Bacillus velezensis T149-19 and Bacillus safensis T052-76 as Potential Biocontrol Agents against Foot Rot Disease in Sweet Potato

Round 1

Reviewer 1 Report

Paper "Bacillus velezensis T149-19 and Bacillus safensis T052-76 as potential biocontrol agents against foot rot disease in sweet potato" is very interesting.

All Tables are perfect, but quality of Figures is very poor. Figures need correction.
Paper needs separately subsection "3.10. Statistical analysis". Sentence "The data were statistically ana-209 lyzed using analysis of variance of means (ANOVA)..." is not enough. Distribution? One-way?, two-way? with/without interaction?

Authors used Dice (not "DICE") as a coefficient of similarity. Why? The Nei and Li method is the best to analysis of similarity based on 0-1 data.

Figures 1 and 2 need LSD of HSD values.

Lack of results of comparison between quantitative and quality data.

The paper is well written. The aim of this paper is interesting and the topic is original but manuscript needs major revision.

Results and conclusions are incorrect because methods used in this paper are wrong.

Author Response

Reviewer 1.

Paper "Bacillus velezensis T149-19 and Bacillus safensis T052-76 as potential biocontrol agents against foot rot disease in sweet potato" is very interesting.

• Thank you very much for the positive feedback. How we dealt with the different points you raised is included below.

All Tables are perfect, but quality of Figures is very poor. Figures need correction.
Paper needs separately subsection "3.10. Statistical analysis". Sentence "The data were statistically analyzed using analysis of variance of means (ANOVA)..." is not enough. Distribution? One-way?, two-way? with/without interaction?

• All the figures now presented are in high resolution - 300 dpi
• A detailed description of the statistical analyses was included in sections 2.6 and 2.9. The qPCR data showed a normal distribution between replicates and treatments and a one-way ANOVA was performed (without interaction). More details were included in the sections mentioned above.

Authors used Dice (not "DICE") as a coefficient of similarity. Why? The Nei and Li method is the best to analysis of similarity based on 0-1 data.

• We used Dice (now corrected) as the coefficient of similarity included in Bionumerics (http://www.applied-maths.com) which is commonly used to analyze DGGE banding patterns. The data typically are coded as (0,1)-vectors, 1 indicating the presence and 0 indicating the absence of a band at a specific position in the gel. Although the Nei & Li coefficient is very much used to determine genetic relatedness based upon DNA restriction fragment patterns (for example), the Dice coefficient is the best-established method in the literature for determining the structure of microbial communities using DGGE.

Figures 1 and 2 need LSD of HSD values.

• The qPCR results were too large for LSD analysis. In that case, according to PAST software’s manual, we were supposed to consider the results of another post hoc test. Tukey’s HSD test is a more conservative statistical analysis and the usage of only Tukey’s HSD to infer statistical significance is extensively used in literature for microbial communities’ analyses.

Lack of results of comparison between quantitative and quality data.

• The numerical matrix generated by the DGGE bands was submitted to NMDS using the PAST statistical program to correlate the DGGE profiles obtained from the different treatments with physiological and plant development parameters (plant height, number of leaves) and fungal abundance (Figure 4). Therefore, both quantitative and quality data have been presented.

The paper is well written. The aim of this paper is interesting and the topic is original but manuscript needs major revision.

• Thank you again for the positive feedback. We did our best to improve the manuscript as suggested.

Results and conclusions are incorrect because methods used in this paper are wrong.

• We do believe on the results we obtained. All statistical analyses were rechecked. The methods described here are extensively used in the literature, including our own previous published papers.

----------------------------------------------------------------------------------------------------------

Reviewer 2 Report

The authors present potentially valuable bacteria for biocontrol for a sweet potato fungal disease. The results are interesting and I will try to contribute with some questions and insights.

Introduction

This section is clear. However, I would add some extra information on 1) how is the control of this disease nowadays in the field; 2) the authors mention breeding strategies do minimize the disease. how are they going?

