ORIGINAL RESEARCH

published: 05 November 2020

doi: 10.3389/fpls.2020.594272

Plastid Phylogenomics of
Dendroseris (Cichorieae;
Asteraceae): Insights Into Structural
Organization and Molecular
Evolution of an Endemic Lineage
From the Juan Fernández Islands
Myong-Suk Cho 1 , Seon-Hee Kim 1 , JiYoung Yang 2 , Daniel J. Crawford 3 , Tod F. Stuessy 4,5 ,
Edited by: Patricio López-Sepúlveda 6 and Seung-Chul Kim 1*
Stefan Wanke,
1
Technische Universität Dresden, Department of Biological Sciences, Sungkyunkwan University, Suwon-si, South Korea, 2 Research Institute for Dok-do
Germany and Ulleung-do Island, Kyungpook National University, Daegu, South Korea, 3 Department of Ecology and Evolutionary
Biology and the Biodiversity Institute, The University of Kansas, Lawrence, KS, United States, 4 Department of Evolution,
Reviewed by:
Ecology, and Organismal Biology, The Ohio State University, Columbus, OH, United States, 5 Department of Botany
Sònia Garcia,
and Biodiversity Research, University of Vienna, Vienna, Austria, 6 Departamento de Botánica, Universidad de Concepción,
Botanical Institute of Barcelona, Spain
Concepción, Chile
Domingos Cardoso,
Federal University of Bahia, Brazil
Matthias Jost, Dendroseris D. Don comprises 11 species endemic to the Juan Fernández islands in
Technische Universität Dresden,
Germany Chile. They demonstrate spectacular and unusual growth forms of rosette trees with
*Correspondence: extremely variable morphology and occupy wide ecological ranges on the islands. These
Seung-Chul Kim unique plants are now highly threatened with extinction with very small population
sonchus96@skku.edu;
sizes, typically consisting of 10 or fewer individuals in wild. Despite morphological
sonchus2009@gmail.com
and ecological divergence among species of Dendroseris, their monophyly has
Specialty section: been supported in previous studies, but with little resolution among subgeneric
This article was submitted to
Plant Systematics and Evolution,
groups. We assembled seven complete plastome sequences from seven species
a section of the journal of Dendroseris, including representatives from three subgenera, and carried out
Frontiers in Plant Science comparative phylogenomic analyses. The plastomes are highly conserved in gene
Received: 12 August 2020 content and order, with size ranging from 152,199 to 152,619 bp and containing 130
Accepted: 14 October 2020
Published: 05 November 2020 genes (87 coding genes, 6 rRNA genes, and 37 tRNA genes). Plastid phylogenomic
Citation: analyses based on both the complete plastome sequences and 81 concatenated
Cho M-S, Kim S-H, Yang JY, coding genes only show Dendroseris nested within Sonchus sensu lato, and also that
Crawford DJ, Stuessy TF,
López-Sepúlveda P and Kim S-C
inter-subgeneric relationships are fully resolved. Subg. Phoenicoseris is resolved as
(2020) Plastid Phylogenomics sister to the remaining species of the genus and a sister relationship between the two
of Dendroseris (Cichorieae; subgenera Dendroseris and Rea. Ten mutation hotspots from LSC and SSC regions and
Asteraceae): Insights Into Structural
Organization and Molecular Evolution variable SSRs are identified as potential chloroplast markers for future phylogenetic and
of an Endemic Lineage From the Juan phylogeographic studies of Sonchus and related groups.
Fernández Islands.
Front. Plant Sci. 11:594272. Keywords: Asteraceae, adaptive radiation, plastome evolution, insular woodiness, Juan Fernández Islands,
doi: 10.3389/fpls.2020.594272 critically endangered

Frontiers in Plant Science | www.frontiersin.org 1 November 2020 | Volume 11 | Article 594272

Cho et al. Plastid Genomes of Dendroseris

INTRODUCTION Arn., and D. pruinata (Johow) Skottsb. (Stuessy et al., 1984;

