Maintenance of quiescent oocytes by noradrenergic signals

All females adopt an evolutionary conserved reproduction strategy; under unfavorable conditions such as scarcity of food or mates, oocytes remain quiescent. However, the signals to maintain oocyte quiescence are largely unknown. Here, we report that in four different species – Caenorhabditis elegans, Caenorhabditis remanei, Drosophila melanogaster, and Danio rerio – octopamine and norepinephrine play an essential role in maintaining oocyte quiescence. In the absence of mates, the oocytes of Caenorhabditis mutants lacking octopamine signaling fail to remain quiescent, but continue to divide and become polyploid. Upon starvation, the egg chambers of D. melanogaster mutants lacking octopamine signaling fail to remain at the previtellogenic stage, but grow to full-grown egg chambers. Upon starvation, D. rerio lacking norepinephrine fails to maintain a quiescent primordial follicle and activates an excessive number of primordial follicles. Our study reveals an evolutionarily conserved function of the noradrenergic signal in maintaining quiescent oocytes.

For C. elegans evidence comes from altering two enzymes required for OA synthesis and a receptor identified here as relevant. Phenotypes included mitotic division of haploid oocytes after sperm exhaustion in hermaphrodites and in genetically feminized animals that produce no sperm. A similar phenotype was seen in females of C. remanei where there are natural females. A few shortcomings and questions are likely answered relatively easily: Are Emo phenotypes seen for older hermaphrodites and females also when males are present? What explanation might be offered for the limited rescue by OA (Fig. 2c)? What assurances can be offered that tbh-1 and tdc-1 dysfunction affect only OA synthesis? Most of the data in Fig. 3 lacks clear quantitation, including the proportion of oocytes with Emo phenotypes in feminizing tests. The penetrance of the latter phenotype appears to be quite low in fog2; tbh1 mutants relative to other situations (and no DAPI is shown for these). Any explanations? Are feminized C. elegans with OA defects sterile, like C. remanei and is there a significant difference in the timing (onset) of Emo responses presumed to cause sterility in C. remanei? Rescue with SER-3 using a ceh-18 promoter is used to suggest gonadal sheath cells as the cells receiving OA signals. Are there additional cells that express ceh-18 and might plausibly be important? Are gonadal sheath cells important for any part of the flux of oocytes contributing to oogenesis, fertilization or egg-laying, potentially co-ordinating such roles with influencing oocyte quiescence?
In Drosophila the major phenotype reported is the progression of oogenesis beyond stage 8 in young adults starved for protein for tbh mutants and an identified receptor, octb2R. A strong aspect of these results is the robust rescue of the former, but not the latter mutant phenotype by provision of OA (Fig.  5c). There is a notable weakness or curiosity to be explained better. Protein starvation elicits arrest prior to stage 8 and degeneration of stage 8 egg chambers, and the authors state that if degeneration is blocked that maturation continues. I am not sure of the origin of this statement (it is not referenced) but it suggests that regulation of progression beyond stage 8 is entirely dependent on a decision to initiate degeneration or not. Yet the authors main thesis is (or appears to be) that interfering with OA signaling does not affect degeneration but does promote progression. The authors should be clearer about the context of current understanding and its origins, and whether they are proposing that current understanding is incorrect and that there is in fact an OA-enforced growth regulation beyond stage 8. If the main thesis is instead that arrest prior to stage 8 is absent in OA-deficient animals then I think the phenotype of earlier egg chambers should be reported quantitively with respect to characteristic P granules and microtubules. Indeed, it is in any case a central issue to establish which starvationsensitive checkpoints are or are not affected by OA (that includes effects on GSC division). Currently, the focus of OA action and supporting evidence is not clear to me. With regard to data, I think the various presentations in Fig. 4b-e are potentially confusing. I presume that b, e, f are per fly and that, because stages 2-7 might be evenly represented and are aggregated in b, the average frequency of each of those stages in most genotypes shown is about 14-15.
Assuming that d-f show non-degenerating egg chambers, the frequency of intact stage 8 chambers seems remarkably high (for controls, given that both prior arrest and degeneration should reduce their frequency). Does stage 8 degeneration initiate only after a significant delay, at least for some egg chambers? When degeneration is explicitly measured and reported in Fig. 5d the age of flies is not given but the big difference between mutant and control st. 8 intact egg chambers is unexpected because it was not apparent in Fig. 4. Also, if these data are accurate, then the proportion of stage 8 egg chambers undergoing degeneration is much lower for mutants than controls, indicating a defect in degeneration. Indeed, it is hard to understand how so many egg chambers proceed further unless they are spared from degeneration. Also, the authors note that the investigated OA receptor is expressed in nurse cells but that remains a speculation regarding site of action since that was not investigated experimentally. It would also be helpful to know whether the allele used (and the tbh allele) is known to be a null mutation or its molecular basis.
The zebrafish studies use one mutant gene (and allele) and one phenotypic measure, and are therefore more limited with regard to site of action, receptor or assurances that only NE action (in a relevant location) is being disrupted.
