The age- and sex-specific decline of the 20s proteasome and the Nrf2/CncC signal transduction pathway in adaption and resistance to oxidative stress in Drosophila melanogaster

Hydrogen peroxide resistance is diminished with age

Prior studies suggest an age-related decline in stress resistance of D. melanogaster to stresses such as starvation [45] and the redox-cycling agent paraquat [46], resulting in increased mortality. The present study sought to compare the age-associated changes in survival of both male and female flies following exposure to toxic levels of hydrogen peroxide (H₂O₂). A prior study found an increase in H₂O₂ sensitivity within 35 day old flies [47], arguably ‘middle-age’ in typical control strains of D. melanogaster [48]. Here, we sought to extend this, by comparing sensitivity in 3 day and 60 day old flies. 3 day old flies were selected to represent young [43,44], and 60 day old flies were selected to represent aged animals [49]. The selection of 60 day old flies was influenced by the fact that more than 80% of the population was still alive at this point. Had an even older population been chosen to study, these results might have suffered from a selection bias in favor of the oldest survivors.

Flies were provided food (5% sucrose) without H₂O₂, or with the addition of H₂O₂ at concentrations from 1.0M to 8.0M, on Kimwipes as previously described [43]. Survival was scored every 8 hours. High concentrations of H₂O₂ caused roughly similar toxicity in both young males and females (Fig. 1A and 1B). With age, however, males and females became more sensitive to H₂O₂ toxicity, with survival decreasing by more than 30 hours at the lowest concentration in females, and by more than 60 hours in males (Fig. 1C and 1D). Thus, females appear to be significantly more stress-resistant with age than do males.

Figure 1. Hydrogen peroxide resistance declines with age. (A-D) Males and females of the Actin-GS-255B strain crossed to the w[1118] strain were aged to 3 days or 60 days and then transferred to vials containing kimwipes soaked in 1.0M to 8.0M H₂O₂ dissolved in 5% sucrose solution, and flies were scored as dead when completely immobile, as described previously [43]. (A) 3 day old females. (B) 3 day old males. (C) 60 day old females. (D) 60 day old males.

Hydrogen peroxide adaptation is specific to females and is lost with age

Adaptive homeostasis [8] is the transient physiological activation of the cellular stress response pathways, which allow cells [50,51], tissues [52], and whole organisms [43,44,53] to withstand future oxidative insult for a period of time; the effectiveness of adaptive homeostasis diminishes with age. A prior study assessing H₂O₂-mediated adaptation, found 35-day old (‘middle-age’) females were no longer able to adapt [47]. Here, we sought to expand these findings and test the age-related changes in adaptation between 3 day (young) and 60 day (old) flies. Flies from both sexes and age groups were either not pretreated with H₂O₂, used as controls, or flies were pretreated with 10µM H₂O₂ for 8 hours, and then allowed a further 16 hours to adapt, prior to exposure to a toxic dose of H₂O₂. Previous studies had confirmed similar uptake and ingestion of H₂O₂ by both males and females over the 8-hour period [43]. Young females, showed increased survival upon exposure to the 4.4M toxic dose of H₂O₂ if they were first pre-treated with the 10µM adaptive dose of H₂O₂ (Fig. 2A and Supplementary Table S1). In contrast, aged females no longer exhibited increased survival after pretreatment (Fig. 2C and Supplementary Table S1). Remarkably males, regardless of age or pretreatment, displayed no adaptive response to H₂O₂ (Fig. 2B, 2D, and Supplementary Table S1). Moreover this finding built upon a previous result from our lab, which also demonstrated lack of adaptation explored only in young males [43]. It should be noted that, since aged flies were no longer capable of withstanding the same 4.4M H₂O₂ toxic dose used with young flies (Fig. 1A, 1B, 1C, and 1D), 2.0M H₂O₂ was used as the toxic dose in aged females and 1.0M H₂O₂ as the toxic dose in males.

Figure 2. Adaptation to hydrogen peroxide declines with age in females. Progeny of the Actin-GS-255B strain crossed to the w[1118] strain were aged to 3 days and 60 days prior to H₂O₂ pretreatment. Following recovery, flies were fed H₂O₂ challenge dose: 4.4M in 3 day old males and females, 2M in 60 day old females, and 1M in 60 day old males. (A) 3 day old females showed increased survival following pretreatment. (B) 3 day old males showed no change in survival following pretreatment. (C) With age, 60 day old females no longer show increased survival. (D) 60 day old males show no change in adaptation following pretreatment. Statistical difference in survival (p < 0.05) was calculated using the Log-Rank test. Statistical summary is located in Supplementary Table S1.

Adaptive de novo expression of the 20S proteasome diminishes with age in a sex-dependent manner

Prior studies of mammalian cell cultures have demonstrated that the 20S proteasome is inducible, following oxidative stress [23,51,54], but proteolytic capacity diminishes with nominal age or senescence [55]. To understand the age-related changes in the adaptive expression of the 20S proteasome, 3 day old and 60 day old male and female fruit flies were studied. Following H₂O₂ [0µM-100µM] pretreatment, entire flies were homogenized and the lysates were analyzed by western blot. The blots were incubated with an antibody directed against the D. melanogaster α-subunit of the 20S proteasome, and protein loading was normalized to actin. Total lysates from young females showed robust increase in proteasome expression, however, aged females exhibited no H₂O₂ inducible increase in proteasome levels (Fig. 3A). Importantly, although they were no longer capable of inducing 20S proteasome expression, basal proteasome levels in aged females actually matched, or exceeded, the maximum H₂O₂ inducible induction achieved by young females (Fig. 3A)_. In contrast, total lysates from males showed no change in 20S proteasome expression regardless age or pretreatment (Fig. 3B). To be able to directly compare both age-related and sex-specific differences in basal proteasome expression, young and old female and male lysate were next analyzed by western blot, under identical conditions in a single experiment. Interestingly, aged females reproducibly demonstrated a higher basal proteasome expression compared to young females, whereas, males showed no difference with age (Fig. 3C).

