Skip to main content
PMC home page
eLife logoLink to eLife
. 2019 Nov 22;8:e49455. doi: 10.7554/eLife.49455

Prolonged ovarian storage of mature Drosophila oocytes dramatically increases meiotic spindle instability

Ethan J Greenblatt 1, Rebecca Obniski 1, Claire Mical 1, Allan C Spradling 1,✉
Editors: Michael B Eisen2, Michael B Eisen3
PMCID: PMC6905857  PMID: 31755866

Abstract

Human oocytes frequently generate aneuploid embryos that subsequently miscarry. In contrast, Drosophila oocytes from outbred laboratory stocks develop fully regardless of maternal age. Since mature Drosophila oocytes are not extensively stored in the ovary under laboratory conditions like they are in the wild, we developed a system to investigate how storage affects oocyte quality. The developmental capacity of stored mature Drosophila oocytes decays in a precise manner over 14 days at 25°C. These oocytes are transcriptionally inactive and persist using ongoing translation of stored mRNAs. Ribosome profiling revealed a progressive 2.3-fold decline in average translational efficiency during storage that correlates with oocyte functional decay. Although normal bipolar meiotic spindles predominate during the first week, oocytes stored for longer periods increasingly show tripolar, monopolar and other spindle defects, and give rise to embryos that fail to develop due to aneuploidy. Thus, meiotic chromosome segregation in mature Drosophila oocytes is uniquely sensitive to prolonged storage. Our work suggests the chromosome instability of human embryos could be mitigated by reducing the period of time mature human oocytes are stored in the ovary prior to ovulation.

Research organism: D. melanogaster

Introduction

Animal oocytes grow extensively to become the largest body cells, but at a few specific stages ovarian follicles can persist in a non-growing state. Following recombination and prophase arrest at the diplotene stage of meiosis, mammalian oocytes within primordial follicles cease demonstrable development to establish the ‘ovarian reserve,’ whose slow utilization over multiple decades in humans determines the duration of female fertility. Both oocytes and granulosa cells within primordial follicles remain able to transcribe and translate genes and are bathed in maternal nutrients, which may assist in maintaining their long period of quiescence. Eventually, arrested oocytes resume growth and develop to their final size while remaining in meiotic diplotene. Shortly before fertilization, oocytes mature, during which meiosis resumes and progresses to an arrest at metaphase I or II (Coticchio et al., 2015; Hughes et al., 2018).

In many species, fully grown oocytes also have a period of quiescence. In a good nutritional environment, which is common in the laboratory but rare and transient in the wild, mated Drosophila females ovulate mature oocytes shortly after they reach their final size. However, Drosophila store metaphase I-arrested oocytes for multiple days if adequate protein or sperm are unavailable, despite a lack of transcription. Analysis of polysomes suggests that stored oocytes maintain protein production, though at a reduced level (Lovett and Goldstein, 1977). Likewise, mammalian oocytes routinely cease transcription and become quiescent sometime after reaching their full size (Abe et al., 2010; Jukam et al., 2017). Oocytes remain transcriptionally inactive until zygotic genome activation at the two-cell stage (mouse) or at the 4-cell stage (human). It has been difficult to study the exact duration and biological significance of mature oocyte storage in mammals because of asynchrony and oocyte to oocyte variation (reviewed in Conti and Franciosi, 2018).

Storing oocytes is generally associated with a significant risk of functional impairment. In humans, where all oocytes are stored to some extent, a portion of oocytes develop meiotic segregation errors including non-disjunction that are the major cause of miscarriage. Past the age of 35, chromosome mis-segregation further increases as reflected in exponentially growing rates of Down's syndrome (Webster and Schuh, 2017). However, studies of in vitro fertilized human oocytes suggest that spindle-related errors in mitotic chromosome segregation during early embryonic cell cycles are frequent even in embryos derived from donor eggs of young women (McCoy et al., 2015). The high frequency of meiotic defects in human oocytes has been explained by the exceptional length of time they spend as arrested primordial follicles after the establishment of sister chromatid cohesion (Chiang et al., 2010; Herbert et al., 2015). In Drosophila oocytes, genetic studies also support a role of cohesion loss in meiotic chromosome instability (Hughes et al., 2018; Subramanian and Bickel, 2008). However, cohesion loss may not fully explain the high frequency of non-disjunction, and evidence in mice supports the proposal that altered microtubule dynamics leading to aberrant spindle formation also contributes to non-disjunction (Nakagawa and FitzHarris, 2017).

Here, we show that mature Drosophila oocytes remain capable of supporting embryonic development for many days while stored in the ovary, providing a system for the molecular genetic analysis of oocyte aging. Oocytes stored only briefly develop with high fidelity. However, as aging continues, completing meiosis successfully following fertilization becomes the major factor limiting oocyte viability. Cytologically detectable spindle defects increase during storage and early developmental arrest gradually become the predominant fate of the resulting embryos. Translation of mRNAs encoding meiotic metaphase and spindle-related proteins decline as part of a general 2.3-fold reduction during aging in the absence of bulk changes to mRNA levels. Our findings show that storage of highly functional mature oocytes in vivo is sufficient to destabilize chromosome segregation, suggesting that the prolonged storage of mature oocytes may be an important source of meiotic chromosome instability in human females.

Results

A general method for studying Drosophila oocyte aging

Drosophila ovaries are organized into highly regulated ovarioles that preserve the order in which follicles develop (Figure 1A). Ovarian biology allowed us to develop a method to obtain mature oocytes that have been stored in the ovary for a known period of time. Newly eclosed virgin female flies with immature ovaries are fed a nutrient-rich yeast paste that stimulates exactly two young follicles per ovariole to develop to maturity past a nutrient-sensitive checkpoint at stage 8 (Figure 1A,B). Withdrawal of the yeast food after 24 hr prevents any additional follicles from passing the checkpoint, however oocyte and maternal physiology are not adversely affected (Drummond-Barbosa and Spradling, 2001). In the absence of mating, the mature eggs are stored in the ovary indefinitely and not replaced, as shown by the continuing absence of post-checkpoint stage 10 oocytes (Figure 1B; Greenblatt and Spradling, 2018).

Figure 1. Oocytes age reproducibly in a temperature-dependent manner.

Figure 1.

Open in a new tab

(A) A schematic of the ovarioles that make up a Drosophila ovary (above) and the structure of a single ovariole (below) showing the germline stem cells (left) and a string of increasingly mature follicles. Stages 8, 10, and 14 follicles are labeled. Prophase I arrested oocytes undergo meiotic resumption at stage 10, progressing to a secondary arrest point at metaphase I which is maintained until ovulation. (B) DAPI stained ovaries from females that were fed 1 day (left), fed 1 day then protein restricted for 1 day (middle), or fed 1 day then protein-restricted 13 days (right). Oocytes are colored as in A, revealing the stable storage of two mature stage 14 oocytes per ovariole. Each mature follicle is about 450 μM in length. (C) Eggs laid per day by females containing stored stage 14 follicles, that were provided with males after 3 (green), 6 (blue), 9 (orange) or 12 (red) days. Mating stimulates deposition of the stored oocytes as fertilized embryos. (D) Aging curves (days) for follicles stored in vivo at 29°C (magenta), 25°C (green) or 20°C (blue). For each point, stored oocytes were recovered as in (C) and the hatch rate determined (N > 100). (E) Protein content of mature oocytes that were unstored, stored in vivo for 1 day or for 13 days (‘oocyte age’). Protein restriction of mothers does not affect the protein content of stored mature oocytes. (F) Hatch rate of embryos developing from fresh mature (stage 14) oocytes from 5-day-old females, mature oocytes stored 14 days (during days 2–16) from 16 day-old females, or fresh mature oocytes from 20-day-old females. Oocyte age during storage, but not maternal age, is associated with reduced hatch rate. Error bars in (C), (E), and (F) denote SD.

We measured the stability of stored oocytes over time by placing females with held, mature oocytes of known age with males, which stimulates the rapid fertilization and deposition of their mature oocytes following mating (Figure 1C). We found that while oocytes stored for 5 days or less support development to hatching at high rates (90–96%), more extensively stored oocytes generate embryos with lower levels of hatching. Loss of developmental capacity follows highly reproducible sigmoidal kinetics over the course of 1–4 weeks depending on temperature (Figure 1D). At 29°C, 25°C or 20°C, 50% of eggs fail to hatch after about 7, 12, or 23 days of storage, respectively (Figure 1D). Thus, by appropriately feeding newly eclosed females and delaying mating, we are able to obtain an abundant supply of oocytes for study of known age that are at a known point on an aging curve. Stored mature oocytes are thought to be impermeable to macromolecules; they showed no visible changes and retained the same protein content during 13 days of storage in the ovaries of protein-restricted females (Figure 1E). Their eventual loss of developmental capacity correlated with the duration of storage (intrinsic aging), but was unaffected by maternal age (Figure 1F).

Protein translation declines during oocyte aging

In order to investigate the gene products actively translated by mature oocytes as they aged, we isolated 2, 8, and 12-day-old stage 14 follicles (at 25°C), whose hatch rates are 96%, 89%, and 44%, respectively, removed their follicle cells (see Methods), and performed mRNA-seq and ribosome profiling in triplicate. We added a constant amount of ovarian lysate from D. pseudoobscura (Dpse) as a spike-in to allow us to measure changes in translation globally as well as at the single gene level. Dpse is sufficiently diverged from D. melanogaster (Dmel) that > 97% of 30 nucleotide ribosome footprints contain at least one polymorphism that can be distinguished by sequencing (see Materials and methods). Ribosome profiling experiments were highly reproducible (Figure 2—figure supplement 1A–F). Total mRNA levels, as determined by the ratio of Dmel to Dpse sequencing reads, did not change significantly between 2 and 12 days of oocyte age (Figure 2A). Not only the amount of mRNA, but mRNA composition was also unchanged during aging. RNA-seq transcripts per million (TPM) values from oocytes at day 2 or day 12 of aging correlated very strongly (Figure 2B). By contrast, bulk translation in 8 and 12-day-old oocytes was reduced to 59% and 43%, respectively, of levels in 2 day oocytes (Figure 2C). Thus, the overall level of translation declines significantly as oocytes age in vivo over a ten-day period, due to widespread declines in translational efficiency (Figure 2D).

Figure 2. Stability of mRNA levels and translation in stored stage 14 follicles.

(A) Total mRNA per mature follicle normalized to spike-in control after oocyte aging for 2 (blue), 8 (orange) or 12 (purple) days. Differences are not significant (Student’s t-test, p=0.78, p=0.48, and p=0.49 for 2 vs. 8, 8 vs. 12, and 2 vs. 12, respectively). (B) Log-Log plot showing high correlation (R2 = 0.97) of mRNA-seq values (TPM, transcripts per million) from stage 14 follicles stored at 25°C for 2 days (96% viability) vs 12 days (44% viability). Equal expression (dashed red line). (C) Total translation levels per mature follicle normalized to spike-in control were compared between oocytes aged for 2 (blue), 8 (orange) or 12 (purple) days. Differences are significant as shown (Student’s t-test). (D) Volcano plot showing global reduction in translation efficiency in 12 day versus 2 day oocytes. (E) R-GFP serves as a reporter of nascent protein levels. Steady state levels of N-end rule proteasomal substrate R-GFP are greatly decreased as compared to stable control M-GFP, consistent with rapid degradation of R-GFP but not M-GFP. Scale bar = 50 μm. (F) Plot showing that R-GFP levels decrease ~30% during oocyte aging, consistent with reduced translation. Error bars in (A) and (C) denote SD.

