Transcriptomic Analysis of Female Panicles Reveals Gene Expression Responses to Drought Stress in Maize (Zea mays L.)
Round 1
Reviewer 1 Report
Reviewer Comments for “Transcriptomic analysis of female panicles reveals gene expression responses to drought stress in maize (Zea mays L.)”
The manuscript entitled “Transcriptomic analysis of female panicles reveals gene expression responses to drought stress in maize (Zea mays L.)” by Jia et al. investigated the influence of drought stress on female panicales in maize.
Although the paper was well written and the figures appropriate, the manuscript must be improved before it is ready for publication.
Some comments:
Sentence beginning line 41: “When external drought stimuli are perceived..” is difficult to follow. Consider revising or breaking into multiple sentences.
Introduction: In general, the introduction is well written but feels incomplete. It would be good to go into more detail regarding the stages of panicle differentiation and development as well as developing out the paragraph on studies in maize in response to drought stress.
Methods section 2.1 Plant material and growth conditions:
This section is lacking some important information. First off, I am not exactly certain what is meant by ‘proving ground mobile canopy’ or ‘soil growth pool’. These need to be reworded or clarified. More information also needs to be given about the field experiment- soil type, average temperatures throughout the field trial, etc. should be described. How many plots were there? A diagram may help to clarify some of these things.
Throughout the text: the author goes between the terms dry matter and biomass accumulation- it would be better to keep it consistent.
Figure 1: May be a good idea to reformat the figures (two on top and two on the bottom in a square instead of all in a row) so that they are a bit easier to read. The legend also needs to be larger. Also, the figures do not need the break on the y-axis as the y-axis does not have to start at zero as long as the lower limit is labeled.
Figure 4: labels need to be larger
In general, there are too many abbreviations within the manuscript, which were difficult to keep track of. Consider removing a few.
Lines 169-172: Keep formatting of numbers consistent (different number of decimal points)
Author Response
Comment 1. Sentence beginning line 41: “When external drought stimuli are perceived..” is difficult to follow. Consider revising or breaking into multiple sentences.
Response: We have accepted this valuable comment, and in the revised manuscript, the sentence has been revised into two sentences. For example, the previous sentence was revised to “When external drought stimuli are perceived and captured by sensors on cell membranes, the signals are transmitted through multiple signal transduction pathways. Then, plant can regulate the expression of drought-responsive genes to protect themselves from the harmful effects of external stimuli”.
Comment 2. Introduction: In general, the introduction is well written but feels incomplete. It would be good to go into more detail regarding the stages of panicle differentiation and development as well as developing out the paragraph on studies in maize in response to drought stress.
Response: Thanks for your suggestion. In the third paragraph of Introduction, details of panicle differentiation were added, but the studies on panicle differentiation and development in response to drought stress through transcriptome analyses are not sufficient. Please see the details in the revised manuscript.
Comment 3. Methods section 2.1 Plant material and growth conditions:
This section is lacking some important information. First off, I am not exactly certain what is meant by ‘proving ground mobile canopy’ or ‘soil growth pool’. These need to be reworded or clarified. More information also needs to be given about the field experiment- soil type, average temperatures throughout the field trial, etc. should be described. How many plots were there? A diagram may help to clarify some of these things.
Response: In the revised manuscript, ‘proving ground mobile canopy’, ‘soil growth pool’ have been changed to ‘a movable awning’ and ‘experimental plot’. Field experiment soil type was Fluvo-aquic soils, which were added in line No. 77-78. The air temperature and precipitation over the growing season were shown in following Figure 1. In addition, there were nine plots with a diagram in following Figure 2.
Comment 4. Throughout the text: the author goes between the terms dry matter and biomass accumulation- it would be better to keep it consistent.
Response: We have accepted this suggestion and checked all text of the manuscript. In the revised manuscript, ‘biomass accumulation’ have been changed into ‘dry matter’.
