Reversibility of changes in nucleic acid methylation and gene expression induced in rat liver by severe dietary methyl deficiency.

As we have reported previously, both DNA and tRNA become hypomethylated in livers of rats fed a cancer promoting, methyl-deficient diet (MDD) for as short a period as one week. Within the same period, activities of tRNA and DNA methyltransferases (MTases) increase and levels of mRNAs for several genes believed to have roles in growth regulation are altered. These diet-induced changes in nucleic acid methylation and gene expression increased in extent when MDD was fed continuously for four weeks. We also observed hypomethylation of specific CCGG sites within several genes for which mRNA levels were increased. These included c-myc, c-fos and c-Ha-ras. To investigate the reversibility of such diet-induced alterations in methylation and gene expression, animals were fed MDD for four weeks, after which a diet supplemented with adequate sources of methyl groups (CSD) was fed for 1-3 weeks. One to two weeks after the restoration of an adequate diet, the overall extent of methylation of tRNA and DNA from livers of these rats did not differ from that of tRNA and DNA from livers of age matched animals continually maintained on CSD. At the same time, activities of MTases in the liver dropped to normal values. Levels of mRNAs for all genes studied returned to control levels within three weeks after ending MDD feeding, although at different rates. In contrast, MDD-induced hypomethylation of some HpaII sites in c-myc, c-fos and c-Ha-ras genes persisted after 3 weeks refeeding of an adequate diet. These results, which demonstrate that most of the effects of MDD on the parameters we have studied occur rapidly and are essentially reversible, are consistent with the role of MDDs as promoters of hepatocarcinogenesis. However, the finding that unmethylated sites persist in genes that play a role in growth regulation suggests a mechanism by which intermittent or long term exposure to MDDs could result in heritable phenotypic changes in some hepatocytes that lead to hyperplasia and tumorigenesis.

Methyl groups in carcinogenesis: effects on DNA methylation and gene expression.

Lipotrope-deficient (methyl-deficient) diets cause fatty livers and increased liver-cell turnover and promote carcinogenesis in rodents. In rats prolonged intake of methyl-deficient diets results in liver tumor development. The mechanisms responsible for the cancer-promoting and carcinogenic properties of this deficiency remain unclear. The results of the experiments described here lend support to the hypothesis that intake of such a diet, by causing depletion of S-adenosylmethionine pools, results in DNA hypomethylation, which in turn leads to changes in expression of genes that may have key roles in regulation of growth. In livers of rats fed a severely methyl-deficient diet (MDD), lowered pools of S-adenosylmethionine and hypomethylated DNA were observed within 1 week. The extent of DNA hypomethylation increased when MDD was fed for longer periods. The decreases in overall levels of DNA methylation were accompanied by simultaneous alterations in gene expression, yielding patterns that closely resembled those reported to occur in livers of animals exposed to cancer-promoting chemicals and in hepatomas. Northern blot analysis of polyadenylated RNAs from livers of rats fed control or deficient diets showed that, after 1 week of MDD intake, there were large increases in levels of mRNAs for the c-myc and c-fos oncogenes, somewhat smaller increases in c-Ha-ras mRNA, and virtually no change in levels of c-Ki-ras mRNA. In contrast, mRNAs for epidermal growth factor receptor decreased significantly. The elevated levels of expression of the c-myc, c-fos, and c-Ha-ras genes were accompanied by selective changes in patterns of methylation within the sequences specifying these genes. Changes in DNA methylation and in gene expression induced in livers of rats fed MDD for 1 month were gradually reversed after restoration of an adequate diet. In hepatomas induced by prolonged dietary methyl deficiency, methylation patterns of c-Ki-ras and c-Ha-ras were abnormal. Although human diets are unlikely to be as severely methyl deficient as those used in these experiments, in some parts of the world intake of diets that are low in methionine and choline and contaminated with mycotoxins, such as aflatoxin, are common. Even in industrialized nations, deficiencies of folic acid and vitamin B12 are not uncommon and are exacerbated by some therapeutic agents and by substance abuse. Thus, it seems possible that interactions of diet and contaminants or drugs, by inducing changes in DNA methylation and aberrant gene expression, may contribute to cancer causation in humans.

Reduced membrane protein methylation in red cells of the McLeod blood group phenotype.

