Comparison of Endophytic and Epiphytic Microbial Communities in Surviving and Dead Korean Fir (Abies koreana) Using Metagenomic Sequencing

Round 1

Reviewer 1 Report

In the manuscript “Comparison of Endophytic Bacterial and Fungal Communities in Surviving and Dead Abies koreana using metagenomic sequencing” the authors report a study on the comparison of endophytic bacteria and fungi in live and dead plants of Abies koreana.

The paper is interesting although there are some important points that can be improved:

1- Introduction: too generic, little focused on the disease and the causes of dead of the considered plants. It would be useful to implement it with a more specific bibliography reflecting the problem and similar case studies.

2- Materials and methods: in general, this whole section needs to be improved. In “Study site and tree sampling” more details on plants conditions (especially for dead samples) are needed in order to explain their conditions.

The samples analyzed (3) is very limited for a statistical approach particularly if the aim is to highlight the causes of a disease using a metagenomic approach. There are many inaccuracies and unreported instances (example: no references for some primers and protocols adopted). On the other hand, the pure statistics part seems well done and explained.

3- Results, discussion, and conclusions: apart from the improvements here is there, in my opinion they are written quite well

 

From the methodological point of view, I would like to underline the following points mainly considering the aim proposed by the authors: “elucidate the possible relationship between endophyte communities and the mortality of Korean fir trees”.

1- The site (s) of action of the pathogens involved in the disease is not considered in sampling. For ex. if it were a root disease, and the samples are taken in the trunk, it is obvious that only pathogens capable of moving through the vascular system could be identified.

2- The use of only 3 biological replicas in this type of study (microbiome) could be a limit in order to obtain robust and reliable results. It seems to me a lot of reductive and can lead to various artifacts. In fact, a low number of total reads can be appreciated in the results.

3- Concerning to DNA amplification; specifically, temperature 55 °C was used for 30 s in the annealing step. I consider that very few for both 16S and ITS. From my experience, with such a low T °, specificity of reaction is little, leading to a lower overall amplification efficiency, which can generally translate into fewer ASVs per organism. Again, the total number of reads is few: this is due to the few samples (3 replicates) and to the fact that the authors, after filtering the reads, obtained real sequences relating to microorganisms in a very low number (moreover, probably influenced, as reported, by co-amplification with plant ribosomal DNA). Put simply, there is not such a number of reads for adequate genomic coverage.

Concerning to the general approach of the work, which leads to the Results, Discussion and Conclusion, I don’t fully agree on the use exclusive of the metagenomic approach. Metagenomics fits well in other contexts is correct when it is useful to provide a broad vision: (i.e. carposphere microbiome in berries, soil microbiome, rhizosphere microbiome). In this paper the focus is a disease, in which generally the involved microorganisms (pathogens) are one or few. Therefore, being able to obtain outcomes as families or genders is not enough detailed. This leads only to speculations on who may actually be the actors involved, as is reported in the discussion and in the conclusion by the authors, without really going to the detail and without, therefore, being able to conclude properly.

Comments for author File: Comments.pdf

Author Response

Reviewer1

Comments and Suggestions for Authors

In the manuscript “Comparison of Endophytic Bacterial and Fungal Communities in Surviving and Dead Abies koreana using metagenomic sequencing” the authors report a study on the comparison of endophytic bacteria and fungi in live and dead plants of Abies koreana.

The paper is interesting although there are some important points that can be improved:

1- Introduction: too generic, little focused on the disease and the causes of dead of the considered plants. It would be useful to implement it with a more specific bibliography reflecting the problem and similar case studies.

Answer: Thank you for your comments on the introduction. We have now included some introductory sentences on how alteration of endophyte communities affects the plant susceptibility to diseases: “For example, the microbial diversity in horse chestnut (Aesculus hippocastanum) trees is negatively correlated with bleeding canker disease symptom severity [1]. In Pinus monticola, exogenous inoculation of fungal endophytes to seedlings confers resistance to the pathogen Cronartium ribicola, the fungal causal agent of white pine blister rust disease [2].” (lines 58-62)   

 

2- Materials and methods: in general, this whole section needs to be improved. In “Study site and tree sampling” more details on plants conditions (especially for dead samples) are needed in order to explain their conditions.

