Lifespan extension and elevated hsp gene expression in Drosophila caused by histone deacetylase inhibitors

The lifespan of an organism is influenced by both genetic and environmental factors. To date, our understanding of the mechanisms of aging is limited because of the biological complexities of the process. There has been evidence to support the hypothesis that aging is associated with the accumulation of abnormal and/or malfolded proteins and oxidatively damaged proteins in many organisms, including nematodes, flies and mammals(Gershon and Gershon, 1970; Stadtman, 1992). The heat shock proteins (Hsps) are induced in response to protein damage caused by heat and other stresses (Parsell and Lindquist,1993). Induction of hsp genes during Drosophilaaging might have a beneficial effect on the lifespan of flies(Wheeler et al., 1995; King and Tower, 1999). Mild heat stress of Drosophila transgenic for extra copies of the hsp70 gene produced increased expression of the gene and extended the lifespan (Tatar et al., 1997; Tatar, 1999). Heat-induced expression of hsp70 may reduce age-related mortality rates(Minois et al., 2001) and hsp22 and hsp23 genes were upregulated in selected lines for increased longevity (Kurapati et al.,2000), implicating the role of Hsp proteins in longevity determination. Molecular chaperones must be playing important roles in maintaining cellular functions during aging, through promoting protein renaturation and preventing proteins from aggregating and denaturing(Hartl, 1996; Marin et al., 1993).

Expression of hsp genes represents a special model of gene regulation involving basal and inducible expression. The acetylation/deacetylation modifications of N-terminal tails of core histones play critical roles in activation/repression of many eukaryotic genes,including hsp genes. In vitro, histone acetylation has been shown to be involved in hsp gene activation(Nightingale et al., 1998; Reid et al., 2000). Li et al.(1998) demonstrated that in Xenopus oocytes, p300 participated in the inducible transcription of the hsp70 gene both as a coactivator and an acetyltransferase. During Xenopus development, the presence of histone deacetylase (HDAC)inhibitors enhanced the heat shock-induced accumulation of both hsp70and hsp30 mRNA in post-midblastula transition (MBT) staged embryos and resulted in the expression of hsp30 immediately after MBT rather than at the late neurula/early tailbud stage under normal conditions(Ovakim and Heikkila, 2003). In a previous study, we showed that HDAC inhibitors trichostatin A (TSA) and sodium butyrate (BuA) were able to affect the chromatin structure at the site where the hsp70 gene is located along the polytene chromosome and significantly enhanced both the basal and the inducible expression of hsp70 gene in Drosophila(Chen et al., 2002).

Histone acetylation has also been shown to be involved in lifespan determination. In Saccharomyces cerevisiae, sir2 and rpd3,which encode two important histone deacetylases, were found to be involved in the regulation of lifespan, and deletion of sir2 shortened the lifespan (Kaeberlein et al.,1999), while the rpd3 knockout increased lifespan(Kim et al., 1999). Additional copies of sir2 in Caenorhabditis elegans also increased the lifespan of the worm (Tissenbaum and Guarente, 2001). The lifespan extension was achieved by feeding the HDAC inhibitor phenylbutyrate (Kang et al., 2002), or by a hypomorphic mutation of rpd3(Rogina et al., 2002) in Drosophila. These results indicate that histone acetylation is associated with lifespan. But the molecular mechanisms of HDAC-dependent changes in aging and lifespan remain unclear.

The aim of this study was to investigate the roles of histone hyperacetylation in expression of hsp genes during aging and longevity determination in Drosophila. Our results show that the increase in acetylation level of histone H3 enhances the basal and inducible expression of hsp22 and hsp70 during aging, and extends both mean and maximum lifespan, to variable extents, in different lines of Drosophila. This strongly implicates a correlation among histone acetylation, hsp gene expression and longevity determination.

Wild-type strain of Drosophila melanogaster, Canton-S, was used. Flies were raised on standard cornmeal/yeast/agar medium under a 12 h:12 h light:dark cycle.

