Mireille Bétermier, Mick Chandler, Gaël Cristofari
The session on transposition mechanisms included talks, which addressed cellular control of transposon integration, transposase oligomerization and the control of transposition, mechanisms of reverse transcription and retrotransposition and transposase domestication and genome dynamics.
Vincent Parissi (CNRS, University of Bordeaux, Bordeaux, France) presented studies on the integration of human immunodeficiency virus (HIV) and prototype foamy virus (PFV) retroviruses. He described the effect of nucleosome positioning in vitro on concerted (double end) integration using purified components. HIV integration into stable nucleosomal regions was reduced, but this inhibition was partially alleviated by the ATP-dependent chromatin-remodeling complex SWI/SNF known to interact with the retroviral integrase IN. This is consistent with high throughput sequencing studies indicating that HIV integration is enriched in nucleosome-poor regions. On the other hand, integration of PFV was shown to be significantly less sensitive than HIV to the presence of stable nucleosomes. This pattern is consistent with the DNA bending (observed in structural studies) required for PFV integration but not for that of HIV.
Control of integration was also addressed by Bernard Hallet (University of Louvain, Louvain la Neuve, Belgium), for the bacterial Tn4430 transposon, a member of the Tn3 family whose transposase, like retroviral IN, belongs to the DDE superfamily. Tn4430 transposes using a replicative cointegration mechanism and its insertion appears tightly coupled to target DNA replication. Slowing down the target replication fork results in preferential upstream insertions. Moreover, Tn4430 transposase binds with high affinity to artificial DNA forks in vitro and uses these as a specific target for joining the transposon ends. The data suggest a ‘replication fork hijacking’ mechanism whereby Tn4430 would recruit the cellular replication machinery by jumping into replication intermediates.
The theme was also taken up by Bao Ton-Hoang (CNRS, Laboratoire de Microbiologie et Génétique Moléculaires, Toulouse, France), who showed that the single strand DNA insertion sequence, IS608, which uses tyrosine rather than DDE chemistry for transposition, is excised from and inserts preferentially into the lagging strand template of both plasmid and chromosomal replication forks. Its transposase, TnpA, also localizes to blocked replication forks in vivo and preferentially binds branched DNA structures in vitro, such as forks, D-loops and Holliday junctions, suggesting a mechanism for assuring transposition activity at the fork.
Two presentations addressed the role of transposase oligomerization in regulating transposition. Ronald Chalmers (University of Nottingham, Nottingham, UK) described a model explaining overproduction inhibition (OPI), a phenomenon observed for several eukaryotic transposons where high levels of transposase result in a reduction in transposition activity. Using the human mariner TE, Hsmar1, he provided strong biochemical support for a model in which the key controlling element is a transposase dimer, which first binds a single transposon end via one of the component monomers before using the second to bind the other end to form an active transpososome. OPI is proposed to arise from occupation of each end of the transposon by a transposase dimer, preventing formation of the transpososome. The model was rigorously tested using computer simulations. It can be used to explain the strong tendency for TEs to undergo decay in eukaryotes.
Fred Dyda (National Institutes of Health, Bethesda, MD, USA) presented a set of crystallographic structures of the transposase of another eukaryotic TE, Hermes belonging to the hAT family. He identified an octomeric unit in which four dimers are intertwined via their C-terminal domains. This was found to have limited activity in vitro. However, the structure predicts that removal of an α-helix should destroy the interface holding the octomer together and result in the formation of transposase dimers. It was observed that such deletions render the transposase hyperactive in vitro. Although the exact details remain to be elucidated, it seems probable that the ‘closed’ octomeric form may represent a downregulated transposase species.
A third theme of this session was the mechanism(s) involved in reverse transcription and retrotransposition. Thomas Eickbush (University of Rochester, Rochester, NY, USA) presented the results of studies of R2 non-LTR retrotransposons. These elements integrate in a sequence-specific way into 28S RNA genes by target-primed reverse transcription (TPRT). R2 RNA, the transposition intermediate, is processed from a 28S/R2 cotranscript by a self-cleaving ribozyme located at the 5' R2 RNA end. Eickbush has now shown that this activity is present in all R2 elements from Drosophila to hydra. The position of cleavage varies from organism to organism and determines insertion site selectivity. When cleavage occurs within the 28S RNA moiety, subsequent insertion into a 28S DNA target is precise - presumably due to base pairing between the 28S gene and the remaining 28S RNA sequence attached to the R2 transcript. For those organisms in which cleavage occurs at the exact R2-28S RNA junction, insertion shows addition of non-templated nucleotides. Eickbush also localized by mutagenesis the RNA binding domain of the R2 protein in a region upstream of the reverse transcriptase domain conserved in other non-LTR retrotransposons and in telomerase.
