SEARCH

SEARCH BY CITATION

Keywords:

  • aging;
  • protein quality control;
  • protein carbonylation;
  • heat-shock proteins;
  • trade-offs;
  • proteasome

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. Author contributions
  9. References

In organisms with a soma–germ demarcation, the germline must be ‘preserved’ such that harmful damage is not transmitted to the offspring. Keeping the progeny free of damage may be achieved by gametes enjoying elevated, and/or more functional, homeostatic maintenance systems. This possibility was approached here by testing whether the soma and maturating oocytes (eggs) dissected from female Drosophila melanogaster in reproductive ages display differential capacities for protein quality control and whether these capacities change during aging and mating. Eggs exhibited a high capacity to prevent protein aggregation, strong capacity for 26S proteasome-dependent degradation and reduced levels of oxidatively damaged (carbonylated) proteins compared to the soma. The capacity to prevent protein aggregation was not affected in either soma or eggs by age and/or mating, while the 26S proteasome capacity declined in the soma but was maintained in the eggs of aged females. However, the levels of carbonylated proteins increased with age in both soma and eggs, and this increase was more pronounced in females allowed to mate continuously. Furthermore, the levels of carbonylated proteins in the eggs of mated flies correlated negatively with the propensity of the eggs to develop into an adult fly. In young flies, mating caused a decrease in 26S proteasome capacity and an increase in protein carbonylation in the soma, but not in the eggs. These results are in line with trade-off theories of aging where aging is considered a consequence of investment in reproduction over somatic maintenance.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. Author contributions
  9. References

Keeping the proteome in a functional state is essential for the survival of cells of all organisms. For this purpose, organisms invest resources in protein quality control systems designed to secure that individual proteins are accurately produced, folded, transported, and compartmentalized (Balch et al., 2008; Koga et al., 2011). The protein quality control network consists of molecular chaperones and various proteases, such as the proteasomes, that recognize and repair damaged proteins and remove aberrant proteins when repair is impossible. The efficiency of this network declines gradually during aging (Koga et al., 2011), and many tissues of aged organisms show signs of failure in the control of protein homeostasis. This includes progressive accumulation of aberrant protein species in the form of oxidized proteins [e.g., carbonyl- and 4-hydroxynonenal (HNE)-modified proteins], advanced glycation end-product (AGE)-modified proteins, protein aggregates, amyloids, and inclusion bodies (Levine, 2002; Nyström, 2005; Friguet, 2006; Demontis & Perrimon, 2010; Jacobson et al., 2010; Koga et al., 2011). Intriguingly, boosting protein quality control functions by genetic means has, in some cases, been shown to postpone the development of age-related protein conformational disorders, such as neurodegenerative disorders (Balch et al., 2008), and even to extend the lifespan of some model organisms (Hsu et al., 2003).

A question of special importance for understanding aging is how the protein quality control network is partitioned between the soma and germline. The separation of germ cells from the soma is argued to be a pivotal event in the evolution of aging. Germ cells, in contrast to somatic cells, must be optimized for youthfulness, whereas the soma, in a sense, is disposable and serves as the ‘steward’ of the germline (Weismann, 1893; Williams, 1957; Kirkwood, 1977). Consequently, it has been argued that the germline, or at least the gametes, must be protected from damage but it is not clear how or whether this is indeed the case. Nevertheless, different principal means of protecting the germline has been hypothesized, including germ cells enjoying elevated, or more functional, homeostatic maintenance systems (Hernebring et al., 2006), an asymmetrical partitioning of damage during gametogenesis, and/or a selection process that rids the organisms of germ cells exhibiting high levels of damage (Medvedev, 1981). In conjunction with such ideas, it has been stressed that there are ‘trade-offs’ between reproductive success and somatic maintenance (Williams, 1957; Kirkwood, 1977; Barnes & Partridge, 2003). At the evolutionary genetic level, survival costs of reproduction are envisioned to arise from alleles with opposing effects of reproduction and somatic survival ‘antagonistic pleiotrophy’ (Williams, 1957; Leroi et al., 2005), whereas the ‘disposable soma theory’ emphasizes that trade-offs are caused by competitive allocation of limited resources; that is, investment of resources into germline homeostasis and reproductive activity is traded for somatic maintenance (Kirkwood, 1977). Many studies have demonstrated that the rate of aging in Drosophila subobscura and D. melanogaster is indeed part of a life-history trade-off with reproductive rate (see e.g., (Smith, 1958; Luckinbill & Clare, 1985; Rose, 1989; Sgro & Partridge, 1999). However, it has been shown also that longevity–reproduction trade-offs can in some cases be uncoupled (Dillin et al., 2002; Partridge et al., 2005).