Comment 5. Figure 1: May be a good idea to reformat the figures (two on top and two on the bottom in a square instead of all in a row) so that they are a bit easier to read. The legend also needs to be larger. Also, the figures do not need the break on the y-axis as the y-axis does not have to start at zero as long as the lower limit is labeled. Figure 4: labels need to be larger
Response: Thanks! In revised MS, Figure 1 and Figure 5 had been reformatted according to your valuable suggestions. In Figure 1, two on top and two on the bottom in a square were instead of all in a row. Due to the addition of the new figure, Figure 4 has been changed to figure 5 and labels of Figure 5 became larger.
Comment 6. In general, there are too many abbreviations within the manuscript, which were difficult to keep track of. Consider removing a few.
Response: Thanks for your suggestions. In revised MS, ‘PDI’ and ‘PDL’ were removed.
Comment 7. Lines 169-172: Keep formatting of numbers consistent (different number of decimal points)
Response: We have accepted your careful check. And in revised MS, ‘32.0%’ has been changed to ‘32.00%’, and we have checked all numbers of the manuscript and kept number of two decimal points consistent.
Author Response File: 
Author Response.docx
Reviewer 2 Report
In this manuscript, Jia et al. describe phenotypic and transcriptomic response of Zea mays female panicles in response to light and mild drought stresses by conducting RNA-seq analyses. The transcriptome approach is important to understand the drought stress response mechanism at tissue-specific level. However, one major drawback of this manuscript is that it lacks validation/follow-up from the transcriptome result. Also, some clarification and data deposition need to be done before publication. Detail of the concerns are outlined below.
The raw RNA-seq data must be deposited to a repository before submission. Please follow the instructions for authors and provide their accession number sin the text.
RNA-seq results need to be validated by other methods such as reverse transcription quantitative PCR (RT-qPCR), at least for genes important to this paper (TE1, CUC2, DLF1, ARFs, GH3, TIR1, SAURs, GSTU6, GST12, etc.)
It’s interesting that LD vs CK show only 40 DEGs, despite that there were clear phenotypic differences similar to MD at 30 DAD (Fig. 2). Why?
l.121: definition of “clean data (high-quality reads) need to be clarified. How were adapter-containing reads removed? Were these trimmed? Also, were the p-values false-discovery rate (FDR) corrected?
l.138-139: Here it states that all data are three independent replicates, but different numbers are indicated in some places (l.160 and l.179). I’d suggest adding “unless otherwise stated” to l. 139.
l.275-278: Expression of CUC2 is temporary and spatially regulated during development/morphogenesis. Is it possible that LD and MD delay the ear maturation process? If CUC2 is specifically expressed during maturation, higher abundance of CUC2 transcript in LD and MD-treated tissues compared to CK may be simply due to maturation-level difference, rather than “up-regulation” in response to stress which implies a distinct mechanism to activate gene expression through stress-responsive cis-elements.
l.290-291: In addition to RT-qPCR validation, auxin and auxin-conjugate levels should be examined to assess if the drought stress affect auxin biosynthesis/metabolism in female panicles.
l.308-313: GSTs are both up- and down-regulated. Are there change in the overall enzyme activity? GST activity in extract from LD, MD and CK treated female panicles should be examined.
Fig.1 (b) and (c): 30-day plots show a, ab and c. Shouldn’t these be a, a and b? Not sure why “c” is needed here.
Figs. 4 and 5: Labels are too small and very difficult to read. Please enlarge the labels to make them readable.
Fig. 6: What do the heat-maps and numbers represent? What does the color-coding mean? What are the meaning of green/yellow highlight in some boxes?
Table S2: It seems odd that all 12 libraries show the same “49.42” million raw reads. Please double check. If they are correct, show the entire number like columns 5 and 6.
Author Response
Comment 1. The raw RNA-seq data must be deposited to a repository before submission. Please follow the instructions for authors and provide their accession number sin the text.
Response: We have uploaded the files to NCBI SRA (Submission ID: SUB6906596. BioProject ID: PRJNA604094.) and submitted an application to NCBI to generate the accession number and web link, but we haven't received a reply yet. And, we will add the results to the paper once receive the accession number and web link.
Comment 2. RNA-seq results need to be validated by other methods such as reverse transcription quantitative PCR (RT-qPCR), at least for genes important to this paper (TE1, CUC2, DLF1, ARFs, GH3, TIR1, SAURs, GSTU6, GST12, etc.)
