An evolutionarily conserved pathway mediated by neuroparsin-A regulates reproductive plasticity in ants

Phenotypic plasticity displayed by an animal in response to different environmental conditions is supposedly crucial for its survival and reproduction. The female adults of some ant lineages display phenotypic plasticity related to reproductive role. In pharaoh ant queens, insemination induces substantial physiological/behavioral changes and implicates remarkable gene regulatory network (GRN) shift in the brain. Here, we report a neuropeptide neuroparsin A (NPA) showing a conserved expression pattern associated with reproductive activity across ant species. Knock-down of NPA in unmated queen enhances ovary activity, whereas injection of NPA peptide in fertilized queen suppresses ovary activity. We found that NPA mainly affected the downstream gene JHBP in the ovary, which is positively regulated by NPA and suppression of which induces elevated ovary activity, and shadow which is negatively regulated by NPA. Furthermore, we show that NPA was also employed into the brain–ovary axis in regulating the worker reproductive changes in other distantly related species, such as Harpegnathos venator ants.

The authors report a series of experiments that reveal the anti-gonadotropic function of NPA.These are well-executed manipulations that provide experimental evidence that NPA inhibits reproduction.It is also interesting to note that NPA may be colocalized with ILP2, which was recently shown to activate reproduction.However, the second pillar of the manuscript, namely that NPA regulates behavior as well, is not well supported by the data.The behavioral experiments have several flaws that prevent from making firm conclusions on the behavioral role of NPA.First, all behavioral results may be associated with reproductive changes (e.g., locomotory activity being reduced in reproductive individuals) and would simply be the indirect consequence of the effect of NPA on reproduction.Response: Thanks for the reviewer's comment.In our manuscript, we have not excluded the possibility that NPA may indirectly affect behaviors, actually, we propose that the brain-ovary axis underlying reproductive regulation may implicate both forward and backward signalings.We speculate that the signalings from the ovary (or fat body) may feedback to regulate brain functions, so that behavioral outputs of the ants would be affected.There is growing evidence in insects as well as mammals that female reproductive states could affect their behavioral outputs through the actions of circulating hormones in modulating the neural circuits of the brains [1][2][3][4][5].The reproductive ground-plan hypothesis proposes that the co-option of the regulatory components for female reproduction may underlie social evolution [6,7], and the central prediction of the hypothesis is that the reproductive physiology and social behavior, in particular division of labor, are linked by pleiotropic genetic network [8,9].In agreement with this hypothesis, our study demonstrated that this conserved pathway mediated by NPA has been used to regulate the division of labor on both reproduction and social tasks between castes in ants.Consistently, previous study in Harpegnathos saltator ants also demonstrated that another brain peptide corazonin negatively regulates vitellogenin expression in the fat body of the workers, and the altered vitellogenin signaling could affect reproduction as well as hunting behaviors of worker ants [10].While we agree with the reviewer that further study on whether the behavioral effects of these neuropeptides are direct or indirect is an interesting question to follow up, such question could only be addressed with advanced genetic tools for tissue specific manipulations (e.g.restricting the functions of NPA within the brain, or alternatively, specifically disrupting the signalings from the ovary or fat body), which is currently nondeveloped yet in ant systems.We re-wrote the introduction to highlight the potential role of brain-ovary axis in regulating developmental plasticity in ants.
Second, because of the way they are scored, all behavioral measurements (activity, foraging and brood care) may simply reflect variation in locomotion (more active individuals are more likely to interact with food or brood).Proper behavioral investigations would control for such variation.Response: We scored foraging behavior as more than 3 seconds contacting with the food and brood care behavior as more than 3 seconds touching the brood or pupae.To address this question, we applied the same measurements to score the contact frequencies of the ants to the similar-sized blank areas near the food patch as controls.If the high foraging is simply due to the locomotion, we would expect the high locomotion group would also have high frequent contact in the control region.We found that the ants display comparable contact frequencies to the blank areas, indicating that the higher foraging of ds-eGFP ants are not directly caused by their higher locomotion activities.The new results are appended to the S4 Figure .Another evidence is that, the NPA peptide-injected ants display significantly higher locomotion activities (Fig. 3F), however, they rarely display any foraging activities (S3 Fig. ).If the foraging activities directly reflect locomotion differences, then we should obtain similar measurements for this scenario, however, our actual observation indicates the speculation is not true.Third, all behavioral experiments are performed one after the other using the same individuals in the same box, thus variation in one behavior may directly trigger variation in the next one that was tested (e.g., higher foraging activity could result in better nutritional status or larger food reserves to be shared, resulting in apparent increase in brood care).Again, proper behavioral investigations would separate the different experiments to disentangle the effect of the different behaviors.Response: We performed the behavior assays one after the other with the same cohorts of ants primarily due to the technical challenges and limited gyne samples for experiment.Collecting enough simultaneously eclosed gynes requires synchronized set up of hundreds of ant colonies, which would take almost one year, since the gynes are only occasionally reared by the workers in the absence of the queens.Besides, multiple injections must be performed to induce longterm NPA knockdown, which will cause notable losses of the gynes during the manipulations.Considering these difficulties to address this question, we removed the brood contact behavior data in the current revision.This change does not affect our conclusion on other behaviors.