In this work, we tested whether gametes and soma have different capacity for protein quality control by analyzing protein damage levels and key activities of the protein quality control network including capacity to prevent protein aggregation and proteasome activity. Further, we assessed whether aging and reproduction influence somatic and germline protein quality control. The results are discussed in light of trade-off theories of aging.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. Author contributions
  9. References

Protein damage and quality control differ among body parts

Protein carbonylation serves as a general diagnostic marker for performance of the protein quality control network and increases as a consequence of, for example, diminished oxidative stress defenses, increased translational errors, reduced proteolytic activities, and mutations in genes encoding chaperones and proteases (Dukan et al., 2000; Ballesteros et al., 2001; Fredriksson et al., 2006). Carbonylation of proteins has previously been shown to gradually increase with the age of different organisms, including D. melanogaster (see e.g., (Levine, 2002; Jacobson et al., 2010). To evaluate whether protein damage differs among body parts in aged, but not yet dying female flies with a reproductive potential, a quantitative analysis of the distribution of carbonylated proteins among body parts in the whole fly was performed. Of the total amount of carbonylated proteins in a 35-day-old female, 13% (±1.58) originates from the head, 41% (±2.18) from the thorax, and 45% (±3.27) from the abdomen (Table 1). This distribution was very similar in young (5-day-old) females (not shown) and reflects the relative size and total protein content of the body parts, rather than differential damage levels (Table 1). A qualitative analysis of protein carbonylation in head, thorax, and abdomen of 35-day-old females revealed that the carbonyl content of the proteins in the abdomen was ∼ 2-fold lower than in head and thorax (Fig. 1A,B). This implies that the abdominal protein pool is more protected against such damage. Hence, the capacity for protein quality control was examined in head, thorax, and abdomen of 35-day-old flies. To assess the general chaperone potential, the capacity to prevent heat-induced aggregation of luciferase was calculated as percentage of nonaggregated luciferase after 20 min at 42 °C. A ∼ 2-fold higher aggregation prevention capacity was observed in protein extracts from abdomen compared to head and thorax (Fig. 1C). Measurements of the levels of individual chaperones revealed a complex and body part–specific pattern. The cytosolic small heat-shock protein (Hsp) 26 and the nuclear small Hsp27 were highly abundant in abdomen, compared to head and thorax (Fig. 1D,E,G). A similar, but less pronounced pattern was seen for Hsp70 and heat-shock cognate (Hsc70) proteins by using an antibody that recognizes both Hsp70 and Hsc70s (Fig. 1F,G). On the contrary, high levels of the mitochondrial small Hsp22 were seen in the head as reported previously (King & Tower, 1999), while it was undetectable in thorax and abdomen (Fig. 1G). Another mitochondrial chaperone Hsp60 displayed comparatively similar levels in all three body parts (Fig. 1G). Thus, while the capacity to prevent protein aggregation is clearly elevated in the abdomen, this capacity cannot yet be attributed to a single chaperone.

Table 1.   Distribution of the total amounts of carbonylated proteins among the body parts of 35-day-old flies
Body partPercent carbonylated proteinsPercent total proteinsP-value
  1. All values given are means ± SEM after exclusion of the highest and lowest data point of each group (n = 9). For statistical analyses, the paired two-tailed t-test was used after arcsine transformation of the data.

Head13 (±1.58)8 (±1.12)0.027
Thorax41 (±2.18)31 (±1.39)0.003
Abdomen45 (±3.27)61 (±1.84)0.002
image

Figure 1.  The capacity for protein quality control is elevated in abdomen compared to head and thorax. All panels show analyses of protein extracts from head ‘H’, thorax ‘T’, and abdomen ‘A’ of 35-day-old female flies, correlated with head samples. Error bars represent the standard error of the mean (SEM) when n ≥ 3 and standard deviation (STDEV) when n = 2 (see below). For statistical analyses, the paired two-tailed t-test was used and asterisks mean; *P < 0.05, **P < 0.01, and ***P < 0.001, and n.s. = no significant difference. (A) Relative levels of protein carbonyls per total protein (n = 9). (B) Representative immunoblot showing protein carbonyls (left) and protein pattern, visualized by Coomassie in a corresponding protein gel (right). (C) Relative capacity in protein extracts to prevent heat-induced luciferase aggregation, calculated as percentage of nonaggregated luciferase after 20 min at 42 °C (n = 3). (D) Relative levels of Hsp26 (n = 2), (E) Hsp27 (n = 5), (F) Hsc70 and Hsp70 proteins (n = 2). (G) Representative immunoblots showing Hsp22, Hsp26, Hsp27, Hsp60, and Hsc/p70 in 35-day-old females and, as a control, in heat-shocked Drosophila Schneider 2 cells (S2 cells) as indicated. (H) Relative 26S proteasome capacity in protein extracts measured as the rate of hydrolysis of the fluorogenic peptide Suc-LLVY-AMC. (I) Relative levels of the proteasome 19S regulatory particle subunit PSMC5 and (J) α-subunits of the proteolytic core of the 20S proteasome. (K) Representative immunoblots showing PSMC5 and α-subunits.

