Sea Buckthorn Hippophae rhamnoides and Fruit Flies Rhagoletis batava: Search for Volatile Semiochemicals Involved in Pest Attraction

Round 1

Reviewer 1 Report

The manuscript entitled „Sea buckthorn Hippophae rhamnoides and fruit flies Rhagoletis batava: search for volatile semiochemicals involved in pest attraction” describes results of behavioural studies and GC-EAD analyses, which were performed to better understand the process of host plant location by R. batava flies. These outcomes could represent a significant contribution to the field of interest. The topic is important, as the sea buckthorn, according to its pro-health properties, is gathering more and more attention by both consumers and producers, especially organic ones. R. batava on the other hand, is an aggressive and economically important pest of sea buckthorn. Currently there is no effective pest control strategy for this insect. Thus, this manuscript seemed very promising.

 

First of all, I have no comments on the introduction section, which contains sufficient information. and is a good foundation for the whole manuscript. Unfortunately, the presented research flow has some gaps.

According to the behavioural tests:

• the Authors excluded from the analysis the flies, which did not make a choice, but it is important to know how many flies did not make a choice. If you test e.g. 50 flies and 25 are not making a choice… the experimental set-up should be modified. Or only 5 did not make a choice? It is very important question! Even if the ratio is similar to the variant of fly behaviour in an empty olfactometer, if the ratio is still very high it strongly suggests that the experiment set-up is not correct.
• On average 20 individuals could be not enough to make valuable conclusions…
• The results of the behavioural tests may be different when the batava flies would not be fed before the experiment.
• Figure 1 in Results section suggests that different experimental variants were tested for males and females, why? Why males were tested only for choice between fruit (in one from three selected ripening stages) versus air? Why the Authors didn’t compare the choices between different kind of fruits, like in the female case? And, on the other hand, why the females were not tested for choice between fruit (in all three ripening stages, not only the unripe fruits) and air? Both males and females should be first tested for preferences toward three fruit stages when compared to pure air. What were the female preferences towards semi ripe or ripe fruits, when compared to air? How about testing the female preferences between ripe and unripe fruits - as they were so interested in unripe fruits (in comparison to air)?

According to the GC and GC-EAD analyses:

• Why the Authors select only unripe and ripe fruits for the analyses? The results of behavioural tests suggested that the female preferr semi-ripe fruits over the other kind of fruits. The GC and GC-EAD analyses seems then uncomplete…
• How about the information about compounds emitted by fruit but not triggering fly response? It could be also interesting.
• How many repetitions for GC analysis for each fruit type were performed?

According to the discussion:

• The conclusions from the manuscript are not fully confirmed by the results:
  • the Authors wrote that “For males, only ripe fruits are attractive, while females prefer green soft fruits that are not ripe yet, but already at an early stage of ripening” (lines 302-304). Where are the results for female and the semi-ripe and ripe fruits with air? From the viewer perspective You checked only the unripe fruits for females. They preferred them from air, but we have no knowledge about semi-ripe and ripe fruits. Thus, this statement is not fully correct.
  • the Authors wrote that “In contrast to males, ripe fruits were not attractive to females” (lines 304-305). You can not compare results from “fruit-to-air” variant and the results of choice between two kinds of fruits!

Table 2 could be moved to supplementary material.

In general, the design of the study and the presented results did not fully investigate the comprehensive process of host plant localization by R. batava flies. The manuscript needs clarification, evaluation of additional experimental variants.

Author Response

Responses to the first reviewer’s comments

Authors. We thank the reviewer for his valuable comments. Please, find our response point by point.

Reviewer. •      the Authors excluded from the analysis the flies, which did not make a choice, but it is important to know how many flies did not make a choice. If you test e.g. 50 flies and 25 are not making a choice… the experimental set-up should be modified. Or only 5 did not make a choice? It is very important question! Even if the ratio is similar to the variant of fly behaviour in an empty olfactometer, if the ratio is still very high it strongly suggests that the experiment set-up is not correct.

Authors. There were a few flies only, which did not make a choice (2 flies of 166 tested) and the number was additionally supplied in Fig 1.

Reviewer. •      On average 20 individuals could be not enough to make valuable conclusions…

Authors. The conclusions were based on statistical analysis (chi-square test, p value 0.05 or less), and if the result was considered significant, there was no sense to increase the number of tested flies.

Reviewer. •      The results of the behavioural tests may be different when the batava flies would not be fed before the experiment.

Authors. In behavioural tests, 4-15 days old flies were used. We aimed to conduct the study on insects maintained under conditions close to natural ones. Adult flies feed in nature.