Sanders et al., 1987; Crawford et al., 1992, 1998). These unique
The currently circumscribed Sonchus subg. Dendroseris (D. Don) plants are now highly threatened with extinction, as they are
S.-C. Kim & Mejías (tribe Cichorieae; Asteraceae) includes 12 exceedingly rare with very small population sizes in the wild
highly threatened island endemics to the Juan Fernández and the (almost always fewer than 10 plants) and widely scattered
Desventuradas Islands in the Pacific Ocean. The Juan Fernández (Skottsberg, 1953; Stuessy et al., 1998). Faced with the threats
archipelago is composed of three volcanic islands, i.e., Robinson of extinction, all 11 Dendroseris species on the Juan Fernández
Crusoe (also known as Masatierra, located about 667 km west archipelago have been categorized as Critically Endangered in the
of continental Chile; 48 km2 ), Alejandro Selkirk (also known as wild, CE BI +2c, on the IUCN Red List of Threatened Species
Masafuera, 181 km further westward, 50 km2 ), and Santa Clara (Walter and Gillett, 1998).
(close to Robinson Crusoe, only 2 km2 ). Located 750 km north Dendroseris on the Juan Fernández Islands has been the
of the Juan Fernández archipelago, the Desventuradas Islands subject of numerous systematic and biogeographic studies
consist of the small islands of San Ambrosio and San Félix, along over several decades as one of the most striking examples of
with several small islets. adaptive radiation on oceanic islands in the Pacific Ocean (e.g.,
For several decades, these 12 species in subg. Dendroseris Skottsberg, 1953, 1956; Carlquist, 1967; Sanders et al., 1983,
have been recognized at the generic rank (Dendroseris D. Don 1987; Crawford et al., 1987, 1992; Sang et al., 1994; Anderson
and Thamnoseris F. Phil.) in subtribe Dendroseridinae Benth. et al., 2001; Bernardello et al., 2001). Despite the morphological,
(Stebbins, 1953; Bremer, 1994). However, recent molecular anatomical, and ecological divergence, Dendroseris has been
phylogenetic studies of Sonchus and related genera revealed strongly supported as a monophyletic group by several molecular
that Dendroseris (comprising 11 species on the Juan Fernández markers (Sanders et al., 1983, 1987; Crawford et al., 1987, 1992;
archipelago) is deeply embedded within the genus Sonchus Spooner et al., 1987; Pacheco et al., 1991; Sang et al., 1994; Kim
(Kim et al., 1996a,b, 2007). As a consequence, Dendroseris and et al., 1996a,b; Esselman et al., 2000). Although the monophyly
Thamnoseris were merged in Sonchus as one of four recognized of Dendroseris was well-established, phylogenetic relationships
subgenera in a newly delimited subtribe Hyoseridinae Less (Lack, among the three traditionally recognized taxonomic groups
2007; Kilian et al., 2009; Mejías and Kim, 2012). The genus and species remained uncertain probably due to their rapid
Sonchus is an ideal group to study patterns and processes of plant speciation. In addition, the phylogenetic position of Dendroseris
evolution given its diverse morphological and life history traits relative to other Sonchus groups has been elusive (Sang et al.,
(e.g., weedy annuals/biennials, herbaceous perennials, and woody 1994; Kim et al., 1996a,b, 1999, 2007). While the two subgenera
perennials) and peculiar geographic distribution (e.g., Atlantic Dendroseris and Phoenicoseris were resolved as monophyletic,
and Pacific Oceans, Africa, and cosmopolitan). subg. Rea was not and relationships among the three subgenera
Regardless of its taxonomic rank, the Dendroseris lineage were unresolved based on chloroplast (cpDNA) restriction
has attracted numerous taxonomists and evolutionary biologists fragment length polymorphisms (RFLPs), nuclear ribosomal
for nearly a century. Hereafter, while the lineage is nested DNA (nrDNA) internal transcribed spacer (ITS) sequences
within Sonchus, we treat Dendroseris in the traditional sense (Crawford et al., 1992; Sang et al., 1994), and allozymes (Crawford
as a distinc genus. Within the typically herbaceous family et al., 1998; Esselman et al., 2000). The neighbor-joining (NJ)
Asteraceae, Dendroseris shows spectacular and unusual tree- tree based on randomly amplified polymorphic DNAs (RAPDs)
like growth-forms with extremely variable morphology and showed the monophyly of each three subgenera, without further
occupies wide ecological ranges on the islands (Figure 1 and resolutions among them (Esselman et al., 2000). Lastly, the
Table 1). The spectrum of life form ranges from palmiform phylogenetic position of Dendroseris within the Sonchus group
rosette trees to sparsely branched rosette trees and succulent could not be determined confidently (Kim et al., 1999, 2007).
rosette shrubs (Carlquist, 1967). Initially, this diversity in In particular, the partial use of several coding and non-coding
growth forms led Skottsberg (1953) to recognize four segregate plastid regions in previous studies have been insufficient to
genera (Dendroseris, Rea, Hesperoseris, and Phoenicoseris), but provide robust phylogenetic relationships among Dendroseris
recognition of one cohesive genus is now followed by most and other closely related Sonchus groups (Sang et al., 1994; Kim
workers (e.g., Wodehouse, 1935; Stebbins, 1953). Anatomical et al., 1996a,b, 1999, 2007; Lee et al., 2005).
and molecular evidence provides support for a species-level Fortunately, massive amounts of data have now become
classification that concur with the three traditionally recognized available with the advent of high-throughput sequencing
subgenera within Dendroseris (Carlquist, 1967; Sang et al., technologies of next-generation sequencing (NGS) has
1994). Subsequent treatments have recognized three subgenera; revealed considerable genome-wide variation in sequences
(1) subg. Dendroseris Skottsb., comprising the four species and structures of entire plastid genomes. The benefits of
D. litoralis Skottsb., D. macrantha (Bertero ex Decne.) Skottsb., genome-wide data have increased phylogenetic resolution and
D. macrophylla D. Don and D. marginata Hook. & Arn., (2) significantly enhanced our understanding of plant evolution
subg. Phoenicoseris Skottsb., with the three species D. berteroana and diversity in the field of plastid genetics and genomics
(Decne.) Hook. & Arn., D. pinnata (Bertero ex Decne.) Hook. (Daniell et al., 2016). Whole plastome sequencing is now
& Arn., and D. regia Skottsb., and (3) subg. Rea (Bertero ex an efficient option for increasing phylogenetic resolution
Decne.) Skottsb., consisting of the four species D. gigantea at lower taxonomic levels that are currently hindered by
Johow, D. micrantha Hook. & Arn., D. neriifolia Hook. & limited sequence variation due to recent divergence, rapid

