurred primarily in the guanosine residues in the 100p positions (61). 4. Biochemical Effects of 2-Acety1aminofluorene Binding to Macro- molecules Several biochemical effects of covalent binding of AAF to cellu- lar macromolecules have been described, with the hope of correlating a biochemical action with the biologic effects of the carcinogen. Covalent binding of relatively low levels of AcAAF to E, 2911 transfer RNA's produces specific modifications in their amino acid acceptance capacities, codon recognition and ribosomal binding (51). Binding of AcAAF to ribosomal RNA impaired its ability to hybridize to homologous DNA (116). Extensive carcinogen binding also occurred on cytoplasmic and nuclear proteins both in_yiyg_and in_yitrg, and there is evidence that specific proteins are particularly highly modified (107). Some of these modifications can presumably interfere with normal protein and enzyme functions, though the role, if any, of protein binding in chemical carcinogenesis is not known at present (107). 18 Most studies have concentrated on the effects of carcinogen modification of DNA. EXposure both in cell culture and in_yi££g_to AcAAF results in macromolecular complexes between DNA and protein (52,137,152). These complexes apparently are covalent DNA-protein crosslinks, and they have been postulated to be related to DNA single strand breaks (52) and chromosomal aberrations (152). DNA reacted with AcAAF was a less effective primer-template for DNA polymerase- catalyzed DNA synthesis (127). A single injection of AAF into male rats inhibits liver regeneration following partial hepatectomy, an effect which may be due, in part, to inhibition of DNA synthesis (74). Evidence from thermal elution chromatography and pyrimidine isotich analysis suggests that damaged DNA in human cells exposed to AcAAF is faithfully restored by DNA repair synthesis to the original nucleotide sequence (117). The methods used in these studies, however, could not detect shuffled sequences or occasional DNA base substitutions (117). One of the most striking biochemical effects of AAF is its ability to block RNA synthesis. This action may be due, in part, to alterations in RNA polymerase activity (64,66,67,82,210,211), as well as a result of direct modifications of the DNA template (2,73,96,150, 195,210,212). The inactivation of the DNA template for RNA synthesis is noteworthy in that it: 1) occurs rapidly upon carcinogen reaction with DNA (202); 2) can result in complete blockade of RNA synthesis when extensive carcinogen modification of DNA occurs (212); 3) can substantially reduce the priming activity of DNA for RNA polymerase at levels of carcinogen modification that do not affect the activity of DNA for DNA synthesis (195). The difference in sensitivity for 19 inhibition of DNA and RNA synthesis is approximately 20-fold, with RNA synthesis more sensitive to inhibition (195). DNA template inactiva- tion occurs after carcinogen exposure both in_gigg_(73,l95,210) and in 3i££g_(64,150,208,212) and also when assayed with bacterial (150,195, 208,212) or mammalian hepatic RNA polymerases (64). In_yiyg, the greatest inhibition appears to occur in the synthesis of 453 ribosomal RNA precursor (65,96), apparently due to an effect on nucleolar DNA template function (73,210). Direct inhibition of RNA polymerase activity by AAF is dose-dependent, however, in_yiyg_RNA polymerase II (nucleoplasmic) is more sensitive to inhibition than is RNA polymerase I (nucleolar) (64). A single injection of N-OH—AAF also results in an inhibition of poly-A RNA synthesis, although the inhibition of poly-A RNA was somewhat less than that of ribosomal RNA (65). Poly-A RNA is a measure of the processing of nuclear RNA to its cytoplasmic forms, a process which involves post-transcriptional events such as methylation and polyadenylation. Using a variety of approaches, it has been demonstrated that the inactivation of the DNA template for RNA synthesis is due to a reduc- tion of RNA elongation, but not RNA synthesis initiation (64,150,195, 212). It has been postulated that most of the AAF residues bound to the coding strand of the DNA cause premature termination of transcrip- tion, at or near the site of modification. This results in the production of shorter RNA chains; the synthesis of RNA presumably cannot continue distally to the site of the AAF-modification (150). This inhibition of DNA template capacity persists when the DNA is packaged in its native chromatin structure (208). DNA modified in vitro was reconstituted with unmodified chromatin proteins (208). 