I think these two pieces of information can be useful for the reader, since they can show how important is the development of the biological control the authors are presenting: is it extremelly necessary, because there are absolutely no other means of control? can it be useful even if breeding succeds, because (maybe) no true resistance could be found yet?

M&M

line 142: the substrate choice was based uniquely on potato´s requirements or some changes were necessary for bacterial survival? in any case, it can the useful for the readers to understand these choices

line 170: the authors mention that symptom evaluation occured always after 60 days of the last innoculum. However, in this case, the pathogen infection time is different in 60 days, which represents half of the experiment time (as mentioned in lines 136-144). How does that impact the results evaluation?

Results

line 291: I believe that the Y axis of Fig 1B showld be "height" instead of "lenght"

lines 316-326: all microrganism names must be italicized

Discussion

This section is sound. But I wonder whether the increased tolerance found is due exlusively by an antimicrobial effect of the Bacillus. Even if previously published results showed an in vitro antagonism (if yes, this could be better explained in this section), in the plant it is not necessarily what is happening. It is possible that bacterial PAMPs are inducing some sort of plant response that can help it to fight the fungal infection. The causality is not clear to me. I don´t think it is necessary to repeat the experiments at this point, but I think it is prudent, for your next experiments, that another control be added: 1) a similar bacillus without antagonism in culture plates OR 2) the addition of a neutralized suspension of bacteria (by boiling the culture, for example). These controls can help us to understand whether the in planta results are directly due to bacterial colonization and an "in planta antagonism" insted of a side-effect of bacterial presence.

Author Response

Reviewer 2

The authors present potentially valuable bacteria for biocontrol for a sweet potato fungal disease. The results are interesting and I will try to contribute with some questions and insights.

• Thank you very much for the positive feedback. How we dealt with the different points you raised is included below.

Introduction

This section is clear. However, I would add some extra information on 1) how is the control of this disease nowadays in the field; 2) the authors mention breeding strategies do minimize the disease. how are they going?I think these two pieces of information can be useful for the reader, since they can show how important is the development of the biological control the authors are presenting: is it extremelly necessary, because there are absolutely no other means of control? can it be useful even if breeding succeds, because (maybe) no true resistance could be found yet?

• Extra information was added to the introduction section concerning the breeding programs and the difficulties on the use of fungicides in different countries/regions.

M&M

line 142: the substrate choice was based uniquely on potato´s requirements or some changes were necessary for bacterial survival? in any case, it can the useful for the readers to understand these choices

• The substrate choice was based uniquely on sweet potato´s requirements.

line 170: the authors mention that symptom evaluation occured always after 60 days of the last innoculum. However, in this case, the pathogen infection time is different in 60 days, which represents half of the experiment time (as mentioned in lines 136-144). How does that impact the results evaluation?

• As all the treatments were evaluated comparatively, we considered a bias of the experimental model of the study.

Results

line 291: I believe that the Y axis of Fig 1B showld be "height" instead of "lenght"

• Corrected

lines 316-326: all microrganism names must be italicized

• Done

Discussion

This section is sound. But I wonder whether the increased tolerance found is due exlusively by an antimicrobial effect of the Bacillus. Even if previously published results showed an in vitro antagonism (if yes, this could be better explained in this section), in the plant it is not necessarily what is happening. It is possible that bacterial PAMPs are inducing some sort of plant response that can help it to fight the fungal infection. The causality is not clear to me. I don´t think it is necessary to repeat the experiments at this point, but I think it is prudent, for your next experiments, that another control be added: 1) a similar bacillus without antagonism in culture plates OR 2) the addition of a neutralized suspension of bacteria (by boiling the culture, for example). These controls can help us to understand whether the in planta results are directly due to bacterial colonization and an "in planta antagonism" insted of a side-effect of bacterial presence.