Download figure to PowerPoint

The ubiquitin-proteasome system degrades a large proportion of damaged and irreparable proteins (Tonoki et al., 2009). This process works less well in aged organisms, including D. melanogaster that, through a progressive decline, loses ∼ 50% of the proteasome activity in head at 30 days of age (Vernace et al., 2007; Tonoki et al., 2009; not shown). To elucidate whether the proteasome capacity differed in head, thorax, and abdomen of 35-day-old flies, proteasome peptidase activity was assayed in protein extracts by measuring the rate of hydrolysis of the fluorogenic peptide Suc-LLVY-AMC under conditions which favors the 26S proteasome (see Experimental procedures). This analysis revealed that the 26S capacity was higher in abdomen compared to head (> 4-fold) and thorax (> 8-fold; Fig. 1H). The levels of the 19S regulatory particle subunits, p42C (PSMC5; ATPase; Fig. 1I,K), and p39A (PSMD7; non-ATPase; not shown) were essentially identical and showed the same fold difference in each body part as the 26S capacity. Also the levels of six of the seven α-subunits of the 20S proteolytic core correlated with the 26S capacity, although the differences between the levels in abdomen and head (> 2-fold) and thorax (> 3-fold) were less pronounced than for 19S (Fig. 1J,K). Together, these data demonstrate that the abdomen exhibits a superior capacity for protein quality control.

The levels of protein damage are lower in eggs compared to the soma

A large proportion of the female abdomen consists of the reproductive organs including the ovaries with maturating oocytes (eggs). Hence, it is possible that the relatively low levels of carbonylated proteins in the abdomen are because of eggs having low protein carbonyl content. To test this, protein carbonylation was analyzed separately in eggs, removed by dissection from adult, not mated (virgin) 5-day-old (young) and 35-day-old (aged) female flies, and the soma. This revealed that the levels of protein carbonyls (per total protein) are significantly lower (∼ 35%) in eggs from both young and aged females (Fig. 2A) compared to the corresponding soma. Analysis of protein HNE levels confirmed a lower level of protein damage in eggs (Fig. 2B). However, protein carbonylation increased with age of the fly to the same extent in both soma and the eggs (Fig. 2A). Proteomic immunodetection of carbonylated proteins followed by mass spectrometric identification revealed that Hsc70-4, TBA1 (α-tubulin), and TBB1 (β-tubulin) are among major targets of protein carbonylation in eggs and also in the female soma (Fig. 2C). Differently from tubulins, another major cytoskeletal protein actin was only moderately carbonylated relative to its abundance (Fig. 2C). The abundant yolk proteins vitellogenins (VIT) 1, 2, and 3 were not carbonylated when migrating at their normal molecular weights (∼ 46–50 kDa; Fig. 2C). However, a highly carbonylated, low-abundance subfraction of VIT2/3 migrated at a high molecular weight (Fig. 2C), indicating that damaged VIT2/3 may form multimeric aggregates. Taken together, while protein carbonylation is generally lower in maturating eggs compared to the female soma, the specificity of targets seems similar between the two.

image

Figure 2.  Protein damage is lower in maturating oocytes (eggs) compared to the soma. For all panels, eggs dissected from the females are indicated by ‘E’ and the egg-depleted soma by ‘S’. Error bars represent SEM. For statistical analyses, the paired two-tailed t-test was used and asterisks mean; *P < 0.05, **P < 0.01, and ***P < 0.001, and n.s. = no significant difference. (A) Relative levels of protein carbonyls per total protein (n = 5). (B) Relative levels of 4-hydroxynonenal (HNE) per total protein (n = 3). (C) Identification of carbonylated proteins in eggs and egg-depleted female soma using two-dimensional gel electrophoresis and mass spectrometric analysis. The upper panels show representative protein carbonyl immunoblots, and the lower panels the corresponding Coomassie-stained gels. Arrows indicate proteins with a strong carbonylation signal in both eggs and soma; 1: Hsc70-4, 2: α-tubulin, 3: β-tubulin, and 4: the yolk proteins vitellogenins (VIT) 2 and 3 in a high molecular weight constellation. Boxed proteins are abundant, but not much carbonylated; 5: actin, 6: VIT 1, 2, and 3; 7: VIT 3, while the white oval shape marks a set of polypeptides, strongly carbonylated in females, which could not be detected by Coomassie in gels.

Download figure to PowerPoint

Gametes exhibit a markedly elevated capacity for protein quality control

Next, it was tested whether the reduced concentration of protein carbonyls in eggs was associated with an elevated capacity for protein quality control. As depicted in Fig. 3A; a thermolabile luciferase did not aggregate at 42 °C in the presence of protein extracts from eggs, while it did in the presence of somatic extracts. Boiling of the egg extracts strongly reduced (∼ 2/3) their ability to prevent luciferase aggregation (Fig. 3A;b), suggesting the effect being mainly proteinaceous. Incubation of somatic or egg extracts at 42 °C without luciferase had no effect on light scattering (Fig. 3A;c), demonstrating that the assay specifically recorded the aggregation of luciferase rather than proteins of the extracts. The luciferase aggregation prevention capacity was calculated as percentage of nonaggregated luciferase after 20 min at 42 °C (Fig. 3B). This calculation demonstrates that protein extracts from eggs were almost one order of magnitude more efficient in preventing heat-induced luciferase aggregation compared to the soma (Fig. 3B). This aggregation prevention capacity did not change with age (Fig. 3B) and coincided with elevated levels of Hsp26, Hsp27, and Hsp/Hsc70 in eggs (Fig. 3C). These chaperones are known to participate in aggregation management (Morrow et al., 2006). Interestingly, inhibiting the Hsp26/27 function by incubating egg protein extracts with antibodies toward these chaperones prior to assaying luciferase aggregation at 42 °C markedly reduced the capacity to prevent aggregation (Fig. 3D).