Frontiers in Plant Science | www.frontiersin.org 2 November 2020 | Volume 11 | Article 594272

Cho et al. Plastid Genomes of Dendroseris

FIGURE 1 | Species of Dendroseris from Robinson Crusoe Island in the Juan Fernández Archipelago, showing divergent morphological features. (A,B) D. litoralis
(subg. Dendroseris); (C,D) D. micrantha (subg. Rea); (E,F) D. berteroana (subg. Phoenicoseris). Photo Credit: Tod F. Stuessy (Ohio State University, Columbus).

radiation and conservative genome evolution of plastomes markers for further population genetic or phylogeograhic studies
(Parks et al., 2009). of Dendroseris.
In this study, we sequenced and assembled the whole plastid
genomes of seven species of Dendroseris, representing three
subgenera (Dendroseris, Rea, and Phoenicoseris) on the Juan MATERIALS AND METHODS
Fernández Islands. Based on the complete plastome sequences,
we tested the previous phylogenetic hypotheses proposed by Plant Materials and DNA Extraction
various molecular markers, specifically focusing on inter- Plant materials of seven Dendroseris species were collected
subgeneric relationships within the genus. We also performed previously in the field during four expeditions of the
comparative plastome analyses based on the phylogenetic Universidad de Concepción, Chile and the Ohio State University,
framework to determine the structure, gene content, and United States to the Juan Fernández Islands. As acknowleged in
rearrangements in the plastid genomes. Furthermore, we wanted the previous studies (Crawford et al., 1987, 1992; Pacheco et al.,
to identify highly variable plastid regions and microsatellites or 1991), the CONAF (Corporación National Forestal) of Chile
simple sequence repeats (SSRs), which could be utilized as useful issued the permission to collect in the Robinson Crusoe National

Frontiers in Plant Science | www.frontiersin.org 3 November 2020 | Volume 11 | Article 594272

Cho et al. Plastid Genomes of Dendroseris

TABLE 1 | Morphological and ecological characteristics of seven Dendroseris species sequenced and analyzed in this study.

Subg. Dendroseris Subg. Rea Subg. Phoenicoseris

Species D. litoralis D. macrantha D. marginata D. pruinata D. micrantha D. berteroana D. pinnata

Habit Branched rosette tree with solid stems Branched rosette shrub with Palmiform with hollow stem
solid and
diaphramed stem
Inflorescence Heads few, large (3–5 mm wide), orange Heads many, small, whitish Heads many, small, whitish
Achene Winged, large, thick-walled Non-winged, small, thin-walled Non-winged, small, thin-walled
Leaves Entire, thin to thick, glabrous Entire, elongate, very thick, glabrous Pinnate, elongate, thick, glabrous
Locality and Santa Clara and Robinson Crusoe Robinson Crusoe Robinson Crusoe Robinson Crusoe Robinson Crusoe Robinson Crusoe
habitat small rock off the Island in moist Island, cliffs at Island and Santa Island, common in Island, in wet forest Island, higher
south coast of forest higher elevations Clara, open areas, the east and central (extreme shade and ridges, at the limit
Robinson Crusoe either at low parts, in forests and moisture), tree/fern of the forest,
Island, coastal cliff, elevations near thickets at middle forests at higher among shrubs and
among coastal coast, or windy elevations, altitudes small trees
rocks in low islet cliffs at higher 400–600 m alt
elevations

