Bis-naphthopyrone pigments protect filamentous ascomycetes from a wide range of predators

It is thought that fungi protect themselves from predation by the production of compounds that are toxic to soil-dwelling animals. Here, we show that a nontoxic pigment, the bis-naphthopyrone aurofusarin, protects Fusarium fungi from a wide range of animal predators. We find that springtails (primitive hexapods), woodlice (crustaceans), and mealworms (insects) prefer feeding on fungi with disrupted aurofusarin synthesis, and mealworms and springtails are repelled by wheat flour amended with the fungal bis-naphthopyrones aurofusarin, viomellein, or xanthomegnin. Predation stimulates aurofusarin synthesis in several Fusarium species and viomellein synthesis in Aspergillus ochraceus. Aurofusarin displays low toxicity in mealworms, springtails, isopods, Drosophila, and insect cells, contradicting the common view that fungal defence metabolites are toxic. Our results indicate that bis-naphthopyrones are defence compounds that protect filamentous ascomycetes from predators through a mechanism that does not involve toxicity.

Overall comments: This paper attempts to show that aurofusarin is an important secondary defense compound in fungi to protect against insect predators. This is a nice study that deserves to be published, but why it belongs in Nature Communications -the authors have not convinced me. To me, this paper belongs in a much more specialized journal. The way it is written is not to engage a general audience and one paragraph of introduction was not enough to grab me and convince me of the larger importance of this story and its greater relevance. In general, the way this paper was written was extremely difficult to follow. This paper requires a reader to have familiarity and specific knowledge of 1.Fungal pathogens 2.RNA seq and transcriptomics 4. Many insect genera AND 5. esoteric secondary defense/chemical compounds. Honestly, I don't know anyone who has enough expertise in all of these areas to be able to assess this paper fully. I am comfortable with fungi and RNAseq to a certain extent, but in order to have a reader understand and follow a paper that spans all of these fields MUCH more effort needs to be put in to explain and define terms and help the reader follow the paper. The paper would benefit greatly from a more expanded and broader introduction and a direct reference to the methods section (which I did not find until after I had read the paper). Also, the figures need to be split up and simplified. I suggest that you pick your favorite point to highlight, and highlight it, and don't try to fit so much into a single figure. I also think many of the statements i.e. "aurofusarin content in fungal mycelia is astonishingly high" sounded rather hyberbolic in absence of citations or comparative data so we can know how high they are relative to other defense compounds.
Line comments: Line 56: You need to make it clear that you are talking about your experiment now. It would be easier to follow if you said "we sequenced the transcriptomed in order to study…" In general, One paragraph of introduction and then going straight into your results is very quick. I would like more introduction and background for why you did this experiment, why did you choose Fusarium gramineareum? What is the significance of that fungus? Is it just something you happened to have in culture? How many replicates did you have? Is this just one culture? I need more details in the methods here to assess your experimental design. Figure 1: You need to explain your abbreviations. What is PKS? What is NRPS? There must be a shorter way to list your accession numbers. Also there are many ways to analyze transcriptome data so I would like to see more detail in the supplemental methods or post scripts to github and give us the link.
Line 59-60: You need to introduce these metabolites and why you decided to look for them. I've never heard of aurofusarin before -why is it of interest and how did you know to look for it? Line 63: this is a lot of jargon and complicated defense agent names. This is seeming very specialized, perhaps belongs in a more specialized journal? Lines 86-106: In general, an extremely dense paragraph with a TON of jargon. Words I've never heard before reading this paper aurofusarin, dimeric naptho-gamma-pyrone pigment, those nematode genera ….. you use HPLC abbreviation without ever introducing it first. I can't see that you made a huge effort to make this paper interesting to a more general audience such that you might get with nature communications.
Line 100: Which five Fusarium species did you use? Why did you choose them? How many replicates?
Lines 102-104: Not sure why you used ELSD here in addition to the previous HPLC Line 106-108: To me this sounds like a just so story. I don't think you can say the "extraordinarily high amounts of aurofusarin …explain why aurofusarin synthesis incurs substantial fitness costs". I think it's more cautious to say these high levels may be a fitness cost, as evidenced by increased growth rates of mutants.." Especially if you didn't measure anything else about the fungus so you don't know what other trade-offs there might be. Figure 2a. Until you showed me the corn ear I had no idea that F. graminearum was a corn pathogen or that was how you were growing it. In general, you need more background, more methods, and more explanation of your experimental set up in order for anyone outside your lab group to be able to follow this story and understand what is happening.
In general figure 2 is extremely dense and includes way too much information for a single figure.
Line 130: are these cultures in liquid culture or are they in corn?
Line 132: why are you switching to the woodlouse when before you were using nematodes?
Line 133: you need to explain this more. Does aurofusarin synthesis cause things to turn red? How would I as a reader know that if you don't tell me?
Line 138: Previously in this paragraph F. was used to abbreviate Fusarium. I'm assuming the collembolan is another genus but which one? You can't abbreviate like that if you haven't introduced it before.
Line 137-138: How many of each species? Also, how many replicates of each experiment do you have? Did you conduct statistics?
Line 145: this sounds very anecdotal Line 146: when did you disrupt the biosynthetic pathways for deoxynivalenol and zearalenone and how did you do it?
Line 148: are we back to talking about Fusarium with the F. now?
Lines 150-151: I don't understand how you can say aurofusarin is the only defense metabolite that has deterred springtails -how many metabolites did you test? Without an exhaustive test I don't see how you can make such statements. Seems hyberbolic to me. Figure 3. The pictures would be more helpful if you had labels next to them i.e "mealworm" next to the picture of the mealworm for example, to help out those of us who have no idea what a mealworm looks like.
Line 169-170: I fail to see how that experiment answered the question. Please explain further. Also you can't just say "the larvae's strong preference for flour without aurofusarin demonstrated.." without showing data so I can assess the magnitude and effect size Line 180: Ok so if aurofusarin did not affect meal worm growth, it doesn't seem like it's very toxic to them. How does this help your argument?  Line 234-235: How do you know that no arthropods have adapted to this defense metabolite? You have tried a handful of insect species, it could be that many species are adapted to it and you just didn't happen to try them.
Line 236: If this is for Nature communications, you need to link this sentence to a larger picture. Why should anyone care about fungal bis-naphthopyrones? What is the greater relevance of this research? I'm not saying it's not important, I'm saying that you need to explain to the audience why it's important and what is the bigger picture.
Line 337: I'm only now seeing that there is a methods section. How come it was not referred to in the main document? Why is it after the references and hidden somewhere I would never find it?

