Aquatic Hemiptera in Southwest Cameroon: Biodiversity of Potential Reservoirs of Mycobacterium ulcerans and Multiple Wolbachia Sequence Types Revealed by Metagenomics

Round 1

Reviewer 1 Report

After the new changes, I consider the ms could be accepted for publication in present form

Author Response

Thank you.

Reviewer 2 Report

Since this is a review of resubmission, I have nothing to add to my first review comments.  I recommend the publication of this paper.

Author Response

Thank you.

Reviewer 3 Report

This original research article presents the findings of an intensive molecular study intended to the identification of aquatic insects (Orden Hemiptera) in areas suspected to be endemic of Buruli ulcer (BU). Authors have collected a relatively large number of insects in water bodies from various communities thought to be mesoendemic and hypopendemic for BU. Later on, they proceeded to identify the collected specimens according to their morphology and also using molecular methods. They also explored the potential presence of Mycobacterium spp and mycolactone-producing mycobacteria in a subsample of collected insects (belonging to the families Gerridae and Veliidae). Finally, they complemented the aforementioned analysis with a phylogenetic study of the Wolbachia symbiont within the selected Gerridae and Veliidae specimens. Despite the extensive and intensive molecular analysis conducted and the relative ecological interest of a phylogenetic study of aquatic Hemiptera specimens, the whole work lacks epidemiological interest and necessary rigorousness to conclude anything concerning the potential role of aquatic hemipteran insects as vector or reservoirs of Mycobacterium ulcerans in BU endemic areas of Southwest Cameroon. I found these major reasons for this assertion.

The study area for this study has been identified as potentially endemic for BU based on reported BU cases which have been clinically confirmed across Cameroon between 2001 and 2014 (Tabah et al. 2016). Looking in more detail at this long-trend study, we can see that PCR confirmation rate was never above 57% (this achieved in 2010) and reaching the lower positive rate in 2014 with 20% of BU clinically suspected cases. Authors should present more conclusive proof of the actual presence of BU in the area (endemicity status). From the number of recorded BU cases in the Southwest region, how many were confirmed by PCR? In absence of PCR confirmation we cannot conclude that ulcerative skin lesions are actually caused by a mycolactone-producing bacteria. We cannot conclude that surveyed communities are actually located in mesoendemic or hypoendemic BU areas. Therefore, not having found many Hemiptera specimens (only 1) with a trace of M ulcerans DNA it is inconclusive of their role as potential vector or reservoirs. However, authors concluded that “aquatic Hemiptera may have a minor role (if any) in the spread of BU in Southwest Cameroon.”

 

It is unclear whether the authors sampled water bodies from communities reporting BU cases. They referred to health districts which have been classified as mesoendemic and hypoendemic but have not specified whether sampled water bodies were selected from communities, or surroundings, reporting BU cases. BU has proven to be quite focal so that human cases may come from restricted areas (i.e. a small group of communities) within geographical areas (districts) classified as endemic.  

 

Surprisingly, authors have decided to leave out from the mycobacterial DNA analysis the specimens which have been recognised as biting insects (i.e. Belostomatidae, Naucoridae, Corixidae), and therefore stronger candidates to play some role on the transmission of BU in endemic areas. Neither Gerridae nor Veliidae species are considered biting insects, despite some Gerridae spp have been found harbouring ulcerans.

 

Authors hypothesize about the potential interaction of Wolbachia symbiont and ulcerans, the former may act as either a facilitator or competitor. This role is far from being elucidated and by no means, this work is addressing this issue. However, authors have proven the presence of this symbiont in the collected Hemiptera specimens, which is not surprising considering that this symbiont has been seen to be present in Hemiptera insects, and even in high rate, among other orders such as Coleoptera, Diptera, and Odonata in previous and more systematic ecological studies.

Sazama EJ, Bosch MJ, Shouldis CS, Ouellette SP, Wesner JS. Incidence of Wolbachia in aquatic insects. Ecol Evol. 2017;7(4):1165–1169. Published 2017 Jan 24. doi:10.1002/ece3.2742

 

Minor comments

Line 48. Reference 10. I do not think this reference is rather appropriate to provide some insights into the geographical distribution of BU. I suggest the authors use a more comprehensive and updated. For instance, “Mapping the global distribution of Buruli ulcer: a systematic review with evidence consensus. Simpson, Hope et al. 2019. The Lancet Global Health, Volume 7, Issue 7, e912 - e922.” Line 59. Double full stop. Revise some typos throughout the draft.

