The Proteasome Beta 5 Subunit is Essential for Sexually Divergent Adaptive Homeostatic Responses to Oxidative Stress in D. melanogaster

Abstract

Our studies center on the physiological phenomenon of adaptive homeostasis in which very low, signaling levels of an oxidant can induce transient expansion of the baseline homeostatic range of protective mechanisms, resulting in transient stress protection. The 20S proteasome is a major element of such inducible defense enzymes against oxidative stress but the relative importance of each of its three proteolytic subunits, β1, β2, and β5, is only poorly understood. We focused the present studies on determining the role of the β5 subunit in adaptation, survival, and lifespan. Decreased expression of the 20S proteasome β5 subunit (with RNAi) blocked the adaptive increase in the catalytic activities of the 20S proteasome response to signaling levels of H₂O₂ in female flies. Similarly, female-specific adaptive increases in survival following H₂O₂ pretreatment and subsequent toxic challenge was blocked. In contrast, direct overexpression of the 20S proteasome β5 subunit enabled an increased 20S proteasome proteolytic response, but prevented further adaptive homeostatic increases through H₂O₂ signaling, indicating there is a maximum ‘ceiling’ to the adaptive response. Males showed no adaptive change in proteasomal levels or activity whatsoever with H₂O₂ pretreatment and exhibited no significant impact upon the other 2 proteolytic subunits of the proteasome. However, chronic loss of the β5 subunit led to shortened lifespan in both sexes. Our exploration of the importance of the 20S proteasome β5 subunit in adaptive homeostasis highlights the interconnection between signal transduction pathways and regulated gene expression in sexually divergent responses to oxidative stimulation.

Graphical Abstract

INTRODUCTION

The highly-conserved 20S proteasome is the primary non-lysosomal mechanism for protein degradation, which is crucial for cellular homeostasis (1). Mutations or impaired functions of subunits of the 20S proteolytic core result in lower rates and efficiency of protein turnover (2, 3), increased sensitivity to oxidative stress (4, 5), and shortened lifespan (6, 7). Declining 20s proteasome activity is associated with increasing age (8-12) and models of accelerated aging (13-17), as poor clearance of oxidized proteins leads to accumulation of misfolded and non-functioning protein aggregates, a hallmark of many age-associated pathologies (18-20).

The ability of cells, tissues, and organisms to transiently up-regulate the activities and components of stress-protective pathways, such as the 20S proteasome, encapsulates the concept of ‘adaptive homeostasis,’ which is a widely-characterized phenomenon, that protects against damage accrual from environmental and physiological stresses (21, 22). Adaptive homeostasis is a concept to describe the short-term induction/activation of various stress-protective enzymes, including de novo synthesis of the 20S proteasome (23), in response to stimulatory, non-toxic, levels of an agent or condition. In turn, rapid, transient elevation of stress-protective mechanisms ensures protection against further oxidative insult, which otherwise may have been highly damaging or even lethal. However, with age, the ability to rapidly induce the adaptive response shrinks, thus aging leads to a compression in the adaptive capacity (6, 11, 13, 24). Moreover, multiple studies have suggested that although baseline levels of stress-protective enzymes actually increase with age, the ability to stimulate a further induction with exposure to stimulatory, non-toxic, levels of an agent or condition is blunted or even lost with age (6, 11, 13, 22, 24).

The four-ringed barrel-shaped 20S proteasome contains three ATP-independent catalytic subunits within its interior. The beta 1 (β1) subunit displays caspase-like activity, the beta 2 (β2) subunit exhibits trypsin-like activity, and the beta 5 (β5) subunit has chymotrypsin-like activity. Although the 20S proteasome is frequently considered in relation to its role as the proteolytic core of the 26S proteasome, abundant evidence demonstrates that it is the predominant form for the degradation of oxidized proteins (25). During oxidative insult, the 19S regulatory components of the 26S complex, which cap each end of the 20S proteasome to form the 26S proteasome, are susceptible to damage and actually dissociate from the 20S core (26, 27). The 19S regulatory components are also not essential for survival during oxidative stress, as evidenced by mutant deletion studies. Moreover, the 20S proteasome is far more efficient at degrading misfolded, oxidized, or denatured proteins arising as a consequence of aging, oxidative stress, or mutation, compared to the ATP-dependent 26S form (28, 29).

The importance of the 20S proteasome becomes immediately evident during periods of proteotoxic stress. Stress-dependent activation/induction of the 20S proteasome occurs due to a myriad of different stressors (30). In response to cellular insult, the 19S regulatory caps dissociate from the 20S proteasome, providing an immediate and available pool for degradation of oxidized proteins (31). Concurrently, Nrf2-mediated de novo synthesis of the 20S proteasome further elevates the pool of 20S complexes available to conduct protein degradation (23). Exposure of mammalian cells (23, 32, 33), C. elegans (24, 34), and D. melanogaster (6, 34-36) to signaling doses of various oxidants leads to rapid and short-term induction in the 20S proteasome which is sex- and age-dependent (34, 35, 37, 38).

Much work has centered on exploring the interplay between the three proteolytic subunits of the 20S proteasome. Early studies in yeast suggested that the β5 subunit may serve as the predominant catalytic subunit (39). Yeast mutants defective in either the trypsin- or caspase-like activities were not associated with substantial declines in protein degradation, whereas mutants lacking chymotrypsin-like activity exhibited significant inhibition of proteolysis, and growth defects (39). In addition, improper processing of the β5 subunit during 20S proteasome assembly led to slower cellular growth and reduced proteolytic capacity (40). Mammalian studies have shown that loss of the β5 subunit results in a shortened lifespan (41) and increased sensitivity to oxidative insults (42). Conversely, overexpression of the β5 subunit has been found to increase stress resistance (43, 44) and upregulation of either the β1 or β5 subunits leads to an orchestrated increase in other 20S subunits (42).