• Thank you for your concerns. The antagonism (dual) test in plates has previously been performed and one strain did not inhibit the growth of the other strain. However, at this point, we cannot exclude the possibility that the bacterial strains are inducing any plant response.

Author Response File: 

Reviewer 3 Report

In this manuscript the authors described the application of two Bacillus strains against foot rot disease of sweet potato. These Bacillus strains were able to increase plant survival in the presence of the pathogen, increase plant growth and reduce the number of ITS copies of the pathogen in treated samples. The  authors presented also information on some gene clusters encoding the potential production of antibiotic peptides that may be involved in disease reduction. The manuscript is well written, but has a small number of improvements to be made, which are detailed below. 

1. General comments:

One of the limitations of this study is that the authors did not analyse the amount of disease in the treated plants. It is difficult to have an idea of the reduction in disease incidence or severity provided by these bacterial strains. This point should be included in the discussion.

From what I understood, the pot experiment was done only once and data such as the qPCR would benefit from a biological replicate of this experiment. This is also the case with the other data. This limitation should also be included in the discussion.

The combination of the bacterial strains does not seem to provide any additional advantages to the parameters evaluated in the experiment. However, in the discussion the authors argued that the metabolites may be produced in a synergistic way (L395-397). The authors could perform a simple plate experiment to very the compatibility of these strains and add this information to the discussion.

Figure 4 could be deleted. There is no useful information that can be extracted from it. An alternative is to show it in supplementary material, but a topic on this in the results section is unnecessary. 

2. Specific comments:

L. 21 - are not, instead of is

L.41 - The term resistance is used for biotic stresses (all plant pathogens are included here). On the other hand tolerance is used only for abiotic stresses such as drought and temperature. Please, use these terms in the correct way.

L.99-100 - Readers are not interested in who identified the fungus, but in how was it done

L.145 - sweet potato plantlets or cuttings are more appropriate than seedlings for this plant

L. 217 - It would be preferred if the authors used identity instead of similarity when comparing DNA sequences, unless the comparison is being done with amino acids and in this case it should be mentioned here. The term identity will give a more precise indication to the reader. Please, verify the whole text.

Table 1. Some of the identities are too low to be considered here. Please delete them. Are they low because the genomes are not complete?

Additionally, these two tables are very poor in information. They could easily go to supplementary data and the authors could present a simple chart with the gene clusters present in each strain.

L.316-321 - there are some scientific names that should have been in italics.

L. 330 - What does four replicates of each control treatment mean?

L329-335 - What were the samples here? per g of what? it needs to be here.

L. 334 - what is a Tukey’s pairwise test? Tukey’s test is supposed to be a multiple comparison method, not a pairwise method.

L276 -357 - the name of the treatments needs to be standardised to facilitate understanding. The authors need to use the same names in all figures and in the text as they all resulted from the same experiment.

Author Response

Reviewer 3

-In this manuscript the authors described the application of two Bacillus strains against foot rot disease of sweet potato. These Bacillus strains were able to increase plant survival in the presence of the pathogen, increase plant growth and reduce the number of ITS copies of the pathogen in treated samples. The  authors presented also information on some gene clusters encoding the potential production of antibiotic peptides that may be involved in disease reduction. The manuscript is well written, but has a small number of improvements to be made, which are detailed below. 

• Thank you very much for the positive feedback. How we dealt with the different points you raised is included below.

1. General comments:

One of the limitations of this study is that the authors did not analyse the amount of disease in the treated plants. It is difficult to have an idea of the reduction in disease incidence or severity provided by these bacterial strains. This point should be included in the discussion.

• Only the presence or absence of disease symptoms was considered. This is explained in m&m. The reduction (or any alterations) in disease incidence was not considered in this study.

From what I understood, the pot experiment was done only once and data such as the qPCR would benefit from a biological replicate of this experiment. This is also the case with the other data. This limitation should also be included in the discussion.