Author Response

-This original research article presents the findings of an intensive molecular study intended to the identification of aquatic insects (Orden Hemiptera) in areas suspected to be endemic of Buruli ulcer (BU). Authors have collected a relatively large number of insects in water bodies from various communities thought to be mesoendemic and hypopendemic for BU. Later on, they proceeded to identify the collected specimens according to their morphology and also using molecular methods. They also explored the potential presence of Mycobacterium spp and mycolactone-producing mycobacteria in a subsample of collected insects (belonging to the families Gerridae and Veliidae). Finally, they complemented the aforementioned analysis with a phylogenetic study of the Wolbachia symbiont within the selected Gerridae and Veliidae specimens. Despite the extensive and intensive molecular analysis conducted and the relative ecological interest of a phylogenetic study of aquatic Hemiptera specimens, the whole work lacks epidemiological interest and necessary rigorousness to conclude anything concerning the potential role of aquatic hemipteran insects as vector or reservoirs of Mycobacterium ulcerans in BU endemic areas of Southwest Cameroon. I found these major reasons for this assertion.

-The study area for this study has been identified as potentially endemic for BU based on reported BU cases which have been clinically confirmed across Cameroon between 2001 and 2014 (Tabah et al. 2016). Looking in more detail at this long-trend study, we can see that PCR confirmation rate was never above 57% (this achieved in 2010) and reaching the lower positive rate in 2014 with 20% of BU clinically suspected cases. Authors should present more conclusive proof of the actual presence of BU in the area (endemicity status). From the number of recorded BU cases in the Southwest region, how many were confirmed by PCR? In absence of PCR confirmation we cannot conclude that ulcerative skin lesions are actually caused by a mycolactone-producing bacteria. We cannot conclude that surveyed communities are actually located in mesoendemic or hypoendemic BU areas. Therefore, not having found many Hemiptera specimens (only 1) with a trace of M ulcerans DNA it is inconclusive of their role as potential vector or reservoirs. However, authors concluded that “aquatic Hemiptera may have a minor role (if any) in the spread of BU in Southwest Cameroon.”

Our response: We were not involved in the study of Tabah et al. (2016), but it is clear from reading this work that they did not sample any site for Buruli ulcer (BU) nor M. ulcerans presence. Their study analyzed existing data on BU from 2001 to 2014. The data were collated and activity reports of the National Buruli ulcer Control Programme (NBUCP) were reviewed. In their retrospective study, case-detection rates and key Buruli ulcer control indicators were computed and plotted on a time scale to elucidate trends in the performance of the NBUCP. Their objective was to evaluate the NBUCP and identify areas for improvements. By WHO classification, Cameroon is one of the 16 countries in sub-Saharan African endemic for Buruli ulcer (BU). Within Cameroon, many BU endemic foci have been identified. Within the context of the National Buruli ulcer Control Programme (NBUCP), the six health districts included in our study are referred to as "endemic foci". However, in our study we suggested the terms 'hypoendemic' and 'mesoendemic' to highlight the relatively low numbers of BU cases reported from these Health Districts between 2001 and 2014 (compared with BU cases in other endemic foci). Buruli Ulcer Detection and Treatment Centres "BU-DTCs" have been created in major endemic foci in Cameroon and one of these BU-DTCs is in Mbonge Health District in the South West region of Cameroon, where some of our sampling of aquatic bugs took place. All BU-DTCs are provided with well-trained personnel, geographic-area specific case-definitions from WHO and diagnostic guidelines for the clinical diagnosis of BU cases. The drop in the PCR confirmation rate of BU cases to 57% (in 2010) and even lower to 20% (in 2014) has been solely attributed to the dwindling of external partner funding on which the NBUCP relied. Before 2010, the confirmation of BU cases by PCR was well above 57% (the WHO target is 70%) (Tabah et al., 2016).