Figure 2.

Figure 2—figure supplement 1. Reproducibility of ribosome profiling data.

Figure 2—figure supplement 1.

(A–D) Log-log plots showing high reproducibility of individual gene ribosome profiling TPM values from replicate ribosome profiling experiments of 2 day (A) 8 day (B) and 12 day (C) oocytes and 0–2 hr embryos from non-aged oocytes (D). (E) Log-log plot showing that translation of individual genes is less correlated when comparing 12 day to 2 day oocytes versus replicate experiments. (F) Principle component analysis (ClustVis) showing tight clustering of replicate ribosome profiling experiments.
Open in a new tab

A reduction in ribosome footprints might theoretically result from an increase in ribosome elongation rates with age rather than a decrease in ribosome initiation. While decreased initiation would lead to decreased overall translation levels, increased elongation would lead to the opposite outcome (increased overall translation). We reasoned that measuring the levels of nascent proteins would allow us to infer relative translation rates and discriminate between these possibilities. We generated Drosophila lines expressing the proteasomal substrate R-GFP, which is rapidly degraded by the N-degron pathway (Dantuma et al., 2000). Due to its reduced stability, R-GFP staining in the ovary is weaker than the corresponding stable GFP control (‘M-GFP’) (Figure 2E). We then characterized the levels of R-GFP expression in oocytes stored for 1 day or for 12 days, when the level of footprints has declined to ~50% overall of starting levels, with some variation from transcript to transcript. We found that the level of R-GFP fluorescence was reduced by an average of ~30% – as expected if less protein is produced in aged oocytes due to decreased translation initiation (Figure 2F). These data strongly support our interpretation that translation levels decline during oocyte aging.

Aging reduces the translation of all classes of genes

In order to determine whether declining translation in aging oocytes was a general phenomenon or preferentially affected a subset of genes, we analyzed changes to translation and mRNA levels from individual genes. The translation of germline expressed genes such as Hsp26, vtd, and me31B, but not their mRNA levels, reproducibly declined, much like total translation (Figure 3A). The changes in translation were not caused by premature egg activation. Genes such as CycB, CycA, and bora, whose translation is substantially upregulated at the start of embryogenesis did not increase in aged oocytes (Figure 3B). Rather, translational efficiency declined globally (Figure 2D, Supplementary file 1).

Figure 3. Translation is reduced globally during oocyte aging.

Figure 3.

Open in a new tab

(A) Relative read depths from replicate ribosome footprinting and mRNA-sequencing experiments of Hsp26, vtd, and me31B from 2-, 8-, and 12-day-old oocytes. Data were normalized to spike-in controls. Relative translational efficiency (right panel) falls 2.5–5 fold between day 2 and day 12. (B) Read depths and translational efficiency values are shown as in (A) for genes preferentially translated in embryos CycB, CycA, and bora from 2-day oocytes, 12-day oocytes, and 0–2-hr embryos from non-aged oocytes. (C) Heat maps showing reduced translation, but similar mRNA levels, of genes of various GO categories in 12-day oocytes as compared to 2-day oocytes. Gene classes of interest are indicated on the left, along with specific genes on the right; see Flybase for information on each gene (http://flybase.org). Error bars in (A) and (B) denote SD.

We grouped genes in several functional categories with potential relevance to oocyte aging and compared changes in the mRNA levels and translation levels (Figure 3C). Widespread reductions were seen among genes with the GO categories spindle assembly checkpoint, meiotic/mitotic spindle organization, chaperone-dependent protein folding, electron transport chain, and mRNA binding proteins (Figure 3C). For each category, the location of several well-known genes is indicated (Figure 3C), including genes shown previously to be dose-sensitive for chromosome stability, such as sub, ncd, nod, and SMC1 (Knowles and Hawley, 1991; Moore et al., 1994; Subramanian and Bickel, 2008; Zhang et al., 1990).

A small subset of mRNAs are translationally upregulated in arrested mature oocytes

Mature oocytes stockpile mRNAs, some of which are translated preferentially in oocytes and some of which are translated preferentially during early embryogenesis (Kronja et al., 2014a). We reasoned that some genes preferentially translated in arrested oocytes may function to maintain viability during prolonged arrest. In order to identify genes that are preferentially translated in arrested oocytes, we performed ribosome profiling and mRNA-seq experiments on 0–2 hr embryos with Dpse ovarian extract spike-in and compared these data to those from arrested oocytes. Total normalized ribosome footprints increased 2.7-fold in developing embryos as compared to mature oocytes stored for two days (Figure 4A). While total translation is lower in arrested oocytes than early embryos, we identified a small subset of candidate oocyte ‘pilot light genes’ (243, representing 6.0% of oocyte mRNAs) that are preferentially translated during oocyte arrest (Figure 4B; Supplementary file 2).

Figure 4. Translation of a small group of genes is boosted during oocyte arrest.

(A) Normalized total translation levels from ribosome profiling in 0–2 hr embryos compared to oocytes stored for two days. (B) Plot showing that most genes are translated at higher levels in the 0–2 hr embryo than the 2 day stored oocyte. Examples of the 243 more highly translated ‘pilot light genes’ are labeled in orange. (C) The hatch rate of stored oocytes from wild type (blue), Df(sHSP)/+ (green), or Df(sHSP)/Df(sHSP) (purple) females after indicated storage period (‘oocyte age’). Deletion of Hsp26 and Hsp27 accelerated the rate of decline during storage (N = 3 at each point). (D–G) GO analysis (PANTHER) of genes with significantly (p<0.01, Student’s t-test) increased (D,F) or decreased (E,G) translation in 2- day-old mature oocytes compared to 0–2-hr embryos (D,E) or growing follicles (F,G). Error bars in (A) and (C) denote SD. FDR = false discovery rate.

Figure 4.

Figure 4—figure supplement 1. Deletion of the small heat shock protein gene locus.

Figure 4—figure supplement 1.

(A) Normalized ribosome footprinting (upper tracks) and mRNA-seq (lower tracks) read depths in the 67B small heat shock protein gene cluster, are compared for replicate experiments from oocytes stored for 2 days (orange) and from 0–2-hr embryos (blue). Above the tracks is a map of the gene cluster, as well as the position of an FRT recombination-induced deletion of all but the left-most two genes that were generated. Two genes, Hsp26 and Hsp27, are transcribed and translated in both stored oocytes and 0–2-hr embryos. (B) Wild type (WT) and Df(sHSP)/Df(sHSP) ovarioles were stained with antibodies specific for Hsp26 and Hsp27. Both genes are expressed during oogenesis and are abundant in later stages (stage 10 shown). However, expression was absent above background in stage 10 follicles from the Df(sHSP)/Df(sHSP) ovarioles. (C) Plot showing similar stage distributions of 4–8-hr embryos derived from non-aged or aged Df(sHSP)/+ or Df(sHSP)/Df(sHSP) oocytes, suggesting that the premature loss of viability of oocytes lacking small heat shock proteins is not due to an acceleration of meiotic spindle defects observed in aged wild type oocytes.
Figure 4—figure supplement 2. Genes preferentially translated during oocyte arrest are not protected from widespread age-associated reduced translation efficiency.

Figure 4—figure supplement 2.

(A) Cumulative distribution plot showing that genes that are translationally upregulated in arrested oocytes (orange) as compared to 0–2-hr embryos show a slightly greater reduction in translation efficiency during aging as a group as compared to the distribution of all genes translated in oocytes. (B) Volcano plot showing that genes preferentially translated in oocytes (orange) as compared to early embryos are globally reduced in translation during oocyte aging.
Open in a new tab

The two small heat shock protein chaperones Hsp26 and Hsp27 are highly expressed in mature oocytes (Fredriksson et al., 2012; Zimmerman et al., 1983), and qualified as potential ‘pilot light’ products since they were translated at higher levels in oocytes than in early embryos (Figure 4B). Small heat shock proteins (sHSPs) are also highly expressed during yeast meiosis (Kurtz et al., 1986), suggesting a potential conserved function for sHSPs during gametogenesis. In order to test for a role of Hsp26 and Hsp27 during prolonged oocyte arrest, we used FRT-mediated recombination to construct a deficiency strain, Df(sHSP), that eliminates Hsp26 and Hsp27 (Figure 4—figure supplement 1A,B). The strain was backcrossed seven times to the yw strain to homogenize the genetic background.

The survival of stored wild type oocytes was compared to stored Df(sHSP) homozygous oocytes after between 2 and 14 days of storage in vivo to determine if Hsp26 and Hsp27 contribute to oocyte stability. Whereas wild type and Df(sHSP)/+ oocytes showed normal stability reductions during storage, homozygous Df(sHSP) oocytes lost developmental capacity more quickly (Figure 4C). Thus, Hsp26 and Hsp27 are important for the survival of oocytes during prolonged storage. We found that Hsp26 and Hsp27 are induced in stage 10 egg chambers several hours before the completion of oocyte development (Figure 4—figure supplement 1B), consistent with prior data (Zimmerman et al., 1983). These data suggest that a subset of genes critical for arrested oocytes are developmentally induced starting just prior to oocyte maturation, in preparation for a prolonged arrest.

Many genes preferentially translated in arrested oocytes are known or are likely to play important roles in the mature oocyte and shortly after the onset of embryogenesis. For example, Fmr1 is required to optimally maintain stored oocytes (Greenblatt and Spradling, 2018). Others including the thioredoxin-like dhd, are required to remove sperm protamines following fertilization (Emelyanov and Fyodorov, 2016; Tirmarche et al., 2016).

In addition, many kinetochore, spindle assembly checkpoint, and meiotic maturation genes including Ndc80, Nek2, Zw10, gnu, and mos (Lee, 2003; Radford et al., 2015; Sagata et al., 1989; Uto and Sagata, 2000; Williams et al., 1996) are also preferentially translated in oocytes (Supplementary file 2). GO analysis of the 243 genes showed a significantly enrichment for cell cycle-related processes, including mitotic cell cycle, spindle organization, and DNA repair (Figure 4D,E). We found that the translation of putative pilot light genes declined during aging slightly more than bulk translation as a whole (Figure 4—figure supplement 2A,B).

To investigate whether some functional categories of mRNAs are preferentially stockpiled in advance of oocyte completion, we also gathered information on how gene expression changes as oocyte growth ceases in preparation for storage, ovulation and embryonic development. We carried out RNA-seq and ribosome profiling on the ovaries of young flies 12–16 hr post-eclosion that still lack mature oogenic stages and compared them to day 2 mature oocytes. Genes whose translation is significantly upregulated in mature compared to growing oocytes are summarized in Supplementary file 3. These studies were consistent with previous analyses of gene expression during oogenesis and oocyte maturation (Cui et al., 2013; Kronja et al., 2014b; Sieber and Spradling, 2015; Tootle et al., 2011). Translation changes late in oogenesis analyzed by gene ontology reflect completion of follicle growth, reduced ribosomal production, nurse cell dumping, and reactivation of oocyte meiotic progression from diplotene to metaphase I (Figure 4F,G).