The broader context of the study as a whole really concerns what was known before, what is now understood better and what remains unresolved. Here I am being guided largely by the authors' narrative. Mostly, I found the narrative to be informative, especially given the challenge of describing relevant background for three experimental organisms. One sticking point for me, which is perhaps most readily phrased for the Drosophila studies, is whether the authors are suggesting or concluding that there is an active response to protein deficiency (as opposed to the absence of a response to protein sufficiency). The authors say it is not known whether the absence of insulin and JH signals is sufficient to maintain quiescent ovaries or another inhibitory signal is necessary. This needs discussion or refinement as my understanding is that several (maybe all?) facets of quiescence are observed in the absence of those signals (and in some cases are reversed by addition). Later (p12) the idea is repeated, suggesting that there is indeed also a nutrient-responsive neuronal signal, followed by a chain of conditional sentences about expected scenarios. Do the authors have any evidence that OA presentation to relevant cells is altered in response to nutrition? Also, have the authors tested whether OA signaling disruption alters insulin peptide secretion or responses or whether one type of signaling overrides the other (under genetic conditions that permit such tests)? It appears to me that the idea of a nutrient-responsive signal is at this stage speculative and that it is not yet clear whether OA responses act independently of insulin-like signals. Another significant unknown (in each setting, possibly with the exception of C. elegans) appears to be the exact cells responsible for receiving the OA signal, how they then relay the signal and what might be the significance of those choices (why they are suited to the purpose). The evidence in Fig. 1 does not really answer these questions and I am particularly unconvinced from the images about the statement that the stained processes reach germarial regions or why that may be relevant for the reported phenotypes.
Another issue that seems debatable is what is the significance of finding somewhat related roles for OA and NE in three organisms. Much about how OA acts and in which cells remains to be discovered and the manifestations of failures of quiescence and unfavorable conditions examined are quite diverse. It is therefore hard to assess what might be the evolutionary mechanisms and potential common advantages of employing OA/NA. Altogether, I expect that the authors will be able to resolve most of the issues I have raised with more careful explanations and making available quantitative data more explicit. While I am not personally convinced that investigation of three systems makes the conclusions especially noteworthy, it does seem that there are robust data for more than one system supporting some new insights into an important area of biology that will likely not be very familiar to most readers but nonetheless interesting and informative.
Reviewer #3 (Remarks to the Author): Remarkably, it is shown in this manuscript that Octopamine/Norepinephrin regulates oogenesis quiescence during starvation conditions in four species, from C.elegans, Drosophila, to zebrafish. It is well known that oogenesis undergoes quiescence during starvation to save nutrients for survival. However, how the quiescence is mediated molecularly was not known. In C.elegans and Drosophila, the authors found that oocytes progressed in oogenesis in starvation or protein-deficient conditions when OA (Octopamine) fails to be produced, unlike wild type, which arrest in their development at previtellogenic stages. Importantly, exogenous OA could rescue the defect, demonstrating the specific role of OA in mediating this quiescence. In both C.elegans and Drosophila, the authors also identified the receptor of OA, which exhibited the same phenotype as the OA-deficient animals. In zebrafish, they show that lack of norepinephrine (OA in vertebrates) causes the same defect that oogenesis proceeds even during starvation conditions. The data strongly support the conclusions.
The authors hypothesize that it is the balance of OA to nutrient signals (Insulin related) that mediates either quiescence or oogenesis progression. It would add important depth to this study, if it were possible to test this in one of the organisms. In Drosophila and C.elegans, the nutrient signaling pathway is well studied, so it wouldn't be too difficult to test this model. We deeply appreciate the constructive comments from the reviewers. We believe that we addressed all the concerns and that the revised manuscript is improved due to those comments.
The summary of the changes: • One of the major concerns was how OA signal fits into the known oocyte maturation process in C. elegans (i.e., how OA interacts with the sperm signal that controls oocyte maturation). To address this issue, we added new results from three additional experiments, where we examined MAPK activation in OA mutants (now Fig. 2c), constructed the double mutants of fog-2; tbh-1, and determined the genetic interaction between the sperm signal and OA by measuring ovulation rates and scoring the Emo phenotype (now Fig. 3c and 3d). We also confirmed the double mutants did not maintain quiescent oocytes even when they contained stacked oocytes, by mating the double mutants with males and showing that the oocytes become Emo (now Fig.  4). We confirmed that the defect is only in maintaining quiescence in the absence of sperm; when we mated the double mutant at the L4 stage so that sperm are available before the first oocyte matures, the double mutant produced intact embryos without producing Emo (now Fig.  4f).
• Another major concern was whether OA signal could be an active signal to override the nutrient signals. We added new results from experiments in D. melanogaster where we examined how OA and nutrient signals interacted using different combinations of OA and nutrients (now Fig  6e-g). When we treated flies with OA and nutrients varying the concentration of each, OA and nutrients balance each other, indicating that OA is an active signal and thus supporting our hypothesis; A medium concentration of OA (5 mg/ml) inhibits a low level of nutrient signaling (1% yeast). This OA effect is overridden by a high level of nutrient signaling (2.5% yeast). A high concentration of OA (10 mg/ml) inhibits a high level of nutrient signaling (2.5% yeast).
• We revised the text and removed the part regarding egg chamber degradation in Drosophila to clarify ambiguity and avoid distraction. We removed the part because the idea that nondegenerated egg chambers would become matured egg chambers was indeed a baseless assumption, and therefore the increased number of matured egg chambers in OA mutants is not due to failure of degeneration. Indeed, even if there are variations in numbers, we always observe a significant number of degenerating egg chambers in OA mutants, showing that the degeneration process does occur. Most importantly, our new results show that the stage 8 egg chambers of the OA mutants actively undergo vitellogenesis whereas the stage 8 egg chambers of the control barely do. If the increased number of vitellogenic egg chambers in OA mutants were due to a defect in degeneration, the ovaries of OA mutants should contain an increased number of stage 8 egg chambers with the same appearance of that of the control (pale with reduced vitellogenesis). This difference clearly indicates that the increased number of matured egg chambers in OA mutants is not due to defects in degeneration but due to failure in remaining quiescent. Therefore, regardless of any concerns about degeneration, our conclusion remains valid especially after our newly added experiments (now Fig. 6e-g).
The reviewer's comments are shown in blue and our responses are in black.
