The draft genome of Primula veris yields insights into the molecular basis of heterostyly

Abstract

Background: The flowering plant Primula veris is a common spring blooming perennial that is widely cultivated throughout Europe. This species is an established model system in the study of the genetics, evolution, and ecology of heterostylous floral polymorphisms. Despite the long history of research focused on this and related species, the continued development of this system has been restricted due the absence of genomic and transcriptomic resources.

Results: We present here a de novo draft genome assembly of P. veris covering 301.8 Mb, or approximately 63% of the estimated 479.22 Mb genome, with an N50 contig size of 9.5 Kb, an N50 scaffold size of 164 Kb, and containing an estimated 19,507 genes. The results of a RADseq bulk segregant analysis allow for the confident identification of four genome scaffolds that are linked to the P. veris S-locus. RNAseq data from both P. veris and the closely related species P. vulgaris allow for the characterization of 113 candidate heterostyly genes that show significant floral morph-specific differential expression. One candidate gene of particular interest is a duplicated GLOBOSA homolog that may be unique to Primula (PveGLO2), and is completely silenced in L-morph flowers.

Conclusions: The P. veris genome represents the first genome assembled from a heterostylous species, and thus provides an immensely important resource for future studies focused on the evolution and genetic dissection of heterostyly. As the first genome assembled from the Primulaceae, the P. veris genome will also facilitate the expanded application of phylogenomic methods in this diverse family and the eudicots as a whole.

Figure 1

Venn diagrams of orthologous gene clusters. The number of putatively unique transcripts in each transcriptome is shown below the taxon name. The value in parentheses represents the number of orthologous genes that are most likely to be single-copy in all of the five transcriptomes sampled. (A) Orthologous groups identified in comparative transcriptome analysis of five species of Primula. (B) Orthologous groups identified in comparative transcriptome analysis of five Euasterid species with draft genome assemblies. See also Additional files 4 and 5.

Figure 2

Distribution of genes differentially expressed in L- and S-morph plants assigned to functional annotation clusters. Annotation clusters are labeled with a single GO term that is representative of all of the terms defining cluster membership. See Additional file 1: Table S5 for a complete listing of GO terms associated with each annotation cluster.

Figure 3

Sequence alignment and tree showing the relationship of GLOBOSA / PISTILLATA genes in Primulaceae. (A) Amino acid positions that are shaded are identical to PveGLO2. (B) The tree was rooted with the PI gene sequence from the kiwifruit draft genome (A. chinensis, [69]). Branch support values were calculated based on 1,000 neighbor-joining bootstrap replicates. Our labelling of GLO1 and GLO2 follows the convention established by Viaene et al. [91]. See text for details regarding gene names.

Figure 4

Identification of S -locus linked SNPs. Graphical representation of genotypes of (A) 48 P. veris plants genotyped for 13 PCR-based markers derived from the bulk segregant RAD sequencing analysis and three SNPs in SLL1, and (B) 91 P. veris plants genotyped for four PCR-based SNP markers and three SNPs in the SLL1 gene. Phenotype: yellow = L-morph, green = S-morph. Genotypes: yellow = homozygous for allele 1, green = heterozygous, blue = homozygous for allele 2.