All reviews of published articles are made public. This includes manuscript files, peer review comments, author rebuttals and revised materials. Note: This was optional for articles submitted before 13 February 2023.
Peer reviewers are encouraged (but not required) to provide their names to the authors when submitting their peer review. If they agree to provide their name, then their personal profile page will reflect a public acknowledgment that they performed a review (even if the article is rejected). If the article is accepted, then reviewers who provided their name will be associated with the article itself.
Thank you for your diligent efforts to further improve your manuscript.
# PeerJ Staff Note - this decision was reviewed and approved by Dezene Huber, a PeerJ Section Editor covering this Section #
The reviewers appreciated your efforts to improve the manuscript, but have identified some additional edits. Namely, the quality of the writing (grammar) in the edited sections needs improvement, and the incorporation of "generation" into the two-way ANOVA as if it were simply an independent variable. Please address these issues and others identified by the reviewers.
This is a second review of this paper. The authors have incorporated recommendations and made edits I suggested in the first review. The manuscript is improved but the quality of the writing in the edited sections is somewhat inferior to that of the original manuscript. This is not a major concern because this can be easily addressed with basic editing.
A major stumbling block with this paper, as pointed out with all three reviewers, is the lack of a positive control. The authors have stated that they tried potassium arsenate as a positive control, but it was a poor choice for a multi-generational study. This now is explained in the manuscript. Also, the authors used ELISA to document high concentrations of Cry1F in the BT diet, which shows the toxin was present.
I stand by my first assessment “This paper makes an important contribution to understanding how focusing on gene expression with RNA-Seq can be a powerful tool for evaluating non-target effects of Bacillus thuringiensis (Bt) derived products that are commonly used in insect-resistant genetically engineered (IRGE) crops.”
Perhaps the authors should consider discussing the challenges with finding positive controls in a multi-generational study. Presumably a sensitive insect would not survive a multi-generational study. Should one consider new sensitive insects for each generation to show exposure or should one try to find a species that can survive multiple generations with sublethal effects, or other approaches?
The experimental design is adequate (flow chart in Figure 1 was useful) and the analyses and interpretation of results are appropriate. The conclusions are supported by the data.
Many minor edits are needed, for example:
75: “detective some potential” should be “detect potential”
83: don’t use “thing”
150: “into baker’s yeast granules with” consider “into baker’s yeast granules”
See previous review and response below.
See previous review and response below
See previous review and response below
I thank the authors for their thoughtful responses to reviewer concerns and suggestions. I remain skeptical that multigenerational research into effects of a plant-expressed Bt protein on an insensitive non-target organism. I also remain skeptical that transcriptomic analysis is warranted where no physiological or ecological effects are observed. Nevertheless, the quality of the research and its contribution to the vast literature on the ecological safety of Bt crops make the manuscript worthy of publication. The authors have added sufficient additional information to address the major concerns.
My remaining concern is in the incorporation of "generation" into the two-way ANOVA as if it were simply an independent variable. The authors would be better justified in conducting three one-way ANOVAs, one for each generation.
The reviewers have raised serious concerns about the risk hypothesis. If you are not able to state a clear risk hypothesis for this research (both the multi-generation effects of exposure to Cry1F and the use of transcriptomics when there are no ecological effects), the paper cannot be accepted.
Secondly, if you cannot provide evidence that the Cry1F diet preparation was active against sensitive species, the paper will be rejected as not being valid. Essentially, there is no positive control supporting the study conclusions of no effect from prolonged exposure to Cry1Ab because there is no evidence (within the study) that the test system was capable of detecting adverse outcomes.
Please address these and the other reviewer comments in your revised manuscript. One of the reviewers has also included annotated comments on an attached file.
