The Methylcytosine Dioxygenase Ten-Eleven Translocase-2 (tet2) Enables Elevated GnRH Gene Expression and Maintenance of Male Reproductive Function

Animals

Mice were housed under a 12-hour light, 12-hour dark cycle and had access to water and chow (Harlan/Teklad 2019) ad libitum. All experimental procedures adhered to guidelines provided in the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committee at the University of Wisconsin-Madison.

Cell culture

Quantitative real-time PCR

Total RNA was extracted from pelleted cells or POA-MBH tissue using TRIzol reagent (Invitrogen) according to the manufacturer's protocol. POA-MBH tissues were dissected using the following confines. Anterior, the anterior limit of the nucleus of the vertical limb of the diagonal band (bregma, +1.34 mm); posterior, the posterior limit of the mammillary nucleus (bregma, −3.40 mm); lateral, the external borders of the medial POA (rostral) and mammillary nucleus (caudal); and dorsal, 1.5 mm in depth. Two micrograms of RNA were reverse transcribed using the StrataScript first-strand synthesis system kit (Stratagene). Real-time quantitative PCR was performed in triplicate using SyberGreen MasterMix (Bio-Rad) and the ICycler quantitative PCR machine (Bio-Rad). 18S RNA was used as internal control for cDNA input; 18S critical threshold (Ct) values were stable within each analysis between conditions. Primer sets are listed in Supplemental Table 1. PCR efficiency was determined with a 10-fold serial dilution of cDNA as previously described (24). PCR conditions were optimized to generate more than 95% and less than 105% efficiency. Fold changes in relative gene expression were calculated by 2^−ΔCt or 2^−Δ(ΔCt) according to Livak and Schmittgen (25). Briefly, for each sample, a ΔC_t was calculated to normalize for the internal control (18S) using the equation: ΔC_t = C_t(gene) − C_t(18S). For comparisons without a specified CON sample for comparison, expression levels were calculated as 2^−ΔCt. To obtain differences between experimental and CON conditions (ie, GN11 cells), ΔΔC_t was calculated: ΔC_t(sample) − ΔC_t(GN11). Relative mRNA levels were then calculated using the equation fold difference = 2^−ΔΔCt, with GN11 relative expression (ΔC_t(GN11) − ΔC_t(GN11)) consequently set to 1.

Chromatin immunoprecipitation

Enrichment of tet2, H3K4me3, and H3K27me3 was measured at 4 regulatory regions of the mouse GnRH gene (previously described in Ref. 3) in GN11, GT1–7, GN11 cells overexpressing tet2, and GT1–7 cells with crispr/cas9-mediated tet2 gene disruption. Forty-eight hours after transfection, 4 cultures of each condition (confluent 150-mm plates) were treated with 1% formaldahyde for 10 minutes at room temperature. Cell lysis, sonication (4 × 30-s pulses on ice in a Fisher Scientific 550 Sonic Dismembranator set at level 4 to shear DNA into 300- to 600-bp segments), immunoprecipitation and DNA purification were carried out according to manufacturer protocol (MAGNA ChIP A/G; catalog number 17–10085; Millipore) Briefly, lysates were prepared by exposure of cells to ice cold lysis buffer for 20 minutes, followed by gentle disruption through shaking and repeated pipetting up and down. The antibodies were covalently cross linked to the magnetic beads with disuccinimidyl suberate, then cleared lysates were incubated with the beads while rolling at 4°C overnight. Beads were then captured and washed using a magnetic stand. Immunoprecipitated protein were then eluted and analyzed by Western blotting to determine enrichment of tet2. Eluted protein were sent to MS Bioworks for in gel digestion followed by mass spectrometry and bioinformatic analysis/identification of peptides measured in each sample. Tet2-binding partners were considered proteins with at least 4 unique peptides detected and no less than 3 times the unique peptides detected in IgG-precipitated eluates. Antibody information for all methods is presented in Table 1.