Response to the reviewers' comment on the manuscript "Bis-naphthopyrone pigments protect filamentous ascomycetes from a wide range of predators"
We appreciated the comments of the reviewers, which we found very helpful. The manuscript was revised and new results were added. Originally the manuscript was submitted as a Letter but the editor of [redacted] recommended us direct transfer to Nature Communications, which is why the original submission was not formatted for the latter journal. In the following we respond to the comments of the reviewers.

Comment: The manuscript by Xu et al. sheds light onto the chemical defense of filamentous fungi against predators. The authors refer to the unresolved issue that most attempts to correlate the biosynthesis of toxic secondary metabolites with deterral of fungal predators were not successful. Based on their results, they put forward the idea that in terms of antifeedant activity, the biosynthesis of large amounts of a specific class of comparably 'nontoxic' pigment is much more effective and used by many filamentous ascomycetes.
The overall conclusions of this manuscript are fascinating and the presented data is convincing.

Response
Thank you for this positive assessment.
Comment:...I see the only major flaw of the study in the mechanistic aspects of deterrence and pigment induction. The deterrence mechanism is experimentally not addressed other than by the -negative -toxicity assays. In this regard, it would be very nice if the authors could present lack of toxicity of the purified pigment for another organism besides mealworms for which deterrence is demonstrated; if this is not possible, they could extend their test organisms by Drosophila melanogaster larvae both for toxicity and deterrence assays since the pigment could be easily mixed in the food similar to mealworms. In my opinion, however, the (very low) toxicity cannot be the cause for the observed universal deterrence.