Response: Thanks for your suggestion. We have randomly selected eight genes for validation of RNA-seq by qRT-PCR experiment, as shown in Figure 4. & Supplementary Materials Table S4. of the revised manuscript. This validate the differential expression of the genes identified as being under drought stress.
Comment 3. It’s interesting that LD vs CK show only 40 DEGs, despite that there were clear phenotypic differences similar to MD at 30 DAD (Fig. 2). Why?
Response: We have carefully checked the pictures of Figure 2a and statistics of Figure 2b in details. Because Figure 2a could not represent adequately the sample used for transcriptome sequencing, one suitable picture has been replaced. As for Figure 2b, we have added 5 replicate sample data and then recalculated all the data in the revised manuscript. Combining Figure 2a, b and transcriptome data, there are not significant differences between CK and LD at 30 DAD. However, the drought stress continued to the flowering period, and drought stress had a great influence on maize flowering and spinneret spacing and pollination, therefore, LD greatly affected yield for of mature maize plants (Figure 2c).
Comment 4. l.121: definition of “clean data (high-quality reads) need to be clarified. How were adapter-containing reads removed? Were these trimmed? Also, were the p-values false-disc overy rate (FDR) corrected?
Response: We have accepted your valuable suggestion. The source of ‘clean data’ and the method of adapter-containing reads removed have been clarified in line 120-122. When RNA-seq data are used to compare and analyze the differential expression of the same gene in two samples, there are two criteria can be selected: one is FoldChange, which is a multiple of the expression level of the same gene in two samples. The second is p-value (p-value of negative binomial distribution test H0: A=B) or FDR (adjusted p-value). We choose p-value instead of p-adj (p-value corrected by FDR) in the manuscript.
Comment 5. l.138-139: Here it states that all data are three independent replicates, but different numbers are indicated in some places (l.160 and l.179). I’d suggest adding “unless otherwise stated” to l. 139.
Response: Thanks for your careful check. In revised MS, “three independent replicates” have changed to “All data are expressed as the mean ± SD value from three independent experiments unless otherwise stated”, and accurate replicates were added in each figure legend.
Comment 6. l.275-278: Expression of CUC2 is temporary and spatially regulated during development/morphogenesis. Is it possible that LD and MD delay the ear maturation process? If CUC2 is specifically expressed during maturation, higher abundance of CUC2 transcript in LD and MD-treated tissues compared to CK may be simply due to maturation-level difference, rather than “up-regulation” in response to stress which implies a distinct mechanism to activate gene expression through stress-responsive cis-elements.
Response: Thanks for your comment. The genes te1 and cuc2 are related with growth and development panicle. We have found that cuc2 is specifically expressed during maturation (Data not listed in MS), thus “Here, although te1 and cuc2 were up-regulated under moderate drought stress (Figure 4, Supplementary Table S5), implying a distinct mechanism to activate gene expression due to MD induced maturation-level difference, rather than “up-regulation”.” was added in revised MS.
Comment 7. l.290-291: In addition to RT-qPCR validation, auxin and auxin-conjugate levels should be examined to assess if the drought stress affect auxin biosynthesis/metabolism in female panicles.
Response: Your suggestion is valuable for the paper. Auxin biosynthesis/metabolism in plants under drought have been reported in the past research [1,3]. Under drought stress, the change of IAA content increases in some plants [2], while decreases in others [4]. Through RNA-seq, we have found that ‘the expression levels of auxin biosynthesis genes of female panicles were down-regulated after MD stress (Figure 7). And auxin level was also reduced (Supplementary Materials Figure S1a) according to the method [2].
References:
[1] Yang H, Gu X, Ding M, et al. Activities of starch synthetic enzymes and contents of endogenous hormones in waxy maize grains subjected to post-silking water deficit [J]. Scientific reports, 2019, 9(1): 1-8.
[2] Zhao, M. R., Han, Y. Y., Feng, Y. N., Li, F., Wang, W. (2012): Expansionsare involved in cell growth mediated by abscisic acid and indole-3-acetic acid under drought stress in wheat. –Plant Cell Reports 31(4): 671-685.
