RPD3 histone deacetylase and nutrition have distinct but interacting effects on Drosophila longevity

Nutrients and RPD3 have additive effects upon longevity

This study utilizes the same alleles as our previous report [8], rpd3^def (null), rpd3^P-UTR (hypomorphic), and rpd3^P-1.8 (control), and varies complete nutrients (both sucrose and yeast) from 0.3N to 1.5N relative to standard 1.0N conditions. The rpd3^def and rpd3^P-UTR alleles are lethal when homozygous [10]. In all survival experiments using the rpd3^def allele, including the study of flies that are double mutants for rpd3 and Thor, crosses prior to the life span measurements yielded flies with matched genetic backgrounds (a hybrid mix of the rpd3 stock and the yw strain, the latter being wild type for all genes related to this study). rpd3^P-UTR has greatly reduced transcription in all tissues, whereas rpd3^P-1.8 only has reduced transcription in the eye [10, 28]. All survival experiments using the rpd3^P-UTR and rpd3^P-1.8 alleles also had matched genetic backgrounds, derived from crosses to CS flies. Heterozygous male flies (20 days old) carrying the rpd3^def and rpd3^P-UTR alleles have 40% and 25% reduced levels, respectively, of rpd3 mRNA (Supplementary Figure 1).

Long-lived rpd3 mutant flies exhibited nutrient-mediated changes in longevity. While a diet containing corn meal and yeast is often considered standard (hereafter called corn), the standard food used in this study contained equivalent calories but was formulated using sucrose and yeast, called 1.0N [26, 29]. Males heterozygous for rpd3^def have extensions of 37% and 42% in mean longevity on corn and 1.0N food, respectively, compared to genetic controls (Figure 1A, B). Female flies heterozygous for the same allele have 4% and 31% longevity extension on corn and 1.0N food (Figure 1A, B, Table 1, 2). Mean life spans at different nutrient levels are shown for heterozygotes carrying the rpd3^def allele in Figure 1C and heterozygotes carrying the rpd3^P-UTR and rpd3^P-1.8 alleles in Figure 1D (Table 1, 2). Male and female flies carrying the rpd3^def allele have extended longevity at all nutrient levels compared to genetic controls, while also showing increases and decreases in mean life span mediated by nutrients. Life span extensions for males vary between 36-41% compared to controls at the same food levels, while for females they vary between 12% at 1.5 food and 33% at 0.7N food. Male and female flies carrying the second allele, rpd3^P-UTR, have smaller but consistent extensions in life span compared to flies carrying the control rpd3^P-1.8 allele (at most food levels). Females have 10-13% extensions at 0.5N, 1.0N, and 1.5N food levels, whereas males have extensions of 6-7% at 0.3N, 0.5N and 1.0N food levels.

Some nutrient-mediated changes in life span differed for flies carrying the two rpd3 alleles. In male and female flies carrying the rpd3^def allele and control flies for both alleles (rpd3^def controls, and rpd3^P-1.8) the peak extension due to DR (maximal extension of mean longevity compared to 1.0N food) occurs at 0.7N (Figure 1C,D and Table 1–4). However, in male and female flies carrying the rpd3^P-UTR allele the peak extension due to DR occurs at 0.5N (Figure 1D, and Table 3, 4). The effect of high calories (1.5N) also differed. The mean life span is unchanged for rpd3^P-UTR and rpd3^P-1.8 flies on 1.5N and 1.0N food, but for rpd3^def and rpd3^def controls the mean life span is lower on 1.5N compared to 1.0N food. Since all of the survival studies reported here were performed at the same time with the same batches of food, precluding food preparation and fly culture as sources of variability, and since the two controls for each allele exhibited different responses to high calorie food (1.5N), it is likely that the response to high calorie food was affected by the genetic background. It has been shown by other groups that flies with different genetic backgrounds can differ in their responses to nutrients [30].

It is of interest that there is a DR peak for flies carrying both the null and hypomorphic rpd3 alleles, despite the potentially confounding effects of the two hybrid genetic backgrounds used in these experiments. For males carrying either allele, the DR peak is 7-9% higher than the mean life span at 1.0N, whereas for females carrying either allele the DR peak is 12-15% higher. Male controls for both alleles have a DR peak of 11%, whereas the female controls for rpd3^def have a DR peak of 11% and the female controls for rpd3^P-UTR have a DR peak of 20%. Finally, there is a trend when comparing mutants to controls. For both males and females carrying the rpd3^P-UTR allele, the DR peak is lower in mutants compared to controls. A similar trend is seen for males but not females carrying the rpd3^def allele.

rpd3 mutants live longer under conditions close to starvation

Decreasing food below the level that yields DR induces increasing starvation. Flies raised on 0.1N food since eclosion have much shorter life spans compared to flies raised on all higher nutrient levels, indicating a condition close to starvation. rpd3^P-UTR flies live longer on 0.1N food compared to controls, indicating increased resistance to severe food deprivation (Figure 1E,F).