Response
We carried out three additional toxicity assays. D. melanogaster larvae were fed on food with aurofusarin for two days and allowed to accomplish their development on medium without aurofusarin (Fig. 5b). Aurofusarin has not reduced the number of adults emerging from pupae, showing the lack of developmental toxicity. Our attempt to establish a deterrence assay with -2 -Drosophila larvae failed because boring of larvae into food interfered with counting. Aurofusarin toxicity was also investigated on the springtail Folsomia candida and the isopod Trichorhina tomentosa, fed on fungal cultures with and without aurofusarin for 5 weeks (Table 1). The lack of mortality and only slight growth suppression observed in these experiments corroborated our previous results with mealworms and insect cells.
Comment: I rather hypothesize that the pigment either binds, as a signal, to some kind of receptor that is conserved in all tested organisms and this signal recognition leads to the observed avoidance behavior (without toxicity). The large amounts of pigment needed for deterrence might be explained by a low affinity of the pigment for the receptor which, on the other hand, allows binding to the slightly different variants of the same receptor in all tested organisms.

Response
Thank you very much for these ideas. We suggested that large amounts of aurofusarin prevented adaptation, writing "high levels of aurofusarin... may saturate degradation or other activities counteracting the antifeedant effect" (orig. man. L222-223), but we missed the idea of saturating low-affinity receptors. In the new discussion we wrote: Revision L399-402: "We hypothesise that high levels of aurofusarin in fungal mycelia prevented predators from adaptation by saturating molecular targets of aurofusarin with binding affinities reduced by mutations, or by overwhelming enzymatic degradation. Mycotoxins never accumulate in comparably high concentrations..." The instructions for authors prevent us from including thanks to referees in the Acknowledgment, but we feel that the saturation of low-affinity receptors was an important idea, and we therefore acknowledged the source.
Comment:....Identification of the putative receptor is beyond the scope of this study but the authors could at least formulate some hypotheses in this regard.

Response
We offered such a hypothesis, based on the wide range of predators that are affected: Revision L411-415: "The intriguing question for future research is how a single metabolite class deters a wide range of predators. The presence of gustatory receptors triggered by bis-naphthopyrones in phylogenetically distant arthropods, including crustaceans, springtails and insects, indicates that natural ligands of these receptors are compounds common in food substrates of all arthropods, such as proteins or polysaccharides." -3 -Comment: ...Alternatively, since the deterring compound is a pigment, it is possible that not the compound itself but rather its color, is perceived by the organisms. I am not really a specialist but I recently read that even nematodes have light-sensing neurons. A possible experiment to distinguish between these two possibilities, could be to replace aurofusarin with other pigments with similar absorption spectra in the food choice experiments.

Response
To exclude the possibility that the animals recognized and avoided the colour of aurofusarin, we carried out a new food choice experiment with aurofusarin-amended wheat flour in complete darkness (Fig. 4b). Because a single mealworm per arena was used, exposure to dim light for a second was sufficient to locate the animal. The preference for aurofusarin-free flour in the dark was similar to the preference for aurofusarin-free fungal mycelia at ambient light (cf. Fig. 3g).
Comment: ... In addition to the deterrence mechanism, the authors should also elaborate a bit more on the mechanism of pigment induction.
According to the results in Fig. 2b, aurofusarin production is induced upon physical damage of the mycelium e.g. by chewing and puncturing predators. The authors should mention that in the text and eventually confirm this e.g. as an additional panel of Ext. Data Fig. 3 by local damage of the mycelium on plate with a scalpel. This figure also lacks a negative control of punching in a plastic cylinder without adding any predator.