Eggs from a single female fly were collected and used for the screening of iso-female lines, and ten iso-female lines were established and the lifespan was measured. The long-lived and short-lived lines were then screened from these lines. To obtain flies of defined age, newly eclosed flies were collected and maintained at 25°C at a density of 50 per vial and were transferred to fresh vials every 3 days toprevent growth of bacteria or mold,and 200 flies for each line were cultured. The number of dead flies was counted every day.

The HDAC inhibitors trichostain A (TSA; Sigma, MI, USA) and sodium butyrate(BuA; Sigma) were used at final concentrations of 10 μmol l^-1and 10 mmol l^-1, respectively. The third instars larvae were divided into equal aliquots for different treatments. Part of the larvae were cultured for 5 h in physiological brine (130 mmol l^-1 NaCl, 4.7 mmol l^-1 KCl and 1.9 mmol l^-1 CaCl₂) and then allowed to develop directly into flies. For the repeated mild heat shock(RMHS), the flies were induced at 32°C for 1 h every 3 days. Some of the larvae were heat shocked (HS) at 37°C for 30 min before they were allowed to develop into flies. For HDAC inhibitor treatment, the larvae were cultured in physiological brine with TSA or BuA and then allowed to develop into flies either without additional inhibitor feeding (o.TSA and o.BuA) or were continuously cultured on medium containing inhibitors (c.TSA and c.BuA). The remaining flies were treated with both HDAC inhibitors and RMHS (TSA-RMHS and BuA-RMHS).

The flies were cultured and treated as described above until 6 and 30 days old, or induced at 37°C for 2, 5, 10, 20 and 40 min. Total RNA was extracted from flies using the RNA extraction kit supplied by Promega. The reverse transcription (RT) reaction was performed by using a RT system(Promega) following the manufacturer's protocol. Quantification of mRNA was performed using an ABI PRISM® 7700 sequence Detection System (PE Applied Biosystems, Weiterstadt, USA) and SYBR® Green (PE Applied Biosystems) as a double-stranded DNA-specific fluorescent dye. Rp49 (ribosomal protein 49) was used as a housekeeping gene for standardizing hsp mRNA expression.

Amplification mix (50 μl) contained 4 μl cDNA solution, 10×SYBR® Green buffer, 0.24 mmol l^-1 MgCl₂, dNTP (0.2 mmol l^-1 dATP, dCTP, dGTP, 0.4 mmol l^-1 dUTP), 1.25 U AmpliTaq Gold DNA polymerase, 0.5 U Amp Erase UNG and 200 nmol l^-1DNA primers. Amplification primers for hsp22 were 5′-ctttcacgccttcttcc-3′ and 5′-gcggttttgtcttttgg-3′;for hsp70 were 5′-agccgtgccaggtttg-3′ and 5′-cgttcgccctcataca-3′; and for rp49 were 5′-agcacttcatccgccacc-3′ and 5′-atctcgccgcagtaaacg-3′. Samples were incubated for 2 min at 50°C to inhibit carryover of contamination. Thermal cycling conditions consisted of an initial denaturation step at 95°C for 10 min, followed by 40 cycles of 15 s at 95°C and 1 min at 60°C. All PCR reactions were run in duplicate.

The Ct (threshold cycle) was defined as the number of cycles required for the fluorescence signal to exceed the detection threshold. Data were analyzed by using the 2^-ΔΔCt method, which is a convenient way to analyze the relative changes in gene expression(Livak and Schmittgen, 2001). Quantitative results are given as means ± s.e.m.

Flies were cultured and treated as described above. Nuclear protein was extracted as described by Giancotti et al.(1984), or whole flies were homogenized directly into lysis buffer [50 mmol l^-1 Tris-HCl pH 8.8, 300 mmol l^-1 NaCl, 1% NP-40, 1 mmol l^-1 EDTA, 1 mmol l^-1 dithiothreitol (DTT), 1 mmol l^-1phenylmethylsulfonyl fluoride (PMSF)] to obtain total proteins. Following 15%SDS-polyacrylamide gel electrophoresis (PAGE), protein was transferred to nitrocellulose membranes, then the membranes were incubated with antibodies to acetylated histones H3 and H4 (Upstate Biotechnology, Lake Placid, NY, USA),to Hsp22 (produced in rabbit), and to actin (Santa Cruz, CA, USA),respectively. Actin was used as a housekeeping gene product for normalizing the loading materials.

