Heterochromatin Formation Promotes Longevity and Represses Ribosomal RNA Synthesis

1Organismal aging is accompanied by the accumulation of damage to DNA and other macromolecules, and a progressive decline in vitality and tissue function. The underlying mechanisms remain unclear, and many models have been proposed to explain the aging phenomenon. Prominent among these models is the “free radical theory of aging”, which posits that the gradual and collective damage done to biological macromolecules (including DNA and proteins) by reactive oxygen species (ROS) from intrinsic (e.g., metabolism) or extrinsic sources (e.g., radiation), is the major cause of organismal aging,. Other competing (although some are overlapping) models of aging include genetically programmed senescence,, heterochromatin loss, telomere shortening, genomic instability, nutritional intake and growth signaling –, to name a few. In the heterochromatin loss model of aging, Villeponteau (1997) has proposed that heterochromatin domains, which are set up early in embryogenesis, are gradually lost with aging, resulting in derepression of silenced genes and aberrant gene expression patterns associated with old age.

2Experimental tests of the role of heterochromatin formation in animal aging, however, have produced controversial results. On the one hand, cellular senescence is associated with an increase in localized heterochromatin formation in the form of Senescence-Associated Heterochromatin Foci (SAHFs), which are a hallmark of replicative senescence of aged cells in culture, and have also been found in the skin cells of aged animals –. On the other hand, it has been shown that premature aging diseases in human and animal models correlate with global heterochromatin loss –.

3Heterochromatin is important for chromosomal packaging and segregation, and is thus important for genome stability,. Indeed, it has been shown in Drosophila that heterochromatin is essential for maintaining the stability of repeated DNA sequences and of the rDNA locus in particular. Loss of heterochromatin causes disruption of nucleolar morphology and formation of extrachromosomal circular (ECC) DNA, which results from an increase in illegitimate recombination at the rDNA locus. Interestingly, disruption of heterochromatin and nucleolar structure, and the consequent increase in ECC DNA, have previously been shown to cause accelerated aging in yeast,. These reports suggest a positive role for heterochromatin formation in promoting longevity.

4To understand the role of heterochromatin in animal aging, and the underlying molecular mechanisms, we altered heterochromatin levels in Drosophila by genetically manipulating Heterochromatin Protein 1 (HP1) levels and JAK/STAT signaling, and assessed the effects on aging. Our results suggest that heterochromatin formation positively contributes to preventing premature aging and suppresses illegitimate recombination of the rDNA locus and unnecessary rRNA synthesis.

5To investigate whether heterochromatin levels are important for longevity, we examined the life spans of flies with reduced or increased levels of HP1. These flies exhibit reduced or increased levels of heterochromatin, respectively, during development, as HP1 is an integral component of heterochromatin and controls heterochromatin levels,. We found that reducing HP1 levels by half, as in Su(var)2055 heterozygotes, caused a dramatic shortening of life span compared to isogenic controls (p = 2.03−86) (Figure 1A). Similar results were found with a second allele, Su(var)2052 (Figure S1A). Conversely, a moderate overexpression of HP1, caused by basal activity of the hsp70 promoter, significantly extended life span, resulting in a 23% increase in median life span and a 12% increase in maximum life span (p = 6.31−24) (Figure 1A). Similarly, at non-heat shock conditions (25°C), a second (independent) line of hsp70-HP1 flies also lived significantly longer than their control flies (Figure S1B). At basal levels of transcription, hsp70-HP1/+ flies exhibited higher heterochromatin levels. By quantitative real-time polymerase chain reaction (qPCR) measurements, we found that these flies had approximately 20% higher HP1 mRNA expression than control (Figure S2A). Over-expression of HP1 at higher levels, such as under heat-shock inducible conditions, however, caused developmental abnormality or lethality to the animal. These results suggest that heterochromatin levels significantly influence life span, and moderately higher levels of heterochromatin promote longevity.