Park. Our current samples of Dendroseris are representatives of Comparative Plastome Analyses and
those previous collections. For outgroup Reichardia ligulata, we Identification of Highly Divergent
used the same material as previously studied (Kim et al., 1996b,
2007). The fresh leaves were either dried (placed in sealable Regions
plastic bags with silica gel) or placed on ice and retained at 4◦ C We performed several comparative plastome analyses among
until extracted in the laboratory at The Ohio State University, Dendroseris and other closely related Sonchus species on the
Columbus, OH, United States. Total genomic DNAs were Canary Islands, Atlantic Ocean. The analyses also included two
extracted using the CTAB technique of Doyle and Doyle (1987), related Sonchus species (S. asper and S. canariensis). Sonchus
and purified in CsCl/ethidium bromide gradient. asper is a globally distributed herbaceous weed and S. canariensis
is an arborescent shrub endemic to the Canary Islands. Codon
usage frequency was calculated by using MEGA7 (Kumar et al.,
2016) with relative synonymous codon usage (RSCU) value,
Plastome Sequencing, Assembly, and which is the relative frequency of occurrence of the synonymous
Annotation codon for a specific amino acid. The online program predictive
Illumina paired-end (PE) genomic libraries with fragment size RNA editor for plants (PREP) suite (Mower, 2009) was used
of 550 bp were prepared and sequenced using the Illumina to predict the possible RNA editing sites for annotated protein-
HiSeq platform (Illumina, Inc., San Diego, CA, United States) coding genes with 35 reference genes available with known edit
at Macrogen Corporation (Seoul, South Korea). The sequence sites, based on a cutoff value of 0.8 (suggested as optimal for
contigs were assembled by the de novo genomic assembler, PREP-Cp). Overall sequence divergence was estimated using
Velvet 1.2.10 (Zerbino and Birney, 2008) at coverage ranging the LAGAN alignment mode (Brudno et al., 2003) in mVISTA
from 429 to 1489x. Annotation was performed using the (Frazer et al., 2004). Nucleotide diversity (Pi) was calculated using
Dual Organellar GenoMe Annotator (Wyman et al., 2004), the sliding window analysis (window length = 1000 bp and step
ARAGORN v1.2.36 (Laslett and Canback, 2004), and RNAmmer size = 200 bp excluding sites with alignment gaps) to detect
1.2 Server (Lagesen et al., 2007). Using Geneious v8.1.6 the most divergent regions (i.e., mutation hotspots) in DnaSP
(Biomatters, Ltd., Auckland, New Zealand), the draft annotation (Librado and Rozas, 2009).
was inspected and corrected manually, performing blast search
by comparison with homologous genes in Lactuca sativa Repeat Sequence Analysis
(DQ383816), S. canariensis (NC042381), S. acaulis (NC042382), Two types of repeat sequences were identified in the eight plastid
and S. webbii (NC042383) from the GenBank database at genomes of seven species of Dendroseris and R. ligulata. REPuter
the National Center for Biotechnology Information (NCBI) as (Kurtz et al., 2001) was used to detect the various types of
references. The complete plastome sequences were registered in repetitive sequences with search parameters set to: maximum
GenBank under the accession numbers MN893255 (R. ligulata computed repeats = 50, minimum repeat size = 8 bp, and
from Spain, Collection # TE); MK371014 (D. berteroana, hamming distance = 1. SSRs were identified using MISA web1
Collection # 12125); MK371011 (D. litoralis, Collection # 6435); with search parameters of 1–15 (unit size-minimum repeats,
MT157218 (D. marginata, Collection # 12068); MK371008 i.e., mono-nucleotide motifs with 15 minimum numbers of
(D. micrantha, Collection # 11944); MT157219 (D. pinnata, repetition), 2–5, 3–3, 4–3, 5–3, and 6–3 with 100 interruption
Collection # 11132); and MK371005 (D. pruinata, Collection # (maximum difference for two SSRs).
5108). OGDRAW (Lohse et al., 2013) was used to draw circular
plastid genome maps (Figure 2). 1
http://pgrc.ipk-gatersleben.de/misa/