Coimmunoprecipitation

GT1–7 cells were cultured to full confluence in 150-mm dishes. Cells from 2 plates were combined for each immunoprecipitation; 2 independent immunoprecipitations were conducted for each antibody: tet2 (catalog number sc-136926; Santa Cruz Biotechnology, Inc) and normal rabbit IgG (catalog number 12–370; Millipore). Co-IP were performed using the Pierce Crosslink Magnetic IP/Co-IP kit (catalog number 88805) according to manufacturer recommendations. Two micrograms of each antibody (tet2 [sc-136926; Santa Cruz Biotechnology, Inc], H3K4me3 [07–473; Millipore], H3K27me3 [07–449; Millipore], and normal rabbit IgG [12–370; Millipore]) were used for each overnight immunoprecipitation at 4°C. To measure fold enrichment of each histone modification and tet2 binding, quantitative PCR was performed with 2 ng of purified DNA, ABI SYBR green PCR master mix (catalog number 4309155; ABI) and primers listed in Supplemental Table 1. Reactions were run in triplicate using an ABI 7300 system 2 minutes at 50°C, 10 minutes at 94°C then 45 cycles of 15 seconds at 94°C then 1 minute at 60°C followed by a melt curve analysis. Average C_t values were calculated for each primer set and fold enrichment was calculated relative to IgG (2 (average IgG C_t − average-specific antibody C_t), data are presented as mean of 4-fold enrichment values ± SEM.

Generation of GnRH-cre/tet2 ^lox/lox (gTKO) mice

To specifically ablate tet2 activity in GnRH neurons, a mouse strain was used in which exon 3 of the tet2 gene, which encodes greater than half of the murine protein and is included in the 2 described tet2 transcripts, is flanked by loxP sites (tet2^lox/lox). Deletion of exon 3 using this model results in a lack of tet2 transcript in Cre-expressing cells/tissues (26). These mixed B6/129S background mice were backcrossed with a pure C57BL/6J background strain. For GnRH neuron-specific tet2 ablation, GnRHcre male mice (27) were crossed with female tet2^lox/lox mice, and then the resultant GnRH cre⁺/tet2^lox/wt males were crossed with tet2^lox/lox female mice. The F2 progeny included GnRHcre⁺/tet2^lox/lox (gTKO) mice and GnRHcre⁻/tet2^lox/lox mice. The latter were used as CON animals. Genotyping. Toe samples were clipped from mice 1–2 weeks after birth and subsequently incubated in lysis buffer containing proteinase K at 95°C for 3 hours followed by the addition of KCl and centrifugation to pellet cell/tissue debris. The DNA pellet was washed in 70% ethanol and repelleted. DNA pellets dissolved in water were used for PCR analysis using standard conditions. Primer sequences are listed in Supplemental Table 1.

Immunohistochemistry

Perifusion experiments

Methods have been previously described in detail (28). Briefly, cultures containing GnRH cells were mounted in Sykes-Moore chambers as described by Martínez de la Escalera et al (29), ie, 2 coverslips were placed, face to face, separated by a rubber O-ring, forming a chamber with a volume of 200 μL. Cells were then perifused with artificial cerebrospinal fluid containing 0.1% glucose (pH 7.4) under 95% O₂ and 5% CO₂ at 37°C. Perifusates were collected at 25 μL/min in 20-minute fractions for 6 hours using the ACUSYST (Endotronics) perifusion system. To test the viability of cells, 56-mm K⁺ was perifused for 20 minutes during the last hour (h 6) of the experiment. Samples were stored at −80°C until they were assayed for GnRH by RIA.

Radioimmunoassays

Western immunoblotting

Cells were grown to 90% confluency in tissue culture flasks. Cell lysates were prepared by sonication and collected after centrifugation of cellular debris. Protein concentrations were measured using Pierce bicinchoninic acid assay kit (catalog number 23227). Proteins were separated using a 4%–15% SDS-PAGE gel (Bio-Rad) after equal loading of 15 μg of protein per well. Proteins were transferred to PVDF membrane-Immobilin-FL (Millipore) and probed overnight at 4°C with the primary antibody rabbit polyclonal tet2 (S-13) (1:100 dilution, catalog number sc-136926; Santa Cruz Biotechnology, Inc; tet2 detected at molecular weights 133 and 131) or mouse anti-β-actin antibody (1:2000; Sigma-Aldrich) after completion of the initial tet2 western followed by membrane stripping. After 4 washes with 1× PBS+0.1% Tween 20, the membranes were treated with secondary antibody IRDye 800CW goat antirabbit (1:5000) or IRDye680LT goat antimouse (1:10 000) for 1 hour at room temperature with gentle shaking. The blots were analyzed using the LI-COR Odyssey Infrared Imaging system. Quantification of the relative band densities (tet2/β-actin) was completed using ImageJ on 2 separate blots.