6Since both JAK overactivation and STAT loss reduce heterochromatin levels,, we investigated the effect of altering JAK/STAT signaling on aging. JAK/STAT signaling plays two roles: in the canonical pathway, JAK/STAT directly regulates target gene expression,, while in its non-canonical function, unphosphorylated STAT is essential for heterochromatin formation, and genome stability. In the canonical pathway, loss of STAT has effects equivalent to loss of JAK and opposite to JAK overactivation. However, in the non-canonical function, loss of STAT has the same effects as JAK overactivation, causing heterochromatin destabilization, and genome instability. We examined the life span of Stat92E+/− flies and those heterozygous for gain- or loss-of-function mutations of hop. We found that flies heterozygous for either the gain-of-function hopTum-l or the Stat92E mutation exhibited shortened lifespans compared with wild-type control flies (p = 8.87−23; 2.92−53, respectively), while flies heterozygous for a loss-of-function hop allele, hop3, had longer lifespans (p = 7.34−25) (Figure 1B). These results are consistent with the idea that heterochromatin levels influence lifespan.

7A previous study has shown that Drosophila life span was only slightly reduced in Su(var)205 heterozygotes and was not affected by a chromosomal duplication that encompasses the Su(var)205 locus and many other genes. Using chromosomal duplication, any effects associated with higher levels of Su(var)205 might be masked by higher levels of other neighboring genes. In studies with the loss-of-function heterozygotes, the authors in that study ensured isogenicity of their compared strains by extensively back-crossing the mutations into a common genetic background, and relied on the suppression of position-effect variegation (PEV) to determine the presence of the Su(var)205 mutation. PEV results from heterochromatin-mediated gene repression, commonly seen in loss of eye pigmentation. The presence or absence of the Su(var)205 mutation was assumed to correlate 100% with the PEV phenotype. However, we have found that this is not the case. We examined the PEV phenotype of wm4 in the progeny of single pairs of wm4; Su(var)2055/CyO and w flies, and found that the PEV phenotype was not 100% correlated with the Su(var)205 mutation (Figure S3), rather, there was less PEV than would be expected, suggesting that, with regard to suppression of PEV, either the Su(var)205 mutation is not completely penetrant or that there is an incomplete epigenetic reprogramming at the wm4 locus, or both. On the other hand, it has been shown that many Su(var) mutations exhibit maternal-effect suppression of PEV, such that the PEV phenotype can be modified regardless of inheritance of the Su(var) mutation. It has also been shown that HP1 mutations disrupt epigenetic reprogramming, causing transgenerational inheritance of epigenetic information,. Thus, in the aging study by Frankel and Rogina (2005), the presence or absence of the Su(var)2052 mutation in the test flies may not have been accurately determined. In our current studies, we confirmed the presence of Su(var)205 mutations in the coisogenic strains by both suppression of PEV and homozygous lethality (see Materials and Methods). We found that heterochromatin levels are essential for longevity using both gain- and loss-of-function strategies.

8To investigate the cause of altered life span in flies with different heterochromatin levels, we observed the behaviors of these flies by video-recording (see Methods). Video playbacks show that aged flies exhibited a gradual loss of mobility and eventually became immobile (Videos S1, S2, S3). By quantifying their mobility (see Methods), we found that, compared with wild-type controls, flies with reduced heterochromatin levels lost mobility much faster, and those with increased heterochromatin levels maintained their mobility for a longer period of time (Figure 1C; Videos S1, S2, S3).

9It has been shown in C. elegans and Drosophila that old animals die of sarcopenia (muscle degeneration) and impaired muscle function precedes aging, similar to the gradual loss of muscle function and frailty in aging humans. Since we found that heterochromatin levels influence Drosophila life span, and since the altered life span was associated with the animals' mobility, we investigated whether loss of heterochromatin is associated with muscle degeneration.

10We used whole-mount fluorescent immunostaining to examine the integrity of the large intestinal wall muscle, which can be visualized readily in adult flies of different ages after minimal dissection. The fly large intestinal wall muscles consist of longitudinal (thick) and circular (thin) muscle fibers (Figure 1D, top left). We found that, wild-type flies exhibited progressive muscle degeneration as they aged (sarcopenia), such that the gut muscle fibers gradually showed breakage starting around day 20, and extensive breakage was seen in 40-day-old fly gut muscles. We found that heterochromatin levels affected the ability to maintain muscle integrity, with 20-day-old Su(var)205+/− flies showing extensive muscle fiber breakage (Figure 1D, top middle), whereas hsp70-HP1 flies maintained their muscle integrity beyond 40 days after eclosion (Figure 1D, top right). We quantified the breakages in longitudinal muscle fibers in a defined area of the midgut and calculated the muscle integrity index for each genotype and age (see Methods). We found that the muscle integrity indices correlate well with the mobility of flies of different genotype and age (Figure 1D, bottom). These results are consistent with the differences in fly motility that we directly observed. Thus, maintenance of heterochromatin levels is essential for the maintenance of muscle structure and function, which consequently affect animal mobility and lifespan.

