Targeted De Novo Centromere Formation in Drosophila Reveals Plasticity and Maintenance Potential of CENP-A Chromatin

Centromeres are chromosomal structures that are essential for accurate chromosome segregation during cell division. Centromeres are commonly composed of repetitive DNA sequences, however, chromosomes can lack these sequences and segregate normally, giving rise to the current model in which centromeres are determined epigenetically by chromatin containing the histone H3 variant CENP-A. Confoundingly, centromeric repeats are able to assemble functional de novo centromeres on human artificial chromosomes, and neocentromeres (new centromeres that lack centromeric repeats) gain centromere repeats over time, suggesting that DNA sequence can influence centromere function. Elucidating the contributions of genetic and epigenetic components to centromere identity remains difficult as the repetitive nature of centromeric DNA has made it virtually impossible to assemble contiguous centromeres in most eukaryotes. Additionally, we lack an animal model to test whether or not CENP-A alone is sufficient for centromere identity and function at any noncentromeric location. Using Drosophila melanogaster as a model to address these issues, I have A) used FISH probes designed against candidate centromeric contigs to determine whether or not these candidates are in fact the centromeres, and B) developed a fly model in which to assemble de novo centromeres in a site-specific and tissue-specific manner to test the epigenetic model for centromere specification. I have found that the five candidate centromeric contigs, which are composed of islands of retroelements flanked by satellite DNA, correspond to all five Drosophila melanogaster centromeres. Interestingly, the retroelement G2/Jockey-3 is present in multiple copies within each centromere candidate, suggesting a possible role in centromere function. Tethering CAL1, the Drosophila specific centromere assembly factor, to any genomic locus in somatic cells is sufficient at nucleating de novo centromeres. After CAL1 tethering has ceased, these de novo centromeres can be epigenetically maintained. However, their maintenance is dependent upon chromosomal stability. These data emphasize the expendability of centromeric DNA in centromere specification, indicating that centromeres are determined by CENP-A in animals, however, the discovery that all endogenous centromeres contain the G2/Jockey-3 retroelement is intriguing and may point to a supporting role for DNA in centromere function