Expansion of the molecular and morphological diversity of Acanthamoebidae (Centramoebida, Amoebozoa) and identification of a novel life cycle type within the group
Reviewer’s report 2: Sandra Baldauf, Uppsala University, Sweden
Endorse Publication
Reviewer Summary:
-Amoebozoa is a generally poorly understood and very unevenly sampled eukaryote supergroup. This includes the major division Acanthamoebidae, despite the fact that it includes important model organisms. The ms reports substantial interesting new data. The most interesting aspect is the origin of sporocarpy, important for understanding basic properties of especially soil microbes, such as dispersal and dormancy. Therefore, the work potentially appropriate for a scientifically broad journal, such as BiolDirect. However, this manuscript lacks coherence and seems to be 2-3 different manuscripts – classical protist taxonomy (Fig. 1), deep phylogeny of Amoebozoa (Fig. 2) and molecular phylogeny of Acanthamoebidae (Additional file 5: Figure S3 and Additional file 6: Figure S2). It was particularly unclear to me what the point of Fig. 2 is. There are also some problems with presentation, but most of these could be easily fixed. I think the manuscript is important and interesting but needs some major revision. You might also consider moving Fig. 2 to a separate manuscript, or if it is included, some additional analyses are recommended and more information on its relevance (detailed below).
Reviewer recommendations to authors:
In general, the manuscript is somewhat lacking in coherence and seems almost like 2-3 different manuscripts – classical protist taxonomy (Fig. 1), deep phylogeny of Amoebozoa (Fig. 2) and molecular phylogeny of Acanthamoebidae (Additional file 6: Figure S3 and Additional file 7: Figure S2).
Author’s response: We have made every effort to clearly unify these three elements (which we feel are all equally necessary) into a single coherent and concise story.
It was particularly unclear to me what the point of Fig. 2 is.
Author’s response: The Acanthamoebidae and Centramoebida have typically never had high statistical support in molecular phylogenetic reconstructions of Amoebozoa. However, due to the morphological and ultrastructural similarities of the genera that make up the classical composition of the group (i.e. Acanthamoeba and Protacanthamoeba), the validity has never been called into question. We have seen from previous work of ours on Luapeleamoeba hula that due to the drastically different morphology and life cycle of this organism from that of traditional acanthamoebids, reviewers have been highly skeptical of results using SSU alone placing it in Acanthamoebidae. Since the morphologies of both Luapeleamoeba arachisporum and Dracoamoeba jomungandri (deposited as Stereomyxa ramosa ATCC® 50982™ ), and the life cycles of Luapeleamoeba arachisporum and Acanthamoeba pyriformis are equally/more divergent from traditional acanthamoebids than that of Luapeleamoeba hula we chose to use phylogenomics as an additional and possibly more convincing line of evidence for the inclusion of these taxa in the group. Another initial incentive of ours to include this component in our study was the only data available for Dracoamoeba jomungandri (“Stereomyxa ramosa” ATCC® 50982™) was a transcriptome generated by the Marine Microbial Eukaryote Transcriptome Sequencing Project. We wanted to include it in our analyses on Acanthamoebidae since previous phylogenomic studies show it was sister to Acanthamoeba castellanii, but could say no more about the exact phylogenetic placement of this organism due to limited taxon sampling. Despite being able to bioinfomatically find the SSU of this organism in the available transcriptome and thus include it in our SSU analysis, we still feel the above concerns from others in our community, and the traditional lack of support from either ML or Bayesian analyses for the group justify/require the phylogenomic analysis to be included here.
Specific Points Fig. 1 (taxonomy) The discussion of this figure is 50% of Results, but for a general reader the terminology is inaccessible (furcate, hyaloplasm, lamellopodia, pellitids, etc.) and the detail is of limited utility. About half of this is also confirmation of previous descriptions of the same species. This is all unlikely to be useful for other than a specialist and needs some revision to be more widely accessible (e.g., define terms, move less relevant details to SupDat).