20 Carcinogen modification did not grossly affect the ability of the DNA to interact with chromosomal proteins to form normal nucleosome structures. The reconstituted chromatin, however, showed approxi- mately a 70% reduction in template capacity for RNA synthesis. As might be predicted for a chemical that interacts extensively with DNA, AcAAF has been shown to be highly mutagenic in a variety of systems. Maher g£_al. (128) reported an inactivation and an increased mutation frequency in transforming DNA of Bacillus subtilis that had been reacted with AAF esters. Similarly, Corbett 35 a1. (35) examined the mutagenicity of AcAAF in the T4 bacteriophage-E, ggli_system. The types of mutants observed were large deletions, A~T to G-C transi— tions, and frameshift mutations, occurring with a frequency ratio of approximately 1:2:3, respectively. In the presence of a mammalian metabolic activation system, AAF induces frameshift mutations in the Salmonella typhimurium histidine revertant system (5). Mutations were also observed upon the injection of AAF, N-OH-AAF and AcAAF into Drosophila melanogaster (47). The first two compounds were only weakly mutagnic, and were selective for rRNA genes. AcAAF was highly mutagenic and in addition to the rRNA genes induced point mutations and gene eliminations in heterochromatic regions. AcAAF also produces mutations in mammalian cells in_vitro (113). 5. Chromatin Structure and Fractionation There has been an enormous amount of research in recent years into the molecular biology of chromatin, and mechanisms of gene regulation (for example, see the 1978 Cold Spring Harbor Symposium, 21 ref. 33). Therefore, only a selective and limited review will be attempted here. A cornerstone in molecular biology is the concept that a multicellular organism develops with every cell having the same complement of DNA. Thus, the determining factor in the specialized structure and function of each cell is the portion of DNA in the genome which is expressed at any given time. Control of gene ex— pression is believed to occur, to a large extent, at the level of RNA synthesis such that transcription of specific genes occurs in one tissue, while being inhibited in other tissues (7). It has been estimated that 10-30% of the DNA in differentiated tissues is re- presented in total cellular RNA, and about one-tenth of this infor- mation is present in mRNA. This yields a value of approximately 12,000-15,000 genes actively being transcribed in the liver (7). Many investigations have focused on the structure of chromatin as a means of elucidating its function in gene transcription. Nuclear DNA exists as a complex with RNA and proteins. The proteins include five chemically-defined histones, and large, heterogeneous group of non-histone proteins. The basic structural unit of chromatin is the nucleosome, which consists of a well defined length of DNA (about 200 base pairs) complexed with eight histone molecules, 2 each of histones H2A, HZB, H3 and H4. The DNA is wrapped around the outside of the histone octamer, forming a repeating, roughly spherical particle and giving rise to a "beads on a string" appearance (102). The non- histone proteins are then associated in a as yet undetermined manner with the DNA-histone complex. 22 For quite some time it has been suggested that template active and template inactive chromosomal regions can be differentiated based on the density of their staining in gigu. Frenster §E_al, (54,122) have demonstrated that RNA transcription occurs on the more extended, template active, euchromatin, but not on the more condensed, template repressed heterochromatin. This apparent physical difference in chromatin structure has been widely used as a means of fractionating chromatin into putative eu- and heterochromatic sequences for further analysis in_vi££2_(178). Chromatin fractionation is basically a two-step process. The interspersed eu- and heterochromatin segments must first be separated, either by physical shearing, sonication, or enzymatic cleavage. The chromatin fragments are then separated based on physico-chemical differences by a variety of methods including differential centrifu- gation, rate-zonal sedimentation, gel-exclusion chromatography, and salt precipitation (178). The separated chromatin fractions are then characterized as either eu- or heterochromatin based on in 31553 template capacity, association with nascent RNA, satellite DNA se— quences, or endogenous RNA polymerase (178). Although these parameters are logically based on current know- ledge of chromatin function, criticisms