Red cells (RBCs) contain an abundance of protein methylase II, which catalyzes the transfer of methyl groups from S-adenosylmethionine to carboxyl groups of aspartyl and glutamyl residues in proteins. Enzyme-catalyzed transfer of methyl groups, labeled with 14C or 3H, from S-adenosylmethionine to membrane proteins of McLeod, Ko, and control RBCs was assayed by determining the acceptance of labeled methyl groups under standardized conditions. Membranes of control cells and Ko cells showed about 50 percent greater uptake than did those of McLeod cells. However, when ovalbumin was used as a methyl-accepting substrate, the levels of protein carboxymethyltransferase activity in all three types of cells were found not to differ significantly. In addition, no significant qualitative differences were apparent when methyl-labeled polypeptides from control and McLeod cells were separated by slab gel electrophoresis. The mechanisms responsible for changes in membrane protein methylation of McLeod cells remain unclear. However, these observations provide further evidence of the pleiotropic biochemical lesion associated with the acanthocytic morphology that characterizes McLeod RBCs.

Alterations in expression and methylation of specific genes in livers of rats fed a cancer promoting methyl-deficient diet.

We have reported earlier that hypomethylated DNA is rapidly induced in the livers of male Fischer rats fed an extremely methyl-deficient diet (MDD). The early effects of dietary methyl deficiency on the expression of several genes in the livers of such animals have now been investigated. Poly(A)+ RNA was isolated from the livers of rats fed MDD or a similar diet supplemented with adequate supplies of choline, methionine, folic acid and vitamin B12 (CSD) for periods ranging from 1 to 4 weeks. The levels of mRNAs for the c-myc and c-fos protooncogenes in livers of rats given either MDD or the liver carcinogen, 2-acetylaminofluorene (AAF), were compared by Northern blot analysis with those in livers of animals given control diets. Both AAF and MDD induced significant elevations in levels of mRNAs specific for these two genes. After 1 week of MDD intake, large increases in the levels of c-myc and c-fos mRNAs and a smaller increase in the levels of c-Ha-ras mRNAs were observed. In contrast, there were marked decreases in the levels of mRNAs for epidermal growth factor receptor and for epidermal growth factor. These effects on mRNA accumulation persisted and were further enhanced during a 4 week period of MDD feeding. The appearance of hypomethylated DNA in the livers of these MDD-fed rats coincided with the observed changes in levels of mRNA for these genes associated with the regulation of cell growth. Increases in levels of mRNA for c-fos, c-Ha-ras and c-myc were correlated with loss of methylation at specific sites within these genes as early as 1 week after the start of MDD feeding. These combined observations are consistent with the hypothesis that methyl-deficient diets are cancer promoting and/or carcinogenic, at least in part, because they induce hypomethylation of DNA with concomitant alterations in the regulation of gene expression.

Prolonged survival of female AKR mice fed diets supplemented with methionine and choline.

Female mice of the AKR/J (AK) strain were fed a control diet (Purina Rodent Laboratory Chow) or a lipotrope-supplemented diet (Purina Rodent Chow plus 2% D,L-methionine and 1% choline chloride) beginning at 1 day after weaning. Food consumption and weight gain were found to be the same in both groups of animals. Mice of this inbred strain spontaneously develop thymic lymphoma, with close to 100% mortality expected by 12-13 months of age. Two separate experiments were carried out with 50 mice per group in one, and 40 mice per group in the other. The slopes of the survival curves for the animals in the control group and supplemented group of mice diverged after the animals reached 6.5 months of age. In both experiments, 20% of the mice receiving supplemented diet were still alive at 1 year, while 3% in one experiment and 8% in the other experiment survived in the control groups. Each experiment was terminated when the animals reached 13 months of age. At that time the survival rate of the controls was 2 and 4%, and survival in the groups of mice receiving supplemented diet was 14 and 18%. Necropsy revealed that the animals in both groups had advanced malignant lymphoma. Our results demonstrate that intake of a chow diet that is supplemented with moderate quantities of methionine and choline results in enhanced survival of spontaneously leukemic AK mice, in comparison with animals of this strain fed the same diet without supplements of choline and methionine.

Rapid appearance of hypomethylated DNA in livers of rats fed cancer-promoting, methyl-deficient diets.

