Transcriptional and epigenetic regulation of MERVL retrotransposons in mouse embryonic stem cells

At fertilization, both maternal and paternal genomes are transcriptionally silent and development is supported by maternally provided transcripts. During the first cell cycle, the embryonic genome starts to be transcribed, a process which is further boosted at the late 2-cell (2C) stage. During this process, termed zygotic genome activation (ZGA), several types of retrotransposons become de-repressed in a stage specific manner. MERVL elements are strongly and specifically upregulated at the 2-cell stage. MERVL and other 2-cell (2C)-specific transposons also act as regulatory elements for embryonic transcription and have been implicated in shaping the chromatin landscape.
MERVL elements are also transiently expressed in a rare subpopulation of embryonic stem cells, termed 2C-like (2CL) cells. These 2CL-cells have been previously identified by expression of a stably integrated MERVL-based fluorescent “2C” reporter construct. The 2C construct contains the 5’ MT2-Mm LTR, a primer binding site (PBS) sequence, and a short sequence encoding partially for the GAG protein. 2CL-cells are characterized by expression of MERVL elements and a subset of genes specifically expressed during ZGA. Moreover, their relaxed chromatin genome structure resembles that of 2-cell stage embryos.
In recent years, regulation of MERVL expression in vitro has been used as a model system to study embryonic transcription. To date, several factors that modulate the appearance of 2CL-cells have been identified. Among others, the transcription factor DUX was identified as a regulator of MERVL and zygotic gene expression, being required and sufficient to drive the MERVL-transcriptional program in vitro. In vivo, however, deficiency of Dux did not substantially impair expression of MERVL and other transposons nor of ZGA genes during pre-implantation development.
In my PhD research, I aimed at studying mechanisms regulating MERVL expression. By generating series of mono-clonal ESC lines expressing modified 2C-reporter constructs, I could demonstrate that the PBS-gag sequence confers potent repression to the MERVL element, independently of the integration site of the reporter.
By analyzing transcriptional and chromatin accessibility profiles of early embryos, we identified a transcription factor (TF) binding site that might regulate, next to DUX, the 2C-specific transcriptome in vivo and in vitro. Proteins of several transcription factor families can bind to this motif in vitro, termed OBOX.
Accordingly, I observed that absence of the DUX motif significantly reduced, yet not completely abolished expression of the reporter driven by the MT2-Mm LTR promoter only. In contrast, absence of both OBOX and DUX motifs fully prevented MT2-Mm-driven expression. The requirement of both motifs for reporter expression was only observed in a non-permissive chromatin environment, but not in a permissive and accessible genomic region. This finding suggests that the corresponding transcription factors act as pioneering transcription factors in 2CL-cells.
Based on RNA-sequencing data, I conclude that while 2C reporter expression recapitulates the known 2C-like state, reporter expression driven by the MT2-Mm LTR promoter and lacking the PBS-gag sequence does not represent a specific cellular state. Nonetheless, both reporters mark transient populations and their expression oscillates between active and inactive transcriptional states.
Finally, by exploring different growth conditions and chemical modulation of chromatin modifier activities, I demonstrated that MERVL reporters with (2C) or without PBS-gag (MT2-Mm) are controlled to varying degrees by G9a/GLP histone H3 lysine 9 methyltransferases as well as SUMO-conjugating enzymes, in part possibly directly and/or indirectly.