of some of the techniques used have suggested that the separation of different fractions could have been due to the formation of artifacts (178). These studies indicate that physical shearing of chromatin may cause important alterations due to effects on chromatin proteins (43) or RNA (183). Similarly, other studies have concluded that although physically sheared and fractionated chromatin retains many of the structural distinctions 23 between eu— and heterochromatin known to exist in_yigg, these chroma— tins fail to demonstrate any selective fractionation of globin or keratin genes (103), actively transcribed viral nucleic acid sequences (84), or actively transcribed mRNA sequences (93). On the other hand, one study has successfully demonstrated a fractionation of ribosomal genes in sonicated chromatin fractionated on a sucrose gradient (83). In an attempt to avoid potential artifacts induced by random physical shearing, DNA endonucleases have been used to fragment chromatin into template active and inactive fractions. It was hOped that these enzymatic probes could at least partially recognize salient features of chromatin structure as does RNA polymerase in_yiyg_(162). Besides demonstrating a selective distribution of nascent RNA with the euchromatin (15), studies with DNAse I (63,162, 204) and DNAse II more importantly reveal a fractionation of a variety of gene sequences in various differentiated tissues. Several of the studies with DNAse II are listed in Table 1. Thus, the nuclease digestion procedures demonstrate chromatin fractionation intoeu- and heterochromatin as judged by a more stringent criteria, a localization of actively transcribed or repressed segments of the genome in chromatin derived from differentiated tissues. 6. Intragenomal Distribution of DNA Damage and Repair Interest has been increasingly focused in the past few years on the influence of chromatin structure on the interactions of carcino— gens and DNA. Several approaches have been used to analyze the distribution of carcinogen damage and repair, including: 1) Staphylo— coccal nuclease digestion of chromatin. Nuclease-sensitive regions 24 TABLE 1 Evidence for Gene Localization After DNAse II - MgCl2 Chromatin Fractionation "Pro" "Con" 1) 2) 3) 4) 5) 6) 7) 82 contains specific subset of 1) of whole genomal DNA. 82 DNA codes for 60% of total cellu— lar RNA (72). 32 DNA of liver differs from 2) 82 DNA of brain (70). 82 enhanced for globin gene DNA in friend erythroleukemia cells (131). $2 enhanced for globin gene in reticulocytes, but not in liver (81). Nucleolar 82 enriched in rDNA (131). Thyroid hormone nuclear receptor enriched in $2 of rat pituitary cells (181). Estrogen nuclear receptor en- riched in $2 of hen oviduct chro- matin (80). No selective globin gene localization in Friend erythroleukemia cell S2 DNA (114). No selective globin gene localization in chick reticulocyte 82 DNA (75). $2 = MgCl -soluble euchromatin fraction, as described in Methods. rDNA = DNA sequence which codes for ribosomal RNA. 25 are primarily the linker DNA, the 60 DNA base pairs found between the 140 base pairs that are intimately associated with spherical histone particle (”core DNA") (123). 2) DNAse I digestion. Nuclease-sensi- tive regions after a short digestion are primarily the template active euchromatin DNA (63,204). 3) Sheared chromatin fractionated by gradient sedimentation. Chromatin is fractionated based on apparent differences in physical structure. 4) Separation of repetitive vs. unique DNA sequences by DNA-DNA hybridization. Approximately 15-40% of eukaryotic DNA is composed of reiterated sequences which may play a role in the regulation of the structural genes, which consist of unique sequences (37). 5) Separation of satellite DNA from mainband DNA by Ag+-CsSO4 centrifugation. Satellite DNA comprises 8-9% of the total DNA and is comprised of a highly specialized set of hetero- chromatic, highly reiterated sequences (119). 