Tuesday, November 24, 2015

Q&A on Holobionts and Hologenomes

Image from Brucker and Bordenstein, 2013, Zoology
In September, a few friends on twitter entered into a lively discussion on holobionts and hologenomes that was subsequently borne out in a series of Q&A emails. Below, we summarize this discourse - whose goal was to flesh out some of the nuances and foundations of the holobiont and hologenome terms (related blog post  - What are holobionts and hologenomes?) and concepts. David Baltrus poses queries to Seth Bordenstein and Kevin Theis. What follows is very informal, but useful for the three of us and perhaps our colleagues. We all hope it is helpful from the various angles that one may look at this stimulated area. 

Q1: The first point that I think still needs clarifying after the hologenome article and the perspective piece in Science is what the point of promoting “the hologenome” actually is. I’m probably just too close...but to me it’s not a controversial idea that microbes matter for evolution/ecology of macro-organisms.

A1 (SB): There’s two ways to answer this question. First, I think that this is an uncertain argument as one could raise that same question for virtually any term. Why do we need the term “metagenome” if we knew there were previously microbial genomes in soil, or why do we need the term "selfish gene” if we already knew there were transposons, etc. We could stop there, but I think these terms clearly have uses now that one would not question. Indeed, the selfish gene debate early on is very similar to the holobiont and hologenome terms; moreover I think it goes deeper than just this. We’re not just using words willy nilly :) There hasn’t been a good singular term yet to describe the entity of the multispecies organism and their genomes. In fact, we argue its been quite the opposite as we describe in principle 2 (link to paper):

There appears to be a considerable number of misplaced characterizations and colloquialisms used to refer to host-microbiota symbioses, and these misnomers can potentially act as impasses to new advances”.

We need to chuck out all the superorganism and organ system colloquialisms for good reasons. Moreover, the holobionts and hologenomes properly emphasize the diverse and complex array of symbionts and their genomes in one host species. In contrast to binary symbioses, the multispecies consortia are relatively new and universal features of macrobes that emerged within the last decade for many biologists. That’s what the terms best capture in a timely fashion. See Gordon, Knowlton, Reman, Rohwer, Youle for similar arguments (link).  

A1 (KRT): Yes, it is readily accepted that microbes matter for the evolution/ecology of macro-organisms, but importantly the hologenome concept goes beyond this observation in a very substantive way. Specifically, it posits that the host and its microbiota are a potential primary unit of selection. It is this component of the concept, along with the emphasis on the diversity of the symbiotic communities to be considered, as SB already noted, that requires the use of new terminology. All macro-organisms are populated by microbes, these associations are often not random but rather are replicated across macro-organismal generations and are reflective of their phylogeny, and these associations also often profoundly affect the phenotypes of what we typically think of as individual macro-organisms. As we note in the paper:

the debatable and testable issue of the hologenome is whether nuclear genes and microbes are coinherited to a degree that evolution can operate on their interaction.”

Q2: In light of my own point #1 above, almost all of the PLoS principles paper is couched in terms of microbial “symbionts”. While I gather that you are using the broad term “symbionts” here which includes pathogens, it seems to me that all the points that the piece brings up can be flipped and looked at from the point of view of pathogens as well. I don’t think any of this is controversial if it’s couched in terms of microbial pathogens, especially long term chronic ones. HIV can be environmentally acquired or vertically inherited and certainly causes changes in human phenotypes/evolution/ecology. These phenotypic changes (i.e. loss of T cells) can appear to “reboot” some aspects of Lamarck’s ideas just as much as beneficial microbes, only in this case the acquired phenotype is disease. Do pathogenic infections count for rebooting Lamarckian principles if they are inherited across generations? Seems like it's the same story, but you guys put the Larmarckian thing in there to be a bit provocative, which is completely understandable, it just kind of sticks out to me. I don’t think that you see a distinction between beneficial and pathogenic microbes, but I think at least some of the ideas mentioned have been pretty thoroughly worked out for pathogens (speciation for one). If anything, I might argue that pathogens have more power to coevolve with humans than beneficial microbes because of stronger selection pressures on both ends (handwaving over a lot of assumptions).

A2 (SB): I'm trying to follow the logic here, as it seems you are playing both sides of the Q/A. First, you are right in that we have always used symbionts in our publications to refer to all types of associations. And we do not subscribe to overselling one phenotype for another. Quite the contrary, we write that

In the microbiome, selection favors the spread of beneficial microbes involved in nutrition, defense, or reproduction, while pathogenic microbes are either purged by holobiont selection or the pathogens deploy adaptations….”

