Identification, sequence divergence, relative abundance, and transcription of centromeric retrotransposon subfamilies in maize

Despite the conserved function of centromeres in cell division, the underlying DNA sequence at the centromeres is not conserved. Centromeres were originally thought to be impervious to transcription, but now there is increasing evidence of transcription of centromeric repeats, albeit at low levels. Plant centromeres, in general, are composed of tandem arrays of centromeric satellite repeats interspersed with centromeric retrotransposons (CR) that integrate exclusively at centromeres. Two subfamilies each of centromeric retrotransposons of maize (CRM) and rice (CRR) have been described so far. To understand the evolution of CRs, we analyzed the available Zea mays inbred B73 genomic sequence and the rice genomic sequence and identified an additional four CRs, two each from maize and rice. We determined the orthologous relationships between maize and rice CRs using extensive phylogenetic comparisons. Members of CRM1 subfamily, many of which have integrated recently into the genome, were at least 5 times more numerous than members of any other CRM or CRR subfamily. The CRM1 homolog, CRR3 could not be identified in Oryza sativa ssp. japonica but was present in the wild rice O. officinalis. Further, examination of CR homologs in Arabidopsis suggests a correlation between the abundance of active CRs and centromere/chromosome size. In addition, we show that CRM1 subfamily has recently expanded at the maize centromeres due to the formation of at least 5 major subgroups (R1 through R5---in the order of their formation) derived by recombination among two parent alleles CRM1 A and CRM1 B and their derivatives. The parent alleles, A and B appear to have been brought together during the formation of present day maize by allotetraploidization. Furthermore, to assess the transcription of CRs in maize we analyzed the publicly available ESTs (expressed sequence tag). Our results indicate that CRM transcripts are rare but are expressed at a relatively high level in sperm cells, shoot apical meristem, and immature ear. In addition, we determined the start and end of CRM transcripts. Overall, the characterization of CRMs has increased our understanding of CRM structure, diversity and evolution as well as facilitated the efficient junction mapping of centromeres.