Our prior work in assessing changes in the adaptive response of the Drosophila 20S proteasome revealed the necessity of either the 20S β1 or β2 subunits as being crucial for transient induction of the protein amount and proteolytic capacity of the 20S proteasome, in a female-specific manner (6). Here we sought to explore how modulation of the 20S β5 subunit impacts the adaptive homeostatic response. We utilized two fly strains to determine the impact of either limiting the adaptive increase of 20S β5 protein levels (using an RNAi strain) or exploring whether further adaptive increases are possible upon basal upregulation of the 20S β5 subunit (using an overexpressor strain) in response to the oxidative stimulus of hydrogen peroxide (H₂O₂). Additionally, we explored the indirect impact of changes in β5 levels upon the two other 20S proteolytic subunits (β1 and β2).

RESULTS

The Beta 5 subunit is necessary for the transient adaptive increase in the amount of the 20S proteasome in D. melanogaster

The 20S proteasome is the primary means for degradation of oxidized proteins, mediated by its three catalytic subunits, β1, β2, and β5. Under oxidative stress, both the activity and the amount of the 20S proteasome are transiently upregulated (6, 24, 33-35), ensuring rapid clearance of damaged proteins. However, less is known as to whether each subunit is equally necessary for these transient adaptive responses. Loss of either the β1 or β2 subunit eliminates transient induction of the 20S proteasome, resulting in poor protein clearance and decreased survival (6). Earlier evidence suggested that the β5 subunit may orchestrate the parallel upregulation of the two other catalytic subunits (42), with its loss being detrimental to survival (40). To assess how changes in the β5 subunit impact the adaptive stress response, we explored whether partial knockdown or over-expression of the β5 subunit affects the response of the other two catalytic subunits, β1 and β2 in adult males and females.

The ‘Gene-Switch’ driver system was utilized to either limit β5 levels or explore if it could be further transiently induced in response to an H₂O₂ oxidant signal following over-expression (45). Males of the β5 RNAi strain were mated to virgin females of the Actin-GS-255B driver strain and the adult male and virgin female progeny were propagated in the absence or presence of RU486. Virgin females were used to avoid the stimulatory effect of mifepristone on life span often observed in mated females (46). The continual presence of RU486 led to 50% reduction in mRNA levels of the β5 subunit in females and males (Supplemental Figure 1A, B). In females, levels of the β1 and β2 subunits, with or without RU486, were significantly lower compared to β5, and this was also evident for male β2 mRNA levels (Supplemental Figure 1A, B). Conversely, adult progeny of the male β5 overexpressor strain mated to virgin females of the Actin-GS-255B strain exhibited at least a two-fold increase following propagation on RU486 (Supplemental Figure 1C, D). In females, β5 over-expression led to concurrent transcriptional upregulation of the β2 subunit (Supplemental Figure 1C). Males had lower mRNA levels of the β2 subunit, irrespective of RU486, and showed no β5-dependent change in the β1 or β2 mRNA levels (Supplemental Figure 1D).

We sought to explore how changes in the β5 subunit impacted the de novo induction of the 20S proteasome in response to a non-damaging oxidant stimulus. Prior work has shown that the 20S proteasome is transiently induced in yeast (47), cell culture (13, 33), C. elegans (24, 34), and D. melanogaster (6, 34, 36) in response to a non-damaging oxidant stimulus. Conversely, removal of individual subunits results in blunting of the adaptive response (6, 24). Here, we sought to expand these findings by exploring how changes in the β5 subunit may affect the transient upregulation of the 20S proteasome. Adult male and female progeny of the β5 RNAi strain crossed to the Actin-GS-255B strain were propagated either without or with RU486 to blunt any adaptive increase in 20S expression. Flies were either fed no H₂O₂, used as controls, or were pretreated with stimulatory amounts [10μM or 100μM] of H₂O₂ to assess the transient change in the levels of the 20S proteasome. In the absence of RU486, female flies exhibited higher levels of the 20S proteasome following H₂O₂ stimulation, compared to their non-pre-treated counterparts (Figure 1A). However, the adaptive response was eliminated in females fed RU486 to generate β5 RNAi (Figure 1A). The elimination of the adaptive response was independent of RU486, as age-matched female progeny of the w[1118] strain mated to the Actin-GS-255B strain showed the same adaptive increase of the 20S proteasome following H₂O₂ stimulation (Supplemental Figure 2A,C). Consistent with prior findings (6, 34), males evidenced no adaptive increase in the amount of the 20S proteasome, regardless of H₂O₂ stimulation or RU486 treatment. This male-specific lack of response was independent of RU486 as the same response arose in age-matched males of the w[1118] mated to the Actin-GS-255B strain (Supplemental Figure 2B,D).

Figure 1. The Beta 5 subunit is necessary for the transient adaptive increase in the amount of the 20S proteasome in D. melanogaster females but overexpression of Beta 5 only increases baseline proteasome levels and not inducible levels.

Progeny of the Actin-GS-255B strain mated to (A-D) Beta 5 RNAi or (E-H) Beta 5 overexpressor strains were incubated for 5 days with ethanol (black, denoted with ‘control’) or RU486 (red for females or green for males, denoted with ‘+RU486’) to block only the adaptive H₂0₂ induction of transcription/translation-dependent increase in the 20S proteasome amount, without depleting the basal level. The Beta 5 RNAi prevented the adaptive increase in the H₂O₂ induced 20S proteasome α subunit expression. (A,C) In the absence of RU486 (black), females show an adaptive increase in 20S proteasome amount in response to H₂O₂ pre-treatment, which is lost in the presence of RU486 (red). (B, D) In males, irrespective of the absence or presence of RU486, show no change in the amount of the 20S proteasome, irrespective of H₂O₂ pre-treatment. In the presence of RU486, the Beta 5 overexpressors demonstrated an ‘adaptive ceiling’ effect following H₂O₂ pretreatment. (E,G) In the absence of RU486, females showed an adaptive increase in the 20S α subunit amount. Whereas in the presence of RU486, basal amounts of the 20S α subunit increased, but were not further induced following H₂O₂ pre-treatment. (F,H) Males show no significant change in the 20S α subunit, irrespective of the absence or presence of RU486 or H₂O₂ pre-treatment. Western blots were performed in triplicate, normalized to Actin-HRP, and analyzed using two-way ANOVA. Significance is indicated with * p < 05, relative to 0μM control.