Longevity is dependent on both genetic and environmental factors. Under the same environmental conditions, diversity of longevity occurs among the different individuals. We established 10 iso-female lines of Drosophila melanogaster, and the number of dead flies was counted every day. From these lines, we selected the long-lived line iso2 and the short-lived line iso4. The maximum and mean lifespan of the long-lived iso2 line were 73 days and 45.53 days, respectively (Fig. 1). In comparison, the maximum and mean lifespan of the short-lived iso4 line were 62 days and 38.47 days, respectively(Fig. 1).

A real-time quantitative PCR procedure was used to analyze the RNA expression of hsp genes in long- and short-lived lines during aging. It is clear from Fig. 2 that the hsp22 (Fig. 2A)and hsp70 (Fig. 2B)basal expression declined with aging and this decline was more distinct in the long-lived line than in the short-lived line(Fig. 2). On the other hand, hsp22 and hsp70 mRNA levels were higher in the long-lived line relative to the short-lived line in young flies(Fig. 2), but became less distinct in old flies (Fig. 2). No prominent differences were seen in the expression of other hspgenes, such as hsp26, during aging and between long- and short-lived lines (data not shown).

The hsp RNA expression after 37°C heat induction for different periods of time was also examined by real-time quantitative PCR. As shown in Fig. 3, differences in the response to heat induction and in the level of expression of hsp22(Fig. 3A) and hsp70(Fig. 3B) gene products in fly strains with different lifespan were observed. The long-lived line was more sensitive to heat shock than the short-lived line. Heat shock for 2 min caused rapid increase in the expression of hsp22 and hsp70 in long-lived line (Fig. 3, iso2),but in the short-lived line, a detectable change in hsp22 and hsp70 expression did not occur until 5 min of heat shock(Fig. 3, iso4). Also, hsp22 expression increased by approximate 130-fold in short-lived flies at 40 min of induction, while only an approximate 80-fold increase was detected in long-lived flies at the same time(Fig. 3A). It can also be seen from Fig. 3B that the intensity of hsp70 gene did not change significantly upon heat treatment.

The experimental results described above revealed a correlation between hsp gene expression and longevity in flies. We then investigated the tolerance of flies to persistent high temperature, and we found that the long-lived flies had a greater resistance to high temperature stress. Within 30 min of the treatment at 37°C rapid death occurred in the short-lived line, whereas in long-lived flies, the rapid death occurred after 60 min of the heat treatment (Fig. 4).

To test the action of HDAC inhibitors (TSA and BuA) and heat shock in flies with different longevity, we cultured and treated flies as described in Materials and methods, and the lifespan was measured. The repeated mild heat shock (RMHS) resulted in an extension of both the mean and maximum lifespan of the flies (Fig. 5A,E), but the degrees of extension were different between the long- and short-lived lines. The maximum lifespan increased only slightly, by 5.6%, in long-lived flies,and moderately, by 11.5%, in short-lived flies; the mean lifespan increased moderately, by 10.3%, in long-lived flies and considerably, by 25.8%, in short-lived flies compared with the control. It can be seen from Fig. 5B,F that the HDAC inhibitor TSA strikingly influenced the lifespan in both long- and short-lived lines, but with variable degrees. For the long-lived flies, one-off TSA treatment (o.TSA) had no obvious effect on both mean and maximum lifespan,while continuous TSA treatment (c.TSA) only increased mean lifespan by 15.6%(Fig. 5B). For short-lived flies, o.TSA resulted in an extension of mean lifespan by 5%, and c.TSA significantly increased mean lifespan by 24.4% and maximum lifespan by 16.4%(Fig. 5F). Data in Fig. 5C,G demonstrated that BuA treatment only affected the lifespan of the short-lived line, as one-off BuA treatment (o.BuA) increased the mean lifespan by 25.8% and the maximum lifespan by 11.5%. When BuA was added continuously throughout their lifetime(c.BuA), the life extension effect became less obvious(Fig. 5G). When the flies were fed with HDAC inhibitor and RMHS was applied every 3 days, we found no significant differences among RMHS, TSA-RMHS and BuA-RMHS treatments, and a similar extension to that of the controls was seen(Fig. 5D,H).

