In Vitro Studies on the Methylation of Histones in Rat Brain Nuclei*

When isolated nuclei from 12-day-old rat brains were incubated with S-adenosyl-~-[meth&3Hlmethionine, significant amounts of 3H-methyl were incorporated into lysyl residues in histones H3 and H4. About 0.024% of the total methylation sites on histone H3 and 0.013% of the sites on histone H4 were unmethylated at the time the nuclei were isolated. Methylation of these sites proceeded stepwise, pro-gressing to a stable ratio of 0.93:1.0:0.17 for w-mono-, N’- di-, andN’-trimethyllysine in histone H3 and 0.19:l.O for N’-mono- and N’-dimethyllysine in histone H4. The K, values of the enzyme for S-adenosyl+methionine were 11.5 YZ 1.1 PM and 12.5 f 1.3 PM with histones H3 and H4 as methyl acceptors, respectively. The V,,, values were 11.1 and 5.3 pmol of 3H-methyl incorporatedlminfmg of histone H3 and H4, respectively. Since histone H3 contains 2 mol of N’-methyllysinelmol and histone H4 contains 1 mollmol, no difference in the overall rates of methylation

It is now well established that specific lysyl residues on the arginine-rich histones are methylated. Histone H4 from calf thymus contains a single methylated lysyl residue at position 20, while histone H3 from calf thymus or carp testis contains methylated lysyl residues at positions 9 and 27 (l-3). Both sites on histone H3 contain W-mono-, W-di-, and NC-trimethyllysine, while NC-trimethyllysine is absent from histone H4. Honda et al. (4) have obtained similar results for the arginine-rich histones from trout testis; however, histone H3 contains an additional minor methylated residue at position 4.
The methylation sites and amino acid sequences of histone H3 and H4 are highly conserved (l-13). However, there appear to be differences in the extent of methylation of lysyl residues at these sites. This is particularly striking in the pea, where histone H4 is entirely unmethylated and trimethyllysine is absent from histone H3 (6,(14)(15)(16). Some of the differences in methylation appear to be genetic, while others may be related to the mitotic state of the cells at the time the histones were isolated from a particular organ. In HeLa cells, methylation * This investigation was supported by Grant NS-09725-05 from National Institute of Neurological Diseases and Stroke.
occurs mainly in the late S cycle (17,18). The ratio ofN'-monoto N'-dimethyllysine in the malignant cell was about twice that found in the normal cell. In the rat, the extent of methylation of the arginine histones does not vary from organ to organ, but varies significantly with age (19). In histone H3 from lo-day-old rats, the molar ratio of mono-:di-:trimethyllysine was 0.55:1.0:0.35. In histone H4 from these same age animals, the molar ratio of mono-:dimethyllysine was 0.1:O.g. These ratios shift towards the more highly methylated forms with age. If methylation of histones is a late event occurring after the histones are transported into the nuclei, then the newly synthesized polypeptide chains should not be fully methylated in rapidly proliferating tissue. In this communication, we employed nuclei from the brain of young rats in order to determine: (a) the number of unmethylated lysyl residues in histones, (b) if methyl groups are added sequentially or simultaneously, (c) if histones are methylated prior to or after they are bound to DNA, and (d) some of the kinetic properties of the histone-lysine methyltransferase. S-Adenosyl-L-homocysteine, S-adenosyl-L-homocysteine sulfoxide, and S-ribosyl-L-homocysteine were prepared by the method of Duerre et al. (21). S-Adenosyl-n-homocysteine was a gift from Dr. Fritz Schlenk, Argonne National Laboratories, Argonne, Ill. S-Adenosyl-thio-cr-ketobutyrate was prepared enzymatitally from S-adenosyl+homocysteine using the L-amino-acid oxidase from Proteus rettgeri (22) and purified according to the procedure of Duerre et al. (23). Methylthioadenosine was prepared following the procedure of Schlenk and Ehninger (24). L-Homocysteine was prepared from L-homocysteine thiolactone with alkali (25). Cyclic adenosine 3':5'-monophosphoric acid was obtained from Sigma Chemical Co. Preparation of Nuclei -Long-Evans rats, 12 to 14 days old, were killed by decapitation and their brains were removed. The brains were placed in cold 0.32 M sucrose containing 1.0 mM MgCl, and 1.0 rnM potassium phosphate buffer, pH 7.6 (buffered sucrose). The brains were homogenized and the crude nuclei were harvested by centrifugation (26). on Bio-Gel P-10 as described in Fig. 1. Proteins under the peaks were quantitated by the method of Lowry et al. (27) using a Technicon auto-analyzer. Radioactivity under peaks was determined by placing 0.2-ml aliquots of each fraction in 10 ml of scintillation fluid and counting in a Packard Tri-Carb scintillation spectrometer (28). Fractions under the peaks with uniform specific activities were pooled and concentrated by lyophilization. Polyacrylamide gel electrophoresis based on the method of Panyim and Chalkley (29) was used to check the purity of the histone fractions. From 1.0 to 2.0 mg of histone H3 and H4 were hydrolyzed and the basic amino acid composition was ascertained on Beckman PA-35 resin as previously described (19). Similar experiments were attempted using brain nuclei purified by centrifugation through 1.77 M buffered sucrose. However, nuclei so prepared were difficult to resuspend in either 0.32 or 1.77 M sucrose and had a tendency to lyse upon incubation at 37". Preparation of Chromatin -Crude nuclei were prepared from 12to 14-day-old rats as described above. After purification by centrifugation through 1.77 M sucrose, the nuclei were lysed with 20 mM phosphate, pH 7.6. The chromatin was harvested by centrifugation and washed twice with the same buffer.