Second, there is the issue of Lamarck. You agree with Lamarckian aspects of pathogens, but then questions if using Lamarck in the paper is too provocative. Mmmm…can’t have it both ways :) Our goal is to get the science right in its most accurate way, not be provocative in a purposeful sense. If that accuracy or view of science happens to be taken as provocative, that might say something more about where the community is now, than our intent per se. Indeed, Margaret McFall-Ngai once told similar stories for some of her perspectives on the immune system’s role in regulating beneficial microbes. The immunologists were initially up in arms. Finally and to your last point, it is the very fact that many aspects of symbioses have been worked out that we can reason the holobiont and hologenome concepts are the best model to explain the complexity of the systems in a comprehensive light. I think Kevin and I both agree also that its’ not perfect. It needs to be made better with more concepts, theory, and experimentation that is the stated intent of the essay.

A2 (KRT):  We do employ “symbiont” broadly, referring to resident microbes: beneficial, pathogenic, or otherwise. The hologenome concept accommodates each of these interaction types. Lamarckian evolution is simply heritability of acquired characteristics. If acquired symbionts have fitness consequences and are thereafter heritable, they are consistent with Lamarckian evolution. I do not see that as provocative. It should, however, be noted that although we present the argument for a Lamarckian element to the hologenome concept, the Rosenbergs (and many others) of course originally did so (link).

In general, strength of selection may be greater on pathogenic interactions as you suggest, but it can be substantial on beneficial interactions as well. The Rosenbergs’ coral example comes readily to mind (link). Regardless, this is not consequential to the validation or refutation of the hologenome concept as again it accommodates all interaction types. Lastly, I concur with SB, the hologenome concept is not perfect, and in the essay we discuss where future efforts should be focused, but it is the most comprehensive evolutionary framework we currently have for discussing and investigating the ecological and evolutionary complexities of host-microbial systems.

Q3: We have a pretty good vocabulary for ideas like co-evolution, multi-level selection, phenotypic plasticity. What makes the hologenome different enough to warrant special mention in light of a lot of previous theory and literature? If it’s combination of all these things, all of these ideas must inherently incorporate each other in nature anyway, no?

A3 (SB): I think A1 mostly applies here too, but let me know if I'm missing something. Most importantly, we haven’t had a word/singular framework that captures the fact that all of these evolutionary and ecological processes are happening at once across the complexity of the holobiont. We’ve traditionally seen host-microbe interactions as interactions between individual host and individual microbe. The holobiont and holgoenome concepts instead place emphasis on the complex systems biology of the whole system driving evolution. For many phenotypes (and as Charles Goodnight initially put in an email to me), so-called "individual selection" will actually be selection on the community, with many players affecting the selection, and that is what we as a community haven’t appreciated enough.

A3 (KRT): Nice explanation SB. I will only add that, for me at least, the holobiont concept is a redefining of that which constitutes an individual animal or plant. This new construct, or unit, requires a name and a clear definition. That definition is itself ‘evolving,’ hopefully catalyzed to some degree by our essay.

Q4a: You say that the hologenome is “not complex” and point out that they are really only referring to microbes (see microbiome in Title).

A4a (SB): Mmmm, you’re taking words out of context here. We actually state the opposite:

The holobiont and hologenome concepts upgrade this conventional vision to encompass the vast ecological and genomic COMPLEXITY of a host and its total microbiota

And the particular quote excised above was from the following paragraph:

For instance, do the genomes of an insect pollinator and flower constitute a hologenome? The ANSWER HERE IS NOT COMPLEX. Holobionts and their hologenomes are exclusive to the hosts and their associated microbiota. Different holobionts, such as the aforementioned pollinator and flower, clearly interact, but these interactions are not new to biology, as they form the basis for all past and present ecological investigations. They are simply holobionts themselves interacting with each other."

Q4b: My read of this has always been that I don’t know why microbes should be viewed any differently than macrobes that have close associations with humans. The presence of worms can affect distributions of gut microbes across animal life. Macrobes can have just as much an effect on human evolution as microbes. Why the distinction? Is there a distinction? Is it because that humans/etc...are hosts for microbes? Humans are hosts for worms and mites over the course of their whole lives too. Worms and mites can show patterns of vertical inheritance just like microbes. Why are worms and mites “holobionts themselves interacting with one another”? Microbes inside of the microbiome interact within one another, and can be hosts to other microbes and viruses (see Bdellovibrio and prophage), why isn’t each microbe itself a holobiont? <head briefly explodes>. There has to be a distinction somewhere along the way, and I’m still not getting it even after multiple discussions and reading the PLoS paper. Maybe I’m just prone to thinking everything in Biology is complex and worry too much about the specifics? Even so, there has to be a simple demarcation point going forward that delineates what the hologenome is.

A4b (SB): A4a applies here in addition to the salient difference that while worms live in and outside humans, the worm and human can live without themselves, but neither can live without their microbiomes. This is a key point for why the holobiont and hologenome stop at the host and its associated microbes. We as a community just haven’t appreciated this aspect enough and there’s far more science to delve into because of it.

