Genomic studies on floral and vegetative development in the Genus Streptocarpus (Gesneriaceae)

Abstract

The genus Streptocarpus consists of around 180 species with diverse morphologies. At least three main types of vegetative growth forms can be distinguished: caulescent, rosulate (acaulescents with multiple leaves), and unifoliate (acaulescents with one leaf). Floral size, shape, and pigmentation pattern are also highly variable between species.

Previous studies have suggested that some of the morphological characters are inherited as Mendelian traits. For instance, the rosulate growth form is dominant over the unifoliate, and the rosulate / unifoliate growth form was hypothesised to be determined by two genetic loci, based on the Mendelian segregation ratios recorded in backcross and F2 populations.

However, the identity of the loci and the underlying molecular mechanisms remain unknown. In this study, Streptocarpus rexii (rosulate) and Streptocarpus grandis (unifoliate) were used to study the genetic basis of morphological variation in Streptocarpus. The aim is to use modern next generation sequencing (NGS) technologies to build draft genomes, transcriptomes, and genetic maps for the non-model Streptocarpus plants, and carry out quantitative trait loci (QTL) mapping to locate the causative loci.

First, suitable DNA and RNA extraction methods for obtaining NGS-quality nucleic acids from Streptocarpus were established. For DNA extraction this was a modified protocol of the ChargeSwitch gDNA Plant Kit, and for RNA extraction a TRIzol reagent plus phenol:chloroform:isoamyl alcohol wash protocol was devised. The nucleic acid samples extracted were subsequently used for library preparation and NGS sequencing experiments.

Whole genome shotgun sequencing was performed for S. rexii and S. grandis using Illumina HiSeq 4000 and HiSeq X. De novo assembly of the sequence data produced a S. rexii draft genome of 596,583,869 bp, with 95,845 scaffolds and an N50 value of 35,609 bp.

The S. grandis draft genome had a total span of 843,329,708 bp, with 127,951 scaffolds and an N50 value of 31,638 bp. The genome assemblies served as references for subsequent NGS data analysis.

The RNA samples derived from various vegetative and floral tissues of S. rexii and S. grandis were sequenced on MiSeq and HiSeq 4000 platforms. The transcriptome assembly was carried out using de novo and reference-based methods (i.e. mapped to the obtained draft genomes), followed by putative protein-coding open reading frame identification and annotation. For S. rexii, 60,500 and 53,322 transcripts were constructed in the de novo and reference-based assemblies respectively. For S. grandis, 51,267 and 46,429 transcripts were constructed respectively.

A Streptocarpus genetic map was constructed using restriction-site associated DNA sequencing (RAD-Seq) genotyping of a backcross population ((S. grandis × S. rexii) × S. grandis). The RAD-Seq data were analysed using a de novo approach and reference-based approaches with two different aligners, and the RAD-markers recovered from the three approaches were combined to maximise the genetic map density. Different marker-filtering strategies with varying stringencies were also tested and compared. The results showed that the most stringently filtered map had 377 mapped markers in 17 linkage groups, and a total distance of 1,144.2 cM. On the other hand, the densest map consisted of 853 markers in 16 linkage groups (matching the basic haploid chromosome number of the Streptocarpus species used here), and a total distance of 1,389.9 cM.

The maps constructed were used for QTL mapping of growth form variation, identifying up to 5 effective loci for the rosulate / unifoliate phenotypes, with two of the loci on LG2 and LG14 consistently found in all mapping attempts. The results suggest that the variation in growth form may be regulated by two major loci, but a few additional minor loci might also be associated with the trait. Several QTLs for floral dimension, flowering time, and floral pigmentation patterns were also found, and the genetic regions associated with the floral traits of Streptocarpus were revealed for the first time.

During this study valuable genomic resources were generated for future research to identify the genes underlying different morphologies in the genus Streptocarpus. The reported QTLs narrow down the genetic region for fine-mapping studies, and the genome and transcriptome resources will aid the isolation of candidate gene sequences.

Identifying the genetic loci and their crosstalk behind the variable morphologies in future work will greatly add to our knowledge on how the highly diverse genus Streptocarpus has evolved and on how fundamental developmental processes of plants are regulated.