The chromatin was suspended 20 mM phosphate buffer (final pH 6.8) at a concentration of approximately 5 mg of DNA/ml. DNA was quantitated by the method of Burton (30).

RESULTS
When isolated rat brain nuclei were incubated with Sadenosyl-n-[methyZ-3Hlmethionine, significant amounts of 3Hmethyl groups were incorporated into histones H3 and H4 (Fig. 1). There was a measurable amount of radioactivity associated with a protein which eluted just after histone Hl. This may be the result of dimerization of histone H3; however, this has not been fully established.
The time course of incorporation of 3H-methyl groups into histones H3 and H4 in intact nuclei is presented in Fig. 2. Methylation of these histones proceeded at a linear rate for 10 to 15 min, reaching saturation after 40 min. The further addition of S-adenosyl-L-[methyZ-3H]methionine at this point has no effect; therefore, all available sites appeared to be fully methylated.
Methylation of histone H3 proceeded about 2.6 times faster than methylation of H4.
The distribution of methylated basic amino acid residues was ascertained by amino acid analysis. All the radioisotope incorporated into histones H3 and H4 occurred as [methyl-3HlW-methyllysines.
There was no radioactivity detectable in the regions where methylarginine or methylhistidine eluted from the column. The rates of incorporation of 3H-methyl groups into N'-methyllysines from histones H3 and H4 are presented in Fig. 3. Monomethyl groups were added to both histones at a rapid linear rate, then decreased with time. There was a short lag prior to the formation of W-dimethyllysine in both histones, while there was a much more pronounced lag prior to the formation of NC-trimethyllysine in histone H3. In histone H3, the ratio of mono-:di-:trimethyllysine shifted toward the higher methylated lysine forms with time (Table I). After 40 min, the ratio approached that which we had previously observed in vivo (0.55:1.0:0.35); however, there appeared to be somewhat more NC-monomethyllysine in the in vitro experiment. As with histone H3, the ratio of mono-:dimethyllysine in histone H4 shifted toward the higher methylated forms (Table II). The ratio after 40 min approached that which we had previously observed in vivo (19). From the above data, it appeared that methylation proceeded stepwise, that is mono-to di-to trimethyllysine in histone H3 and mono-to dimethyllysine in histone H4. At saturation, 0.48 and 0.13 mmol of "H-methyl were incorporated/mol of histone H3 and H4, respectively. The total amount of NC-methyllysine was 2.0 mol/mol of histone H3 and 1.0 mol/mol of histone H4. From the radioisotope data, it can be calculated that 0.024% of the histone H3 molecules and 0.013% of the histone H4 molecules remained unmethylated at the time the animals were killed. This probably represents newly synthesized histone which has not as yet been methylated.
The unmethylated histones could possibly be free in the nucleoplasm.
However 1. In vitro incorporation of 3H-methyl groups into rat brain histones.
Freshly prepared rat brain nuclei (about 20 mg of DNA) were incubated at 37" with 15 FM S-adenosyl-L-[methyl-3H]methionine (1.0 Ci/mmol) 0.32 M sucrose, 1.0 rnM MgCl,, and 1.6 mM phosphate buffer, pH 7.6. The final pH of reaction mixture was 6.8. After 5 min of incubation, the reaction was terminated and histones isolated as outlined under "Experimental Procedures." The histones were dissolved in 0.01 M HCl, containing 1.0 mM dithiothreitol and 6.0 M urea. Ten-milligram samples were placed on a Bio-Gel P-10 column (1.0 cm x 3.5 m). The proteins were eluted with 0.01 M HCl and collected in 30-min fractions.
Aliquots of each fraction were used for the determination of radioactivity and protein concentration.    K, for S-adenosyl-L-methionine was 11.5 k 1.1 PM (Fig. 4).
