Genomic analysis reveals rich genetic variation and potential targets of selection during domestication of castor bean from perennial woody tree to annual semi-woody crop

Dynamics of imprinted genes and their epigenetic mechanisms in castor bean seed with persistent endosperm

Genomic imprinting refers to parent-of-origin-dependent gene expression and primarily occurs in the endosperm of flowering plants, but its functions and epigenetic mechanisms remain to be elucidated in eudicots. Castor bean, a eudicot with large and persistent endosperm, provides an excellent system for studying the imprinting. Here, we identified 131 imprinted genes in developing endosperms and endosperm at seed germination phase of castor bean, involving into the endosperm development, accumulation of storage compounds and specially seed germination. Our results showed that the transcriptional repression of maternal allele of DNA METHYLTRANSFERASE 1 (MET1) may be required for maternal genome demethylation in the endosperm. DNA methylation analysis showed that only a small fraction of imprinted genes was associated with allele-specific DNA methylation, and most of them were closely associated with constitutively unmethylated regions (UMRs), suggesting a limited role for DNA methylation in controlling genomic imprinting. Instead, histone modifications can be asymmetrically deposited in maternal and paternal genomes in a DNA methylation-independent manner to control expression of most imprinted genes. These results expanded our understanding of the occurrence and biological functions of imprinted genes and showed the evolutionary flexibility of the imprinting machinery and mechanisms in plants.

Genome-wide characterization and evolutionary analysis of linker histones in castor bean (Ricinus communis)

H1s, or linker histones, are ubiquitous proteins in eukaryotic cells, consisting of a globular GH1 domain flanked by two unstructured tails. Whilst it is known that numerous non-allelic variants exist within the same species, the degree of interspecific and intraspecific variation and divergence of linker histones remain unknown. The conserved basic binding sites in GH1 and evenly distributed strong positive charges on the C-terminal domain (CTD) are key structural characters for linker histones to bind chromatin. Based on these features, we identified five linker histones from 13 GH1-containing proteins in castor bean (Ricinus communis), which were named as RcH1.1, RcH1.2a, RcH1.2b, RcH1.3, and RcH1.4 based on their phylogenetic relationships with the H1s from five other economically important Euphorbiaceae species (Hevea brasiliensis Jatropha curcas, Manihot esculenta Mercurialis annua, and Vernicia fordii) and Arabidopsis thaliana. The expression profiles of RcH1 genes in a variety of tissues and stresses were determined from RNA-seq data. We found three RcH1 genes (RcH1.1, RcH1.2a, and RcH1.3) were broadly expressed in all tissues, suggesting a conserved role in stabilizing and organizing the nuclear DNA. RcH1.2a and RcH1.4 was preferentially expressed in floral tissues, indicating potential involvement in floral development in castor bean. Lack of non-coding region and no expression detected in any tissue tested suggest that RcH1.2b is a pseudogene. RcH1.3 was salt stress inducible, but not induced by cold, heat and drought in our investigation. Structural comparison confirmed that GH1 domain was highly evolutionarily conserved and revealed that N- and C-terminal domains of linker histones are divergent between variants, but highly conserved between species for a given variant. Although the number of H1 genes varies between species, the number of H1 variants is relatively conserved in more closely related species (such as within the same family). Through comparison of nucleotide diversity of linker histone genes and oil-related genes, we found similar mutation rate of these two groups of genes. Using Tajima's D and ML-HKA tests, we found RcH1.1 and RcH1.3 may be under balancing selection.

Development of Novel Markers for Yield in Hevea brasiliensis Muell. Arg. Based on Candidate Genes from Biosynthetic Pathways Associated with Latex Production

Scarcity of functional genetic markers associated with candidate genes (CGs) is a serious constraint for marker-assisted selection in the natural rubber producing tree, Hevea brasiliensis. In order to develop markers associated with rubber yield, five CGs involved in latex biosynthesis were characterized from 16 popular Hevea varieties. Novel SNPs and indels were identified and developed into markers using simple genotyping techniques like allele-specific PCR, CAPS, etc. A progeny population was genotyped using these markers to validate them, to understand their segregation pattern and to map them to a genetic linkage map. Parent-specific maps were constructed using pseudo-test cross strategy with the help of additional markers. The sequence structure information generated will be useful for future studies on gene mapping, functional relevance of coding SNPs and evolution of rubber biosynthesis genes in Hevea. Concurrently, the markers developed may serve as powerful tools for yield-based selection and for genetic diversity and pedigree studies in Hevea. Above all, the marker assays designed for genotyping could be economically carried out in any laboratory having basic molecular biology infrastructure and expertise.