Next, we sought to address if there is a maximum ‘ceiling’ that occurs in the inducibility capacity of the 20S proteasome. Prior work in aged fibroblasts (13) C. elegans (24), D. melanogaster (6), and mice (11) indicated that baseline amounts of the 20S proteasome undergo an age-dependent increase, but proteasome inducibility appears to decline significantly. These findings suggest both a decrease in adaptive capacity and in the maximal amount of proteasome that can be induced: possibly an age-dependent ceiling effect? To explore whether upregulation of basal levels of the 20S β5 subunit reaches a maximum (ceiling) amount, or if the 20S proteasome can be further induced following H₂O₂ stimulation, we used adult progeny from a β5 overexpressor strain mated to the Actin-GS-255B strain fed either no H₂O₂, used as controls, or pretreated with stimulatory amounts [10μM or 100μM] of H₂O₂. In the absence of RU486, females showed an adaptive increase in the amount of the 20S proteasome following H₂O₂ pretreatment. However, in the presence of RU486, the baseline amount of the 20S proteasome increased (to levels matching pre-treated females) but was unable to be further stimulated following H₂O₂ pretreatment (Figure 1E,G). In contrast, males evidenced no change in basal or adaptive amounts of the 20S proteasome, irrespective of H₂O₂ pretreatment, or the absence or presence of RU486 (Figure 1F,H).

The Beta 5 subunit is necessary for the transient adaptive increase in proteolytic capacities of the 20S proteasome in D. melanogaster females

The 20S proteasome represents approximately 1% of the total soluble protein content of eukaryotic cells, with the free 20S proteasome core identified as the predominant proteasomal form within cells (48, 49) and serves to degrade 20% of cellular proteins (50). During oxidative stress, the pool of available pre-existing and de novo synthesized free 20S proteasomes also significantly supersedes the levels of the 26S proteasome and serves as the major protease for removing oxidized proteins (31, 51). Additionally, under non-damaging oxidative stimulating conditions, the proteolytic capacity of the 20S proteasome is transiently upregulated in multiple model systems (6, 13, 24, 26, 33, 34, 36, 52). However, this transient upregulation is significantly restrained following forced reduction of proteasomal catalytic subunits (24, 33), in the presence of proteasome inhibitors (31, 33), or upon inhibition of the Nrf2 transcriptional regulator (23). Moreover, prior studies have demonstrated that, although there is a significant increase in the basal amounts of 20S proteasomes with age, there is an actual decrease in proteolytic capacity (6, 11, 13, 24). Here, we sought to determine how changes in the amount of the 20S β5 subunit may impact overall transient adaptive proteolytic capacity as well as the proteolytic capacities of two other catalytic subunits, β1 and β2.

Adult progeny of the 20S β5 RNAi strain mated to the Actin-GS-255B strain were propagated in the absence or presence of RU486 to limit the proteolytic capacity after a non-damaging oxidative stimulus consisting of H₂O₂ pretreatment. Females pretreated with H₂O₂ exhibited increased proteolytic capacity in all three catalytic subunits (Figure 2A-C). In the presence of RU486, limitation of the 20S β5 subunit shows no adaptive increase in chymotrypsin-like activity following H₂O₂ pretreatment (Figure 2C), and shows a concurrent blunting of the trypsin-like activity (Figure 2B), whereas the caspase-like activity retained increased proteolytic activity in response to H₂O₂ pretreatment (Figure 2A). Adaptive increase of individual proteolytic capacities were induced in control females following H₂O₂ pretreatment, irrespective absence or presence of RU486 (Supplemental Figure 3 A, C, E). Males showed no change in proteolytic capacity following H₂O₂ pretreatment (Figure 2D-F and Supplemental Figure 3 B, D, F).

Figure 2. The Beta 5 Proteasome subunit is necessary for the transient adaptive increase in proteolytic capacity of the 20S proteasome in D. melanogaster females.

Progeny of the Actin-GS-255B strain mated to either (A-F) Beta 5 RNAi or (G-L) Beta 5 overexpressor strains were propagated for 5 days in the absence (black, denoted with ‘control’) or presence of RU486 (red for females and green for males, denoted with ‘+RU486’). Proteolytic capacity of the individual subunits of the 20S proteasome (caspase/peptidyl glutamyl-peptide hydrolyzing-like activity, trypsin-like, and chymotrypsin-like activity) was measured in whole fly lysate. The intent of the experiment was to either block the adaptive increase of the Beta 5 subunit (A-F) or assess if the Beta 5 subunit activity could be further induced (G-L) by H₂O₂ pre-treatment. (A-C) Following H₂O₂ pre-treatment control females (black) exhibited an adaptive increase in the proteolytic capacity of the (A) caspase-like (B) trypsin-like and (C) chymotrypsin-like activities. In the presence of RU486 (red) the adaptive increase in the (B) trypsin-like and (C) chymotrypsin-like activities were blocked in Beta 5 RNAi females. (D-F) Males exhibited no adaptive increase following H₂O₂ pre-treatment, irrespective of RU486. Proteolytic activity in (D) caspase-like and (F) chymotrypsin-like activity decreased in Beta 5 RNAi males in the presence of RU486. (G-I) In control females (black), an adaptive increase in proteolytic capacity occurred in all three subunits after H₂O₂ pre-treatment. In the presence of RU486 (red), overexpression of Beta 5 in females exhibited a basal increase in proteolytic capacity in all three subunits. (J-L) Control males (black) showed no adaptive increase in proteolytic capacity following H₂O₂ pre-treatment. In the presence of RU486, Beta 5 overexpressor males had a basal increase in proteolytic capacity in (L) chymotrypsin-like activity. Proteolytic activity was assessed in triplicate, analyzed using two-way ANOVA. Significance is indicated with * p < 0.05, ** p < 0.01, *** p < 0.001, relative to 0μM control or *** p < 0.001, relative to 0μM +RU486 fed female flies. * p < 0.05, ** p < 0.01, *** p < 0.001, relative to 0μM +RU486 fed male flies.