Vaginal opening (V.O.) and estrous cyclicity

Beginning at 20 days of age, females were examined for the onset of V.O. To examine the possible effects of genotype on estrous cyclicity, vaginal lavages from female mice (beginning at V.O. through approximately 55 d of age) were obtained and viewed under a microscope daily (9–10 am) for at least 21 consecutive days. Estrous stages were defined based on the primary cell types present in vaginal cytology samples as follows: leukocytic (L), metestrus/diestrus; nucleated (N), proestrus; and cornified (C), estrus. First estrus was determined as the first complete cycle of the pattern proestrus-estrus-2 consecutive days of metestrus/diestrus, as described previously (see reference 43 below).

Body weight measurements

Body weights were recorded daily (9–11 am).

Reproductive function (fecundity)

At p60, each mouse was paired with 1 CON animal of the opposite sex of proven fertility. When females were visibly pregnant, animals were separated and the number of days from initial pairing to the birth of a litter was recorded. The total number of pups per litter was also recorded.

Data analyses and statistics

Data are presented as the mean ± SEM. One-way ANOVA followed by Tukey post hoc was used to determine statistical significance related to tet2 overexpression and knockdown and GN11 cell stimulation (see figures 3 and 4 below). Two-way ANOVA followed by Tukey post hoc tests were used to determine statistical significance in chromatin-immunoprecipitation (ChIP) experiments and mouse hypothalamic mRNA content. Difference in the colocalization rate in CON and gTKO brains was tested with χ² test. Two-tailed, unpaired t tests were used to determine statistical significance for the remaining comparisons. Differences were considered significant when P < .05. Statistics were performed using GraphPad Prism software, version 6.0e.

Developmental expression of tet enzymes in vitro

To examine whether the process of GnRH neuronal differentiation requires active DNA demethylation, we hypothesized that tet enzyme expression or activity increases during the course of development. We first evaluated relative tet1, tet2, and tet3 gene expression in 2 GnRH-secreting cell lines; GN11 and GT1–7 cells. This comparison capitalizes on the distinct stages of development between these cells. GN11 cells, originally isolated from the mouse nasal region (31), express GnRH at very low levels and represent GnRH neurons at the earliest stages of differentiation, whereas GT1–7 cells, which were isolated from the adult mouse hypothalamus (32), are characterized by mature activity patterns, including elevated GnRH mRNA levels. We found that tet1 is expressed in both cell lines, with a trend toward higher expression in GT1–7 cells (n = 3 culture for each cell line; GN11, 0.536 ± 0.077 vs GT1–7, 0.858 ± 0.126; P = .09) (Figure 1). Tet3 is also expressed in both cell lines with no difference in expression (n = 3 cultures for each cell line; GN11, 0.203 ± 0.046 vs GT1–7, 2.15 ± 0.050) (Figure 1). Tet2 expression was scarcely detectable in the immature cells but strongly expressed in the mature cells (n = 3 cultures for each cell line; GN11, 0.020 ± 0.008 vs GT1–7, 0.648 ± 0.165; P = .02) (Figure 1).

Developmental expression of tet2 in vivo

Using quantitative PCR and total RNA isolated from POA-MBH tissues, we also found that tet2 expression increases in male and female mice between postnatal day (p)15 and p25 (n = 5 for each sex and time point; male p15, 0.442 ± 0.048 vs p25, 0.822 ± 0.10; t test, P = .007; and female p15, 0.190 ± 0.03 vs p25, 1.117 ± 0.39; t test, P = .047) (Figure 2). Sex differences in tet2 expression were not observed (two-way ANOVA, sex F_1,16 = 0.0112, P = .9171; age F_1,16 = 10.33, P = .0054, Tukey post hoc test at α = 0.05: p15 female < p25 female).

