Dietary and genetic effects on age-related loss of gene silencing reveal epigenetic plasticity of chromatin repression during aging

Gene silencing is lost with age in repressive heterochromatin regions of the adult fly

To test the hypothesis that loss of repressive heterochromatin marks with age leads to a loss of gene silencing in regions associated with repressive heterochromatin, we utilized a novel PEV reporter containing a heat shock-inducible HS-lacZ gene that allows for visualization of gene expression at the level of single cells, in individual tissues, at any time during the life of the adult fly [27, 28]. We examined two lines where the PEV transposon, containing a heat shock inducible HS-lacZ gene and mini-white gene, is located in association with one of two different heterochromatin regions. The In(3L)BL1 line (BL1) is associated with 3L heterochromatin, while the Tp(3;Y)BL2 (BL2) line transposon region has been translocated to the heterochromatic Y chromosome [27, 28]. Both lines demonstrate the PEV phenotype of mosaic eye pigmentation, caused by partial repression of the mini-white gene, and mosaic lacZ expression from PEV during developmental stages [27, 28]. As a control, we used the original P[w+ HS-lacZ] (65E) line from which the two PEV lines were derived, where the transposon is in the 3L euchromatin region [27, 28].

Upon a heat shock and recovery period (37°C for 1 hour, 25°C for 1 hour), flies of the control line, where the reporter construct is in normal euchromatin, were universally stained blue in all major tissues when reacted with X-gal to reveal lacZ expression, while the two PEV lines showed mosaic staining patterns, suggesting stochastic induction of HS-lacZ in individual cells by heat shock (Fig 1, S1).

We selected three distinctly different tissue types with ready accessibility to X-gal in whole mount preparations, and sufficient optical clarity to allow us to count individual stained and unstained cells in whole mount preparations. These included: Malpighian tubules (renal system), oenocytes (nutrient storage, pheromones) and the cells at the root of the major dorsal thoracic bristles (neuronal sensory and bristle support cells). Malpighian tubules, with their tubular, linearly aligned arrangement of cells, permit ready counting of individual lacZ-expressing cells as a percent of all cells present in the tubule, when these are visualized by a DAPI nuclear stain (Fig. 1, S1). Malpighian tubules have the additional distinction of being one of the few somatic tissues in the adult fly for which there are active stem cells and replacement of Malpighian tubule cells throughout adult life [29]. Oenocytes, with their linear and segmentally arranged distribution in the fly abdomen, can be easily counted with X-gal staining, as individually lacZ positive or negative, directly through the abdominal cuticle (Fig. 1, S1). The 10 major thorax bristles selected for counting (two pairs of dorsocentral setae, two pairs of postalar setae, and one pair of anterior scutellar setae) permit direct visualization of the cells at the root of their bristles, through the cuticle of the dorsal thorax.

The level of cell-specific gene silencing with age was quantified by counting the number of lacZ-expressing cells as a percentage of total cells in each specific tissue. After heat shock, young 10-day old flies ranged from 5-35% of cells positive for lacZ expression in the Malpighian tubule, oenocytes, or thoracic bristle roots in the two different PEV lines, while similarly-treated 50-day old flies show between 40-80% lacZ expression in Malpighian tubule or oenocytes for BL1 and BL2 (Fig. 2). At present, the thoracic bristle cells can only be evaluated up to day 30, as visualization of lacZ expression in these cells is obscured by the age-associated increase in expression of the HS-lacZ reporter in the underlying thoracic muscles that is seen by day 30. Together, these data document the dynamic change in age-related loss of gene silencing in two different heterochromatin regions, 3L and Y heterochromatin.

Basal level of activity of HS-lacZ is unaffected by age in Malpighian tubules and oenocytes

Our hypothesis is that the observed age-related loss of gene silencing is due to an age-related decrease in repressive heterochromatin in the region of the PEV reporters. Alternatively, an increase in the percent of cells expressing lacZ with age could result from a change in the basal level of activity of the HS promoter with age. We ruled out the possibility that the increase in percent of cells expressing lacZ with age seen in the PEV lines is the result of a change in the basal level of activity of the HS promoter with age by showing that without heat shock, lacZ expression in the BL1 and BL2 lines showed a low level of expression, ~10%, that did not change with age in these tissues (Fig. 2).