Author’s response: Agreed. The morphological details that would allow experts to feel confident in our identification of previously described species isolated from nature due to their unavailability from any culture collection (i.e. Acanthamoeba pyriformis and Luapeleamoeba arachisporum) have been moved to the supplementary results section. We have also gone through and either replaced specialist terms with more widely understood synonyms or defined them upon their initial use.
Figure 2 (global rooted phylogeny of Amoebozoa) Most of the Results for this figure focus on the root, which differs from previous work. However, you don’t explain why this is relevant here. Perhaps it is meant to show monophyly of Acanthamoebidae Centramoebids, but you’ve not made it clear that this is in question.
Author’s response: The purpose of the figure was to show the monophyly of Acanthamoebidae centramoebids when including organisms we now show to be acanthamoebids, but differ greatly with respects to their morphology and life cycles in some cases from that of traditional acanthamoebids (Acanthamoeba spp. and Protacanthamoeba spp.). In most phylogenetic reconstructions of Amoebozoa using the SSU gene where L. hula is included, neither Centramoebida nor Acanthamoebidae are strongly supported (posterior probability???.95 and ML bootstrap???80 in our opinion) in both ML and Bayesian analyses (i.e. Shadwick et al. 2009, Lahr et al. 2011, and Berney et al. 2015). In order to clarify this was indeed the purpose of this figure, we have tried to introduce this topic more clearly in the abstract and background sections. Also any discussion about the overall topology or root of the tree has been removed from the results section.
Also, please explain why other deeply sequenced in-group taxa are not included (e.g. Stereomyxa?)
Author’s response: The transcriptome of “Stereomyxa ramosa” ATCC® 50982™ generated by the Marine Microbial Eukaryote Transcriptome Sequencing Project is included in our analysis as mentioned above. We renamed this strain Dracoamoeba jormungandri, which is a major emphasis of this manuscript, as our light microscope observations on this organism were dramatically different than those of Grell in his original description of S. ramosa. We chose to discuss these inconsistencies between our observations on this organism and those of Grell 1966 in greater detail in a supplementary discussion section as these findings are likely only of interest to specialists. To help with confusion related to old and proposed new names we have also added a table at the suggestion of Reviewer 3 (Purificacion Lopez-Garcia) that includes the names of all isolates used in this study and our suggested new names based on our phylogenetic analyses if relevant. We refer to this table early on.
Alternatively, if the question is whether more Acanthamoebid data affect deep resolution in Amoebozoa (which seems unlikely), then it would make more sense if this was the last figure. In that case, I also strongly recommend additional controls, such as alternative hypothesis (e.g. AU) tests, and analyses with different outgroups, particularly given the very short internal branches and extremely long terminal ones.
Author’s response: This is not the case. See above.
Some additional information also – e.g., clearly state that only 2 of 6 PhyloBayes runs converged.
Author’s response: This section has been changed to read:
After 1200 generations convergence was achieved for
two of the six chains
.
These two chains
were summarized……
How were problematic sequences vetted (i.e. for what)?
Author’s response: We are not sure how to further clarify this passage from our methods:
“To test for undetected paralogy or contaminants, we constructed a consensus tree (ConTree) representing phylogenetic groupings of well-established eukaryotic clades [35]. The resulting individual protein trees that placed taxa in conflicting positions relative to the ConTree with more than 70% ML bootstrap support, with a zero-branch length, or with extremely long branches were checked manually. All problematic sequences identified using these methods were removed from the dataset.”
Except maybe to say that the sequences that fit the criterion above were deleted from our dataset due to the suspicion that they might represent paralogs or contamination from another organism. These sequences, for obvious reasons, would result in an erroneous phylogenetic signal for a particular organism.
The methods all use complex models, which come at the expense of adequate search algorithms and rigorous statistical tests. However, the latter are especially important for complex trees (many taxa, widely different rates). If you want to make a strong case for an alternative root, a more comprehensive bootstrap analysis as well as AU tests would be good.
