Retrotransposon instability dominates the acquired mutation landscape of mouse induced pluripotent stem cells

Abstract

Induced pluripotent stem cells (iPSCs) can in principle differentiate into any cell of the body, and have revolutionized biomedical research and regenerative medicine. Unlike their human counterparts, mouse iPSCs (miPSCs) are reported to silence transposable elements and prevent transposable element-mediated mutagenesis. Here we apply short-read or Oxford Nanopore Technologies long-read genome sequencing to 38 bulk miPSC lines reprogrammed from 10 parental cell types, and 18 single-cell miPSC clones. While single nucleotide variants and structural variants restricted to miPSCs are rare, we find 83 de novo transposable element insertions, including examples intronic to Brca1 and Dmd. LINE-1 retrotransposons are profoundly hypomethylated in miPSCs, beyond other transposable elements and the genome overall, and harbor alternative protein-coding gene promoters. We show that treatment with the LINE-1 inhibitor lamivudine does not hinder reprogramming and efficiently blocks endogenous retrotransposition, as detected by long-read genome sequencing. These experiments reveal the complete spectrum and potential significance of mutations acquired by miPSCs.

Comprehensive profiling of L1 retrotransposons in mouse

“…Although ONT is prone to high error rates as compared to short-read sequencing methods, the long reads still usually allow unambiguous alignment, even within repetitive regions. High coverage ONT whole genome sequencing has been used to detect L1 insertions (34, 37, 38); however, whole genome sequencing may not be cost effective when examining a large number of samples, such as studies involving single cell analysis. Sequencing an L1-enriched TIPseq reaction is expected to decrease costs considerably.…”

LINE-1 retrotransposon activation intrinsic to interneuron development

“…Despite apparent potential for SOX6-mediated L1 transcriptional activation, DNA methylation in somatic cells is expected to silence L1 promoters 6,43,44 . Therefore, to further probe the apparent specificity of L1 transcription to PV + neurons, we performed L1 T F 5ʹUTR monomer bisulfite sequencing 30,35,45 on neonate hippocampal cell populations. L1 T F was significantly ( P =0.03) less methylated on average in PV + neurons (83.9%) than in PV - neurons (91.8%) ( Fig.…”

“…As an orthogonal approach, we electroporated cultured primary mouse neurons with an L1-mCherry retrotransposition reporter based on L1spa, a mobile mouse L1 expressed from its native monomeric 5′UTR promoter [28][29][30] . Whilst we observed PV + /mCherry + cells (Extended Data Fig.…”

“…b, L1 T F monomer methylation was significantly lower (*P<0. Long-read Oxford Nanopore Technologies (ONT) sequencing allows genome-wide analysis of TE family methylation, as well as that of individual TE loci 30,35,50 . We ONT sequenced PV + , PV -, PV -/Tub + and PV -/Tubcells from neonatal hippocampus samples at ~16× average genome-wide depth, and ~29× and ~34× respectively for the PV + and combined PVpopulations (Supplementary Table 1).…”

OGT prevents DNA demethylation and suppresses the expression of transposable elements in heterochromatin by restraining TET activity genome-wide

“…In mouse embryonic stem cells (mESC), TET1 and TET2 bound the regulatory regions of diverse classes of TEs, including young L1 subfamilies that are not yet repressed by KAP1; TET binding at these elements was associated with increased 5hmC and decreased 5mC [45], implicating TET-mediated DNA demethylation in TE expression. In “naïve” mESC cultured under 2i conditions, which display lower 5mC levels and higher TE expression than “primed” mESC grown in serum plus LIF [47, 48], depletion of the H3K9 methyltransferase SETDB1 increased the expression of IAPs and young L1 (T F subfamily) elements, and this increase required TET catalytic function [46]. KAP1 and TET proteins are implicated in controlling TE expression via H3K9 and DNA methylation [49].…”