[3] Yu J, Jiang M, Guo C. Crop Pollen Development under Drought: From the Phenotype to the Mechanism[J]. International journal of molecular sciences, 2019, 20(7): 1550.
[4] Zhang, H. Y., Duan, W. X., Xie, B. T., Dong, S. X., Wang, B. Q., Shi, C. Y., Zhang, L. M. (2018): Effects of drought stress at different growth stages on endogenous hormones and its relationship with storage root yield in sweet potato. –Acta Agronomica Sinica 44(1): 126-136.
Comment 8. l.308-313: GSTs are both up- and down-regulated. Are there change in the overall enzyme activity? GST activity in extract from LD, MD and CK treated female panicles should be examined.
Response: We have accepted your valuable suggestions. GST activity in extract from LD, MD and CK treated female panicles were examined according to AbdElgawad et al. method [1]. The GST activity was significantly up-regulated in maize female panicles under MD (Supplementary Materials Figure S1b), which was consistent with previous study [2]. Accordingly, the sentence about GST activity was added in line 336-338 in revised MS.
[1] AbdElgawad H, Zinta G, Hegab M M, et al. High salinity induces different oxidative stress and antioxidant responses in maize seedlings organs[J]. Frontiers in plant science, 2016, 7: 276.
[2] Wang C, Lu G, Hao Y, et al. ABP9, a maize bZIP transcription factor, enhances tolerance to salt and drought in transgenic cotton[J]. Planta, 2017, 246(3): 453-469.
Comment 9. Fig.1 (b) and (c): 30-day plots show a, ab and c. Shouldn’t these be a, a and b? Not sure why “c” is needed here.
Response: Thanks, we have corrected the 30-day plots of Fig.1 (b) and (c), please see Figure 1 in revised MS.
Comment 10. Figs. 4 and 5: Labels are too small and very difficult to read. Please enlarge the labels to make them readable.
Response: Labels of Figures 5 &6 (Figs. 4 and 5 have changed to Figs.5 &6) have been enlarged and became more explicit.
Comment 11. Fig. 6: What do the heat-maps and numbers represent? What does the color-coding mean? What are the meaning of green/yellow highlight in some boxes?
Response: Explanatory legend of Figure 7 (Fig.6 have changed to Fig.7) has been added in the revised manuscript.
Comment 12. Table S2: It seems odd that all 12 libraries show the same “49.42” million raw reads. Please double check. If they are correct, show the entire number like columns 5 and 6.
Response: The size of the data volume is uneven, some have 7 G, some have 10 G clean data, in order to ensure that the data volume is similar, the same data volume is intercepted uniformly. In this way, the quantitative analysis of genes will not be affected. If the data volume is not intercepted, it will affect the quantitative calculation of genes.
Author Response File: Author Response.docx
Round 2
Reviewer 1 Report
Overall, the authors have greatly improved the manuscript.
I have a few minor comments that could further improve the document.
In response to comment 3: While you have added figures for experimental design and average temperatures to the comment, I do not see anything reflected in the document. please add information regarding temperatures during the trial to the text.
Line 121: it is good you have clarified how the raw reads were trimmed, but you need a reference for the program Trimmomatic.
Author Response
Comment 1. In response to comment 3: While you have added figures for experimental design and average temperatures to the comment, I do not see anything reflected in the document. please add information regarding temperatures during the trial to the text.
Response: Thanks for your suggestion. In revised MS, we have added the average daily maximum and minimum temperatures in line 85-86.
Comment 2. Line 121: it is good you have clarified how the raw reads were trimmed, but you need a reference for the program Trimmomatic.
Response: Thanks for your suggestion. That was a mistake when I compiled the references. Reference [35] should be the reference of sentence ‘Raw data (raw reads) were processed using Trimmomatic’, which included the usage of Trimmomatic. We have changed the position of [35] from line 121 to 122 in revised MS.
[35] Bolger, A.M.; Lohse, M.; Usadel, B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics 2014, 30, 2114.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
The authors have improved their manuscript considerably by adding some new data and clarifying technical details. There are a few concerns left, which need to be addressed before publication in Agronomy.