image

Figure 3.  Protein quality control capacity is elevated in maturating oocytes (eggs) compared to the soma. For all panels, eggs dissected from the females are indicated by ‘E’ and the egg-depleted soma by ‘S’. Error bars represent SEM when n ≥ 3 and STDEV when n = 2 (see below). For statistical analyses, the paired two-tailed t-test was used and asterisks mean; *P < 0.05, **P < 0.01, and ***P < 0.001, and n.s. = no significant difference. (A) Representative graphs showing heat-induced aggregation of luciferase during 20 min at 42 °C in the absence or presence of native (a) or boiled (b) protein extracts as indicated. In (c), native protein extracts (no luciferase) were incubated for 20 min at 42 °C, demonstrating no background interference on light scattering by the extracts. (B) Relative capacity in protein extracts to prevent heat-induced luciferase aggregation calculated as percentage of nonaggregated luciferase after 20 min at 42 °C (n = 2). (C) Representative immunoblots of Hsp26, Hsp27, and Hsc70/Hsp70. (D) Luciferase aggregation after 20 min at 42 °C in the absence or presence of protein extracts that had been incubated for 30 min at 25 °C with or without antibodies recognizing either Hsp26 or Hsp27 prior to the assay (n = 3). (E) Relative 26S and (F) 20S proteasome capacity in protein extracts measured as the rate of hydrolysis of the fluorogenic peptide Suc-LLVY-AMC (see Experimental procedures for details). (G) Proteasome complexes separated on a nondenaturing gel followed by western blot detection of the 19S subunit PSMD7. (H) Relative amounts of assembled 20S proteasome quantified from western blots of proteins separated on nondenaturing gels using antibodies recognizing six of seven 20S α-subunits (n = 2).

Download figure to PowerPoint

The proteasome capacity was also elevated in eggs (Fig. 3E,F) and was accompanied by elevated levels of mature 26S proteasomes (Fig. 3G). As shown in Fig. 3E, aging attenuated the 26S capacity of the egg-depleted soma of 35-day-old females with 40% compared to the 5-day-old counterparts, whereas the 26S capacity in the eggs was not affected by female age (Fig. 3E). The 20S particle showed no age-dependent changes in egg or soma (Fig. 3F,H).

Proteasome inhibition in fertilized eggs increases protein carbonylation

To determine whether the proteasome is involved in controlling the levels of carbonylated proteins in vivo in mature eggs, newly laid, fertilized eggs (0–1 h old) were permeabilized and submersed for 10 min in Schneider cell medium supplemented with sublethal levels of the proteasome inhibitors MG132 and epoxomicin. Protein carbonylation was analyzed after 1 h of inhibition. Both inhibitors caused a significant (> 1.6-fold) increase in protein carbonyls (Fig. 4A,B) and a concomitant increase in the levels of polyubiquitinylated proteins (Fig. 4B). The levels of the α-subunits of the 20S proteolytic core and the 19S ATPase subunit PSMD7 of the 26S proteasome were not affected (Fig. 4C). These results imply that proteasome activity is involved in keeping protein carbonyl levels at bay in mature eggs/early zygotes.

image

Figure 4.  Proteasome activity is involved in keeping protein carbonyls at bay in mature eggs and early zygotes in vivo. (A) Relative protein carbonylation in fertilized, newly laid eggs (0–1 h of development), treated with the proteasome inhibitors MG132 (‘M’, 800 μm), epoxomicin (‘E’, 1 μm), or DMSO (‘D’; control). Error bars represent SEM with n = 5. For statistical analysis, the paired two-tailed t-test was used and asterisks mean; *P < 0.05 and ***P < 0.001. (B) Representative immunoblots showing protein carbonyls, polyubiquitinylation, and total protein pattern visualized by Coomassie in a corresponding protein gel. (C) Representative immunoblots showing 20S proteasome α-subunits and PSMD7 of the 19S as indicated.