In order to test if changes in histone acetylation are involved in HDAC inhibitor-mediated lifespan extension and hsp gene induction, we examined the acetylation levels of core histones (H3 and H4) after HDAC inhibitor treatment and heat shock using antibodies against acetylated H3 and H4. As shown in Fig. 6,acetylated lysine of histone H3 were detected. The photodensitometric analysis of the bands revealed that both TSA and BuA caused an increase in the acetylation level of H3 in long-lived (Fig. 6A) and short-lived (Fig. 6B) lines. More specifically, o.TSA and o.BuA treatments moderately increased the acetyl-H3 level (P<0.05)(Fig. 6C, o.BuA and o.TSA). c.TSA treatments brought about a significant increase (P<0.01)(Fig. 6C, c.TSA), whereas c.BuA treatment moderately increased the acetyl-H3 level in long-lived flies(P<0.05), and significantly increased it in the short-lived line(P<0.01) (Fig. 6C,c.BuA). Moreover, RMHS also increased acetyl-H3 (P<0.01)(Fig. 6C, RMHS). When a combined treatment of both HDAC inhibitors and heat shock was applied, there was an additive effect (P<0.01)(Fig. 6C, TSARMHS and BuARMHS). However, no changes were detected in the acetylated lysine of histone H4 under the same treatments (data not shown).

Next, we wanted to know whether the hyperacetylation of histone H3 influences hsp expression during aging. The hsp mRNA levels were assayed by quantitative real-time PCR, and it was found that after c.TSA and o.BuA treatments, hsp22 and hsp70 basal expression increased moderately in young flies (6-day old) of the long-lived line(P<0.05) (Fig. 7A,C, iso2), while it increased considerably in young flies (6-day old) of the short-lived line (P<0.01)(Fig. 7A,C, iso4). With the aging of the flies, the degree of increase in these gene products declined gradually. In 30-day old short-lived flies, only the basal transcription of hsp22 increased moderately compared with control (P<0.05)(Fig. 7B, iso4). For the inducible expression of hsp genes, only TSA treatment gave rise to the noticeable effect (Fig. 7E-H). In both long- and short-lived young flies, TSA-RMHS treatment caused a higher expression of hsp22 and hsp70 than RMHS (Fig. 7E,G). Similarly,the increase in inducible hsp expression declined gradually during aging (Fig. 7E-H). Until 30 days after eclosion, TSA-RMHS treatment only increased hsp22expression in the short-lived line (Fig. 7F, iso4), while no increase in hsp70 expression was detected (Fig. 7H). In addition, to determine whether changes in mRNA levels actually reflect changes in the relative levels of proteins, we assayed Hsp22 protein content in young flies after HDAC inhibitor treatment and heat shock. Total protein was isolated from 6-day-old flies and Hsp22 protein was detected by western blot(Fig. 8). Consistent with changes in the hsp22 mRNA level, Hsp22 protein expression was up-regulated by the HDAC inhibitors (Fig. 8, c.TSA and o.BuA). Similarly, continuous TSA treatment (c.TSA)and one-off BuA treatment (o.BuA) strikingly increased Hsp22 protein basal expression and only TSA treatment produced a rise in inducible expression of Hsp22 (Fig. 8, TSA-RMHS). Additionally, it is noticeable that the HDAC inhibitor-induced up-regulation of Hsp expression was more prominent in the short-lived line than in the long-lived line (Fig. 8, iso2 and iso4).

