The ecology of the genome and the dynamics of the biological dark matter

Highlights

• Transposable elements (TEs) are DNA parasites that, in asexual taxa, are predicted to either vanish or proliferate until host extinction.

• New Eco-genomic models reveal original stable persistence of TEs in asexual genomes.

• Stability arises from density-dependent selection and TEs over-dispersion among lineages.

• TE-induced host extinction was over-estimated by earlier models with no molecular silencing.

• More stable TEs dynamics in asexual taxa weakens the advantage of maintaining sex.

Abstract

Transposable elements (TEs) are essential components of the eukaryotic genomes. While mostly deleterious, evidence is mounting that TEs provide the host with beneficial adaptations. How ‘selfish’ or ‘parasitic’ DNA persists until it helps species evolution is emerging as a major evolutionary puzzle, especially in asexual taxa where the lack of sex strongly impede the spread of TEs. Since occasional but unchecked TE proliferations would ultimately drive host lineages toward extinction, asexual genomes are typically predicted to be free of TEs, which contrasts with their persistence in asexual taxa. We designed innovative ‘Eco-genomic’ models that account for both host demography and within-host molecular mechanisms of transposition and silencing to analyze their impact on TE dynamics in asexual genome populations. We unraveled that the spread of TEs can be limited to a stable level by density-dependent purifying selection when TE copies are over-dispersed among lineages and the host demographic turn-over is fast. We also showed that TE silencing can protect host populations in two ways; by preventing TEs with weak effects to accumulate or by favoring the elimination of TEs with large effects. Our predictions may explain TE persistence in known asexual taxa that typically show fast demography and where TE copy number variation between lineages is expected. Such TE persistence in asexual taxa potentially has important implications for their evolvability and the preservation of sexual reproduction.

Introduction

How organisms are able to adapt to new environmental conditions is central to evolutionary biology, and to unravel the determinants of such adaptation has increasing socio-economic implications in the context of global changes (Carroll et al., 2014). The scientific knowledge used to address those challenges is rooted in standard genetic and ecological studies. Genetic approaches provide insights into traits adaptive changes by accounting for a description of the underlying sets of genes and their interactions (Mauricio, 2005). Ecological approaches are essential in identifying the specific forms of frequency- and density-dependent selection that emerge from detailed descriptions of ecological interactions (Dieckmann et al., 2004). In both cases, the underlying Neo-Darwinian view implies the evolution of genes determining life-history traits and/or reproductive isolation (Barrett and Schluter, 2008).

The revolution in sequencing technologies has revealed that such genes only represent a minor fraction of eukaryotic genomes as non-genes typically account for about two-thirds of our own genome (de Koning et al., 2011) and up to 85% of the maize genome (Tenaillon et al., 2010). Non-coding sequences were soon referred to as the genomic ‘dark matter’ by analogy with the hypothetical substance that is predicted to account for around five-sixths of the matter of the universe (Rubin et al., 1980). Comparison does not hold far behind such figures as this part of the genome is anything but made of slow moving massive particles that weakly interact with normal matter. Instead, genomic and post-genomic studies have started to shade light on the importance of those non-genic sequences that can potentially dwarf the information of the genes. Not only can non-genes contribute to the regulation of gene expression (Elbarbary et al., 2016), but they can also be highly mutagenic actors altering genome structure (Slotkin and Martienssen, 2007), standing genetic variations and species evolvability (Lynch, 2016).

These drastic changes in our perception of genomes structure and dynamics are driven by terabytes of data generated by ever-higher throughput genomic, transcriptional and post-transcriptional studies. With such an unprecedented accumulation of information, new opportunities are emerging to better understand the mechanisms underlying micro-evolution (van’t Hof and Campagne, 2016) and biological innovations (Lynch et al., 2011). As in other fields of biology, theoretical approaches are essential in digging up knowledge from complex datasets (Cohen, 2004). This has led to repeated calls to develop an ‘Ecology of the Genome’, an approach that aims at adapting ecological concepts and models to comprehend the interactions between the genic and non-genic entities shaping genomes (Brookfield, 2005, Venner et al., 2009).

The broad objective of the present study is to contribute to the emergence of this approach by a theoretical investigation of the dynamics of retro-transposons, i.e. class I transposable elements, thereafter simply referred to as ‘TEs’ although, sensu stricto, the definition ‘TEs’ also include class II elements. Those TEs are widely spread in eukaryotic species, typically representing 10–50% of their DNA (Tenaillon et al., 2010) and constituting a substantial amount of the genomic ‘dark-matter’. They indeed have a high potential to spread via ‘copy and paste’ mechanisms, which raises a fundamental question at the heart of genome ecology; what is regulating TE proliferation and the correlated increase in genome size by restraining their copy number to an equilibrium level?

Most theoretical answers to this question have been focused on two hypothetical regulatory mechanisms. Modelling studies have shown that i) selection against deleterious effects of TEs can allow for a stable number of copies if such effects show negative synergistic epistasis (Charlesworth and Charlesworth, 1983, Charlesworth and Langley, 1989), and that ii) regulation of transposition can stabilize the number of TE copies through competitive interactions between copies (Langley et al., 1983, Charlesworth and Langley, 1986, Le Rouzic and Capy, 2006, Basten and Moody, 1991). Comparisons with the theory of host-parasite interactions suggest that several features lacking in the above genetic models could have an impact on the predicted TE dynamics. First, host population size is typically considered as a constant in these models, so that they not allow tracking the effect of the spread of TEs on the population size. The corresponding models therefore do not account for the density-dependent feedback between the spread of TEs and the host population dynamics, while it is clearly established in theoretical epidemiology that such feedback plays a key role in regulating ‘true’ parasite populations (Diekmann and Heesterbeek, 2000). Second, genetic models lack a flexible description of the distribution of the number of TE copies per host individual that is considered to be Poisson (Brookfield and Badge, 1997), while experimental (Kalendar et al., 2000, Vielle-Calzada et al., 2009) and theoretical (Wright and Schoen, 1999, Iranzo et al., 2017) studies have shown that over-dispersed TE distribution emerge from self-fertilization or heterogeneities between lineages. The theoretical evidences that aggregation of ‘true’ parasites in hosts significantly affect the stability of their interactions (Anderson and May, 1978, Dobson and Hudson, 1992, Gaba and Gourbiere, 2008) suggest that such over-dispersion of TE distribution is likely to have an impact on TE persistence. Third, most genetic models do not consider TE epigenetic silencing despite its ubiquitous effect on transposition (Tenaillon et al., 2010, Elbarbary et al., 2016, Slotkin and Martienssen, 2007, Lynch et al., 2011) and while similar mechanism of within-host developmental delays, such as dormancy of ‘true’ parasites, have been shown to significantly change host-parasite dynamics (Dobson and Hudson, 1992, Gaba and Gourbiere, 2008). While TE copies can also be inactivated by mutations, such as insertions or deletions, we restrained ourselves from considering the distribution of non-transposing remnant copies (as in (Sheinman et al., 2016, Banuelos and Sindi, 2018), although those can potentially contribute to host adaptation (Casacuberta and Gonzales, 2013) and genome size variations (Fischer et al., 2014).

In this contribution, we developed original ‘Eco-genomic’ models of host-TE population dynamic based on analogies between TE dynamics and the transmission of ‘macro-parasites’, such as helminths and parasitic arthropods, in order to provide theoretical insights into the potential effects of the three above determinants that are currently not accounted for in our theoretical understanding of TE population dynamics and evolution.