Transposon domestication versus mutualism in ciliate genome rearrangements

Abstract

Ciliated protists rearrange their genomes dramatically during nuclear development via chromosome fragmentation and DNA deletion to produce a trimmer and highly reorganized somatic genome. The deleted portion of the genome includes potentially active transposons or transposon-like sequences that reside in the germline. Three independent studies recently showed that transposase proteins of the DDE/DDD superfamily are indispensible for DNA processing in three distantly related ciliates. In the spirotrich Oxytricha trifallax, high copy-number germline-limited transposons mediate their own excision from the somatic genome but also contribute to programmed genome rearrangement through a remarkable transposon mutualism with the host. By contrast, the genomes of two oligohymenophorean ciliates, Tetrahymena thermophila and Paramecium tetraurelia, encode homologous PiggyBac-like transposases as single-copy genes in both their germline and somatic genomes. These domesticated transposases are essential for deletion of thousands of different internal sequences in these species. This review contrasts the events underlying somatic genome reduction in three different ciliates and considers their evolutionary origins and the relationships among their distinct mechanisms for genome remodeling.

Figure 1. Nuclear dimorphism and genome rearrangements in ciliates.

A) From left to right: Oxytricha trifallax, Paramecium tetraurelia, Tetrahymena thermophila. DNA is shown in cyan, yellow represents tubulin staining. Images were kindly provided by Wenwen Fang (Princeton University, Princeton), Kensuke Kataoka (IMBA, Vienna), and Janine Beisson (CNRS, Gif sur Yvette). Abbreviations: i = micronucleus, a = macronucleus. In Oxytricha trifallax, two lobes of a macronucleus are connected by a thin nuclear bridge (not visible in the image). B) Genome rearrangements in all ciliates shown include elimination of micronucleus (MIC)-limited sequences (i, purple IES) and chromosome breakage, which in Tetrahymena occurs at specific chromosome breakage sites (labeled c). After religation of the flanking macronuclear (MAC) sequences, Tetrahymena chromosomes undergo endoreplication to produce 50 identical copies. C) DNA unscrambling in Oxytricha involves the reshuffling and occasional inversion of precursor micronuclear (MIC) sequences (numbered blue boxes) to assemble them in the correct macronuclear order.

Figure 2. TBE transposases in Oxytricha are germline-limited sequences and they participate in their own removal.

The encoded transposases of the Tc1/mariner family have a DDE catalytic motif. Cleavage of the germline-limited sequences starts with a 3 nucleotide 5′ overhang at an ANT recognition site; the second target site serves as the integration site .

Figure 3. Transposases in Tetrahymena and Paramecium belong to the PiggyBac family.

As domesticated transposases, they are present as single-copy genes in the micronuclear and macronuclear genomes. After expression from the somatic genome, they facilitate IES excision from the new macronuclear genome. IES removal occurs via a 4-base 5′ protruding end. In Paramecium, all deleted sequences have a TA dinucleotide at both boundaries, whereas Tetrahymena displays no consensus sequence.

Figure 4. Phylogenetic analysis of representative transposases of the DDE/DDD superfamily.

This tree supports the conclusion that TBE elements belong to the Tc1/mariner superfamily of transposons and also that there appear to be TBE-like elements present in Tetrahymena (labeled “42 kDa transposase”). Additionally, this analysis supports the conclusion that the two PiggyBac-like transposases, Pgm and Tpb2p, in Paramecium and Tetrahymena are homologous to each other. The grouping of the MULE family representative within hAT transposases is unexpected and possibly the result of an alignment artifact due to its disproportionately long sequence. Recently discovered Paramecium transposases Sardine, Thon, AnchoisA, and AnchoisB were omitted because their inclusion in the analysis significantly lowered confidence scores for a majority of branches. The tree was created with MRBayes phylogenetic inference software using the alignment shown in Dataset S1, which was edited to remove regions with gaps in the consensus sequence. The phylogeny was generated using a mixed amino acid substitution model and invariable gamma distribution rate model over 200,000 iterations with a burn-in of 25%. Branch confidence values represent conditional probabilities generated by the Bayesian inference process. The scale bar corresponds to 0.05 expected substitutions per site of the unmasked alignment positions. Domain and motif annotations were produced using the Pfam web server .