Proteasome β5 subunit overexpression improves proteostasis during aging and extends lifespan in Drosophila melanogaster

Abstract

The β5 subunit of the proteasome has been shown in worms and in human cell lines to be regulatory. In these models, β5 overexpression results in upregulation of the entire proteasome complex which is sufficient to increase proteotoxic stress resistance, improve metabolic parameters, and increase longevity. However, fundamental questions remain unanswered, including the temporal requirements for β5 overexpression and whether β5 overexpression can extend lifespan in other species. To determine if adult-only overexpression of the β5 subunit can increase proteasome activity in a different model, we characterized phenotypes associated with β5 overexpression in Drosophila melanogaster adults. We find that adult-only overexpression of the β5 subunit does not result in transcriptional upregulation of the other subunits of the proteasome as they do in nematodes and human cell culture. Despite this lack of a regulatory role, boosting β5 expression increases the chymotrypsin-like activity associated with the proteasome, reduces both the size and number of ubiquitinated protein aggregates in aged flies, and increases longevity. Surprisingly, these phenotypes were not associated with increased resistance to acute proteotoxic insults or improved metabolic parameters.

Figure 1

Overexpression of the β5 subunit of the proteasome causes increased chymotrypsin-like activity without consistent transcriptional upregulation of other proteasome subunits. Strong, ubiquitous, and constitutive overexpression of the proteasome β5 subunit in mated female flies at 10 days post eclosion (dpe) does not result in consistently altered transcript levels of different subunits of the proteasome core (β1, β2, α2, α3) or cap (rpn10, rpn13, rpt2, rpt3, rpn6, rpn11) when normalized to four different reference genes: (a) actin42a [CG12051], (b) α-tubulin [CG1913], (c) RpL32 [CG7939], and (d) eEF1α2 [CG1873]. 5 replicates, 5 flies per replicate. *p < 0.05, ***p < 0.001, t-test. (e) Overexpression of the β5 subunit in mated female flies using a drug-inducible driver produces a moderate increase in β5 transcript level at 10 dpe. Expression of the β5 subunit is not influenced by the presence of the inducing drug (RU486/mifepristone) in driver-only controls. At least 9 replicates, 3 flies per replicate. *p < 0.05, t-test. (f) Induction of β5 subunit overexpression during adulthood in mated female flies results in a significant elevation in 26S and/or 30S proteasome specific chymotrypsin-like activity by 30 dpe. 5 replicates, 5 heads and thoraces per replicate. *p < 0.05, t-test. All error bars represent standard error.

Figure 2

Moderate overexpression of the β5 subunit extends longevity without decreasing feeding. (a) β5 subunit overexpression throughout adulthood results in 10–15% increases in mean lifespan of mated females in two independent replicates (left, at least 230 flies per condition, right, at least 150 flies per condition). p < 0.001, log-rank test. Presence of mifepristone alone caused a significant decrease (b, p < 0.001, log-rank test) or no effect (c, p > 0.05, log-rank test) in survivorship of two different control flies (driver crossed to (b) w¹¹¹⁸ or (c) attp33 background lines). (d) Feeding at 13 dpe was increased in flies overexpressing β5 but not in control flies. 10 replicates, 10 flies per replicate, *p < 0.05, t-test. (e) Increased feeding did not result in significant differences in weight at 13 dpe (11 replicates, 10 flies per replicate, p > 0.05, t-test), but resulted in (f) improved wet starvation resistance (approx. 120 flies per condition, p < 0.001, log-rank test). All error bars represent standard error.

Figure 3

Moderate overexpression of the β5 subunit does not improve stress resistance or metabolic parameters. (a) Under severe hyperoxia conditions (>92% O₂), presence of mifepristone in driver-only control flies caused a significant increase in survival at 10 dpe (~4% increase in mean, p = 0.015), no significant changes at 30 dpe (p = 0.72), and a significant decrease in survival at 45 dpe (~8% decrease in mean, p = 0.006). Under the same conditions, overexpression of the β5 subunit resulted in no significant increase in survival at any time point (no significant difference at 10 dpe, a significant decrease at 30 dpe (~8% decrease in mean, p = 0.008), and no significant difference at 45 dpe). Approx. 120 flies per condition, log-rank test. (b) Under elevated temperatures (37 °C), presence of mifepristone in driver-only control flies caused a decrease in survival over time (no significant difference at 11 dpe, significant decrease at 30 dpe (~7% decrease in mean, p < 0.0001) and 45 dpe (~15% decrease in mean, p = 0.0001). Under the same conditions, overexpression of the β5 subunit resulted in no significant differences in survival at any time point. Approx. 150 flies per condition, log-rank test. (c) Increased β5 expression during adulthood did not correlate with altered CO₂ production (4–5 replicates, 5 flies per replicate, t-test) and (d) did not influence climbing rates (5 technical replicates of 5 experimental replicates, 15 flies per experimental replicate, t-test). All error bars represent standard error.

Figure 4

Increased chymotrypsin-like activity of the proteasome improves proteostasis during aging. (a) Representative images of the indirect flight muscles stained with phalloidin (red) and ubiquitin antibodies (anti-FK2, green) at three different time points (10, 30, and 45 dpe) are shown. (a,b) Overexpression of the β5 subunit results in significantly lower proportions of the tissues being labeled with anti-ubiquitin antibodies. (c,d) Drug treatment alone resulted in no significant differences at all time points. The lower proportion of anti-ubiquitin staining in flies that overexpress β5 was due to a significant reduction in the number of aggregates (e) and in the average size of aggregates (g). Both parameters were unchanged in drug control flies (f,h). At least 7 different hemi-thoraces imaged per time point. *p < 0.05, ***p < 0.001, t-test. All error bars represent standard error. Scale bar represents 10 μm.