In conclusion, this study clarifies the correlation between the elevated expression of hsp genes and the longevity in Drosophila melanogaster. The results revealed the higher hsp basal expression level, higher thermotolerance and higher response to heat shock,but lower hsp22 induction under continuous high temperature in long-lived than in short-lived flies. Drosophila lifespan could be extended by feeding the flies with HDAC inhibitors TSA and BuA, and/or repeated mild heat shock. The range of lifespan extension differed between long- and short-lived lines. HDAC inhibitor treatment promoted hspbasal expression, but with the time of aging, the extent of this increase declined gradually. Chromatin modulation may be involved in HDAC inhibitor-mediated hsp gene activation, since hyperacetylation was detected in core histone H3 upon TSA and BuA treatments. Thus, a close correlation among histone acetylation, hsp gene regulation and aging in D. melanogaster can be established.

It has been shown that aging is associated with the accumulation of inactive enzymes, partially denatured and/or damaged proteins(Gershon and Gershon, 1970; Stadtman, 1992). Heat shock proteins are thought to reduce protein denaturation and aggregation,facilitate re-folding of partly denatured proteins, direct the entry of damaged proteins into proteolytic pathways and protect the organisms from additional stress (Wheeler et al.,1995; Parsell and Lindquist,1993). Results presented in this report indicated that hsp22 and hsp70 genes were upregulated and more promptly induced in the long-lived iso-female line (Figs 2 and 3). Also, the thermotolerance was higher in the long-lived line (Fig. 4). Lin et al.(1998) demonstrated that the long-lived D. melanogaster mutant displayed resistance to starvation,high temperature and oxidative stress. A similar increased resistance to heat was observed in long-lived C. elegans lines(Lithgow et al., 1995). It is possible that higher expression of hsp genes may help maintaining cellular functions during aging and prolong the lifespan of the fly. Meanwhile, in long-lived flies, the higher response to heat shock perhaps results in the higher thermotolerance. On the other hand, both the basal expression and induction speed of hsp genes in short-lived flies were low. When the flies were kept in harsh conditions, such as high temperature,more Hsp22 was induced to repair the damage(Fig. 3A). This process presumably consumes more energy and therefore results in quicker death.

Previously, two patterns of hsp expression during aging were reported. With Oregon-R wild-type and W¹¹⁸ flies cultured at low density in culture bottles, hsp22 RNA levels were low or undetectable in young flies, but hsp22, hsp23, hsp70 were upregulated with age (King and Tower,1999). In contrast, when the same Oregon-R flies were cultured at high density using population cage protocol, the starting level of hsp22 RNA in young flies was high, and the increase in hsp22RNA during aging was minor or undetectable(Kurapati et al., 2000). In the experiments of this study, Canton-S wild-type flies were cultured using the high-density population cage protocol. In contrast to the previous observations, we found that hsp22 and hsp70 mRNA levels were relatively higher in young flies, and declined with age(Fig. 2). We presume that this may be partially due to the different strain used. Also, it is important to note that this difference was observed under conditions that were more stressful than the optimum.

Moreover, data in this study revealed that the response of hspgenes to heat induction varied in flies with different longevity. In the long-lived flies, hsp genes were more sensitive to heat than in the short-lived line. In addition, different hsp genes exhibited a variable response to heat, i.e. hsp22 was more responsive to heat than hsp70. This suggests that a more positive correlation may exit between hsp22 and longevity, while hsp70 may affect longevity to a lesser extent. In contrast, the expression of hsp26did not differ significantly in expression between long- and short-lived flies(data not shown), implying that it may not be closely related to longevity. These results implicate the existence of a mechanism for preferential induction of hsp genes during aging.