Answer: We chose trees having at least 25 cm of diameter at breast height (DBH), indicating that these trees were at least 50 years old trees. Since Korean fir trees are coniferous tree species, living trees maintain vivid green leaves regardless of seasons, whereas dead trees exhibit a standing dead appearance showing that none of the leaves were attached to the branches of trees. To date, the causal agents of the declining Korean fir trees are largely unknown. Therefore, we determined the health conditions of trees based on the apparent conditions of leaves on the branches. For living trees, we chose three trees that have vivid green leaves on all branches of trees. To collect dead tree samples, we carefully chose likely dying trees that have both greenish and yellowish leaves on 10% of branches whereas the remaining 90% of the branches have no leaves. We believe that these dead trees can represent the declining trees and be suitable for our study, since they might be almost dead but not completely dead trees. To collect sawdust samples, we made the 6 cm of four holes in four different directional spots showed a clean appearance without any contaminations. We intensively rewrote the ‘2.1. Study site and tree sampling’ section as follows: “We randomly selected three dead and live trees with at least 25 cm of diameter at breast height in the same sites at intervals of > 10 m to choose trees in similar growth environmental conditions (Supplementary Figure 1). Since the causal agents of the declining Korean fir trees are largely unknown, we determined the health conditions of trees based on the apparent conditions of leaves on the branches. We chose three live trees that had vivid greenish leaves on all branches of trees. For the selection of dead trees, we carefully chose three dying trees that had both greenish and yellowish leaves on 10 % of branches whereas the remaining 90% of the branches had no leaves. To extract endophytic microbial DNA from the Korean fir trees, we collected sawdust samples from three live (A1-A3) and three dead (D1-D3) trees by drilling the 6 cm of four holes on the trunks at four different directional spots with a clean appearance. The sawdust samples were mainly composed of sapwood and trace amounts of cambium and bark.” (lines 93-105)

 

The samples analyzed (3) is very limited for a statistical approach particularly if the aim is to highlight the causes of a disease using a metagenomic approach.

Answer: The three samples of our 16S rRNA and ITS sequencing data were distinctly clustered depending on the health condition of trees, according to our principal coordinate analysis, indicating that these three samples are enough to perform statistical analysis to understand how the health conditions of trees affect endophyte communities. In the present manuscript, we cannot connect any particular microbial taxa to the causal agent of declining trees as well as diseases. The main goal of this paper is to understand the composition of endophytic microbial compositions depending on the health condition of trees and predict their potential roles in tree fitness using bioinformatic analyses. We modified sentences in the introduction section that may mislead that our aim: “In this study, we analyzed the composition of bacterial and fungal endophytes of live and dead Korean fir trees to understand how the health conditions of trees affect endophyte communities and predict their ecological roles.” (lines 75-77). “Our findings would be useful for understanding the potential role of endophytic microorganisms in the endangered Korean fir trees and improving the sustainability of trees from ecologically vulnerable environments.” (lines 87-89)

 

There are many inaccuracies and unreported instances (example: no references for some primers and protocols adopted). On the other hand, the pure statistics part seems well done and explained.

Answer: To comply with your comments on primer sequences, we added two references [3-5] in the Materials and Methods section for primers used in the amplification of bacterial 16S rRNA genes and fungal ITS sequences (lines 114 and 117).

 

3- Results, discussion, and conclusions: apart from the improvements here is there, in my opinion they are written quite well

Answer: Thank you for your comments, which helped improve our manuscript.

 

From the methodological point of view, I would like to underline the following points mainly considering the aim proposed by the authors: “elucidate the possible relationship between endophyte communities and the mortality of Korean fir trees”.

1- The site (s) of action of the pathogens involved in the disease is not considered in sampling. For ex. if it were a root disease, and the samples are taken in the trunk, it is obvious that only pathogens capable of moving through the vascular system could be identified.

Answer: Since a large portion of biotic factors, particularly pathogens and diseases, are largely unknown in Korean fir trees, it is very difficult to sample the specific part of trees suffering from pathogen infections. To explain the sites of sampling position in detail, we modified the Materials and Methods section as above (lines 93-105).

 

2- The use of only 3 biological replicas in this type of study (microbiome) could be a limit in order to obtain robust and reliable results. It seems to me a lot of reductive and can lead to various artifacts. In fact, a low number of total reads can be appreciated in the results.

3- Concerning to DNA amplification; specifically, temperature 55 °C was used for 30 s in the annealing step. I consider that very few for both 16S and ITS. From my experience, with such a low T °, specificity of reaction is little, leading to a lower overall amplification efficiency, which can generally translate into fewer ASVs per organism. Again, the total number of reads is few: this is due to the few samples (3 replicates) and to the fact that the authors, after filtering the reads, obtained real sequences relating to microorganisms in a very low number (moreover, probably influenced, as reported, by co-amplification with plant ribosomal DNA). Put simply, there is not such a number of reads for adequate genomic coverage.

