The secret lives of mobile genetic elements: a multi-level perspective
I became interested in the ecology and evolution of microbial communities a little over a decade ago. Despite being invisible to the naked eye, the microverse is indispensable for life on Earth. Microbial communities are responsible for completing nutrient cycles in terrestrial and aquatic ecosystems, form intricate symbiotic relationships with larger organisms, and have more direct implications for human well-being in terms of the production of antibiotics, gut and skin health, and its relationship to the human immune system. If we wish to protect the biodiversity on this planet — or protect ourselves from dangerous pathogens — we better get a grasp on what makes these systems tick.
Microbial communities: a paradigm system for multi-level evolution
During the first year of my bachelor’s degree in biology, my intuitions on how evolution works were mostly shaped by textbook examples of plants and animals. That is to say: selection acts on large, clearly countable, sexually reproducing individuals, resulting in endless different species. However, concepts like “species” and “individuals” do not translate easily to the microverse. While one does not have to know what exactly a “species” is to appreciate the diversity of life, individuals are part and parcel of Darwinian evolution through natural selection. So what in the microbial world counts as an individual? How do collectives of individuals give rise to “Darwinian entities” in their own right?
It is very tempting to look at microbes and focus on cells as the “true” individuals. Others have argued that a gene-centric approach is more useful. I argue, however, that it may be counter-productive to focus on any one particular level alone. Let me elaborate.
Firstly, the behaviour of a bacterial strain or species typically only makes sense in light of its community. In fact, many bacterial species will not grow when taken out of their community context. It is even possible that community-level properties are so important, that communities themselves are, by Darwinian means or otherwise, evolving “individuals”. This perspective is supported by data that shows how the gene content of individual bacterial species can change rapidly, while the gene content of complex microbial communities appears to be strongly conserved. For example, the distribution of microbial taxa associated with individual animals and plants can vary substantially, but the functional gene content found in each sample is nevertheless very similar. In fact, the assembly of microbial communities appears to be driven by functional genes, not species. How can we explain these observations?
On the one hand, this outcome can be the result of ecological sorting, where lineages naturally assemble into communities with complementing gene repertoires. On the other hand, community assembly may also be driven by evolutionary processes, with species establishing into a new niche through adaptive gene loss, and by picking up new genes from nearby cells or the environment. This process of picking up new genes, called horizontal gene transfer (HGT), may have huge consequences for microbial evolution:
“Taken to extremes, the preponderance of HGT could even imply that microbiomes are better conceptualized as collections of locally adaptive genes, rather than communities of locally adapted species” — James P.J. Hall (2021)
With the quote above, James Hall illustrates how HGT can draw our attention to selection operating on genes, rather than species. Interestingly, however, HGT does not transfer each and every gene with the same likelihood. Certain genes have been shown to transfer a lot, while other genes rarely transfer. So, what determines which genes move horizontally, and which genes are simply inherited vertically from parent to offspring? How can genes live their own lives like that? Well… as it turns out, microbial cells are not the rock-bottom of individuality either. On an even smaller scale, a zoo of mobile genetic elements (MGEs) is replicating, mutating, and differentially surviving (see image below). Through a variety of different mechanisms, these MGEs make it possible for genes to move from one cell to the next.
On a scale smaller than microbes, a rich zoo of mobile genetic elements (MGEs) is replicating,
mutating, and differentially surviving (image by Ellie Harrison)
In a recently published special issue in Philosophical Transactions (Royal Society B) called “The secret lives of mobile genetic elements”, a large cohort of papers was published where MGEs were put in the spotlight. However, I argue that similar to microbial species making no sense outside of their community context, MGEs make no sense outside of their host’s context. Instead, I would argue that MGEs transform microbial communities into a “babushka doll” of Darwinian entities, with all the players shaping one another in concert.
In other words, my fellow apes, microbial communities are a paradigm system for multi-level evolution! Experimentalists are already working hard to come up with new methods to shine a light on these multiple levels. For example, many leading scientists are combining forces in the CRC (collaborative research consortium) 1182: Metaorganism research, trying to get a grasp on how the small impacts the large, and how the large impacts the small. Similarly, theoreticians have been developing mathematical models to describe and understand how multi-level evolution works. Personally, however, I still feel like something is missing…
Embracing complexity in “virtual laboratories”
Considering the large zoo of MGEs found in microbial communities, the many potentially conflicting selection pressures operating on different levels, and the precise spatial and temporal scales, it will be very hard to disentangle the cobweb of interactions in microbial systems. So how can we nevertheless study such a system? I argue that in order to make sense of any of this, we need to dare and embrace not only multi-level evolution, but also the true complexity and “messiness” that comes with biology. This does not mean we all need to become experts on the underlying theory and math on multi-level evolution. In fact, despite having written numerous papers on multi-level evolution in microbial ecosystems, I have never directly done the math myself. Instead, I have mostly relied on computational simulations to figure out what makes these systems tick. Heck, if 70 years ago, computation can save 21 million lives during world war II, it is likely to be highly valuable today. In fact, it was Alan Turing himself who realised that the true power of computation is to teach us by surprise:
“Machines take me by surprise with great frequency.” — Alan Turing (1950)
Indeed, my personal experience with simulations of multi-level evolution is that, more often than not, I am surprised by the outcome. In fact, I often deliberately set myself up to be surprised (see how I do that in the intermezzo below). For example, I found that genes that are less beneficial naturally evolve to be more mobile. I found that bacteria will take up environmental DNA even when exposed to very deleterious genetic parasites. I found that self-replicating DNA inside chromosomes can result in… smaller chromosomes. None of these results was what I initially expected, and I had something entirely different in mind when starting these projects. It all appeared very counter-intuitive to me when I first observed it.
