Sequence preference for de novo centromere formation on an artificial chromosome in Caenorhabditis elegans

Abstract

Canonical centromere DNA sequence is essential for de novo centromere formation on artificial chromosomes (ACs) in regional and point centromeric species. However, neocentromeres on endogenous chromosomes in these species have also been found on non-centromeric sequences. To further analyze the contribution of DNA sequence, chromatin contexts and the mechanism of de novo centromere formation, we took advantage of the robust AC-forming system in Caenorhabditis elegans embryos. We injected different DNA sequences to form artificial chromosomes in C. elegans and monitored the mitotic segregation of ACs in embryos to elucidate if there is a sequence preference for forming de novo centromere. We demonstrated that high-AT content sequences, but not repetitive sequences accelerated the process of de novo centromere formation on newly formed ACs. We further introduced exogenous genomic DNA from another species into C. elegans to construct an AC that can propagate through generations. We co-injected the restriction enzyme-digested budding yeast genomic DNA and a low ratio of C. elegans marker genes. We applied MinION and Mi-seq platforms, performed whole-genome sequencing and de novo genome assembly of the AC. The assembled AC contigs are about 10.8 Mbases in total, which is closed to the cytological estimation. Interestingly, the sequence structure of AC reveals that the AC was formed through random end-joining of the injected yeast genomic DNA fragments and the co-injection marker. We also demonstrated that short DNA fragments (<500bp) were selectively excluded from the process of AC formation. We mapped CENP-A (HCP-3) in the embryonic cells of this strain using ChIP-seq and found that both the CENP-A (HCP-3) domains on endogenous chromosomes and the AC showed a positive correlation of large domain size with high AT-content. In addition, de novo centromere prefers to form on non-expressed or silenced marker gene loci, which is consistent to the organization of endogenous centromeres