Reticulate evolution, lateral gene transfer, and innovation in plants

The evolutionary diversification of organisms is commonly pictured as following a bifurcating tree of life. As time passes, species divide and follow their own evolutionary path along separate branches. In this strict bifurcating tree scenario, any evolutionary novelty or innovation is restricted to the lineage in which it arose. However, it is well known that branches of the tree of life can be porous, with episodic or even frequent genetic exchanges between species. Hybridization is especially frequent in some groups of plants and allows various amounts of gene flow between closely related taxa (Goulet et al., 2017). While hybridization involves sexual reproduction, recent evidence indicates that genetic material is also transferred between distantly related plant species by means other than sexual reproduction in a process known as lateral gene transfer (LGT). The cases of LGT reported to date involve various groups of plants and donors from different parts of the plant phylogeny or different phyla (reviewed by Richardson and Palmer, 2007; Gao et al., 2014; Wickell and Li, 2020), and genes received from bacteria have even been linked to the emergence of land plants (Cheng et al., 2019). In this short essay, we discuss the potential impacts of these genetic transfers for our understanding of plant evolution. We focus on LGTs, because they are readily tractable for assessing the effect of reticulations on plant functional diversification. The impacts of hybridization and introgression are, however, similar and also briefly discussed. The remarkable adaptations that have allowed plants to colonize most environments around the globe did not arise over night. In some cases, protracted evolutionary periods were required to incrementally modify genes either inherited from their ancestors or originating de novo in the genome (Bianconi et al., in press). Reticulation events have the potential to speed up this process and even provide access to new traits that would otherwise have been out of reach of natural selection. For the recipient genome, LGT represents the insertion of foreign DNA, effectively generating a major structural variant that natural selection can act upon. The fate of each LGT depends on its selective value in the recipient, and the frequency we observe will be significantly lower than the actual rate in nature as most transfers will be either deleterious or selectively neutral. However, given a high enough frequency, random transfers of DNA will occasionally increase the fitness of the recipients. Such beneficial LGTs include those transferring functional genes that have undergone substantial modification through prolonged periods of positive selection in the donor genomes before being transferred (Christin et al., 2012). A gene that is selected for may also be accompanied by hitchhikers on the same fragment of foreign DNA, which can act as standing genetic variation for secondary selection (Olofsson et al., 2019). Lateral gene transfer thus provides a mechanism by which plants can effectively “steal” the genetic blueprints for molecular adaptation, thereby recycling the product of natural selection and altering the evolutionary trajectory of a species. Among plants, lateral gene transfer has spread functional genes involved in metabolism, disease resistance and, abiotic stress tolerance (Dunning et al., 2019). These transfers can have an adaptive effect and bypass the necessary time required for the repeated action of natural selection to adapt the orthologous gene in the recipient species. For example, phosphoenolpyruvate carboxylase (PEPC) is a key enzyme in the C4 photosynthetic pathway that has repeatedly undergone convergent positive selection in independent C4 origins to boost the enzymes efficiency (Christin et al., 2012). The grass Alloteropsis semialata represents a recent C4 origin, and its native PEPC lacks the adaptive amino acid substitutions typically observed in older C4 groups (Christin et al., 2012; Dunning et al., 2017). Given enough time, it is likely the native gene from A. semialata will undergo similar convergent changes as witnessed in other species. However, through LGT certain populations of this species acquired a PEPC that had been adapted for the C4 function in an older C4 lineage of grasses (Christin et al., 2012; Dunning et al., 2017). In this example, LGT acts as an evolutionary shortcut to a novel phenotype. Lateral gene transfer may also allow an organism to evolve beyond its innate capability. The adaptive potential of an organism is limited by its evolutionary history, and the necessary material for specific innovations may be missing from some lineages if certain genes have been lost or never evolved (Blount et al., 2012). Lateral gene transfer can bypass this problem, allowing organisms to reach previously inaccessible areas of the adaptive landscape. For example, the chimeric neochrome photoreceptor for growing in low light was unique to the hornwort lineage, a small group of brypophtes that diverged early during land plant evolution (Li et al., 2014). A gene encoding the bryophyte-restricted neochrome was laterally acquired by ferns over 170 million years ago and is cited as a key innovation underpinning the Cretaceous diversification of this group under