Download figure to PowerPoint

Reproduction affects somatic protein quality control

Female flies allowed to reproduce continuously through constant presence of males, have a shorter lifespan [not shown; e.g., (Smith, 1958; Rogina et al., 2007)] and have a reduced fertility at old age compared to virgin females (Fig. 5A; (Rogina et al., 2007). While ∼ 80% of the eggs laid by 5-day-old females developed into adults, only ∼ 25% of the eggs laid by 35-day-old, continuously mated, females resulted in adult offspring (Fig. 5A). Replacing old males with young (3 days old) at day 28 for continuously mated females had only minor effects on the percentage of eggs that developed into adults. However, ∼ 45% of the eggs laid by females mated for the first time at day 30 resulted in adult offspring. These aged virgins showed as high (or higher; ∼ 87%; n = 577) propensity to mate within 1 h after the addition of young males as young (2- to 5-day-old) virgins (∼ 81%n = 109). The low development into adults of eggs laid by continuously mated females was preceded by a more pronounced, relative increase in protein carbonylation in both the soma and maturating eggs of these females at day 35, compared to virgin flies (Fig. 5B). In addition, the age-related reduction in 26S proteasome capacity seen in egg-depleted virgins was augmented in continuously mated females (Fig. 5C), while there was no such effect in the eggs (not shown). To test whether mating resulted in an immediate impairment of protein quality control, females were mated for 48 h at day 1 or day 30 of age and sampled 5 days after the addition of males. Interestingly, mating of young flies resulted in a 25% reduction in 26S capacity of the egg-depleted soma (Fig. 5D) accompanied by a 20% increase in protein carbonyls (Fig. 5E). The eggs displayed the same levels of protein carbonyls and 26S capacity as eggs dissected from young virgins (not shown). In contrast to young females, mating at day 30 of virgin females did not affect either 26S capacity or protein carbonyl levels in eggs or soma of 35-day-old flies (not shown). However, mating at day 30 reduced the somatic (but not the egg) levels of Hsp26 and Hsp27 with ∼ 50% in 35-day-old females compared to the corresponding virgins (Fig. 5F), while mating at day 1 had no such effects in 5-day-old flies (not shown). Luciferase aggregation prevention capacity was not affected by mating in either eggs or soma in 5- or 35-day-old females (not shown). As depicted in Fig. 5G, mating at day 30 had a negative effect on lifespan (50% survival reduced ∼ 9%). Taken together, there seems to be a clear correlation between somatic protein quality control capacity and egg quality.

image

Figure 5.  Aging and mating have negative effects on somatic protein quality control. For all panels, error bars represent SEM when n ≥ 3 and STDEV when n = 2 (see below). For statistical analyses, the paired two-tailed t-test was used and asterisks mean; *P < 0.05, **P < 0.01, and ***P < 0.001, and n.s. = no significant difference. (A) Percentage of the eggs, laid by ∼ 5-day-old (young) and ∼ 35-day-old (aged) mated females, that developed into adult flies (eggs collected at days 3–7 after the addition of males): Bar 1, eggs laid by young females mated continuously from day 1 by constant presence of sibling males (n = 4; total no of eggs: 1000); bar 2, eggs laid by aged females, mated with 3-day-old (young) males for the first time at day 30 (n = 9; total no of eggs: 2275); bar 3, eggs laid at days 33–37 by continuously mated females (n = 4; total no of eggs: 1038); and bar 4, eggs collected at days 33–37 laid by females mated continuously with sibling males that, at day 28, were replaced by young males (n = 19; total no of eggs: 4578). (B) Relative levels of protein carbonyls in the egg-depleted soma ‘S’ and the eggs ‘E’ dissected from continuously mated 5- and 35-day-old females (n = 4). For comparison, the dashed lines show the protein carbonyl levels in virgin females. (C) 26S proteasome capacity in the egg-depleted soma of 5-day-old (n = 7) and 35-day-old (n = 3) continuously mated females (dashed line: 26S capacity of aged virgins) and (D) in the soma of 5-day-old virgin ‘V’ and mated females ‘M’ (n = 4). (E) Relative levels of protein carbonyls in the soma of egg-depleted 5-day-old virgin ‘V’ and mated ‘M’ females (n = 5). (F) Relative levels of Hsp27 (n = 3) and Hsp26 (n = 2) in egg-depleted 35-day-old females mated for the first time at day 30 ‘M’ and the corresponding 35-day-old virgins ‘V’ (G) Effects on lifespan in virgins ‘V’ and females mated for the first time at day 30 ‘M’. Dashed line indicates 50% survival (day 52, ‘M’ and day 57, ‘V’). Statistical analysis of the 50% survival < 0.001 was performed using paired two-tailed t-test with n = 7.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. Author contributions
  9. References

An age-dependent decrease in protein quality control and increase in protein damage in head and thorax of D. melanogaster have previously been associated with shortened lifespan, decreased mitochondrial function, and loss of muscle strength (Das et al., 2001; Neretti et al., 2009; Tonoki et al., 2009; Demontis & Perrimon, 2010). The data presented in this work expand on these observations and show that in aged, but reproductively capable female D. melanogaster, a large part of the animal’s capacity for protein quality control is assigned to the abdomen, which hosts the reproductive system. Both the capacities for 26S proteasome-dependent proteolysis and those for the prevention of heat-induced protein aggregation were stronger in the abdomen, compared to head and thorax in aged (35-day-old) females. In line with this, the levels of irreversibly damaged (carbonylated) proteins were lower in abdomen compared to head and thorax.

Based upon the idea that the germline should be better equipped to fend of damage, the high capacity for protein quality control and low damage levels in the abdomen may be due to that a rather large part of the female abdomen consists of the ovaries and maturating oocytes (eggs). Indeed, compared to the soma, eggs dissected from both young (5-day-old) and aged (35-day-old) females displayed (i) a superior ability to prevent heat-induced aggregation of luciferase, (ii) a markedly higher capacity for 26S proteasome-dependent degradation of proteins, and (iii) lower levels of carbonylated proteins. In contrast to luciferase aggregation prevention, however, the somatic 26S capacity declined with age (Fig. 3E). This decline is consistent with what has been reported previously for mixed samples of females and males (Vernace et al., 2007).