6) In_§i£u_evaluation of DNA repair after carcinogen exposure by electron microscopy and autoradiography. These studies indicate that carcinogen binding is initially greatest in euchromatin regions determined by sucrose gradient sedimen- tation (153, 155) or by DNAse I digestion (40,94,139,17l,172) after exposure to BP (94), AAF (17,139,172), DMN (140,171) or UV light (187). This region of the genome also demonstrates a greater extent of DNA repair as revealed by loss of bound carcinogen from chromatin fractionated by DNAse I digestion (171,172) or sucrose gradient frac- tionation (155). Similarly, the extent of incorporation of 3H—thy- midine during DNA repair synthesis is also greater in the DNAse I 26 susceptible chromatin (18) and in electron microscopically-identified euchromatin (75). In contrast, no selectivity of binding of MNU (l7), MMS (17), AcAAF (119), or UV-induced damage (119) between mainband and satellite DNA sequences could be detected. Similarly, there was no difference in the amount of 7-methy1-guanine damage or repair between repetitive and unique DNA sequences after in_!iyg_exposure to DMN (62). An equal distribution of bound adducts between repetitive and unique DNA sequences was also reported to result from exposure of human fibro- blasts to AcAAF (118) or murine skin cells to a high concentration of DMBA (184). At low carcinogen doses, however, a concentration-depen— dent preferential binding to reiterated DNA sequences occurred (184). These investigators also demonstrated an inverse linear relationship between the enrichment of hydrocarbon adducts in reiterated DNA sequences and the logarithm of the amount of total hydrocarbon bound to DNA. Intriguing results have also been reported in the distribution of DNA damage and repair within each nucleosome, as revealed by Staphy- lococcal nuclease digestion. Early studies demonstrated an enhanced binding of BP (94), AAF (139,172) and alkylating agents (140,171) to the "linker" regions of the nucleosome. However, conflicting results were obtained when chromatin was isolated at various times after carcinogen exposure. More recent studies have carefully examined this phenomenon using pulse-chase incorporation of 3H-thymidine during DNA repair synthesis (l8,186,l87,195). It appears that from 70-100% of the DNA repair initially occurs in the linker DNA region; with time 27 (4-12 hr), however, the chromatin within the nucleosome rearranges such that the initially repaired lesions are now located in the core DNA, while previously unrepaired damage moves into the linker DNA region. This nucleosome rearrangement repair process is nonspecific in that it occurs in either mouse mammary cells (18) or human cells (31,186,187,l95) after exposure to either AcAAF (195), UV light (31,186, 187) or alkylating agents (18). A general conclusion from the studies cited above is that chroma- tin structure influences carcinogen damage and DNA repair synthesis both at the nucleosome level, and also at higher levels of chromatin organization. Furthermore, a rearrangement of the DNA within the nucleosome can occur, resulting in an enhanced exposure of damaged DNA to repair enzymes with time. At the present time it is not known whether this rearrangement process is constitutive (i.e., occurring in the absence of DNA damage) or is induced by the repair process itself (195). 7. Experimental Objectives The objectives of this investigation were to examine the nature of the interaction of the carcinogen AAF and its N-hydroxy metabolite with target organ DNA ig_giyg, These included studies of quantitative and qualitative aspects of carcinogen-DNA binding as well as an assessment of potential functional DNA damage as measured by effects on RNA synthesis. Carcinogen binding or damage was assessed shortly after a single exposure to N-OH—AAF (2 hr), and after carcinogen ingestion for 4, 7 and 14 days. Repair of carcinogen damage was 28 measured 4, 7 and 10 days after a single injection of N-OH-AAF, and 7 days after a 4 day exposure to AAF in the diet. The studies were initiated to test the premise that carcinogen- DNA interactions do not occur randomly in_giyg, and have specifically focused on potential differences occurring between DNA in template active and DNA in template repressed portions of the genome. Chroma— tin was fractionated by either sonication-glycerol gradient, soni- cation-MgCl2 precipitation, or DNAse II digestion-MgCl precipita- 2 tion. Carcinogen DNA adducts were examined either as radioactivity in a 100°C-acid hydrosylate or by determining chemically identified adducts with silica gel thin layer chromatography. These studies might be useful in understanding the relationship between carcinogen- induced DNA damage and later stages (i.e., promotion) of carcino- genesis, as well as hOpefully helping to pinpoint the location of critical DNA lesions responsible for initiating the progression of a normal cell to a neoplastic cell. MATERIALS AND METHODS 1. Animals; Carcinogen Administration Male, Sprague-Dawley rats (150-200 g), purchased from Spartan Farms (Haslett, Michigan) were used in all studies. Animals were housed 2 per cage in a room with a controlled 12 hour light cycle beginning at 7 p.m. The rats were fed either a basal diet (Carcino— genic Basal Diet; Teklad Mills, Madison, Wisconsin) or one containing 2-acetylaminofluorene (Aldrich Chemical Milwaukee, Wisconsin). When indicated rats were injected i.p. with [ring-3H]-N-hydroxy-N-acetyl-2- aminofluorene (50.2 mCi/mmole, purchased from New England Nuclear), or 1 [9- 4C]-2-acetylaminofluorene (10 mCi/mmole, New England Nuclear) which was dissolved in 0.9% NaCl containing 12% ethanol. 