Figure 3. Adaptive de novo expression of the 20S proteasome diminishes with age in a sex-dependent manner. (A,B) Virgin females of the Actin-GS-255B strain were crossed to males of the w[1118] strain and progeny were assayed for 20Sα expression without H₂O₂ pretreatment, or after pretreatment with [10µM or 100µM] H₂O₂. (A) 3 day and 60 day aged females. (B) 3 day and 60 day aged males. All samples were compared to the 3 day old 0µM H₂O₂ controls (C) Basal expression of the 20S proteasome α subunits was measured between 3 day old females, 60 day old females, 3 day old males, and 60 day old males, with samples normalized to the 3 day old female. Western blots were performed in triplicate, normalized to Actin-HRP, and quantified using ImageJ. Error bars denote standard error of the mean (S.E.M) values. * P <0.05 and ** P < 0.01, relative to control using one-way ANOVA. Asterisks color indicates the age of the sample, young female (pink *) and aged female (black *). All statistical significance was calculated relative to the young controls.

Adaptive proteolytic capacity of the 20S proteasome diminishes with age in a sex-dependent manner

Because proteasome expression was inducible in a sex-specific manner, proteolytic activity of the 20S proteasome was also measured. 3 day old and 60 day old males and females were either not pretreated with H₂O₂ (controls), or were pretreated for 8 hours, with either 10µM or 100µM H₂O₂ as adaptive doses, then allowed a further 16-hour interval to adapt, before sample collection commenced. Proteolytic capacity was measured in total lysates by the ability to degrade model fluorogenic peptides, designed to be specific for each of the proteasome’s three proteolytic activities: Z-LLE-AMC for caspase-like/β1 activity, Z-ARR-AMC for trypsin-like/β2 activity, and Suc-LLVY-AMC for chymotrypsin-like/β5 activity. After exposure to increasing concentrations of H₂O₂, young females (pink) showed a marked increase in proteolytic capacity in all three proteasomal subunits (Fig. 4). In contrast, aged females (pink checkered), showed loss of proteasome induction and basal decline in proteasome degradation (Fig. 4A, 4B, and 4C). Increased proteolytic capacity was shown to be proteasome-dependent, as inhibition of proteolytic activity by the proteasome selective inhibitors lactacystin and epoxomycin blocked the inductive response evident in young pretreated females (Supplementary Fig. S1A and S1B). Both, young (green) and aged males (green checkered) showed no increase in proteolytic capacity following H₂O₂ pretreatment, nor a basal change in proteolytic capacity with age (Fig. 4A, 4B, and 4C). As oxidized proteins are the primary substrates of the 20S proteasome [5,6], lysate from whole flies were incubated with ([³H]-labeled) oxidized hemoglobin ([³H]Hb_oxdized) and proteolysis was assessed by increased acid-soluble, low molecular-weight, scintillation counts (from the originally acid-precipitable small ³H-peptides and ³H-amino acids released by proteolysis). Young females, pretreated with H₂O₂, showed increased proteolytic degradation compared to controls, whereas with age, H₂O₂-induced proteolytic capacity was lost. Males, irrespective of pretreatment or age, showed no change in proteolysis following pretreatment (Fig. 4D).

Figure 4. Adaptive proteolytic capacity of the 20S proteasome diminishes with age in a sex-dependent manner. Virgin females of the Actin-GS-255B strain were crossed to males of the w[1118] strain and progeny were assayed for the proteolytic activity of the three catalytic subunits of the 20S proteasome in 3 day-old (red or blue) and 60 day-old flies (checked pattern). Caspase-like activity in (A) Females and males. Trypsin-like activity in (B) Females and males. Chymotrypsin-like activity in (C) Females and males. (D) Proteolytic degradation of oxidized [³H] hemoglobin in flies pretreated with hydrogen peroxide at 3 days and 60 days. Statistical significance for proteolysis of oxidized substrate was compared to young control females. Error bars denote standard error of the mean (S.E.M) values. * P <0.05, ** P < 0.01, *** P < 0.001 relative to control using one-way ANOVA. Asterisk color corresponds to young females (pink *), aged females (black *), young males (green #), and aged males (black #). Statistical significance was calculated relative to the young control females (A-D).

The adaptive expression of the 20S proteasome is age and tissue-dependent in females

Next, it was important to determine whether various tissues displayed differences in the adaptive response of the 20S proteasome. 3 day old and 60 day old flies were either used as controls, or were pretreated with 10µM or 100µM H₂O₂ before body segments (head, thorax, and abdomen) were collected. Expression of the 20S proteasome was assessed in tissue homogenate by western blot, as in Figure 3. Young females showed a marked increase in 20S proteasome expression within the abdomen and the head (Fig. 5A and 5C). However, thorax from young females showed no induction of the 20S proteasome, regardless of pretreatment, (Fig. 5E). With age, the adaptive response visible in the 3 day old abdomen and head were lost (Fig. 5B and 5D), and there was again no change in 20S proteasome expression after pretreatment in thoraxes from 60 day old females (Fig. 5F).