Despite the high capacity for protein quality control in eggs, protein carbonylation was increased to the same extent (∼ 35%) in both eggs and soma of aged females compared to young females and their eggs. Hsc70-4 and α- and β-tubulins were identified as major carbonylated proteins in both eggs and soma. These proteins (when present) have been identified as oxidatively modified in a variety of organisms, including bacteria (Hsp70), mouse embryonic stem cells, plants, and humans (Johansson et al., 2004; Fredriksson et al., 2005; Hernebring et al., 2006; Sultana et al., 2010), but it is not yet known why they are targets of such damage or what the specific cellular consequences are. The abundant yolk proteins VIT 1, 2, and 3 were carbonylated, only in a multimeric aggregate form, suggesting that carbonylation is associated with protein aberrancies in both eggs and soma.

The age-dependent increase in protein carbonyls in maturating eggs was followed by a decreased propensity of eggs laid by aged, mated females to develop into adults. This is in line with other studies, which demonstrate that the germline, like the soma, displays aging-related features, including increased oxidative damage (Goudeau & Aguilaniu, 2010). For example, the function of D. melanogaster germline stem cells declines upon aging, and this can, in part, be counteracted by overproduction of superoxide dismutase (Pan et al., 2007). Further, in the mouse ovary, oxidative damage increases during aging, and this is accompanied by attenuated expression of antioxidant genes (Lim & Luderer, 2011). It is possible that the proteomes of the soma and the germline age at a similar rate, at least in terms of accumulating oxidative damage, but because of the germline starting off with lower levels of damage, it takes longer for the germline to reach a nonfunctional state. Possibly, it is more economical in terms of total reproductive success and resource allocation, to equip gametes with a high potential for removing potentially harmful damage after fertilization, rather than maintaining a germline entirely free of damage. A role of the proteasome in such a ‘self-rejuvenating’ process is supported by our data because inhibition of the proteasome in newly laid, fertilized eggs resulted in the accumulation of carbonylated proteins. In line with these data, mating was found to have specific effects on protein quality control in the soma, but not in the eggs.

Mating of virgin females at day 30 decreased the lifespan and reduced the levels of Hsp26 and Hsp27. Together with a previous study, showing that overexpression of either hsp26 or hsp27 extends the mean lifespan by ∼ 30%, increases oxidative stress resistance, and reduces fertility (Wang et al., 2004), these results suggest that small Hsps can govern a trade-off between maintenance and reproduction.

Intriguingly, while mating of virgins at day 30 did not further augment the age-dependent somatic decline in 26S proteasome capacity seen in 35-day-old females, mating at day 1 attenuated the 26S proteasome capacity by 25% in 5-day-old females (Fig. 5D). This was accompanied by a 20% increase in protein carbonylation in the soma (but not in the eggs) of young females.

This negative effect of mating on protein quality control in young females may be explained by mating of young females resulting in a surge of egg laying (∼ 10-fold increase), whereas mating of aged females does not. Thus, the large investment in egg production may, according to the ‘disposable soma theory’ (Kirkwood, 1977), leave less resources available for protein quality control in the soma. In support of this reasoning, it has previously been shown that stimulating egg production in young females by adding live yeast to the food increases the sensitivity (propensity to die) of the females to the oxidative agent methyl viologen, while for sterile females, additional yeast does not affect the sensitivity to methyl viologen (Wang et al., 2001). Another possibility is that seminal fluid proteins (e.g., Acps) may have direct proteotoxic effects (Chapman et al., 1995; Barnes & Partridge, 2003) and that the proteome of young females for some reason is more vulnerable than the proteome of aged females. However, mating of 1- or 30-day-old females with males lacking Acp70A [responsible for the increased egg production after mating in young females and implicated in reduced mother survival (Wigby & Chapman, 2005)] had no effect on protein carbonylation, in either soma or eggs (results not shown). This does of course not exclude that other Acps (∼ 130 identified this far by the M. Wolfner laboratory) or seminal fluid proteins affect the protein quality. Moreover, because 13% of the female fly genome responds to mating (McGraw et al., 2004), much research remains before the intricate mechanisms of somatic maintenance and germline preservation/rejuvenation in aging and reproduction can be better understood.

As summarized in Fig. 6, the data obtained in this work suggest that, in comparison with the 5- and 35-day-old female soma, maturating oocytes (eggs) have I) a less damaged proteome, II) a superior capacity for protein quality control, III) a similar age-dependent increase in protein damage (carbonylation), and IV) a potential for protein quality control that is not influenced by age or mating of the female.

image

Figure 6.  Schematic representation of the effects of aging and mating on protein quality control in soma and maturating oocytes (eggs). The data suggest that (A) in comparison with the female soma, maturating oocytes (eggs) have (i) a less damaged proteome (showed in gray), (ii) a superior capacity for protein quality control (showed in green), (iii) a similar age-dependent increase in protein damage (carbonylation) as the soma and (iv) a potential for protein quality control that is not influenced by age or (B) mating of the female. However, both aging and mating have negative and seemingly additive effects on somatic protein quality control. In addition, mating results in a slight increase in the levels of protein carbonyls in the maturating oocytes in both young and aged females.