Heat shock response of the HS-lacZ promoter is not age-sensitive

To examine the possibility that the increase in lacZ expression with age could be caused by an age-related increased sensitivity of the HS promoter to heat shock, we examined the effect of age on the euchromatin control line, in which the reporter construct is free of the confounding effects of heterochromatin repression. We found no significant difference in lacZ staining with age in Malpighian tubules or oenocytes after a brief heat shock for the euchromatin control line (Fig. 3).

Taken together, these results suggest that the observed increase in lacZ expression with age in the BL1 and BL2 PEV lines is most likely due to an age-related de-repression of heterochromatin in the region of the reporter gene, and not to an increase in lacZ expression from an age-related increase in the basal activity of the HS-lacZ reporter, or to a direct effect of age on the sensitivity of heat shock-inducible promoter.

Loss of gene silencing of lacZ and mini-white increases globally with age

In addition to determining loss of gene silencing at the level of single cells in specific tissues of individual flies, we examined gene silencing with age globally, by measuring mRNA expression from the HS-lacZ gene and the mini-white gene that make up the PEV reporter construct [27, 28]. We found a two-fold increase in HS-lacZ (after heat shock) and mini-white (without heat shock) mRNA expression with age, demonstrating that for two different promoters, HS-lacZ and white minigene, both repressed by heterochromatin, there is a loss of silencing with age (Fig. 4). There is no change in level of mRNA expression for the HS-lacZ (after heat shock) or white minigene (without heat shock) when the reporter construct is in euchromatin (Fig 4). The finding of a two-fold increase in HS-lacZ and white minigene mRNA expression with age for the heterochromatin-positioned reporter further shows that in a variety of tissue types, there is a loss of gene silencing with age in heterochromatin regions.

Loss of repressive heterochromatin marks is associated with loss of gene silencing with age

To determine whether the loss of gene silencing seen with age results from a decrease in heterochromatin repression in the region of the PEV construct, we measured relative enrichment of the repressive heterochromatin marks HP1a and H3K9me2 in the region of the PEV reporters, using chromatin immunoprecipitation and qPCR (ChIP-qPCR). We found a significant reduction in the enrichment of HP1a and H3K9me2 with age over the entire PEV construct (Fig. 5). Reduction in repressive heterochromatin marks over the PEV reporter was the same with or without heat shock, demonstrating that the heat shock used for inducing lacZ expression does not itself alter repressive heterochromatin marks in this region. Although it has been reported that heat shock leads to recruitment of HP1a to the native hsp70 gene regions [30], we saw no further increased enrichment of HP1a in the HS-lacZ reporter region with heat shock. These data support the hypothesis that loss of gene silencing of the lacZ and white minigene reporters is likely due to a decrease in repressive heterochromatin with age.

Increased expression of the histone deacetylase, dSir2, opposes age-related loss of gene silencing

In yeast, the normal age-related reduction in expression of the histone deacetylase Sir2 results in loss of gene silencing during aging that can be rescued by increasing SIR2 expression [3]. To examine the effect of increased dSir2 expression on the age-related loss of gene silencing in Drosophila, we measured and compared loss of gene silencing with age in our PEV reporter line containing the normal two copies of dSir2, and in flies of the same PEV line, but with an added copy of a dSir2 transgene [31], which increased dSir2 expression (measured by qPCR) level to approximately 150% of control. Elevated dSir2 expression led to a statistically significant decrease in lacZ-expressing cells by day 20 in oenocytes (Fig. 6). Thus, increasing the level of expression of a known histone deacetylase enzyme, dSir2, may delay age-related loss of gene silencing in flies as it does in yeast [3].

Calorie restriction delays the age-related loss of gene silencing

To further investigate the relationship between aging and loss of gene silencing, we examined the effect of the life span-extending intervention CR on age-associated loss of gene silencing by comparing the loss of gene silencing with age in Malpighian tubules, oenocytes, and thorax bristle root cells for flies grown on high or on low calorie food (Fig. 7). At an early age, flies on either high or low calorie food showed the same percent of stained cells (Fig. 7). CR resulted in significantly lower levels of age-related HS-lacZ loss of gene silencing from ages 20 through 40 days for Malpighian tubules and ages 10 through 30 days for oenocytes, after which time, the maximal level of lacZ staining for the PEV line in each of these tissues is reached. The age of increase in loss of gene silencing is delayed under low calorie conditions, as indicated by the finding of a statistically significant increase in loss of gene silencing in oenocytes on high calorie food from day 2 to day 10 (P<0.001) and from day 10 to day 20 (P<0.001), as well as a statistically significant increase in loss of gene silencing in Malpighian tubules on high calorie food from day 10 to day 20 (P<0.05) that is not seen until a later time on low calorie food (Fig. 7). These analyses indicate that CR delays the age-related loss of gene silencing, providing strong evidence for the hypothesis that decreases in repressive heterochromatin and loss of gene silencing are a component of the aging process.