Figure 5. The adaptive expression of the 20S proteasome is age and tissue-dependent in females. Body segments collected from females of the Actin-GS-255B strain crossed to the w[1118] strain were used as controls, or were pretreated with either 10µM or 100µM hydrogen peroxide. (A,B) 20Sα expression in female abdominal tissue following pretreatment. (A) 3 day old. (B) 60 day old. (C,D) 20Sα expression in female head following pretreatment. (C) 3 day old. (D) 60 day old. (E,F) 20Sα expression in female thorax following pretreatment. (E) 3 day old. (F) 60 day old. Western blots were performed in triplicate, normalized to Actin-HRP, and quantified using ImageJ. The bar charts represent the quantification. Error bars denote standard error of the mean (S.E.M) values. * P <0.05, ** P <0.01, *** P < 0.001, relative to the female control using one-way ANOVA. Statistical significance is indicated in young females with pink asterisks (pink *).

The proteasome core is crucial to the rapid removal of damaged proteins from cells and tissues. However, loss of protein turnover has been considered to be a hallmark of aging [55,56], accompanied by diminished basal proteolytic capacity of the 20S proteasome [57,58], though this finding is somewhat controversial [20,53,59–61]. To explore this matter further, tissue lysate (abdomen, head, thorax) from 3 day old and 60 day old females, pretreated with various concentrations of H₂O₂ [0µM-100µM] were assessed for changes in proteolytic capacity (measured fluorometrically) following the addition of fluorogenic peptides specific for the three proteolytic activities of the 20S proteasome core. Similar to the increased proteasome expression (seen in Fig. 5), all three proteasomal proteolytic activities in the abdomen and head exhibited significant increases in 3 day old H₂O₂ pretreated females (Fig. 6A, 6B, 6C, 6D, 6E, and 6F). No such adaptive increase in proteolytic capacities was observed, however, in 60 day old females (Fig. 6A, 6B, 6C, 6D, 6E, 6F). Thoracic tissue showed no change in proteolysis, regardless of age or pretreatment (Fig. 6G, 6H, and 6I). Thus the results of Figures 5 and 6 demonstrate an excellent correlation between the loss of proteasome protein inducibility (Fig. 5) and diminished ability to increase actual proteasome proteolytic activity (Fig. 6) with age and tissue type.

Figure 6. Tissue-specific differences of the adaptive proteolytic capacity, and age-dependent changes in basal activity, of the 20S proteasome in females. Body segments collected from female progeny of the Actin-GS-255B strain crossed to the w[1118] strain were used as controls, or were pretreated with either 10µM or 100µM hydrogen peroxide. Individual proteolytic capacity of the 20S proteasome (caspase/peptidyl glutamyl-peptide hydrolyzing-like activity, trypsin-like, and chymotrypsin-like activity) was measured in the abdomen, head, and thorax. (A-C) Abdomen isolated from 3 day old (pink) and 60 day old (black) females following hydrogen peroxide pretreatment. (A) Caspase-like activity. (B) Trypsin-like activity. (C) Chymotrypsin-like activity. (D-F) Head isolated from 3 day old (pink) and 60 day old (black) females following hydrogen peroxide pretreatment. (D) Caspase-like activity. (E) Trypsin-like activity. (F) Chymotrypsin-like activity. (G-I) Thorax isolated from 3 day old (pink) and 60 day old (black) females following hydrogen peroxide pretreatment. (G) Caspase-like activity. (H) Trypsin-like activity. (I) Chymotrypsin-like activity. Error bars indicate the standard error of the mean (S.E.M) values. * P <0.05, ** P <0.01, *** P < 0.001, relative to the young female control using one-way ANOVA. Statistical significance is shown with pink asterisks (pink *) for young females and black asterisks (black *) in aged females.

Males show no tissue-specific or age-related adaptation of the 20S proteasome

Whole tissue extracts showed no change in proteasome expression or activity in males following any H₂O₂ pretreatment tested. It should also be noted that it was tested at several higher concentrations in males, none of which caused adaptation (data not shown) [43]. Although no proteasome induction was observed in whole tissue extracts from males, we sought to explore if any differences existed within the three major body segments (head, thorax, and abdomen) in D. melanogaster males, because male body segments are not equal in size are not equal in size. Therefore, it was important to test for possible proteasome changes in the three major body segments separately. For these experiments, 3 day old and 60 day old males were pretreated with varying amounts of H₂O₂ [0µM-100µM], and allowed an opportunity to adapt for an additional 16 hours, prior to body segment collection (head, abdomen, and thorax). Following pretreatment, 20S expression in the abdomen, head, and thorax remained unchanged, irrespective of the pretreatment regimen used (Fig. 7A, 7C, and 7E) or age (Fig. 7B, 7D, and 7F). The lack of body segment-specific adaptive responses in males recapitulates the earlier finding of whole male lysate showing no adaptive response to H₂O₂ pretreatment [43].