The V,,, was 11.1 pmol of 3H-methyl incorporated/min/mg of histone H3. This value was about twice that observed with histone H4 (5.3 pmol of 3H-methyl/min/mg). Since hi&one H3  After 4 min, the reaction mixtures were made 10.0 PM with respect to S-adenosyl+homocysteine and chilled to 0". The histones were prepared and fractionated as outlined under Fig. 1.
In Vitro Methyl&ion of Histones contains two methylation sites and histone H4 only one site, no difference in overall rates of methylation of the two histones can be deduced from this data. The effect of several compounds on the rate of methylation of histones H3 and H4 was measured. Of all the compounds tested, including S-adenosyly-thio-a-ketobutyrate, S-ribosyl-L-homocysteine, S-adenosyl-L-homocysteine sulfoxide, adenosine, adenine, methylthioadenosine, CAMP, ATP, homocysteine, homoserine, and methionine, only the D and L isomers of S-adenosylhomocysteine had any effect on that rate of methylation of histones H3 and H4. Inhibition of the enzyme by Sadenosyl-L-homocysteine was of the competitive type with respect to S-adenosyl-L-methionine (Fig. 4). The inhibition constants (K,) for S-adenosyl-n-homocysteine were 5.5 2 0.4 pM and 5.9 & 0.5 FM with H3 and H4 as methyl acceptors, respectively.

DISCUSSION
One of the major difficulties in studies on the in vitro methylation of histones is obtaining a suitable substrate. Histones H3 and H4 do not turn over in adult brain, nor do the methyl groups turn over independently of the polypeptide chain (31). The irreversibility of methylation of histones has also been reported in other tissues (32,33). Consequently, histones from such tissues should be fully methylated. This has been substantiated in the brain by both in vitro and in viva studies. When brain nuclei from adult animals were incubated with radiolabeled S-adenosyl-L-methionine, histones H3 and H4 failed to incorporate significant amounts of labeled methyl groups (34). Furthermore, when adult rats are given radiolabeled lysine and methyl-labeled methionine, only trace quantities of radioactivity were incorporated into brain histones.' In contrast, significant amounts of 3H-methyl groups were incorporated into histones when the nuclei were prepared from brains of young rats (Fig. 2). Even in these nuclei, the number of unmethylated lysyl residues in histones H3 and H4 are quite limited (Tables I and II). Apparently a small fraction of the cells are in a state in which the newly synthesized histones have condensed with DNA, but are not yet fully methylated.
Following partial hepatectomy, methylation of histones was found to be a late event occurring for a significant time after DNA synthesis (35,36). In tissue culture, methylation occurs throughout the cell cycle with the maximum rate ranging from late S phase through G2, prior to mitosis (17,37,38). During the first few days of life, the rat brain develops rapidly. There are many mitotic cells around the ependymal wall of the ventricle and in the subpial germinal zone of the cerebellar cortex, the external granular layer (39)(40)(41).
Methylation of the lysyl residues in histones H3 and H4 within isolated nuclei was found to proceed stepwise, progressing from mono-to di-to trimethyllysine in histones H3 and from mono-to dimethyllysine in histone H4. From the kinetic data presented in Fig. 3, it can be concluded that the lower methylated derivatives of lysine serve as precursors for the higher methylated forms. Thomas et al. (38) came to similar conclusions while studying methylation of histones in Ehrlich ascites human cells. At saturation, the ratios of 3H-methyl in the different methyllysine forms approximated those which we had obtained in uivo (19). These was no significant difference in the rates of methylation of H3 and H4, if we take into account the number of sites. There was also no significant 1 J. A. Duerre, J. C. Wallwork, D. P. Quick, and K. M. Ford, unpublished data. difference in the K,, values of the enzyme for S-adenosyl-nmethionine with histone H3 and H4 as methyl acceptor, nor was there any significant difference in the Ki values for Sadenosyl-chomocysteine with the two histones as methyl acceptors.
The differences in the extent to which the lysyl residues in both histones H3 and H4 are methylated may be the result of more than one histone methyltransferase.
However, it is more probable that the extent to which these lysyl residues are methylated is conferred by the arrangement of the histones on the chromatin. Similarities in methylation sequences, -X-Arg-Lys-X-common to both histones H3 and H4 and -Ala-Arg-Lys-Ser-common to both sites in histone H3, have been observed by DeLange et al. (1, 2).
Loss of specificity of the enzyme for specific recognition sites on soluble histones is evident by the fact that free histones from the brains of old rats are as good methyl acceptors as the histones from young rat brains when incubated with soluble enzyme (34). Loss of specificity is also evident by the finding that free lysine-rich and slightly lysine-rich histones will accept methyl groups in a soluble system (34,(42)(43)(44). These histones do not accept methyl groups when bound to chromatin in the presence of soluble enzyme (44). The loss of specificity of the enzyme for specific recognition sites may be attributed to the loss in original steric confirmation of the histones during isolation, especially removal from DNA.