Tet2 activity in GnRH neurons in vitro

We reasoned that if GnRH gene expression is the consequence of the elevated levels of tet2 in GnRH cell lines, then GnRH mRNA levels should increase when tet2 is overexpressed in GN11 cells; conversely, GnRH mRNA levels might decline after disruption of tet2 expression in GT1–7 cells. We transfected GN11 cells with either a plasmid that encodes murine tet2 or a CON plasmid (GN11+tet2); GT1–7 cells were either cotransfected with a plasmid that encoded tet2 gene-specific guide RNAs and cas9 and a plasmid encoding selection genes (GT1–7(−)tet2) or with the CON CRISPR/Cas9 plasmid alone (CON GT1–7). Tet2 mRNA expression was significantly higher in tet2-transfected GN11 and GT1–7 cells compared with CON-transfected GN11 and GT1–7(−)tet2 cells (n = 3 cultures for each condition, GN11: 1, GT1–7: 11.21 ± 1.8, GT1–7(−)tet2: 0.515 ± 0.10, GN11+tet2: 15.95 ± 3.19; ANOVA F_3,8 = 17.37, P = .0007; Tukey post hoc at α = 0.05: GT1–7 and GN11+tet2 > GN11, GT1–7 and GN11+tet2 > GT1–7(−)tet2, GT1–7 = GN11+tet, GN11 = GT1–7(−)tet2) (Figure 3A). Expression differences were confirmed by Western blotting (Figure 3A, inset) followed by quantification of tet2/β-actin band intensity ratios using ImageJ (n = 2 blots, GN11: 0.193 ± 0.02, GT1–7: 0.490 ± 0.04, GT1–7(−)tet2: 0.205 ± 0.06, GN11+tet2: 0.563 ± 0.06; ANOVA F_3,4 = 15.67, P = .011; Tukey post hoc at α = 0.05: GT1–7 and GN11+tet2 > GN11, GT1–7 and GN11+tet2 > GT1–7(−)tet2, GT1–7 = GN11+tet, GN11 = GT1–7(−)tet2). GnRH mRNA levels rose after tet2 overexpression in GN11 cells (n = 3 cultures for each condition, GN11: 0.110 ± 0.02, GN11+tet2: 1.03 ± 0.29; t test, P = .034) (Figure 3B), whereas CRISPR/cas9-mediated disruption of GT1–7 cell tet2 expression caused a significant decrease in GnRH mRNA levels (n = 4 cultures per condition, GT1–7: 3.11 ± 0.069, GT1–7(−)tet2: 0.79 ± 0.12; t test: P < .001) (Figure 3C), indicating that tet2 not only enables GnRH expression in immature cells but also has substantial influence over the maintenance of GnRH mRNA level in mature cells. Notably, using a perifusion method we found that the rise in GN11 cell GnRH mRNA after tet2 overexpression had an apparent effect on peptide release, as GN11+tet2 cells released more GnRH peptide compared with CON-transfected cells (n = 3 cultures per condition; GN11: 2.10 ± 0.26 vs GN11+tet2: 2.67 ± 0.12 pg/mL per 10 minutes (P = .043) (Figure 3D).

Tet2 binding and histone modifications in GnRH cell lines

In addition to, and possibly as a consequence of manipulating DNA methylation dynamics, tet2 promotes H3K4me3 (see reference 56 below), an activating histone modification. Consequently, using chromatin immunoprecipitation followed by quantitative PCR, we measured levels of H3K4me3, H3K27me3 (a restrictive histone modification), and tet2 binding across 3 enhancer regions and the promoter of the GnRH gene (Figure 4A) in GN11, GT1–7, GN11+tet2, and GT1–7(−)tet2 (n = 4 cultures for each cell line). Tet2 binding was low in GN11 cells but significantly higher in GT1–7 and GN11+tet2 cells, where binding was most pronounced at enhancer E1 and the promoter; tet2 binding significantly declined at those regions in GT1–7(−)tet2 cells (two-way ANOVA, cell line F_3,55 = 39.44, P < .001; gene region F_4,55 = 56.66, P < .001) (Figure 4B). H3K4me3 levels were also low in GN11 cells and higher in GT1–7 and GN11+tet2 cells. H3K4me3 was most pronounced at the promoter and to a lesser extent at E1 in GT1–7; H3K4me3 levels also increased in GN11 cells after tet2 overexpression at E1 and the promoter. On the other hand, GT1–7(−)tet2 cells exhibited significantly reduced H3K4me3 abundance at the promoter (two-way ANOVA, cell line F_3,55 = 18.91, P < .001; gene region F_4,45 = 32.53, P < .001) (Figure 4C). H3K27me3 levels were similar between the cell lines with some evidence of greater enrichment near E1 and the promoter (two-way ANOVA, cell line F_3,55 = 1.468, P = .1316; gene region F_4,45 = 7.775, P < .001) (Figure 4D).