Response
We conducted the suggested experiment, damaging fungal mycelium with an array of razor blades. The results showed that mechanical damage was sufficient to induce aurofusarin synthesis, the synthesis was restricted to the damaged area, and the stimulation of aurofusarin synthesis lasted for at least 120 h (new Fig. 6). These results pointed out at interesting differences between the chemical defence against predators in fungi and plants. We elaborate on these differences in the discussion (revised manuscript L380-389); therefore Fig. 6 was placed into the main manuscript rather than becoming an additional panel in previous Ext. Data Fig. 3 (now Supplementary Fig. 4).
Negative controls for the experiments with plastic cylinders are shown for two fungal species in Supplementary Fig. 4 in panels a and b on the right.
Because shaken cultures of F. venenatum accumulated less aurofusarin than still cultures (Orig . Fig 4b; Revised Supplementary Fig. 3), apparently contradicting the hypothesis that mechanical damage triggers aurofusarin synthesis, we injured mycelium of F. venenatum by cutting ( Supplementary Fig. 6) and observed pigmentation of the mycelia. We described these results in the revised manuscript on L272-277 .
We believe that the manuscript merits publication in Nature Communications because of the conceptual advancements that it provides. Chemical defence in fungi has been studied for decades and several lines of evidence corroborated the defence role of secondary metabolites, but attempts to confirm that fungal defence agents are mycotoxins remained inconclusive. Our works show that rather than mycotoxins, the sought-for defence metabolites are bisnaphthopyrones of low toxicity that are ubiquitous in ascomycetes. The striking features of the new defence mechanism are an unprecedented diversity of predators that are deterred by these metabolites and the lack of adaptation in predators, in spite of persistent exposure. The hypothesis explaining these features by high levels of bis-naphthopyrones in fungal hyphae saturating low-affinity receptors and detoxification activities is a new concept in the biology of fungal defence and in chemical ecology in general. In the original submission we reported our findings on aurofusarin. In the revised manuscript we added new data on two additional bis-naphthopyrones that are produced by many fungal species, corroborating the general validity of the concept. In our opinion, the discovery of a fundamental ecological function of a widespread group of fungal metabolites is important for mycologists, soil ecologists as well as biological chemists, and it is well suited for a wide audience of Nature Communications.
Comment:...This paper requires a reader to have familiarity and specific knowledge of 1.Fungal pathogens 2.RNA seq and transcriptomics 4. Many insect genera AND 5. esoteric secondary defense/chemical compounds. Honestly, I don't know anyone who has enough expertise in all of these areas to be able to assess this paper fully. I am comfortable with fungi and RNAseq to a certain extent, but in order to have a reader understand and follow a paper that spans all of these fields MUCH more effort needs to be put in to explain and define terms and help the reader follow the paper.

Response:
We are aware of the challenge posed by a multidisciplinary treatise written by a group of authors with expertise in different fields. If the manuscript is published, the readers will face the same challenge, we therefore appreciate every advice how to present our results intelligibly. In the revised manuscript, we defined all technical terms except those that are generally known. We also provided additional explanations to improve the readability, for instance regarding the use of ELSD.
Comment:...The paper would benefit greatly from a more expanded and broader introduction and a direct reference to the methods section (which I did not find until after I had read the paper). Also, the figures need to be split up and simplified. I suggest that you pick your favorite point to highlight, and highlight it, and don't try to fit so much into a single figure. I also think many of the statements i.e. "aurofusarin content in fungal mycelia is astonishingly high" sounded rather hyberbolic in absence of citations or comparative data so we can know how high they are relative to other defense compounds.