Frontiers in Plant Science | www.frontiersin.org 4 November 2020 | Volume 11 | Article 594272

Cho et al. Plastid Genomes of Dendroseris

FIGURE 2 | Gene maps of the plastid genomes of seven Dendroseris and Reichardia ligulata sequenced and analyzed in this study. The other six Dendroseris
plastomes were consistent with D. litoralis in the structures and gene contents, but differed in their total length, having 152,263 bp for D. macrantha (MT157217),
152,261 bp for D. marginata (MT157218), 152,290 bp for D. pinnata (MT157219), 152,327 bp for D. micrantha (MK371008), 152,199 bp for D. berteroana
(MK371014), and 152,348 bp for D. pruinata (MK371005). The genes inside and outside of the circle are transcribed in the clockwise and counterclockwise
directions, respectively. Genes belonging to different functional groups are shown in different colors. The thick lines indicate the extent of the inverted repeats that
separate the genomes into small single copy (SSC) and large single copy (LSC) regions.

Phylogenetic Analysis similarity in sequences (99.8% among seven Dendroseris species

Phylogenetic relationships of the newly sequenced accessions of only) (Figures 2, 3). Within Dendroseris, the total length of
Dendroseris were investigated with other closely related Sonchus seven plastomes ranged from 152,199 (D. berteroana; subg.
species using R. ligulata as the outgroup. Seven representative Dendroseris) to 152,348 (D. pruinata; subg. Rea) base pairs
complete plastid sequences belonging to the major lineages of (bp), and consisted of four typical regions: large single copy
Sonchus were obtained from GenBank, including the woody (LSC), small single copy (SSC), and a pair of inverted repeat
Sonchus alliance species (Sonchus subg. Dendrosonchus) in (IR) regions. The overall guanine-cytosine (GC) content of each
Macaronesian Islands and globally occurring herbaceous weedy plastid genome was 37.6%, with LSC, SSC, and IR regions
species (Sonchus subg. Sonchus). In total, full sequences of 15 having 35.7–35.8, 31.1–31.3, and 43.1% GC contents, respectively
plastid genomes, including an outgroup taxon, R. ligulata, were (Table 2). Each of the eight cp genomes contained 130 genes,
aligned using MAFFT v.7 (Katoh and Standley, 2013). Maximum including 87 protein-coding genes (excluding pseudogenes), six
likelihood trees based on both the complete plastid genome rRNA genes, and 37 tRNA genes (Table 3). Eighteen genes
sequences and the concatenated sequences of 81 coding genes contained introns, including seven tRNA genes. Three genes of
(excluding six repeated in IR) were produced with 1000 replicate clpP, rps12, and ycf3 exhibited two introns. The trnK-UUU tRNA
bootstrap (BS) analyses by IQ-TREE (Nguyen et al., 2014). The gene harbored the largest intron, which contained the matK gene
best fit evolutionary model was chosen as TVM + F + I, which in between. In total, 17 genes were duplicated in the IR regions,
was scored according to the Bayesian information criterion (BIC) including seven tRNAs, three rRNAs, and seven protein genes.
scores and weights by testing 88 DNA models of ModelFinder The trans-splicing gene rps12, consisting of three exons, was
(Kalyaanamoorthy et al., 2017) implemented in IQ-TREE. located in the LSC region for exon 1, but exon 2 and exon 3 of
the gene were imbedded in the IR regions. Part of ycf 1 and rps19
duplicated in IR region were annotated as pseudogenes in all cp
RESULTS genomes sequenced in this study.

Gene Content, Order, and Organization Codon Usage and RNA Editing Sites
of the Plastomes of Dendroseris The frequency of codon usage was calculated based on the
Despite the great morphological and ecological differences sequences of protein coding genes, of which RSCU (the
among them, the seven plastomes of Dendroseris species relative frequencies of occurrence of the synonymous codon
and one outgroup taxon, R. ligulata were highly conserved usages for a specific amino acid) values were reported in
in gene content and arrangement, displaying 99.3% pairwise Supplementary Figure 1 (see Supplementary Table 1 for

Frontiers in Plant Science | www.frontiersin.org 5 November 2020 | Volume 11 | Article 594272

Cho et al. Plastid Genomes of Dendroseris

FIGURE 3 | Comparison of the border positions of the large single copy (LSC), small single copy (SSC), and inverted repeat (IR) regions among seven Dendroseris
and Reichardia ligulata plastid genomes. Gene names are indicated in boxes, and their lengths in the corresponding regions are displayed above the boxes. 9
indicates a pseudogene.