[# PeerJ Staff Note: It is PeerJ policy that additional references suggested during the peer-review process should only be included if the authors are in agreement that they are relevant and useful #]
The writing is excellent. This is a very clearly written and well-organized presentation and a pleasure to read. The structure of the paper is excellent and conforms to journal standards, figures and tables are useful and relevant, and raw data are provided. The introduction and background emphasize literature to support the long-term study for Cry1Ab toxicity to Collembola. The argument for conducting this research is overstated and does not holistically represent understanding of toxicity of Cry1Ab protein to non-target organisms. An example is the statement “The possibility that Bt plants might have long-term effects on the soil biota remains a major concern (Icoz & Stotzky, 2008).” The statement is more strongly voiced than in the source publication, the publication is dated, and there is substantial literature that challenges the statement. There is little rationale for chronic testing of Cry1Ab based on (1) well-established specificity of the protein which would preclude toxicity to Collembola, (2) extensive prior tests of Collembola which do not support concerns for toxicity (the authors themselves cite long-term studies which they themselves do not consider to be particularly meaningful), and (3) decades of field deployment showing no adverse ecological effects and very low levels of exposure.
This is original primary research within the scope of the journal. The reviewer is exceptionally impressed with the quality, experience and capabilities of the research lab undertaking this study and of the state-of-the art methodology employed for the research. The research question is well defined but as previously mentioned it does not seem clearly relevant and meaningful as a research consideration. This is a concern for this reviewer because justifying this type of advanced studies for Cry1Ab justifies continued skepticism regarding a technology that has been extensively investigated and has a well-established record of safe use. This is a rigorous investigation performed to a high technical and ethical standard with methods that are well described and fully detailed. It is somewhat surprising that the study does not include a positive control to show that the experimental protocol was indeed capable of ascertaining adverse effects, since positive controls are the norm in ecotoxicity testing.
Overall the outcomes of this research should be viewed as valid, but since there is no positive control the study conclusions of no effect from prolonged exposure to Cry1Ab could be challenged because there is no evidence (within the study) that the test system was indeed capable of detecting adverse outcomes. The authors to provide data to show exposure of the test organism, but it is not clear as to how the collembolans were prepped for ELISA (that is, was the exposure external or has the Cry1Ab been ingested?). These not a major limitations since the authors cite use of the testing approaches used elsewhere in the literature where adverse outcomes were measured. The conclusions are well stated, linked to the original research question and are limited to the supporting results.
This reviewer wishes to emphasize that the quality of this research is very high and I am impressed with the research capabilities and approaches shown in this work. My concerns are restricted to the rationale for such extensive studies directed toward a protein with such extensive prior analysis.
This paper reports on a three generational study to evaluate possible non-target effects of Cry1F protein on the collembolan, Folsomia candida. This is the first attempt to use RNA-Seq to evaluate the impact of Bt toxin on gene expression of this insect. Although Cry1F protein was continuously digested and was detected in F. candida throughout the study, there were no significant differences in the global gene expression between F. candida fed diets with and without Cry1F. Over the three generations, there were no consistently different co-expressed genes. In addition, Cry1F did not obviously alter the expression profiles of seven sensitive biological markers. The authors conclude that the Cry1F protein had no long-term harmful effects on F. candida.
This paper makes an important contribution to understanding how focusing on gene expression with RNA-Seq can be a powerful tool for evaluating non-target effects of Bacillus thuringiensis (Bt) derived products that are commonly used in insect-resistant genetically engineered (IRGE) crops. These methods probably also could be used to evaluate other types of arthropod stressors, including next generation RNAi-based stressors for IRGE crops. This is an important contribution and step toward using biological markers for such studies. I hope overtime such studies will become more common and more refined as more specific biological markers are found based on the nature of the tested stressors. The authors should be commended because they focused on major trends and not the few differentially expressed genes (DEGs) (potential false positives) that may be due to random events caused by uncontrolled conditions.
Literature citations, however, could be improved. More key papers outside of China should be referenced. More specific recommendation are provided below. The paper is very well written and should be interesting and valuable to readers of PeerJ. I highly recommend publication.
By all accounts, this is a first-rate comprehensive study. The experimental design is excellent (flow chart in Figure 1 was useful) and the analyses and interpretation of results are appropriate. The conclusions are supported by the data.
The authors should consider the following recommendations to improve the manuscript.
37: There are numerous references that could be included here with Liu et al.
