The Evolutionary Volte-Face of Transposable Elements: From Harmful Jumping Genes to Major Drivers of Genetic Innovation

Abstract

Transposable elements (TEs) are self-replicating DNA elements that constitute major fractions of eukaryote genomes. Their ability to transpose can modify the genome structure with potentially deleterious effects. To repress TE activity, host cells have developed numerous strategies, including epigenetic pathways, such as DNA methylation or histone modifications. Although TE neo-insertions are mostly deleterious or neutral, they can become advantageous for the host under specific circumstances. The phenomenon leading to the appropriation of TE-derived sequences by the host is known as TE exaptation or co-option. TE exaptation can be of different natures, through the production of coding or non-coding DNA sequences with ultimately an adaptive benefit for the host. In this review, we first give new insights into the silencing pathways controlling TE activity. We then discuss a model to explain how, under specific environmental conditions, TEs are unleashed, leading to a TE burst and neo-insertions, with potential benefits for the host. Finally, we review our current knowledge of coding and non-coding TE exaptation by providing several examples in various organisms and describing a method to identify TE co-option events.

Keywords: epigenetics; exaptation; silencing; transposable elements.

Figure 2

A model to explain the occurrence of TE exaptation. This hypothetical model describes the exaptation of a class II TE-derived transposase. At stage 0, TEs are repressed by silencing pathways to ensure host genome integrity (1). Nevertheless, as purifying selection and genetic drift negatively impact the transgenerational persistence of TEs within the host genome, these latter might use anti-silencing strategies (2) to maintain some basal level of activity (3), and to be inherited as functional elements. Among the pool of TEs, one of them can be considered as an ETE precursor. At stage 1, upon, for instance, the perception of an environmental stimulus, such as heat, this TE is activated. As a young and functional element, it can transpose and mobilize in new genomic regions (neo-insertions). At stage 2, the neo-inserted TE is expressed, producing transposase proteins that will recognize TE-derived CREs genome-wide and potentially recruit additional chromatin factors to regulate host genes in trans. At this stage, although the neo-inserted TE is still capable of mobilization, it can enter selective pressure when providing positive advantage to the host. However, this feature must be followed by a swift transition to stage 3, in which the neo-inserted TE becomes immobilized. At stage 3, the neo-inserted TE is immobilized by mutations occurring in its TIRs (red stripes), turning into an ETE gene. Undoubtedly, TE immobilization is essential in the process of TE exaptation. Although TE immobilization is an absolute prerequisite for TE exaptation, it is not sufficient. ETE genes must remain stably expressed, most likely because environmental conditions persist. The ETE gene still undergoes positive selection and CREs that are not conferring adaptive traits are negatively selected. At stage 4, ETE gene and CREs co-evolve, accumulating mutations (*) under purifying selections, and together create a new host gene regulatory network that will be perpetuated only if the environmental changes persist. At stage 5, the ETE gene can be considered a genuine host gene, providing phenotypic value. Following relaxed selection pressure, the TE-derived sequence may inherit point mutations (*), or gene duplication may occur. ETE diversification and neo-functionalization ensure “long-term” positive adaptation of ETE sequence to ultimately create genetic variants with new cellular functions.

Figure 1

Several silencing pathways cooperate to silence TEs in A. thaliana. (A). Initiation and maintenance of DNA methylation involve specific DNA methyltransferases to perform 5mC in three cytosine contexts. (B). Histone modifications such as methylation of histone 3 at lysine 9 or lysine 27 involve specialized histone methyltransferases. Deacetylation of histone tails requires HDAC proteins, such as HDA6. (C). DDM1 is involved in the deposition of the heterochromatic histone variant H2A.W, which interferes with linker histone H1 occupancy at constitutive heterochromatin. The connections between DDM1, H3K27me3 and 5mC remain unclear. (D). PMD proteins MAIN and MAIL1 interact with PP7L, forming a complex involved in TE silencing through an unknown mechanism. (E). SMC4 cooperates with other epigenetic pathways to repress TEs, presumably by promoting high-order chromatin organization. (F). MOM1 interacts with PIAL proteins to repress TEs using an elusive process. (G). MORC proteins promote constitutive heterochromatin compaction to ensure TE silencing by topological DNA loop trapping mechanism. CC: coiled-coil domain (H). The SILENZIO factor interacts with MDB5 and MDB6 and HSP factors to read mCG and silence TEs. Crosstalk and synergistic effects have been described between several epigenetic pathways.

Figure 3

From single locus positive selection of TE-derived sequences to the elaboration of complex regulatory networks. (A) Co-option of a TE gene can positively impact the host through the regulation of specific gene with the ETE proteins acting as transcriptional activators or repressors, most likely by recruiting other chromatin factors (not represented here). At the genomic scale, ETE proteins can participate in large regulatory networks ensuring the fine tuning of gene expression. (B) Co-option of non-coding TE sequences acting as CREs can regulate the expression of single locus through proximal or distal interactions with the transcriptional machinery. Genome-wide, these TE-derived CREs can be part of interconnected gene regulatory networks. TFBS: transcription factor (TF)-binding sites. Co-option of non-coding TE sequences acting as ncRNAs can either target a discrete host gene as siRNAs or produce lncRNAs with trans regulatory functions. In the case of SINE Alu elements located in the 3′UTR of host genes, they can form IRAlu secondary structure of mRNAs involved in splicing, mRNA subcellular localization and other processes. In a hypothetical model, TE-derived siRNAs can engage in complex regulatory networks repressing several host genes.