Download figure to PowerPoint

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. Author contributions
  9. References

Fly stocks and culturing conditions

The D. melanogaster strain W1118 (http://flystocks.bio.indiana.edu) was raised at 25 °C on a 12-h light/12-h dark cycle on standard cornmeal/agar medium. Newly eclosed virgin (V) female flies were separated on ice from males and kept together for 5 or 35 days and sampled by instant freezing (CO2-ice), or allowed to mate for 48 h with sibling (day 1) or 3-day-old males (day 30) after eclosure after which males were removed on ice. Mated (M) females were sampled at day 5 or 35, respectively (corresponding virgins were iced similarly to mated flies). For continuously mated flies, newly eclosed sibling flies were kept together. In egg development experiments, eggs laid overnight on applejuice/agar plates were collected, counted, and put in cornmeal/agar food vials to develop. After 14 days, eclosed flies were counted. To check the propensity to mate, females put together with two 3-day-old males in test tubes with cornmeal/agar food were scored for mating during 1 h. For life spans, seven replicates of a total sum of 827 virgin or 1043 females mated for 48 h at day 30 were counted, put in cornmeal/agar food vials (20 flies/vial), and changed to new vials every second day. The % surviving flies was determined for each experiment after which the data were combined by averaging the survival for each day of the seven biological replicates. Statistical analysis of the 50% survival data point, in average at days 52 (V) and 57 (M) < 0.001, was performed using paired, two-tailed t-test with n = 7.

Immunochemical detection of protein modifications and Hsps levels

Extracts were made from 25 females after removal of eggs, the eggs from these females, 25 μL of embryos, 25 heads, thoraces, or abdomens (with eggs). The extracts were prepared by homogenization on ice using a Teflon pestle either in buffer 1 containing 50 mm Tris–HCl, pH 7.4, 5 mm MgCl2, 5 mm ATP, 1 mm DTT, and 10% glycerol (Vernace et al., 2007) or in a similar buffer 2 containing 25 mm Tris–HCl, 100 mm NaCl pH 7.4, 5 mm MgCl2, 1 mm ATP, 1 mm DTT, and 5% glycerol (Lipinszki et al., 2009). Both buffers were supplemented with the protease inhibitors, such as 0.7 μg mL−1 pepstatin (Sigma-Aldrich, Stockholm, Sweden), 0.5 μg mL−1 leupeptin, 0.5 μg mL−1 aprotinin, and 1 mm PEFAblock (Roche, Bromma, Sweden), which preserve the 26S proteasome assembly and native protein conformations and gave equal results for all analyses performed. Extracts were cleared by centrifugation (16 100 g, 10 min at 4 °C), and protein concentrations were determined using the Bradford assay (Bio-Rad, Sundbyberg, Sweden) in a TCA-precipitated aliquot. The carbonyl groups of the protein side chains were derivatized to 2,4-dinitrophenylhydrazone by reaction with 2,4-dinitrophenylhydrazine (DNPH), as referenced in the study by Levine (2002). For polyubiquitinylation, 4-HNE, and Hsp levels, protein extracts were mixed with NuPAGE LDS sample buffer (Invitrogen, Stockholm, Sweden) with DTT and heated to 70 °C for 10 min. Ten to twenty micrograms of proteins was separated by SDS-PAGE using precast NuPAGE 12% acrylamide Bis-Tris gels (Invitrogen) and blotted onto a nitrocellulose membrane (GE Healthcare, Stockholm, Sweden). For two-dimensional gel electrophoresis, 500–700 μg of proteins was separated by isoelectric focusing using 18 cm pH 3–10 NL strips (GE Healthcare) according to the manufacturers’ instructions after which the strips were derivatized with 10 mm DNPH in 2N HCl for 20 min before separation on SDS-PAGE using 11.5% polyacrylamide gels (Conrad et al., 2001) and blotted onto polyvinylidene difluoride membranes (Millipore, Solna, Sweden). The DNP moiety on the derivatized proteins was detected using a rabbit polyclonal anti-DNP (D9656; Sigma-Aldrich). The other antibodies used were mouse monoclonal anti-polyubiquitin (ab7254-50; Abcam, Cambridge, UK), goat polyclonal anti-4HNE (ab46544, Abcam), rabbit polyclonal anti-Hsp70/Hsc70 (ab69412; Abcam), mouse MCP231 (Enzo Life Sciences, Farmingdale, NY, USA) toward six of the seven 20S proteasome α-subunits and to the 19S; the rabbit polyclonals anti-PSMC5 (ARP38192; AVIVA, San Diego, CA, USA) and PSMD7 (ab11436; Abcam). Antibodies toward Hsp22 and Hsp60 (rabbit polyclonals) and toward Hsp26 and Hsp27 (mouse monoclonals) were kind gifts from the R.M. Tanguay laboratory (Morrow et al., 2004, 2006). As secondary antibodies, goat anti-rabbit, anti-mouse, and donkey anti-goat IRDye 680/800CW (LI-COR, Cambridge, UK) were used. As loading controls for both one- and two-dimensional analyses and to be used for protein identification by mass spectrometry (below), gels parallel to the blotted ones were stained with Coomassie Brilliant Blue. Blots and gels were scanned using the LI-COR Odyssey Infrared scanner. Quantifications were made using the Odyssey quantification software. Quantified data are presented as ratio of antibody signal to Coomassie. Statistical analyses were performed using paired two-tailed t-tests. Error bars represent standard error of the mean (SEM) with n ≥ 4 unless otherwise indicated.