2 Reviewer #1 (Remarks to the Author): Summary/Critique This manuscript from Kim and colleagues reports evidence from three biological systems, nematodes, Drosophila, and zebrafish suggesting that octopamine signals from cells of the somatic gonad help maintain oocyte quiescence. A strength of the manuscript is the investigation in multiple species. This is also a weakness because the experiments are very preliminary, somewhat superficial, and largely descriptive. The studies do not leverage the wealth genetic, cell biological, and mechanistic information and tools available from longstanding studies of oogenesis in these systems. This is particularly true of the studies in Drosophila and zebrafish. The C. elegans studies are better and might stand alone with additional clarification and better supporting evidence as discussed in Major Points below.
Major Points 1. The Drosophila and zebrafish studies would benefit from molecular endpoints and better integration with the literature in these fields.
We appreciate the reviewer's comment, and revised the introduction by adding oogenesis of zebrafish and discussing common molecular pathways. However, because there is not much known about 'quiescence' signaling especially in zebrafish, and also because there are many diverse mechanisms due to the big differences among the three models' reproductive strategies and environments, it is not easy to integrate the fields in terms of molecules. In addition, oocyte maturation, especially in response to endocrine signaling, has been intensively studied by many excellent groups. Our discovery suggests that OA/NE acts prior to those signals as a safeguard to protect the oocyte pool when the environment is harsh and unpredictable. We hope the reviewer understands our effort to focus on what is unknown related to our discovery instead of reviewing the vast amount of previously reported findings.
2. Regarding the C. elegans studies, instead of using RNAi, the authors should construct double mutants between strong loss-of-function mutations in fog-2 (e.g. oz40) and null alleles of tbh-1 and tdc-1. The concern is that the authors will observe that oocytes stack up in the germline and fail to undergo spontaneous meiotic maturation. The reviewer is aware that others have made these strains and obtained that result.
We agreed and generated the double mutant fog-2(q71); tbh-1(n3247). The double mutants showed the same phenotype as that of the RNAi in first 10 hours from late L4s; they failed to contain quiescent oocytes and ovulated at a faster rate than control (Fig. 3c). We also noticed, as the reviewer mentioned, about 40% of the double mutants did contain stacked oocytes whereas, the remaining 60% did not. However, when we mated them with wild-type males, the stacked oocytes became Emo or produced unhatched embryos, indicating the stacked oocytes were defective and not quiescent ( Fig. 4c-f).
The new results are described as: "Consistent with the RNAi results, fog-2; tbh-1 double mutants failed to maintain quiescent oocytes from as young as 10 h after L4, and the oocytes became round instead of remaining cylinder-shaped as in fog-2 single mutants (Fig. 3c). fog-2; tbh-1 double mutants laid oocytes at a higher ovulation rate than fog-2 single mutants ( Fig. 3d) and became Emo at 1 d (19% Emo, n = 31), 2 d (50% Emo, n 3 = 24) and 3 d (67% Emo, n = 24) after L4. In contrast, fog-2 females maintain quiescent oocytes and produce no Emo for 3 days (0% Emo, n = 35)." and "We observed that fog-2; tbh-1 double mutants occasionally stacked oocytes in their gonad arms. To determine whether those stacked oocytes are quiescent and intact, we selected 1-d old females that contained stacked oocytes (42%, n = 31) and mated them with wild-type males. Mated fog-2; tbh-1 mutants produced the Emo phenotype within 5 h of mating (76.5% Emo, n = 17), whereas mated fog-2 females or OA-treated fog-2; tbh-1 did not (0% Emo, n = 13) ( Fig. 4c-e)." 3. The authors should not measure the formation of endomitotic oocytes as an endpoint, but should directly measure rates of meiotic maturation and ovulation. A concern is that the authors are looking at subtle effects on the spontaneous meiotic maturation rate at late time points.
We agreed on the concern and measured the rates of meiotic maturation by counting ovulated unfertilized oocytes. The new data are in Fig. 3d. Specifically, to avoid any sampling errors, we measured oocyte maturation rate within 1 d from L4 before over 50% of the females carry Emo (females carrying Emo cannot lay eggs). The oocyte maturation rate of fog-2; tbh-1 double mutant females is approximately 5 times higher than that of fog-2 control females. We think this is an obvious phenotype (not subtle) as a similar fold change was reported when ceh-18(mg57) mutant females were compared to fog-2(q71) or fog-3(q443) in the previous study (Govindan et al. Current Biol. 16, 1257-1266, 2006. 4. The studies should employ molecular readouts of meiotic maturation signaling and better attempt to fit tbh-1 and tdc-1 into a genetic pathway. What happens when double mutants are made with genes required in the gonadal sheath cells for oocyte meiotic maturation (e.g., gsa-1 and acy-4).
With all due respect, although we appreciate the reviewer's insight and suggestion, and tested gsa-1 interaction with OA signal, we do not think the OA signal directly interacts with the sperm signal. Rather, our study suggests the two events of 'awakening' and 'sperm signal-mediated maturation' are sequential and independent, sharing aspects of oocyte maturation process in all other male/female animals, where oocyte awakening happens before the oocyte meets sperm. As a hermaphrodite, C. elegans does not need the 2 nd meiotic arrest, the waiting step for sperm, and thus the two events of 'awakening' and 'sperm signal-mediated maturation' look connected. However, none of OA mutants show any defects in oocyte maturation in the presence of sperm, and the double mutant of fog-2; tbh-1 only produces Emo when sperm are unavailable, or when sperm are introduced at a late adult stage (Fig. 4f). This indicates that the OA signal is not necessary for 'sperm signal-mediated maturation' step. Indeed, our result ( Fig. 4) suggests the two signaling pathways are independent.