TABLE 2 | Genomic features of the complete plastid genomes of seven Dendroseris species and Reichardia ligulata sequenced and analyzed in this study.

Taxa Accession Total plastid size LSC size (bp)/ IR size (bp)/GC SSC size (bp)/GC No. of No. of protein No. of No. of rRNA
number (bp)/GC content GC content (%) content (%) content (%) genes coding genes tRNA genes genes
(%)

Reichardia (outgroup)
R. ligulata MN893255 152,620/37.6 84,205/35.7 24,945/43.1 18,525/31.1 130 87 37 6
Dendroseris subg. Dendroseris
D. litoralis MK371011 152,263/37.6 84,213/35.8 24,737/43.1 18,576/31.2 130 87 37 6
D. macrantha MT157217 152,263/37.6 84,213/35.8 24,73743.1 18,576/31.2 130 87 37 6
D. marginata MT157218 152,261/37.6 84,236/35.8 24,737/43.1 18,551/31.2 130 87 37 6
Dendroseris subg. Rea
D. pruinata MK371005 152,348/37.6 84,291/35.8 24,737/43.1 18,583/31.2 130 87 37 6
D. micrantha MK371008 152,327/37.6 84,287/35.8 24,738/43.1 18,564/31.2 130 87 37 6
Dendroseris subg. Phoenicoseris
D. berteroana MK371014 152,199/37.6 84,187/35.8 24,730/43.1 18,552/31.3 130 87 37 6
D. pinnata MT157219 152,290/37.6 84,258/35.8 24,732/43.1 18,568/31.2 130 87 37 6

LSC, large single copy region; SSC, small single copy region; IR, inverted repeat.

details). The 10 plastomes including seven Dendroseris, two (1.88) followed by UUA for leucine (1.81–1.82), while the lowest
Sonchus species (S. asper and S. canariensis), and outgroup taxon, one, AGC for serine (0.34–0.35) followed by CUG for leucine
R. ligulata showed very similar frequencies of codon usage despite (0.37–0.39) among all of them in common.
morphological and evolutionary divergence among them. We The total number of possible RNA editing sites predicted
found that all possible codons for all amino acids were used among 10 plastomes ranged from 93 to 104 sites in 35 protein-
in their plastomes as specified in Supplementary Table 1. The coding genes. Compared to the herbaceous species S. asper (98
highest RSCU value was in the usage of AGA codon for arginine sites) and R. ligulata (93 sites), most woody species of Dendroseris

Frontiers in Plant Science | www.frontiersin.org 6 November 2020 | Volume 11 | Article 594272

Cho et al. Plastid Genomes of Dendroseris

TABLE 3 | Genes present in the complete plastid genomes of seven Dendroseris species and R. ligulata sequenced in this study.

Category Gene name

Photosystem I psaA, psaB, psaC, psaI, psaJ, ycf3b , ycf4

Photosystem II psbA, psbB, psbC, psbD, psbE, psbF, psbH, psbI, psbJ, psbK, psbL, psbM, psbN, psbT, psbZ
Cytochrome b6/f complex petA, petBa , petD, petG, petL, petN
Cytochrome C synthesis ccsA
ATP synthase atpA, atpB, atpE, atpFa , atpH, atpI
RuBisCO rbcL
NADH oxidoreductase ndhAa , ndhBa,d , ndhC, ndhD, ndhE, ndhF, ndhG, ndhH, ndhI, ndhJ, ndhK
Large subunit ribosomal proteins rpl2a,d , rpl14, rpl16a , rpl20, rpl22, rpl23d , rpl32, rpl33, rpl36
Small subunit ribosomal proteins rps2, rps3, rps4, rps7d , rps8, rps11, rps12b,d,e , rps14, rps15, rps16a , rps18, rps19
RNA polymerase rpoA, rpoB, rpoC1a , rpoC2
Translation initiation factor infA
Others accD, cemA, clpPb , matKa
Unknown function genes (conserved reading frames) ycf1, ycf2d , ycf15d
Ribosomal RNAs rrn5d , rrn16d , rrn23d
Transfer RNAs trnA-UGCa,d , trnC-ACAa , trnC-GCA, trnD-GUC, trnE-UUC, trnE-UUCa,d , trnF-GAA, trnG-GCC, trnH-GUG,
trnK-UUUa , trnL-CAAd , trnL-UAAa , trnL-UAG, trnM-CAUC,d , trnN-GUUd , trnP-UGG, trnQ-UUG, trnR-ACGd ,
trnR-UCU, trnS-CGAa , trnS-GCU, trnS-GGA, trnS-UGA, trnT-GGU, trnT-UGU, trnV-GACd , trnW-CCA, trnY-GUA
a Gene containing a single intron.
b Gene containing two introns.
C Two gene copies in LSC.
d Two gene copies in IRs.
e Trans-splicing gene.