11If heterochromatin levels are important for longevity and tissue integrity, then normal aging should be accompanied by gradually decreasing heterochromatin levels. Indeed, it has been shown that normal aging in C. elegans, as well as the premature aging observed in human progeric syndromes, is correlated with changes in nuclear architecture and loss of heterochromatin –,. Since pericentromeric heterochromatin is readily observable in enterocytes, we examined HP1 foci in enterocytes of young and old adult flies. We found that, in contrast to young flies, whose enterocytes had prominent chromocenter enriched with HP1 (Figure 2A; top), old flies had much reduced levels of heterochromatin, with many nuclei in the gut epithelia lacking pronounced HP1 foci (Figure 2A; bottom). Since HP1 is recruited to heterochromatin by binding to histone H3 di- or tri-methylated at lys9 (H3K9m2 or H3K9m3), heterochromatin-specific chromatin modifications, we further investigated changes in the levels of H3K9m2 in flies of different ages. Interestingly, we found that total histone H3 levels decreased with age when compared with the non-histone nuclear protein HP1, which remained nearly constant relative to α-tubulin (Figure 2B). However, total levels of H3K9m2 showed a more dramatic decrease with age, and the decrease was obvious even relative to H3 levels (Figure 2B). Total H3K4m3 levels, on the other hand, showed a less dramatic decrease (Figure S4). Our results are consistent with previous reports that the levels of total histone H3 and heterochromatin marks decrease when animals age –,,.

12Taken together, the above observations suggest that, total histone H3 levels and their modifications by methylation, especially methylation of K9, exhibit gradual decline when animals age. The presence of excess HP1 throughout life might help preserving H3K9 methylation, thus delaying its decline. Although HP1 protein levels control heterochromatin levels during development, HP1 is not the sole factor determining heterochromatin formation post-development, especially in aged adult flies, where we have observed a decrease in the levels of H3K9m2, but not of HP1.

13To further confirm that HP1 is not localized on heterochromatin sequences in old flies, we carried out chromatin immunoprecipitation (ChIP) experiments to determine HP1 occupancy on the transposable element 1360, which is highly enriched in constitutive heterochromatin. Transposable element 1360 is present in >300 copies in the fly genome and has been used as a representative sequence for global heterochromatin,. The ribosomal protein 49 (rp49) gene is a constitutively transcribed gene normally not associated with heterochromatin and can be used as a negative control. By performing ChIP experiments using anti-HP1 antibodies followed by PCR amplification, we assessed the levels of HP1 occupancy in these sequences. Indeed, we found that HP1 was found associated with 1360 in young but not old wild-type flies (Figure 2C), whereas in flies carrying the hsp70-HP1 transgene, HP1 was also found associated with 1360 even in old flies. These results are consistent with the idea that heterochromatin levels decrease with aging, and that over-expressing HP1 prevented heterochromatin decline. Taken together, these results suggest that there is a gradual loss of heterochromatin when flies age, as with C. elegans and humans, that the lack of HP1 localization to heterochromatin foci in old flies is likely due to the loss of H3K9 methylation, and that over-expression of HP1 throughout life can prevent or delay heterochromatin loss.

14Heterochromatin loss could cause re-expression of genes that are normally repressed by heterochromatin. We thus examined the expression of a heterochromatinized lacZ transgene in young and old DX1 flies. These flies carry a tandem array of seven P[lac-w] transgenes, but lacZ (and white+) expression from these transgenes is normally repressed by DNA repeat-induced heterochromatin formation. A reduction in heterochromatin can cause derepression of the lacZ gene contained in the P[lac-w] elements of DX1 flies,. Indeed, we found that old, but not young, DX1 flies expressed lacZ (Figure 2D, 2E), consistent with the idea that heterochromatin levels decline with age. Thus, normal aging is accompanied by a gradual loss of heterochromatin in Drosophila as well.