Figure 7. Males show no tissue-specific or age-related adaptive changes in 20S proteasome protein levels. Body segments were collected from males of the Actin-GS-255B strain crossed to the w[1118] strain that were used as controls, or that were pretreated with either 10µM or 100µM hydrogen peroxide. (A,B) 20Sα expression in male abdominal tissue following pretreatment. (A) 3 day old. (B) 60 day old. (C,D) 20Sα expression in male head following pretreatment. (C) 3 day old. (D) 60 day old. (E,F) 20Sα expression in male thorax following pretreatment. (E) 3 day old. (F) 60 day old. Western blots were performed in triplicate, normalized to Actin-HRP, and quantified using ImageJ. The bar charts represent the quantification. Error bars denote standard error of the mean (S.E.M) values, relative to the male control using one-way ANOVA.

Next, proteolysis was assessed within the young and aged pretreated male tissues (head, abdomen, and thorax). Pretreated lysates were incubated in the presence of fluorogenic substrates, specific for each of the proteasome’s caspase-, trypsin-, and chymotrypsin-like activities. Irrespective of which body segment was studied, there were no adaptive changes in caspase-like activity (Fig. 8A, 8D, and 8G), trypsin-like activity (Fig. 8B, 8E, and 8H), or chymotrypsin-like activity, following H₂O₂ pretreatment (Fig. 8C, 8F, and 8I).

Figure 8. Males show no tissue-specific differences in the adaptive proteolytic capacity of the 20S proteasome, but do exhibit age-dependent changes in proteasomal basal activity. Body segments were collected from male progeny of the Actin-GS-255B strain crossed to the w[1118] strain that were used as controls, or that were pretreated with either 10µM or 100µM hydrogen peroxide. Individual proteolytic capacity of the 20S proteasome (caspase/peptidyl glutamyl-peptide hydrolyzing-like activity, trypsin-like, and chymotrypsin-like activity) was measured in the abdomen, head, and thorax. (A-C) Abdomen isolated from 3 day old (green) and 60 day old (black) males following hydrogen peroxide pretreatment. (A) Caspase-like activity. (B) Trypsin-like activity. (C) Chymotrypsin-like activity. (D-F) Head isolated from 3 day old (green) and 60 day old (black) males following hydrogen peroxide pretreatment. (D) Caspase-like activity. (E) Trypsin-like activity. (F) Chymotrypsin-like activity. (G-I) Thorax isolated from 3 day old (green) and 60 day old (black) males following hydrogen peroxide pretreatment. (G) Caspase-like activity. (H) Trypsin-like activity. (I) Chymotrypsin-like activity.

Age-dependent changes occur in the basal activities of the 20S proteasome in both females and males

Interestingly, both females and males exhibited marked age-dependent changes in basal (non-induced) levels of the three proteasomal activities within the three body segments tested. Specifically, aged females showed basal increases in the caspase-like activity in all three tissues (Fig. 6A, 6D, and 6G), and no significant change in the trypsin-like activity within the abdomen or thorax (Fig. 6E and 6H). In contrast, aged females showed a robust decrease in the chymotrypsin-like activity in all tissues (Fig. 6C, 6F, and 6I). Similarly, basal caspase-like activity showed a robust increase with age within all male tissues (Fig. 8A, 8D, and 8G). As well, male head and abdominal tissue showed marked increases in basal trypsin-like activity with age (Fig. 8B and 8E). And, similar to females, basal chymotrypsin-like activity showed a significant decrease in aged male tissue (Fig. 8C, 8F, and 8I). This dramatic difference between the individual activities of the 20S catalytic core with age (basal elevation in the caspase- and trypsin-like, whereas a dramatic decline in the chymotrypsin-like activity) may hint at the underlying decrease in protein degradation, regardless of the increase in basal protein expression [20]. Furthermore, this finding may point to proteasome’s chymotrypsin-like activity being the potential rate-limiting step in protein degradation by the 20S core.

Adaptation is dependent upon the 20S proteasome

To assess the role of the 20S proteasome in the adaptive response, the ‘Gene-Switch’ system was employed to limit the adaptive increase of the 20S proteasome beta subunit expression [62]. Males of the RNAi strains for the β₁ and β₂ subunits were mated to virgin females of the Actin-255B driver strain and adult progeny were raised in the absence or presence of RU486. After continual RU486 exposure, mRNA levels decreased by at least 50% in both sexes (Supplementary Fig. S2A-S2D). In addition, the adaptive increase in the amount of the 20S proteasome was blunted in females upon RU486 exposure (Supplementary Fig. S2E and S2G). Males showed no change in the amount of proteasome expression following H₂O₂ pretreatment or RU486 exposure (Supplementary Fig. S2F and S2H). Induction of the proteolytic capacity of the individual subunits was also blocked in females cultured in the presence of RU486 following H₂O₂ pretreatment (Fig. 9A and 9C), whereas males showed no change in proteolytic capacity (Fig. 9B and 9D). Similarly, the adaptive response of the 20S proteasome was prevented in females cultured on RU486 and pretreated with H₂O₂ (Fig. 9E and 9G). The purpose of these experiments (Fig. 9 and Supplementary Fig. S2) was not to completely knockdown the entire pool of 20S proteasome, but only to block the transcription/translation-dependent adaptive increase in proteasome expression following hydrogen peroxide pretreatment. Thus, we used RNAi conditions that blocked increased proteasome expression, without depressing basal proteasome protein levels. Using this approach, we found at least a 50% decrease in mRNA in RNAi strains, and within proteasome western blots and activity, we found blockage of the adaptive increase. Although males showed no change in the adaptive response upon pretreatment following the loss of either the β1 or β2 subunit, limiting the adaptive increase of either subunit decreased survival (Fig. 9F and 9H). Importantly, blocking the adaptive increase in proteasome expression in females also blocked the adaptive increase in survival against a subsequent toxic H₂O₂ exposure (Fig. 9E and 9G).