The new result was described as: 4 "Next, we examined the genetic interaction between OA and sperm signaling by comparing the ovulation rates between fog-2; tbh-1 and fog-2 under the condition of reduced sperm signaling of Gas, which is encoded by gsa-1. We reasoned that if OA directly interacts with sperm signaling, the gsa-1 phenotype would be epistatic to that of tbh-1 in maintaining quiescent oocytes; the phenotype of gsa-1RNAi would be the same as that of gsa-1RNAi in the tbh-1 background. Because Gas is also required for spermatheca contraction, the oocytes of gsa-1 RNAi-treated fog-2; tbh-1 animals became trapped in the spermatheca, where they matured and became Emo in the gonadal sheath, whereas the oocytes of fog-2 single mutant did not (Fig. 4a, b). Although it is unclear why oocytes of fog-2 mutants were able to pass spermatheca in gsa-1 RNAi treatment, the fact that gsa-1 RNAi produces different phenotypes between fog-2 and fog-2; tbh-1 indicates gsa-1 is not epistatic to tbh-1." 5. The authors model isn't clear because both tdc-1 and tbh-1 are expressed in gonadal sheath cells in both the presence and absence of sperm.
We apologize, but this concern is not clear to us. If the reviewer suggests OA production has to be induced only in the absence of sperm, then we think OA should be present as a default signal to maintain quiescence oocytes. We are not aware of signaling pathways that allow the C. elegans hermaphrodite to gauge the sperm level and to control their OA production. Our new experiment where we show fog-2; tbh-1 mutants produced Emo faster after mating also suggests that maintaining quiescent oocytes before meeting sperm is critical.
6. The discussion of the mammalian system in the Introduction could be improved by consulting an expert in mammalian oogenesis.
We really thank the reviewer for this comment. We have revised the Introduction.
The authors investigate effects of reducing or eliminating octopamine (OA) signaling on Drosophila and nematode oogenesis, and inhibiting noradrenalin (NA) signaling in zebrafish. In all cases they report a marked consequence of releasing oocyte development or activation that is otherwise suppressed by conditions unfavorable for reproduction, while reproduction under optimal conditions is largely unaffected. The authors therefore infer a general role for OA/NA in preserving oocyte "quiescence" under unfavorable conditions.
Here I consider the quality of the evidence and how the results fit into a broader picture.
For C. elegans evidence comes from altering two enzymes required for OA synthesis and a receptor identified here as relevant. Phenotypes included mitotic division of haploid oocytes after sperm exhaustion in hermaphrodites and in genetically feminized animals that produce no sperm. A similar phenotype was seen in females of C. remanei where there are natural females. A few shortcomings and questions are likely answered relatively easily: 1. Are Emo phenotypes seen for older hermaphrodites and females also when males are present?
We tested this idea in a feminized mutant (fog-2) and C. remanei females. The current understanding is that constant present of sperm signal prevents Emo by coordinating oocyte maturation and ovulation. Consistent with this understanding, both the fog-2 C. elegans and female C. remanei without OA did not produce Emo phenotype if they are mated from the final larval (L4) stage -and thus guarantee the constant sperm signal. In this case, 3-d old females did not produce Emo. However, they produce Emo if males are present after the female has been an adult for at least one day. The fog-2; tbh data are presented in Fig. 4f and C. remanei data are mentioned in the text.
2. What explanation might be offered for the limited rescue by OA (Fig. 2c)?
We appreciate this comment. We found the original graph included (inadvertently) all the time points including several earlier timepoints we used to observe the time course. When we corrected this error and plotted the graph with the data from 89 h after hatch, as written in the figure legend, the analysis showed that exogenous OA rescues Emo phenotype in both tbh-1 and tdc-1 mutants nearly completely (Fig. 2d). The full rescue on Emo phenotype by OA is seen repeatedly in other backgrounds (Fig. 3c, d and Fig. 4e, g).
3. What assurances can be offered that tbh-1 and tdc-1 dysfunction affect only OA synthesis?
It is our understanding that as these are enzymes to process octopamine and tyramine, and we cannot think of any other biological processes these mutants would affect. If the concern is about potential tyramine function, our experiments adding octopamine that rescued all of the phenotypes that we examined suggests that the oocyte phenotype is due to lack of octopamine.
4. Most of the data in Fig. 3 lacks clear quantitation, including the proportion of oocytes with Emo phenotypes in feminizing tests. The penetrance of the latter phenotype appears to be quite low in fog2; tbh1 mutants relative to other situations (and no DAPI is shown for these). Any explanations?
Thank you. To address this and the reviewer #1's concerns, we generated a feminized mutant lacking OA (fog-2; tbh-1). This allows us to precisely quantify relevant phenotypes including Emo (Fig. 3c-e). In summary, (a) we measured an oocyte maturation rate in the fog-2; tbh-1 mutant 6 females and found it is higher than fog-2 mutant females ( Fig. 3c-d). (b) we quantified the fraction of Emo, stacked oocytes, and awaken oocytes in the fog-2; tbh-1 females and mated fog-2; tbh-1 mutants and found most of stacked oocytes (even before they show apparent Emo) are not functional (Fig. 3c-e; Fig. 4a-g).
Although DAPI is better to examine large number of animals for Emo phenotype, it cannot be used to examine live animals (since it requires fixation) and thus cannot provide high-resolution DIC images necessary to observe the oocyte maturation process.
5. Are feminized C. elegans with OA defects sterile, like C. remanei and is there a significant difference in the timing (onset) of Emo responses presumed to cause sterility in C. remanei?
Thank you for the insightful question. While we are addressing the question of, "Are Emo phenotypes seen for older hermaphrodites and females also when males are present?", we also address the timing issue. Specifically, (a) when fog-2; tbh-1 females are mated with males from late L4 so that sperm will be present as soon as they become young adults, they are completely normal and fertile (Fig. 4f). (b) However, when 1-d old fog-2; tbh-1 females were mated with wild type males, 76% of females produced Emo. These data implicate that even if a certain fraction of 1-d old fog-2; tbh-1 females accumulate oocytes, those stacked oocytes are not quiescent, and thus cannot undergo normal oocyte maturation, resulting in Emo ( Fig. 4c-d).