2. Isolation of Chromatin The method of Rodriguez and Becker (179) for the isolation of chromatin was modified as follows. Livers were homogenized in 250 mM sucrose, 50 mM Tris (pH 7.9), 25 mM KCl, 5 mM MgCl2 (STKM) and fil- tered through cheese-cloth. Nuclei were isolated by washing twice in 2% Triton X-100 STKM followed by centrifugation at 750 x g for 10 min. The nuclear pellet was washed once in STKM. The washed nuclei were homogenized in 10 mM Tris (pH 7.9). Chromatin was purified by layering 10 m1 on a sucrose-Tris (10 mM, pH 29 30 7.9) step gradient consisting of 14 ml 1.3 M sucrose and 14 ml 1.6 M sucrose, and sedimented by centrifugation (112,000 x g for 2 hours) in a Beckman SW27 rotor. The chromatin was resuspended in 8% glycerol, 10 mM Tris (pH 7.9). Chromatin prepared in this manner had a pro- tein:DNA ratio of 3.9, and a RNA:DNA ratio of 0.27. 3. Chromatin Fractionation Isolated chromatin was fractionated by three different methods. In the first method, chromatin was sonicated for two 10 second bursts at 70% maximal power using an Ultrasonics sonicator equipped with a microtip. Approximately 2.5 mg of purified, sonicated, chromatin DNA was layered on a 12-90% (v/v) linear glycerol gradient in 10 mM Tris (pH 7.9), and centrifuged at 72,000 x g for 15 hours in a Beckman SW27 rotor. Gradients were monitored at 260 nm and 12 fractions, 3 ml each, were collected. Fractions of fast and slow sedimenting chroma- tin were pooled, as indicated in Figure 2, for further analysis. The portion of the chromatin which sedimented to the bottom of the gra— dient (designated pelleted heterochromatin) was also analyzed. This material represents approximately 15% of the total chromatin DNA. Alternatively, chromatin was fractionated by selective aggrega- tion with MgCl2 (6). Sonicated chromatin, prepared as described above, was centrifuged (2,000 x g for 15 min) in the presence and absence of 1.75 mM MgC12. The supernatant and pelleted portions were analyzed. Chromatin fractions were classsified as follows: "Euchro- matin", chromatin in the supernatant after centrifugation in the presence of 1.75 mM MgC12; "pelleted heterochromatin", chromatin in the pellet after centrifugation without MgClZ. The portion of chromatin 31 .afiumsouzoououoz wouoaaom mm on wmuuomou on Hamcm mfinu mam undo onu mo aouuon ecu um voucoefivom omam cfiumEoupo onu mo cowuuoa < .afiumaopnooumuos voumcwfimow ouo3 42m usoomm: a“ ucofiofimov mao>fiumaou mdoauomum omocu mafiumEousoso woumswfimow whoa >\rdn:o: AXE. Control C 15 min. 70° is O V .E: CON .0 uncontamnix 0.6 r- Froction No. Figure 3 44 TABLE 3 Chromatin Fractionation by Selective MgCl2 Precipitation a , Nascent RNA Chromatin Fraction DNA (A) (DPM/ug DNA) I. MgCl2 solubleb 9.4:2.5 851i206 II. M3012 insolublec 76.5:3.6 126: 45 III. Insolubled 13.5i0.3 40i 1.3 aSonicated chromatin was prepared as described in Methods. Supernatant after centrifugation at 2000 x g for 15 min in the presence of 1.75 mM MgClZ. cPrecipitate after centrifugation at 2000 x g for 15 min in the presence of 1.75 mM MgCl2 and after subtraction of values from III. dPrecipitate after centrifugation at 2000 x g for 15 min in the presence of MgClz. 45 apparent improved separation over that obtained with the glycerol gradient fractionation procedures. The relative purity of the MgCl2 soluble chromatin, as measured by the amount of nascent RNA, is 7-20— fold when compared to the MgCl2 insoluble or soluble chromatin. Because the MgCl2 and the glycerol gradient fractionation proce- dures take advantage of related, but not necessarily identical physico- chemical properties of chromatin as the basis for fractionation, it was of interest to directly compare these two methods. Sonicated chromatin was prepared and fractionated on a glycerol gradient. Gradient fractions were pooled as indicated in Figure 2. These were then fractionated by selective precipitation with MgCl2 (1.75 mM, final concentration), and the amount of chromatin DNA that was MgCl2 soluble, and MgCl2 insoluble was determined (Table 4). To examine whether the pelleted chromatin represents only in- sufficiently sonicated chromatin, samples were resuspended, divided into five aliquots of approximately 20 ug chromatin-DNA, and resoni— cated for various times (Figure 4). The amount of pelletable material was then determined. Following a very brief sonication (15 sec), a large portion of the chromatin DNA is no longer pelletable. However, even with extensive sonication, 40% of the initial pelleted chromatin cannot be transformed into MgCl soluble or MgCl2 insoluble chromatin. 