-
How do you allow a computational simulation to surprise you? Here’s a brief description of my process. First, take a minute to visualise, in your head, your current scientifc project. Do you have an idea of how it works? Which proteins move where? When cells divide? How plasmids transfer between cells? You are likely already aware that the cartoon in your head is flawed in many ways, as it is a major simplification of reality. But, precisely how is the “cartoon in your head” wrong?
The perfect way to test the cartoon in your head, is to get it out of your head. We can do that, by trying to implement a simulation of your system, ideally from simple rules. So: open up Rstudio, a python notebook, or my online modelling toolkit, and try and describe your system based on simple rules: cells take up glucose, produce important metabolites, and grow in volume, division happens when a certain volume is reached, etc. Perhaps you are uncertain about some of these steps (e.g. the division volume is debatable), but that’s okay. You can always investigate that assumption later.
Finally, be open to surprise. If your model does not behave how you expected it to, first figure out why. Do not try and build your own expectations into the model. Explore, and most importantly, have fun! :)
Of course, the journey does not end by merely listing a set of counter-intuitive results. Instead, by continuing to experiment with these “virtual laboratories”, one can gradually reshape their (initially flawed) intuition. What happens if we mix the system? What happens if we double mutation rates? What if cells can evolve their rate of taking up DNA from the environment? Eventually, you will understand very well what makes and breaks your virtual ecosystem. Of course, it is still possible that your model is very unrealistic, or even “wrong”. However, models don’t need to be realistic or “correct” in order for them to be useful. Paulien Hogeweg puts it like this in her lecture series on computational biology:
“A is a good model of B, if by studying A you learn something about B.” — Paulien Hogeweg
In other words, the relevant question isn’t whether your model is realistic or correct. Instead, ask yourself this: did you learn something about the system you modeled?
The hypermobility of chromosomal DNA
Alright, now that you (hopefully) appreciate the importance of MGEs, multi-level evolution, and computational simulations, let’s include a new puzzle piece that has been blowing my mind of late. Personally, I have always implicitly assumed that in order for genes to be mobilised, they have to be part of an MGE. In other words: the mobilising influence of MGEs is restricted to the genes that reside on them, right? Right? Well… maybe not.
Three years ago, Chen et al. (2018) discovered that in Staphylococcus aureus, large sections of the bacterial chromosome are mobilised to transfer horizontally. Interestingly, this chromosomal DNA does not encode the ability to transfer itself, but is transferred by the actions of another MGE: bacteriophages. Bacteriophages (or phages) are viruses that infect bacteria, carrying genetic material to replicate itself within their hosts. They inject their DNA into bacterial cells, hijacking the host machinery to produce more copies of themselves. Chen et al. show how during the production of viral particles, many viral particles are produced that encode only DNA of the host chromosome. But the ability of this particle to act as vehicles and infect other cells is no less efficient! As such, hundreds of kilobase pairs of the host chromosome can be effectively “mobilised”. This process has been called lateral transduction.
In a recent follow up study, Humprey et al., (2021) have shown that through lateral transduction, large sections of the chromosome can be much more “mobile” than canonical MGEs like plasmids and transposons. Clearly, lateral transduction forces us to once again reconsider what separates one bacterial lineage from another. Clearly, the barriers which separate microbial strains or species, are filled with holes. If genes can cross these barriers all the time, what is to keep them from becoming truly “selfish”, and replicate at the expense of their host?
To understand how the fitness interests of MGEs and bacteria may or may not align, let us consider for a minute how conflicting fitness interests pose a massive risk. If MGEs exploit their host — or even kill their host — they could eventually drive a complete system collapse. A problem with our current theory is the assumption that once a system collapses, it is somehow “game over”. This of course makes no sense given how we know that >99% of all species that ever lived are now extinct. The extinction of a biological system is far from “game over”. Instead, it makes room for other systems. A system that is less prone to extinction will persist and may be able to spread to the vacant niche space made available by collapsing (neighbouring) systems. Over long timescales, we may therefore expect a long-term selection effect for robustness and ecosystem persistence.
