The retrotransposon Ylt1 was described previously from the genome of the dimorphic fungus Yarrowia lipolytica. Remarkably, Ylt1 is currently the largest LTR retrotransposon reported from fungal genomes. However, little was known about its biology and its interactions with host genome. So, the aim of this work was the characterization of properties of Ylt1.Analysis of proteins encoded by Ylt1 (Gag protein and integrase) was carried out during this work. To enable their detection, both proteins were tagged with HA epitopes. The sizes of Gag protein and putative precursors of Gag protein and integrase were estimated, and a model for the proteolytic processing of the polyprotein of Ylt1 was proposed. It was shown that Gag protein of Ylt1 is about 2-fold larger than Gag proteins of other studied yeast retrotransposons. An analysis of Ylt1 expression was also performed. Production of the Ylt1 Gag protein under different conditions was analyzed by Western blotting. Expression of Ylt1 occurred on all tested carbon sources. The amount of Ylt1 decreased rapidly upon transition to stationary growth phase, in the presence of copper sulfate and under heat shock conditions. It is suggested that Ylt1 is expressed in actively growing cells, whereas stress conditions have a negative impact on its expression. Such expression pattern was not previously reported for other yeast retrotransposons. Activity of Ylt1 in vivo was characterized using an Ylt1 elements tagged with SUC2 gene of Saccharomyces cerevisiae. Mobilization of the marked Ylt1 element and its transposition from autonomous plasmid into host genome was observed in performed experiments. Obtained results strongly support the idea that Ylt1 is transpositionally active. Formation of tandem repeats by newly inserted Ylt1 elements was observed in several cases. It is suggested that integrase function was affected in this case, and that the integration was mediated by homologous recombination instead. Analysis of the Ylt1 insertion specificity and of the Ylt1 distribution in the genome of Y. lipolytica E150 was done. The remarkable sequence specificity of Ylt1 insertions, which is unusual for LTR retrotransposons, was revealed during this analysis. Also, it was shown that Ylt1 insertions are found mainly in intergenic regions, often at a significant distance (>500 bp) from the next reading frame. No association of Ylt1 insertions with tRNA genes was observed. Searches for Ylt1-related elements in the Y. lipolytica genome database were performed. The novel Ty3/gypsy element Tyl6 was found in the genome of Y. lipolytica E150. The sequence analysis of this element was carried out. It was shown that structural properties of Tyl6 resemble the properties of the Ty3 element of S. cerevisiae. However, two reading frames of Tyl6 (gag and pol) are separated by -1 frame-shift, which was not previously reported for retrotransposons of hemiascomycetous yeasts. Phylogenetic analysis placed Tyl6 within chromoviruses, and the Tse3 element of S. exiguus was shown to be the closest relative of Tyl6. The distribution of Tyl6 among Y. lipolytica strains was analyzed. Interestingly, the novel element was found only in strains derived from the strain YB423-12. The strains of independent origin included in the analysis were shown to be Tyl6-free. The same distribution was previously reported for the retrotransposon Ylt1 and for the DNA transposon Mutyl. Two models of the evolution of transposable elements in Y. lipolytica genome were proposed based on these results.

The retrotransposon Ylt1 was described previously from the genome of the dimorphic fungus Yarrowia lipolytica. Remarkably, Ylt1 is currently the largest LTR retrotransposon reported from fungal genomes. However, little was known about its biology and its interactions with host genome. So, the aim of this work was the characterization of properties of Ylt1.Analysis of proteins encoded by Ylt1 (Gag protein and integrase) was carried out during this work. To enable their detection, both proteins were tagged with HA epitopes. The sizes of Gag protein and putative precursors of Gag protein and integrase were estimated, and a model for the proteolytic processing of the polyprotein of Ylt1 was proposed. It was shown that Gag protein of Ylt1 is about 2-fold larger than Gag proteins of other studied yeast retrotransposons. An analysis of Ylt1 expression was also performed. Production of the Ylt1 Gag protein under different conditions was analyzed by Western blotting. Expression of Ylt1 occurred on all tested carbon sources. The amount of Ylt1 decreased rapidly upon transition to stationary growth phase, in the presence of copper sulfate and under heat shock conditions. It is suggested that Ylt1 is expressed in actively growing cells, whereas stress conditions have a negative impact on its expression. Such expression pattern was not previously reported for other yeast retrotransposons. Activity of Ylt1 in vivo