Prolonged intake of diets deficient in sources of methyl groups leads to development of hepatomas in rats and promotes chemical carcinogenesis in both rats and certain strains of mice. Since methylation of cytosine residues in regulatory regions can affect gene activity, several investigators have postulated that the effects of methyl-deficient diets on tumorigenesis result from the inability of cells to maintain normal patterns of DNA methylation. However, significant decreases in the 5-methylcytosine content of liver DNA have not been reported to occur until rats have consumed methyl-deficient diets for several months. To determine whether methyl-deficient diets have immediate effects on nucleic acid methylation, we assessed the degree to which hepatocyte DNA and tRNA were methylated in vivo, by measuring their ability to act as methyl acceptors in vitro. Hypomethylation of DNA and tRNA was detected within 1 week after rats were started on a diet deficient In methionine, choline, folic acid, and vitamin B12 and it persisted throughout the 4 weeks of study. A significant elevation in liver DNA synthesis occurred in parallel with increased hypomethylation of DNA. Chronic failure to fully methylate DNA that is newly synthesized in response to liver damage induced by methyl-deficient diets provides a feasible mechanism for changing patterns of DNA methylation. Our results indicate that such changes could occur rapidly enough to play a causal role in the cancer-promoting and, in some instances, cancer-inducing properties of the diet.

Altered expression of retrovirus-like sequences and cellular oncogenes in mice fed methyl-deficient diets.

Methyl-deficient (lipotrope-deficient) diets enhance liver carcinogenesis in rodents. Although the mechanisms responsible for the cancer-promoting activity of such diets have not been identified, they have been observed to cause impaired immune response, alterations in methylation of liver RNA and DNA, and enhanced susceptibility to oxidative damage. Since alterations in gene expression may also play a critical role, the present studies examined the expression of the c-myc, c-H-ras, epidermal growth factor receptor, and ornithine decarboxylase genes, as well as endogenous retrovirus-like sequences, in C57BL/6J x C3H/HeJ F1 mouse liver during the first 2 weeks of feeding of a methyl-deficient diet. The kinetics of liver cell proliferation was investigated in parallel. Increased [3H]thymidine incorporation into liver DNA was found at day 4 and reached a maximum at days 7-11 after commencement of the methyl-deficient diet, when compared to age-matched mice fed a complete diet. Northern blot analysis of polyadenylated liver RNA samples indicated an increase in the levels of RNA homologous to Moloney murine leukemia virus and intracisternal A particle sequences but no significant change in the level of VL30 retrovirus-related RNA in the samples from mice fed methyl-deficient diets. A marked increase in the levels of c-myc and a slight increase in the levels of ornithine decarboxylase and c-H-ras transcripts were seen in the liver RNA samples from the treated mice. Of particular interest was a decrease in the abundance of epidermal growth factor receptor transcripts in the liver RNA samples from the treated mice. These changes in cellular levels of specific RNA resemble, in several respects, those we have previously described in rodent liver during regeneration and tumor promotion and also those seen in rodent hepatomas. They may reflect, therefore, a common profile of gene expression relevant to cell proliferation.

Comparison of methyltransferase activities of pair-fed rats given adequate or methyl-deficient diets.

The short-term effects of a lipotrope-deficient (methyl-deficient) diet on tRNA and protein methyltransferase activities have been studied using pair-fed male Fischer rats. The activity of liver N2-guanine tRNA methyltransferase II (NMG2) of animals receiving the methyl-deficient diet (MDD) for 2 weeks was found to be elevated more than 2-fold. This is in agreement with the results of earlier experiments in which the animals were fed ad libitum. These data indicate that the effects of lipotrope-deficient diets on NMG2 activity observed in the earlier studies can be attributed to the nature of the diet, and not to differences in caloric intake. In the same pair-fed animals, very little effect of MDD on the activity of NMG2 of either brain or spleen was observed. In liver, the activity of one of the enzymes that catalyze protein methylation--protein methylase I (S-adenosyl-methionine: protein-arginine N-methyltransferase)--was significantly elevated in response to the lipotrope-deficient diet. In contrast, the activities of protein methylase II (S-adenosylmethionine: protein-carboxy-O-methyltransferase), from control and experimental animals did not differ significantly. Lipotrope-deficient diets are thus seen to induce, within a short period of time, selective changes in the activities of some, but not all, of the liver enzymes that catalyze the methylation of tRNA and protein.

