Transposase domestication and epigenetics: Impact on genome dynamics – PIGGYPACK

For the functional analysis of PgmLs and other Pgm partners identified in the study (e.g. Ku70/Ku80), we performed systematic RNAi experiments and monitored the ability of silenced cells to develop a functional new MAC (survival tests in sexual progeny). The effect of RNAi on IES excision and TE elimination was tested through high throughput DNA seq from new MACs extracted from RNAi-treated cells and bioinformatic quantitative analysis of DNA seq data. Nuclear localization of Pgm and each putative partner has been confirmed using GFP fusions in live cells, or by immunostaining of endogenous or tagged proteins using custom or commercial specific antibodies. Protein-protein interactions have been validated using protein co-precipitation assays.
To search for Pgm-interacting proteins directly in Paramecium cells, two major strategies have been used: protein immunoprecipitation from soluble nuclear extracts using specific anti Pgm antibodies, or proximity-driven protein modification.
Nuclear Magnetic Resonance (NMR) was used to determine the structure of the CR domains of the PB transposase (used as a reference) and Pgm. Mass spectrometry has been used to establish a catalog of all histone modifications present in Paramecium during MAC development. Gel shift assays have been used to test the interaction of PB, Pgm and PgmL CR domains with DNA. Interaction with differentially modified histones has been tested using commercial histone peptide arrays, pull-down assays and immunostaining of fixed cells with anti Pgm antibodies and custom antibodies raised against specific Paramecium histone modifications.
All DNA seq and RNA seq datasets have been stored and processed using a private secured ParameciumDB interface.

• PGML genes are grouped in five families, each specifically transcribed during MAC development. PgmL proteins localize in the developing new MAC during IES excision and interact with Pgm when co-expressed in a heterologous system. Each PgmL family is essential for correct localization of Pgm in the developing MACs, suggesting that Pgm acts within a large multi-component complex. A novel software (ParTIES) was developed to facilitate IES detection and quantitative analysis of excision efficiency and error rates. Depletion in PgmL2, 4 and 5 completely inhibits IES excision, suggesting that these families are core components of the complex. Depletion in PgmL1 and 3 allows partial IES excision, but with more frequent excision errors, suggesting that the latter two families play a more accessory role.
• The CR domain of the PB transposase adopts a compact fold and binds two zinc ions in a cross-brace zinc finger motif similar to PHD and RING motifs. Unexpectedly, this domain is a DNA binding domain that specifically recognizes a motif within the terminal inverted repeats of the piggyBac transposon. Interaction models have been proposed according to NMR studies of the peptide/DNA complex and molecular-docking simulations. The structure of the Pgm CR domain was also determined: preliminary analyses of NMR data indicate that two Zn2+ are coordinated by two His and six Cys residues, in a novel cross-brace arrangement of the two zinc-binding motifs that is very different from the PB CR domain.
• Mass spectrometry analysis of Paramecium histone modifications confirmed the existence of H3K9me3 and H3K27me3 in the developing MAC. Specific antibodies against Pgm; PgmL2, PgmL5 and Paramecium H3K27me3 and H4K20me1 have been raised and validated.
• A novel procedure was set up to sort new developing MACs to high purity from PGM-silenced cells using flow cytometry, according to size, granularity and DNA content.

Future studies will aim at determining the stoichiometry and organization of Pgm and PgmLs in the Pgm-associated complex purified from Paramecium nuclear extracts or following in vitro reconstitution using purified proteins. These complexes will be tested in vitro for DNA binding or cleavage activity. Immunostaining using available specific antibodies against endogenous or tagged proteins will be used to follow the relative timing of appearance/disappearance of each components of the complex during MAC development, together with interesting histone modifications. Colocalization will be monitored at the highest possible resolution using confocal microscopy.
Binding of the Pgm and PgmL CR domains to histones will be tested in pull-down experiments, using GST-tagged CR domains and purified Paramecium histones. Binding to interesting candidate modifications will be analyzed in detail using NMR spectroscopy to identify critical residues involved in the interaction. New purification strategies will be tested to obtain soluble preparations of PgmL CR domains suitable for NMR spectroscopy.
We will pursue our efforts to develop experimental strategies to identify novel Pgm partners directly from Paramecium cells.