The mechanisms of hsp gene induction by heat stress in Drosophila have been studied in some detail(Lis and Wu, 1993). Binding sites for the GAGA transcription factor and other cis-acting sequences at hsp promoters are required for generation of an accessible promoter structure in unstressed cells, which involves a transcriptionally engaged RNA polymerase paused ∼25 bp downstream of the start site for transcription (Gilmour and Lis, 1986; Lee et al.,1992; Rougvie and Lis,1988). Upon heat shock, the heat shock factor (HSF) is converted into an active trimer from inactive monomers, and binds to the heat shock element (HSE), facilitating the transcription of hsp genes. However,the role of histone acetylation in transcription regulation of hspgenes is still obscure. In this study, we showed that the acetylation level of H3 was increased upon treatment with the HDAC inhibitors, TSA and BuA(Fig. 6). We also demonstrated that hsp genes were upregulated after HDAC inhibitor treatment(Fig. 7). Interestingly, we noticed that heat shock also increased the acetylation level of H3(Fig. 6), further confirming the involvement of histone acetylation in hsp gene transcription in D. melanogaster. TSA and BuA are the two widely used HDAC inhibitors,but they may function through different mechanisms(Gibson, 2000; Finnin et al., 1999). TSA was originally identified as a fungal antibiotic with differentiation inducing properties, and it can strongly increase the acetylation level of histone H3(Zhong et al., 2003). BuA is a simple chemical that can raise the acetylation level of H3 and H4 by suppressing HDAC activity via a noncompetitive mechanism(Gibson, 2000). TSA and BuA affected hsp gene expression differently in this study. TSA was able to induce the expression of hsp22 and hsp70 genes above their basal level, and it worked cooperatively with heat shock to raise the inducible transcription intensity of hsp22 and hsp70 genes(Fig. 7). BuA only promoted the basal expression of hsp genes(Fig. 7A,C). In addition,continuous TSA treatment (c.TSA) and one-off BuA treatment (o.BuA) increased hsp22 and hsp70 expression more in short-lived flies, and this result was consistent with the influences of these HDAC inhibitors on the longevity of the flies (Fig. 5F,G). Moreover, as aging proceeded, HDAC inhibitor treatments became less effective in promoting hsp gene expression. This may presumably be due to the increasingly inactive response and repair mechanisms in aged flies.

A balance of activation and repression of various genes regulates an optimal physiological and cellular environment for longevity. The acetylation/deacetylation modifications of histones play critical roles in activation/repression of many genes, and hence probably regulate the lifespan. In yeast and worms, the HDAC Sir2 and Rpd3 have been shown to be closely related to lifespan (Kaeberlein et al.,1999; Kim et al.,1999; Tissenbaum and Guarente,2001), and feeding of the HDAC inhibitor phenylbutyrate(Kang et al., 2002) or a hypomorphic mutation of rpd3(Rogina et al., 2002) in the Drosophila extended the lifespan. However, mutation in Drosophila sir2 (dsir2) gene did not shorten lifespan, as predicted from yeast and worms (Newman et al.,2002). It appears that the influence of histone acetylation on longevity is a complex process and may involve different pathways and mechanisms. We treated flies with a one-off and continuous feeding of HDAC inhibitors, heat shock, as well as combinations of both HDAC inhibitors and heat shock, in an attempt to elucidate the effects of HDAC inhibitor-mediated histone acetylation on longevity. The results showed that TSA strikingly extended the lifespan in different ways, as o.TSA treatment only increased the mean lifespan in the short-lived line, and c.TSA treatment extended the mean lifespan by 15.6% in the long-lived line; while in the short-lived line, it extended both the mean and maximum lifespan by 24.4% and 16.4%, respectively(Fig. 5B,F). Thus, a prolonged TSA action in vivo is probably necessary for longevity determination. However, BuA had less effect on lifespan. Only in the short-lived line did o.BuA treatment extend both the mean and maximum lifespan. When BuA was applied throughout the lifetime of the flies, its life-extension effect became insignificant (Fig. 5G). We presume that prolonged BuA feeding may be toxic, which reduced the survival rate. As a rule, both TSA and BuA were more effective in the short-lived line than in the long-lived line. It appears that different genetic background may determine the action as well as the type and optimal concentrations of the HDAC inhibitors. Furthermore, combined treatments of both inhibitors and heat shock (TSA-RMHS and BuA-RMHS) and heat shock alone (RMHS) had basically the same effect on life extension (Fig. 5D,H). We reason that the life extension by RMHS treatment may mask the influences of HDAC inhibitors. These results suggest that establishment of an altered cellular environment by HDAC inhibitors and heat shock changed the lifespan of flies, possibly through a pathway that involves the change in expression of certain genes resulting in inhibition of the accumulation of damaged proteins, and/or stimulation of repair mechanisms.

This work was supported by grants from The National Natural Science Foundation of China (30370316) and The Major State Basic Research Project of China (G1999053902).