Figure 9. Adaptation is dependent upon the 20S proteasome. Progeny of the Actin-GS-255B strain crossed to the β₁ or β₂ RNAi strains were aged for 5 days in the absence or presence of RU486 prior to H₂O₂ pretreatment. (A-D). The purpose of the experiment was not to completely knockdown the entire pool of 20S proteasome, but only to block the transcription/translation-dependent adaptive increase in proteasome expression following hydrogen peroxide pretreatment. Thus, we used RNAi conditions that blocked increased proteasome expression, without depressing basal proteasome protein levels. Using this approach, we found at least a 50% decrease in mRNA in RNAi strains, and within proteasome western blots and activity, we found blockage of the adaptive increase. After pretreatment, proteolytic capacity of the individual subunits of the 20S proteasome (trypsin-like, caspase/peptidyl glutamyl-peptide hydrolyzing-like activity, and chymotrypsin-like activity) were measured in whole fly lysate. (A-B) Proteolytic capacity in β₁ RNAi flies in the absence (black) “control” or presence (pink in females or blue in males, denoted with “+RU486”) of RU486. (A) Females. (B) Males. (C-D) Proteolytic capacity in β₂ RNAi flies in the absence (black) “control” or presence (pink in females or blue in males, denoted with “+RU486”) of RU486. (C) Females. (D) Males. (E,G) Females of the β₁ and β₂ RNAi strains raised in the absence of RU486 were either not pretreated “control” (black circle) or were pretreated with either 10µM H₂O₂ (grey squares) or 100µM H₂O₂ (grey circles) for 8 hours, followed by a 16-hour recovery prior to H₂O₂ [4.4M] challenge. Females of the β₁ and β₂ RNAi strains raised in the presence of RU486 were either not pretreated “+RU486” (pink triangle) or were pretreated with either 10µM H₂O₂ (pink diamonds) or 100µM H₂O₂ (pink squares) for 8 hours, followed by a 16-hour recovery prior to H₂O₂ [4.4M] challenge. (F,H) Males of the β₁ and β₂ RNAi strains raised in the absence of RU486 were either not pretreated “control” (black circle) or were pretreated with either 10µM H₂O₂ (grey circles) or 100µM H₂O₂ (grey triangles) for 8 hours, followed by a 16-hour recovery prior to H₂O₂ [4.4M] challenge. Males of the β₁ and β₂ RNAi strains raised in the presence of RU486 were either not pretreated “+RU486” (green circle) or were pretreated with either 10µM H₂O₂ (blue square) or 100µM H₂O₂ (blue circle) for 8 hours, followed by a 16-hour recovery prior to H₂O₂ [4.4M] challenge. Statistical difference in survival (p < 0.05) was calculated using the Log-Rank test. Statistical summary is located in Supplementary Table S2.

Decline in 20S proteasome adaptation is accompanied by an accumulation of oxidized proteins

Next, as increased 20S expression and activity was shown to occur in pretreated lysate, it was important to assess whether an inductive increase in the levels of functional 20S proteasome would impact the clearance of oxidized proteins. Indeed, early studies suggested that increased proteasome expression resulted in decreased accumulation of damaged proteins [63]. As well, prior studies in cell culture demonstrated increased protein removal in cells pretreated with an adaptive dose prior to being subjected to a challenge dose of an oxidant [53,64,65].

Having shown the inducibility of both protein expression and activity of the 20S proteasome, following H₂O₂ pretreatment of female flies (Fig. 3 and Fig. 4), it was important to determine if the adaptive increase in proteasome expression would limit damage accumulation upon subsequent exposure to a (much higher) toxic amount of H₂O₂ [4.4M]. Western blotting was used to examine protein carbonylation levels. Protein carbonylation is an irreversible protein modification, resulting from oxidative damage, that often leads to loss of protein function [66]. Pretreatment with a very low, adaptive dose of H₂O₂ [10µM or 100µM] resulted in a much lower accumulation of total protein carbonyls when 3 day old females were subjected to the toxic H₂O₂ dose, compared with matched controls that were not pretreated before H₂O₂ challenge (Fig. 10A). This was presumably because the increased proteasomal capacity of adapted animals allowed for increased rates of degradation of the oxidized, carbonyl containing, proteins generated during exposure to the toxic dose of H₂O₂. However, 60 day old females demonstrated no adaptive decrease in the accumulation of protein carbonyl levels after pretreatment with toxic levels of H₂O₂ (Fig. 10B), following a similar trend in the loss in proteasome induction with age. Similarly, pretreatment with low levels of H₂O₂ failed to prevent the significant accumulation of carbonylated proteins in either young or old male flies (Fig. 10C and 10D), both of whom also failed to adapt to H₂O₂ in any other way measured in this study. Thus, these results for cumulative protein damage (carbonyl accumulation) reflect the changes (or lack of change) in 20S proteasome expression and activity observed in pretreated flies (Fig. 3A, 3B, and Fig. 4).