The above-mentioned concept of “survival of the systems” has been recently described by Tim Lenton (2021), and has come forward time and time again in my simulated microbial ecosystems. My simulations repeatedly showed how multi-level evolution resulted in more stable and persistent systems in the long run, instead of one level eventually dominating or driving the system towards extinction. The video below is an example of such an eco-evolutionary simulation, where bacteria and selfishly replicating transposable elements (TEs) gradually evolve to form a more persistent system. The dynamics leading to this outcome involve multiple levels: MGEs, cells, and spatially separated communities.
So, back to lateral transduction. How does it impact microbial evolution, and how important are the above-mentioned levels of selection?
Lateral transduction: questions in need of a multi-level perspective
In a thought-provoking comment, James P. J. Hall (2021) elegantly scrutinizes the idea that lateral transduction renders the bacterial chromosome to be an MGE in itself. But nevertheless, the facts are there: chromosomal genes transfer with great frequency when lateral transduction is involved. So what are the implications? One question addressed in the comment piqued my interest in particular: assuming LT is (somewhat) costly for the phage, can it nevertheless be a phage-level adaptation? And what would that entail? Below I speculate on some follow-up questions, which I believe can only be understood from a multi-level perspective:
Does LT help in purging genetic parasites? (e.g. Croucher et al. (2016) show a similar mechanism)
Is there a trade-off between phage immunity and purging genetic parasites?
Alternatively, are accessory genes / DNA repair the “currency” (Wein et al, 2019) with which a symbiotic exchange is consolidated?
Will bacterial chromosomes (re)organise to have “rescuable genes” (van Dijk et al., 2020) close to phage attachment sites, further solidifying the symbiosis?
Is it necessary (for stability) that the recipient of LT-mediated benefits is itself infected/infectable by the same phage?
Does LT-mediated symbiosis depend on / influence the phage host range?
Alright. Time for another model. Or six.
References
“The secret lives of mobile genetic elements” https://royalsocietypublishing.org/toc/rstb/377/1842
Black, Andrew J., Pierrick Bourrat, and Paul B. Rainey. "Ecological scaffolding and the evolution of individuality." Nature Ecology & Evolution 4.3 (2020): 426-436.
Ereshefsky, Marc, and Makmiller Pedroso. "Rethinking evolutionary individuality." Proceedings of the National Academy of Sciences 112.33 (2015): 10126-10132.
Doolittle, W. Ford, and Austin Booth. "It’s the song, not the singer: an exploration of holobiosis and evolutionary theory." Biology & Philosophy 32.1 (2017): 5-24.
Puigbò, Pere, et al. "Genomes in turmoil: quantification of genome dynamics in prokaryote supergenomes." BMC biology 12.1 (2014): 1-19.
Burke, Catherine, et al. "Bacterial community assembly based on functional genes rather than species." Proceedings of the National Academy of Sciences 108.34 (2011): 14288-14293.
Morris, J. Jeffrey, Richard E. Lenski, and Erik R. Zinser. "The Black Queen Hypothesis: evolution of dependencies through adaptive gene loss." MBio 3.2 (2012): e00036-12.
Quistad, Steven D., Guilhem Doulcier, and Paul B. Rainey. "Experimental manipulation of selfish genetic elements links genes to microbial community function." Philosophical Transactions of the Royal Society B 375.1798 (2020): 20190681.
Jaspers, Cornelia, et al. "Resolving structure and function of metaorganisms through a holistic framework combining reductionist and integrative approaches." Zoology 133 (2019): 81-87.
Chen, John, et al. "Genome hypermobility by lateral transduction." Science 362.6411 (2018): 207-212.
Humphrey, Suzanne, et al. "Bacterial chromosomal mobility via lateral transduction exceeds that of classical mobile genetic elements." Nature Communications 12.1 (2021): 1-13.
Van Nimwegen, Erik, James P. Crutchfield, and Martijn Huynen. "Neutral evolution of mutational robustness." Proceedings of the National Academy of Sciences 96.17 (1999): 9716-9720.
Hogeweg, Paulien. "Toward a theory of multilevel evolution: long-term information integration shapes the mutational landscape and enhances evolvability." Evolutionary systems biology (2012): 195-224.
Croucher, Nicholas J., et al. "Horizontal DNA transfer mechanisms of bacteria as weapons of intragenomic conflict." PLoS biology 14.3 (2016): e1002394.
van Dijk, Bram, and Paulien Hogeweg. "In silico gene-level evolution explains microbial population diversity through differential gene mobility." Genome biology and evolution 8.1 (2016): 176-188.
van Dijk, Bram, et al. "Slightly beneficial genes are retained by bacteria evolving DNA uptake despite selfish elements." Elife 9 (2020): e56801.
van Dijk, Bram, et al. "Transposable elements drive the evolution of genome streamlining." bioRxiv (2021).
Hall, James PJ. "Is the bacterial chromosome a mobile genetic element?." Nature Communications 12.1 (2021): 1-4.
Wein, Tanita, et al. "Currency, exchange, and inheritance in the evolution of symbiosis." Trends in microbiology 27.10 (2019): 836-849.