Author’s response: Again, the intentions of the phylogenomic analyses in the manuscript were not to evaluate where the root of Amoebozoa may lie. Our goal was merely to add an additional layer of support for the monophyly of the Acanthamoebidae clade of Centramoebida that includes our morphologically diverse organisms of interest.
Figure 3 (SSU phylogeny) Most of this section of Results focuses on Additional file 6: Figure S2. This is a nice figure, informative and well-presented. The purpose of using Fig. 3 instead, which does not include all the taxa in question, is not clear. You also don’t even mention this figure until half way through this section of Results, which I found confusing.
Author’s response: We agree, and have now moved an aesthetically modified version of Fig. S2 (Now Fig.
3
) into the main text. However, we choose to maintain our centramoebid enriched tree (formally Fig.
3
now Fig.
4
) to show a more precise and more well resolved phylogeny of our group of interest.
Results The description of trees in Results is mostly a repetition of the names in the figures. This is hard work to read and didn’t add much to my understanding of the main points of the figure. It is _very_ frustrating that taxon labels in the trees are different from those in the text (e.g. Protostelium pyriformis is referred to as such throughout the text but labelled as Acanthamoeba pyriformis in all the trees). This is only explained in the last paragraph of Results. Some higher level taxon names used in the text are also not defined or labeled in the figures (e.g Gocevidae – no indication it includes Endostelium).
Author’s response: We have revised both results sections that discuss tree topologies to take these constructive comments into consideration. We have edited the manuscript in a way that establishes new names of taxa early on and consistently refer to our new names which are displayed on all trees throughout.
Discussion There are some very interesting points and additional informal observations. However, this is quite long and tends to ramble in places. I think this would be easier to read and have much more impact if you tightened it up a bit.
Author’s response: We have made every attempt to streamline the discussion to focus on the main points.
In some places there are also multiple layers of speculation, which you should probably keep to a minimum. Homology of sporocarps is critically important, but this is simply stated as a fact. It would greatly help to have documented evidence from micrographs and maybe also diagrams.
Author’s response: We certainly did not intend to word our discussion in a way that would lead readers to assume that homology of sporocarpy across Amoebozoa is a proven fact. This is simply working hypothesis that is a future emphasis of our research endeavors. We have edited this section to make this as clear as possible.
When we discuss similarities in sporocarp development or morphology that may indicate potential homology of acanthameobid/centramoebid sporocarpy, numerous citations where these observations were originally documented/discussed or followed up on. As this work has already been done, the need to include additional light micrographs or diagrams of these observations in this particular study.
I don’t think it’s wise to dismiss this under the assumption that genome sequences will solve it. A genome sequence is still a long way from identifying genes responsible for specific traits, particularly for an erratically expressed one. So I expect that this is going to have to rely on ultrastructural evidence for some time.
Author’s response: We agree fully and have edited all such statements to suggest a more holistic approach to tackle this question. This includes techniques that are old and some that very new. We absolutely agree and understand that genomes and transcriptomes are useful and informative tools, but as you point out, especially in non-model organisms we are a long way away from being able to pinpoint exact molecular machinery responsible for particular traits. Although the generation of that type of data is a logical step towards an answer to the intriguing question of sporocarp homology and ancestral amoebozoan complexity.
I would suggest some caution in putting too much emphasis on the fact that sporocarps are unknown outside Amoebozoa, since you show they are often missed when present, even in Amoebozoa.