Maintaining meiotic spindles limits oocyte longevity

Given the large reductions we observed in the translation of genes related to meiotic spindle organization and the spindle assembly checkpoint, we investigated whether defects in the meiotic spindle could explain the reduced oocyte viability we observed during extended storage. We examined 1-, 7-, and 13-day-old Drosophila oocytes expressing α-tubulin-GFP, a construct which has been previously used to analyze meiotic spindles (Colombié et al., 2008). We found that the meiotic spindles of day 1 and day 7 oocytes were usually bipolar and highly tapered as previously described (Theurkauf and Hawley, 1992) (Figure 5A and A’). By day 13 however, many of the spindles were abnormal, such as unipolar, tripolar or fragmented (Figure 5B–D). Similar defects are seen in mutants of many of the meiotic spindle maintenance genes (e.g. sub, polo, and 14-3-3ε) whose translation declined substantially (>2 fold) during oocyte aging.

Figure 5. Stored oocytes lose developmental competence primarily due to problems completing meiosis.

Figure 5.

Open in a new tab

Meiotic spindles of oocytes stored for 1, 7 or 13 days at 25°C were visualized using α-tubulin-GFP (green) and DAPI (magenta) (A–C) or using α-tubulin-GFP alone (A'–C'). Normal bipolar spindles predominate at 1 day (A) but tripolar (B) and unipolar spindles (C) increase, and predominate by 13 days (D) (30 oocytes analyzed per timepoint). (E) Spindle structure correlates closely with oocyte function. (orange bars) The hatch rate of oocytes from wild type animals with the same oocyte and maternal ages measured in parallel. (gray bars) The percentage of oocytes that contain bipolar spindles (measured using α-tubulin-GFP) at the indicated age of storage, produced by females of the indicated ages. (Hatch rates were measured in triplicate and the meiotic spindles of 30 oocytes were analyzed per timepoint). (F–H) Stored oocytes that fail to develop show problems of meiotic completion and preblastoderm arrest. (F) DAPI stained 0–1-hr embryo from an oocyte stored <1 day shows normal cleavage stage nuclei and condensed polar body (arrowhead) visible at the 8 cell stage. (G,G’) 0–1-hr embryo from 12-day-old oocyte shows arrest at the first mitotic division; arrested mitotic spindle (arrow) and polar body (arrowhead). (G') higher magnification of the spindle in (G) with tubulin-GFP (green) and DAPI (magenta). (H,I) 0–1 hr embryos from a 12-day stored oocyte showing chaotic, arrested meiotic divisions with abnormal, tripolar/fragmented spindles. (J) Stage distribution of embryos from non-stored or 12-day-old or 17 day-old stored oocytes (N>30 at each point). Embryos fell into two categories; embryos from non-aged oocytes developed to stages 8–12 (blue), whereas embryos from 12-day and 17-day oocytes either progressed to stages 8–12 or arrested (orange) during the initial meiotic/mitotic divisions (pre-blastoderm). Scale bars = 10 μM.

In order to determine whether the loss of spindle bi-orientation was the primary cause of oocyte failure, we compared the hatch rate of embryos developing from young (1-day old) or aged (14-day old) oocytes with the proportion of oocytes with bipolar spindles at the same timepoints. We found a striking correlation between the fraction of oocytes able to support development to hatching and the proportion of oocytes with bipolar spindles (Figure 5E). Analyzing embryos derived from aged oocytes, we observed errors in chromosomal segregation during the mitotic divisions that follow pronuclear fusion (Figure 5G,G’), and meiotic divisions (Figure 5H,I). In contrast embryos derived from non-aged oocytes progressed normally through cleavage divisions (Figure 5F).

To investigate whether errors of chromosome segregation are the major cause of reduced embryonic viability, we collected embryos derived from unstored (<1 day), 12-day-old and 17-day-old oocytes and analyzed their level of development 4–8 hr after fertilization. Approximately half of embryos derived from oocytes stored for 12 days at 25°C, are able to develop to hatching, but >95% fail to develop after 17 days (Figure 1D). 100% of embryos derived from unstored oocytes had progressed to stages 8–12, as expected for normal development (Figure 5J). In contrast, the embryos derived from older oocytes showed a bimodal distribution of development. 58% of these embryos from 12-day-old oocytes developed to stages 8–12 like embryos from young oocytes. The other 42% arrested during initial cleavage divisions of pre-blastoderm embryos, failing to progress past the mitotic cell cycles of early embryogenesis preceding zygotic genome activation (Figure 5J). In the case of 17-day-old oocytes, almost all derived embryos arrested at pre-blastoderm stages. Only a few percent continued to develop normally (Figure 5J). This strong correlation between prolonged storage, lost developmental capacity and embryonic arrest prior to the blastoderm stage further implies that oocyte storage preferentially damages meiotic spindles and the ability to segregate chromosomes accurately to complete meiosis. Consequently, an increasing fraction of oocytes give rise to embryos that undergo chromosome mis-segregation shortly after fertilization leading to lethal aneuploidy.

Discussion

Drosophila females can be used to study how mature oocytes age during storage in the ovary

We developed a general system for studying the expression and genetic function of genes involved in the aging of completed Drosophila oocytes held in the ovary. Using our approach we determined precise aging curves for mature oocytes and showed they varied with temperature. Identifying the genes required for mature oocyte storage in the absence of transcription will elucidate mechanisms that enhance female fertility in many animals, define the limits of these mechanisms, and provide insight into why rare species such as humans are unable to maintain functional oocytes throughout adulthood.

Our studies also address more fundamental questions about the aging of cells that utilize long-lived mRNAs. Oocytes rely heavily on the regulated translation of relatively stable mRNA populations, especially towards the end of egg production. In this they resemble many other cells, including neurons, that utilize relatively stable mRNA at synapses, and male germ cells, which following meiosis transform into sperm by an elaborate translational program (Besse and Ephrussi, 2008; Fuller, 2016). Normally, an mRNA turns over in a matter of hours, not days (Sharova et al., 2009), and it remains unclear exactly how the functional capacity of mRNAs can be maintained for extended periods. The close association of long-lasting mRNAs in oocytes, neurons and sperm with P bodies, themselves derived from RNA turnover machinery, and the ability of mRNA to cycle between active and inactive states are likely to play critical roles that can now be studied more easily in a relatively simply and tractable in vivo system, the mature Drosophila oocyte.

A general genomic analysis of translational changes during aging

Our genomic studies reveal the changes in both mRNA and translation levels of essential Drosophila genes throughout the aging process. Our data show that despite the potential instability of mRNAs, a measurable decline in mRNA levels is not involved in the loss of oocyte biological function during aging. Using spike-in controls, it was possible to quantitatively compare samples between different time points. We found no significant decline in mRNA levels over the first 12 days of aging at 25°C.

Despite the preservation of mRNA, there was a pervasive general decrease in mRNA translation that correlated with the loss of oocyte function. Translation must decline either because of changes to mRNAs that reduce their ability to be translated, changes to the ribosomes, RNP granule dysfunction, or alterations in trans-acting factors required for translation. Analyzing the changes in translated proteins during aging did not reveal which of these mechanisms was likely to be responsible. Genes involved in multiple potentially relevant cellular processes undergo significant decreases in translation. These include genes involved in M phase, in meiotic cohesion, and in spindle formation, maintenance and bipolarity. Some of these genes are dose-sensitive (Knowles and Hawley, 1991; Moore et al., 1994; Subramanian and Bickel, 2008; Zhang et al., 1990), suggesting that a decrease of two-fold in expression would be enough to generate a phenotype. We observed greatly increased spindle instability as the levels of protein translation fell non-specifically into this range.

Drosophila oocyte decline represents aging in the absence of transcription

An important difference between storage of oocytes within primordial follicles and as full-grown oocytes concerns the status of transcription. Primary oocytes can continue to transcribe genes and repair or replace cellular components as needed, while granulosa cells can divide and replace whole cells if necessary. In contrast, mature oocytes without transcription must rely on translation, which despite the presence of sophisticated RNP-based regulatory machinery undergoes a significant decline in translational efficiency over relatively short periods.

What causes the decline in translation over the course of oocyte aging? One possibility is wear and tear on mRNAs that gradually reduces their ability to undergo translation. Even one cleavage usually inactivates an mRNA and targets it for turnover. Most mRNAs decay within a day or less (Sharova et al., 2009) even in growing cells, suggesting that specialized stabilization mechanisms exist during oocyte storage. Unlike proteins, which can be turned over and replaced using mRNA as a template, there is no transcription in mature oocytes and no way to replace damaged mRNA molecules. RNA molecules with expanded trinucleotide repeat sequences can seed the formation of aggregates in a manner analogous to protein aggregation (Jain and Vale, 2017; Querido et al., 2011). It is unknown if mRNAs are generally susceptible to misfolding and aggregation as has been well-characterized for proteins. We hypothesize that P bodies and stress granules, which form during periods of cellular stress as a response to the accumulation of untranslated mRNAs (Eulalio et al., 2007), participate in the long-term preservation of stored mRNAs by decreasing mRNA aggregationor damage.

Our findings suggest new insight into the strategy of oocyte maintenance in mammals

Our studies have several possible implications for understanding and potentially mitigating the increasing instability of chromosome segregation in mammalian and especially in human oocytes and early embryos. Because of the decades long delay between the onset of meiosis and its completion, some slow decay of an important meiotic process during the primordial follicle stage has been suspected. A logical candidate is the process of sister chromatid cohesion. After forming during pre-meiotic S phase, meiotic cohesion complexes do not appear to turn over or incorporate freshly synthesized protein subunits (Revenkova et al., 2010; Tachibana-Konwalski et al., 2010). Reduced levels of the meiotic cohesin complex components Rec8 and Sgo2, which protect cohesin from separase-mediated cleavage, were observed in aged oocytes (Chiang et al., 2010; Lister et al., 2010) and interkinetochore distances increase with age (Merriman et al., 2013).

However, our results suggest that defects arising during the storage of fully grown oocytes have been under-appreciated as an additional source of meiotic and early embryonic mitotic instability. The production of new transcripts in the oocyte GV strongly drops or ceases after oocytes reach full size, and does not begin again until the 4-cell stage in humans. During this period, oocytes would be largely or entirely dependent on their existing mRNA pool, like stored Drosophila mature oocytes (see model, Figure 6). Currently, there are insufficient studies using cell marking techniques to follow how long individual full-size mammalian oocytes remain in a quiescent state, what the consequences of late storage are on translation, and whether the average length of mature oocyte storage changes with maternal age and increased incidence of menstrual irregularities. Given the high sensitivity of mature Drosophila oocytes to storage shown here, we suggest that a significant fraction of human chromosome instability is caused by the duration of late storage, rather than by defects that occur at the primordial follicle stage. This would imply that the problems of human chromosome instability may be more susceptible to intervention than previously believed.

Figure 6. Model for prolonged mRNA storage during human oocyte development.

Figure 6.

Open in a new tab

Accumulation of meiotic mRNAs occurs prior to the cessation of oocyte growth in pre-antral secondary follicles. Meiotic maturation of prophase I arrested human oocytes requires the translation of mRNAs that have been stored for a prolonged period of development.