6. Rescue with SER-3 using a ceh-18 promoter is used to suggest gonadal sheath cells as the cells receiving OA signals. Are there additional cells that express ceh-18 and might plausibly be important?
Thank you. We clarified, "SER-3 is expressed in head muscles, several neurons, intestine, spermatheca and gonadal sheath cells. To locate the SER-3 action, we targeted ser-3 expression to gonadal sheath cells using a ceh-18 promoter. CEH-18 is a Pit-1/Oct-1,2/Unc-86 (POU) domaincontaining transcription factor required for gonadal sheath cell differentiation. Although CEH-18 is broadly expressed including in muscles, neurons, and the gonads 41 , in the gonad, it is only expressed in gonadal sheath cells and not detected in sperm or oocytes. In addition, the only tissues where both ceh-18 and ser-3 are expressed are gonadal sheath cells and spermatheca. This construct rescues the Emo phenotype in ser-3 mutants (Fig. 2e, ser-3 TG). Thus, we suggest SER-3 function in gonadal sheath cells is sufficient to maintain quiescent oocytes." 7. Are gonadal sheath cells important for any part of the flux of oocytes contributing to oogenesis, fertilization or egg-laying, potentially co-ordinating such roles with influencing oocyte quiescence?
Thank you for an insightful and important question. It is known that the gonadal sheath cells communicate with the oocytes via gap junctions and are involved in ovulation by mechanically pushing oocytes toward to the spermatheca. However, mutants simply causing slow contraction (i.e. egl-30 which encodes Gaq) do not produce Emo, although they lay significantly fewer progeny than wild-type (Govindan et al., Development 136, 2211Development 136, -2221Development 136, , 2009Brundage et al., Neuron, 16, 999-1009, 1996. Considering that the wild-type and egl-30 hermaphrodites have the same amount of sperm, and that reduced sheath contraction itself does not produce Emo, it is implicated that oocyte maturation and the rate of the gonadal sheath contraction are well-coordinated in these mutants. The detailed mechanism especially whether or how any such coordination would affect oocyte quiescence in specific, however, is not well understood. Currently, we do not feel it is the focus of our manuscript and hope to address it in the future. 7 8. In Drosophila the major phenotype reported is the progression of oogenesis beyond stage 8 in young adults starved for protein for tbh mutants and an identified receptor, octb2R. A strong aspect of these results is the robust rescue of the former, but not the latter mutant phenotype by provision of OA (Fig. 5c). There is a notable weakness or curiosity to be explained better.
Protein starvation elicits arrest prior to stage 8 and degeneration of stage 8 egg chambers, and the authors state that if degeneration is blocked that maturation continues. I am not sure of the origin of this statement (it is not referenced) but it suggests that regulation of progression beyond stage 8 is entirely dependent on a decision to initiate degeneration or not. Yet the authors main thesis is (or appears to be) that interfering with OA signaling does not affect degeneration but does promote progression. The authors should be clearer about the context of current understanding and its origins, and whether they are proposing that current understanding is incorrect and that there is in fact an OA-enforced growth regulation beyond stage 8.
Thank you for pointing out this important issue. We do not suggest a novel mechanism of OA in the progression beyond stage 8 egg chambers, or its role in degeneration. Please see our response to the next.
If the main thesis is instead that arrest prior to stage 8 is absent in OA-deficient animals then I think the phenotype of earlier egg chambers should be reported quantitively with respect to characteristic P granules and microtubules.
The characteristic P granules and microtubules under protein-starvation condition were mainly observed in stage 6/7 egg chambers (Shimada et al, 2011, Burn et al, 2015. However, as shown in the images below and explained next, stage 7 egg chambers were rarely seen in the tbh -/-mutant ovaries under starvation. This impedes any comparison of the distribution of the P granule marker (Ypsilon schachtel: Yps) and Tubulin between the controls and tbh -/-mutants.
Indeed, it is in any case a central issue to establish which starvation-sensitive checkpoints are or are not affected by OA (that includes effects on GSC division). Currently, the focus of OA action and supporting evidence is not clear to me.
The currently known starvation check points are GSC and stage 6/7 egg chambers. Because it takes approximately 7 days from a germ line stem cell (GSC) to develop to a stage 2 egg chamber (Spradling, 1993), and because all our OA experiments were done using flies in 2 dpe (2 days after eclosion), we can safely exclude the possibility that the increased number of vitellogenic egg chambers (~5´ more than control in tbh -/-mutants, Fig 6b, d) in OA mutants was from any potential increased number of GCS.
We tested a complete starvation condition where we removed all nutrient sources including sugar and cornmeal. This complete starvation blocks almost all egg chambers from awakening in control (Canton-S) of age of 2 dpe (Fig. 6d, n=15). Most of the egg chambers of Canton-S remained at stage 6/7 or earlier stages, confirming that stage 6/7 is a starvation checkpoint (marked as Q in the below Figure R1, R2). The egg chambers of tbh -/-mutants, however, contain fewer stage 6/7 egg chambers but instead many vitellogenic egg chambers including stage 14 egg chambers (Fig. 6d,  n=56), confirming the role of OA in maintaining the stage 6/7 egg chambers.
To support our explanation, we provide below images of oogenesis of Canton-S and tbh -/-mutants under complete starvation. Because Fig. 6a already shows the clear difference in oogenesis between 8 Canton-S and tbh -/-mutants under the protein starved conditions, we did not add them to the main figure, to avoid repetition. (Left) The ovary of Canton-S contains five "quiescent stage 7 egg chambers" (Q), which can be determined by the size of stage 7/8 egg chambers and lack of vitellogenesis, and four small vitellogenic egg chambers (*) at 2 dpe, suggesting low level of spontaneous vitellogenesis under complete starvation at 2 dpe.