2 This suggests that a portion of the pelleted chromatin exists in a different conformation than the remainder of the chromatin, not related to differences in the size of the chromatin fragments. An attempt was made to identify particular chromatin fractions based on the presence of specific endogenous RNA polymerases (RNA 46 .mumn m mo .z.m.m H some mum mooam> Ham mm oupmfim ou mammoa CH ponfiuomom mm woumouumua mummo .m oaane oome .N oupwfim aw woumawamov mm woaoom mdoauomnmu «Hams «zm usoommz saunaounoououom A.Ham.mo m.oao.o~ m.oaa.ma ANV <2o ooooaaom mama ~2u>3 I no. 2.930.333.53330 a Ausuegm § amnion 52 .voumuumpaaw omam own meow Hmpwfl>fiwcfi o>Humuoomouaou ”swam aofi Hosea onu ma House Hoaoa 05H .muwmcouafi o>fiumaou ma oumofiwno onu .Aoooom ImNV oupumuooaou onouo mafimmouoafi um voofimuno mcmom Hmpwa>flwaw mammouomu mmmfiomnm ecu so muonapz .odouopamoafiamamuoomu~n%%ouv%5n2:”mmH mo coausnwuumflv so“ unapooaoz .o oupwwm 53 o madman 54 300 with a cyclohexane:t—butanol:acetic acidzHZO (l60:5:15:20) sol- vent, and Silica Gel G with a isoprOpranol:ammonium hydroxyde (2:1) solvent. In both cases, all the radioactivity co-chromatographed with the major UV absorbing spot (Rf = 0.5), demonstrating that the tritium label was indeed associated with N—OH-AAF. 3. Analysis of Carcinogen Binding and Loss Carcinogen binding to cellular components was determined 2 hours after a single injection of 3H-N-OH-AAF, the time of maximal binding. Approximately 2% of injected carcinogen is covalently bound to ma- terial in the liver (Table 5). Of this, less than 5% is located in the nucleus, and less than 10% of the carcinogen in the nucleus is associated with the DNA. The time course of carcinogen binding to chromatin protein, RNA and DNA was examined after a single injection of tracer amounts (0.4 umoles/lOO g body weight) of [3H]-N-hydroxy-2-acetylaminofluorene (3H- N-OH—AAF). Two hours after injection, most of the carcinogen on a weight basis (i.e., per mg RNA, protein or DNA) was associated with the chromatin RNA, with smaller amounts associated with the chromatin proteins and DNA (Figure 7). As illustrated in Figure 7, loss of the carcinogen from chromatin proteins appears to be first order over the time period examined. In contrast, no carcinogen could be detected on the chromatin RNA after 7 and 10 days. A portion of the carcinogen appears to be persistently bound to the DNA based on the 10-15% of the amounts observed at 2 hours remaining bound for 7-10 days. There were no significant 55 TABLE 5 Analysis of Initial Carcinogen Binding to Cellular Componentsa (Male Sprague-Dawley Rats) Carcinogen injected i.p. 3.9x108 dpm (100 uCi/0.48 mg/100 g BW) Hepatic Acid-Insoluble 7.8x106 dpm 100% Hepatic Nuclear Acid-Insoluble 3.7x10 dpm 4.7% Nuclear Protein 2.5x102 dpm 3.2% Nuclear RNA 0.75x105 dpm 0.92% Nuclear DNA 0.30x10 dpm 0.38% aAnimals were injected with [ring 3HJ-N-hydroxy-Z-acetyl— aminofluorene (50.2 mCi/mmole) and were sacrificed after 2 hours. Cellular components were separated as described in Methods. 56 Figure 7. Binding of carcinogen to chromatin components. Male rats were injected with 3H—OH-AAF (20 uCi/lOO g) and sacrificed after 2 hours and 4, 7 and 10 days. Chromatin was isolated from hepatic nuclei and separated into component protein, RNA, and DNA as de— scribed in Methods, and the amount of radioactivity in each was determined. The rats were maintained either on a control diet (0), or one containing 0.03% (w/w) AAF (0). Each point represents the mean : SEM of 3-7 rats. 57 1000f 100' 10" HCONTROI. DIET 0—0 AAF DIET 10.000 " 1000 .100- DPM/mg 2000- DNA 1000' 100- L. 1 I 1 1 2 4 6 DAYS AFTER INJECTION Figure 7 i 58 differences in the loss of carcinogen from chromatin protein, RNA or DNA between rats maintained on a control diet or one containing AAF. To analyze the location of carcinogen binding within the genome, sonicated chromatin isolated 2 hrs after an injection of 3H-N-OH-AAF was centrifuged on a glycerol gradient. The DNA-bound carcinogen in each fraction was then determined. As seen in Figure 8, two hours after a single injection of 3H-N-OH-AAF