During the first 18 months of the project, the collaborative work of the PIGGYPACK consortium was presented at 5 international conferences (4 invited talks, 2 oral presentations and 1 poster presentation) and 4 french meetings (1 invited talk, 4 oral presentations and 1 poster presentation).
Individual partners published 3 articles in peer-reviewed journals (2 for partner 3, 1 for partner 4) and gave presentations in 7 international conferences (4 for partner 1, 1 for partner 2, 1 for partner 3, for partner 4) and 8 french meetings (3 for partner 1, 5 for partner 4).

Transposases mobilize their cognate transposons through specific binding to terminal inverted repeats (TIR) at transposon ends, phosphodiester-bond cleavage and strand-transfer into integration sites. Domesticated transposase genes have been identified in numerous genomes but whether they serve a cellular function is an open question. These genes are generally no longer embedded within mobile elements. The best known domesticated transposase is RAG1, which initiates assembly of immunoglobulin genes in lymphocytes, by cleaving specific DNA sequences with homology to the TIRs of its ancestral transposon. Many other examples of programmed genome rearrangements (PGR) have been described during development or somatic differentiation in various organisms and an attractive hypothesis would be that they are also carried out by domesticated transposases.

To address the question of the part taken by domesticated transposases in catalysis of PGR, we will focus on the ciliate Paramecium. This unicellular eukaryote provides an extraordinary model for studying the mechanisms involved in PGR, because it rearranges its entire somatic genome at each sexual cycle. Paramecium harbors two distinct nuclei within its cytoplasm. A diploid micronucleus (MIC) contains the germline genome that is transmitted to the progeny at each sexual cycle. A polyploid somatic macronucleus (MAC) contains a rearranged version of the germline genome, and is responsible for gene expression. At each sexual cycle, the MAC is lost and a new MAC develops from a copy of the MIC. MAC development involves elimination of two types of germline-limited sequences: (i) repeated DNA is removed in association with chromosome fragmentation, and (ii) at least 45,000 short single-copy noncoding sequences (IES, or Internal Eliminated Sequences) are excised precisely to reconstitute functional open reading frames. IES excision is initiated by double-strand cleavages that depend on PiggyMac (Pgm), a domesticated PiggyBac transposase.

Our major objective is to understand how IES ends are recognized for Pgm-dependent cleavage. The requirement for a domesticated PiggyBac transposase for PGR in Paramecium is intriguing, because IESs are not related to piggyBac transposons and do not carry TIRs. Published lines of evidence suggest that noncoding RNAs mediate genome-wide comparison of the old MIC and MAC genomes and guide the elimination of MIC-restricted sequences in the new developing MAC, through the deposition of epigenetic heterochromatin marks. However, recent reports indicate that this mechanism may control the excision of only a subset of IESs. Interestingly, we recently discovered nine other Pgm-related domesticated transposases, some of which, although apparently catalytically inactive, are required for IES excision and interact with Pgm in vitro. In the present project, we will explore two novel non-exclusive hypotheses that may explain how Pgm targets precise excision of all IESs: (i) Pgm may associate with additional partners and form a variety of protein complexes that direct precise cleavage at the ends of distinct classes of IESs; (ii) Pgm may be guided to its DNA cleavage sites through the “reading” of alternative chromatin modifications. We will also consider the possibility that Pgm is complexed with as yet unidentified “guide” RNA molecules that direct the complex to its targets.

We propose a multidisciplinary work plan to (i) characterize the components of the Pgm-associated machinery; (ii) dissect their function in IES excision using in vivo and in vitro approaches combined with large scale genomic analyses; and (iii) determine the structural basis of Pgm complex recognition of histone marks. To achieve these goals, our consortium brings together four partners with highly complementary expertise in Paramecium molecular and cellular genetics, biochemistry, bioinformatic analysis of genome-wide datasets and structural analysis of protein interactions based on NMR.