Protein identification

Two-dimensional blots showing carbonylated proteins were stained with Coomassie Brilliant Blue after scanning and used as references to match the carbonylated proteins with the corresponding Coomassie-stained gel protein spots. The matched protein spots were subsequently cut out and used for mass spectrometric analysis. Samples were analyzed using a matrix-assisted laser desorption ionization linear (MALDI) reflection mass spectrometer (Micromass, Manchester, UK) in reflection mode. Tryptic digest (0.5 μL) was mixed with 0.5 μL matrix solution (12 mg mL−1α-cyano-4-OH-cinnamic acid in acetonitrile/water [1:1]–0.1% trifluoroacetic acid) directly on the MALDI probe and allowed to dry at ambient conditions. Monoisotopic mass values from MALDI were used for protein identification by MASCOT against the no database of the National Center for Biotechnology Information. Proteins identified all had a score equivalent to a 95% confidence level or higher.

In vitro proteasome capacity assay

The proteasome chymotryptic activity was assayed according to standard procedures as the rate of hydrolysis of the fluorogenic peptide suc-LLVY-AMC (Bachem, Bubendorf, Switzerland). 20 μg of protein (protein extracts prepared as described, but without protease inhibitors) was incubated with 200 μm suc-LLVY-AMC in a 26S optimizing buffer (50 mm Tris–HCl, pH 8, 5 mm MgCl2, 5 mm ATP, 1 mm DTT, and 10% glycerol) in a total volume of 200 μL. 20S core activity was assayed in the absence of ATP after the addition of 0.02% SDS. AMC fluorescence was read on a spectrofluorometer using 390-nm excitation and 460-nm emission filters using free AMC as a standard (Bachem). Statistical analyses were performed using paired two-tailed t-tests. Error bars represent SEM with n ≥ 4.

In vivo proteasome inhibition

Because proteasome activity is essential for oogenesis, laid eggs were used for this experiment. Eggs/embryos laid during 1 h by young flies on applejuice/agar plates with live yeast paste were collected and dechorionated in 2% NaClO according to standard procedures and permeabilized for 2 min in octane (Sigma-Aldrich) after which the octane was evaporated. Permeabilized embryos were immersed in Drosophila Schneider cell medium (GE Healthcare) supplemented with 8% DMSO (inhibitor solvent; controls) or either of the proteasome inhibitors MG132 (800 μm) and epoxomicin (1 μm) for 10 min, left to recover for 50 min on applejuice/agar plates prior to analyses (∼ 70% of the embryos could hatch into larvae after all treatments). Statistical analyses were performed using paired two-tailed t-tests. Error bars represent SEM with n ≥ 5.

Native PAGE

Protein extracts were prepared as described above to preserve the 26S proteasome assembly prior to separation on precast 4–16% Bis-Tris gels (Invitrogen) according to the manufacturer’s instructions. Gels were blotted onto nitrocellulose membranes, and blots were scanned as described above. To detect the proteasome, the anti-PSMD7 and the MCP231 20S proteasome α-subunits antibodies were used as described above. Experiments were repeated at least two times. Representative results are shown.

Luciferase aggregation prevention capacity

Heat-sensitive luciferase (200 nm; L9506; Sigma-Aldrich) was heat-denatured at 42 °C in 50 mm Tris pH 7.6, 2 mm EDTA, and < 0.001%β-mercaptoethanol in the absence or presence of 10 μg of protein extracts prepared in buffer 2 (see above), extracts that had been boiled for 30 min, or extracts that had been incubated at room temperature for 30 min with anti-Hsp26 or 27. Aggregation of luciferase was determined as light scattering at 340 nm. Maximum aggregation (100%) was obtained after 20-min incubation of luciferase in the absence of protein extracts. Luciferase aggregation prevention capacity was calculated as percentage of nonaggregated luciferase after 20 min at 42 °C. Error bars represent standard deviation (n = 2).

Acknowledgments

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. Author contributions
  9. References

We thank R.M. Tanguay, G. Morrow, A. Uv, M. Wolfner, and K. Kvint for strains and helpful suggestions and discussion. Protein identification was carried out by the Proteomics Core Facility at Gothenburg University. This work was sponsored by an EC Marie Curie Outgoing Fellowship (Acronym: Prot.ox. and Hsps) and a grant from ‘KVVS i Göteborg’ to ÅF and grants from the Swedish Natural Research Council, KAW (Wallenberg Scholar), an Advanced ERC grant (QualiAGE), and the EC (Acronym: Proteomage) to TN.

Author contributions

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. Author contributions
  9. References

ÅF supervised the project, drafted and edited the manuscript, designed and conducted the experiments, collected, quantified, and analyzed the data, raised and maintained flies. EKG designed and performed the aggregation experiments. MH conducted the preparative experiments, performed the HNE analysis, and edited the manuscript. EP gave technical assistance, protein extraction, 2D gels, raised and maintained flies. AJ contributed to Hsp level and protein aggregation experiments. AA contributed to in vivo proteasome activity experiments. TN drafted and edited the manuscript.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. Author contributions
  9. References