Finally, multiple ants were put together in the same box for behavioral essays, resulting in dependency of the data, which is very problematic in social animals (the behavior of one individual would be greatly influenced by the behavior of the rest of the group).In my opinion, the behavioral experiments are very weak in comparison to the molecular work and are not on par with the usual standards of Plos Biology.Response: Social insects live their entire life centered around colony activities and they must constantly interact with other nestmates.The reason we performed the behavioral experiments with groups of ants is because isolation is an extremely stressful environment for social insects.As observed in our study as well as reported by other research groups, isolation could cause pronounced behavioral abnormalities, as well as gene expression fluctuations in social insects, and their health states are impaired [11][12][13], therefore, we have to perform the behavior experiment under the natural condition as a social group to avoid the impact of the unnatural manipulation by isolation.And since we measured the behavior scores for all individuals in the knockdown groups and the control groups, the significant difference results between the two groups should be sufficient to reflect the behavior effects of the gene knockdown.
Apart from this main concern, I have other concerns that I detail below.L2-3: I think that the title is not very clear, and could be improved.Furthermore, there is a typo in neuroparsin… Response: Thank you for the suggestion and careful reading.We have corrected the spell of neuroparsin.
Generally, I think that the writing could be improved substantially.The manuscript was very difficult to read and understand in some parts.Response: Sorry for the confusion, we carefully polished our writing.L40-59: I find the use of terms such as "queen-like" or "worker-like" in the abstract very vague, and not appropriate for a generalist journal like Plos Biology.Response: We have rephrased the sentence to "Knock-down of NPA in gyne enhances ovary activity and inhibits activity for locomotion and foraging behavior.Conversely, injection of NPA peptide in fertilized queen suppresses ovary activity and causes higher locomotion.".(L46-48) L40-59: I do not think that the claim "we found the function of NPA in regulating reproductive role differentiation is conserved across the ant phylogenetic tree" is supported by the data.The number of species studied is too low and the species list is not representative of the ant phylogeny (e.g., two Camponotus species are used to represent the entire subfamily Formicinae).Response: We do acknowledge that the number of species we studied was limited compared to the huge diversity of ants.Since the gyne individuals of ants only appear in very short periods in the wild, it is rare that we can collect the gyne samples from many species for experiments.However, to satisfy this reviewer, we have added other two species Oecophylla smargdina and Lasius alienus from subfamily Formicinae for rt-qPCR analysis.The new results are included in the Figure 1.We found that the NPA expression levels were also significantly higher in gyne than queen in Oecophylla smargdina and Lasius alienusrom.All together, the currently available species have covered 5 subfamilies and 8 genus.L100-103: From the introduction, it sounds like NPA has already been identified as a candidate for reproductive division of labor in previous studies, which appears to be redundant with some of the results presented in this manuscript.I think that the only added value of the manuscript is the experimental evidence that NPA regulates reproduction.While it is very interesting for ant researchers, I am not convinced that this finding will elicit a broad interest.Response: We have edited the introduction to specify what is known about NPA and what questions have not been addressed.Previous studies only demonstrated the potential correlative relationship between NPA expression levels and reproductive levels.Whether and how NPA is causally responsible for reproductive role differentiation, and whether the functions are conserved across different ant species with distinct reproductive differentiation scenarios remain to be clarified.In this study, we expanded the number of species to cover five different subfamilies and eight genus, and provided experimental evidences to reveal the causal role of NPA in regulating reproductive physiologies as well as behavioral performances.Our study suggested that NPA functioned as a crucial component through the brain-ovary axis for the reproductive role differentiation, and demonstrated this function was conserved across ant species that have distinct reproductive role differentiation scenarios.As NPA is a conserved IGFBP (IGF binding protein) across diversified arthropods, the findings of our study demonstrating the function of NPA as a neuro-hormone integrating juvenile hormone and ecdysone pathways would be of interest to a wide range of researchers in insect physiology as well as to a general readership.