15We next investigated the possible mechanism(s) by which heterochromatin formation promotes life span extension. It has been shown previously that H3K9 methylation and RNA interference regulate nucleolar stability. Loss of HP1 or Su(var)3–9 levels causes fragmentation of the nucleolus, as revealed by the nucleolar marker Fibrillarin. When we examined the effects of heterochromatin levels on nucleolar morphology, which can be seen most easily in 3rd instar larval salivary gland giant nuclei, we found that conditions that decreased heterochromatin levels, such as JAK over-activation or loss of STAT, were associated with nucleolar instability (Figure 3A). Conversely, conditions that increased heterochromatin formation, such as hop loss-of-function or HP1 over-expression, were associated with a stable nucleolus: the presence of a single, round nucleolus (Figure 3A). Moreover, HP1 over-expression suppressed the nucleolar fragmentation associated with hopTum-l (Figure 3A). These results are consistent with previous findings that JAK overactivation disrupts heterochromatin formation and that heterochromatin formation is important for nucleolar stability,.

16Nucleolar fragmentation has been attributed to illegitimate recombination of repeated DNA sequences, resulting in instability of the highly repeated rDNA locus. Illegitimate recombination events can be assessed quantitatively by measuring the levels of extrachromosomal circular (ECC) DNA. We isolated ECC DNA from flies of different genotypes and quantified ECC levels by calculating ECC index (see Methods). Indeed, we found increased ECC levels in mutants with decreased heterochromatin, such as hopTum-l, Stat92E, and Su(var)205, and decreased ECC levels in mutants with increased heterochromatin formation, such as hop+/− (Figure 3B).

17Interestingly, increased ECC formation due to instability of the rDNA locus has previously been shown to cause accelerated aging in yeast,. Taken together with our results from Drosophila, we suggest that the rDNA locus (or nucleolus) might be an important cellular target regulated by heterochromatin formation.

18Finally, we investigated the functional consequences of nucleolar instability. The nucleolus is the site of rRNA biogenesis, where precursor rRNA molecules are transcribed and processed to give rise to 18S, 5.8S and 28S rRNAs (Figure 4A). In Drosophila, as well as in mammals, the rDNA locus consists of a few hundred rRNA transcriptional units in tandem repeats. The number of rDNA genes vastly exceeds what is needed for adequate rRNA transcription. Normally only 20 to 25 units (<10% of the total) are actively transcribed, while the majority of the rDNA locus is silenced presumably by unknown epigenetic mechanisms. Since it has been shown that loss of heterochromatin, as in Su(var)205 transheterozygotes, leads to illegitimate recombination and thus instability of the rDNA locus, we investigated whether heterochromatin loss also leads to derepression of rDNA transcription.

19The levels of rDNA transcription can be more sensitively detected by examining the transcription of a class of transposons (e.g., R2 elements) that are specifically inserted into, and are cotranscribed with, the 28S rDNA gene. Normally the host “selects” a region of the rDNA locus free of R2 insertions for transcription and represses the R2-inserted rDNA units by unknown mechanisms. We found that in Su(var)205 transheterozygous mutants, however, transcription of R2 elements was dramatically increased by >40 fold compared to their sibling heterozygous control flies (Figure 4B). This suggests that in preserving the structural integrity of the nucleolus, heterochromatin formation plays a crucial role in silencing the transcription of the majority of rDNA genes, and that loss of heterochromatin causes a dramatic increase in rRNA transcription, which could lead to an increased capacity for protein synthesis, conducive to growth and accelerated aging.

20To investigate whether the moderately altered heterochromatin levels that have been shown to alter lifespan (Figure 1), affect the rate of rRNA synthesis, we measured pre-rRNA transcript levels in flies heterozygous for Su(var)205 or carrying an hsp70-HP1 transgene by quantitative real-time PCR (qRT-PCR). Indeed, we found that flies in which HP1 is moderately over-expressed (by basal activity of the hsp70 promoter without heat shock) contained >50% less pre-rRNA, whereas Su(var)205 heterozygous flies had levels of pre-rRNA transcripts that were >2 fold higher (Figure 4C).