-It is unclear whether the authors sampled water bodies from communities reporting BU cases. They referred to health districts which have been classified as mesoendemic and hypoendemic but have not specified whether sampled water bodies were selected from communities, or surroundings, reporting BU cases. BU has proven to be quite focal so that human cases may come from restricted areas (i.e. a small group of communities) within geographical areas (districts) classified as endemic.  

Our response: All water bodies sampled in our study were from communities that reported BU cases from 2001 to 2014.

-Surprisingly, authors have decided to leave out from the mycobacterial DNA analysis the specimens which have been recognised as biting insects (i.e. Belostomatidae, Naucoridae, Corixidae), and therefore stronger candidates to play some role on the transmission of BU in endemic areas. Neither Gerridae nor Veliidae species are considered biting insects, despite some Gerridae spp have been found harbouring ulcerans.

Our response: We have discussed the decision to concentrate on Gerridae and Veliidae in our study already (lines 481 – 500, Discussion), as well as the biases arising in the prior literature due to an unjustified focus on biting bugs. However, we have now screened the biting bugs from our study with the IS2404 PCR assay and none was found to be positive (lines 317 - 320). We have updated the manuscript accordingly and note this reinforces our original conclusion that “almost no evidence for a role for these insects as an environmental reservoir of the pathogen was uncovered”.

-Authors hypothesize about the potential interaction of Wolbachia symbiont and ulcerans, the former may act as either a facilitator or competitor. This role is far from being elucidated and by no means, this work is addressing this issue. However, authors have proven the presence of this symbiont in the collected Hemiptera specimens, which is not surprising considering that this symbiont has been seen to be present in Hemiptera insects, and even in high rate, among other orders such as Coleoptera, Diptera, and Odonata in previous and more systematic ecological studies.

Sazama EJ, Bosch MJ, Shouldis CS, Ouellette SP, Wesner JS. Incidence of Wolbachia in aquatic insects. Ecol Evol. 2017;7(4):1165–1169. Published 2017 Jan 24. doi:10.1002/ece3.2742

Our response: We are aware of the Sazama study and indeed, we cite it. That paper contains no information for aquatic insects from Africa and some taxa are very poorly represented worldwide (e.g., a grand total of two veliid specimens are included in the screening!). Even in light of the estimate of 50% of aquatic insects being infected with Wolbachia, our results are quite surprising and additional novelty is provided by our analysis showing the diversity of symbiont strains in the veliids and gerrids.

-Minor comments

Line 48. Reference 10. I do not think this reference is rather appropriate to provide some insights into the geographical distribution of BU. I suggest the authors use a more comprehensive and updated. For instance, “Mapping the global distribution of Buruli ulcer: a systematic review with evidence consensus. Simpson, Hope et al. 2019. The Lancet Global Health, Volume 7, Issue 7, e912 - e922.” Line 59. Double full stop. Revise some typos throughout the draft. 

Our response: We have included this new reference and corrected the typos.

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

Line 42: Arguably, there are several papers that document fairly well what the incubation period

 

Line 54: limnic not a commonly used term

 

Line 56, 69: more careful review of literature, citations not appropriate for statements being made.

Paragraph info not always congruent with main focus of paragraph ex. Lines 53-71. This paragraph is all over the map and the sequence of thought does not flow. For example, line 55, the authors state that the mode of transmission is controversial and explain why very well, but then return to the to the possibility of them being vectors at the end of the paragraph and into the next paragraph at line 72.

 

Lines 84-92: Bringing in Wolbachia into this conversation is distracting and a stretch as a reservoir

or method of disseminating M. ulcerans.

 

Methods

The prose describing the sample sites do not match up with the figures. There are several streams and a few that look fast-flowing yet these are not mentioned.

 

Lines 116-127: This paragraph is poorly written and should be revised, it is wordy and should be more concise.

 

Figure numbers unclear. What does Figure S1? Mean, why S?

Figure S1A-D: These figures do not indicate much about the sampling points, these points should be described in the caption or text. There are 2-4 panels per figure but the panels are not described.