L129-138: Similar to my previous comment, it is not clear what is novel vs published data in this part of the results.Response: These sentences in the introduction show three models of reproductive role differentiation in adults ants proposed in Nagel et.al [14].In these three models, a correlative relationship between NPA expression and reproductive activation was reported in three species.In this study, we surveyed an expanded list of ant species covering 5 subfamilies and 8 genus, and demonstrated that the correlative relationship between the NPA expression and reproductive activation is conserved across all these ant lineages.More importantly, we provided experimental evidence to show the functional roles of NPA in both physiological as well as behavioral changes associated with reproductive role differentiation.
Figures: Maybe it is due to the pdf conversion, but the figures are of very bad quality, to the point where it is difficult to decipher the labels and assess the results.Response: We are sorry for the low-quality of the figures that made the reading difficult.The issue was caused by the improper conversion from a word to a pdf document.We fixed the issue with improved figure quality in the new version of the manuscript.
Figure 1: The evidence for differential expression of NPA is actually weak in many of the species.In particular, in Monomorium pharaonis -the main study system -it looks like the effect is entirely driven by a single data point.This should be clearly acknowledged.I am not convinced that the authors can make a general, robust claim of a central regulator of role differentiation when there is very weak evidence that it actually differs between gynes and queens.Response: We have repeated this experiment with Monomorium pharaonis, and the result was consistent with the previous one.We have updated the result in Figure 1.The weak differential expression is mainly due to the small sample size, which was mainly caused by the difficulty in sample collection, since most of the ant species were collected from the fields and the gyne/queen samples were very limited in the colonies.Response: The model 3 represents some of the secondarily derived ant lineages such as clonal raider ant Ooceraea biroi where their workers regained reproductive ability without insemination.Ooceraea biroi reproduce by thelytokous parthenogenesis, alternating between an egg-laying reproductive phase to a brood-caring non-reproductive phase.In this species, the components of GRN for reproductive role differentiation that is normally not expressed in the non-reproductive phenotypes might have been co-opted for their respective worker reproductive role transitions.Although in this model, the transition is triggered by the presence/absence of the larval signals rather than the insemination.Although we did not perform a functional experiment on this species，our own expression data also supported the NPA expression is also correlated with the reproductive status in this species, which might also through this conserved brain-ovary axis GRN.L141: "p value" is mentioned twice.Why? Response: Sorry for the confusion.Actually, we described the results of four different species here.To clear the vagueness, we deleted the descriptions of p-values in the main text and appended the exact corrected p-values in the supplementary notes.
More generally, throughout the manuscript, the authors should report the effect size and the exact corrected p-value, not just "P<0.05".This does not follow standard scientific practice and is misleading, especially as the evidence here is very weak.Response: We have appended the exact corrected p-values, the Cohen's d-values for effect sizes, and 95% CI to the S5 Data.L153: I am not convinced by the general use of "reproductive activated role" in this manuscript.This is not a well-accepted term in the literature and it does not have an established meaning.The authors could simply state that it correlates with reproduction (especially as they provide very weak support for a behavioral role, see before).Or that it differs between reproductive and non-reproductive individuals.I understand that they want to avoid the confusion with worker vs queen comparison but I think they make too much out of a simple comparison.Response: Thanks for the suggestion, we have changed "reproductive activated role" simply to the "reproductive activation in ants".(L149-150) Maybe I missed it, but I did not find proper evidence for all the experiments that the experimental manipulations did result in a downregulation of NPA expression or increase in the NPA peptide quantity.Response: The NPA expression levels in knockdown manipulation were added in the S5 Figure .The NPA peptide level has not been quantified, since the endogenous NPA peptide and injected NPA peptide were too low to be detected by western blot.We have added another control with heated inactive NPA for the reproduction experiment in Monomorium pharaonis, which showed no effect in suppressing reproduction.The ovarian activities of the queen injected with heat-inactivated NPA were similar to those injected with BSA, indicating that the reproduction suppression was caused by active NPA effects.
Finally, coming back to my main concern, I cannot help but notice that the analyses of the transcriptomic response to the experimental manipulation and the discussion in general (including Figure 6) focus almost exclusively on the reproductive role of NPA.To me this study is about the impact of that gene on reproduction, not on a complex behavioral repertoire associated with reproductive role differentiation.Response: Thanks for the reviewer's comments.These constructive comments greatly help to improve our rationale of the research design, as well as the presentation of our findings.We have extended our responses to these major concerns in previous paragraphs.Please refer to our point-to-point responses to these major concerns in the beginning paragraphs.