species and S. canariensis showed relatively more numbers of 0.0039). The SSC regions, where the most variable gene, ycf 1 was
RNA edited sites (102–104 sites) except for D. berteroana (99 located, showed the highest nucleotide diversity in both analyses
sites). The numbers of potential editing sites were not correlated (0.00573 for 10 plastomes vs. 0.00121 for seven of Dendroseris),
with gene length; the highest numbers of potential editing sites while the lowest value was in the IR boundary regions (0.00083
were found in the psaB gene (10–12 sites), followed by the ndhB vs. 0.0003). Ten divergence hotspots among 10 plastomes were
gene (9–10 sites). Only minor sites were predicted for longer suggested as potential plastid markers for phylogenetic studies
genes, i.e., two sites for the longest gene rpoC2 and one for the of Dendroseris species and closely related Sonchus groups. Eight
second longest rpoB. Any potential RNA edited sites were not intergenic regions (trnS-trnC, trnC-petN, trnT-trnL, trnL-trnF,
predicted for seven genes of atpF, petD, petG, petL, psaI, psbB, ndhC-trnC, psbE-petL, ycf 1-rps15, and rpl32-ndhF), one intron
and psbF consistently from all analyzed plastomes. R. ligulata region (rpl16 intron), and one protein coding region (ycf 1) were
differed at two more codon position in the psaB gene, from serine found in LSC and SSC regions (Figure 4). Additionally, seven hot
(S) to leucine (L), and from leucine (L) to phenylalanine (F), spots identified among Dendroseris plastomes were found in LSC
while S. asper changed one more codon position in the ndhB and SSC regions (Supplementary Figure 3).
gene, from leucine (L) to phenylalanine (F) when compared The result of mVISTA plotted against R. ligulata also exhibited
with the rest of the species of Dendroseris and Sonchus. Most a high degree of synteny and gene order conservation among
editing sites were distributed at the 2nd and 1st codon positions the plastomes of Dendroseris and Sonchus species (Figure 5).
(Supplementary Table 2). The highest conversions in editing A total of 1,873 polymorphic sites, which were identified in the
frequencies of codons associated with the corresponding amino DnaSP analysis, were visualized in mVISTA graph from mostly
acid changes were represented by the changes from proline (P) non-coding and intron regions, but also from several protein
to leucine (L) (average score of 21.606) followed by serine (S) to coding regions. The divergent coding genes of Dendroseris and
leucine (L) (average score of 19.12) (Supplementary Figure 2). Sonchus species against R. ligulata were atpA, accD, rpoA, ycf 2,
ycf 1, and ndhF.
Sequence Divergence and Hotspot
The divergence level of nucleotide diversity was compared using SSRs and Large Repeat Sequences
DnaSP in both ways among seven Dendroseris plastomes only vs. All of the eight plastid genomes of Dendroseris and R. ligulata
among 10 plastomes including reference plastomes of Sonchus contained very similar numbers and distribution patterns of
species (S. asper and S. canariensis) and the outgroup species repeated sequences. There were 71–74 SSRs detected by MISA
R. ligulata. Overall nucleotide diversity value (Pi) among 10 based on search parameters set for 1–15 (mono-nucleotide motifs
plastomes (average value of 0.00283, ranging from 0 to 0.01593) with 15 minimum numbers of repetition), 2–5, 3–3, 4–3, 5–
was much higher than the one comparing seven Dendroseris 3, and 6–3. The vast majority of the SSRs were tri-nucleotide
plastid genomes (average value of 0.00061, ranging from 0 to motifs (63–64 SSRs, 87%), while other repeat types were relatively

Frontiers in Plant Science | www.frontiersin.org 7 November 2020 | Volume 11 | Article 594272

Cho et al. Plastid Genomes of Dendroseris

FIGURE 4 | Ten most hotspot regions found in seven Dendroseris, two Sonchus (S. asper, and S. canariensis), and Reichardia ligulata.

