Downloaded from genome.cshlp.org on September 29, 2021 - Published by Cold Spring Harbor Laboratory Press 1 2 3 4 5 6 7 Diversification and collapse of a telomere elongation mechanism 8 9 10 Bastien Saint-Leandre1,2, Son C. Nguyen2,3, and Mia T. Levine1,2* 11 12 13 1Department of Biology, 2Epigenetics Institute, University of Pennsylvania, Philadelphia, PA 14 3Department of Genetics, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, 15 19104. 16 17 18 19 20 Running title: diversification of Drosophila telomeres 21 22 Keywords: telomere, retrotransposon, domestication, Drosophila 23 24 *Corresponding Author: 25 Mia Levine, Department of Biology, University of Pennsylvania, 433 S. University Avenue, Philadelphia, 26 PA 19104. p: 215-573-9709 e: [email protected]. 1 Downloaded from genome.cshlp.org on September 29, 2021 - Published by Cold Spring Harbor Laboratory Press 27 ABSTRACT 28 In most eukaryotes, telomerase counteracts chromosome erosion by adding repetitive sequence to 29 terminal ends. Drosophila melanogaster instead relies on specialized retrotransposons that insert 30 exclusively at telomeres. This exchange of goods between host and mobile element—wherein the mobile 31 element provides an essential genome service and the host provides a hospitable niche for mobile element 32 propagation—has been called a ‘genomic symbiosis’. However, these telomere-specialized, jockey family 33 retrotransposons may actually evolve to ‘selfishly’ over-replicate in the genomes that they ostensibly 34 serve. Under this model, we expect rapid diversification of telomere-specialized retrotransposon lineages 35 and possibly, the breakdown of this ostensibly symbiotic relationship. Here we report data consistent with 36 both predictions. Searching the raw reads of the 15-million-year-old melanogaster species group, we 37 generated de novo jockey retrotransposon consensus sequences and used phylogenetic tree-building to 38 delineate four distinct telomere-associated lineages. Recurrent gains, losses, and replacements account for 39 this retrotransposon lineage diversity. In D. biarmipes, telomere-specialized elements have disappeared 40 completely. De novo assembly of long reads and cytogenetics confirmed this species-specific collapse of 41 retrotransposon-dependent telomere elongation. Instead, telomere-restricted satellite DNA and DNA 42 transposon fragments occupy its terminal ends. We infer that D. biarmipes relies instead on a 43 recombination-based mechanism conserved from yeast to flies to humans. Telomeric retrotransposon 44 diversification and disappearance suggest that persistently ‘selfish’ machinery shapes telomere elongation 45 across Drosophila rather than completely domesticated, symbiotic mobile elements. 46 2 Downloaded from genome.cshlp.org on September 29, 2021 - Published by Cold Spring Harbor Laboratory Press 47 INTRODUCTION 48 Transposable elements (TEs) infest eukaryotic genomes, ever-evolving to increase in copy number over 49 time (Feschotte and Pritham 2007; Beauregard et al. 2008). These so-called ‘selfish genetic elements’ 50 enhance their own transmission relative to other elements in the genome, imposing neutral or deleterious 51 consequences on the host (Werren 2011). Deleterious consequences arise when TE insertions disrupt host 52 genes (Hancks and Kazazian 2012), nucleate local epigenetic silencing (Slotkin and Martienssen 2007; 53 Lee and Karpen 2017), and trigger catastrophic recombination between non-homologous genomic regions 54 (Langley et al. 1988; Beck et al. 2011). TEs also provide raw material for genome adaptation (Jangam et 55 al. 2017): across eukaryotes, host genomes re-purpose TE-derived sequence for basic cellular and 56 developmental processes, from immune response (van de Lagemaat et al. 2003) to placental development 57 (Lynch et al. 2015) to programmed genome rearrangements (Cheng et al. 2010; Cheng et al. 2016). These 58 diverse ‘molecular domestication’ events share a common feature—the degeneration or deletion of the 59 TE’s capacity to propagate. Consequently, the TE-derived sequence resides permanently at a single 60 genome location, just like any other host gene sequence. Inability to increase copy number via 61 transposition resolves prior conflict of interest between the host and the TE (Jangam et al. 2017). 62 However, not all adaptive molecular domestication events necessitate TE immobilization; in rare cases, 63 essential host functions rely on retention of the mobilization machinery that promotes recurrent TE 64 insertions into host DNA. The non-canonical telomere elongation mechanism of Drosophila is exemplary 65 (Pardue and DeBaryshe 2003; Casacuberta 2017). 