Materials and methods

Oocyte aging assay

Newly eclosed virgin wild type females of indicated genotypes were placed in standard food vials containing added yeast paste made by mixing live yeast with water until the mixture acquires the consistency of peanut butter, but does not trap flies. After feeding on the yeast for 24 hr, flies were transferred to "molasses plates" containing agar-sugar medium (44 g agar, 180 mL molasses, 37 mL in 5% Tegosept, 1112 mL water) to provide humidity, but no edible yeast. After one additional day, ovarioles contain two stage 14 oocytes and are considered to have begun day 1 of quiescence. For study, ovaries were dissected after the desired period of quiescence and the mature stage 14 oocytes were collected. To study oocyte viability, 10 females were transferred to chambers with fresh molasses egg laying plates and 10 males were added. Molasses plates were scored with a needle to increase the number of eggs laid. Males were isolated from females for at least 2 days prior to addition and were aged for 3–8 days from eclosion. Laid embryos were counted and recovered for study after various periods of time. Eggs were collected 16 hr after addition of males and hatch rates were determined 48 hr after collection. Protein measurements were performed using the BCA assay (Pierce).

Generation of antibodies and deletion alleles of Hsp26 and Hsp27

Antibodies were generated against the C-termini of Drosophila melanogaster HSP26 and HSP27 using peptides KLHcarrier-cys-KANESEVKGKENGAPNGKDK and KLHcarrier-cys-APEAGDGKAENGSGEKMETSK respectively (Proteintech) and were used at a concentration of 1:4000. A deletion of the sHSP region was generated via FLP-mediated recombination of FRT-bearing lines d00797 and d05052 from the Harvard Exelixis collection (Thibault et al., 2004). Deletion of Hsp26 and Hsp27 was confirmed by PCR analysis and immunostaining.

Drosophila ovary and embryo immunostaining

Ovaries were hand-dissected in Grace’s Insect Medium (Life Technologies) from flies fed for 3 days with wet yeast paste. Ovaries were fixed in 4% formaldehyde (37% formaldehyde diluted in PBST (0.2% BSA, 0.1% Triton X-100 in 1X PBS) for 12 min. Ovaries were incubated with primary antibodies diluted in PBST with gentle agitation overnight at 4°C. Ovaries were then washed 3 times in PBST for at least 20 min and incubated with secondary antibodies overnight. Ovaries were then washed 3 times with PBST for at least 20 min each, and DAPI (1:20,000-fold dilution of a 5 mg/mL stock) was added to the last wash.

Embryos were dechorionated for 2 min in bleach (50% diluted fresh Clorox bleach). Embryos were fixed for 25 min in a 1:1 mixure of fixative (50 mM EDTA, 9.25% formaldehyde, 1XPBS buffer) and heptane with gentle agitation. The lower fixative layer was removed and an equal volume of methanol was added. Embryos were devitellinized by shaking vigorously by hand for 4 min, removing the heptane layer, and shaking for an addition 1–2 min. Embryos were washed three times in methanol and rehydrated in 50% methanol in PBST, and washed three times in PBST. Embryos were blocked for one hour in PBST and then processed as described for ovaries.

Ribosome profiling and mRNA-seq library preparation

Ribosome profiling and mRNA sequencing was carried out as described in Greenblatt and Spradling (2018) with the following modifications. Aged oocytes were defolliculated by treating ovaries with 5 mg/mL collagenase (Sigma-Aldrich) in PBST for 10 min at room temperature with gentle agitation and then washed three times in PBST. Oocytes were isolated and separated from debris by filtration. Following extraction of defolliculated oocytes or 0–2 embryos with lysis buffer (0.5% Triton X-100, 150 mM NaCl, 5 mM MgCl2, 50 mM Tris, pH 7.5, 1 mM DTT, 20 ug/mL emetine (Sigma-Aldrich), 20 U/mL SUPERaseIn (Ambion), 50 uM GMP-PNP (Sigma-Aldrich)) Drosophila melanogaster oocyte extract containing 80 ug RNA was combined with Drosophia pseudoobscura whole ovary extract containing 1.6 µg RNA. The combined extract was then processed as in Greenblatt and Spradling (2018).

Ribosome profiling and mRNA-seq data analysis

Analysis of ribosome profiling and mRNA sequencing data was conducted as in Greenblatt and Spradling (2018) with the following modifications. For quantification of bulk mRNA/ribosome footprint levels, adaptor-trimmed reads were mapped to a file containing coding sequences of combined Dmel and Dpse transcripts using Bowtie v2.3.2 and filtered for only uniquely mapping reads (lines containing string ‘NH:i:1’) and total reads mapping to either Dmel or Dpse were counted (Supplementary file 4). Unique Dmel reads were then re-mapped to the Dmel release 6.02 genome with HISAT2 ver2.1.0. Transcripts per million (TPM) values for coding sequences (ribosome profiling) or exons (mRNA sequencing) were obtained using Stringtie v1.3.5 with the files dmel-CDS-r6.02 gtf or dmel-exons-r6.02.gtf used as a reference annotation for ribosome profiling or mRNA sequencing analysis respectively. Ribosome profiling TPM values from 8 day oocytes, 12 day oocytes, and 0–2 embryo samples were then adjusted by factors of 0.595, and 0.427, and 2.76 respectively to account for bulk changes in translation as determined by the ratios of Dmel to Dpse reads. Of 26,223 potential 30mer footprints from the top 30 translated Drosophila genes in oocytes, we found that 25,324 (97%) sequences contained at least one polymorphism when comparing sequences from orthologous Dpse transcripts. Gene ontology analysis was performed using the PANTHER server (Mi et al., 2019).

Acknowledgements

We are grateful to Kamena Kostova, Steve DeLuca, Chenhui Wang, and members of the Spradling lab for support and comments on the manuscript.

Funding Statement

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Contributor Information

Allan C Spradling, Email: spradling@ciwemb.edu.

Michael B Eisen, HHMI, University of California, Berkeley, United States.

Michael B Eisen, HHMI, University of California, Berkeley, United States.

Funding Information

This paper was supported by the following grant:

  • Howard Hughes Medical Institute to Allan C Spradling.

Additional information

Competing interests

Reviewing editor, eLife.

No competing interests declared.

Author contributions

Conceptualization, Data curation, Software, Formal analysis, Validation, Investigation, Visualization, Methodology, Project administration.

Data curation, Formal analysis, Visualization.

Data curation.

Conceptualization, Formal analysis, Supervision, Investigation, Methodology.

Additional files

Supplementary file 1. Translational and mRNA changes during oocyte aging.
elife-49455-supp1.xlsx (3.6MB, xlsx)
Supplementary file 2. ‘Pilot light’ genes translationally upregulated during oocyte arrest.
elife-49455-supp2.xlsx (232.7KB, xlsx)
Supplementary file 3. Genes upregulated during oocyte maturation.
elife-49455-supp3.xlsx (78.1KB, xlsx)
Supplementary file 4. Counts of total reads mapping to Dmel and Dpse for ribosome profiling and mRNA-sequencing experiments.
elife-49455-supp4.xlsx (9.9KB, xlsx)
Transparent reporting form

Data availability

Data has been uploaded to BioProjects at NCBI under PRJNA573922.

The following dataset was generated:

Greenblatt EJ, Obniski R, Michael C, Spradling AC. 2019. Ribosome profiling and mRNA sequencing of aging oocytes in Drosophila. NCBI Bioprojects. PRJNA57392