the largest amount of car- cinogen is bound to DNA in the less condensed chromatin. There is progressively less DNA-bound carcinogen in the more condensed chroma- tin on the glycerol gradient, while the level of binding of the carcinogen to DNA in the pelleted heterochromatin is intermediate. Although the extent of carcinogen binding appears to be inversely correlated with the density of the chromatin, several fractions within the eu- and heterochromatin (e.g., 3-6 and 9-12) appear to be uni- formly labelled. Individual gradient fractions were pooled as indi- cated at the top of Figure 8, and the results from several experiments were compared (Table 6). When the genome was fractionated on a gly— cerol gradient, there was significantly more carcinogen bound to the DNA in the euchromatin, when compared to the DNA in the heterochro- matin. Chromatin can also be fractionated based on the enhanced suscep— tibility of the more condensed portions of the genome to precipitation by divalent cations. The euchromatin prepared by this method remains soluble in the presence of 1.75 mM MgClz, and contained significantly more DNA-bound carcinogen when compared to the heterochromatin or the pelleted heterochromatin DNA. 59 .mo.v9 .umou meDMIGma3o2Iuaowpum .<29 nauseouno Iououos wouoaaom pom :Humaounoououon noon Scum usouowwfiv haucmowmaswwm moumowpsfi% oncovg nummu meoMIGMEBDZIuowwnum .<29 :Humaousoououo: Eoum uaoHCMMAw adudmofimwawfim moumofivaHm .mcosuoz ca wonfluomov How: 25 mm.a sufis doaumufl9aoo99 o>HuooHom so momma woumCOHuomum mp3 nauseounow oaxm.H wdfiapmwm woumapoamou mm .N .<29 we\movfiuooaopa ma .9 ouowfim ca woumoflmsfi mm meofiuomum unofiumuw voaoo99 .mumu “no mo 29m H memos ucomoumou mopam> .mafiuosoo coaumaafiuawom prvHH pom pom: mm3 uouoflam on men .vaSuoz cw wonfiuomop mm mommaouvmn mp3 cofiuomum nauseouno some as <29 was .Aw ooaxfio: 99v 9<<|90I2|2 wo coauoonaa oawafim m noumm mason N mums aoum woumCOHuomuw mp3 caumaouzoo m mm.o 9mm Homqm <29 sfiumaousooumuom wouoaaom mw.m ooa Homma <29 aHumaounoououom oa.om kennaaowom <29 :Humaounopm Baowumnofiuomum Naowz mH.o mma quam <29 nauseounooumuom vouoaaom mm.q mum “mama <29 sfiumaounoououom ow.“ omen HmHNN <29 nauseounoom QdOHuNCOfiuomum uaofivmuw Houoomau omouwuomaooz coaxm<< <29 wa\299 oazm canmaonao on accumoumm mo monocam HoaoaoH o m9m pom .uswwu ou umoa Boum mp3 coaumuomaawmm .mwonuoz 9H ponauomov mm ueowwmuw Houoohaw a do woumcowuomnm was woumHomfi mp3 cwumaouso .mupon N Houmm wooHMHuomm was .9<) w 0.3th 3h. c.3825 oz 20:05: 33...... canon an... o. .. .13 A, QN r 1100-. w .. 0.» m ion. “.11 co .. icon 9w . G6... I HUI-T ill! ON? 2=<20¢1UO¢EI z=<¢ nuH3 v0u0H00H00 003 uamu0doo 0umu 9009 .m0dflau0u0v 00H0 003 09HH 090 mo ARV ua0fiofimm0oo 00H00H0nuoo 099 Imwoafia00 0 8099 U000H00H00 003 A200\u:0ou09v dowumofiawa0 mo ua0u0doo 0u0~ H0090 umuamo .08Hu .03 00:09 d0woafioumo mo 0099 oaanufin Ne.o o.ma me.o a.a~ mm.o mm.aa. eaomaonaoonooom ooooaaom no.0 a.ma mm.o o.m ae.o m.am annoaonaoonoaom mm.o «.mm am.o m.oa om.o o.o~ eaonaoneoom none mam Nmo.o no.0 me.aa mm.o a.ea ao.o oa.aa caomeonaoonooom ooooaaom am.o o.o~ aa.o m.oa Hm.o m.om canneoneoonooom ma.o p.mm Am.o a.m~ an.o m.mm aaaoaonaoom 00H9 Houuaoo H 0u02 u 0umm H 0u02 aze anemone czo oNHowz 9u00fivmuu Hou0omao 3 n mqme can con can Gnu CON oo— no— 2... un— fl In N iAIlV'lill r— 53 ms AllSNilNI cow 73 to aminofluorene (180), fluorene (16S), and guanine (152) can be seen. A minor peak of radioactivity (Rf = 0.67) was also reproducibly obtained from DNA purines. The minor peak had a R value identical to f that reported by Westra g£_§l, (205) for 3-(guanin—N2-yl)—acetyl- aminofluorene. Furthermore, as illustrated in the lower panel of Figure 11, this adduct could not be detected in purines isolated from RNA, in agreement with previously published reports (106). When the distribution of carcinogen bound to these two guanine sites was examined, the pelleted heterochromatin DNA (prepared by MgCl2 precipi- tation) had a significantly greater proportion of the Nz-guanine adduct than did either the eu- or heterochromatin DNA (Table 8). 