Suppression by methionine and choline of onco-fetal patterns of liver tRNA methyltransferase activities in carcinogen-treated rats.

When male Fischer rats were fed Purina chow supplemented with 2% D,L-methionine and 1% choline chloride, the rapid increase in N2-guanine tRNA methyltransferase II (NMG2) activity otherwise seen in response to cancer-promoting doses (0.02% in the diet) of 2-acetylaminofluorene (AAF) was prevented, and the increase in NMG2 activity otherwise caused by carcinogenic doses of AAF (0.06% in the diet) was decreased by 50%. In addition, the return of NMG2 activity to a normal level after completion of a 3-week regimen of 0.06% AAF was accelerated in animals fed the methionine plus choline supplemented diet. As shown earlier in this laboratory, liver tRNA methylating enzyme activities are shifted rapidly to an onco-fetal pattern in rats receiving methyl-deficient diets. This pattern is characterized by selectively elevated NMG2 activity while the activities of other base-specific tRNA methylating enzymes are relatively unchanged. Our combined results indicate that the exogenous supply of methyl groups is a factor in regulating NMG2 activity and can modulate at least one response of animals to carcinogens.

Altered tRNA methylation in rats and mice fed lipotrope-deficient diets.

A diet that is deficient in methionine, choline, folic acid and vitamin B12 has been found to induce alterations rapidly in liver tRNA methylation in male Fischer rats. In vitro assays indicated that activity of N2-guanine tRNA methyltransferase II (NMG2) was increased to 150% of controls levels in 1 week and 300% of control levels after 2 weeks or longer on this diet. Incompletely methylated tRNA was isolated from livers of these same animals, indicating that there was impairment of methylation in vivo. The effects on liver tRNA methylation of this methyl-deficient diet were thus seen to mimic those of the liver carcinogen, ethionine, which also causes production of hypomethylated tRNA and increased activity of NMG2. The effect of the same diet on liver tRNA methyltransferase activity of C57BL/6J and C3H/HeJ inbred mice were also studied. Intake of the lipotrope-deficient diet induced elevation in activity of liver N2-guanine tRNA methyltransferase II activity in C57BL/6J mice, similar to that seen in rats. In contrast, the methyl-deficient diet had very little effect on the same enzyme activity in C3H/HeJ animals.

Differences in activity of N2-guanine tRNA methyltransferase II among several inbred strains of mice.

The tRNA methyltransferase activities of C57BL/6J, C57L/J, C58/J, AKR/J, and C3H/HeJ inbred mice were studied with the use of various amino acid-specific Escherichia coli tRNA's as substrates. Mice from two strains with high incidence of spontaneous leukemia (AKR/J and C58/J) exhibited levels of liver N2-guanine tRNA methyltransferase II (N2-MeGII) activity that were double those of two strains of mice with low incidence of spontaneous leukemia (C57BL/6J and C57L/J). Activities of liver and kidney N2-MeGII of the high spontaneous hepatoma strain C3H/HeJ were also found to be twice as high as those of C57BL/6J mice. The activities of other tRNA base-specific liver tRNA methyltransferases were very similar in all strains studied. The N2-MeGII activity of the F1 progeny of a cross between C57BL/6J and C3H/HeJ showed levels of activity intermediate to those of the parental strains. Activities of liver N2-MeGII of two inbred strains of mice that differ in their H-2 haplotype (C57BL/10SnJ and the congenic strain B10.BR/SgSnJ) were also compared. Both C57BL/10SnJ and B10.BR/SgSnJ strains exhibited low levels of liver N2-MeGII activity, indicating that H-2 does not directly control the activity of this enzyme.

Increased activity of rat liver N2-guanine tRNA methyltransferase II in response to liver damage.