Quinoa genome assembly employing genomic variation for guided scaffolding

We propose to use the natural variation between individuals of a population for genome assembly scaffolding. In today's genome projects, multiple accessions get sequenced, leading to variant catalogs. Using such information to improve genome assemblies is attractive both cost-wise as well as scientifically, because the value of an assembly increases with its contiguity. We conclude that haplotype information is a valuable resource to group and order contigs toward the generation of pseudomolecules. Quinoa (Chenopodium quinoa) has been under cultivation in Latin America for more than 7500 years. Recently, quinoa has gained increasing attention due to its stress resistance and its nutritional value. We generated a novel quinoa genome assembly for the Bolivian accession CHEN125 using PacBio long-read sequencing data (assembly size 1.32 Gbp, initial N50 size 608 kbp). Next, we re-sequenced 50 quinoa accessions from Peru and Bolivia. This set of accessions differed at 4.4 million single-nucleotide variant (SNV) positions compared to CHEN125 (1.4 million SNV positions on average per accession). We show how to exploit variation in accessions that are distantly related to establish a genome-wide ordered set of contigs for guided scaffolding of a reference assembly. The method is based on detecting shared haplotypes and their expected continuity throughout the genome (i.e., the effect of linkage disequilibrium), as an extension of what is expected in mapping populations where only a few haplotypes are present. We test the approach using Arabidopsis thaliana data from different populations. After applying the method on our CHEN125 quinoa assembly we validated the results with mate-pairs, genetic markers, and another quinoa assembly originating from a Chilean cultivar. We show consistency between these information sources and the haplotype-based relations as determined by us and obtain an improved assembly with an N50 size of 1079 kbp and ordered contig groups of up to 39.7 Mbp. We conclude that haplotype information in distantly related individuals of the same species is a valuable resource to group and order contigs according to their adjacency in the genome toward the generation of pseudomolecules.

Genomic insights into the origin, domestication and genetic basis of agronomic traits of castor bean

BACKGROUND

Castor bean (Ricinus communis L.) is an important oil crop, which belongs to the Euphorbiaceae family. The seed oil of castor bean is currently the only commercial source of ricinoleic acid that can be used for producing about 2000 industrial products. However, it remains largely unknown regarding the origin, domestication, and the genetic basis of key traits of castor bean.

RESULTS

Here we perform a de novo chromosome-level genome assembly of the wild progenitor of castor bean. By resequencing and analyzing 505 worldwide accessions, we reveal that the accessions from East Africa are the extant wild progenitors of castor bean, and the domestication occurs ~ 3200 years ago. We demonstrate that significant genetic differentiation between wild populations in Kenya and Ethiopia is associated with past climate fluctuation in the Turkana depression ~ 7000 years ago. This dramatic change in climate may have caused the genetic bottleneck in wild castor bean populations. By a genome-wide association study, combined with quantitative trait locus analysis, we identify important candidate genes associated with plant architecture and seed size.

CONCLUSIONS

This study provides novel insights of domestication and genome evolution of castor bean, which facilitates genomics-based breeding of this important oilseed crop and potentially other tree-like crops in future.

Bulked segregant analysis reveals candidate genes responsible for dwarf formation in woody oilseed crop castor bean

Plant dwarfism is a desirable agronomic trait in non-timber trees, but little is known about the physiological and molecular mechanism underlying dwarfism in woody plants. Castor bean (Ricinus communis) is a typical woody oilseed crop. We performed cytological observations within xylem, phloem and cambia tissues, revealing that divergent cell growth in all tissues might play a role in the dwarf phenotype in cultivated castor bean. Based on bulked segregant analyses for a F2 population generated from the crossing of a tall and a dwarf accession, we identified two QTLs associated with plant height, covering 325 candidate genes. One of these, Rc5NG4-1 encoding a putative IAA transport protein localized in the tonoplast was functionally characterized. A non-synonymous SNP (altering the amino acid sequence from Y to C at position 218) differentiated the tall and dwarf plants and we confirmed, through heterologous yeast transformation, that the IAA uptake capacities of Rc5NG4-1Y and Rc5NG4-1C were significantly different. This study provides insights into the physiological and molecular mechanisms of dwarfing in woody non-timber economically important plants, with potential to aid in the genetic breeding of castor bean and other related crops.