References

  1. Abe K, Inoue A, Suzuki MG, Aoki F. Global gene silencing is caused by the dissociation of RNA polymerase II from DNA in mouse oocytes. Journal of Reproduction and Development. 2010;56:502–507. doi: 10.1262/jrd.10-068A. [DOI] [PubMed] [Google Scholar]
  2. Besse F, Ephrussi A. Translational control of localized mRNAs: restricting protein synthesis in space and time. Nature Reviews Molecular Cell Biology. 2008;9:971–980. doi: 10.1038/nrm2548. [DOI] [PubMed] [Google Scholar]
  3. Chiang T, Duncan FE, Schindler K, Schultz RM, Lampson MA. Evidence that weakened centromere cohesion is a leading cause of age-related aneuploidy in oocytes. Current Biology. 2010;20:1522–1528. doi: 10.1016/j.cub.2010.06.069. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Colombié N, Cullen CF, Brittle AL, Jang JK, Earnshaw WC, Carmena M, McKim K, Ohkura H. Dual roles of incenp crucial to the assembly of the acentrosomal metaphase spindle in female meiosis. Development. 2008;135:3239–3246. doi: 10.1242/dev.022624. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Conti M, Franciosi F. Acquisition of oocyte competence to develop as an embryo: integrated nuclear and cytoplasmic events. Human Reproduction Update. 2018;24:245–266. doi: 10.1093/humupd/dmx040. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Coticchio G, Dal Canto M, Mignini Renzini M, Guglielmo MC, Brambillasca F, Turchi D, Novara PV, Fadini R. Oocyte maturation: gamete-somatic cells interactions, meiotic resumption, cytoskeletal dynamics and cytoplasmic reorganization. Human Reproduction Update. 2015;21:427–454. doi: 10.1093/humupd/dmv011. [DOI] [PubMed] [Google Scholar]
  7. Cui J, Sartain CV, Pleiss JA, Wolfner MF. Cytoplasmic polyadenylation is a major mRNA regulator during oogenesis and egg activation in Drosophila. Developmental Biology. 2013;383:121–131. doi: 10.1016/j.ydbio.2013.08.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Dantuma NP, Lindsten K, Glas R, Jellne M, Masucci MG. Short-lived green fluorescent proteins for quantifying ubiquitin/proteasome-dependent proteolysis in living cells. Nature Biotechnology. 2000;18:538–543. doi: 10.1038/75406. [DOI] [PubMed] [Google Scholar]
  9. Drummond-Barbosa D, Spradling AC. Stem cells and their progeny respond to nutritional changes during Drosophila oogenesis. Developmental Biology. 2001;231:265–278. doi: 10.1006/dbio.2000.0135. [DOI] [PubMed] [Google Scholar]
  10. Emelyanov AV, Fyodorov DV. Thioredoxin-dependent disulfide bond reduction is required for protamine eviction from sperm chromatin. Genes & Development. 2016;30:2651–2656. doi: 10.1101/gad.290916.116. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Eulalio A, Behm-Ansmant I, Schweizer D, Izaurralde E. P-body formation is a consequence, not the cause, of RNA-mediated gene silencing. Molecular and Cellular Biology. 2007;27:3970–3981. doi: 10.1128/MCB.00128-07. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Fredriksson Å, Johansson Krogh E, Hernebring M, Pettersson E, Javadi A, Almstedt A, Nyström T. Effects of aging and reproduction on protein quality control in soma and gametes of Drosophila melanogaster. Aging Cell. 2012;11:634–643. doi: 10.1111/j.1474-9726.2012.00823.x. [DOI] [PubMed] [Google Scholar]
  13. Fuller MT. Chapter Twenty - Differentiation in Stem Cell Lineages and in Life: Explorations in the Male Germ Line Stem Cell Lineage. In: Wassarman P. M, editor. Current Topics in Developmental Biology. Academic Press; 2016. pp. 375–390. [DOI] [PubMed] [Google Scholar]
  14. Greenblatt EJ, Spradling AC. Fragile X mental retardation 1 gene enhances the translation of large autism-related proteins. Science. 2018;361:709–712. doi: 10.1126/science.aas9963. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Herbert M, Kalleas D, Cooney D, Lamb M, Lister L. Meiosis and maternal aging: insights from aneuploid oocytes and trisomy births. Cold Spring Harbor Perspectives in Biology. 2015;7:a017970. doi: 10.1101/cshperspect.a017970. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Hughes SE, Miller DE, Miller AL, Hawley RS. Female meiosis: synapsis, recombination, and segregation in Drosophila melanogaster. Genetics. 2018;208:875–908. doi: 10.1534/genetics.117.300081. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Jain A, Vale RD. RNA phase transitions in repeat expansion disorders. Nature. 2017;546:243–247. doi: 10.1038/nature22386. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Jukam D, Shariati SAM, Skotheim JM. Zygotic genome activation in vertebrates. Developmental Cell. 2017;42:316–332. doi: 10.1016/j.devcel.2017.07.026. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Knowles BA, Hawley RS. Genetic analysis of microtubule motor proteins in Drosophila: a mutation at the ncd locus is a dominant enhancer of nod. PNAS. 1991;88:7165–7169. doi: 10.1073/pnas.88.16.7165. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Kronja I, Yuan B, Eichhorn SW, Dzeyk K, Krijgsveld J, Bartel DP, Orr-Weaver TL. Widespread changes in the posttranscriptional landscape at the Drosophila oocyte-to-embryo transition. Cell Reports. 2014a;7:1495–1508. doi: 10.1016/j.celrep.2014.05.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Kronja I, Whitfield ZJ, Yuan B, Dzeyk K, Kirkpatrick J, Krijgsveld J, Orr-Weaver TL. Quantitative proteomics reveals the dynamics of protein changes during Drosophila oocyte maturation and the oocyte-to-embryo transition. PNAS. 2014b;111:16023–16028. doi: 10.1073/pnas.1418657111. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Kurtz S, Rossi J, Petko L, Lindquist S. An ancient developmental induction: heat-shock proteins induced in sporulation and oogenesis. Science. 1986;231:1154–1157. doi: 10.1126/science.3511530. [DOI] [PubMed] [Google Scholar]
  23. Lee LA. The Drosophila cell cycle kinase PAN GU forms an active complex with PLUTONIUM and GNU to regulate embryonic divisions. Genes & Development. 2003;17:2979–2991. doi: 10.1101/gad.1132603. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Lister LM, Kouznetsova A, Hyslop LA, Kalleas D, Pace SL, Barel JC, Nathan A, Floros V, Adelfalk C, Watanabe Y, Jessberger R, Kirkwood TB, Höög C, Herbert M. Age-related meiotic segregation errors in mammalian oocytes are preceded by depletion of cohesin and Sgo2. Current Biology. 2010;20:1511–1521. doi: 10.1016/j.cub.2010.08.023. [DOI] [PubMed] [Google Scholar]
  25. Lovett JA, Goldstein ES. The cytoplasmic distribution and characterization of poly(A)+RNA in oocytes and embryos of drosophilia. Developmental Biology. 1977;61:70–78. doi: 10.1016/0012-1606(77)90342-6. [DOI] [PubMed] [Google Scholar]
  26. McCoy RC, Demko ZP, Ryan A, Banjevic M, Hill M, Sigurjonsson S, Rabinowitz M, Petrov DA. Evidence of selection against complex Mitotic-Origin aneuploidy during preimplantation development. PLOS Genetics. 2015;11:e1005601. doi: 10.1371/journal.pgen.1005601. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Merriman JA, Lane SI, Holt JE, Jennings PC, García-Higuera I, Moreno S, McLaughlin EA, Jones KT. Reduced chromosome cohesion measured by interkinetochore distance is associated with aneuploidy even in oocytes from young mice. Biology of Reproduction. 2013;88:31. doi: 10.1095/biolreprod.112.104786. [DOI] [PubMed] [Google Scholar]
  28. Mi H, Muruganujan A, Ebert D, Huang X, Thomas PD. PANTHER version 14: more genomes, a new PANTHER GO-slim and improvements in enrichment analysis tools. Nucleic Acids Research. 2019;47:D419–D426. doi: 10.1093/nar/gky1038. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Moore DP, Miyazaki WY, Tomkiel JE, Orr-Weaver TL. Double or nothing: a Drosophila mutation affecting meiotic chromosome segregation in both females and males. Genetics. 1994;136:953–964. doi: 10.1093/genetics/136.3.953. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Nakagawa S, FitzHarris G. Intrinsically defective microtubule dynamics contribute to Age-Related chromosome segregation errors in mouse oocyte Meiosis-I. Current Biology. 2017;27:1040–1047. doi: 10.1016/j.cub.2017.02.025. [DOI] [PubMed] [Google Scholar]
  31. Querido E, Gallardo F, Beaudoin M, Ménard C, Chartrand P. Stochastic and reversible aggregation of mRNA with expanded CUG-triplet repeats. Journal of Cell Science. 2011;124:1703–1714. doi: 10.1242/jcs.073270. [DOI] [PubMed] [Google Scholar]
  32. Radford SJ, Hoang TL, Głuszek AA, Ohkura H, McKim KS. Lateral and End-On kinetochore attachments are coordinated to achieve Bi-orientation in Drosophila oocytes. PLOS Genetics. 2015;11:e1005605. doi: 10.1371/journal.pgen.1005605. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Revenkova E, Herrmann K, Adelfalk C, Jessberger R. Oocyte cohesin expression restricted to predictyate stages provides full fertility and prevents aneuploidy. Current Biology. 2010;20:1529–1533. doi: 10.1016/j.cub.2010.08.024. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Sagata N, Watanabe N, Vande Woude GF, Ikawa Y. The c-mos proto-oncogene product is a cytostatic factor responsible for meiotic arrest in vertebrate eggs. Nature. 1989;342:512–518. doi: 10.1038/342512a0. [DOI] [PubMed] [Google Scholar]
  35. Sharova LV, Sharov AA, Nedorezov T, Piao Y, Shaik N, Ko MS. Database for mRNA half-life of 19 977 genes obtained by DNA microarray analysis of pluripotent and differentiating mouse embryonic stem cells. DNA Research. 2009;16:45–58. doi: 10.1093/dnares/dsn030. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Sieber MH, Spradling AC. Steroid signaling establishes a female metabolic state and regulates SREBP to control oocyte lipid accumulation. Current Biology. 2015;25:993–1004. doi: 10.1016/j.cub.2015.02.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Subramanian VV, Bickel SE. Aging predisposes oocytes to meiotic nondisjunction when the cohesin subunit SMC1 is reduced. PLOS Genetics. 2008;4:e1000263. doi: 10.1371/journal.pgen.1000263. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Tachibana-Konwalski K, Godwin J, van der Weyden L, Champion L, Kudo NR, Adams DJ, Nasmyth K. Rec8-containing cohesin maintains bivalents without turnover during the growing phase of mouse oocytes. Genes & Development. 2010;24:2505–2516. doi: 10.1101/gad.605910. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Theurkauf WE, Hawley RS. Meiotic spindle assembly in Drosophila females: behavior of nonexchange chromosomes and the effects of mutations in the nod kinesin-like protein. The Journal of Cell Biology. 1992;116:1167–1180. doi: 10.1083/jcb.116.5.1167. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Thibault ST, Singer MA, Miyazaki WY, Milash B, Dompe NA, Singh CM, Buchholz R, Demsky M, Fawcett R, Francis-Lang HL, Ryner L, Cheung LM, Chong A, Erickson C, Fisher WW, Greer K, Hartouni SR, Howie E, Jakkula L, Joo D, Killpack K, Laufer A, Mazzotta J, Smith RD, Stevens LM, Stuber C, Tan LR, Ventura R, Woo A, Zakrajsek I, Zhao L, Chen F, Swimmer C, Kopczynski C, Duyk G, Winberg ML, Margolis J. A complementary transposon tool kit for Drosophila Melanogaster using P and piggyBac. Nature Genetics. 2004;36:283–287. doi: 10.1038/ng1314. [DOI] [PubMed] [Google Scholar]
  41. Tirmarche S, Kimura S, Dubruille R, Horard B, Loppin B. Unlocking sperm chromatin at fertilization requires a dedicated egg thioredoxin in Drosophila. Nature Communications. 2016;7:13539. doi: 10.1038/ncomms13539. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Tootle TL, Williams D, Hubb A, Frederick R, Spradling A. Drosophila eggshell production: identification of new genes and coordination by pxt. PLOS ONE. 2011;6:e19943. doi: 10.1371/journal.pone.0019943. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Uto K, Sagata N. Nek2B, a novel maternal form of Nek2 kinase, is essential for the assembly or maintenance of centrosomes in early xenopus embryos. The EMBO Journal. 2000;19:1816–1826. doi: 10.1093/emboj/19.8.1816. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Webster A, Schuh M. Mechanisms of aneuploidy in human eggs. Trends in Cell Biology. 2017;27:55–68. doi: 10.1016/j.tcb.2016.09.002. [DOI] [PubMed] [Google Scholar]
  45. Williams BC, Gatti M, Goldberg ML. Bipolar spindle attachments affect redistributions of ZW10, a Drosophila centromere/kinetochore component required for accurate chromosome segregation. The Journal of Cell Biology. 1996;134:1127–1140. doi: 10.1083/jcb.134.5.1127. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Zhang P, Knowles BA, Goldstein LS, Hawley RS. A kinesin-like protein required for distributive chromosome segregation in Drosophila. Cell. 1990;62:1053–1062. doi: 10.1016/0092-8674(90)90383-P. [DOI] [PubMed] [Google Scholar]
  47. Zimmerman JL, Petri W, Meselson M. Accumulation of a specific subset of D. melanogaster heat shock mRNAs in normal development without heat shock. Cell. 1983;32:1161–1170. doi: 10.1016/0092-8674(83)90299-4. [DOI] [PubMed] [Google Scholar]

Decision letter

Editor: Michael B Eisen1

In the interests of transparency, eLife publishes the most substantive revision requests and the accompanying author responses.

Acceptance summary:

The decay of viability of stored oocytes contributes significant to miscarriage and limits reproductive lifespan in humans. This study establishes the Drosophila oocyte as a model to study aging of mature oocytes, and provides high-quality data on transcription and translation during the oocytes aging process.

Oocytes are unique cells that can persist for days (or years in the case of humans) in an arrested state. In most animals, oocytes arrest twice during normal development: a first arrest in prophase of meiosis I (which is typically the longer arrest) and a second, typically shorter, arrest after oocyte maturation when the oocyte is in metaphase of Meiosis I or II. How oocytes maintain viability and competency to support embryogenesis during these arrest stages is an area of great interest and practical import.

This paper applies a system the authors previously developed to study Drosophila oocytes during the second mature oocyte arrest. This arrest is normally very brief as mature oocytes are normally efficiently ovulated and fertilized. Here, the authors manipulated nutrition and the availability of males to artificially prolong the second arrest by an order of magnitude. They find that, even though transcription has been halted, mRNA levels remain steady in aging oocytes over this period, but that translation rates decrease. They go on to offer evidence that this translational decay plays a role in declining oocyte viability, specifically in the development of abnormal mitotic spindles.

In an effort to identify genes required for viability of arrested oocytes, the authors identify a set of candidate "pilot light" genes that may play a role in maintaining higher level of translation in arrested oocytes compared to embryos. They confirm two of these candidates, Hsp26 and Hsp27, showing that loss of these genes affects oocyte viability, although the precise role of the genes in the process remains to be illuminated.

Overall this study defines a potentially very interesting consequence of aging oocytes after oocyte maturation: a progressive decline in translation rate, and it should be interesting to anyone studying gonadal and germline development as well as infertility. Many questions remain, such as whether this phenotype is specific or reflects a general decline in cell viability, what is the primary cause of the translational slow down and does this process play some role in maintaining viability even as it decreases? This system appears to have great potential for illuminating these and other questions related to the oocyte maturation process.