Reviewer. •      Figure 1 in Results section suggests that different experimental variants were tested for males and females, why? Why males were tested only for choice between fruit (in one from three selected ripening stages) versus air? Why the Authors didn’t compare the choices between different kind of fruits, like in the female case? And, on the other hand, why the females were not tested for choice between fruit (in all three ripening stages, not only the unripe fruits) and air? Both males and females should be first tested for preferences toward three fruit stages when compared to pure air. What were the female preferences towards semi ripe or ripe fruits, when compared to air? How about testing the female preferences between ripe and unripe fruits - as they were so interested in unripe fruits (in comparison to air)?

Authors. First, our goal was to clarify which fruit ripening stage is attractive to males and females. The unripe but already soft fruits were attractive to females, so it was further tested which of the three stages (unripe, semi-ripe, and ripe) is preferred by females. The males were attracted to ripe fruits only, and not to unripe or semi-ripe. Thus, we did not see any point to test the preference of attractive and non-attractive stages of the fruits. Since unripe and ripe fruits appear at a different time of the season, a simultaneous test of fresh intact fruits is hardly possible.

 Reviewer. According to the GC and GC-EAD analyses:

• Why the Authors select only unripe and ripe fruits for the analyses? The results of behavioural tests suggested that the female preferr semi-ripe fruits over the other kind of fruits. The GC and GC-EAD analyses seems then uncomplete…

Authors. For GC-EAD analysis, we selected the extreme stages of fruit ripening: unripe and ripe. The study revealed little difference in VOC quality of ripe and unripe fruits (a single compound only). The differences were quantitative only. We assume that the intermediate stage of fruit ripening should not produce more pronounced differences.

Reviewer. •      How about the information about compounds emitted by fruit but not triggering fly response? It could be also interesting.

Authors. The compounds emitted by fruit but not triggering fly response were presented in detail in the Results subsection 3.3.

Reviewer. •      How many repetitions for GC analysis for each fruit type were performed?

Authors. Sampling and GC analysis of volatiles from unripe and ripe fruits were carried out in triplicate. This information was presented, see lines 187-188.

Reviewer. According to the discussion:

• The conclusions from the manuscript are not fully confirmed by the results:
• the Authors wrote that “For males, only ripe fruits are attractive, while females prefer green soft fruits that are not ripe yet, but already at an early stage of ripening” (lines 302-304). Where are the results for female and the semi-ripe and ripe fruits with air? From the viewer perspective You checked only the unripe fruits for females. They preferred them from air, but we have no knowledge about semi-ripe and ripe fruits. Thus, this statement is not fully correct.

Authors. The statement was corrected: “For males, only ripe fruits are attractive, while females prefer green soft fruits that are not ripe yet.” (lines 314-315).

Reviewer. •      the Authors wrote that “In contrast to males, ripe fruits were not attractive to females” (lines 304-305). You can not compare results from “fruit-to-air” variant and the results of choice between two kinds of fruits!

Authors. The statement was corrected to “In contrast to males that were attracted to ripe fruits, females preferred semi-ripe to ripe fruits.” (lines 316-317)

Reviewer. Table 2 could be moved to supplementary material.

Authors. Table 2 moved to supplementary material as the reviewer suggested.

Reviewer. In general, the design of the study and the presented results did not fully investigate the comprehensive process of host plant localization by R. batava flies. The manuscript needs clarification, evaluation of additional experimental variants.

Authors. We agree with the reviewer that our study did not fully investigate the comprehensive process of host plant localization by R. batava flies. The aim of our study was much narrower: to search for volatile semiochemicals involved in pest attraction (the title of the paper). As no information has been available so far on the compounds involved in the interaction between R. batava and sea buckthorn, we identified a number of VOCs perceived by the flies. This opens new opportunities for further research.

Reviewer 2 Report

The paper is interesting and well written despite the fact that the same authors have recently published relatively similar papers (refs. 14, 29, 30).

However, there is an apparent inconsistency between Tables 1 and 2 due to the different techniques used to identify the volatile compounds in the fruit. In fact, Table 1 includes 27 different compounds, all of which were tested for EAG activity, whereas Table 2 lists 76. therefore, the 49 volatile compounds produced by Hippophae rhamnoides fruits not included in Table 1 were evaluated or not for EAG activity?

Figure 2 would indicate a no as answer, as only about forty peaks are represented in Figure 1 while compound 27 of Table 1 corresponds to compound 71 in Table 2.

In addition: a) Table 1 shows the retention index and Table 2 the retention time; b) the ratio in terms of percentage area between the different volatile compounds does not correspond between the two tables either between compounds or in the trend of some compounds between unripe and ripe fruit.