A4b (KRT): I agree with SB’s broader sentiment here, and will add that the hologenome concept explains (or seeks to explain) the evolution of animals and plants, organisms (i.e. holobionts) whose phenotypes are necessarily products of interactions between their own genomes and those of their symbiotic microbes. The concept does not explain the evolution of microbes.

Q5: What about microbes that are completely neutral to the hologenome? Are these incorporated into the concept (I assume they are)? I’m coining the term “junk microbes” here for these very special cases so that we can have all the fun of the ENCODE folks about arguing what percentage of the hologenome is composed of junk microbes. If you include “junk” microbes into the discussion, then the hologenome includes cases where selection doesn’t act. Therefore, the hologenome idea would include post-reproductive stages of hosts as well, where (hand waving about grandmother’s nurturing effects on evolution aside) the hologenome doesn’t matter for evolution and is just there in the present time for the host. The microbes will of course still evolve with each other. Maybe one prediction is bacterial cheaters arising in older hosts? Just fuel for discussion,  another added layer to think about whether selection is implicit in the concept of the hologenome (I don't think it is).

A5 (SB): Are we really on Q5? Wow :) To your first point, principle VIII and III is all about neutrality and selection in the hologenome. I think we’re on the same page here. And this is an important point of the essay as the original holobiont and hologenome concepts did not distinguish between neutral and selected microbes.

A5 (KRT): Ah, see the “Microbiome Mutiny Hypothesis” (link).

Q6: “Antibiotic or axenic experiments in speciation studies must be routine”. This doesn’t make sense to me, even though I understand exactly where the data is coming from and why you say it. Speciation has never occurred in the absence of microbes in any eukaryote. Taking the microbes away is a completely unrealistic environment and wouldn’t ever happen under natural conditions. I’d be on board with saying microbe swaps are necessary for speciation studies, but to me the axenic experiments don’t tell you anything about speciation in naturo because it’s artificial.

A6 (SB): Would you believe that you disagree with Jerry Coyne too, as he stated the same thing in his Nature review article in 1992! The axenic experiments will importantly tell the relative fraction of reproductive isolation phenotypes that are microbe dependent, something that is assumed by most speciation geneticists to be rare if you were to poll them. So, its a starting point to get us a “symbiotic heritability” if you will for the effects of symbionts vs. host genes on reproductive isolation traits. The microbe swapping is certainly a good next step, but the former is where we lack the most data right now and the easiest to do for wide-ranging systems.

Q7: Phylosymbiosis...Again, I completely get the concept in the broad sense, but this is another place the absolutely needs to be fleshed out. Different branches in the phylogeny should have their own microbiome signals. Where do you draw the line? Microbiomes can easily differ within a species. Microbiomes can easily differ within one individual within one species over time, sometimes in really random ways…….My further guess is that, like most coevolution studies, it’s going to be hard to distinguish between co-evolution and (for lack of a better term) shared environments. Humans/chimps have largely divergent microbiomes, enough to see phylosymbiotic signals. I’m not sure that the same could be said for different species of lemurs on Madagascar because they all fill different niches. Correlations make things messy.

A7 (SB): The question of “where do you draw the line” of microbiome divergence between species for within species is both an experimental question that we’re trying to answer right now and an intellectual question that scales just as well to good old fashion genes. The blurriness in transition from within to between species variation is evidence for evolutionary change over time rather than problematic. Phylosymbiosis is really only best done under “controlled” studies that avoid confounding issues of diet, gender, age, etc. That is why the Nasonia and Hydra studies stand out in my mind. We’ve got more data on other systems as well now. And its important to note here that phylosymbiosis does not equal coevolution. I think we've been pretty clear about this since its inception. I actually came up with the word phylosymbiosis precisely because the concordance between the host phylogeny and microbiome communities does not necessarily mean coevolution.   

A7 (KRT): As the field biologist, or at least once upon a time field biologist, I will suggest that, while I agree with everything SB wrote, broad phylosymbiotic investigations in the wild have merit in that they reveal the extent of microbiota/microbiome variation among hosts in natural systems (descriptive data that are needed), and, if no host phylogenetic signal in symbiotic microbial community composition or structure is evident (while taking into account other host characteristics like diet, sex, etc.), then it strongly suggests that these host-microbial associations are not as pronounced and persistent as the holobiont and hologenome concepts posit.

Q8:  I’m “in like” with the analogy to mitochondria and microbiomes, I’m not in love with it. You can lose some of your microbes and still be functional, you can’t lose your mitochondria and still be a functional human. I think there is a fundamental difference, given phenotypic redundancy in the microbiome. Always willing to be proven otherwise:)

A8 (SB): Right…it's a gradient of dependency. I take the fact that certain microbes can be lost, but some microbes are required as a compelling reason for why the holobiont and hologenome concepts are relevant and novel. We know that the microbes need to be there to some degree and with some specificity, but we also know that any given microbe many not be required. Fascinating!

No comments:

Post a Comment