Alterations in rat liver transfer RNA (tRNA) methyltransferase activities have been observed after liver damage by various chemicals or by partial hepatectomy. The qualitative and quantitative nature of these activity changes and the time course for their induction have been studied. Since homologous tRNAs are essentially fully modified in vivo, E. coli tRNAs were used as in vitro substrates for the rat liver enzymes in these studies. Each of the liver-damaging agents tested rapidly caused increases in activities of the enzyme(s) catalyzing methyl group transfer to tRNAs that have an unmodified guanine at position 26 from the 5' end of the molecule. This group of tRNAs includes E. coli tRNANfmet, tRNAAla1, tRNALeu1, or Leu2, and tRNASer3 (Group 1). In each case N2-methylguanine and N2,N2-dimethylguanine represented 90% or more of the products of these in vitro methylations. The product and substrate specificity observed are characteristic of N2-guanine methyltransferase II (S-adenosyl-L-methionine : tRNA (guanine-2)-methyltransferase, EC 2.1.1.32). In crude and partially purified preparations derived from livers of both control and treated animals this enzyme activity was not diminished significantly by exposure to 50 degrees C for min. The same liver-damaging agents induced little or no change in the activities of enzymes that catalyze methyl group transfer to various other E. coli tRNAs that do not have guanine at position 26 (Group 2). The results of mixing experiments appear to rule out the likelihood that the observed enzyme activity changes are due to stimulatory or inhibitory materials present in the enzyme preparations from control or treated animals. Thus, our experiments indicate that liver damage by each of several different methods, including surgery or administration of chemicals that are strong carcinogens, hepatotoxins, or cancer-promoting substances, all produce changes in liver tRNA methyltransferase activity that represent a selective increase in activity of N2-guanine tRNA methyltransferase II. It is proposed that the specificity of this change is not fortuitous, but is the manifestation of an as yet unidentified regulatory process.

Comparisons of liver transfer RNA methyltransferase and adenosylmethionine decarboxylase activities of male and female rats.

The activities of liver transfer RNA (tRNA) methyltransferases from control or ovariectomized female rats were found to be higher than those of control or castrated males. Administration of testosterone to ovariectomized females caused activity to decrease to the level of the males. Conversely, administration of estrogen to castrated males resulted in liver enzyme levels similar to those of the females. When the substrates for in vitro methylation were either mixed heterologous tRNA's from Escherichia coli or mixed homologous methyl-deficient tRNA from livers of ethionine-treated rats, the difference in activity between males and females was about 35%. When amino acid-specific tRNA's from E. coli were used as substrates, the ratios of activity of enzymes from females to that of males were: tRNANfMet 1.5; tRNAMetMet 1.1; tRNASer3 1.85; tRNAPhe 1.1; and tRNATyr 1.25, indicating that there are qualitative as well as quantitative differences in the liver tRNA methyltransferases of the two sexes. The adenosylmethionine decarboxylase activity of female rat liver preparations was approximately double that found for males. Testosterone, given to ovariectomized females, lowered the activity of this enzyme to about the same level as that of males. It is not clear whether the observed sex-related differences in activity of several adenosylmethionine-utilizing liver enzymes represent isolated phenomena or are indicative of a sex-related difference in the rate of liver adenosylmethionine turnover.

Selective inhibition of uracil tRNA methylases of E. coli by ethionine.

L-ethionine has been found to inhibit uracil tRNA methylating enzymes in vitro under conditions where methylation of other tRNA bases is unaffected. No selective inhibitor for uracil tRNA methylases has been identified previously. 15 mM L-ethionine or 30 mM D,L-ethionine caused about 40% inhibition of tRNA methylation catalyzed by enzyme extracts from E. coli B or E. coli M3S (mixtures of methylases for uracil, guanine, cytosine, and adenine) but did not inhibit the activity of preparations from an E. coli mutant that lacks uracil tRNA methylase. Analysis of the 14CH3 bases in methyl-deficient E. coli tRNA after its in vitro methylation with E. coli B3 enzymes in the presence or absence of ethionine showed that ethionine inhibited 14CH3 transfer to uracil in tRNA, but did not diminish significantly the 14CH3 transfer to other tRNA bases. Under similar conditions 0.6 mM S-adenosylethionine and 0.2 mM ethylthioadenosine inhibited the overall tRNA base methylating activity of E. coli B preparations about 50% but neither of these ethionine metabolites preferentially inhibited uracil methylation. Ethionine was not competitive with S-adenosyl methionine. Uracil methylation was not inhibited by alanine, valine, or ethionine sulfoxide. It is suggested that the thymine deficiency that we found earlier in tRNA from ethionine-treated E. coli B cells, resulted from base specific inhibition by the amino acid, ethionine, of uracil tRNA methylation in vivo.