Effects of rpd3 on the expression of 4E-BP and Tor

Our results indicate that decreased rpd3 mRNA levels has a minor effect on DR-mediated longevity extension and a major effect on resistance to food deprivation. In order to gain more insight into the interaction between RPD3 and nutrient responses, we determined whether as a transcriptional regulator RPD3 might be affecting the mRNA levels for two post-transcriptional modulators of nutrient signaling, 4E-BP and Tor (Figure 2).

4E-BP regulates mRNA translation and the Tor kinase has a central role in nutrient-mediated signal transduction [31]. We first determined whether 4E-BP and Tor mRNA levels are affected by nutrients in wild type Drosophila. Heads and thoraces were assayed separately, in case changes were localized to one area of the body. Young and old flies were also assayed, in case changes were affected by aging. Young (10 day) wild type female flies have modestly increased 4E-BP mRNA levels at 0.5N compared to 1.5N (similar trends in head and thorax, though the changes are not signifi-cant), whereas older (40 day) flies show no changes (Figure 2A,B). Another group has shown that 4E-BP protein levels are increased under DR conditions [32].

Tor mRNA shows much smaller changes, comparing 0.5N and 1.5N of young and old flies (Figure 2C,D), though the change in 40 day thoraces is significant. It was then determined whether rpd3 mutations affect the transcription of 4E-BP and Tor. When heterozygotes for the rpd3^def allele are compared to genetic controls, young mutant male flies (20 days) show no changes in 4E-BP mRNA levels but older mutant flies (40 days) have 60% less mRNA (Figure 2E). Male flies heterozygous for the rpd3^P-UTR allele have a 40% reduction in 4E-BP mRNA levels when younger and a smaller decrease relative to controls when older (only thoraces were assayed, see Methods) (Figure 2F). Flies heterozygous for the rpd3^def allele have no significant changes in Tor mRNA level at both ages compared to genetic controls (Figure 2G). Flies heterozygous for the rpd3^P-UTR allele have a 60% reduction in Tor mRNA when young and tend towards a decrease relative to controls when older (though this is not significant) (Figure 2H). In summary, the transcription of 4E-BP is affected by both rpd3 alleles, whereas only one allele affects Tor transcription. In all cases where there were significant changes, reduced dRPD3 levels led to decreased transcription. DR and rpd3 mutants differed in their effects upon 4E-BP and Tor transcription, providing further evidence that dRPD3 and nutrient status differ in their mechanism of life span extension.

RPD3 and 4E-BP have distinct effects on longevity

Several groups have reported that flies lacking any 4E-BP have lowered life spans at all nutrient levels [24, 32–34]. Since rpd3-mutant flies have moderately reduced levels of 4E-BP mRNA, we asked whether a moderate reduction in 4E-BP mRNA might affect life span. This was measured by using a hypomorphic allele of the 4E-BP gene, Thor^k07736. Flies carrying this allele were backcrossed to yw flies for 10 generations. Hetero-zygotes carrying the backcrossed Thor^k07736 (4E-BP) allele have a 30% reduction in 4E-BP mRNA (Figure 3A). Heterozygous Thor males have mean life spans that are reduced by 19%, 7%, and 5% relative to yw controls at 1.0N, 0.7N, and 0.5N food levels, respectively (no change at 1.5N) (Figure 3B, Table 5, 6). Heterozygous Thor females have mean life spans that are reduced by 12%, 33%, and 9% relative to yw controls at 1.5N, 1.0N, and 0.5N (no change at 0.7N) (Figure 3B, Table 5, 6). There were several additional differences between Thor mutants and their controls. Over a food range of 0.5N to 1.5N, yw controls show a peak life span at 1.0N, whereas Thor mutants have a peak at 0.5N (Figure 3B, Table, 5, 6). In addition, 1.5N food decreased the mean life spans of yw controls relative to 1.0N food but not the mean life spans of the Thor mutants. These differences between Thor mutants and the controls provide some evidence that 4E-BP can affect the response to nutrients.

We next examined the longevity of flies heterozygous for both Thor^k07736 (4E-BP) and rpd3^def to directly assess the combined effects of these gene products (Figure 3C,D, Table 7, 8). Preliminary experiments indicated that rpd3 mutations mediated a decrease in 4E-BP mRNA when the flies also carried a Thor mutation (double heterozygotes) (Fig. Supplemental 2), similar to the decrease mediated by rpd3 mutations in the absence of Thor (Fig. 2). Double heterozygotes, single hetero-zygotes, and control flies shared the same genetic background. rpd3-mediated life span extension was attenuated but not eliminated by the presence of the Thor allele (Figure 3 C,D, Table, 7, 8). The reduction in mean life span of the double mutants relative to flies singly mutant for rpd3^def ranged from 4% - 14% for males and 3% - 12% for females. While female flies that were doubly heterozygous maintained a smaller DR peak of 0.7N, male flies that were doubly heterozygous had no DR-mediated increase in life span. Doubly heterozygous flies of both sexes had decreased life spans at 1.5N relative to 1.0N. The simplest interpretation would be an overall superimposition of the life span decrease mediated by reduced 4E-BP on the life span extension mediated by reduced dRPD3.