Author’s response: We fully understand this recommendation and have edited our introduction of this concept to be more cautious. However, with the results of this body of work considered, of the 33 described species of amoebae known to exhibit protosteloid sporocarpic fruiting the phylogenetic home of only one (Microglomus paxillus) is truly a mystery. All protosteloid amoebae that have molecular data available have found a phylogenetic home in Amoebozoa. The few (aside from M. paxillus) which have none are clearly close relatives to one or more that have been placed with high statistical support within Amoebozoa by molecular phylogenies. Also, all myxogastids sequenced so far form a monophyletic group in Amoebozoa. Many of us who have worked with protosteloid amoebae have also worked with amoebae from a variety of locations across the tree of life. We have grown these amoebae, if possible, in conditions that would facilitate sporocarpic fruiting and have seen none. At least for now we see no reason to believe that sporocarpy has evolved outside of Amoebozoa; however, we do not deny this possibility. The original avenues for some of us, FWS, MWB, LLS, into an interest in amoebozoans was from our surveys for protosteloid amoebae. We have searched the world thoroughly using techniques designed to maximize our chances of finding protosteloid organisms on the basis of their fruiting. We have looked both at traditional terrestrial substrates and submerged freshwater substrates. When they are plated out in a manner to yield sporocarps we have never found any taxa that appear to be other than amoebozoans. Certainly, should it ever come to pass that a sporocarpic non-amoebozoan is discovered, we will happily revise our thinking, but our vast experience suggests that such a finding would have been expected by now.
Suggestions:
Taxonomy (Fig. 1) – move detailed confirmatory descriptions to SupDat – for the remainder, simplify and define terminology or summarize only and put details in SupDat
Author’s response: Done. As stated above the confirmatory descriptions are now in supplementary discussion. We have also gone through and defined any specialist terms upon their initial use.
– establish new names early on and use consistently (since these aren’t formal descriptions you could even add them to section headings).
Author’s response: Done. Section headings now include purposed new names for all taxa of interest. Names have also been changed throughout the entirety of text (with the exception of the “Abstract” and “Background” sections) to reflect what is used in the figures. We also hope the previously mentioned table that has been added will help with this.
Deep phylogeny (Fig. 2) – simplify or delete – if retained – include all Acanthamoebidae/Centramoebidae with substantial genomic data – simplify rest of tree (collapse nodes, delete irrelevant problematic taxa (e.g. archamoebae, probably also outgroup) – or reduce to Discosea and apply root from separate analyses (place latter in SupDat) – clarify what the purpose of these analyses are, and, if the root is the main concern, include additional controls and full bootstrap analyses.
Author’s response: As previously mentioned all deeply sequenced Acanthamoebidae/Centramoebidae lineages were included in our analyses. Again, trying to determine the location of the root of Amoebozoa was not the goal of these analyses. We have made all possible efforts to clarify our intentions in the text. We choose to not remove the figure or alter it beyond the addition of a label for “Centramoebida”, but have made a great effort to clarify early on what the true intention of these analyses were (show that Acanthamoebidae includes are taxa of interest and is fully supported in both ML and Bayesian analyses).
SSU tree (Fig. 3) – replace Fig. 3 with Fig. 2S, and collapse a few of the more heavily sampled outgroup clades (e.g. most P. mycophaga) (retain full version of 2S in SupDat) Use consistent names – rename OTUs early in Results (can refer forward to SSU tree, if needed) – make sure all higher taxon names used in text are also labelled in the figures (if possible)
Author’s response: As stated above. An aesthetically modified version of Fig. S2 is now included in the main text as Fig.
3
. The full version of this tree is retained in supplement as Fig. S2. However, we choose to retain our centramoebid enriched phylogeny (formally Fig.
3
now Fig.
4
) for reasons mentioned above. All higher taxon names discussed in the text are now labelled on appropriate tree figures.
Tighten the Discussion – more information on sporocarp morphology in different taxa, including illustrations.
Author’s response: See above.
-Style – suggest you avoid excessive use of first person and most narrative writing – best to minimise use of hyperbole (greatly increase, deeply affect, cutting-edge…, ) –
Author’s response: All instances have been removed.
avoid incendiary language (Li. 416 “…based on a profound ignorance of the literature”) –
Author’s response: This statement has been changed.
avoid referring to hypotheses as “not yet proven” (as opposed to “not yet tested”)
Author’s response: All instances have been removed or changed.