Time dependence of ethionine-induced changes in rat liver transfer RNA methylation.

Methyl-deficient transfer RNA (tRNA) and subnormal levels of tRNA-methylating enzymes were found in the livers of female rats that had received injections of 250 mg DL-ethionine per kg body weight per day and 120 mg adenine per kg body weight per day for 2 days. Adenine alone had no effect. When the ethionine plus adenine injections were continued for longer periods of time, liver tRNA-methylating enzyme activity measured in vitro gradually increased and exceeded that of the controls. Concurrently, the relative methyl deficiency of liver tRNA decreased. The latter was evident because of the decreased ability of the tRNA to accept methyl groups during in vitro methylation catalyzed by homologous enzymes. Liver tRNA from animals that were treated with ethionine for 7 days could accept only about 40% as many methyl groups as could tRNA from animals that had received ethionine for only 2 days. No further significant change in methyl deficiency of the tRNA was seen when ethionine administration was extended to a total of 14 days. Enzyme preparations from ethionine-treated, but not control, rat livers contained dialyzable substances that inhibited the tRNA methylases and altered the base specificity of these enzymes. Although S-adenosylhomocysteine and S-adenosylethionine were found to be present in the liver preparations, neither of these substances could account for the observed changes in specificity.

Ethionine-induced changes in rat liver transfer RNA methylation.

We have confirmed the finding by Rajalakshmi that transfer RNA (tRNA) from livers of ethionine-treated rats can act as a substrate for homologous tRNA-methylating enzymes in vitro. This methyl-deficient tRNA from liver can be methylated in vitro by enzymes from normal or ethionine-treated rats. The in vitro inhibition of tRNA methylation that follows ethionine treatment can be at least partially relieved in vitro. The liver extracts from ethionine-treated animals contained a low-molecular-weight inhibitor of tRNA methylation. Dialysis of enzyme preparations from ethionine-treated, but not control, rats resulted in large increases in tRNA methylase activity, with either Escherichia coli or homologous tRNA's as substrate. Furthermore, the tRNA methylase activity of control rat liver enzyme extracts was greatly depressed by dialysate from liver homogenates of ethionine-treated rats. After 5 days of ethionine administration the liver tRNA methylase activities were significantly higher than those of control preparations despite the continued presence of the dialyzable inhibitor(s). The liver tRNA's from these animals were poorer methyl acceptors than those from 3-day-treated rats, although still better than tRNA's from untreated rats. These observations have been interpreted to indicate that ethionine causes the accumulation in the liver of inhibitors of tRNA methylation. Early in the course of ethionine administration, methyl-deficient tRNA can be isolated. When the period of ethionine treatment is extended, the organism attempts to maintain homeostasis by production of increased amounts of tRNA-methylating enzymes. The increased quantities of these enzymes are able to overcome, at least partially, the effects of the inhibitors and to decrease the extent to which methyl-deficient tRNA is produced.

Inhibition of methylated nucleoside synthesis in vivo: accumulation of incompletely methylated transfer RNA in ethionine-treated cells of Escherichia coli B.

tRNA prepared from cells of E. coli B that had been incubated with 0.5% DL-ethionine (Ethio sRNA) was found to accept methyl groups from 14CH3-S-adenosyl-methionine in the enzymatic reaction catalyzed in vitro by tRNA methyl transferases from untreated cells of the same organism. tRNA from cells that were not exposed to ethionine did not accept a significant level of methyl groups when incubated with the same enzyme system. Base ratio analysis of the product obtained after in vitro addition of methyl groups to Ethio sRNA by enzymes from normal E. coli B indicated that a high proportion of uracil sites in this tRNA were available for enzymatic methylation. These results indicated that tRNA from ethionine-treated organisms was recognized by the homologous enzymes to be incompletely methylated, while, as previously shown, all methyl-acceptor sites on tRNA from normal cells were already filled, and that Ethio sRNA was preferentially deficient in methyl groups on uracil moieties in the RNA molecules. Ethionine thus appears to interfere with normal tRNA modification in vivo.