66 67 In most eukaryotes beyond Drosophila, telomerase-added DNA repeats (Greider and Blackburn 1989; 68 Zakian 1989; Blackburn 1991; Zakian 1996) counteract the ‘end-replication problem’ that otherwise 69 erodes unique DNA sequence at chromosome termini (Watson 1972). However, telomeres of select fungi 70 (Starnes et al. 2012), algae (Higashiyama et al. 1997), moths (Osanai-Futahashi and Fujiwara 2011), 71 crustaceans (Gladyshev and Arkhipova 2007), and DNA repair-deficient mammalian cells (Morrish et al. 72 2007) encode not only telomerase-added repeat elements but also TEs that insert preferentially at 3 Downloaded from genome.cshlp.org on September 29, 2021 - Published by Cold Spring Harbor Laboratory Press 73 chromosome termini. These telomeric mobile elements are typically derived from a single class of TE— 74 the non-long terminal repeat (‘non-LTR’) retrotransposons (Beck et al. 2011), which mobilize via reverse 75 transcription and insertion into new genomic locations. The most extreme example of this co-option is 76 found in Drosophila melanogaster, whose telomeres harbor no telomerase-added repeats. In fact, the 220 77 million-year-old ‘true fly’ insect Order, Diptera, completely lacks the genes encoding the telomerase 78 holoenzyme (Pardue and DeBaryshe 2003; Casacuberta 2017). Instead, D. melanogaster harbors three 79 telomere-specialized retrotransposons–HeT-A, TART, and TAHRE—that preserve distal, unique 80 sequence (Pardue and DeBaryshe 2011). These three retrotransposons represent a monophyletic clade 81 within the larger ‘jockey’ element family (Villasante et al. 2007), whose members are more typically 82 found along chromosome arms (Xie et al. 2013). The telomere-specialized jockey subclade, in contrast, 83 rarely inserts outside the telomere (Pardue and DeBaryshe 2003; Berloco et al. 2005; Pardue and 84 DeBaryshe 2011). 85 86 This molecular domestication of still-mobile retrotransposons into an essential genome function is often 87 referred to as a ‘genomic symbiosis’ (Pardue and DeBaryshe 2008). Evidence for such a mutualism is 88 compelling. The elements that maintain telomere ends in D. melanogaster comprise a monophyletic clade 89 ostensibly specialized to replicate only at chromosome ends (Pardue and DeBaryshe 2008). Elements 90 from this jockey clade also appear at terminal ends in distant Drosophila species, consistent with a single 91 domestication event >40 million years ago followed by faithful vertical transmission (Casacuberta and 92 Pardue 2003; Villasante et al. 2007). Moreover, mobile elements from other families rarely appear at 93 terminal ends (Mason and Biessmann 1995; Biessmann et al. 2005; Mason et al. 2016). These data 94 suggest that retrotransposon-mediated chromosome elongation represents a long-term, conserved 95 relationship between a host genome and its domesticated, but still-mobile, retrotransposons. 96 97 This picture of cooperativity was complicated by the discovery that the telomere-associated subclade of 98 jockey elements evolves rapidly across Drosophila. Leveraging 12 Drosophila genomes that span 40 4 Downloaded from genome.cshlp.org on September 29, 2021 - Published by Cold Spring Harbor Laboratory Press 99 million years of evolution, Villasante, Abad, and colleagues detected at least one jockey-like, candidate 100 telomeric retrotransposon in all 12 species (Villasante et al. 2007). These data implicated a single 101 evolutionary event in a common ancestral sequence that conferred telomere-specificity. However, these 102 elements represent distinct phylogenetic lineages rather than species-specific versions of the HeT-A, 103 TART, and TAHRE elements well-studied in D. melanogaster. For example, D. melanogaster and its 104 close relatives encode HeT-A, TART, and TAHRE, while the 15 million year-diverged D. ananassae 105 encodes a single, phylogenetically distinct candidate retrotransposon lineage, ‘TR2’. The 30 million-year 106 diverged D. pseudoobscura species encodes yet other phylogenetically distinct lineages within this jockey 107 subclade. This expansive evolutionary lens revealed a previously unappreciated, dynamic evolutionary 108 history of these jockey subclade retrotransposons (Villasante et al. 2008). However, the large evolutionary 109 distances between species left fine scale dynamics unknown and telomere-specific localization was not 110 explored (Villasante et al. 2007). The evolutionary origin(s), ages, between-species differences, and 111 genome locations of these candidate telomere elongators remain obscure. Elucidating the evolutionary 112 history of these elements is essential to address the possibility that this molecular domestication event is 113 less a stable, long-term genomic symbiosis and instead an ever-evolving relationship between the 114 domesticator and the domesticated. Here we investigate the