 

Lines 142-157: This paragraph is poorly written and oddly phrased. Highly recommend a major rewrite. Much is unclear, e.g., how many are several sweeps? Why didn’t the authors standardize these methods? For example, why not conduct a timed sweep.

 

Line 173-178: Work on wording, e.g., each insect, not bug. To which level of taxonomy were insects identified, family, genus, species? Wording is confusing. It is unclear why morphotypes are being determined for this study, why did the authors go to this extent? Was it expected that M. ulcerans presence might differ among morphotypes?

 

Lines 180-188: No attempt to determine analyze internal v. external contamination/infection of M. ulcerans, insects were cut/ground, the authors did not look at what could have been internal v. external contamination of M. ulcerans.

 

Lines 293: The Wolbachia paragraph is out of place.

 

Results

Figure captions are simply not informative nor the prose in the text that set up some of the figures. For example, in Figure 4, why illustrate multiple examples of the same genus?

Figures are redundant and at times difficult to read.
Figure axes require rewording and are at time confusing.

Table S1 confusing and what does S mean?

Figure 3 caption confusing. Figure actually shows location, not distribution.

Table 2: cannot read this table at all. It appears to be raw data and has no place in the manuscript.

Figure S2 is unclear, what is this caption saying?

Authors should either submit a paper on hemiptera associated with M. ulcerans - which veliids are not really OR a paper on wolbachia but not both. It was confusing to follow both threads in this paper. 

 

Spacing off in some of the lines, e.g., line 350.

 

Lines 321-326: What does ‘most bugs’ mean? Of the insect sampled for M. ulcerans, families Gerridae and Veliidae are both water surface dwellers that can move significant distances on the water surface, why are the major hemiptera (such as notonectids (not commonly human biters), belostomatids, naucorids and nepids involved with macroinvertebrate surveys associated with M. ulcerans ignored with regards to M. ulcerans DNA presence/absence in this study? There were 149 individuals within these individuals collected, why were they not sampled for M. ulcerans?

 

A major issue is the low sample number of specimens (n=1) testing positive for putative M. ulcerans DNA.

 

Lines 394-399: Unclear as to why these data are included, they add nothing to the M. ulcerans conversation.

 

Discussion

The first paragraph is not relevant to this paper, this study is not about transmission, it should be about diversity, so to begin this section in this manner is not a focused approach. In fact, paragraph two is a rehash of the introduction. The fact that the authors find that their data markedly contrast with other literature in Cameroon is not surprising. The authors sampled insects that are feeding on organic material either plant or terrestrial insects on the water surface. Those hemiptera found to be associated with the presence of M. ulcerans are most likely picking it up from the prey they consume under the water that either filter feed the bacterium from the water or scrape it off of biofilms under the water. Therefore, from an aquatic ecological perspective, they have selected taxa that would in fact not support prior literature – thus to conclude that their data do not agree, of course they do not, it was comparing apples to oranges.

The entire Wolbachia prose is an insert into a paper that does not flow with the reason for the study in the first place.

Author Response

Author response to Referee 1.

Line 42: Arguably, there are several papers that document fairly well what the incubation period

We have simplified the sentence to say that “The incubation period of ulcerans infection varies from weeks to several months, with a median of 4 - 5 months [5,6]”.

Line 54: limnic not a commonly used term

With respect to the referee, this is the most succinct and appropriate term to use for all bodies of freshwater, and is used quite widely (hence the term “limnic eruption” and journals such as the International Journal of Limnology and Limnology & Oceanography ).

Line 56, 69: more careful review of literature, citations not appropriate for statements being made.

For line 56, citations 23 and 24 seem appropriate for the point we are making about the routes of transmission of BU remaining unclear. However, as citation 25 does not directly address questions of transmission in aquatic environments, we have replaced it with Garchitorena et al. (2015) DOI: 10.7554/eLife.07616. In line 69, we wrote that “Indeed, it is highly unlikely they are directly responsible for the majority of BU cases” and cited two references [32,33]. One of these is a case report directly implicating a belostomatid in transmission of ulcerans via a bite, and the other is a large entomological survey finding no difference in M. ulcerans positivity in aquatic bugs between BU-endemic and non-endemic areas. We have kept the same citations, but have reworded the sentence to say, “Thus, while it is possible that bites from aquatic bugs could transmit M. ulcerans to humans [32], there is no evidence that this is epidemiological relevant on a wide scale [33]”.