Diet switch experiments demonstrate the plasticity of the adult chromatin state

The finding that CR delays age-related loss of gene silencing suggests a relationship between CR life span extension and maintenance of repressive heterochromatin. However, it does not distinguish between CR having a direct effect on maintenance of repressive heterochromatin, or an indirect effect, by slowing down the rate of aging. To test whether CR may directly mediate repressive heterochromatin maintenance, we measured the level of gene silencing in flies switched from high calorie food to low calorie food and vice versa (diet switch experiments), changes known to rapidly change mortality rates [32],. We found that within 72 hours after either 20 or 30 day-old HS-lacZ PEV reporter flies (BL2) are switched from a low to a high calorie food, or from a high to low calorie food, the percent of cells showing lacZ expression in tissues of these “switched” flies changes toward the level of gene silencing for flies that have been kept on the “switched-to” diet throughout their lives.

To illustrate the change following diet switch for each of the eight biological replicates, we plotted the percent of lacZ expression in oenocytes for flies switched from low calorie to high calorie food, or from high calorie to low calorie food, as a percent of the total difference between the lacZ expression value for flies living continuously on high calorie food (1.0), and the value for flies living continuously on low calorie food (0), for each of the biological replicates (Fig. 8). We performed two-sided chi square analyses on the individual replicates as well as on the pooled data, using the raw counts of the number of positive cells and negative cells within the context of a binomial comparative trial model to determine whether the change in the level of gene silencing for the switched flies is in the same direction as the level of gene silencing for flies fed continuously on the diet to which they were switched. Analysis of the pooled data showed a statistically highly significant change away from the level of gene silencing of the food they had been living on (P<0.005) and toward the level of gene silencing they were moved to. Analysis of the individual replicates showed that for those flies moved from high calorie to low calorie food, 6 of the 8 replicates exhibit a statistically significant change in gene expression away from the level of gene expression of flies on high calorie food (P≤0.02) while 6 of the 8 replicates moved from low calorie to high calorie food exhibited a statistically significant change in gene expression away from the level of gene expression of flies on low calorie food (P≤0.05). The finding that flies at 20 or 30 days of age can rapidly (within 72 hours) respond to a change in diet by altering the level of gene silencing of a PEV reporter gene, a direct functional measure of the state of repressive heterochromatin, suggests a close relationship between the life span-extending intervention CR, and maintenance of repressive heterochromatin. Furthermore, the rapidity of the response to diet seen in the change in gene silencing demonstrates the epigenetic plasticity of chromatin during aging and the potential role of diet and metabolism in chromatin maintenance.

Fly lines and medium

The two PEV lines, In(3L)BL1 (BL1) and Tp(3;Y)BL2 (BL2), and the euchromatin control line P[w+ HS-lacZ] (65E) (Eu) were a gift from Joel C. Eissenberg and are described in [27, 28]. The Sir2 UAS-Sir2-4 transgenic line was backcrossed to a w¹¹¹⁸ control line for 10 generations [31]. Flies were grown and kept at 25 °C on standard cornmeal food (approximately 12% sucrose, 3% yeast, 5% corn meal, and 1% agar). For studies on dietary restriction, after eclosion, flies were grown on either high calorie 1.5N (1.5N: 15% sucrose, 15% yeast, 5% corn meal, 2% agar), or low calorie (0.5N: 5% sucrose, 5% yeast, 5% corn, 2% agar) food, similar to [38].

Quantification of HS-lacZ expression

Flies were collected and heat shocked [1h at 37 °C, followed by 1h recovery at 25 °C for Malpighian tubules, 40 min heat shock at 37 °C and 40 min recovery for oenocytes, and 10 min heat shock at 37 °C and 30 min recovery for thorax bristle roots], fixed with 4% paraformaldehyde in PBS for 10 min, washed in PBS, and incubated in X-gal staining solution [39] at room temperature overnight. Presence of blue staining of a cell represented a positive lacZ-expressing cell. Different levels of intensity of blue staining were not determined, only presence or absence of blue staining was noted.