In summary, DR-mediated longevity extension is maintained in the presence of rpd3 mutations, though it may have a moderately reduced magnitude. Our epistasis experiments using flies doubly mutant for rpd3 and 4E-BP suggest potential crosstalk between dRPD3 and nutrient signaling.

Fly strains and maintenance

rpd3-deficient (rpd3^def24/TM6,Sb), rpd3–hypomorphic (rpd3^P-UTR/TM3,Sb,Ser), and control for rpd3–hypomorphic (rpd3^P-1.8/TM6,Tb) strains were used in our experiments. The hypomorphic rpd3^P-UTR allele has a P-element inserted in the 5′UTR region of the rpd3 gene, close to the transcription start site, which affects expression throughout the fly's body [10]. The control rpd3^P-1.8 allele has a P-element inserted 1.8 kb upstream from the transcriptional start site, which only decreases expression in the eye [10]. The rpd3^P-UTR and rpd3^P-1.8 alleles are derived from the same mutagenesis. The rpd3^def allele was derived by excision of the P-element from rpd3^P-UTR [10]. Canton S (CS), yw and Thor^k07736 flies were kindly provided by the Bloomington Stock center. Thor^k07736 flies were backcrossed to the yw strain for 10 generations to minimize the effects of genetic background.

Maintenance of aging flies

Flies were collected within 24 hours of eclosion and maintained using standard culture media in plastic vials. They were kept at 25°C in a humidified incubator on a 12 hour light: dark cycle. About 25 males and 25 females are kept together in each vial. For life span studies flies were passed every 2 days up to age 30 days and every day after that and the number of dead flies were counted. The number of flies in each life span study is listed in Tables 1 – 8. Standard corn food was prepared as previously described [43, 44]. The nutrient media used are dilutions (0.1N, 0.3N, 0.5N, 0.7N) or concentrations (1.5N) of an arbitrary normal condition, 1.0N, containing 100 g/L brewer's yeast (MP Biomedicals, Inc.), 100 g/L sucrose (MP Biomedicals, Inc.), 20 g/L agar, and 10 mg/L tegosept. Our nutrient protocol coordinately changes both sugar and protein by varying sucrose and yeast [26, 29]. The concentrations of agar and tegosept are not varied for the various media.

Genetic crosses

Male yw flies were crossed to female rpd3^def24/TM6,Sb. F1 progeny not carrying the TM6 balancer were used for life span measurements (flies carrying the mutation). F1 progeny carrying the balancer were crossed to each other and F2 progeny not carrying the balancer and not yw were used as the life span controls. Male flies homozygous for Thor^k07736 (backcrossed into a yw genetic background) were crossed to female rpd3^def24/TM6,Sb flies. F1 progeny not carrying the balancer were used for life span measurements (doubly mutant flies). Life span measurements with flies only heterozygous for Thor^k07736 were performed with progeny from crosses of Thor^k07736 homozygotes and yw. For studies using the other rpd3 alleles, rpd3^P-UTR/TM3,Sb or rpd3^P-1.8/TM6,Sb were crossed to wild type CS and F1 progeny not containing the balancer were used for life span measurements.

Quantitative PCR (Q-PCR)

Flies were separated by sex and frozen at the appropriate age. Thoraces of rpd3^P-UTR /+ and rpd3^P-1.8/+ flies were dissected since the rpd3^P-1.8 allele has reduced expression in the eye. Heads and thoraces were dissected from rpd3^def/+ and +/+ flies since the rpd3 is reduced throughout the body in rpd3^def/+. RNA was isolated by using the standard Chomczynski protocol and Trizol reagent (Gibco BRL) [44]. cDNA was synthesized from RNA. Using TaqMan primers and the Applied Biosystems Thermal Cycler (AB 7500 System), levels of gene expression were determined. Ankyrin was used as an endogenous control [46].

Statistical analysis

Statistical analysis of Q-PCR results was done using the Student's T-Test from at least three independent experiments and expressed as P values. Error bars represent the standard errors, and P values are specifically indicated in the figures. Life span data were analyzed by Log-Rank tests using the JMP-11 program (SAS). All survivorship data were censored for the first 10 days. Maximum life span was calculated as the mean life span of the longest surviving 10% of the population.