Paragraph info not always congruent with main focus of paragraph ex. Lines 53-71. This paragraph is all over the map and the sequence of thought does not flow. For example, line 55, the authors state that the mode of transmission is controversial and explain why very well, but then return to the to the possibility of them being vectors at the end of the paragraph and into the next paragraph at line 72.

We are sorry that the referee has found some of the writing in the Introduction difficult to follow. For this particular paragraph, we have moved the text describing the origin and evolution of the pathogen into the first paragraph, so that this paragraph (originally lines 53-71) considers only the question of the transmission of ulcerans and the role of aquatic bugs. Regarding the vector “controversy”, we use the word deliberately to highlight the unresolved nature of the debate. While it is the personal view of the authors that biting bugs transmit M. ulcerans very infrequently, the case report we cite [Marion et al. (2014) doi.org/10.1016/j.ijmyco.2014.01.004] is relatively recent and follows almost 20 years of debate on the topic that our own study cannot hope to settle. However, for the sake of clarity, we have removed the term “vector” from the Introduction after our statement about the controversy, and only use the term “reservoir”.

Lines 84-92: Bringing in Wolbachia into this conversation is distracting and a stretch as a reservoir or method of disseminating M. ulcerans.

We do not bring Wolbachia into the conversation here as a discursion but because our analyses revealed extensive Wolbachia infections in the aquatic bugs, contrasting with a virtual absence of ulcerans. We are unsure what the referee means about Wolbachia being a “reservoir or method of disseminating M. ulcerans.” We do not make such a statement; rather, we say, “…interactions between M. ulcerans and Wolbachia in aquatic bugs have the potential to impact on the dissemination of the pathogen in the host and perhaps the wider environment”. In case this was unclear, we have changed this statement to say, “Therefore, M. ulcerans and Wolbachia may interact within the microbiome of aquatic bugs, either in a facilitative or competitive manner as observed with Wolbachia in other systems, leading to potential impacts on dissemination of the pathogen in the host and perhaps the wider environment.” There are numerous recent publications demonstrating that Wolbachia can have profound effects on other components of the microbiome in arthropods as diverse as fruit flies, mosquitoes and ants, and we have now cited three such studies.

Methods

The prose describing the sample sites do not match up with the figures. There are several streams and a few that look fast-flowing yet these are not mentioned.

We have added further clarification about the sampling sites to the text and figure legends. As implied in line 121, while some water bodies were slow-flowing when precipitation was low, they became flooded rapidly during heavy rain. We were sampling at the height of the wet season in a region with some of the highest levels of precipitation in the world.

Lines 116-127: This paragraph is poorly written and should be revised, it is wordy and should be more concise.

We have shortened this section, moving some of the details into the legend for Fig. S1.

Figure numbers unclear. What does Figure S1? Mean, why S?

It has come to our attention that the journal transferred supplemental figures and tables into the main text from a file intended as a separate supplement without our knowledge. I assume this was done in the belief that it would facilitate review of the manuscript, but it appears to have caused several distractions and some unfortunate confusion.

Figure S1A-D: These figures do not indicate much about the sampling points, these points should be described in the caption or text. There are 2-4 panels per figure but the panels are not described.

We have added further descriptions of the sites for these supplemental images.

Lines 142-157: This paragraph is poorly written and oddly phrased. Highly recommend a major rewrite. Much is unclear, e.g., how many are several sweeps? Why didn’t the authors standardize these methods? For example, why not conduct a timed sweep.

We apologise if this section is unclear. The methods used were similar to other studies of the role of aquatic fauna in Buruli ulcer epidemiology [papers we have cited, such as Marion et al. (2010) and Garchitorena et al. (2014)]. These other authors report the number of sweeps conducted, but not precise timings, nor the number of individuals involved in sampling. We state that four individuals sampled for 40 minutes per site. Sampling was intensive for the time spent at each site, as for financial and logistical reasons, the fieldwork was limited to less than two weeks in total and we were facing both extreme weather and an uncertain security situation following the Anglophone crisis in the region. Moreover, this was the first intensive sampling of aquatic bugs in Southwest Cameroon, so the primary goal was to obtain sufficient numbers of bugs to allow assays for ulcerans to be conducted on a wide range of species from areas experiencing different levels of endemicity. In other words, our study was not intended as a systematic ecological analysis between sites. We have now clarified the study goals in this section.