Figure 10. Decline in 20S proteasome induction is accompanied by an accumulation of oxidized proteins. Carbonyl content was detected with a DNP antibody in progeny of the Actin-GS-255B strain crossed to w[1118] strain following [0, 10, or 100µM] H₂O₂ pre-treatment for 8 hours, followed by a 16-hour recovery to allow for adaption before challenged with H₂O₂ [4.4M] for an additional 24 hours. (A) Carbonyl content showed significant decrease following H₂O₂ pretreatment in 3 day old females. (B) Carbonyl content showed no significant change in 60 day old females, irrespective H₂O₂ pretreatment. (C) Carbonyl content was measured in 3 day old males that were pre-treated with H₂O₂. (D) 60 day old males showed no change in carbonyl content upon H₂O₂ pre-treatment and subsequent recovery. Western blots were performed in triplicate and carbonyl content was normalized to Actin-HRP. Error bars denote standard error of the mean (S.E.M) values. * P <0.05, ** P <0.01, *** P < 0.001, relative to the young control using one-way ANOVA. Statistical significance is shown with asterisks (pink *) in young females.

Loss of proteasome subunits or proteasome regulators impacts lifespan

Previous studies in yeast have shown that loss of the 20S proteasome dramatically decreases lifespan [67]. Additional work has also shown that continual feeding of proteasome inhibitors dramatically reduces lifespan [68]. Moreover, as these results indicate that limiting the induction of the 20S proteasome, blocks adaptation in a sex-specific manner (Fig. 9 and Supplementary Fig. S2), the impact of limiting expression of proteasomal subunits or proteasome regulators upon Drosophila lifespan was addressed. The 20S RNAi β1 and β2 strains were utilized to characterize the impact of decreased 20S expression on lifespan. The CncC and Keap1 RNAi strains were also used. Virgin flies were cultured in the presence or absence of RU486 for the entirety of their lifespan [48,69]. The constitutive feeding of RU486 had no impact on lifespan in control flies (Fig. 11A and 11B). Moreover, the RU486 induced limitation of either β1 or β2 subunit expression (pink in females or blue in males) was detrimental to lifespan in both sexes (Fig. 11C, 11D, 11E, 11F, and Supplementary Table S2). Next, as Nrf2/CncC has been indicated to transcriptionally activate the 20S expression [43,51], the impact of limiting the expression of CncC upon the lifespan was explored. Upon removal of CncC, both males and females showed a marked decrease in lifespan, albeit not as detrimental as upon direct removal of the 20S beta subunits (Fig. 11G, 11H and Supplementary Table S2). Conversely, upon removal of Keap1, which normally sequesters Nrf2/CncC in the cytosol, males and females showed a slight increase in lifespan (Fig. 11I, 11J and Supplementary Table S2), matching earlier findings [70].

Figure 11. Loss of proteasomal subunits or regulators impacts lifespan. (A,B) To control for the effect of RU486 on males and females, lifespan of progeny from the Actin-GS-255B strain crossed to w[1118] strain raised in the absence/control (black line) or presence of RU486, pink line for females and blue line for males. (A) Females. (B) Males. (C-F) Effect of removal of proteasome subunits on life span. The Actin-GS-255B strain was crossed to the β₁ RNAi or β₂ RNAi strains and the progeny were assayed for life span in the absence/control (black line) or presence (pink line in females and blue line in males) of RU486, as indicated. (C) β₁ RNAi females. (D) β₁ RNAi males. (E) β₂ RNAi females. (F) β₂ RNAi males. (G,H) Effect of removal of the Cap-n-collar (CncC)/Nrf2 orthologue upon lifespan. The Actin-GS-255B strain was crossed to the CncC RNAi strain. Male and female lifespan was measured in the absence/control (black line) or presence (pink line for females and blue line for males) of RU486. (G) CncC RNAi female. (H) CncC RNAi male. (I,J) Effect of removal of Keap1 upon lifespan. The Actin-GS-255B strain was crossed to the Keap1 RNAi strain. Males and female lifespan was assessed in the absence/control (black line) and presence (pink line for females and blue line for males) of RU486. (I) Keap1 RNAi female. (J) Keap1 RNAi male. Statistical difference in survival (p < 0.05) was calculated using the Log-Rank test. Statistical summary is located in Supplementary Table S3

Stress resistance of aged (60 day old) flies improves upon continual loss of Keap1

Adaptive homeostasis [8] is necessary for dynamic regulation of protein quality control. As the 20S proteasome is an integral mechanism for protein turnover, loss of its inducibility, in response to mild oxidative stress, is an indicator of protein dysregulation [53]. The loss of 20S proteasome induction with age is potentially attributed to changes in Nrf2/CncC transcriptional regulation, as this transcription factor determines 20S proteasome expression upon oxidative insult [43,51,68,70]. To begin to address this disparity between chronic elevation of the 20S proteasome and its dynamic response to stress, an earlier study explored the impact of chronic over-expression of the C. elegans Nrf2/CncC homolog, Skn-1, throughout the lifespan, and found improved stress tolerance, yet inability to activate the proteasome adaptive response with age [53]. In turn, indicating a dichotomy between chronic basal expression and transient induction of the proteasome. Additionally, RNAi studies against Keap1, the cytosolic inhibitor of CncC, resulted in an elevation of proteasomal subunits [57]. Therefore, the continual removal of Keap1, throughout the lifespan, was employed to determine whether chronic knockdown would be beneficial in mediating an adaptive response in aged flies. Male and female flies of the Keap1 RNAi strain were aged, continually, in the absence (blue) or presence (pink) of RU486. After 60 days, flies were used as controls/no pretreatment, or were pretreated with 10µM H₂O₂, prior to being challenged with 2M (females) or 1M (males) H₂O₂. Unfortunately, adaptation (increased survival) was not restored upon continual knock-down of Keap1 following H₂O₂ pretreatment, compared to flies with normal Keap1 levels (Fig. 12A and 12B). However, comparison between flies fed only the challenge dose, showed overall stress resistance was improved upon constant Keap1 removal in females and males (Fig. 12A and 12B). Nor was this improved stress resistance due to RU486, alone, as control males and females, aged continually in the absence or presence of RU486, showed no difference in stress resistance (Supplementary Fig. S3). Together showing that although adaptation was not restored, stress resistance was improved by lifelong Keap1 knockdown. Therefore, further solidifying the disconnect between the adaptive and basal changes of the stress response.