Answer: Since previous studies revealed that the optimal annealing temperature for primers (341F, 805R, ITS3, ITS4) used in this study is 55 ℃ [3-5], we used the same annealing temperature to amplify bacterial 16S and fungal ITS sequences using PCR. For readers who might have a similar question on the annealing temperature of DNA amplification, we added these references in the materials and methods section (lines 114 and 117).

Since we amplify 16S and ITS sequences using sawdust samples from the trunks of trees, it is inevitable to amplify DNA fragments from plastidial and mitochondrial DNA. We filtered the amplified DNA fragments derived from plant plastidial and mitochondrial DNA as described in the material and methods (lines 131-132). After filtering, we obtained at least 50,000 high-quality 16S rRNA reads and 14,000 high-quality ITS reads per tree sample. We believe that the number of filtered reads per sample is suitable to represent the genomic coverage of endophyte in trees, since our rarefaction analysis showed that bacterial and fungal communities were about to reach saturation at approximately 8,000 and 3,800 sequencing depths (Figure 1, lines 161-165). To provide detailed information about the number of high-quality 16S rRNA and ITS reads per sample, we added a supplementary table in the result section (line 159, Supplementary Table 1).

Sample ID	16S rRNA reads	ITS reads
A1	59,192	53,715
A2	52,530	63,444
A3	53,054	41,082
D1	84,800	17,217
D2	87,310	30,411
D3	76,987	14,542
sum	413,873	220,411

Supplementary Table 1. Number of high-quality 16S rRNA and ITS reads after filtering

 

Concerning to the general approach of the work, which leads to the Results, Discussion and Conclusion, I don’t fully agree on the use exclusive of the metagenomic approach. Metagenomics fits well in other contexts is correct when it is useful to provide a broad vision: (i.e. carposphere microbiome in berries, soil microbiome, rhizosphere microbiome). In this paper the focus is a disease, in which generally the involved microorganisms (pathogens) are one or few. Therefore, being able to obtain outcomes as families or genders is not enough detailed. This leads only to speculations on who may actually be the actors involved, as is reported in the discussion and in the conclusion by the authors, without really going to the detail and without, therefore, being able to conclude properly.

Answer: The aim of our manuscript is not to reveal the connection between any specific microbial taxa and the causal agent of declining Korean fir trees, since none of the studies has reported the pathogenic effect on declining Korean fir trees to date. In the present manuscript, we’d like to dissect the composition of endophytic microbial compositions depending on the health condition of trees, thereby providing a broad vision of endophytes in Korean fir trees. Furthermore, we tried to predict the potential ecological roles of preferentially enriched endophytic microbe using several bioinformatic analyses. Therefore, we believe that our manuscript can provide a comprehensive understanding of the endophytic microbiome in endangered Korean fir trees using a metagenomic approach.

 

Author Response File: Author Response.pdf

Reviewer 2 Report

This study aims to elucidate the role of endophytic bacteria and fungal communities in the decline and mortality of Korean fir. The authors report a comparison of endophytic bacterial and fungal communities between living and dead Abies koreana trees using metagenomics. The metagenomic and community analyses are well conducted and well written.

In my opinion, the main weakness of the work is the study design and sampling. Firstly, the sampling included only six trees (three alive; A-C, three dead; D-F), and the sampled living trees seem to be close to each other and located separate from the sampled dead trees, suggesting that local growing conditions may potentially affect the microbiome in addition to tree condition. The investigation would have benefitted from larger more randomized sampling.

Furthermore, it is not clear how and why the sampled trees were selected. Did the living trees appear healthy? What kind of symptoms do the declining trees have?  Since the aim of the study was to understand the role of endophytic microorganisms in the mortality of Korean fir trees, it would have been more beneficial to also sample healthy trees as well as trees suffering from the decline, but not dead. If there are pathogenic organisms involved in the decline they are likely not detected anymore in dead trees; instead you find more saprophytes who require dead material (as was reported in this study). The authors should be careful when drawing conclusions about microbes enriched in the dead trees having an effect on tree health. They may simply be saprotrophs colonizing the already dead material.

In addition, sampling is not well described. The sampling was done by collecting saw dust by drilling the tree trunks, but it is not mentioned if the samples included bark, cambium, sapwood and/or heartwood, which likely host different endophyte populations. Was the bark removed first? How was it ensured that the samples did not include epiphytes? Is it possible that the lichen forming fungi detected in this study originate from epiphytic growth on the trunks of the living trees?

In my opinion, the interpretation of the results should be reconsidered/revised according to the points raised above. Some additional comments are also included in the pdf.