Decision letter after peer review:

Thank you for submitting your article "Prolonged ovarian storage of mature Drosophila oocytes dramatically increases meiotic spindle instability" for consideration by eLife. Your article has been reviewed by two peer reviewers, and the evaluation has been overseen by Michael Eisen as Reviewing and Senior Editor. The other reviewers have opted to remain anonymous.

I have drafted a review that synthesizes the comments of myself and the two reviewers, who have reviewed and approved it.

The authors have developed a system (previously described in Greenblatt and Spradling, 2018) to study Drosophila oocytes during the second arrest ("mature oocyte arrest"). This arrest is normally very brief, as mature oocytes are normally efficiently ovulated and fertilized. By manipulating nutrition and the availability of males, they artificially prolong the second arrest to ~12 days. These oocytes experienced declines in hatch rate with increasing periods of arrest that was correlated with increasing levels of MI spindle abnormalities.

They find that mRNA levels remain steady in aged oocytes over this period, but ribosome footprints gradually decline. They interpret these results as showing that prolonged arrest of mature oocytes leads to a gradual decline in global translation (but see below). The decline is general and includes genes with essential functions such as chaperone dependent protein folding and spindle assembly, consistent with increase in spindle abnormalities as oocytes age.

In an effort to identify genes required for viability of arrested oocytes, the authors identify so-called "pilot light" genes that maintain a higher level of translation in arrested oocytes compared to embryos. They identify two "pilot-light" genes Hsp26 and Hsp27 and show that loss of these genes affects oocyte viability, but not spindle assembly. Whether these genes are specifically required in oocytes to maintain viability during arrest, or are more generally required for germ cell viability, is not clear and so the significance of these findings remains uncertain.

Overall this study defines a potentially very interesting consequence of aging oocytes after oocyte maturation: a progressive decline in translation rate. But there are several important issues/questions that need to be addressed:

1) Is the phenomenon specific to aging oocytes? There are several reasons to be concerned that this is not the case.

First the authors utilize starvation to arrest the oocytes. While the premise is that late stage oocytes have already stockpiled the proteins and mRNAs it needs prior to the starvation, previous studies have shown that starvation induces changes that affect the entire organism (for example halting development of earlier oocytes and affecting life span). It is not clear if late stage oocytes are affected by the signaling cascades induced by starvation and if some of the changes observed are the oocytes response to the cue that food will be absent in the environment the oocyte will be deposited. This could be addressed at least partially by exploiting the fact that Drosophila females can be made to hold late stage oocytes for 4-5 days without starvation by preventing access to males. Examining such 4-5 day arrested oocytes without starvation would strengthen the argument that translational changes are due specifically to arrest.

Second, the authors point out that in humans oocytes from younger women still showed increased abnormalities when implanted into older woman indicating mother's age may influence developed oocytes as well. Some of the upregulated genes and phenotypes they observed may be influenced by the age of the females (general aging) rather than actual oocyte arrest that would occur in other species. This could be addressed by examining stage 14 oocytes from fed and mated females that are aged 14 days for both translational and cytological assays. Using oocytes from 2 day starved females meant changes may have already occurred in response to starvation and/or arrest, oocytes from non-starved females is the better control.

Finally, it is unclear why the authors were making comparisons to immature follicles and embryos when the focus of the paper is written to focus on changes in arrested oocytes. These comparisons only indicate how a cell in metaphase I is different from cells either early in meiosis or mitosis. Studies have looked at these differences before and were only mentioned in passing. To understand specifically oocyte arrest the authors need to compare oocytes of the same stage but arrested for long periods compared to not arrested stage 14 oocytes. The current studies primarily show what is needed in a metaphase I oocyte.

2) Is the effect specific to translation?

The paper focuses specifically on translation, but it is unclear if the effect is specific to translation. Perhaps the oocytes are just slowly dying and the translational slowdown is a consequence of a general loss in cell viability, drop in ATP levels, etc… Several of the controls suggested in (1) above might address this, but in general it would be very helpful for the authors to demonstrate that the effects on translation are specific and that other metabolic activities are not affected or affected only secondarily, as a consequence of the translational shut down. Otherwise, the data could be interpreted as simply reporting on the slow death of oocytes caused by an artificially extended arrest that may or may not be physiologically relevant.

3) Throughout the paper, the authors rely on ribosomal profiling to measure translation. The data beautifully document that ribosome footprints/mRNA decline with age, but whether this means that translational output also declines is not a foregone conclusion. Ribosome footprints are not necessarily a direct measure of translational output: a decrease in ribosome footprints could mean a decrease in the rate of translation initiation (fewer ribosomes initiating, fewer ribosomes on message, less translational output) or an increase in the rate of translation elongation (faster ribosomes, fewer ribosomes on message, more translational output). Without a separate assay to directly measure translational output, it is not possible to distinguish between these two options.

4) There are a number of other prior studies that need to be discussed in detail.

The prior work examining oocyte aging by the Bickel lab needs to be incorporated into the Discussion.

In Fredriksson et al., 2012, Hsp26 and Hsp27 were already found to be strongly expressed at 5 and 35 days of age in late stage eggs which the authors need to mention. The authors show expression of these genes by antibody in stage 10 oocytes but the authors needed to show these genes by cytology or quantitative western of stage 14 oocytes that were arrested for 14 days with starvation and from non-starved 14 day females. This would support that these genes are specifically up-translated in oocytes arrested long periods. The past and current data suggest the proteins are loaded prior to stage 14 oocyte arrest.

If the authors are to continue to include the comparisons to mitotic embryos and early stage follicles their results should be compared and contrasted to the results looking at total protein changes by the Orr-Weaver lab.

Specific suggestions:

The authors mention a number of genes that shows changes in translation but fail to always explain what the genes are and/or provide references. Not all the readers interested in oocyte aging would be familiar with Drosophila gene names.

Figure 4. To study proteins that are preferentially translated in arrested oocytes and identify "pilot light" genes, the authors compare translation in 2-day old arrested oocytes to early embryos. However, since embryos must have a vastly different translational program than oocytes, it might be more enlightening to compare arrested oocytes to growing oocytes. It seems this experiment has already been done (Figure 4E, 4F) but further analysis of specific genes that are more highly translated in arrested oocytes compared to growing oocytes could potentially give more insight into other "pilot light" genes that contribute to meiotic spindle stability.

Figure 5. In humans, primordial follicles are stored for long periods of time in meiotic prophase before re-entering the cell cycle and briefly pausing again at metaphase II before fertilization. It is my understanding that this metaphase II pause lasts only 1-2 days in vivo. In this study, the authors show that mature Drosophila oocytes arrested at metaphase of meiosis I for long periods (> 6 days) exhibit meiotic spindle instability and abnormal chromosome segregation. They imply that this may be relevant to chromosome instability in human females ("Our findings show that storage of highly functional mature oocytes in vivo is sufficient to destabilize chromosome segregation, suggesting that the prolonged storage of mature oocytes may be an important source of meiotic chromosome instability in human females"). Whether mature human oocytes ever experience such a long arrest is not clear. A diagram that clearly compares oogenesis in Drosophila versus humans with the different arrest points could help clarify their argument.

Figure 6. The authors hypothesize that "One potential function of sHSPs might be to stabilize the meiotic spindle and thereby extend the functional lifetime of oocytes." However, they then seem to disprove this very statement in Figure 6D. Although loss of Hsp26 and Hsp27 has an impact on oocyte maintenance, the mechanism is unknown and as the final figure, this result is a bit disappointing.

eLife. 2019 Nov 22;8:e49455. doi: 10.7554/eLife.49455.sa2

Author response

I have drafted a review that synthesizes the comments of myself and the two reviewers, who have reviewed and approved it.

The authors have developed a system (previously described in Greenblatt and Spradling, 2018) to study Drosophila oocytes during the second arrest ("mature oocyte arrest"). This arrest is normally very brief, as mature oocytes are normally efficiently ovulated and fertilized. By manipulating nutrition and the availability of males, they artificially prolong the second arrest to ~12 days. These oocytes experienced declines in hatch rate with increasing periods of arrest that was correlated with increasing levels of MI spindle abnormalities.

They find that mRNA levels remain steady in aged oocytes over this period, but ribosome footprints gradually decline. They interpret these results as showing that prolonged arrest of mature oocytes leads to a gradual decline in global translation (but see below). The decline is general and includes genes with essential functions such as chaperone dependent protein folding and spindle assembly, consistent with increase in spindle abnormalities as oocytes age.

In an effort to identify genes required for viability of arrested oocytes, the authors identify so-called "pilot light" genes that maintain a higher level of translation in arrested oocytes compared to embryos. They identify two "pilot-light" genes Hsp26 and Hsp27 and show that loss of these genes affects oocyte viability, but not spindle assembly. Whether these genes are specifically required in oocytes to maintain viability during arrest, or are more generally required for germ cell viability, is not clear and so the significance of these findings remains uncertain.

We thank the reviewers for their comments on our manuscript. We have made changes, including experiments reported in 5 new and 1 modified main figure panels that have addressed all the requested issues as detailed in this guide.

However, we first respond to a major misunderstanding regarding the physiological role of mature egg storage in Drosophila. Regarding the in vivo stage 14 arrest we studied the reviewers state:

"This arrest is normally very brief, as mature oocytes are normally efficiently ovulated and fertilized. By manipulating nutrition and the availability of males, they artificially prolong the second arrest to ~12 days."

"Otherwise, the data could be interpreted as simply reporting on the slow death of oocytes caused by an artificially extended arrest that may or may not be physiologically relevant."

No evidence supports and much evidence contradicts the idea that mature oocytes in the wild are normally ovulated and fertilized after a "very brief" arrest. We are concerned that a sentence in the Abstract that referred to rapid ovulation by laboratory Drosophila maintained in good conditions misled the reviewers regarding the critically important biological function of mature oocyte storage for Drosophila in their natural environment.

Natural conditions of Drosophila reproduction in the wild differ drastically from the conditions Drosophila experience in the laboratory (Markow, 2015, eLife), and make a system for storing mature eggs critically important for female reproduction. Adult females in the wild face long odds at reproduction. With adequate protein each female can produce many hundreds of eggs, but in the wild only two progeny enter the next generation on average. Wild females do not have ready access to protein-rich yeast, but have to find this transient, rare resource and compete for access, on multiple occasions. Wild caught females almost never display the large ovaries replete with rapidly developing oocytes such as seen after special feeding in the lab, a sign of chronic protein limitation. In the wild, if stage 14 oocytes are present, the ovarioles usually contain multiple stage 14 oocytes, a sign of the late arrest we studied. Thus, Drosophila face the same problem that drove mosquitos to evolve blood feeding.

The ability to store mature oocytes during multi-day searches for ovulation sites probably represents the rate-limiting factor determining female reproductive success. The fact that mature eggs can be stored for 14 days at 25°C (14 times longer than necessary to complete embryonic development) is completely implausible as an unselected "slow death." This is a complex biologically programmed capability. There are extensive preparations during oogenesis for quiescence, including the disassembly of mitochondria, and shut down of oxidative phosphorylation, processes which are then reversed after storage is completed at ovulation. Thus, the ability to store stage 14 oocytes represents an evolved state molded by strong evolutionary pressure to maximize the opportunity for successful reproduction and to avoid wasting resources. Our experiments, which are carried out entirely in vivo and utilize the same protein limitation commonly experienced by wild flies, make it possible to precisely study this critical process of Drosophila reproductive biology for the first time. They also serve as a model for understanding oocyte storage in other species, where it likewise often plays a major role.