Reviewer #2: This manuscript builds on previous work (Nagel et al. 2020) that implicated a gene encoding neuroparsin A (NPA), among others, in association with early regulation of reproductive activation in adult female Monomorium ants.The prior results were largely correlational, but the NPA correlation was established on a fine spatiotemporal scale in the brain.The current study expands upon the prior results by showing that the association between reproductive activation and NPA decrease is persistent across 5 subfamilies of ants encompassing three distinct modes of reproductive activation (queen/post-mating, gamergate/dueling, clonal/cyclical), and interrogates the function of NPA with a series of experiments involving NPA knockdown and NPA peptide injection in i) an ant from the myrmecine subfamily with a "typical" ant reproductive life cycle and ii) an ant from the ponerine subfamily with dominancebased reproductive gamergates.The results reveal that NPA suppresses reproduction by being involved in signaling from the brain to the ovaries and regulates behavior shifts in the myrmecine ant M. pharaonis and the ponerine ant H. venator, consistent with a model in which NPA has a conserved function in regulating reproductive role differentiation across ant species.The work is impressive in scope and design.The authors have leveraged phylogenetic variation in ant life histories to reveal an apparent conserved molecular underpinning of ovarian activation and affiliated behaviors that may be highly important to the temporal regulation of reproduction and behavior, and thus division of labor, across ants.I would definitely suggest this paper as a reading for any lab group working on ant biology, social insects, or temporal regulation of reproduction and behavior in animals (especially in insects).The experiments that involve multiple rounds of treatment with RNAi and peptides are particularly impressive and are not something most well-funded research groups could pull off.The work required extensive expertise in animal husbandry, behavioral experimentation, ant injection/dissection, molecular genetics, and bioinformatics.The results were clearly presented in figures and clearly described in text.I commend the researchers and very much enjoyed carefully reading this important research.I think it is suitable for PLOS Biology.Response: Thanks for the encouraging comments.We are very appreciative of these positive statements.
I did not review the supplementary files; they were in a .rararchive and I would have had to download a new app on my laptop to decompress.I did not evaluate the manuscript for grammar.

Reviewer #3:
This study provides an interesting and overall straightforward exploration of the effects of NPA on ant behavioral plasticity.Strengths of the study include initially considering 7 ant species across the ant phylogeny and the relatively simple and straightforward experimental design.Unfortunately, the figure quality in the pdf I had access to for review was relatively very low so that I couldn't make out axes labels and figure details for all of the figures.That said, the results appear to support the conclusions.I did not find any major problems with the study design but had some detailed questions (e.g., about mating status of gynes).My main comments / suggestions for improvement are regarding several statements that are incorrect or overstatements throughout the manuscript.Details below: Response: We are sorry for the low-quality of the figures, which was caused by improper conversion of a word to a pdf document.We have fixed the issue and the quality of the figures is improved in the new version of the manuscript.
Figure 1 indicates that all red species have workers with "ovary lost".This is not correct.Response: Thanks for your careful reading.We realized "ovary lost" could not generally be applied to all the species labeled in red, we have deleted "ovary lost" in the new Figure 1.
Several statements in the Introduction that provide background / motivation are unsupported or incorrect: L68-70 and the Abstract states that reproductive caste differentiation and "reproductive role transition" are different.The latter is described as "phenotypic plasticity" L70 and Abstract, while the former is described as L43-44 ". realized by canalized development".Both are clear examples of phenotypic plasticity.Response: Thank you for this comment.Reproductive caste differentiation is a bifurcation developmental process under canalization that results in permanent and irreversible morphologically differentiated castes.While reproductive role differentiation is induced by insemination or other social signals in adulthood which entail physiological and behavioral plastic changes.In this manuscript, we have emphasized the conceptual differences regarding reproductive caste differentiation and reproductive role differentiation, since these two concepts are easily confused to a general readership.Reproductive caste differentiation primarily refers to the differentiation between gyne and worker, which is an irreversible consequence of development.Reproductive role differentiation happens in adulthood, after the morphological caste has already been determined.We have discussed this difference in our previous study [14].In our revised manuscript, we have rephrased the relevant descriptions and made it clear that the reproductive role differentiation is an example of phenotypic plasticity which entails physiological and behavioral plastic changes in adulthood.

Figure 1 :
Figure 1: I do not understand why reproductive model 3 is in this manuscript.It seems out of context because it is outside the scope of the main point of the introduction: that insemination