In response to Re: Comment #2; the authors confirmed differential expression of randomly selected genes, instead of the ones discussed further in the paper. The approach may validate the overall RNA-seq quality, it is still better to validate genes directly relevant to this study, rather than the randomly chosen genes. I request to validate the expression pattern of TE1, CUC2 and DLF1 by qPCR, because there is no additional lines of support provided for these genes in the manuscript.
In response to Re: Comment #3; Fig. 2b should include error bars to show the variability. Also, Fig. 2c radar chart seems to be the same as the original, even though after adding 5 more replicates. Table S2 presents 10 replicates. Did the additional 5 replicates apply only to Fig. 2a and 2b? If so, it needs to be clarified in the legend. Or, revise Fig. 2c and Table S2 with the latest 15-replicate data.
In response to Re: Comment #6; Perhaps my previous comment 6 was not clear enough. I meant that the differential expression can be explained simply by the fact that the development/maturation is delayed, WITHOUT any stress-responsive gene regulation mechanism. I would rephrase the sentence (l.299-300) to state that the observed expression level difference may be simply due to difference in the degree of maturation during the stress.
In response to Re: Comment #12; Difference in the library size is normalized by the FPKM (per million reads) method for the quantitative analysis. What do you exactly mean by "data volume"? In the method and the Table S2 legend, please describe any normalization measures used (other than FPKM), if any.
Author Response
Comment 1. In response to Re: Comment #2; the authors confirmed differential expression of randomly selected genes, instead of the ones discussed further in the paper. The approach may validate the overall RNA-seq quality, it is still better to validate genes directly relevant to this study, rather than the randomly chosen genes. I request to validate the expression pattern of TE1, CUC2 and DLF1 by qPCR, because there is no additional lines of support provided for these genes in the manuscript.
Response: Thanks, we have accepted your suggestion, and have supplemented the three genes’ expression level by qPCR. Thus, we have added the three genes (TE1, CUC2 and DLF1) in corresponding parts of revised MS. Meanwhile, we have modified the Supplementary Material Table S1 & S4 accordingly.
Comment 2. In response to Re: Comment #3; Fig. 2b should include error bars to show the variability. Also, Fig. 2c radar chart seems to be the same as the original, even though after adding 5 more replicates. Table S2 presents 10 replicates. Did the additional 5 replicates apply only to Fig. 2a and 2b? If so, it needs to be clarified in the legend. Or, revise Fig. 2c and Table S2 with the latest 15-replicate data.
Response: Thanks for your comment. In the revised manuscript, Fig. 2b has been added error bars.
We have added 5 replicates to Fig. 2b, which were clarified in the legend in line 192-193. But there was not additional 5 replicates in Fig. 2c.
Comment 3. In response to Re: Comment #6; Perhaps my previous comment 6 was not clear enough. I meant that the differential expression can be explained simply by the fact that the development/maturation is delayed, WITHOUT any stress-responsive gene regulation mechanism. I would rephrase the sentence (l.299-300) to state that the observed expression level difference may be simply due to difference in the degree of maturation during the stress.
Response: Thank you for your kindly guidance. We are sorry for wrong understanding in round 1 revision. According your suggestion, “combined with developmental change (Figure 2a), implying that the differential expression of the gene under MD treatments may be related to the mature delay of the FPs tissue” has replaced the previous sentence in the revised MS.
Comment 4. In response to Re: Comment #12; Difference in the library size is normalized by the FPKM (per million reads) method for the quantitative analysis. What do you exactly mean by "data volume"? In the method and the Table S2 legend, please describe any normalization measures used (other than FPKM), if any.
Response: Data volume means the size of original data from Illumina. After drawing randomly, the data volume will be relatively average. In special cases, raw reads will be the same when counted with two decimals after M and G. Because of random selection, it will not affect the subsequent analysis. As for the standardization of quantification that you mentioned, we use FPKM to express standardization exactly. Some samples’ size of original data from Illumina vary widely, on the one hand, drawing randomly acts as standardization for raw data, on the other hand, drawing randomly to a certain extent can avoid some tasks running slower and occupying more resources because of the large amount of data (original data of some samples is 12 G, some is 7 G under extreme cases).
In the revised manuscript, we have described the randomly selection of raw data by SEQTK software in Table S2 legend.
Author Response File: Author Response.pdf