Next, we sought to address whether transient proteolytic capacity could be further induced following H₂O₂ pretreatment using a 20S β5 overexpressor strain. In the absence of RU486, adult females pretreated with H₂O₂ [10μM and 100μM] exhibited increased proteolytic capacity in all catalytic subunits (Figure 2G-I). Upon overexpression of the 20s β5 subunit, females exhibited basal increase in caspase-like (Figure 2G), trypsin-like (Figure 2H) and chymotrypsin-like (Figure 2I) activities, which were not further induced following H₂O₂ pretreatment. Males had no increase in proteolytic capacity following H₂O₂ pretreatment in any of the individual subunits (Figure 2J-L). In the presence of RU486, males showed increased basal proteolytic capacity in the chymotrypsin-like activity (Figure 2L), with no change in trypsin-like activity (Figure 2K) and a decrease in caspase-like activity (Figure 2J). Adaptive proteolytic capacity was not present in males following H₂O₂ pretreatment (Figure 2J-L), similar to prior findings (6, 34, 36).

The Beta 5 subunit of the 20S proteasome is necessary for adaptation in D. melanogaster females

Adaptation to oxidative stress is evident in studies ranging from in vitro bacterial, yeast, and mammalian cell culture experiments, to studies of multiple model organisms. In pretreatment experiments, exposure to various non-damaging amounts of oxidants or to mild heat leads to improved tolerance against future, semi-lethal heat stress or oxidant challenge, a process characterized as ‘adaptive homeostasis’ (21). In vitro cell culture studies reveal that pretreatment with a non-damaging stimulatory amount of H₂O₂ provides protection, in the form of decreased protein aggregation and improved overall survival, if those same cells are subsequently challenged with (an otherwise) toxic dose of H₂O₂, (31-33, 53, 54). Similarly, in model organisms, pre-treated animals show improved survival (6, 34-36, 55), decreased damage accumulation (6, 24), and in some instances, extended lifespan (56-58). Each of these instances involve the activation and induction of stress-protective enzymes, including the 20S proteasome. In addition, prior work has demonstrated the crucial role of the 20S proteolytic core in mediating the adaptive homeostatic response, because removal of proteolytic subunits results in blunting or loss of the adaptive homeostatic response (6, 24, 33).

In the present study we sought to understand whether the 20S β5 subunit was vital in mediating the adaptive homeostatic response in D. melanogaster. Previous studies have shown the importance of the 20S β1 and β2 subunits for survival in female flies pretreated with non-damaging amounts of H₂O₂ followed with a toxic dose (6). The 20S β5 RNAi strain was crossed to the Actin-GS-255B strain, and the adult progeny were propagated in the absence or presence of RU486 for 5 days and were either not pretreated [0μM], used as a control, or were pretreated with non-damaging stimulatory amounts of H₂O₂ [10μM or 100μM] for 8 hours and allowed an additional 16 hours to adapt before being exposed to a toxic dose of H₂O₂ [4.4M]. Survival times were then recorded until all the flies were dead. As well, based on our prior work, we chose to use the same pretreatment and subsequent H₂O₂ challenge because pretreated control female flies showed an adaptive increase, while control males showed no adaptive increase, which was unaffected by the absence or presence of RU486 (6, 55). Females pretreated with H₂O₂ showed increased survival compared to non-pretreated females when subjected to a toxic dose of H₂O₂ (Figure 3A and Supplemental Table 1). In contrast, females fed RU486 to induce expression of β5 RNAi, exhibited no improvement in survival, irrespective of H₂O₂ pretreatment (Figure 3B and Supplemental Table 1), indicating loss in the adaptive-mediated survival response. Males had no adaptive increase in survival, irrespective of H₂O₂ pretreatment or limitation of 20S β5 subunit expression by RU486 (Figure 3C,D and Supplemental Table 1), similar to earlier findings (6, 36). We also used adult progeny of the 20S β5 overexpressor strain crossed to to the Actin-GS-255B strain, and found females, pretreated with H₂O₂ and propagated with or without RU486, showed increased survival compared to females subjected only to the toxic H₂O₂ challenge (Figure 3E,F and Supplemental Table 1), but without a further increase in overall survival upon overexpression of the 20S β5 subunit, suggesting a limit or ceiling to the maximum induction of the 20S proteasome.

Figure 3. Proteasome Beta 5 subunit is required for H₂O₂ adaptation in D. melanogaster females but overexpression of Beta 5 does not increase the adaptive response.