Comments for author File: Comments.pdf

Author Response

Reviewer2

 

Comments and Suggestions for Authors

This study aims to elucidate the role of endophytic bacteria and fungal communities in the decline and mortality of Korean fir. The authors report a comparison of endophytic bacterial and fungal communities between living and dead Abies koreana trees using metagenomics. The metagenomic and community analyses are well conducted and well written.

In my opinion, the main weakness of the work is the study design and sampling. Firstly, the sampling included only six trees (three alive; A-C, three dead; D-F), and the sampled living trees seem to be close to each other and located separate from the sampled dead trees, suggesting that local growing conditions may potentially affect the microbiome in addition to tree condition. The investigation would have benefitted from larger more randomized sampling.

Answer: The three samples of our 16S rRNA and ITS sequencing data were clearly clustered depending on the health condition of trees, according to our principal coordinate analysis, indicating that these three samples are enough to perform statistical analysis to understand how the health conditions of trees affect endophyte communities.

We agree with the reviewer’s comments that larger and more randomized tree samples can improve the understanding of microbial structures in the endangered Korean fir trees. However, we believe that understanding microbial communities in trees located in Mt. Jiri is suitable to understand how the health conditions of Korean fir trees affect endophyte communities and predict their ecological roles, since Mt. Jiri is one of the two major habitats of Korean fir trees (the other habitat is Mt. halla) and Korean fir trees form forest habitats in limited subalpine areas [6]. We chose both live and dead trees from two distinct ridge positions in Mt. Jiri due to clearly distinguishing live trees from dead trees. When we monitored the healthy conditions of living trees nearby dead trees, they declined within a few years, suggesting that it is very difficult to figure out authentic healthy live trees nearby dead trees. To collect samples from live trees under healthy conditions, we chose surviving trees from the distinct area from the dead trees where none of the dead trees were observed. When we monitored environmental conditions including moisture and temperature, there was no significant difference in these two areas.

 

Furthermore, it is not clear how and why the sampled trees were selected. Did the living trees appear healthy? What kind of symptoms do the declining trees have?  Since the aim of the study was to understand the role of endophytic microorganisms in the mortality of Korean fir trees, it would have been more beneficial to also sample healthy trees as well as trees suffering from the decline, but not dead. If there are pathogenic organisms involved in the decline they are likely not detected anymore in dead trees; instead you find more saprophytes who require dead material (as was reported in this study). The authors should be careful when drawing conclusions about microbes enriched in the dead trees having an effect on tree health. They may simply be saprotrophs colonizing the already dead material.

Answer: Since Korean fir trees are coniferous tree species, living trees maintain vivid green leaves regardless of seasons, whereas dead trees exhibit a standing dead appearance showing that none of the leaves were attached to the branches of trees. To date, the causal agents of the declining Korean fir trees are largely unknown. Therefore, we determined the health conditions of trees based on the apparent conditions of leaves on the branches. For living trees, we chose three trees that have vivid green leaves on all branches of trees. To collect dead tree samples, we carefully chose likely dying trees that have both greenish and yellowish leaves on 10% of branches whereas the remaining 90% of the branches have no leaves. We believe that these dead trees can represent the declining trees and be suitable for our study, since they might be almost dead but not completely dead trees. To explain the details of how we chose live and dead trees, we intensively rewrote the ‘2.1. Study site and tree sampling’ section: “Since the causal agents of the declining Korean fir trees are largely unknown, we determined the health conditions of trees based on the apparent conditions of leaves on the branches. We chose three live trees that had vivid greenish leaves on all branches of trees. For the selection of dead trees, we carefully chose three dying trees that had both greenish and yellowish leaves on 10 % of branches whereas the remaining 90% of the branches had no leaves.” (lines 96-101)

In the present study, we believe that it is very difficult to link any particular microbial taxa from dead trees to the observed health conditions of Korean fir trees, since the microbial causal agent of declining Korean fir trees has not been studied well. The main goal of this paper is to understand the composition of endophytic microbial compositions depending on the health condition of trees and predict their potential roles in tree fitness using bioinformatic analyses. We toned down sentences in the introduction section that may mislead that our aim: “In this study, we analyzed the composition of bacterial and fungal endophytes of live and dead Korean fir trees located in the same Mt. Jiri habitat to understand how the health conditions of trees affect endophyte communities and predict their ecological roles.” (lines 75-77). “Our findings would be useful for understanding the potential role of endophytic microorganisms in the endangered Korean fir trees and improving the sustainability of trees from ecologically vulnerable environments.” (lines 87-89)