44-45: Two of these references are not appropriate because they are not herbivores, predators, parasitoids, or detritivores. Goa et al. is a non-target study with zebrafish, and Jia is a non-target study with honeybees. The honeybee study is more acceptable, but technically they do not fit into the categories listed. Again, several other references could be used here.
Here is a partial list of additional references to consider:
Neher, D.A., A.W.N. Muthumbi, and G.P. Dively. 2014. Impact of coleopteran-active Bt corn on non-target nematode communities in soil and decomposing corn roots. Soil Biology and Biochem. 76:127-135; doi:10.1016/j.soilbio.2014.05.019.
G. P. Dively. 2005. Impact of transgenic VIP3a x Cry1ab lepidopteran-resistant field corn on the nontarget arthropod community. Environ. Entomol. 34: 1267-1291.
Svobodová, Zdeňka, Oxana Skoková Habuštová, William D. Hutchison, Hany M. Hussein, and František Sehnal. "Risk assessment of genetically engineered maize resistant to Diabrotica spp.: influence on above-ground arthropods in the Czech Republic." PloS one 10, no. 6 (2015): e0130656.
Bitzer et al. Environ. Entomol. 34, 1346–1376, https://doi.org/10.1093/ee/34.5.1346
Authors should check all the references to determine if others are inappropriate.
For future studies, the authors should consider using a positive control in addition to the negative control. This would allow them to test the validity of their biological markers. There are many ways to approach this with acute or chronic stressors, or both.
47: consider “. . . plant residues, collembolans are an important group of soil
171: should be “was used to detect”
234: rewrite “has no any negative effect”
Authros did not use the existing regulatory data for non-target effects of Cry1F in maize and cotton that was reviewed and summarized by ILSI (http://ilsi.org/researchfoundation/wp-content/uploads/sites/5/2016/05/Cry1f-monograph.pdf), including data on Folsomia and Eisenia as representative soil invertebrates.
The authors describe existing data on the effects of Bt protein on life history parameters (mortality and sublethal effects) of non-target organisms including C. candida, which found little to no direct toxicity. Nevertheless, they propose that a multigenerational analysis may identify effects that are not apparent over a single generation, and that transcriptomic investigation may provide more sensitive measures of the effects of Cry1F protein. However, the authors fail to establish a plausible mechanism whereby a non-toxic protein that has no measurable effect on a single generation could have ecologically relevant effects that are only detectable through transcriptomics. The introduction needs to describe the plausible risk hypotheses the authors are investigating, including (1) how Cry1F exposure, based on the known mode of action of Bt Cry proteins, could meaningfully affect transcription without measurable life history effects, and (2) how changes in transcriptome over generations could be ecologically important for risk assessment. If transcriptome differences had been found, what would be the ecological harm that could be concluded? Without setting up the study this way, this study encourages fishing expeditions to find differences, which harms the goals of ecological risk assessment. It would suggest that if we don’t find evidence of ecological harm, we must keep looking, instead of concluding that there is no relevant harm.
The authors did not mention in the materials and methods section how the groups of individuals were selected for RNAseq. In particular, transcriptome is likely to vary significantly as the insects develop; therefore if all individuals selected were not in the same developmental stage (instar) and age within instar, there is expected to be considerable uncontrolled variation. Please specify how individuals were selected and discuss how differences in age and instar might affect the ability to detect treatment effects, as well as the biological relevance of not a finding treatment signal in the age/instar noise.
The statistical analysis of the data assigns “generation” as a fixed effect in the ANOVA. However, generation should not be regarded as fixed as individuals could not be randomly assigned to a generation, as they are to a diet treatment. Likewise, generation is not a random the order of the generations is not random. Instead, the correct model for the ANOVA is Repeated Measures. The authors need to replace the two-way ANOVA with a repeated measures ANOVA.
Was the Cry1F diet confirmed to be active against sensitive insects? Both the source of the protein (intended use as a positive control for detection assays, rather than ecotox studies) and the preparation could lead to reduced or removed activity. Confirmation of activity of diet preps is an essential step in ecotox studies, and results of studies (especially with no-effect findings) without this confirmation are not valid
Additional specific comments are provided in the pdf file
All text and materials provided via this peer-review history page are made available under a Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.