Progeny of the Actin-GS-255B strain mated to either (A-D) Beta 5 RNAi strain or (E-H) Beta 5 overexpressor strain were propagated in the absence (black, denoted with ‘control’) or presence (pink for females and green for males, denoted with ‘+RU486’) of RU486 for 5 days to explore the impact of either knocking-down (Beta 5 RNAi) or over-expressing (Beta 5 overexpressor) upon the adaptive response following H₂O₂ pre-treatment and challenge [4.4M H₂O₂]. Progeny of the Beta 5 RNAi or Beta 5 overexpressor propagated in the absence of RU486, were either not pretreated [0μM H₂O₂, black circles] or were pretreated [10μM, square symbol & 100μM H₂O₂, triangle symbol] for 8 hours, followed by a 16-hour activation period prior to being exposed to a semi-lethal challenge [4.4M H₂O₂]. (A,E) Pre-treated control females had increased survival. (B) In the presence of RU486 and blockage of the adaptive increase, Beta 5 RNAi females had blunted survival. (C,D) Males, irrespective of RU486 or H₂O₂ pre-treatment showed no change in survivorship following H₂O₂ challenge. (F) In the presence of RU486, overexpression of Beta 5 in females did not further improve survival following H₂O₂ challenge. Statistical difference in survival (p < 0.05) was calculated using the Log-Rank Test. Statistical Summary is in Supplemental Table 1.

Loss of the 20S proteasome Beta 5 subunit shortens lifespan in D. melanogaster males and females but Beta 5 overexpression does not extend lifespan

Not only is the 20S proteasome crucial for the clearance of damaged proteins under oxidative insult but it is even necessary just for normal survival. Loss of either the 20S β1 or 20S β2 subunits has been shown to result in shortened lifespan for adult D. melanogaster(6). In other studies, loss of the 20S β5 subunit not only caused substantial inhibition of proteolytic capacity but was also detrimental to survival in yeast (40) and triggered early larval lethality (7). Chronic feeding of proteasome inhibitors, such as PS-341, which targets the 20S chymotrypsin-like proteolytic activity, also leads to reduced survival (7). Here, we sought to explore how lifelong decreases in the expression level of the 20S β5 subunit may impact animal survival. Adult virgin progeny of the β5 RNAi and β5 overexpressor strains crossed to the Actin-GS-255B strain were aged in the continuous absence or presence of RU486 for the entirety of the adult lifespan (46, 59); control flies were generated by crossing the w[1118] strain to the Actin-GS-255B strain. Lifetime feeding, with or without RU486, had no impact on survival in control flies (Figure 4A,B, and Supplemental Table 2). Continual RU486-β5 RNAi-dependent limitation of 20S β5 subunit expression (red in females and green in males) led to shortened lifespan in both sexes (Figure 4C,D, and Supplemental Table 2). However the RU486-induced overexpression of the 20S β5 subunit resulted in no impact on overall survival (Figure 4E,F, and Supplemental Table 2).

Figure 4. Loss of the 20S proteasome Beta 5 subunit shortens lifespan in D. melanogaster males and females but Beta 5 overexpression does not extend lifespan.

(A,B) The absence or presence of RU486 to induce expression of Proteasome Beta 5 RNAi and suppress proteasome levels has no impact on the survival of male or female control flies. Progeny of the Actin-GS-255B strain crossed to the w[1118] strain in the absence (black, denoted as ‘control’) or presence of RU486 (red for females or green for males, denoted as ‘+RU486’). (C,D) Lifespan curves with the removal of the Beta 5 subunit. (C) Beta 5 RNAi Females (D) Beta 5 RNAi males. (E,F) Lifespan curves with the overexpression (OE) of the Beta 5 subunit. (E) Beta 5 overexpressor Females (F) Beta 5 overexpressor males. Statistical difference in survivorship (p < 0.05) was calculated using the Log-Rank Test. Statistical Summary is in Supplemental Table 2.

DISCUSSION

The present study demonstrates that the 20S proteasome β5 subunit is necessary for the adaptive homeostatic response to H₂O₂ signaling in female flies whereas males do not respond to H₂O₂ signaling. Our work also highlights the observation that there is a biological ‘ceiling’ in terms of the 20S proteasome’s maximal induction and that β5 subunit sufficiency and fully functional proteasomes are crucial elements of a normal lifespan. These findings further support the important role of the 20S proteasome, which is regarded as the leading proteolytic enzyme necessary for removal of oxidized and damaged proteins. Its barrel-like structure is ideal for unfolding and rapidly degrading oxidized proteins in an ATP-independent manner via its three catalytic subunits: β1, β2, and β5 (1). In response to physiological or environmental stimuli, the ubiquitindegrading 26S proteasome is disassembled to increase the immediate and available pool of 20S proteasomes for degradation of oxidatively damaged proteins (31, 52). Concurrently, the master transcriptional regulator Nrf2, is translocated into the nucleus, wherein it orchestrates the transient de novo upregulation of 20S proteasome synthesis (23). Evidence in early-passage cells (23, 32) and young organisms (24, 34-36, 60) have revealed a strong and transient upregulation of the 20S proteasome in response to various oxidative signals.

The transient activation of the 20S proteasome highlights the concept of ‘adaptive homeostasis’. A physiologically relevant concept because for the majority of an organism’s lifespan, it is not subject to a singular toxic insult, but must instead be able to cope with day-to-day fluctuations in physiological and environmental stimulatory signals. This is a key distinguishing factor between the concept of hormesis versus adaptive homeostasis. Whereas hormesis requires a damaging but sub-lethal amount of a toxin or poison to initiate a heightened repair response (61), adaptive homeostasis utilizes physiological and non-damaging levels of signaling molecules or conditions to transiently activate the same pathways (21). Thus, although both lead to increased protection against future damaging insult, the latter represents the day-to-day physiological nuances employed by cells, tissues, and organisms needed for cellular homeostasis.