Tet2 disruption in GnRH neurons

Using conditional gene targeting, we next generated mice with specific ablation of tet2 activity in GnRH (GnRH cre X tet2 fl/fl) expressing neurons. All comparisons were between transgenic animals (gTKO: GnRH-Cre positive, tet2 fl/fl) and CON littermates (GnRH-Cre negative, Tet2 fl/fl). Double immunohistochemical staining indicated that tet2 was observed in the cytosol of CON GnRH neurons but not in gTKO neurons (Figure 5). Importantly, adjacent non-GnRH cells had tet2 expression in both CON and gTKO animals. Quantitative analysis in gTKO and CON brains indicated significantly fewer tet2-positive GnRH neurons in gTKO mice compared with CON mice in both sexes (P < .0001) (Table 2). We then evaluated tet2 and GnRH mRNA levels in the MBH/POA of male and female animals at p15, p25, and p100. Tet2 expression increased with age in both sexes across genotypes (female: two-way ANOVA; age, F_2,22 = 18.33, P = .001; Tukey post hoc test at α = 0.05, p15 CON < p25 CON and p15 kTKO < p25 CON; male: two-way ANOVA, age F_2,21 = 3.620, P = .045) (Figure 6, A and B). Genotype had no impact on tet2 expression in either sex (female: two-way ANOVA; age, F_2,33 = 0.466, P = .632; male: two-way ANOVA; age, F_2,31 = 0.1595, P = .853) (Figure 6, A and B), likely due to the maintenance of tet2 expression in non-GnRH expressing cells. An effect of genotype was evident for GnRH expression in male gTKO animals. GnRH expression was lower in gTKO males compared with CON littermates (male: two-way ANOVA; genotype, F_1,23 = 6.798, P = .0157; Tukey post hoc test at α = 0.05, p15, p25, and p100 CON > p100 gTKO; female: two-way ANOVA; genotype, F_1,21 = 0.3569, P = .5566) (Figure 6, C and D), with a trend toward declining GnRH expression with age across genotypes for both sexes (male: two-way ANOVA; age, F_2,23 = 3.152, P = .0617; female: two-way ANOVA; age, F_2,21 = 3.308, P = .0564) (Figure 6, C and D).

Role of Tet2 in GnRH neurons in vivo

If tet2 is necessary for the maturation of GnRH neurons, then disruption of tet2 activity in GnRH neurons should influence reproductive function. We measured body growth, plasma LH (an indirect measure of GnRH neuronal activity) at p30 and p60, and early adulthood fecundity in both sexes. The day of preputial separation was evaluated in males and V.O., vaginal smears for 28 days after V.O., and a first estrus were determined for females. All comparisons were between transgenic (gTKO; GnRH-Cre positive, Tet2 fl/fl) animals and CON littermates (GnRH-Cre negative, Tet2 fl/fl). We did not find any indication of abnormal timing or progression of puberty in either sex. The average age of V.O. (CON [n = 5]: 29.0 ± 0.3, gTKO [n = 7]: 29.6 ± 0.2 days of age; t test, P = .11) (Figure 7A) and first estrus (CON [n = 5]: 34.6 ± 1.03, gTKO [n = 7]: 36.3 ± 1.56 days of age; t test, P = .43) (Figure 7A) were similar and there were no differences in percentage of time spent in each estrus stage between CON and gTKO animals (Figure 7B). Preputial separation was also similar between gTKO animals and CON littermates (CON [n = 4]: 33.0 ± 0.71, gTKO [n = 3]: 33.7 ± 1.85; t test, P = .72) (Figure 7A). Plasma LH levels exhibited typical developmental patterns and were similar for both genotypes and both sexes at p30 (female CON [n = 4]: 0.933 ± 0.24, gTKO [n = 4]: 1.54 ± 0.82 ng/mL; male CON [n = 5]: 1.90 ± 0.33, gTKO [n = 4]: 0.94 ± 0.25) (Figure 7C). There was no difference between CON and gTKO female plasma LH levels at p60 (CON [n = 3]: 0.268 ± 0.10, gTKO [n = 4]: 0.159 ± 0.014 ng/mL) (Figure 7C) or young adult female fecundity (CON [n = 5]: 23 ± 0.32 vs gTKO [n = 5]: 23.4 ± 0.45 days between pairing and birth of litter; CON [n = 5]: 7.2 ± 0.37 vs gTKO [n = 5]: 7.0 ± 0.80 pups per litter) (Figure 7D).