To ensure that readers understand these fundamental facts that underlie our research program, we have rewritten the Abstract to remove the misleading sentence, and added material in the Introduction describing the critical role of mature oocyte storage in Drosophila female reproduction.

Overall this study defines a potentially very interesting consequence of aging oocytes after oocyte maturation: a progressive decline in translation rate. But there are several important issues/questions that need to be addressed:

1) Is the phenomenon specific to aging oocytes? There are several reasons to be concerned that this is not the case.

First the authors utilize starvation to arrest the oocytes. While the premise is that late stage oocytes have already stockpiled the proteins and mRNAs it needs prior to the starvation, previous studies have shown that starvation induces changes that affect the entire organism (for example halting development of earlier oocytes and affecting life span).

Our protocol does not involve starvation, but protein limitation. Use of the term "starvation" was unfortunate lab jargon that we have now removed. The flies at all times have adequate water, sugar, micronutrients etc. as well as the extensive lipid reserves they stored in their fat bodies during their 24 hr of gorging on pure yeast. It is the lack of available sperm that is inhibiting ovulation, not the physiological condition of the females – they are healthy and begin actively laying immediately following mating even after prolonged protein restriction. While females have enough nutrients to live comfortably under our protocol, they do not have enough protein to make new eggs, and this approach has been used extensively to manipulate oogenesis in past studies without detectably impacting maternal or oocyte physiology. We have clarified these points in the text (subsection “A general method for studying Drosophila oocyte aging”).

In order to test whether there are detectable changes in oocytes stored in well-fed vs. protein-restricted animals, we have now compared the protein content of oocytes under each condition. As shown in new Figure 1E, individual oocytes contain 0.27ug of protein under well-fed conditions, and we observe no significant change in protein levels of oocytes stored in animals after 1 or 13 days of protein restriction. These data indicate that there is no signaling leading to reabsorption of protein from mature oocytes after they have formed.

We also looked for potential changes in gene expression induced by our protocol. We compared our ribosome profiling data of stored oocytes in protein restricted animals to prior measurements by the Orr-Weaver lab (Kronja and Orr-Weaver et al., 2014a) of oocytes from well fed animals. As shown in Author response image 1, we found that the translation levels of individual genes from oocytes stored for 2 days in our study are well correlated with translation of genes from non-stored oocytes from the Orr-Weaver study. This is despite significant differences in our ribosome profiling protocol, as oocytes were defolliculated in our experiments prior to lysis while the Orr-Weaver study used oocytes with their follicular layer intact. In addition, ribosome footprints were prepared with MNase in our study rather than RNase I as in the Orr-Weaver study. These data suggest that protein limitation does not substantially alter oocyte gene expression, so that changes seen in older oocytes are an indication of oocyte aging..

Author response image 1.

Author response image 1.

Open in a new tab

Variation in protein availability is a common natural phenomenon for Drosophila, and fly's dependence on transient protein sources has caused them to be well adapted to a regime of feast or famine with respect to protein. It is natural to have a burst of oocyte development followed by a period of protein starvation. This type of treatment has been heavily used in past research, for example in studies of ovarian developmental regulation by nutrients (Drummond-Barbosa and Spradling, 2001, and many subsequent publications by the Drummond-Barbosa lab). The slowed developmental rate caused by low protein nutrition causes no differences in the structure or physiology of the oocytes, is reversible, and is mediated by well-studied insulin, Tor and other pathways. Unlike a few insects whose egg size can be affected by maternal nutrition, no such changes occur in Drosophila melanogaster.

Finally, it actually does not matter if known or currently unknown pathways are induced as a result of protein limitation. These same effects would be induced in the wild when flies stop laying eggs as a result of environmental protein limitation. Whatever the processes are, they will impact oocyte aging in nature and in our model. We argue that it is critical to first study the effects of oocyte storage on the oocyte, before trying to decipher the pathways that may or may not contribute to those effects.

It is not clear if late stage oocytes are affected by the signaling cascades induced by starvation and if some of the changes observed are the oocytes response to the cue that food will be absent in the environment the oocyte will be deposited. This could be addressed at least partially by exploiting the fact that Drosophila females can be made to hold late stage oocytes for 4-5 days without starvation by preventing access to males. Examining such 4-5 day arrested oocytes without starvation would strengthen the argument that translational changes are due specifically to arrest.

We studied whether withholding males alone can cause females to hold eggs sufficiently tightly to allow experimentation for even 4-5 days as claimed. When analyzed quantitatively, we found that withholding males does not create a tight arrest of ovulation. As shown in Author response image 2, in the absence of males but with continual access to a rich protein source, flies lay (and therefore replace) ~15 eggs in 3 days and ~30 eggs by 4 days. Thus, after 3 days 25% of oocytes, and after 4 days, 50% of the initial stage 14 oocytes will have been replaced with younger oocytes, making these time point problematic and all subsequently times entirely unsuitable for study. Since our studies showed that even after 7 days of oocyte storage only about 20% of the MI spindles were defective, it is clear that withholding males is not a viable approach to studying the effects of oocyte storage on spindle function.

Author response image 2.

Author response image 2.

Open in a new tab

Second, the authors point out that in humans oocytes from younger women still showed increased abnormalities when implanted into older woman indicating mother's age may influence developed oocytes as well. Some of the upregulated genes and phenotypes they observed may be influenced by the age of the females (general aging) rather than actual oocyte arrest that would occur in other species. This could be addressed by examining stage 14 oocytes from fed and mated females that are aged 14 days for both translational and cytological assays. Using oocytes from 2 day starved females meant changes may have already occurred in response to starvation and/or arrest, oocytes from non-starved females is the better control.

We thank the reviewers for raising an interesting question: what is the relative contribution of maternal (general) aging vs. the intrinsic aging of the oocyte. We have now addressed this question in two ways. In order to discriminate between maternal age effects vs. oocyte age effects we compared the hatch rate of newly produced oocytes from young (5 day) versus aged (20 day) females. As shown in new Figure 1F, we found that the hatch rate of newly produced and aged oocytes was not significantly different (98.5% vs. 95.9% respectively, p = 0.114), showing that maternal age alone has no effect on oocyte quality, a reflection of the fact oocytes are continuously produced from stem cells. In contrast, held oocytes aged for 14 days starting when their mother was 2 days old had a drastically reduced hatch rate (13.8%, p = 3.8x10-4) even though the female laying them, now 16 days old, was younger than the 20 day old control flies.

In addition, we tested whether maternal age or oocyte storage time influenced the maintenance of bi-oriented meiotic spindles. As shown in new Figure 5E, whereas only 7/30 or 23% of 14 day old held oocytes from 16 day old females have bi-oriented MI spindles, 29/30 or 97% of newly produced MI arrested oocytes from 16 day old females had bi-oriented MI spindles, which is similar to the fraction of bi-oriented MI spindles from newly produced oocytes of young (3 day old) animals (28/30). The% of normal spindles correlated strongly with the hatch rate of equivalently aged vs. newly produced oocytes from the 16 day old females (Figure 5E). We conclude that the reduced ability of long-stored oocytes to support hatching is due to intrinsic oocyte aging leading to the loss of spindle bi-orientation, with little or no contribution from maternal aging. Descriptions of these experiments and their conclusions were added to the text.

Finally, it is unclear why the authors were making comparisons to immature follicles and embryos when the focus of the paper is written to focus on changes in arrested oocytes. These comparisons only indicate how a cell in metaphase I is different from cells either early in meiosis or mitosis. Studies have looked at these differences before and were only mentioned in passing.

We used the immature oocytes in order to compare the proteins translated by a growing oocyte with those produced by a non-growing, mature, "quiescent" oocyte undergoing storage. Oogenesis is a complex developmental program and the meiotic stage differences represent only a part of the program. We wanted to determine if a specific pattern of translation is associated with the ability of oocytes to remain functional while in a non-growing state. Ribosome profiling studies of growing oocytes suitable for comparison were not previously available. These experiments produced a number of candidate genes and the information will be useful to guide future research.

To understand specifically oocyte arrest the authors need to compare oocytes of the same stage but arrested for long periods compared to not arrested stage 14 oocytes. The current studies primarily show what is needed in a metaphase I oocyte.

We agree. The bulk of our paper is a comparison of mature oocytes arrested at stage 14 (metaphase I) that differ only in how long they have been stored in this state within the ovary.

2) Is the effect specific to translation?

The paper focuses specifically on translation, but it is unclear if the effect is specific to translation. Perhaps the oocytes are just slowly dying and the translational slowdown is a consequence of a general loss in cell viability, drop in ATP levels, etc… Several of the controls suggested in (1) above might address this, but in general it would be very helpful for the authors to demonstrate that the effects on translation are specific and that other metabolic activities are not affected or affected only secondarily, as a consequence of the translational shut down. Otherwise, the data could be interpreted as simply reporting on the slow death of oocytes caused by an artificially extended arrest that may or may not be physiologically relevant.

Stored oocytes are in a physiologically relevant, natural state as discussed above, so there is no danger that our results are not relevant to the real world. They are replete with enough nutrients to make a first instar larva, so it is hard to see why ATP levels would drop. Females are healthy and lack only the large supply of protein needed to make complete oocytes. Their fat bodies contain a rich supply of lipids for producing ATP and dietary sugars are available as well.

Determining a single “cause” of cellular failure is difficult for any aging system and in fact represents a holy grail of aging research. We went a lot farther than most aging studies by showing that the loss of developmental capacity in our system is due to a specific physiological process, meiotic chromosome segregation, a striking and unexpected result. We also found that translation was reduced broadly and affected all mRNAs to a greater or lesser extent, which raises interesting questions for future studies of aging in the absence of transcription.

It is unlikely that all or even most of these genetic changes had a significant functional impact on oocyte decline in the time frame studied, nor did we make such a claim. Instead, our finding that translation of meiotic spindle genes, several of which are dose sensitive, is reduced over time led to us to determine whether the meiotic spindles of stored oocytes became defective. Determining whether translation changes are solely responsible for the spindle defect or whether other factors are involved is of great interest to us and is an area of active research for a future publication. Our working hypothesis is that reduced translation in a cell that depends entirely on translational control of stored mRNAs will have a substantially negative impact on physiology.

Determining the exact cause of translational decline represents an interesting but entirely separate research subject that is beyond the scope of the current study.

A major conclusion of our paper is that reduced translation of meiotic spindle proteins is correlated with the increased inability of oocytes to maintain MI spindle. We found that with increasing storage time, failure to complete meiosis and/or very early mitotic divisions were the major reasons why oocytes no longer developed into hatching larvae. We have now added additional data to support this claim. New Figure 5E shows that hatch rate is strongly correlated to the fraction of oocytes with bipolar spindles as discussed above, and we have added a timepoint to Figure 5J showing that 98% of embryos from aged 17 day old oocytes fail to develop past the pre-blastoderm stage, consistent with early defects in chromosome segregation leading to aneuploidy. We hypothesize that reduced translation contributes to meiotic spindle defects, both directly due to reduced synthesis of spindle components and spindle assembly checkpoint proteins, as well as indirectly, by affecting cell cycle arrest and re-entry, which are controlled entirely post-transcriptionally. In addition, we provide a rich dataset in which to analyze the basis for translational decline during aging.

