Editorial Office BMC Biology The Origin and Evolution of a Two-Component System of Paralogous Genes Encoding the Centromeric Histone CENH3 in Cereals

Abstract

Background: The cereal family Poaceae is one of the largest and most diverse angiosperm families. The central component of centromere specification and function is the centromere-specific histone H3 (CENH3). Some cereal species (maize, rice) have one copy of the CENH3 gene, while some (wheat, barley, rye) have two. We applied a homology-based approach to sequenced cereal genomes, in order to finally trace the mutual evolution of the structure of the CENH3 genes and the nearby regions in various tribes. Results: We have established that the syntenic group or the CENH3 locus with the CENH3 gene and the boundaries defined by the Cdpk2 and bZip genes first appeared around 50 Mya in a common ancestor of the subfamilies Bambusoideae, Oryzoideae and Pooideae. This locus came to Pooideae with one copy of CENH3 in the most ancient tribes Nardeae and Meliceae. The βCENH3 gene as a part of the locus appeared in the tribes Stipeae and Brachypodieae around 35-40 Mya. The duplication was accompanied by changes in the exon-intron structure. Purifying selection acts mostly on αCENH3s, while βCENH3s form more heterogeneous structures, in which clade-specific amino acid motifs are present. In barley species, the βCENH3 gene assumed an inverted orientation relative to αCENH3 and the Cdpk2 gene was substituted with Cbp3c. As the evolutionary and breeding processes went on, the locus was growing in size due to an increasing distance between αCENH3 and βCENH3 because of a massive insertion of the main LTR-containing retrotransposon superfamilies, gypsy and copia, without any evolutionary preference on either of them. A comparison of the molecular structure of the locus in the A, B and D subgenomes of the hexaploid wheat T. aestivum showed that invasion by mobile elements and concomitant rearrangements took place in an independent way even in evolutionarily close species. Conclusions: The CENH3 duplication in cereals was accompanied by changes in the exon-intron structure of the βCENH3 paralog, which it was not in other plant taxa. The observed general tendency towards the expansion of the CENH3 locus reveals an amazing diversity of ways in which different species implement the scenario described in this paper.

Expression of Two Rye CENH3 Variants and Their Loading into Centromeres

“…The βCENH3 gene has four exons and produces a protein 151 amino acids in length, which is shorter due to an extensive deletion in the NTT domain of the parental gene. Thus, the structure of the CENH3 locus and the intron-exon structure of the CENH3 paralogs in rye are similar to those in other Triticeae species [15] but differ from those in CENH3 paralogs in legumes [19], Mimulus [20] and Arabidopsis [21].…”

“…An increase in the size of the locus and, consequently, in the distance between βCENH3 and αCENH3 genes was due to the massive introduction of various families of TEs, mainly of the main LTR-containing retrotransposon superfamilies, gypsy and copia. A comparison of the composition and prevalence of these families between this locus and the genomes of different Triticeae species revealed substantial differences [15]. In this respect, the CENH3 locus in rye is different from the rest of the genome too (Table 1).…”

“…Thus, conservation is the most preferred mechanism with these species; it preserves the duplicated copy in the genome during evolution, which leads to an increase in gene dosage [35]. Interestingly, the CENH3 locus consisting of the syntenic genes LHCB3-l, CENH3 and bZIP is also present in B. distachyon and O. sativa [15]. However, in O. sativa, the CENH3 gene occurs in the locus as a single copy, while in B. distachyon, it occurs as two copies-just as in rye (Figure 1)-strongly suggesting that whether or not a particular gene would become duplicated could rather be a chance circumstance.…”

“…A duplication event occurred within the CENH3 locus in Stipeae and Brachypodieae species around 35-40 MYA and produced two genes, αCENH3 and βCENH3. The distance between these genes (the intergenic spacer) is 2.16 kb in Brachypodium distachyon and 5.9 kb in Stipa sibirica and is filled with relatively short fragments of retrotransposons and tracks of simple repeats [15]. Later in the evolution of the CENH3 locus in Triticeae, the CDPK2-l gene located to the left of βCENH3 was replaced with the LHCB-l gene, which encodes the chlorophyll-binding protein 3C.…”

“…An increase in the distance between the βCENH3 and αCENH3 genes occurred mainly due to massive insertion of the elements of the LTR-containing retrotransposon superfamilies gypsy and copia. At the same time, the changes in the composition and contribution of individual gypsy and copia families to the molecular structure of the CENH3 locus display a significant diversity even in evolutionarily close species [15].…”