Genome-wide mapping of chromatin marks from 1,000 cells to study epigenetic reprogramming in primordial germ cells

摘要

Germline development is characterized by genome-widereprogramming of DNA methylation. Recent work hasenlightened the dynamics of DNA methylation in primordialgerm cells (PGCs), but knowledge of histonemodification dynamics at these developmental stagesremains limited, mostly due to the difficulty in obtainingenough high quality chromatin immunoprecipitation(ChIP) material for sequencing. Previous work in ourlaboratory has demonstrated the importance of histonemethyltransferases in silencing retroelements [1] and asubset of germline-specific genes [2] in embryonic stemcells. Here, we sought to develop a reliable ChIPsequencingprotocol to study the dynamics of histonemodification during the DNA methylation reprogrammingthat occurs in PGCs.We have developed a scaled down native ChIP andsequencing library construction protocol that can be performedon small cell numbers. We optimized sample fragmentation,antibody concentration, ChIP conditions,library construction and amplification to generate highquality, high resolution H3K9me3 sequencing librariesfrom as little as 1,000 embryonic stem cells. Paired-endsequencing (Illumina HiSeq) of these pooled and indexedlibraries generated an average of 28 million aligned readpairs (with 6 libraries per sequencing lane), with under20% duplicate reads, for an average of 23 million uniqueread pairs. Under optimized conditions, we found thatover 85% of the peaks identified using standard nativeChIP-sequencing (using 2 million cells as starting material)were also detected by our small cell number nativeChIP-seq protocol. We also found excellent reproducibilitybetween independent ChIP experiments. Using thisoptimized small cell number native ChIP-seq protocol, wegenerated genome-wide H3 and H3K9me3 profiles from1,000 E13.5 PGCs and identified a unique cohort of genesand retroelements enriched for this repressive mark. Integrationof this chip-seq data with DNA methylation/WGBS and transcriptome data generated at the samedevelopmental stage will be presented.