was characterized using an Ylt1 elements tagged with SUC2 gene of Saccharomyces cerevisiae. Mobilization of the marked Ylt1 element and its transposition from autonomous plasmid into host genome was observed in performed experiments. Obtained results strongly support the idea that Ylt1 is transpositionally active. Formation of tandem repeats by newly inserted Ylt1 elements was observed in several cases. It is suggested that integrase function was affected in this case, and that the integration was mediated by homologous recombination instead. Analysis of the Ylt1 insertion specificity and of the Ylt1 distribution in the genome of Y. lipolytica E150 was done. The remarkable sequence specificity of Ylt1 insertions, which is unusual for LTR retrotransposons, was revealed during this analysis. Also, it was shown that Ylt1 insertions are found mainly in intergenic regions, often at a significant distance (>500 bp) from the next reading frame. No association of Ylt1 insertions with tRNA genes was observed. Searches for Ylt1-related elements in the Y. lipolytica genome database were performed. The novel Ty3/gypsy element Tyl6 was found in the genome of Y. lipolytica E150. The sequence analysis of this element was carried out. It was shown that structural properties of Tyl6 resemble the properties of the Ty3 element of S. cerevisiae. However, two reading frames of Tyl6 (gag and pol) are separated by -1 frame-shift, which was not previously reported for retrotransposons of hemiascomycetous yeasts. Phylogenetic analysis placed Tyl6 within chromoviruses, and the Tse3 element of S. exiguus was shown to be the closest relative of Tyl6. The distribution of Tyl6 among Y. lipolytica strains was analyzed. Interestingly, the novel element was found only in strains derived from the strain YB423-12. The strains of independent origin included in the analysis were shown to be Tyl6-free. The same distribution was previously reported for the retrotransposon Ylt1 and for the DNA transposon Mutyl. Two models of the evolution of transposable elements in Y. lipolytica genome were proposed based on these results.

Retrotransposons are major components of plant genomes influencing their genome size, organization and evolution. In the frame of this work, retrotransposons of the Beta vulgaris genome have been identified by molecular methods and whole genome bioinformatics approaches. Neither belonging to the rosids nor asterids, B. vulgaris (cultivated beet including sugar beet, beet root and mangold) is taxonomically placed at a key position at the root of the core eudicots, and considerably different from traditional plant model species such as thale cress or rice. Its genome has been sequenced, and annotation is under way. In order to compare different evolutionary lineages of B. vulgaris retrotransposons, long terminal repeat (LTR) and non-LTR retrotransposon family have been analyzed in detail. Full-length members have been isolated and characterized by bioinformatics, Southern and fluorescent in situ hybridization. Hallmarks of the LTR retrotransposon family Cotzilla are an additional env-like open reading frame (ORF), homogeneity of the members and the very high abundance. Most family members are evolutionarily young, and have most likely been created during recent bursts of amplification during species radiation. In contrast, the non-LTR retrotransposon family BNR has fewer copies and is much more diverged. Although the BNR ORF2 resembles previously analyzed long interspersed nuclear elements (LINEs) of the L1 clade, its ORF1 sequence differs strongly. It lacks the zinc finger domain described for plant LINEs, but contains instead an RNA recognition motif (RRM) likely to have an RNA-binding function. Database searches revealed the presence of similar LINE families in higher plant genomes such as poplar, lotus and soybean. Comparing their reverse transcriptase regions with other retrotransposons, these BNR-like LINEs form a separate group of L1 LINEs designated as BNR subclade. Availability of the B. vulgaris genome sequence allowed retrotransposon analyses on a genome-wide scale. A Hidden Markov Model-based detection algorithm has been developed in order to retrieve retrotransposon information directly from the database. Nearly 6000 B. vulgaris reverse transcriptase sequences have been isolated and classified into LTR retrotransposons of the Ty3-gypsy and Ty1-copia type, and non-LTR retrotransposons of the LINE type. As a result, a comprehensive overview of the retrotransposon spectrum of the B. vulgaris genome has been generated. Since plant LINEs have been only rarely investigated, the B. vulgaris LINE composition was studied in detail. Out of 28 described LINE clades, only members of the L1 and RTE clades have been identified. Based on a minimal shared sequence identity of 60 %, they form at least 17 L1 families and one RTE family. Full-length members of all investigated L1 families have been