Line 173-178: Work on wording, e.g., each insect, not bug. To which level of taxonomy were insects identified, family, genus, species? Wording is confusing. It is unclear why morphotypes are being determined for this study, why did the authors go to this extent? Was it expected that M. ulcerans presence might differ among morphotypes?

We use the term “bug” as the common English word for a hemipteran insect. It is universally used in the literature for hemipterans, including in Buruli ulcer studies [see cited article titles such as Marion et al. (2010) “Seasonal and regional dynamics of M. ulcerans transmission in environmental context: Deciphering the role of water bugs as hosts and vectors.”]. We identified the bugs to the lowest taxonomic level possible, which was facilitated by prior sorting into morphotypes. As other authors have emphasised [e.g., Ebong et al. (2016], the diversity of aquatic Hemiptera in Cameroon is high and comprehensive species keys do not exist. Morphotypes do not necessarily have a 1:1 concordance with putative species (as we show in the analysis presented in Fig. 8), but molecular screening of the bugs was organised in this manner so that any M. ulcerans-positive morphotypes could have been examined in more detail at a later stage. If there is a true biological association (symbiosis) between M. ulcerans and certain aquatic bugs, which is suggested by the specific trafficking events of the bacterium within hemipteran hosts [see Marsollier et al. (2007) DOI: 10.1111/j.1462-5822.2006.00790.x], it is reasonable to assume that it might vary by morphotype. We have now clarified the approach to morphotyping in the text.

Lines 180-188: No attempt to determine analyze internal v. external contamination/infection of M. ulcerans, insects were cut/ground, the authors did not look at what could have been internal v. external contamination of M. ulcerans.

In common with every other field study of aquatic bugs in the context of BU transmission of which we are aware, we did not “surface-sterilise” or otherwise process the bugs before DNA extraction. In any case, surface sterilisation has not been found to significantly alter microbiome profiles of insects in systematic studies [see Hammer et al. (2015) DOI 10.7717/peerj.1190]. More importantly, in predatory species, gut contents are a more important source of potential contamination, but accounting for this would have necessitated mass dissection of fresh insects, which would have been completely impracticable for >700 specimens. Although we had conceded that our single putative positive might have arisen through contamination (see lines 556 – 557), we have now explained why avoiding possible contamination was not possible.

Lines 293: The Wolbachia paragraph is out of place.

Reporting of the Wolbachia-specific methods is necessary as we are arguing for retaining of these critically important data in the manuscript (see below).

Results

Figure captions are simply not informative nor the prose in the text that set up some of the figures. For example, in Figure 4, why illustrate multiple examples of the same genus?

Figure 4 shows different morphotypes of two of the most abundant genera, Limnogonus and Rhagovelia, in order to illustrate the remarkable diversity within these genera. A short explanation is added to the text.

Figures are redundant and at times difficult to read. 
Figure axes require rewording and are at time confusing.

As explained above, the journal included the supplemental figures and tables in the main text, causing some redundancy. The figures are not shown in their final resolution and Fig. 8 was unfortunately corrupted during formatting by the journal. Regarding the “confusing” figure axes, it is very difficult for us to judge what might be confusing to the referee. However, we have responded to the specific points below.

Table S1 confusing and what does S mean?

This table is from the online supplement as wasn’t intended to be shown in the main text. The heading on the far right has been changed slightly so that it is clear we are presenting the number of biting bugs as a percentage of total bugs caught in each location.

Figure 3 caption confusing. Figure actually shows location, not distribution.

The figure shows the distribution of bug families between the six Health Districts (locations). We have clarified this in a legend.

Table 2: cannot read this table at all. It appears to be raw data and has no place in the manuscript.

This is a supplemental table of raw data that was not intended to be shown in the manuscript. It was placed there by the journal editorial staff.

Figure S2 is unclear, what is this caption saying?