In contrast, in young adult gTKO males, plasma LH levels were significantly lower (male CON [n = 4]: 0.338 ± 0.095, gTKO [n = 4]: 0.101 ± 0.03 ng/mL; t test, P = .04) (Figure 7C) and young adult male gTKO fecundity was compromised, with gTKO males taking longer than CON littermates to impregnate a fertile female (CON [n = 4]: 23 ± 0.45 vs gTKO [n = 5]: 25.8 ± 0.85 d between pairing and birth of litter; t test, P = .034) (Figure 7D). There was no difference in number of pups per litter sired by CON or gTKO animals (CON [n = 4]: 7.60 ± 0.51 vs gTKO [n = 5]: 7.75 ± 0.25) (Figure 7D). There were also no differences in body weight observed between genotypes within sex through p100 (data not shown).

Influence of BPA on Tet2 subcellular localization and GnRH gene chromatin environment

Because in vivo data suggests that tet2 is necessary for the maintenance of GnRH neuron activity, at least in males, then factors responsible for a decline in GnRH mRNA might be the result of altered tet2 activity. BPA is an environmental contaminant shown to influence sperm quality and GnRH mRNA levels in adult male rats (33, 34). We evaluated the impact of BPA on tet2 subcellular localization and the GnRH gene chromatin environment in GT1–7 cells. After 2 hours of BPA exposure (10nM), tet2 exhibited a dramatic shift in subcellular localization (Figure 8A). Specifically, whereas tet2 immunoreactivity was largely distributed in both the nucleus and cytoplasm of vehicle treated cells; it dramatically shifted to primarily the cytoplasm during BPA exposure (Figure 8B). This shift was associated with modifications to the GnRH gene, particularly H3K4me3 abundance (Figure 8B). Tet2 binding trended toward a decrease in promoter region binding after 2 hours of BPA (10nM) exposure (two-way ANOVA; BPA vs vehicle exposure, F_1,16 = 5.055, P = .039; gene region, F_3,16 = 63.71, P < .001; interaction, F_3,16 = 2.594, P = .0886) (Figure 8B). This was accompanied by a significant reduction in promoter H3K4me3 abundance (two-way ANOVA; BPA vs vehicle, F_1,16 = 31.68, P < .001; gene region, F_3,16 = 87.25, P < .001; interaction, F_3,16 = 19.31, P < .001) (Figure 8B). There was no significant effect of this BPA exposure on H3K27me3 abundance at the regions evaluated. GnRH mRNA levels did not change during this brief 2-hour period of BPA exposure, although they were substantially lower after 12 hours of sustained BPA exposure (relative GnRH expression in GT1–7 cells by qPCR (2^−ΔCT × 10⁻⁵) vehicle: 3.72 ± 0.09, BPA 10nM 2 h: 3.21 ± 0.19, BPA 10nM 12 h: 1.95 ± 0.14; n = 3 for each culture condition, ANOVA F_2,6 = 39.03, P < .001 Tukey post hoc: vehicle = BPA 2 h, vehicle > BPA 12 h, BPA 2 h > BPA 12 h).

Tet2 protein interactions in GnRH neurons

The shift in tet2 subcellular localization during BPA exposure implicates the presence of a tet2 chaperone and/or cytoplasmic tet2 target in GnRH neurons. Although tet2 interactions have been reported for other cellular systems (23, 35, 36), there is presently no report of neuronal tet2-interacting proteins. Using mass spectrometry after coimmunoprecipitation, we identified novel tet2 interactions, which isolated the SET complex as likely responsible for H3K4me3 deposition at the GnRH promoter, as well as cytoplasmic targets potentially related to the BPA-driven shift in subcellular localization. The primary interactions, shown in Figure 9, suggest 4 potential roles of tet2 in GnRH neurons, including chromatin remodeling/histone modification, DNA repair, mitochondrial function, and RNA metabolism.