21To determine whether these altered rRNA transcription rates affect growth, and thus the body size of the adult fly, which has often been inversely associated with lifespan, we measured the body size and weight of larval and adult flies with altered heterochromatin levels as mentioned above. Indeed, we found that flies moderately over-expressing HP1 had a smaller body weight, and that Su(var)205−/− larvae had a larger body size (length) (Figure 4D), and that Su(var)205+/− flies had a larger body weight (Figure 4E). The differences in body weight were not as pronounced as those in rRNA transcription, suggesting that body size may not be solely regulated by the rRNA transcription rate. Nonetheless, these results are consistent with the idea that changes in the rate of rRNA transcription may impact global protein synthesis and thus the growth of the organism.

22In summary, we have found that heterochromatin formation promotes longevity, genome stability, and suppresses rRNA transcription. The causal relationship between aging and rRNA transcription, however, awaits further investigation. It is interesting to note that factors that promote growth, such as insulin signaling and protein synthesis, usually accelerate aging, whereas inhibition of these pathways extends life span, –. Moreover, it has been shown in yeast that the Sir2 histone deacetylase counteracts aging by inducing heterochromatin formation at the rDNA locus,, thereby suppressing rRNA transcription and maintaining stability of the rDNA locus. In mammals, it has been shown that heterochromatin and Sirt1 epigenetically silence rDNA transcription in response to intracellular energy status. Thus, loss of rDNA silencing due to heterochromatin loss could lead to instability and increased rRNA transcription, which promotes protein synthesis in general. We suggest that suppression of rDNA transcription might be an evolutionarily conserved mechanism essential for animal longevity.

23All crosses were carried out at 25°C on standard cornmeal/agar medium unless otherwise specified. Fly stocks of hopTum-l, Stat92E06346, Su(var)20505, Su(var)20502, hop3, hsp70-Gal4, and UAS-eGFP were from the Bloomington Drosophila Stock Center (Bloomington, IN). Fly stocks of DX1 and 6-2 mini-white+ (J. Birchler), and hsp70-HP1 (G. Reuter; L. Wallrath) were generous gifts. All alleles used for life span analyses were extensively outcrossed before experiments (see below).

24The following outcrossing schemes were used to minimize genetic background effects. hsp70-HP1 (line 1; p[hsp70-HP1-eGFP, ry+] carried on the 2nd chromosome; ) flies were outcrossed to a ry506 stock for ten generations, and the ry+ or CyO marker was followed to derive outcrossed hsp70-HP1/+ and CyO/+ flies, respectively. These flies were crossed to establish new “outcrossed” hsp70-HP1 (p[ry+])/CyO; ry506 flies. Su(var)20505/CyO flies were outcrossed to In(1)wm4 stock for ten generations, and the CyO marker or suppression of In(1)wm4 PEV was followed to derive outcrossed In(1)wm4; Su(var)20505/+ and In(1)wm4; CyO/+ flies, respectively. These flies were crossed in single pairs to derive new “outcrossed” In(1)wm4; Su(var)20505/CyO stocks (the presence of Su(var)20505 was confirmed by both suppression of PEV and homozygous lethality). For lifespan analysis, the “outcrossed” hsp70-HP1 (p[ry+])/CyO; ry506 flies were crossed to In(1)wm4 flies, and the “outcrossed” In(1)wm4; Su(var)20505/CyO flies were crossed to ry506 flies, and the F1 non-CyO flies were collected for lifespan analysis. “Wild-type” control flies were the F1 of In(1)wm4 and ry506 flies. An independent stock of hsp70-HP1 (line 2; an unmarked p[hsp70-HP1-lacI] inserted in the 2nd chromosome; ) was outcrossed to In(1)wm4 stock (with a CyO chromosome “floating”) for ten generations. An isogenic line of In(1)wm4; hsp70-HP1 (line 2)/CyO flies was established from a single male, and the presence of hsp70-HP1 was confirmed by its strong enhancement of PEV. This stock was used for lifespan studies, and a line that did not enhance PEV (which was considered not carrying the hsp70-HP1 transgene) was used as wild-type control.

25To assess life span, 2-day-old females were separated from males and were transferred to fresh vials at 20 flies/vial, and were subsequently transferred to fresh vials every 2–3 days. Dead flies were counted upon each transfer. Only female heterozygotes were analyzed because hop is located on the X chromosome and the hemizygous mutants are not viable. Flies were cultured at 25°C and 70% humidity.