In contrast to R2, mammalian LINEs, such as L1, do not integrate in a specific locus. Most L1 insertions occur in the highly frequent A/TTTT motif. However, individual sites are often degenerate and contain much longer stretches of AT-rich sequences. The L1 machinery is a ribonucleoprotein particle (RNP), which contains an additional RNA-binding protein, ORF1p, encoded by the L1 element. Clement Monot (laboratory of Gaël Cristofari, Inserm, CNRS, University of Nice-Sophia-Antipolis, Nice, France) has developed a direct reverse transcriptase assay with native L1 RNP for studying the initiation of reverse transcription. Using this system, he defined the preferential rules of L1 reverse transcription initiation. He showed that efficient priming can be achieved with as little as four matching nucleotides at the primer 3' end, but also that the L1 RNP can tolerate terminal mismatches if compensated by an increased number of upstream matching nucleotides. Based on these data, he proposed that the terminal bases of the primer act as a specific snap and the upstream ones as a weaker ‘velcro strap’ allowing efficient and flexible retrotransposition into imperfect AT-rich regions as observed in mammalian genomes.
Alan Schulman (Institute of Biotechnology, Helsinki, Finland) described in vivo studies with the BARE LTR-retrotransposons. These elements produce several RNA populations: one that is capped, polyadenylated, and translated but cannot be reverse transcribed; another that is not capped or polyadenylated, but packaged into virus-like particles (VLPs) and reverse transcribed; and a third which is capped, polyadenylated and spliced to produce high amounts of Gag, the capsid protein that forms the VLPs. The relative amount of the spliced and unspliced forms varies from tissue to tissue. These data highlight a unique situation among the retroelements where distinct RNA pools are committed to translation (with or without splicing) or reverse transcription depending on post-transcriptional processing.
The final theme of the session centered around transposase domestication and genome dynamics. Bao Ton-Hoang (Laboratoire de Microbiologie et Génétique Moléculaires, CNRS, Toulouse, France) presented evidence that an IS608-related tyrosine transposase has evolved to manage repeated extragenic palindromes (REP) sequences which are present in many bacteria in very high copy number and are involved in genome structure and gene expression. In Escherichia coli, only a single copy of the gene, tnp AREP, is present. It is located in an identical position in all strains. Phylogenetic evidence suggested that the gene arrived early in the radiation of E.coli and was later lost in some of the present clades. Purified TnpAREP protein exhibited catalytic activity. It is capable of sequence-specific cleavage and strand transfer of REP sequences. This would provide one of the first examples of transposase domestication in a prokaryote.
Perhaps one of the most important examples of transposase domestication is that of the RAG proteins involved in generating V(D)J diversity. The RAG ancestor is thought to resemble the transposase of Transib. Nancy Craig (Johns Hopkins University School of Medicine, Baltimore, MD, USA) presented a functional analysis of the Transib transposase and demonstrated that, like V(D)J recombination, Transib transposition passes through an intermediate involving formation of a DNA hairpin on the flanking DNA.
The V(D)J system was addressed by Martin Gellert (National Institutes of Health, Bethesda, MD, USA), who presented data confirming that the ‘signal end’ complex within which the recombination reactions occur includes RAG1, RAG2, HMGB1 in a 2:2:1 stoichiometry. He also provided structural information from electron microscopy EM studies. This indicated a parallel anchor-shaped RAG1/2 complex with approximately 2-fold symmetry in which RAG2 is located at the head of anchor, RAG1 N-terminus at the ‘shank’ end, along with DNA chains beyond nonamers. Functional experiments showed that there is autoinhibition by interaction of RAG1 and RAG2 C-termini and that autoinhibition can be alleviated by binding a histone H3 tail peptide containing trimethylated lysine 4 (H3K4me3) and known to bind the PHD domain of RAG2 and to target it to recombinationally active loci. This is possibly the first known case where chromatin tethering activates an enzyme.
Ciliates have recently provided a novel illustration of the role played by domesticated transposases in developmentally programmed genome rearrangements. Alexander Vogt (laboratory of Kazufumi Mochizuki, Institute of Molecular Biotechnology (IMBA), Vienna, Austria), described Tpb2p, a domesticated piggyBac-like transposase essential for the excision of internal eliminated sequences (IESs), an obligatory genome rearrangement in the Tetrahymena life cycle. The enzyme introduces a 4-bp staggered cleavage at an IES boundary in vitro. Mutagenesis of the boundary sequence revealed a crucial role for positions 2 and 3 after the cut (in vitro and in vivo). In addition, IES-specific heterochromatin seems to control cleavage site accuracy through a possible interaction between Tpb2p and H3K9me3. Thus both boundary sequences and heterochromatin interactions are probably involved in specifying precise IES excision.
Mireille Bétermier (CNRS, Centre de Génétique Moléculaire, Gif-sur-Yvette, France) presented the analysis of a genome-wide set of IESs identified in Paramecium by high-throughput sequencing of DNA extracted from cells depleted in PiggyMac, the Tpb2p homolog responsible for IES excision in this ciliate. A vast majority of Paramecium IESs (93%) are shorter than 150 bp. Among the longest IESs, recognizable fragments of Tc1-like transposons were identified, indicating that PiggyMac excises DNA sequences unrelated to piggyBac. Again, this raises the question of how the domesticated transposase is targeted to its cleavage sites. The size distribution of Paramecium IESs exhibits a 10.2-bp periodicity that coincides with the helical phase of DNA. This may reflect DNA bending constraints on assembly of the IES excision complex. Evidence was presented for a requirement for Ku70/Ku80 before DNA cleavage, which would possibly favor the precise repair of IES excision sites by the NHEJ pathway.