Significance: The generation of haploids is the most powerful means to accelerate the plant-breeding process. We elucidated whether point mutations in the centromere-specific histone H3 variant CENH3 could be harnessed for the induction of haploids. We identified plants with impaired centromere loading caused by a mutation in the centromere-targeting domain (CATD). The same mutation results in reduced loading of CENH3 in transgenic Arabidopsis and sugar beet. Arabidopsis plants carrying this single point mutation in wild-type CENH3 were used as haploid inducers. Because the identified mutation site is highly conserved and because point mutations can be generated by mutagenesis or genome editing, the described method offers opportunities for application in a wide range of crop species.

The ability to generate haploids and subsequently induce chromosome doubling significantly accelerates the crop breeding process. Haploids have been induced through the generation of plants from haploid tissues (in situ gynogenesis and androgenesis) and through the selective loss of a parental chromosome set via inter- or intraspecific hybridization. Here, we focus on the mechanisms responsible for this selective chromosome elimination. CENH3, a variant of the centromere-specific histone H3, has been exploited to create an efficient method of haploid induction, and we discuss this approach in some detail. Parallels have been drawn with chromosome-specific elimination, which occurs as a normal part of differentiation and sex determination in many plant and animal systems.

During cell division, spindle fibers attach to chromosomes at centromeres. The DNA sequence at regional centromeres is fast evolving with no conserved genetic signature for centromere identity. Instead CENH3, a centromere-specific histone H3 variant, is the epigenetic signature that specifies centromere location across both plant and animal kingdoms. Paradoxically, CENH3 is also adaptively evolving. An ongoing question is whether CENH3 evolution is driven by a functional relationship with the underlying DNA sequence. Here, we demonstrate that despite extensive protein sequence divergence, CENH3 histones from distant species assemble centromeres on the same underlying DNA sequence. We first characterized the organization and diversity of centromere repeats in wild-type Arabidopsis thaliana. We show that A. thaliana CENH3-containing nucleosomes exhibit a strong preference for a unique subset of centromeric repeats. These sequences are largely missing from the genome assemblies and represent the youngest and most homogeneous class of repeats. Next, we tested the evolutionary specificity of this interaction in a background in which the native A. thaliana CENH3 is replaced with CENH3s from distant species. Strikingly, we find that CENH3 from Lepidium oleraceum and Zea mays, although specifying epigenetically weaker centromeres that result in genome elimination upon outcrossing, show a binding pattern on A. thaliana centromere repeats that is indistinguishable from the native CENH3. Our results demonstrate positional stability of a highly diverged CENH3 on independently evolved repeats, suggesting that the sequence specificity of centromeres is determined by a mechanism independent of CENH3.

The centromere is the chromosome region where microtubules attach during cell division. In contrast to monocentric chromosomes with one centromere, holocentric species usually distribute hundreds of centromere units along the entire chromatid. We assembled the chromosome-scale reference genome and analyzed the holocentromere and (epi)genome organization of the lilioid Chionographis japonica. Remarkably, each of its holocentric chromatids consists of only 7 to 11 evenly spaced megabase-sized centromere-specific histone H3-positive units. These units contain satellite arrays of 23 and 28 bp-long monomers capable of forming palindromic structures. Like monocentric species, C. japonica forms clustered centromeres in chromocenters at interphase. In addition, the large-scale eu- and heterochromatin arrangement differs between C. japonica and other known holocentric species. Finally, using polymer simulations, we model the formation of prometaphase line-like holocentromeres from interphase centromere clusters. Our findings broaden the knowledge about centromere diversity, showing that holocentricity is not restricted to species with numerous and small centromere units.