This is another supplemental figure. It shows a widely-used output from initial analysis of sequence data, a “blob plot”, which assigns taxonomic information to individual DNA sequences using BLAST (see Methods description in lines 255 - 258). The figure shows that this particular sample contains a lot of bacterial contigs. We have added some further information to the legend.

Authors should either submit a paper on hemiptera associated with M. ulcerans - which veliids are not really OR a paper on wolbachia but not both. It was confusing to follow both threads in this paper. 

With respect to the referee, we do not agree that it would be worthwhile or logical to separate the Wolbachia data from the Buruli ulcer study. We will return below to why these particular bugs were screened for ulcerans, but we are confident that the discovery of high levels of Wolbachia sequence diversity within only 20 sequenced samples will be of considerable interest to the Wolbachia field. This is especially true in these understudied aquatic taxa, as almost all prior Wolbachia research has focused on terrestrial arthropods or species in which only the larval stage is aquatic. Separation of the study into two manuscripts would involve multiple redundancies, and the sampling strategy leading to the discovery of the Wolbachia only makes sense in light of the goals of the Buruli ulcer study. Moreover, we remind the referee that we submitted this manuscript to a special issue of Diversity on “Insect Symbiosis”, and our Wolbachia data are a perfect fit with this theme.

Spacing off in some of the lines, e.g., line 350.

There are multiple formatting issues with the manuscript, several of which were introduced by the journal before it was sent for review. We apologise for these distractions, which of course will be corrected.

Lines 321-326: What does ‘most bugs’ mean? Of the insect sampled for M. ulcerans, families Gerridae and Veliidae are both water surface dwellers that can move significant distances on the water surface, why are the major hemiptera (such as notonectids (not commonly human biters), belostomatids, naucorids and nepids involved with macroinvertebrate surveys associated with M. ulcerans ignored with regards to M. ulcerans DNA presence/absence in this study? There were 149 individuals within these individuals collected, why were they not sampled for M. ulcerans?

We state that “Most bugs collected belonged to just two families, with 498 (45.2%) from the Gerridae (water striders) and 426 (38.7%) from the Veliidae (riffle bugs).” In other words, 84% (the overwhelming majority) of the bugs we collected belonged to just these two families. We also state (lines 344 - 346) “For example, the 54 notonectids, placed in six putative species, comprised 16 morphotypes with the most abundant morphotype having only five specimens.” So, there were two reasons for concentrating on the gerrids and veliids: (a) they were by far the most abundant bugs at our sampling locations; and (b) the remaining bug families were characterised by very high diversity but low abundance, so if we had found positive specimens among them, the ecological and epidemiological significance of this would have been very difficult to define. Ideally, we would have screened all 1,102 bugs, but in light of time and financial constraints, it seemed sensible to assay 70% of the total and focus on the two dominant bug families. We have clarified our strategy in the text.

A major issue is the low sample number of specimens (n=1) testing positive for putative M. ulcerans DNA.

We would agree that this would be a major issue if our molecular assays were not robust. However, the rpoB and IS2404 PCR assays were run with an ulcerans positive control that always showed clear amplification. The 20 samples subjected to Illumina sequencing displayed good concordance with the PCR results because these specimens were positive for rpoB, but not IS2404. The rpoB assay was known not be specific to M. ulcerans, hence the sequenced samples contained sequences from other non-tuberculous mycobacteria. The one discrepant case showing a low level of M. ulcerans DNA and a negative IS2404 PCR result is dealt with in the Discussion (lines 556 - 557) and above. All-in-all, our data (also including the Wolbachia data) exhibit strong internal consistency.

Lines 394-399: Unclear as to why these data are included, they add nothing to the M. ulcerans conversation.

We are sorry if Wolbachia does not interest the referee, but we can assure the him/her that a good many people are interested in this symbiont, and for good reason (especially, we imagine, scientists reading a special issue entitled “Insect Symbiosis”). As we explained earlier in the rebuttal, the presence of Wolbachia can alter the microbiome of the arthropod host and the dissemination of pathogens within the arthropod. Even if this was not the case or the referee personally believes this is unlikely, it would be beholden upon us to report and analyse the major bacterial species present in the sequenced bug samples. Note that although several co-authors involved in this study are active in the Wolbachia field, finding Wolbachia was not a goal of this study and the authors have not deliberately overemphasised the Wolbachia content of these samples above more dominant bacterial sequences. Rather, the presence of Wolbachia in these samples was discovered serendipitously and was found to dominate the bacterial sequence data in many of these bugs.