analyzed regarding their sequence, structure and diversity. In order to transfer the algorithm tested in B. vulgaris to other angiosperm genomes, twelve additional plant genomes have been queried for LINE reverse transcriptases. Key finding is the presence of only two LINE clades (L1 and RTE) in the analyzed genomes of higher plants. Whereas plant L1 LINEs are highly diverse and form at least seven subclades with members across species borders, RTE LINEs are extremely homogenized and constitute most likely only a single family per genome. In summary, this work’s results help to gain an understanding of the different strategies of retrotransposon evolution in plants, whereas the generated data directly contributes to the B. vulgaris genome annotation project. / Retrotransposons sind eine wesentliche Komponente von Pflanzengenomen, die sowohl die Größe und Organisation als auch die Evolution dieser Genome wesentlich beeinflussen können. Im Rahmen dieser Arbeit wurden verschiedene Gruppen von Retrotransposons des Beta vulgaris Genoms mittels molekularer und bioinformatischer Methoden identifiziert. Innerhalb der dikotyledonen Blütenpflanzen gehört B. vulgaris (kultivierte Rübe einschließlich Zuckerrübe, Roter Beete und Mangold) weder zu den Rosiden noch zu den Asteriden, sondern nimmt eine Schlüsselposition innerhalb der Kerneudikotyledonen ein. Somit zeigt das Rübengenom wesentliche Unterschiede zu traditionellen Modellpflanzen wie Arabidopsis thaliana oder Oryza sativa. Das Genom ist bereits sequenziert, die Annotation jedoch noch nicht abgeschlossen. Um verschiedene evolutionäre Linien von B. vulgaris Retrotransposons vergleichend zu untersuchen wurden insbesondere Long Terminal Repeat (LTR)- und Non-LTR-Retrotransposon-Familien detailliert analysiert. Vollständige Mitglieder wurden isoliert und mittels bioinformatischer Methoden, Southern- und Fluoreszenz-in situ-Hybridisierung untersucht. Die LTR-Retrotransposon-Familie Cotzilla ist durch einen zusätzlichen env-ähnlichen offenen Leserahmen (ORF), Homogenität ihrer Mitglieder und eine hohe Abundanz gekennzeichnet. Die meisten Cotzilla-Kopien sind evolutionär jung und wurden wahrscheinlich innerhalb eines kurzen Zeitraumes während der Artentstehung stark amplifiziert. Im Gegensatz zur Cotzilla-Familie besitzt die Non-LTR-Retrotransposon-Familie BNR weniger Kopien und ist wesentlich divergierter. Während der BNR-spezifische ORF2 starke Ähnlichkeiten zu anderen pflanzlichen Long Interspersed Nuclear Elements (LINEs) der L1-Klade aufweist, unterscheidet sich der BNR ORF1 von diesen sehr stark. Im Gegensatz zu bereits beschrieben pflanzlichen LINEs kodiert er kein Zinkfingermotiv, sondern substituiert dieses durch ein RNA-Erkennungsmotiv (RRM). Durch Datenbanksuche konnten BNR-ähnliche LINEs in den Genomen höherer Pflanzen wie Soja, Lotus und Pappel identifiziert werden. Ein Vergleich der entsprechenden Reversen Transkriptasen (RT) mit den RTs anderer Retrotransposons zeigt, dass die BNR-ähnlichen LINEs eine separate Gruppe innerhalb der L1 LINEs bilden. Diese wurde daher als BNR-Subklade definiert. Die Untersuchung von Retrotransposons auf Genomebene wurde durch die B. vulgaris Genomsequenz ermöglicht. Um Retrotransposon-Informationen direkt aus dem Genom zu extrahieren, wurde ein Hidden Markov Modell (HMM)-basierter Detektions-algorithmus entwickelt. Annähernd 6000 B. vulgaris Reverse Transkriptase-Sequenzen konnten identifiziert und in LTR-Retrotransposons des Ty3-gypsy- beziehungsweise des Ty1-copia-Typs und in Non-LTR-Retrotransposons des LINE-Typs klassifiziert werden. Somit wurde ein umfassender Überblick über die Bandbreite der B. vulgaris Retrotransposons arhalten. Da pflanzliche LINEs bisher nur wenig erforscht sind, wurde die B. vulgaris LINE Zusammensetzung genauer untersucht. Von 28 beschriebenen LINE-Kladen konnten nur Mitglieder der L1- und der RTE-Klade identifiziert werden. Basierend auf einer Identität von mindestens 60 % bilden die Sequenzen 17 L1 Familien und eine RTE Familie. Vollständige Mitglieder aller L1 Familien wurden hinsichtlich ihrer Sequenz, Struktur und Diversität analysiert. Um den in B. vulgaris getesteten HMM-basierten Algorithmus auf andere Angiospermengenome zu übertragen, wurden zwölf weitere Pflanzengenome auf das Vorhandensein von LINE-spezifischen Reversen Transkriptasen untersucht. Wesentlichstes Ergebnis ist der Nachweis von nur zwei LINE-Kladen (L1 und RTE) in höheren Pflanzen. Während pflanzliche L1 LINEs hochgradig divers sind und über Artgrenzen hinaus mindestens sieben Subkladen mit Vertretern verschiedener Pflanzen bilden, sind RTE LINEs extrem homogenisiert und stellen höchstwahrscheinlich nur eine einzelne Familie pro Genom einer Art dar. Zusammenfassend ermöglichen die Ergebnisse dieser Arbeit eine Erweiterung des Verständnisses der unterschiedlichen Evolutionsstrategien von Retrotransposons in Pflanzen. Zusätzlich tragen die gewonnen Daten zur Annotation des B. vulgaris Genoms bei.