Indeed, early studies in yeast (47) and mammalian cells (32) showed that young and early-passage cells have the capability to transiently activate the 20S proteasome, and with sufficient time, could be repeated for multiple cycles (32). Yet, in vitro studies of replicative senescence found human lung fibroblasts to have decreased proteolytic capacity in all three subunits (42). Similarly, increasing age (11, 24, 42, 62) or diseaseburden (3, 18), result in a common biological signature of poor protein degradation, accumulation of oxidized proteins, and diminished 20S proteolytic capacity. Chronic loss of the 20S proteasome (6, 63) or inhibition of its proteolytic capacity (7, 41) was also found to be detrimental to survival.

Importantly, multiple aging studies have found an age-dependent rise in the basal amount of the 20S proteasome, but a concurrent decrease in its proteolytic efficiency (6, 11-13, 24). For example 18-month old female mice exposed to aerosolized nanoparticulate matter (‘smog’), exhibited increased basal levels of the 20S proteasome in multiple tissues compared to age-matched controls, but with dampened proteolytic capacity (11). Aged worms (24) and flies (6) also exhibit a basal rise in the amount of the 20S proteasome, again coupled with poor proteolytic efficiency. A striking component of these results is that although basal amounts of the 20S proteasome increased with age, they could not be further induced following H₂O₂ pretreatment, indicating a potential maximal amount of adaptive induction of the 20S proteasome - a physiological ‘ceiling.’ To test this phenomenon in a different way, the β5 subunit in C. elegans was subjected to chronic overexpression which resulted in increased basal amounts of the 20S proteasome, but neither β5 subunit levels nor proteasome could be further induced following H₂0₂ signaling (24), suggesting a finite maximum inducible level. In the present study, we sought to test if a similar outcome occurs in D. melanogaster. Overexpression of the β5 subunit in female flies led to increased basal amounts of the 20S proteasome (compared to controls), but H₂0₂ pretreatment did not further increase the levels of the 20S proteasome. Similarly, basal proteolytic capacity of all three subunits was elevated by chronic overexpression of the β5 subunit, but could not be further induced following H₂0₂ stimulation. These findings provide additional evidence to suggest there is a finite maximum inducible level of the 20S proteasome. More importantly, aging leads to the compression of the 20S adaptive increase, due to the age-dependent rise in the basal levels of the 20S proteasome.

The 20S proteasome is unique due to its three different catalytic components: the caspase-, trypsin-, and chymotrypsin-like activities. Yet a major limitation is deciphering whether a hierarchical proteolytic framework of the 20s catalytic subunits exist or if catalytic capacity is equally dependent upon all three beta subunits. In lower organisms, such as yeast, loss of either the 20S proteasome β1 or β2 subunits was found to have limited to no impact upon proteolytic capacity or survival (64, 65). Yet in D. melanogaster, loss of either the β1 or β2 subunits led to dramatically shortened lifespan (6). Conversely in yeast, loss or inhibition of the 20S proteasome β5 subunit greatly crippled proteolytic function and capacity, and was also found to be lethal (39, 64). In contrast, chronic over-expression of the β5 subunit led to increased stress resistance (24, 43) and lifespan (43, 66). Strikingly, in human lung fibroblasts, constitutive overexpression of the β1 subunit also led to concurrent increase in the amount and activity of the β5 subunit and vice versa (42).

Here, we assessed how changes in the mRNA expression levels of the 20S proteasome β5 subunit impacts the mRNA expression and activity of the other two catalytic subunits. Using an RNAi strain, we found decreased expression of the β5 subunit in both sexes compared to controls. Although basal mRNA expression of the β1 and β2 subunits were lower (especially in females), they were not significantly reduced upon gene-switch activation of the β5 RNAi strain. Conversely, overexpression of the β5 subunit resulted in increased expression of both the β5 and β2 subunits in females, whereas only the β5 subunit was elevated in males. Additionally, basal proteolytic capacity of the β1 and β2 subunits was not significantly impacted in either sex upon the inhibition of the β5 subunit, but was dramatically elevated upon its overexpression, suggesting an interplay between the three subunits.

The findings presented in the current study further support the importance of the 20S proteasome for survival. Lifespan studies using an RNAi strain against the 20S β5 subunit show decreased survival, equally in both sexes, upon chronic downregulation of the β5 subunit. Prior work has shown that loss of either the 20S β1 or β2 subunits leads to shortened lifespan (6) thus, suggesting all three subunits in flies, are necessary for normal survival. Interestingly, overexpression of the β5 subunit did not result in increased survival as previously suggested (66). These differences may be partially due to the use of virgin versus mated females, as earlier work has demonstrated the impact of the gene-switch activator, RU486 alone, can be beneficial to survival in mated females (46). It is important to highlight a limitation in many transgenic or RNAi-based studies due to the caveat of the target enzyme is chronically dampened or over-expressed for the organism’s adult lifespan (67, 68). Moreover, many of these studies lack a physiologically relevant stressor to perturb the system and assess how it responds. Thus, chronic activation (hence overriding any physiologically relevant control) may limit the ability to deduce the importance of a target enzyme in the age-dependent adaptive response. Future work will need to address these shortcomings and identify the optimal time within an organism’s lifespan to capitalize upon any benefit from chronic activation.

An unique component of the current study is the assessment of how changes in the 20S β5 subunit elicit a sex-specific adaptive homeostatic response. Females pretreated with a non-damaging adaptive amount of H₂O₂ showed increased levels and and activity of the 20S proteasome. In contrast males, irrespective of H₂O₂ pretreatment, exhibited no adaptive increase, matching prior findings (6, 34, 36). Additionally, when subjected to a toxic amount of H₂O₂, pretreated females evidenced improved adaptive capacity (measured by increased survival) compared to non-pretreated females. However, in the absence of the β5 subunit the adaptive response was blocked. This is a finding that is consistent with prior work demonstrating that loss of either the β1 or β2 subunit blunted the adaptive increase. Conversely, in the present study, over-expression of the β5 subunit, lead to increased basal levels and proteolytic activity of the 20S proteasome, but no further induction could be achieved by H₂O₂ pretreatment. In a similar vein, upon β5 over-expression pretreated females exhibited increased survival when subjected to a semi-lethal bolus of H₂O₂, but the adaptive improvement in survival showed no further gains compared to pretreated control females. Together, these results suggest a sex-specific ceiling effect in the adaptive homeostatic response. These findings are important because they may help us to better understand the sexually dimorphic health outcomes and survival of males and females (69, 70). Females typically outlive males, which is partly attributable to sex-dependent differences in mitochondrial and nuclear gene interactions, hormone levels, and the ability to cope with oxidative burden (38, 71). Our current work, along with prior studies (11, 34, 35, 37, 38, 72), suggests that sex-dependent differences in the ability to mount a robust adaptive homeostatic response may also be a key factor in the assessment of sex-dependent differences.

