Date Published: October 23, 2018
Publisher: Public Library of Science
Author(s): Marcin Nowicki, Sarah L. Boggess, Arnold M. Saxton, Denita Hadziabdic, Qiu-Yun Jenny Xiang, Thomas Molnar, Matthew L. Huff, Margaret E. Staton, Yichen Zhao, Robert N. Trigiano, Berthold Heinze.
Chloroplast DNA is a part of plant non-nuclear genome, and is of particular interest for lineage studies. Moreover, the non-coding regions of cpDNA display higher mutation rates than the conserved coding cpDNA, which has been employed for phylogenetic and population research. We analyzed the cpDNA of 332 gDNA samples from collections of Cornus florida and C. kousa (commercial cultivars, breeding selections, and wild kousa accessions from Asia), using the chlorotyping system developed on North America-native, wild accessions of C. florida. Our results indicated significant differences in chlorotype frequencies between the two species. Cornus florida samples were represented by all major chlorotypes previously described, whereas all C. kousa samples analyzed had only one of the chlorotype patterns shown by C. florida. The chlorotyping analytic panel was then expanded by sequencing the targeted three non-coding cpDNA regions. Results indicated a major difference in the maternally-inherited cpDNA between the two closely related Big-Bracted Cornus species. Chlorotype diversity and differences in the proportion of informative sites in the cpDNA regions of focus emphasized the importance of proper loci choice for cpDNA-based comparative studies between the closely related dogwood species. Phylogenetic analyses of the retrieved sequences for the other species of Cornus provided information on the relative utility of the cpDNA regions studied and helped delineate the groups (Big-Bracted, Cornelian Cherries, Blue/White-Fruited) within the genus. Genealogical relationships based on the cpDNA sequences and the inferred chlorotype networks indicated the need for continued analyses across further non-coding cpDNA regions to improve the phylogenetic resolution of dogwoods.
Native to temperate and boreal climates of Europe, Asia, and North America, dogwoods (genus Cornus L.: Cornaceae) include up to 60 species of shrubs and small trees popular for ornamental and commercial uses (historically), wildlife and human food, and for some ethnic medicinal properties[1, 2]. Flowering dogwood, C. florida L., is native to the eastern United States of America (USA) and has been used as a model tree species in studies on plant biology[3, 4], biochemistry[5, 6], pathology[7, 8], and population genetics[9, 10]. Closely related C. kousa F. Buerger ex Hance F of Asian origin (China, Korea, and Japan) was naturalized to the USA mainly for ornamental purposes (including hybridization with C. florida) because of its attractive growth habit and showy bracts, and its overall high resistance to insects and pathogens[7, 8, 11–13]. The ornamental value of both species drove consumer demand and resulted in selection/breeding and commercial release of over 80 cultivars and 24 varieties for C. florida, and over 70 cultivars for C. kousa[14, 15], despite a relatively long breeding process[4, 16]. Both species are economically important in the USA with sales around $30M per year (corresponding to over 3 million trees sold). Tennessee-based growers accounted for about 25% of this value or almost one-half of the trees in commerce.
In this study, we aimed to expand the chlorotyping strategy on the cpDNA of the two Big-Bracted dogwood collections, which extended the recent study in this area by Call et al.. While doing so, we also used an economically attractive classical Sanger sequencing approach, alternative to a typical GBS, to improve the chlorotyping panel published therein. Use of the cpDNA for phylogenetic and phylogeographic studies, based on its particular advantages (i.e., uniparental inheritance, haploidy, very rare recombination, content and size stability) has contributed significantly to the achievements in plant biology over last 20 years[40–42]. Next-generation sequencing added to the momentum with comparisons of the complete cpDNA sequences[43–45]. This new vantage point confirmed the existence of evolutionary hot-spots, particularly in the non-coding cpDNA regions[19, 20, 46, 47].
The presented chlorotyping panel for C. florida and C. kousa suggests that their diversity has been captured in the cultivated forms at significant levels. Species diversity in the commercial cultivars measured with chlorotyping was lower than in the collection of wild C. florida accessions and suggests the species diversity present in nature will be useful for breeding. Similar comparisons for C. kousa dogwood accessions and cultivars can be completed using the chlorotyping protocols described in this study. We were able to improve on the previous Cornus chlorotyping research by sequencing only five C. florida and six C. kousa accessions over the three non-coding cpDNA regions underscoring the efficiency offered by our approach. Although the C. florida genome sequence is not yet available, our study adds sequence resources to the genus cpDNA information. Furthermore, our data suggests the C. florida-based H3 chlorotype as the most ancestral among the species tested, and underscores the need for testing further regions in regards of their usefulness for Cornus species delineation within the genus and molecular diversity.