We agree with the reviewer that microbes preferentially enriched in the dead trees are likely involved in decomposing of dead trees for nutrient recycling instead of pathogenic agents. In the abstract and discussion section, we carefully addressed this possibility several times: “The ecological function of endophytes in dead trees was predicted to be involved in the decomposition of wood for nutrient recycling.” (lines 24-25), “Among them, Actinobacteria, Thermoleophilia, Bacteroidia, Polyangia, and Verrucomicrobiae were previously reported to be abundant taxa in decomposing dead trees, indicating that the decomposition of dead Korean fir trees is likely to have proceeded.“ (lines 327-330). “In dead trees, the enrichment of endophytic bacterial functional pathways involved in nitrogen degradation was consistent with previous studies suggesting that the decomposition of wood can act as a significant nitrogen reservoir in forest nitrogen cycling.” (lines 345-347) “In dead Korean fir trees, only Xenopolyscytalum predicted to be saprotrophic fungi was significantly abundant ASV (Figure 6B), suggesting that Xenopolyscytalum might be important for the decomposition of dead Korean Fir trees for nutrient cycling in the forest ecosystem.” (lines 362-365).

 

In addition, sampling is not well described. The sampling was done by collecting saw dust by drilling the tree trunks, but it is not mentioned if the samples included bark, cambium, sapwood and/or heartwood, which likely host different endophyte populations. Was the bark removed first? How was it ensured that the samples did not include epiphytes? Is it possible that the lichen forming fungi detected in this study originate from epiphytic growth on the trunks of the living trees?

Answer: To comply with the reviewer’s comments on sampling, we explained the details of sawdust samples in the ‘2.1. Study site and tree sampling’ section as follows: “To extract endophytic microbial DNA from the Korean fir trees, we collected sawdust samples from three live (A1-A3) and three dead (D1-D3) trees by drilling the 6 cm of four holes on the trunks at four different directional spots with a clean appearance. The sawdust samples were mainly composed of sapwood and trace amounts of cambium and bark.” (lines 101-105).

We think that the majority of lichenized fungi in our analysis is likely to be originated from endophyte since the main component of sawdust samples was sapwood and we collected sawdust samples by drilling four spots on trunks with a clean appearance (without any contamination and visible lichen). However, we agree with the reviewer’s opinion that some lichenized fungi detected from our analysis might be originated from epiphyte, since our sawdust samples include the bark of trees. In the revised manuscript, we addressed the possibility that some lichenized fungi might be originated from epiphyte (lines 360-362): “Given that sawdust samples from trees include the bark of trees, some lichenized fungi in our metagenomic analysis can be originated from epiphyte.”

 

In my opinion, the interpretation of the results should be reconsidered/revised according to the points raised above. Some additional comments are also included in the pdf.

Answer: Thank you for your comments on our manuscript. We carefully revised the manuscript according to your comments.

 

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The text has been improved according the suggestions. The methodological setup with the analysis of only three samples for each  group still remain a limiting factor which in my opinion gives little robustness to the results, especially in consideration that the samples are taken in natural conditions and not in experimental/controlled conditions.

Author Response

Thank you for your comments and encouragement to revise our manuscript. We agree with the reviewer’s concern that a small sample number limits the robustness of the analysis. To overcome the limited number of samples, we carefully analyzed metagenomic data and performed statistical analyses to reveal how the health conditions of trees affect microbial communities. Since the metagenomic analysis of endophytes and epiphytes in declining trees in natural conditions has not been studied intensively, we believe that our manuscript can provide novel information to readers who are interested in how the health of plants affects the structure of plant endophyte and epiphyte communities.

Reviewer 2 Report

I have reviewed this manuscript before, and feel that the authors have made good improvements in this revised version.

The aims are more clearly defined to investigate the effect of tree health condition to the endophyte community. Also the tree selection and sampling are much better described, and conclusions refined. It is good that also the composition of wood samples is now available, so reader can have an idea where the endophytes are detected.

As a minor point, the authors may still emphasize that since the sampling included only very small relative amounts (how small?) of bark, the authors believe that majority of detected microbes are endophytes. However epiphytes are not totally excluded (also other than lichen fungi may be epiphyte “contaminations” in the samples).

Author Response

Thank you for your comments and encouragement to revise our manuscript. It is difficult to quantify relative amounts of bark in sapwood samples. We agree with your opinion that some microbial taxa from our metagenomic analyses are originated from epiphytes. To comply with your comment, we have rewritten the manuscript and changed the title to indicate that our metagenomic data analysis includes “epiphyte” as well as “endophyte”.