For example, 3-Methylbutyl 3-methylbutanoate and Ethyl benzoate are present in similar peak area % in ripe fruit in Table 1, but in Table 2 (although the values are expressed as ng) ripe fruit the former compound is present at three times the level of Ethyl benzoate ( 232.16 vs. 79. 04); Ethyl hexanoate is by far the majority compound in table 1 for ripe fruit (peak area %) about eight times more than3-metilbutyl 3-methylbutanoate, but in table 2 the ratio drops to 1.5; concerning the trend between uripe and ripe fruit Ethyl heptanoate decreases in terms of peak area % in Table 1 (from 1.34 to 0.49 %), but as % of total weight of metabolites it increases from about 0.25 to 0.6 % (respectively 1.25 : 485.2 x 100, 16.67 : 2588.07 x100). And even if there is some error in calculation or inappropriate comparison, the problem is that two tables obtained with different methodologies cannot coexist unless they are fully explained.

It is necessary to revise the MS.

Author Response

Responses to the second reviewer’s comments

Authors. We thank the reviewer for the valuable comments. Please, find our explanations point by point.

Reviewer. However, there is an apparent inconsistency between Tables 1 and 2 due to the different techniques used to identify the volatile compounds in the fruit. In fact, Table 1 includes 27 different compounds, all of which were tested for EAG activity, whereas Table 2 lists 76. therefore, the 49 volatile compounds produced by Hippophae rhamnoides fruits not included in Table 1 were evaluated or not for EAG activity?

Authors. We agree with the reviewer that there is an inconsistency between Tables 1 and 2 due to the different techniques used to identify the volatile compounds. The sampling of volatiles for the determination of olfactory active compounds was aimed to collect a sufficient amount of volatiles for multiple GC-EAD runs as well as to make stock solutions ensuring as little variation as possible between GC-EAD replicates. Due to the long sampling intervals, the ratio between components was less accurate compared to the sampling of fruit-released volatiles aiming to as accurately as possible quantify the volatiles.

In table 1, only EAD active volatiles, i.e. 27 compounds, are shown, while table 2 reports all 76 volatiles sampled from the berries including EAD active compounds. 49 volatile compounds produced by Hippophae rhamnoides fruits did not elicit EAD responses (they were tested for EAD activity), therefore, these compounds were not included in table 1. To solve the inconsistency between Tables 1 and 2 we moved the data shown in table 2 to the supplementary as suggested by reviewer 1.

Reviewer. Figure 2 would indicate a no as answer, as only about forty peaks are represented in Figure 1 while compound 27 of Table 1 corresponds to compound 71 in Table 2?

Authors. That’s right, compound 27 in table 1 has number 71 in table 2. In figure 2, only the part of GC-EAD record is shown representing all EAD active compounds. The last 12 minutes of recording after elution of 3-methylbutyl benzoate are not shown aiming to have a clearer representation of GC-EAD active compounds. We have moved the data shown in table 2 to the appendix. Moreover, in the legend of figure 2, we have pointed out that only EAD active peaks are numbered according to Table 1. We hope the representation of the data became clearer.

Reviewer. In addition: a) Table 1 shows the retention index and Table 2 the retention time; b) the ratio in terms of percentage area between the different volatile compounds does not correspond between the two tables either between compounds or in the trend of some compounds between unripe and ripe fruit.

For example, 3-Methylbutyl 3-methylbutanoate and Ethyl benzoate are present in similar peak area % in ripe fruit in Table 1, but in Table 2 (although the values are expressed as ng) ripe fruit the former compound is present at three times the level of Ethyl benzoate ( 232.16 vs. 79. 04); Ethyl hexanoate is by far the majority compound in table 1 for ripe fruit (peak area %) about eight times more than3-metilbutyl 3-methylbutanoate, but in table 2 the ratio drops to 1.5; concerning the trend between uripe and ripe fruit Ethyl heptanoate decreases in terms of peak area % in Table 1 (from 1.34 to 0.49 %), but as % of total weight of metabolites it increases from about 0.25 to 0.6 % (respectively 1.25 : 485.2 x 100, 16.67 : 2588.07 x100). And even if there is some error in calculation or inappropriate comparison, the problem is that two tables obtained with different methodologies cannot coexist unless they are fully explained.

Authors. a) in table 1, we have added RT column in addition to RI; b) data presented in tables 1 and 2 (at present S1) were obtained by different methods which were optimal to serve the different aims of the experiments. In addition, the sampling of fruit-released volatiles for quantification lasted for 2 h in comparison to 24 h sampling of the volatiles for GC-EAD analysis. Moreover, the total amount of volatiles sampled from unripe versus ripe fruits was around 6 times lower, therefore, data for the number of volatiles obtained by two different sampling methods differ. We have explained the aims of each method used.

Round 2

Reviewer 1 Report

The manuscript was improved and now could be published.

Author Response

We thank the reviewer for the positive evaluation of our manuscript.