5. DNAse II:MgClq Fractionation of Chromatin and Analysis of Carcino- gen-DNA Bind 16$ To confirm this nonrandom distribution of carcinogen adducts within the genome, a third chromatin fractionation procedure was utilized. As indicated in Figure 1, the selectivity of the DNAse II— MgCl2 technique has been verified by rigorous criteria. This proce- dure has effected a separation of DNA gene sequences coding for specialized functions in differentiated tissue. The fractionation procedure is diagramed in Figure 13, and the tflme course of formation of the 3 chromatin fractions is illustrated in Figure 14. Although it has been previously demonstrated (81) that a very short DNAse II digestion (e.g., 2-6 min) results in the greatest enhancement of euchromatin genes in the 82 fraction, a digestion time of 30 min was chosen to insure a more complete separation of the heterochromatin (P2 fraction) from the nuclease resistant fraction (Pl fraction). As 74 TABLE 8 Thin Layer Chromatography of Carcinogen-Base Adductsa % Total Radioactivityc Source N Adduct C8 Adduct Unfractionated DNA 12:1.SZ 88:1.SZ Euchromatin DNA 8:0.8Z 92:0.8% Heterochromatin DNA lli2.8%d 89i2.8% Pelleted Heterochromatin DNA 23il.2% 77i0.8% aAnimals were sacrificed 2 hours after injection of 3H-N-OH- AAF (100 uCi/lOO g). Silica gel thin layer chromatography and n—butanol:acetic acid: H 0 (50:11:25) were used to separate the adducts. Values are means i SEM of 3 rats. ngCl (1.75 mM) fractionation of chromatin as described in Methods. DNA was hydrolyzed from chromatin fractions and carcinogen modified bases were purified on a Sephadex LH-20 column as described in Methods. cThe radioactivity in each peak expressed as a percent of total CPM recovered on TLC plate. Average recovery was 1074:271 cpm/rat. Separation of carcinogen DNA adducts is illustrated in Figure 11. dlndicates significantly different from both eu- and hetero- chromatin, Student-Newman—Keuls test, p<.05. 75 .NHowszH mm hp wmvfi>am H¢OmNAqmzv on“ aoflumfiuficwmu :uHB mammnunkm «zmm .wmauom maflmno m ago wow mammnuchm mo omusoo maHu mom uaouxm man do comma wmumasoamo mmsHm> m .uaos\ms%uam moasomaoa Haoaxwm.a use .mzn ma non movfiuomaooa mo moaoa: oa.m mum muonu umnu moaommm mam muaammm ca wonauomow ma mouam doaumfiuaofi mo woman: «no mo doaumasoamo .oa onswfim ca voumuumoaafi mm mzn :uH3 ommumahaoa mzm mafia: mH.o mo Goaumuuflu mo unwom mam mnu aoum vmuaaumuoao .oofiumfiuflafimu vmxooan mo maofiufiwaoo Home: 429 mu m so nonemozuckm «zm Hmafixmz Q .oumofiamdc a“ commune mma nofina mo some moamamm <29 q mo sum H omoa onu ma 05Hm> 50mm .mvonumz nH monauomou mm < TIME (min) Figure 17 92 template; protein contamination increased to 23—44% (Table 13). When the rats were returned to a basal diet after 4 days of AAF ingestion, the degree of protein contamination of the DNA was reduced. RNA contamination of the DNA preparations remained consistently low in the control and treated groups. As illustrated in Figure 18, AAF ingestion significantly altered 12,!1££9_transcription of the DNA derived from the eu— and hetero- chromatin fractions. After 4 days of AAF ingestion (single hatched bars), total RNA synthesis was reduced 50 and 41% using the DNA templates derived from the eu— and heterochromatin, respectively. This reduction is not the result of a decrease in the number of initiation sites, but rather is due to a decrease in the average size of the RNA chains formed. The calculated rate of RNA synthesis on the heterochromatin DNA was decreased (this decrease in rate, however, was not statistically significant). When the animals were returned to a basal diet for 7 days after ingestion of AAF for 4 days (cross hatched bars), both the extent and size of RNA synthesized, as well as the rate of synthesis returned to approximately control values. In contrast to the inhibition of RNA synthesis seen with eu- and hetero- chromatin DNA templates, there was no significant decrease in RNA size or extent on DNA of the pelleted chromatin. As illustrated in Figure 19, there was an inverse correlation between the average RNA chain length and the extent of enzyme turn- over (number re-initiations per 60 min), regardless of the source of the DNA template. Statistically significant differences in enzyme 93 .mmowanoumm mm3 <29 mnu ou m>HumHmu ofiououm mam vao> ecu momma umoh waHuSHo Hmfiumuma one .nasaoo am.al< Howlofim m on wmwamam mm3 ucmu Incummom mnH .wmwomauucmo mom .Homz ou uoommmu sows 2m ovma .mmmdoum .ommzm cufis commune mums mcofiuomum ofiumaouno may .aHuNEouno wmooaama can wcfiuamaavmm ummm .wcaucmefiwom 30am ouaa unmawmuw Houmomaw u no woumaofiuomum was afiumeousuu NN.NH m.mH Nm.mem.NH NH.NHN.HH cemeoem N Hmemm ween N Nm.He m.m Ne.HHa.N Nm.NHN.e azm N + m<< mean e NeHmN Nnnam Neaeee cemuoem N No.NH N.N No.NHo.m NH.HHN.N «zm N m<¢ mean e No.oe o.NH NH.mHo.m Nm.Hew.o :Nmuoum N Noo. “No.0 No.Hem.e Nm.ono.H oau:u mm¢um8>aoa .oumafivuo ago so wouuoaa we woufimozu iamm wuwaon any .mwonuoz ca pmnfiuomow mm ammuoahaom mzm Naoo aw sows wmnfiuomamuu mam wounmoua mama moumHmsou mza woumHomH .mfiwoSuazm