Response:
The introduction was expanded and broadened.
Articles in Nature Communications do not provide direct references the methods section, probably to prevent interruptions of the narrative flow. We used the same style but we referred to the methods section explicitly when it was necessary for understanding why a particular method was used (L131). We agree that the statements about the level of aurofusarin in mycelia sounded hyperbolic and would benefit from citations or comparative data. In the revised manuscript, we used more moderate language. We reviewed the literature but have not found any publication reporting comparably high levels of secondary metabolites in fungal mycelia. The manuscript text was revised as follows: Original text, Abstract: "...causing an accumulation of aurofusarin in fungal mycelia at extraordinary high levels..." Revised manuscript: the passage was removed Original text, L218-219: "Aurofusarin content in fungal mycelia is astonishingly high 4d), unparalleled by any non-polymeric fungal polyketide... Revised text, L124-127: "...aurofusarin in grazed mycelia amounted to up to 2.5% of the dry weight (Fig. 2d, e). We were not aware of any non-polymeric secondary metabolite that accumulates in fungal mycelia at such a level..." Original text, L231-232: "...antifeedants accumulating at unprecedentedly high levels appear to protect fungal mycelia from soil-dwelling predators." Revised text, L393-394: "Aurofusarin content in Fusarium graminearum that has been exposed to predation is very high (Figs. 2d-e)."

Line comments:
Comment: Line 56: You need to make it clear that you are talking about your experiment now. It would be easier to follow if you said "we sequenced the transcriptome in order to study…" In general, One paragraph of introduction and then going straight into your results is very quick.

Response:
In the revised manuscript, the passage appears under the heading Results, which shows that we are reporting about our experiment. Thank you for the suggestion of improved wording for the description of RNAseq, we revised the text as follows: Revised manuscript L80-81: "...we sequenced the transcriptome of the fungus Fusarium graminearum that had been exposed to the springtail Folsomia candida to reveal which biosynthetic pathways were induced by grazing." Only a single short paragraph of introduction was allowed in the format of the original manuscript. In the revised manuscript we extended the introduction substantially; see also our response to the first comment on p. 6.
Comment:...I would like more introduction and background for why you did this experiment, why did you choose Fusarium graminearum? What is the significance of that fungus? Is it just something you happened to have in culture? How many replicates did you have? Is this just one culture? I need more details in the methods here to assess your experimental design.

Response:
We have not explained why we have chosen Fusarium graminearum because we believe that many filamentous ascomycetes would be suitable. We worked with F. graminearum because we had it in the lab, the genome is annotated, and many secondary metabolites have been characterized and assigned to gene clusters, but many fungal species fulfil these conditions. Therefore we have not elaborated on the choice of the species.
-9 - The number of replicates in the RNAseq experiment was 4, as shown in captions of Figs. 1 and 2. We will come back to this issue in our response to the last comment below. We used different strains and cultures on solid media and in liquid media, as specified in figure captions and in the description of methods. Figure 1: You need to explain your abbreviations. What is PKS? What is NRPS? There must be a shorter way to list your accession numbers. Also there are many ways to analyze transcriptome data so I would like to see more detail in the supplemental methods or post scripts to github and give us the link.

Response:
We avoided the abbreviations PKS and NRPS in the main text of the revised manuscript and provided explanations when they were needed in the methods.
We shortened the lists of accession numbers by removing prefixes from all except the first accession number for each cluster.
The analysis of the transcriptome is described in the subsection "Transcriptome analysis by RNAseq" of the methods section in the revised manuscript. All scripts that we used are listed here, including version No. and references to original publications, which also provide URLs for the source code. Remote URLs can be found by searching GitHub for the names of the scripts. We have not used custom scripts.
Comment: Line 59-60: You need to introduce these metabolites and why you decided to look for them. I've never heard of aurofusarin before -why is it of interest and how did you know to look for it?

Response:
We have not decided to look for these metabolites. On L59-61 of the original manuscript (L84-85 of the revision) we listed the secondary metabolites that were induced by grazing, as revealed by the RNAseq experiment. We could not introduce these metabolites before because we did not know which metabolites would be induced.
Aurofusarin was one of the metabolites induced by grazing. In the original manuscript, the reason why aurofusarin was selected for further work was explained in the subsequent section. In the revised manuscript we provided this explanation immediately. We also removed the names of the other induced metabolites from this passage: Revised manuscript L63-65: "The biosynthesis pathways for several secondary metabolites were induced via grazing. Among them, bis-naphthopyrone aurofusarin was selected for further investigation as similar metabolites are produced by many fungal species." Comment: Line 63: this is a lot of jargon and complicated defense agent names. This is seeming very specialized, perhaps belongs in a more specialized journal?