Reviewer 2 Report

The answers to the criticisms appear satisfactory, but it is a pity that the text lacks an adequate explanation (understandable to readers) of what was stated. The result is a worse version of the MS than the previous one. For this reason, the judgement can only be negative.
For future acceptance, the MS must be completely reorganised and rewritten because it is based on the use of two different techniques for separating volatile compounds. 
A correct experimental design should have included the use of the best separation technique right from the start and then go on to assess whether the individual (major) volatile compounds give an electroantennographic activity.

In fact, the GC in Figure 1 together with Table 1 seems to indicate the presence of a limited number of volatile compounds (supported by the % peak indication) which is then contradicted by the actual Table S1 without the discrepancy being discussed in the text.

In addition, Table 1 and Table S1 differ in the relative ratios of the different volatile compounds, and the retention index is absent from Table S1, in addition to the fact that the different numbering of the compounds makes it difficult to understand and compare.

Finally, Figure 3 refers to table S1 without any connection with table 1, much more relevant for the purposes of the work, and again without emphasizing the fact that only some of the 79 compounds have been subjected to analysis of electroantennographic activity. In principle, who guarantees that any of the GC peaks in Figure 1 do not represent two overlapping peaks? For example, ethyl 3-idroxybutanoate is present in Table 1 (no. 23) both in immature and ripe fruits and is active on both males and females insect antennas but is absent in table S1 for ripe fruits (no.42). 

Author Response

Authors. We thank the reviewer for the comments. We have corrected the manuscript according to the suggestions. Please, find our explanations.

Reviewer. The answers to the criticisms appear satisfactory, but it is a pity that the text lacks an adequate explanation (understandable to readers) of what was stated. The result is a worse version of the MS than the previous one. For this reason, the judgement can only be negative.

For future acceptance, the MS must be completely reorganised and rewritten because it is based on the use of two different techniques for separating volatile compounds.

A correct experimental design should have included the use of the best separation technique right from the start and then go on to assess whether the individual (major) volatile compounds give an electroantennographic activity.

Authors. In the discussion, see lines 550-559, we have added the explanation regarding differences in the data due to the use of two different methods.

We have reorganised the manuscript and did appropriate changes in the text starting with a complete analysis of fruit volatiles, followed by GC-EAD tests and identification of olfactory active compounds as it was recommended by the reviewer.

Reviewer. In fact, the GC in Figure 1 together with Table 1 seems to indicate the presence of a limited number of volatile compounds (supported by the % peak indication) which is then contradicted by the actual Table S1 without the discrepancy being discussed in the text

Authors. We have reorganized the presentation of the data and applied the same numbering of the compounds in both tables, which we hope will make a reader easier to go through the data. We have added a paragraph in the discussion explaining the differences in data collected by two methods. Indeed, Table 1 (numbering used in the previous manuscript version) is presenting only EAD active compounds and in Fig 2 only EAD active compounds are pointed out. Other, not EAD active volatiles are represented as not labelled peaks. After the reorganization of the manuscript, this becomes easier to see.

Reviewer. Finally, Figure 3 refers to table S1 without any connection with table 1, much more relevant for the purposes of the work, and again without emphasizing the fact that only some of the 79 compounds have been subjected to analysis of electroantennographic activity. In principle, who guarantees that any of the GC peaks in Figure 1 do not represent two overlapping peaks? For example, ethyl 3-idroxybutanoate is present in Table 1 (no. 23) both in immature and ripe fruits and is active on both males and females insect antennas but is absent in table S1 for ripe fruits (no.42). 

Authors. Yes, figure 3 refers to table S1 (numbering used in the previous manuscript version, while Fig 2 and Table 1 in the current version). All volatiles including EAD not active compounds have been attributed to 8 groups based on functional groups. EAD active compounds fall into 3 chemical groups and at a large extent were dominated by esters. We discussed the chemodiversity of EAD active compounds in the text as in our opinion figure will not provide more visualization having in mind 3 chemical groups of the EAD active volatiles.

All 79 compounds have been subjected to analysis of electroantennographic activity. Indeed, it is difficult to see small peaks in the figure but larger not numbered peaks in the figure indicate tested but not EAD active peaks.

No, we did not detect chromatographically overlapping EAD active peaks. There are few not completely separated peaks like 21 and 22 peaks but EAD responses show that there are two compounds near each other as well as mass spectra analysis of the sample confirm two compounds. Usually, mass spectrometric analysis enables the detection of overlapping peaks if their spectra differ qualitatively by one or more ion fragments and coeluting compounds give enough strong signal. Most certainty would be achieved using two-dimensional separation, however, we did not have this equipment at our research facility.

Absence of certain compounds in table S1 versus table 1 we have already discussed.

Round 3

Reviewer 2 Report

No comments. The authors have responded satisfactorily to previous criticisms.