the emerging angiosperm canopy (Li et al., 2014). It is possible that ferns may never have been able to exploit this new niche had they not laterally acquired the hornwort gene. Understanding when, where, and why a trait evolves are the major goals of evolutionary biology. Most methods used to address these questions rely on phylogenetic trees, but gene flow between taxa can disconnect the history of the underlying trait from the species tree (Pardo-Diaz et al., 2012; Meier et al., 2017). Genes encoding one of the C4 enzymes (phosphoenolpyruvate carboxykinase [PCK]) have been passed among a number of grass lineages (Dunning et al., 2019), which therefore share the encoded enzyme yet have not evolved it independently. A detailed analysis of the C4 phenotype of the grass A. semialata showed that its C4 photosynthetic machinery is composed of both native genes and LGT (Fig. 1; Dunning et al., 2017). The number of origins therefore differs among the constitutive elements of a single ecological innovation (Fig. 1; Dunning et al., 2017). Relying on species trees to assess the timing and drivers of adaptive evolution of these genes would mislead the analyses, including assessment of the levels of convergent evolution of the genes. The importance of LGT for plant evolution is currently difficult to quantify. Most cases of LGT have been incidentally identified, and few systematic scans have been performed. In the case of parasitic plants, Striga has 34 LGTs identified among 34,577 nuclear genes (Yoshida et al., 2019), and 108 are expressed in the Cuscuta transcriptome (Yang et al., 2019). Similarly, in the grass Alloteropsis semialata, 59 of its 22,043 nuclear genes were identified as LGTs (Dunning et al., 2019). Because of detection difficulties and extreme caution in these studies, these numbers might underestimate the true number of LGTs. Yet, it is clear that LGTs represent a small fraction of all plant genes. The transfer of genes among close relatives via hybridization would however similarly disconnect the species and gene trees. Hybridization is known to be frequent in some groups (Goulet et al., 2017), but the taxonomic scales concerned are very different. Yet, all types of reticulate evolution create conflicts among gene trees, which can be problematic even when studying functional divergence on deep evolutionary scales (Koenen et al., 2020). There is therefore a need for the field of plant macroevolution to consider the effects of reticulations in the tree of life on the sorting of functional traits among lineages, as elegantly done in baobabs (Karimi et al., in press). Over the past decade, it has been increasingly recognized that individuals from the same species do not always share the same genes. In bacteria, this well-known phenomenon lead to the “pangenome” concept where a single species contains both a core set of genes present in all individuals and a dispensable set of genes with a discrete distribution (Tettelin et al., 2005). Similar pangenomes have been reported in several model plant species, with many thousands of genes absent from the original reference genome (reviewed by Tao et al., 2019). We suggest that LGT and hybridization might be important, yet underappreciated, contributors to the functional diversity in plant pangenomes. Transfers of genes into an established species will result in a patchy distribution, where, initially at least, only some populations possess the new genes (Fig. 2; Dunning et al., 2019). Local adaptation might, moreover, lead to variation in LGT content among members of the same populations (Prentice et al., 2015). The first methodological consequence is that the full extent of LGT cannot be estimated with a single reference genome per species. More importantly, in the case of functional genes of adaptive significance, it means that individuals from the same species can differ while those belonging to distinct species can share the same key genetic elements (Fig. 1). This phenomenon could have important consequences on the sorting of morphological and functional traits within and among species of plants, especially in groups where genetic exchanges are frequent. While hybridization effects are widely studied, the consequences of LGT for intraspecific diversity need to be assessed because it can impact conservation, crop improvement programs, and our understanding of functionally relevant taxonomic units. The transfer of functional genes among distantly related groups of plants can mislead our reconstruction of plant evolution by disconnecting the history of relevant genes from species trees. The impact of genetic exchange at the intra- and interspecific levels should thus be considered in comparative botany, as both LGT and introgression are known to occur and move genes of functional importance among plants. Work on plant LGT is, however, still embryonic, and there is a need to evaluate the prevalence of the phenomenon in different groups. As more genomes become available, comparative analyses will be able to answer this question, potentially moving plant LGT from an obscure oddity to be appreciated as an important contributor of adaptive diversification in plants and beyond. The authors thank the Editor-in-Chief Professor Pamela Diggle for the invitation to contribute this essay and the two reviewers for their insightful comments.