Figure R1. Difference in distribution of quiescent egg chambers between Canton-S and the tbh
(Right) In contrast, the ovary of tbh -/-mutants contains eight vitellogenic egg chambers (4 of s14, 2 of s13, 1 of s12, 1 of s10B; some egg chambers are out of the frame), four stage 9 egg chambers (#), four small vitellogenic egg chambers (*) as well as at least 5 degenerating stage 8/9 egg chambers (*), suggesting 5 times more awakened egg chambers (total 21) than that of Canton-S control.
9. With regard to data, I think the various presentations in Fig. 4b-e are potentially confusing. I presume that b, e, f are per fly and that, because stages 2-7 might be evenly represented and are aggregated in b, the average frequency of each of those stages in most genotypes shown is about 14-15.
We clarified in the figure legends as: "b: The number of previtellogenic egg chambers (stages 2-7) per fly" "d-f: The distribution of each stage egg chambers in the ovaries of each tbh +/-and tbh -/-female virgin" Assuming that d-f show non-degenerating egg chambers, the frequency of intact stage 8 chambers seems remarkably high (for controls, given that both prior arrest and degeneration should reduce their frequency). Does stage 8 degeneration initiate only after a significant delay, at least for some egg chambers?
The reviewer is correct on both accounts. Fig. 4d-f (now Fig. 5d-f) show non-degenerating egg chambers, and although our data are insufficient to clearly address the kinetics of stage-8 degeneration, our observation and estimation suggests a significant delay in degeneration in control. Please see our response to the next concern.
When degeneration is explicitly measured and reported in Fig. 5d the age of flies is not given but the big difference between mutant and control st. 8 intact egg chambers is unexpected because it was not apparent in Fig. 4. Also, if these data are accurate, then the proportion of stage 8 egg chambers undergoing degeneration is much lower for mutants than controls, indicating a defect in degeneration. Indeed, it is hard to understand how so many egg chambers proceed further unless they are spared from degeneration. Fig. 4 (Fig. 5d-f

in the revised manuscript):
Due to technical difficulties to measure kinetics which requires time course of oogenesis ideally in each ovariole, we compared the total number of all countable previtellogenic egg chambers per fly at eclosion at 1 dpe, 1.5 dpe, and 2 dpe between tbh -/-and tbh +/-mutants to gain snapshots of the progress. Although we understand the reviewer's concern regarding that showing duplicate data could be confusing, we also feel individual data would be informative. Fig. 4 and Fig. 5d (we removed):

Regarding degeneration kinetics concerning
According to the literature, at eclosion, an ovary contains 16 ovarioles. Each ovariole contains one stage 6/7 previtellogenic egg chamber as the most advanced. It would grow to stage 14 at 1.5 dpe in the presence of nutrients. The calculation is based on the 34.5 h to progress from stage 6 to 14 (Spradling, "Developmental genetics of oogenesis", 1993). Under the starvation conditions, Canton-S flies rarely undergo vitellogenesis; only 4 vitellogenic egg chambers are identifiable at 2 dpe (Above image: Left, *). They are all stage 8, suggesting arrest of oogenesis process arrests and delayed degeneration.
In our comparison between tbh -/-and tbh +/-mutants under protein starvation conditions, the ovaries of each tbh -/-fly contains 18 vitellogenic egg chambers in average at 1 dpe. The number of vitellogenic egg chambers is the sum of stages 8 to 12 egg chambers. At 1.5 dpe, 7.3 stage 14 egg chambers appear in average. Because it takes approximately 0.5 d from stage 10 to stage 14, we estimate almost all vitellogenic egg chambers of stage 10-12 (7 in average) in 1 dpe grow to stage 14. This is consistent with the literature that there is no halt or degeneration after stage 10 in growth of egg chamber. In contrast, the difference between the average number of the stage 8/9 egg chambers at 1 dpe (11 in average) and the number of stage 10-13 at 1.5 dpe (5.7 in average) which were supposed to grow from the stage 8/9 at 1 dpe. From these data approximately degeneration of 5.3 egg chambers is deduced (11 stage 8/9 at 1 dpe -5.7 stage 10-13 at 1.5 dpe = 5.3 degeneration).
The ovaries of each tbh +/-contains 16 vitellogenic egg chambers (sum of stage 8 to 11 egg chambers, no stage 12 appears) in average at 1 dpe. At 1.5 dpe, no stage 14 egg chambers in average appears. The difference between the average number of the stage 8/9 egg chambers at 1 dpe (14 in average) and the number of stage 10-13 at 1.5 dpe (5 in average) which were supposed to grow from the stage 8/9 at 1 dpe. From these data, approximately degeneration of 9 egg chambers is deduced (14 stages 8/9 at 1 dpe -5 stages 10-13 at 1.5 dpe = 9 degeneration).
The above estimation suggests potentially reduced degeneration in tbh -/-compared to tbh +/mutants (5.3 vs 9). However, we do not know whether the difference is significant. In theory, the pool of stage 8 egg chambers would be reduced by increase of degeneration and/or by increase of progression to the next stage. When it is impossible to trace a single egg chamber for its oogenesis progress, it is hard to discern the two possibilities. Nonetheless, we clearly see substantial number of degenerating egg chambers in tbh -/-, indicating that degeneration process does occur. Most importantly, the stage 7/8 egg chambers of Canton-S appear pale, whereas those of tbh -/-appear dark, indicating active vitellogenesis in tbh -/-mutants. This difference cannot be explained by failure of degeneration; if so, tbh -/-mutants should contain the same pale stage 7/8 seen in Canton-S with an increased number. Instead, the egg chambers in tbh -/-mutants actively progress to vitellogenic stages compared to either tbh +/-or two controls (Canton-S, w 1118 ). This strongly supports our conclusion that OA is required to maintain quiescent egg chambers of stage 6/7. Based on this deduction and that the assumption that non-degenerated egg chambers would become matured egg chambers was baseless, we removed Fig. 5d of the original manuscript to avoid distraction.