Percent of stained Malpighian tubule cells was determined as fraction of stained cells out of total nuclei, visualized by DAPI. Because of the possibility that new cells, migrating from stem cells in the proximal region of the Malpighian tubule [29], could confound the age-related changes in loss of gene silencing, proximal - distal position of lacZ positive and negative cells in the Malpighian tubules was determined. We found no evidence of a positional bias for either lacZ expression or its absence with age, minimizing the possibility that clonality in lacZ expression or distribution related to proximal or distal origin of the cells skewed our results.

The percent of stained oenocytes was determined from days 20-50 by counting lacZ-stained and total brown-pigmented oenocytes in the six dorsal-abdominal segments. Brown-pigmentation is seen by day 20 in BL1, BL2 and control flies. The total number of oenocytes days 2-10 was shown to be the same as days 20-50 by counting the number of lacZ-expressing oenocytes in the control line after heat shock days 2-50 (no significant difference was seen with age) and comparing with the number of brown-pigmented oenocytes in unstained cohorts of control, BL1 and BL2 flies at days 20-50 (no significant differences with age or line). Therefore, the percent lacZ staining for the Day 2-10 was determined by counting the number of lacZ-stained oenocytes in the six dorsal-abdominal segments and comparing them to a “standard” number of oenocytes in each segment: 1=19, 2=22, 3=24, 4=24, 5=20, and 6=18 cells +/− 1.5 cells/segment.

The percent of stained thoracic bristle roots was determined by counting the number of stained thoracic bristle roots of 10 examined (two pairs of dorsocentral setae, two pairs of postalar setae, and one pair of anterior scutellar setae), by direct visualization of the cells at the root of their bristles, through the cuticle of the dorsal thorax.

Chromatin immunoprecipitation (ChIP) and qPCR

For each chromatin sample, 100 male flies were homogenized in PBS, and cross-linked in 1.8% formaldehyde at RT for 10 min. Cross-linking was stopped by addition of glycine (final concentration 0.25M) for 5 min. Subsequent washing, sonication and chromatin immunoprecipitation followed the protocol described in [23]. Anti-HP1a antibody C1A9 was obtained from the Developmental Studies Hybridoma Bank; anti-H3K9me2 antibody ab1220 was purchased from Abcam. qPCR primers were designed for five unique sites spanning the construct length, as well as for the two control sites in euchromatin regions (2L: 22341800 and 3L: 13168600), which were shown to be free of heterochromatin marks in both young and old flies [23].

Primers: P5-F: TACTTCGGTAAGCTTCGGCTATCG; P5-R: CGTGTCACACTAGAATCTTCA; HZ-F: GAAGAACTCACACACAATGCC; HZ-R: AAGGGGATTAAGTTGGGTAACGCC; Z-F: TGCGGGACGCGCGAATTGAATTAT; Z-R: AGATGGCGATGGCTGGTTTCCAT; mW-F: GGTAGCTGTGCTCGCTATATAAGACA; mW-R: ACGAAGTATCCTACGAAGTAGGTTTA; P3-F: ACGTTAAGTGGATGTCTCTTGCCG; P3-R: GCAGGTTGCCCAAGAAGATTCGAC; CTRL1-F: GCCCTGGTGGAAGTGGAAATTGAA; CTRL1-R: GGCTCAGAAACAGAAGCAGGAACA; CTRL2-F: TAGAAAGATCTGAGGAGAGTCCGCCT; CTRL2-R: TAACATGGCCCTCGGCTACGATTT. CTRL1 is at chromosome position: 3L: 22341700 and CTRL2 is at chromosome position: 2L: 13168600. Enrichment for target sites and euchromatic control sites in ChIP DNA relative to input was determined by qPCR (SYBR Green PCR Master Mix, Applied Biosystems). The fold enrichment was derived from ΔΔCt between each target site and the average of the two control sites.

RNA isolation and RT-PCR

mRNA was isolated from whole flies using Dynabeads mRNA DIRECT Kit (Invitrogen). cDNA was synthesized using iScript cDNA Synthesis Kit (Bio-Rad). HS-lacZ and mini-white mRNA levels were normalized to gapdh1. For the heat shock condition, flies were heat shocked at 37 °C for 30 minutes.

Primers: HS-lacZ-F: TGCGGGACGCGCGAATTGAA TTAT; HS-lacZ-R: AGATGGCGATGGCTGGTTTCC AT; mWhite-F: TTTGGGCCAACAACTCACGCAA; mWhite-R: CTCAGGTGAACACATTTATCGTGGC; Gapdh1-F: TTAACCTGGACGCCTACAGCC; Gapdh1-R: CCTCGACGATCTCGAAGTTGTCAT.