MATERIALS AND METHODS

D. melanogaster strains and culture

All flies were cultured in humidified incubators at 25°C, with 12 hour light/dark cycles, on a standard agar/dextrose/corn meal/yeast media as described previously (73). Transgenic and RNAi fly lines with the beta 5 subunit of the 20S proteasome were obtained from the Bloomington Drosophila Stock Center, y[1] w[67c23]; P{y[+t7.7]=Mae-UAS. 6.11Prosbeta5[DP00085] (abbreviated Beta 5 overexpressor), and y[1] sc[*] v[1] sev[21]; P{y[+t7.7] v[+t1.8]=TRiP.HMS00119}attP2 (abbreviated Beta 5 RNAi). Virgin females of the Actin- ‘Geneswitch’−255B (Actin-GS-255B) driver strain were mated to males of the above strains, or w[1118] as a control (74). Virgin progeny were collected over a 48-hour period following eclosion. Adult flies were kept on media with a final concentration of 160μg/mL mifepristone (RU486, catalog # M8046, Sigam-Aldrich) or ethanol as a control (75). Flies were transferred to fresh media every other day.

RNA isolation and quantitative RT-PCR

RNA isolation was performed using 10 flies per treatment group with slight modifications to the manufacturer’s instructions. Flies were homogenized in 500μL TRIzol (catalog #15596–026, Invitrogen). Following homogenization, an additional 500μL TRIzol is added and samples are incubated for 5 minutes at room temperature. Samples were centrifuged at 12,000g for 15 minutes at 4°C. The aqueous phase was transferred to a fresh tube, wherein an additional 500μL of ice-cold 100% isopropanol was added. Samples were subjected to 15 seconds of vigorous shaking before incubating at room temperature for 10 minutes. Samples were centrifuged at 12,000g for 10 minutes at 4°C in order to pellet the RNA. Subsequently, RNA pellet was resuspended in 1mL 70% ice-cold ethanol, briefly vortexed before being centrifuged at 7500g for 5 minutes at 4°C in order to wash the pellet. The supernatant was removed, and dried RNA pellet was resuspended in 20μL DEPC-treated water. RNA concentration was measured using Nanodrop spectrophotometer (Thermo-Scientific).

Isolated RNA was reverse transcribed for cDNA generation using TaqMan® Reverse Transcriptase Reagents (catalog # N808234, Life Technologies). Quantitative PCR was performed using iTaq SYBR Green (catalog # 1725120, Bio-Rad). The following primer sequences were used for amplification. Beta 1 subunit (Forward 5’CAGTCATTTCGTGTTCGTGC Reverse: 5′ TCGAACTCCACTGCCATAATG), Beta 2 subunit (Forward: 5′AGGTGGTGTTATTCTGGGC Reverse: 5′ TCCGTAGTCATCTCAGTGTCC), Beta 5 subunit (Forward: 5’AGTACGGTCACTCTACAAGAGC Reverse: 5’TATCGTAACCGGCCAGCATC), Rp49 Internal Control (Forward: 5′CGGATCGATATGCTAAGCTGT Reverse: 5′ GCGCTTGTTCGATCCGTA). Primers were designed using the NCBI Primer-Blast software (76).

D. melanogaster hydrogen peroxide pretreatment and adaptation assay

24 hours prior to H₂O₂ exposure, flies were transferred to vials containing 1mL of 5% sucrose. Upon treatment initiation, flies were transferred to vials with H₂O₂ [0μM-100μM] for 8 hours (‘pre-treatment’) and then transferred to vials containing only 1mL 5% sucrose for an additional 16 hours (‘adaptive response’). At which point, flies were either snap-frozen for down-stream processing or were transferred to fresh vials containing a semi-lethal dose of H₂O₂ [4.4M], and survival was scored every 8 hours, until all flies were dead to measure adaptation to an oxidant challenge.

Preparation of D. melanogaster

10 flies per treatment group were re-suspended in 200uL proteolysis buffer (50mM Tris/HCl, 20mM KCl, 5mM MgAc, 1mM DTT, pH 7.5) and homogenized using an electric pestle. Subsequently, samples were subjected to three cycles of ‘freeze-thaw’, which consists of 5-minute intervals, wherein samples are placed on dry ice and subsequently incubated in water, to promote further lysis. Next, samples were centrifuged at 10,000g for 10 minutes at 4°C and protein concentration was measured using the Bicinchoninic acid assay (BCA) with reducing agent compatible kit (catalog # 23252, Thermo-Fisher Scientific).

Western Blots

Whole fly lysate (10μg) was run on a 4-15% SDS-PAGE gradient gel (catalog # 4568084, Bio-Rad) for 1 hour at 100V and then transferred to a PVDF membrane at 4°C (catalog # 1620177XTU, Bio-Rad). Blots were blocked in StartingBlock™ (Catalog # 37538, Thermo-Fisher Scientific). The monoclonal antibody against the α-subunit of the 20S core proteasome of D. melanogaster was used (1:100 dilution, ON at 4°C, catalog # sc- 65755, Santa Cruz Biotechnology). The goat polyclonal anti-Actin-HRP antibody, conjugated to horseradish peroxidase (1:1000 dilution, 1 hour at RT, catalog # sc-1616-HRP, Santa Cruz Biotechnology) was used for protein loading control.