• This limitation was included in the text

The combination of the bacterial strains does not seem to provide any additional advantages to the parameters evaluated in the experiment. However, in the discussion the authors argued that the metabolites may be produced in a synergistic way (L395-397). The authors could perform a simple plate experiment to very the compatibility of these strains and add this information to the discussion.

• The antagonism (dual) test in plates has previously been performed and one strain did not inhibit the growth of the other strain. This information can be found in our paper - Antonie Van Leeuwenhoek 2019 Apr;112(4):501-512. doi: 10.1007/s10482-018-1181-y.

Figure 4 could be deleted. There is no useful information that can be extracted from it. An alternative is to show it in supplementary material, but a topic on this in the results section is unnecessary. 

• Figure 4 combines the quantitative and quality data obtained in the study. The numerical matrix generated by the DGGE bands was submitted to NMDS using the PAST statistical program to correlate the DGGE profiles obtained from the different treatments with physiological and plant development parameters (plant height, number of leaves) and fungal abundance. Therefore, we believe it brings important information and we do apologize for not excluding it from the paper.

2. Specific comments:
3. 21 - are not, instead of is

• Corrected.

L.41 - The term resistance is used for biotic stresses (all plant pathogens are included here). On the other hand tolerance is used only for abiotic stresses such as drought and temperature. Please, use these terms in the correct way.

• We changed resistant to tolerant.

L.99-100 - Readers are not interested in who identified the fungus, but in how was it done

• Information added.

L.145 - sweet potato plantlets or cuttings are more appropriate than seedlings for this plant

• Changed for plantlets.

1. 217 - It would be preferred if the authors used identity instead of similarity when comparing DNA sequences, unless the comparison is being done with amino acids and in this case it should be mentioned here. The term identity will give a more precise indication to the reader. Please, verify the whole text.

• We used the antiSMASH software – an automated genome mining pipeline - for the identification of core biosynthetic enzymes using genome mining strategies that are based on the sequence similarity of the involved enzymes/genes.

Table 1. Some of the identities are too low to be considered here. Please delete them. Are they low because the genomes are not complete?

• The in silico analysis using antiSMASH predicted the presence of multiple biosynthetic gene clusters (BGCs) in both Bacillus strains, which showed similarity with those for antimicrobial substances. Low levels of similarity with known BGCs do not mean less importance but, in many cases, may suggest novelty of the metabolites from those predicted gene clusters.

Additionally, these two tables are very poor in information. They could easily go to supplementary data and the authors could present a simple chart with the gene clusters present in each strain.

• The tables show the size (in nt) of the gene clusters, the % of similarity, the access number to the database MIBiG, in addition to the genes coding for antimicrobial substances in the genomes of  both Bacillus strains. It will be difficult to provide all the information with a simple chart. Moreover, other reviewer considered the tables well presented as he/she stated “All Tables are perfect”.

L.316-321 - there are some scientific names that should have been in italics.

• Corrected.

1. 330 - What does four replicates of each control treatment mean?

• Four replicates of each treatment. Corrected.

L329-335 - What were the samples here? per g of what? it needs to be here.

• You are right. per g of soil. Corrected

1. 334 - what is a Tukey’s pairwise test? Tukey’s test is supposed to be a multiple comparison method, not a pairwise method.

• Tukey method: This test uses pairwise post-hoc testing to determine whether there is a difference between the mean of all possible pairs using a studentized range distribution. This method tests every possible pair of all groups. Corrected in the text.

L276 -357 - the name of the treatments needs to be standardised to facilitate understanding. The authors need to use the same names in all figures and in the text as they all resulted from the same experiment.

• The nomenclature T1 – (F); T2 – (19 + F); T3 – (76 + F); T4 – (19 + 76 + F); T5 – (F + 19); T6 – (F + 76); T7 – (F + 19 + 76) was used in all figures and in the text.

Author Response File:

Round 2

Reviewer 1 Report

Now all is perfect.

Reviewer 3 Report

The authors did most of the improvements that were necessary and I consider it acceptable the way it is.