The loss of meiotic spindle bipolarity has also been observed both in aged mouse oocytes (Nakagawa and FitzHarris, 2017) and in oocytes from human IVF clinics (McCoy et al., 2015; Haverfield et al., 2017, Human Reproduction). That aged Drosophila oocytes and oocytes from aged mammals exhibit similar meiotic spindle polarity defects suggest that our studies are physiologically relevant for aging oocytes in humans and in many other species.

3) Throughout the paper, the authors rely on ribosomal profiling to measure translation. The data beautifully document that ribosome footprints/mRNA decline with age, but whether this means that translational output also declines is not a foregone conclusion. Ribosome footprints are not necessarily a direct measure of translational output: a decrease in ribosome footprints could mean a decrease in the rate of translation initiation (fewer ribosomes initiating, fewer ribosomes on message, less translational output) OR an increase in the rate of translation elongation (faster ribosomes, fewer ribosomes on message, more translational output). Without a separate assay to directly measure translational output, it is not possible to distinguish between these two options.

We have added new experiments supporting our interpretation of the footprint declines. If translation elongation rates increase over time rather than translation initiation decreasing, then protein production will also be increased rather than decreased during aging.

In order to differentiate between decreased initiation vs. increased elongation models, we generated new fly lines which ubiquitously express a rapidly degraded proteasomal N-end rule substrate, R-GFP, which allows one to infer instantaneous translation rates from the levels of nascent protein. We validated these lines by demonstrating that R-GFP levels are substantially lower than a stable M-GFP control line when expressed in Drosophila ovaries (new Figure 2E). We found that R-GFP levels were reduced by ~30% in aged (12 day old) vs. young (2 day old oocytes) as shown in new Figure 2F. This is consistent with the overall ~50% reduction in translation which varies to some extent from transcript to transcript, and inconsistent with the model that elongation increases. These data substantially support our (conventional) interpretation of the reduced footprinting results.

4) There are a number of other prior studies that need to be discussed in detail.

The prior work examining oocyte aging by the Bickel lab needs to be incorporated into the Discussion.

We mentioned work from the Bickel lab in the Introduction. The Bickel lab prevented access to males as a mechanism of oocyte arrest, but as described above this approach acts only over a few days and is leaky. A major difference between our work and prior work from the Bickel lab is that our study focuses on the aging of stored stage 14 wild type oocytes, while the Bickel lab has studied the effect of aging exclusively in mutant backgrounds. We found that the cause of premature oocyte failure in mutant backgrounds (i.e. sHSP mutants in this paper and Fmr1 RNAi oocytes in Greenblatt and Spradling, 2018) is often different from the meiotic spindle defects that cause failure of aged stage 14 oocytes in a wild type background. A strong indication of such physiological differences between our work and the research from the Bickel lab despite their focus on genes involved in meiotic segregation is the fact that the age-dependent non-disjunction they observed did not affect the first oocytes laid when repression was relieved, but only affected oocytes fertilized 16-32 hours later (Subramanian and Bickel, 2008). As the authors themselves conclude, this argues that the defects studied were not present in stored mature stage 14 oocytes, but only in much early oocyte stages, which had to mature before becoming detectable. Our studies do not bear on events at these earlier stages.

In Fredriksson et al., 2012, Hsp26 and Hsp27 were already found to be strongly expressed at 5 and 35 days of age in late stage eggs which the authors need to mention. The authors show expression of these genes by antibody in stage 10 oocytes but the authors needed to show these genes by cytology or quantitative western of stage 14 oocytes that were arrested for 14 days with starvation and from non-starved 14 day females. This would support that these genes are specifically up-translated in oocytes arrested long periods. The past and current data suggest the proteins are loaded prior to stage 14 oocyte arrest.

The reviewers are correct to note that sHSP expression in oocytes has been previously noted (both by Fredriksson et al., 2012 and previously by Zimmerman and Meselson et al., 1983). We have added these references to the paper. That small heat shock proteins continue to be translated in arrested oocytes – indeed Hsp26 is the 11th most highly translated protein in 2 day old arrested oocytes (Supplementary file 1) – is borne out in our ribosome profiling data. In addition, we find that not only are sHSPs expressed but also that Hsp26/Hsp27 are strongly required in oocytes during prolonged arrest (Figure 4C). Whereas newly produced oocytes lacking sHSPs have hatch rates only slightly lower than wild type, this effect is greatly increased with prolonged storage (99% vs. 86% for 1 day stored oocytes as compared to 90% vs. 26% hatch rate for 10 day stored oocytes). These data support our hypothesis that oocyte “pilot light” genes preferentially translated during arrest as opposed to early embryonic development are indeed important for supporting oocyte viability during prolonged storage.

If the authors are to continue to include the comparisons to mitotic embryos and early stage follicles their results should be compared and contrasted to the results looking at total protein changes by the Orr-Weaver lab.

This paper focuses on the storage of mature oocytes. The ribosome profiling studies of early embryos were simply one effort to find genes that are important for oocyte storage. Because the paper does not study early embryo development it would not be appropriate to enter into an analysis of data from other labs on a peripheral subject. These data are being put into the public domain, allowing such comparisons by any interested researchers studying the oocyte to embryo transition. We are not aware of any major differences, but our experiments provide a much greater depth of coverage and used replicates to improve statistical validity.

Specific suggestions:

The authors mention a number of genes that shows changes in translation but fails to always explain what the genes are and/ or provide references. Not all the readers interested in oocyte aging would be familiar with Drosophila gene names.

We added a reference to Flybase for information on specific Drosophila genes within the categories that we described in the text and called attention to the grouping of these genes in physiological categories in the legend to Figure 3C.

Figure 4. To study proteins that are preferentially translated in arrested oocytes and identify "pilot light" genes, the authors compare translation in 2-day old arrested oocytes to early embryos. However, since embryos must have a vastly different translational program than oocytes, it might be more enlightening to compare arrested oocytes to growing oocytes. It seems this experiment has already been done (Figure 4E, 4F) but further analysis of specific genes that are more highly translated in arrested oocytes compared to growing oocytes could potentially give more insight into other "pilot light" genes that contribute to meiotic spindle stability.

As the reviewer notes, we did compare arrested oocytes to growing oocytes, and to early embryos. The paper does discuss the identification of candidate "pilot light" genes using these datasets. However, this was carried out before we realized that the decline of wild type oocyte viability upon storage was fully explained by declining spindle function. It was before we discovered that the effects of storage on translation are very widespread and affect virtually all transcripts (Figure 2D) We also learned that individual genes are all likely to require functional study. For example, one such difference we observed, high expression of small heat shock proteins during stage 14, was relevant to prolonged storage, but not to MI spindle maintenance. Hsp26 and Hsp27 mutants were very different from Fmr1 mutants. We concluded that further bioinformatic analysis could become tedious in the absence of functional tests and might detract from the major message of the paper regarding spindle instability. We are providing these datasets for mining by other groups and we will continue to analyze genes identified in this manner in the future when genetic analysis can also be provided.

Figure 5. In humans, primordial follicles are stored for long periods of time in meiotic prophase before re-entering the cell cycle and briefly pausing again at metaphase II before fertilization. It is my understanding that this metaphase II pause lasts only 1-2 days in vivo. In this study, the authors show that mature Drosophila oocytes arrested at metaphase of meiosis I for long periods (> 6 days) exhibit meiotic spindle instability and abnormal chromosome segregation. They imply that this may be relevant to chromosome instability in human females ("Our findings show that storage of highly functional mature oocytes in vivo is sufficient to destabilize chromosome segregation, suggesting that the prolonged storage of mature oocytes may be an important source of meiotic chromosome instability in human females"). Whether mature human oocytes ever experience such a long arrest is not clear. A diagram that clearly compares oogenesis in Drosophila versus humans with the different arrest points could help clarify their argument.

Although widely assumed, the length of the metaphase II pause is not the only relevant issue with regard to late storage. Mammalian oocytes actually reach full size at the end of the primary follicle stage, which is 1-3 weeks before ovulation in mice and 40-85 days before ovulation in humans. During the secondary and especially at the antral stage, they are essentially being stored, while ongoing growth and changes in the follicle appear to be confined to its somatic cells. During much of this time it is likely that oocyte transcription continues. However, during part of the antral stage oocytes develop a SN or "surrounded nucleolus" configuration that is believed to the be an analog of the karyosome, and the nucleus becomes transcriptionally inactive at this point (known as "GV stage oocytes"). We added more information about this aspect of meiotic maturation as well as reference to a recent review (Conti and Franciosi, 2018). This period may be analogous to the stored stage 14 oocytes studied in our manuscript. Unfortunately, it is difficult to measure the length of the GV stage, and it may vary in length between individual follicles. Not enough solid information is currently available to make a reliable timeline to put in our paper. The important conclusion is that in mammals, damage to the spindle leading to meiotic non-disjunction might occur some fraction of the time during late storage, like we see in Drosophila.

In addition, humans have a menstrual rather than estrous cycle, dissociating ovulation from sexual behavior. Human MII arrested oocytes are therefore unlikely to be immediately fertilized following ovulation as suggested by the reviewers. Currently it is unknown how long MII arrested oocytes can survive prior to their fertilization. Our data suggests that lengthening the period of time between oocyte maturation and fertilization in which the oocyte must survive in the absence of transcription, which could be caused by delayed fertilization or defective ovulation, may contribute to chromosome instability.

Figure 6. The authors hypothesize that "One potential function of sHSPs might be to stabilize the meiotic spindle and thereby extend the functional lifetime of oocytes." However, they then seem to disprove this very statement in Figure 6D. Although loss of Hsp26 and Hsp27 has an impact on oocyte maintenance, the mechanism is unknown and as the final figure, this result is a bit disappointing.

See above discussion as to difficulty of determining the cause of decline in any aging system. We do not claim to have discovered the cause of aging, but an important previously unappreciated feature of aging that warrants further investigation. We agree and have restructured the paper by moving panels reporting the sHSP data either to Figure 4 or to the supplement.

Associated Data

    This section collects any data citations, data availability statements, or supplementary materials included in this article.

    Data Citations

    1. Greenblatt EJ, Obniski R, Michael C, Spradling AC. 2019. Ribosome profiling and mRNA sequencing of aging oocytes in Drosophila. NCBI Bioprojects. PRJNA57392

    Supplementary Materials

    Supplementary file 1. Translational and mRNA changes during oocyte aging.
    elife-49455-supp1.xlsx (3.6MB, xlsx)
    Supplementary file 2. ‘Pilot light’ genes translationally upregulated during oocyte arrest.
    elife-49455-supp2.xlsx (232.7KB, xlsx)
    Supplementary file 3. Genes upregulated during oocyte maturation.
    elife-49455-supp3.xlsx (78.1KB, xlsx)
    Supplementary file 4. Counts of total reads mapping to Dmel and Dpse for ribosome profiling and mRNA-sequencing experiments.
    elife-49455-supp4.xlsx (9.9KB, xlsx)
    Transparent reporting form
    elife-49455-transrepform.docx (249.4KB, docx)

    Data Availability Statement

    Data has been uploaded to BioProjects at NCBI under PRJNA573922.

    The following dataset was generated:

    Greenblatt EJ, Obniski R, Michael C, Spradling AC. 2019. Ribosome profiling and mRNA sequencing of aging oocytes in Drosophila. NCBI Bioprojects. PRJNA57392

    Articles from eLife are provided here courtesy of eLife Sciences Publications, Ltd

    RESOURCES