Fluoropeptide proteolytic activity assays

Whole fly lysate (5μg) was pipetted, in triplicate, to 96-well black plates, followed by the addition of 2μM of proteasome-specific subunits: Caspase-like/β₁ activity, Z-LLE-AMC (catalog # 539141, Calbiochem), Trypsinlikeβ₂ activity, Z-ARR-AMC (catalog # 539149, Calbiochem), and chymotrypsin-like/β₅ activity, Suc-LLVY-AMC (catalog # 539142, Calbiochem). Fluorescence readings were recorded every 10 minutes for 4 hours at 37°C using an excitation/emission of 355nm/440nm. Fluorescence units were converted to free 7-amino-4-methylcuomarin (AMC) based on an AMC standard curve (catalog # 164545, Merck) with background subtracted.

Lifespan Assay

Age-synchronized virgin flies were collected over a 48-hour period following eclosion. Each vial contained either 20 females or 25 males. The number of flies dead were recorded every other day when flies were transferred to a fresh vial, as previously described (59). The mean, median, and their percent changes, and the log-rank p-value were calculated using the R statistical software (77). The values are in the supplemental lifespan table 2.

Supplementary Material

S1

Supplemental Figure 1. Proteasome Beta 5 RNAi is effective in suppressing proteasome Beta1, 2, and 5 levels in male and female D. melanogaster and Beta 5 overexpressing strains successfully increase Beta 1, 2, and 5 levels in both sexes (A,B) Males of the Beta 5 RNAi strain or (C,D) males of the Beta 5 overexpressor were mated to virgin females of the Actin-GS-255B driver strain and progeny were placed on food in the absence (black) or presence of RU486 (red bars for females or green bars for males) for 5 days to assess changes in the transcriptional levels of the beta 1, 2, and 5 subunits. (A) Beta 5 RNAi Females. (B) Beta 5 RNAi Males. (C) Beta 5 overexpressor Females. (D) Beta 5 overexpressor Males.

Supplemental Figure 2. RU486 does not influence the increase in proteasome levels induced by adaptive doses of H₂O₂ in D. melanogaster females (A-D) The absence or presence of RU486 does not impact the adaptive increase in the 20S α subunit in male and female control flies following H₂O₂ pre-treatment. Female and male progeny of the Actin-GS-255B crossed to w[1118] strain. (A,C) Females, irrespective of the absence (black, denoted as ‘control’) or the presence of RU486 (red for females, denoted as ‘+RU486’) show an adaptive increase in the 20S α subunit following H₂O₂ pre-treatment. (B,D) Males, irrespective of H₂O₂ pre-treatment, or propagated in the absence (black, denoted as ‘control’) or presence of RU486 (green for males, denoted as ‘+RU486) showed no difference in 20S α subunit amount. Western blots were run in triplicate and quantified using Image J. Statistical analyses were completed using two-way ANOVA. Significance is indicated with * p < 0.05, relative to 0μM control or * p < 0.05, relative to 0μM ‘+RU486’ females.

Supplemental Figure 3. RU486 does not influence the increase in proteasomal proteolytic activities induced by adaptive doses of H2O2 in D. melanogaster females (A-F) The absence or presence of RU486 does not impact the proteolytic adaptive increase in the individual 20S proteasome subunits in male and female control flies following H₂O₂ pre-treatment. Female and male progeny of the Actin-GS-255B crossed to w[1118] strain were either not pretreated [0μM] or pretreated with H₂O₂ [10μM or 100μM]. Proteolytic activity of each of the catalytic beta subunits was assessed: caspase/peptidyl glutamyl-peptide hydrolyzing-like activity, trypsin-like, and chymotrypsin-like activity in whole fly lysate. (A,C,E) Females exhibited an adaptive increase in proteolytic capacity following H₂O₂ pretreatment in the absence of RU486 (black, denoted as ‘control’) or presence of RU486 (red, denoted as ‘+RU486’). (B,D,F) Males showed no change in proteolytic activity in any of the subunits irrespective of H₂O₂ pretreatment or in the absence (black, denoted as ‘control’) or presence of RU486 (green, denoted a ‘+RU486).

S2

Supplemental Table 1. Hydrogen peroxide adaptation in Proteasome Beta 5 RNAi and overexpression (OE) Strains of D. melanogaster females with no adaptation occurring in males Two cohorts of females and males were propagated in the absence or presence of RU486 which was used to induce expression of either Beta 5 RNAi or to activate Beta 5 overexpression (OE) in the relevant genotypes. Flies were either not pretreated with hydrogen peroxide or were pretreated with hydrogen peroxide [10μM or 100μM] before being given a toxic challenge dose of hydrogen peroxide [4.4M]. Hours of survival were recorded every 8 hours until all flies were dead. Crosses are in order males x virgin females.

Supplementary Table 2: Lifespan statistical summary. Two cohorts of female and male progeny were transferred every other day onto fresh food and dead flies were recorded. Crosses are in order: males x virgin females.

HIGHLIGHTS.

• H₂O₂ signaling induces adaptive homeostasis resulting in transient stress protection.

• Female flies adapt to H₂O₂ but males do not, revealing sexual dimorphism in stress responses.

• The 20S proteasome is a major element of such inducible defenses.

• The β5 subunit of the proteasome plays a key role in adaptive homeostasis.

• Decreased levels of β5 also decrease lifespan but overexpression does not increase lifespan.

• There is an upper ‘ceiling’ to adaptive homeostasis that limits inducible stress defenses.