26Mouse monoclonal anti-HP1 (C1A9; Developmental Hybridoma Bank, Iowa; 1∶200), rabbit anti-H3(di)mK9 (07-212; Upstate Biotechnology; 1∶200), rabbit anti-GFP (CloneTech; ), and rabbit anti-Fibrillarin (Abcam; 1∶500) were used as primary antibodies and fluorescent secondary antibodies (Molecular Probes) were used in whole-mount immunostaining. Tissues were fixed in 4% paraformaldehyde/PBS and 0.3% Triton-X/PBS. Stained tissues were photographed with a Leica confocal microscope. Images were cropped and minimally processed using Adobe Photoshop CS.

27For chromatin immunoprecipitation (ChIP), adult flies of appropriate ages were snap frozen with liquid nitrogen, and then were cross-linked with 1.8% formaldehyde. The flies were homogenized in cell lysis buffer, and ChIP was performed as previously described.

28Adult flies of appropriate genotypes were split into two groups, one for Hirt ECC DNA isolation as described, and the other for genomic DNA isolation by standard protocols to use as controls. Typically 10 adult males were ground in 500 µl Hirt lysus buffer (0.6% SDS, 10 mM EDTA, pH 8.0) for ECC isolation or in 200 µl Buffer A (0.5% SDS, 100 mM EDTA, 100 mM NaCl, 100 mM Tris-HCl, pH 7.5) for genomic DNA isolation. Genomic DNA was quantified by spectrometry, and 100 ng of genomic DNA and an equal volume of ECC DNA were used for PCR amplification. Primer sequences used for amplifying ECC and control DNA were as previously reported. The PCR products from ECC sample and genomic control sample were loaded on the same agarose gel. PCR bands were revealed by ethidium bromide staining and photographed. The level of ECC was measured as the ratio of ECC band intensity to that from genomic control run on the same gel. Three independent experiments were done for each genotype. The ECC index was calculated as the sum of each ECC/genomic ratio divided by the total number of ECC bands:Where N denotes the total number of ECC species examined (N = 8 in this experiment).

29Flies were outcrossed as described in “Life span analysis” to minimize genetic background effects. Virgin female flies of a particular genotype were grouped by 5 in a cornmeal food vial (without supplementing yeast) and were passed daily to a fresh vial. Flies of different ages were video recorded for 2 min, around 9:30 AM, with a mounted digital camera. Recording was started after flies fell to the bottom of the vial by knocking the vial on a bench top. Motility scores were assigned to each fly in the video playback according to the speed with which it moved upward in the vial.

30Adult flies of desired ages were dissected and fixed with formaldehyde. The intestines were stained with phalloidin-fluorescein and observed with an epifluorescence microscope. At least 10 flies of each genotype and age were dissected. Each fluorescein-stained large intestine was assigned a Morphology Score (MS) of 0 to 10 based on the number of breakages in the longitudinal muscle fibers in a 3 gut-diameter long stretch of the midgut immediately adjacent to the hindgut. MS = 10−n, where n = the number of breakages in the defined region of the midgut. A score 10 represents prefect morphology: no muscle fiber breakage in the longitudinal muscles (n = 0). When n≥10, usually no continuous longitudinal muscle fibers can be identified and accurate counting breakages becomes difficult. In this case, a Morphology Score of 0 was assigned. So MS represents the worst morphology. The muscle Integrity Index is defined as the total MS of each genotype and age divided by the total number of intestines observed (N).

31Total RNA was isolated from 10 adult male flies (2-day-old) of desired genotypes using the RNeasy Mini Kit (Qiagen) or trizol (Invitrogen) according to the manufacturer's instructions. One µg of total RNA and primers specific for pre-rRNA and rp49 (control) were used to make the first strand cDNA using Superscript III reverse transcriptase (Invitrogen) in 50 µl total reaction volume. The cDNA (at 1∶100 dilution) was used as template for qRT-PCR analysis using SYBR green based detection on a BioRad iCycler. Reactions were carried out in triplicate, and melting curves were examined to ensure the presence of single products. The levels of pre-rRNA were quantified relative to rp49 transcript levels (control) and were normalized to a wild-type control. The following primer pairs (forward and reverse) were used.

32rp49: tcctaccagcttcaagatgac, cacgttgtgcaccaggaact

33pre-rRNA (5′ETS): atcggccgtattcgaatggattta, ctactggcaggatcaaccaga