Response:
The condensed form of the original manuscript required terse wording. In the revised manuscript, we specify that these products are mycotoxins and explain why they could conceivably act as defence agents. We don't think that reasoning about the role of mycotoxins in fungal defence requires a specialized journal: Revised manuscript L88-91: "Pathways for the mycotoxins deoxynivalenol and zearalenone -which are toxic to insects 13, 14 -and for necrosis and ethylene-inducing peptide-like proteins -which we hypothesised to be defence agents due to their similarity to lectins 24 -were not induced by grazing (Supplementary Data File 1)." Comment: Lines 86-106: In general, an extremely dense paragraph with a TON of jargon.
Words I've never heard before reading this paper aurofusarin, dimeric naptho-gammapyrone pigment, those nematode genera ….. you use HPLC abbreviation without ever introducing it first. I can't see that you made a huge effort to make this paper interesting to a more general audience such that you might get with nature communications.

Response:
We made all efforts to make our work accessible to a general audience. An explanation of the abbreviation HPLC was omitted from the original manuscript because we regarded it as a generally known acronym, and because many recent Nature Communications papers used the abbreviation without explanation. We added the explanation to the revised version of the manuscript (L123).
We fully agree that thus far, aurofusarin was an obscure metabolite known only to specialists.
This will certainly change after our work reveals that aurofusarin plays a key ecological role as a defence agent of many fungi against a wide range of predators.
We acknowledge that chemical names may intimidate readers without a background in chemistry. These names, however, provide important information to other readers, and they help understanding structural similarities among the metabolites (Fig. 7).
We agree that the names of the nematodes and other predators will sound obscure to most readers. We have to use accurate species names in the description of our experiments, but no background knowledge on these animals is required. The only information the reader needs is that they can feed on fungi. Our narrative makes clear that the choice of species of nematodes, crustaceans, springtails and insects in our work was not important. The key message is that that aurofusarin and other bis-naphthopyrones (new Fig. 8 Response: We do not say that aurofusarin is the only defence metabolite that has deterred springtail. We say that reversal of food preference by the disruption of aurofusarin synthesis indicates that aurofusarin was the major or only defence metabolite of F. graminearum. The inference is based on the loss of preference for F. verticillioides over F. graminearum in experiments with F. graminearum that was unable to produce aurofusarin. It is conceivable that other metabolites of F. graminearum would deter springtails under different conditions, we therefore added "in this experiment": Original manuscript, L148-151: "The reversal of springtails' food preference for F. verticillioides over F. graminearum via the disruption of aurofusarin synthesis in F. graminearum (cf. Ext. Data Fig. 6a,c with 6b) indicates that aurofusarin had been the major -or only -defence metabolite of F. graminearum that had deterred springtails." Revised manuscript L156-159: "The reversal of the springtails' food preference for F. verticillioides over F. graminearum via the disruption of aurofusarin synthesis in F. graminearum (Supplementary Fig. 5) indicates that aurofusarin had served as the major -or only -defence metabolite of F. graminearum deterring the springtails in this experiment." Comment: Figure 3. The pictures would be more helpful if you had labels next to them i.e "mealworm" next to the picture of the mealworm for example, to help out those of us who have no idea what a mealworm looks like.
Response: The common as well as scientific names of the animals are specified in the caption.
The pictures were added to facilitate fast comprehension. We understand that this will be of no use to readers who do not recognize the animals, but adding names next to the pictures would clutter the figure with excessive redundancy. Readers who do not recognize the animals in the pictures will check the caption.
Comment: Line 169-170: I fail to see how that experiment answered the question. Please ex-