Figure 12. Hydrogen peroxide stress resistance improves with continual knockdown of Keap1 in aged (60 days) flies. In panels A and B, male and female progeny of the Actin-GS-255B strain crossed to Keap1 RNAi strain were collected and aged to 60 days. (A) Females raised in the absence of RU486 were either pretreated with 10µM H₂O₂ (gray line) “10µM H₂O₂” or not pre-treated with H₂O₂ (black line) “No Additions”. Females raised in the presence of RU486 were either pretreated with 10µM H₂O₂ (red line) “10µM H₂O₂ (+RU486)” or not pretreated with H₂O₂ (pink line) “+RU486”. (B) Males raised in the absence of RU486 were either pretreated with 10µM H₂O₂ (gray line) “10µM H₂O₂” or not pre-treated with H₂O₂ (black line) “No Additions”. Males raised in the presence of RU486 were either pretreated with 10µM H₂O₂ (blue line) “10µM H₂O₂ (+RU486)” or not pretreated with H₂O₂ (green line) “+RU486”. Statistical difference in survival (p < 0.05) was calculated using the Log-Rank test. Statistical summary is located in Supplementary Table S4.

Drosophila strains and culture

Flies were cultured on standard agar/dextrose/corn meal/yeast media at 25°C as described previously [69]. Flies expressing RNAi against two proteasome subunits were obtained from the Vienna Drosophila RNAi center (VDRC, Vienna, Austria), w[1118]; P[GD13913]v35923 (abbreviated B₁ RNAi), w[1118]; P[GD10938]v24749 (abbreviated B₂ RNAi). Strains expressing RNAi against the cap ‘n’ collar transcription factor (an orthologue of the mammalian Nrf2) and dkeap-1 (an orthologue of the mammalian Keap1) were kindly donated by Dr. Dirk Bohman [70,104]. Males from these lines (or w[1118] as a control) were crossed to virgin females of the Actin-‘Geneswitch’-255B (Actin-GS-255B) driver strain [62]. Virgin progeny were collected over 48 hours following eclosion. Adult flies were maintained on media adjusted to final concentration of 160µg/mL mifepristone (RU486, no. M8046, Sigma-Aldrich) or ethanol, and transferred to fresh media every other day.

Preparation of Drosophila

Twenty flies were collected per treatment group and re-suspended in 200µL proteolysis buffer (50mM Tris/HCl, 20mM KCl, 5mM MgAc, 1mM DTT, pH 7.5) and homogenized using an electric pestle. Further lysis was performed by three ‘freeze-thaw’ cycles, which consisted of 5min incubation on dry ice, followed by 5min incubation in water, and vortexed. To remove cuticle fragments, samples were centrifuged for 10,000g for 10min at 4°C. Protein concentration was measured using the Bicinchoninic acid assay (BCA) reducing agent compatible kit (no. 23252, Thermo-Scientific).

Western blots

Protein samples were run on a 4-15% gradient SDS-PAGE gel (no. 4568084, Bio-Rad) and transferred to a PVDF membrane (no. 1620177XTU, Bio-Rad). Mouse monoclonal antibody specific for detection of the α subunit of the 20S core of D. melanogaster origin (1:100 dilution, no. sc-65755, Santa Cruz Biotechnology). The goat polyclonal anti-Actin-HRP antibody, conjugated to horseradish peroxidase (1:1000 dilution, no. sc-1616, Santa Cruz Biotechnology) was used for protein loading control.

Fluoropeptide proteolytic activity assays

5µg of lysate from treatment groups were transferred, in triplicate, to 96-well plates, and 2µM of various proteasome specific-subunit substrates were used: Caspase-like/β₁ activity, Z-LLE-AMC (no. 539141, Calbiochem), Trypsin-like/β₂ activity, Z-ARR-AMC (no. 539149, Calbiochem) Chymotrypsin-like/β₅ activity, Suc-LLVY-AMC (no. 539142, Calbiochem). Plates were incubated at 37°C, and fluorescence readings were recorded every 10 minutes for 4h using an excitation wavelength of 355nm and an emission wavelength of 444nm. Fluorescence units were converted to moles of free 7-amino-4-methylcoumarin (AMC), using an AMC standard curve of known amounts (no. 164545, Merck), with background subtracted. To measure proteolytic inhibition, 20µM of the proteasome inhibitors, lactacystin (no. 80052-806, VWR) or epoxomicin (no. E3652, Sigma-Aldrich), were added directly to lysate, and incubated on plate shaker for 30min at 300rpm, after which, substrate was added.