10. Also, the authors note that the investigated OA receptor is expressed in nurse cells but that remains a speculation regarding site of action since that was not investigated experimentally. It would also be helpful to know whether the allele used (and the tbh allele) is known to be a null mutation or its molecular basis.
Thank you. We revised: (a) "tbh nM18 a null allele that does not produce OA", and cited again the paper by Monastirioti M, Linn CE, Jr., White K. Characterization of Drosophila tyramine beta-hydroxylase gene and isolation of mutant flies lacking octopamine. J Neurosci 16, 3900-3911 (1996).
(b) "An in situ hybridization study shows octβ2R expression in the nurse cells of previtellogenic egg chambers, suggesting it might be a receptor for the OA action in oocyte quiescence. The allele octb2R f05679 contains a piggyBac insertion and is considered as a significantly reduced-function mutant." And we added two references.
11. The zebrafish studies use one mutant gene (and allele) and one phenotypic measure, and are therefore more limited with regard to site of action, receptor or assurances that only NE action (in a relevant location) is being disrupted.
We agree. We added sentences to support the use of this mutant, toned down the section regarding NE function in oocyte quiescence in zebrafish, and discussed the limitation in the Discussion as follows, "Although the zebrafish study was based on the observation made from a mutant and we have not identified the receptor(s) and the action site in the ovary, we could observe that the quiescent oocytes were maintained in the control whereas they were not in the mutant, which is the same as in the other animals we observed." Thank you.
12. The broader context of the study as a whole really concerns what was known before, what is now understood better and what remains unresolved. Here I am being guided largely by the authors' narrative. Mostly, I found the narrative to be informative, especially given the challenge of describing relevant background for three experimental organisms. One sticking point for me, which is perhaps most readily phrased for the Drosophila studies, is whether the authors are suggesting or concluding that there is an active response to protein deficiency (as opposed to the absence of a response to protein sufficiency). The authors say it is not known whether the absence of insulin and JH signals is sufficient to maintain quiescent ovaries or another inhibitory signal is necessary. This needs discussion or refinement as my understanding is that several (maybe all?) facets of quiescence are observed in the absence of those signals (and in some cases are reversed by addition). Later (p12) the idea is repeated, suggesting that there is indeed also a nutrientresponsive neuronal signal, followed by a chain of conditional sentences about expected scenarios. Do the authors have any evidence that OA presentation to relevant cells is altered in response to nutrition? Also, have the authors tested whether OA signaling disruption alters insulin peptide secretion or responses or whether one type of signaling overrides the other (under genetic conditions that permit such tests)? It appears to me that the idea of a nutrient-responsive signal is at this stage speculative and that it is not yet clear whether OA responses act independently of insulin-like signals.
We appreciate the insightful summary/points. The same issue was brought up by reviewer #3. To address this concern, we performed a new set of experiments. When we treated flies with OA and nutrients varying the concentration of each, OA and nutrients balance each other, indicating that OA is an active signal and thus supporting our hypothesis; A medium concentration of OA (5 mg/ml) inhibits low level of nutrient signaling (1% yeast). This OA effect is overridden by high level of nutrient signaling (2.5% yeast). A high concentration of OA (10 mg/ml) inhibits high level of nutrient signaling (2.5% yeast) (Fig. 6 e-g).
13. Another significant unknown (in each setting, possibly with the exception of C. elegans) appears to be the exact cells responsible for receiving the OA signal, how they then relay the signal and what might be the significance of those choices (why they are suited to the purpose). The evidence in Fig.  1 does not really answer these questions and I am particularly unconvinced from the images about the statement that the stained processes reach germarial regions or why that may be relevant for the reported phenotypes.
For the concern regarding detailed molecular mechanisms (a significant unknown), please consider our attempt to address the significance issue below (#14). For the expression, we revised the text, "A GAL4 line that drives GFP expression by a tdc2 promoter showed tdc2 is expressed in many varicosities or boutons of neurons covering the ovary (Fig. 1c)." and removed 'germaria'.
14. Another issue that seems debatable is what is the significance of finding somewhat related roles for OA and NE in three organisms. Much about how OA acts and in which cells remains to be discovered and the manifestations of failures of quiescence and unfavorable conditions examined are quite diverse. It is therefore hard to assess what might be the evolutionary mechanisms and potential common advantages of employing OA/NA. We do not attempt to address the detailed molecular mechanisms among all three models. That would require three different papers (we think). Rather, we investigated the molecular and cellular mechanism in C. elegans, and asked whether the same molecule serves a conserved function by testing two other models. We believe the focus and merit of our paper is that novel finding among three such distant animals. Regardless of the detailed molecular mechanisms, the observations we made clearly share a common theme: in the absence of OA or NE, quiescent oocytes are not maintained. This is the first report of such molecule in any animals and also a first report of OA action in oocyte quiescence. In fact, we found the idea that Nature uses the same molecule to protect oocytes from environmental uncertainty is exciting. It is even more so when considering these animals use completely different reproductive strategies.
Altogether, I expect that the authors will be able to resolve most of the issues I have raised with more careful explanations and making available quantitative data more explicit. While I am not personally convinced that investigation of three systems makes the conclusions especially noteworthy, it does seem that there are robust data for more than one system supporting some new insights into an important area of biology that will likely not be very familiar to most readers but nonetheless interesting and informative.
Thank you so much for the critical reading and identifying the points to clarify. We deeply appreciate it.

13
Reviewer #3 (Remarks to the Author): Nutrient and OA balance in flies.