Discussion

The first paragraph is not relevant to this paper, this study is not about transmission, it should be about diversity, so to begin this section in this manner is not a focused approach. In fact, paragraph two is a rehash of the introduction. The fact that the authors find that their data markedly contrast with other literature in Cameroon is not surprising. The authors sampled insects that are feeding on organic material either plant or terrestrial insects on the water surface. Those hemiptera found to be associated with the presence of M. ulcerans are most likely picking it up from the prey they consume under the water that either filter feed the bacterium from the water or scrape it off of biofilms under the water. Therefore, from an aquatic ecological perspective, they have selected taxa that would in fact not support prior literature – thus to conclude that their data do not agree, of course they do not, it was comparing apples to oranges.

We have greatly shortened the initial paragraphs of the Discussion to reduce repetition from the Introduction. With respect to the referee, we disagree that we didn’t find ulcerans because we were looking in the wrong place. We are aware of studies that associate colonisation of bugs by M. ulcerans with particularly niches or feeding behaviour [such as Ebong et al (2017), which we cite]. However, following the publication of the early studies suggesting that biting bugs were vectors of M. ulcerans, subsequent works [e.g., Benbow et al. (2008) and Marion et al. (2010)] have analysed much greater numbers of belostomatids and notonectids etc. than gerrids or veliids. Thus, there is a clear risk of confirmation bias. However, when even small proportions of gerrids have been screened [as in Marion et al. (2010)], they are positive at similar or higher rates to belostomatids and notonectids (i.e., 20 – 40% positive after pooling). Too few veliids have been screened in previous works to make any conclusions, but as their lifestyle is similar to that of gerrids, there was no a priori reason to think they were less likely to be positive. We addressed these points previously in lines 572 – 574 “…M. ulcerans infection of gerrids has been reported from Cameroon previously at rates similar to other bug taxa [16]. Moreover, the pathogen has been isolated from a gerrid from Benin [17].” We have now expanded our argument in this part of the Discussion slightly.

The entire Wolbachia prose is an insert into a paper that does not flow with the reason for the study in the first place.

With respect, we do not consider the Wolbachia data an “insert” for the reasons we discuss above. Sequencing data in particular can often generate surprises that were not anticipated. Just because they were not part of the original study design does not mean such important findings should not be reported; indeed, quite the contrary.

Reviewer 2 Report

- The discussion paragraph must be summarized.

Author Response

We thank the reviewer for the positive evaluation of our manuscript. We have now added a "Conclusions" section after the Discussion.

Reviewer 3 Report

As an investigative study into the association of M. ulcerans and aquatic insects, I find it is important to publish, even if there is no strong evidence of association and this contradicts previous studies.  Because M. ulcerans and Buruli ulcers is an understudied disease and the reservoir and vector for transmission is still not understood, this study has merit.   The discussion does a good job of interpreting the data and the limitations of possible explanations.  All-in-all a well written report for an important addition to the study of M. ulcerans in the environment.

 

I have not minor or major corrections to comment on.

Author Response

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

Reviewer 4 Report

This a nice ms, well organized and written, with clear objectives and results. It could be accepted in the present form, but including specific names for most genus (by comparing with DNA databases) and making an effort to resume the ms, mainly results and discussion could help to improve it.

Author Response

We thank the referee for their positive comments. In light of referee 1's comments, we have shortened the Discussion somewhat. The inclusion of the supplemental figures and tables in the main Results section by the journal editorial team has, in our opinion, made this section appear a bit confusing and disorganised. With regard to identifying more of our specimens to the species level, we have exhaustively reviewed public DNA databases and the coverage of Afrotropical gerrids and veliids is minimal. Indeed, our study has added substantially to what is known about the aquatic hemipteran fauna of the region, but the formal description of any new species is beyond the scope of the current work.