Preparation of [³H]-labeled substrates

Tritium-tagged oxidized-hemoglobin ([³H]-OxHb) was generated as previously described [6]. Briefly, 5mg of hemoglobin was dissolved in 0.1M Hepes buffer with the addition of 6.6uCi [H3]Formaldehyde and 20mM sodium cyanoborohydride [105]. Mixture was incubated at room temperature on an end-over-end shaker for 1 hr. Hydrogen peroxide (H₂O₂) was added at a final concentration of 5mM, and mixture was rocked for an additional hour. Mixture was dialyzed through a 10,000 MWCO filter (Millipore) at 15,000g for 30min, eluent was discarded, and slurry re-suspended in Hepes buffer. This was repeated for an additional 7 washes to remove unbound [H3]Formaldehyde. Protein content was quantified with BCA assay kit (Thermo-Scientific).

[³H]-labeled substrates proteolytic activity assay

5µg of oxidized [³H] hemoglobin (([³H]-OxHb) was added to 15µg of cell lysate. Samples were incubated on a plate shaker at 300rpm for 2 hours at 37°C. To precipitate any remaining intact protein, 20% trichloroacetic acid and 2% BSA was added. Samples were centrifuged at 13,000rpm for 10min and supernatant was collected and added to 10mL of scintillation fluid. The release of acid-soluble counts were read on a scintillation counter. Background was subtracted and the amount of liberated radiolabel was reported.

RNA isolation and quantitative RT-PCR

RNA isolation was performed following manufacturer’s instructions with slight modification. Flies were homogenized in 500µL TRIzol (no.15596-026, Life Technologies) before an additional 500µL TRIzol were added and incubated at room temperature for 5min. Samples were centrifuged at 12,000g for 10min at 4°C to remove cuticle. 200µL of chloroform was added to the supernatant before samples were shaken for 15 seconds, and incubated at room temperature for 5min. Samples were centrifuged at 12,000g for 15min at 4°C. To the aqueous phase, 500µL of ice-cold 100% isopropanol was added, and samples were incubated at room temperature for 10min. To pellet the RNA, samples were centrifuged at 12,000g for 10min at 4°C and supernatant was decanted. To wash the RNA pellet, 1mL of 70% ice-cold ethanol was added, briefly vortexed before being centrifuged at 7500g for 5min at 4°C. The RNA pellet was dried before re-suspended in DEPC-treated water. RNA concentration was measured using a Nanodrop spectrophotometer (Thermo-Scientific).

For cDNA generation, RNA was reverse transcribed using TaqMan® Reverse Transcription Reagents (no. N8080234, Life Technologies). Quantitative PCR was carried out using iTaq SYBR Green (no. 1725120, Bio-Rad). Amplification of the Beta1 subunit was conducted using the following primer sequence (Forward: 5’ CAGTCATTTCGTGTTCGTGC Reverse: 5’ TCGAACTCCACTGCCATAATG). Amplification of the Beta2 subunit was conducted using the following primer sequence (Forward: 5’AGGTGGTGTTATTCTGGGC Reverse: 5’ TCCGTAGTCATCTCAGTGTCC). Primers for Rp49 were used as an internal control (Forward: 5’CGGATCGATATGCTAAGCTGT Reverse: 5’ GCGCTTGTTCGATCCGTA). Primers were designed using the NCBI Primer-Blast software [106].

Carbonyl content

The protein oxidation detection kit, Oxyblot (no. S7150, Millipore) was utilized to perform immunoblot detection of oxidatively modified proteins. 5µg of protein from young flies (3 days) and aged flies (60 days) was prepared in the same manner as samples for western blot analysis. Carbonyl groups in samples were derivatized to 2,4-dinitrophenylhydrazone (DNP-hydrazone) by reaction with 2,4-dinitrophenylhydrazine (DNPH). Samples were run on a 10% SDS-PAGE gel and transferred to a PVDF membrane for western blot analysis as described previously. Blots were incubated with goat polyclonal anti-Actin-HRP antibody, conjugated to horseradish peroxidase (1:1000 dilution, no. sc-1616, Santa Cruz Biotechnology) to assess protein loading, and detection of carbonylated proteins was measured with the mouse monoclonal anti-DNP antibody (no.MAB2223, Millipore).

Drosophila hydrogen peroxide challenge assays

Flies were aged for 3 days (young) or 60 days (old) before being transferred onto Kimwipes containing 5% sucrose, 24 hours prior to H₂O₂ exposure. Flies were subsequently transferred to vials containing various toxic doses of H₂O₂ [0M-8M] and survival was measured. Flies were scored as dead once they became completely immobile [43,44].

Drosophila hydrogen peroxide pretreatment

24 hours prior to H₂O₂ exposure, flies were transferred to vials containing 5% sucrose on a Kimwipe. For pretreatment (to initiate adaptation), flies were placed for 8 hours in vials without hydrogen peroxide, or in vials that contained 10µM or 100µM H₂O₂. Flies were then transferred to vials containing 5% sucrose for 16 hours to allow for an adaptative response to occur. Afterwards, flies were either immediately collected to assess proteasome expression, or were placed into vials containing a toxic dose of H₂O₂ and survival was measured to assess adaptation.

Lifespan assays

Lifespan assays were performed as previously described [107]. Age-synchronized flies of virgin males and females were collected from culture bottles over a 48 hour period following eclosion. 20 females and 25 males were housed per vial. Flies were transferred to fresh media every other day and deaths were recorded. The mean, median, percent change in the mean and median, and the log-rank p value were calculated using the R statistical software [108]. The values are located in the supplementary lifespan tables.