Remarkably, it is shown in this manuscript that Octopamine/Norepinephrin regulates oogenesis quiescence during starvation conditions in four species, from C.elegans, Drosophila, to zebrafish. It is well known that oogenesis undergoes quiescence during starvation to save nutrients for survival. However, how the quiescence is mediated molecularly was not known. In C.elegans and Drosophila, the authors found that oocytes progressed in oogenesis in starvation or protein-deficient conditions when OA (Octopamine) fails to be produced, unlike wild type, which arrest in their development at previtellogenic stages. Importantly, exogenous OA could rescue the defect, demonstrating the specific role of OA in mediating this quiescence. In both C.elegans and Drosophila, the authors also identified the receptor of OA, which exhibited the same phenotype as the OA-deficient animals. In zebrafish, they show that lack of norepinephrine (OA in vertebrates) causes the same defect that oogenesis proceeds even during starvation conditions. The data strongly support the conclusions.
The authors hypothesize that it is the balance of OA to nutrient signals (Insulin related) that mediates either quiescence or oogenesis progression. It would add important depth to this study, if it were possible to test this in one of the organisms. In Drosophila and C.elegans, the nutrient signaling pathway is well studied, so it wouldn't be too difficult to test this model.
We really appreciate this insightful suggestion. After considering how to pursue this idea, we decided to test it directly balancing OA and nutrients instead of using nutrient sensing mutants such as insulin mutants. The main reason is that such mutants exhibit so many diverse phenotypes, including reproduction and metabolism as well as development and growth, in both C. elegans and Drosophila. In addition to the concern that this pleiotropic phenotype creates technical difficulties in matching the ages and conditions, since those mutants are already adapted to a low nutrient state, we were concerned that the results might not be clear to be conclusively interpreted. The approach we took, however, supports our hypothesis that OA and nutrients balance each other; When we treated flies with OA and nutrients varying the concentration of each other, a medium concentration of OA (5 mg/ml) inhibits low level of nutrient signaling (1% yeast). This OA effect is overridden by high level of nutrient signaling (2.5% yeast). A high concentration of OA (10 mg/ml) inhibits high level of nutrient signaling (2.5% yeast). The new data are added in Fig. 6e and g. We thank the reviewer very much. Adding this piece of evidence helps us to clarify our hypothesis and thus improve the paper.
Minor points: We appreciate the comment and indicated the significance on the graph. Thank you so much for pointing out this error! We fixed it.
The authors have responded diligently to various questions raised and the manuscript reads quite smoothly at a superficial level. However, I still have two sets of comments.
The first concern presentation and are easily addressed. "Oocyte quiescence" is used repeatedly and as a unifying expression. However, I think that the term will be understood in various ways and many are not consistent with the evidence. One meaning is failure to progress through cell cycles-for Drosophila egg chambers progressing past stage 8 that is not an appropriate description. Related to that, Fig. 5 implies that stage 8 might involve a transition from prophase 1 to metaphase 1 (because of ambiguous alignment & contraction of stages). Other sources repeatedly state that occurs at stage 13. Also, this maturation is not measured in any of the studies presented, so there is no evidence related to oocyte cell cycle progression. The authors should make clear that they are not claiming any relevance to the prophase to metaphase transition. It may be that "developmental arrest" is a more appropriate term. This is not trivial in the sense that the Drosophila example may not involve events in the oocyte at all.
Order of presentation and specificity is, in my opinion, not optimal in some places: "lacking OA" line 181 and "OA mutants" line 223 should be replaced or supplemented by the actual genotype (they are not literally OA mutants and they have not been shown here to lack OA, though I have no problem with that as an inference from other studies). Mutant phenotypes are often presented before control behaviors-line 211, 304 and other places, including behavior of animals without starvation (first).
There are some odd statements: Line 291 Drosophila has 2nd meiotic arrest. So what? It is after all of the regulated events studied here and responds to sperm (should be clearly stated somewhere). 301 quiescent oocytes not easily staged in well-fed flies-there is not any evidence of arrest of any type in well-fed flies is there, so what is to be staged?
The discussion (and a response to my first review) suggest that the authors believe they have shown that an OA signal actively responds to a lack of nutrition. I strongly disagree. To show that would require demonstrating a change in OA production in response to nutrients and/or genetically inactivating such a response mechanism (without eliminating OA production altogether) and seeing a defect in the overall oogenesis response to nutrient deprivation (eliminating other response pathways if necessary). It is not addressed by synthetically applying different amounts of OA and seeing different responses. This is an important unjustified claim. OA could be produced at constant level; if it is not, then the source signal and its relay would be interesting, but none of that has been established here.
The second set of comments concern the Drosophila evidence.
The aspect of degeneration of stage 8 oocytes has now been eliminated. Consequently, the text is easy to read and sounds convincing. But when I look at the data in Figs. 5/6 I am far from convinced. In essence, this is because the most striking difference is the accumulation of stage 14 egg chambers in mutant conditions. However, I presume that this has two components-arrival at stage 14 and loss of stage 14 egg chambers. I understand that the latter is blocked in tbh mutants because egg-laying is blocked. That blockage could plausibly account for most of the differences in stage 14 egg chamber accumulation, right? Progress beyond stage 8 manifest by the abundance of stage 9-13 egg chambers is barely evident and not significant. In theory, lack of progress for controls would be evident from an accumulation of stage 8 egg chambers. However, that is not seen at all (Fig. 5e, f). That may be because of degeneration of stage 8 egg chambers, so that measuring degenerating stage 8 egg chambers may be important, if it can be done effectively (unfortunately, I expect it is too transient to help much). So, overall, looking at the data in Fig. 5 I am not sure I see convincing evidence of significant differences in progress past stage 8. Fig. 5d is the most convincing and perhaps summing all poststage 8 stages shows significant differences.