fewer; one to two mono-nucleotide motifs (3%), three to four Dendrosonchus, Dendroseris, and Sonchus (100% BS) (Figure 8).
di-nucleotide motifs (5%), and three to four tetra-nucleotide However, phylogenetic relationships among three lineages within
motifs (5%) (Figure 6A). The most abundant repeat motif was Sonchus sensu lato and within Dendroseris were unresolved
‘AAT/ATT’ (30%) followed by ‘AAG/CTT’ (29%) in all eight between two ML trees. The complete plastome sequences
genomes (Figure 6B and Supplementary Table 3). Interestingly, suggested sister relationship between Dendroseris and Sonchus
SSRs were distributed most frequently in the coding regions (92% BS; Figure 8A), while protein-coding genes showed sister
(53%), followed by intergenic regions (42%), with much lower relationships between Dendroseris and Dendrosonchus (56%;
numbers found in the non-coding introns (5%) in each cp Figure 8B). In terms of relationships within Dendrosonchus from
genome (Supplementary Table 4). The coding regions with the Canary Islands, both trees showed that herbaceous perennial
highest number of SSRs were ycf genes; 10 SSRs (five duplicated species with tuberous roots, S. webbii, diverged first within the
in each IR) in ycf 2 and two in ycf 1 (in SSC). Considering the group and also that pachycaulous S. acaulis is sister to the tree
quadripartite regional occupancy of SSRs, the IR and SSC regions species S. canariensis (Figure 8). Within Sonchus, the newly
were lower in overall SSR frequency compared with the LSC described herbaceous perennial species S. boulosii from Morocco
region; 15% from the SSC region and 18% from each of both is sister to the cosmopolitan herbaceous weedy species S. asper
IR regions versus 62% from the LSC region (Supplementary and S. oleraceus.
Table 4). Additionally, we found 50 pairs of large repeats in each
cp genome (excluding duplicated IR region) using the parameters
of maximum computed repeats = 50, minimum repeat size = 8 bp, DISCUSSION
and hamming distance = 1 by REPuter. They contained 19–22
forward, 8–10 reverse, and 19–23 palindromic matches of repeats Highly Conserved Plastome Organization
(Figure 7A). Most of these large repeats were present in the and Evolution in Dendroseris
intergenic spacers, but five repeat matches were found within the The whole plastid genomes of Dendroseris species reported for
ycf coding genes, two in ycf 1 and three in ycf 2. Lengths of 21–25 the first time in this study were highly conserved structurally,
repeats were the most frequent (37%) followed by lengths of 19– sharing most common genomic features such as sequence
20 repeats (35%), while longer repeats of 26–30 (17%) and 39–58 similarity, gene content and numbers, and distributions of
(11%) were rarer than shorter ones (Figure 7B). repeated sequences despite their morphological and ecological
divergences. Interestingly, their genomic features are also very
Phylogenetic Analysis similar to other Sonchus groups with drastically different habits
The complete plastome sequences provide good resolution or geographical distribution (i.e., the woody Sonchus alliance
of inter-subgeneric and interspecifc relationships within of Dendrosonchus in the Macaronesian Islands, Atlantic Ocean
Dendroseris as well as resolving relationships among three major and the globally distributed weedy sow thistles of Sonchus sensu
lineages within Sonchus sensu lato (Figure 8). Two maximum stricto). Minor differences were found only in the total length
likelihood (ML) trees based on the complete sequences and of the herbaceous weedy species (Sonchus s. str.), which were
the protein-coding genes only confirmed the monophyly of shorter (151,808 ∼ 151,849 bp) than the woody shrubs and

Frontiers in Plant Science | www.frontiersin.org 8 November 2020 | Volume 11 | Article 594272

Cho et al. Plastid Genomes of Dendroseris

FIGURE 5 | Comparison of the plastid genomes of seven Dendroseris and two Sonchus species, against R. ligulata by mVISTA. Gray arrows indicate genes with
their orientation and position. Genome regions are color-coded as blue blocks for the conserved coding genes (exon), aqua blue blocks for introns, and orange
blocks for the conserved non-coding sequences in intergenic regions (CNS). Thick lines below the alignment indicate the quadripartite regions of genomes; LSC
region is in beige, IR regions, in pink, and SSC region, in light blue.

FIGURE 6 | Simple sequence repeat (SSR) number per distribution and repeat type of the plastid genomes of seven Dendroseris and Reichardia ligulata.
(A) Variation in the numbers of SSRs per distribution pattern. (B) Number of SSR motifs in different repeat motifs.

Frontiers in Plant Science | www.frontiersin.org 9 November 2020 | Volume 11 | Article 594272

Cho et al. Plastid Genomes of Dendroseris

FIGURE 7 | Repeat numbers per repeat type and repeat length of eight plastid genomes of seven Dendroseris and Reichardia ligulata. (A) Variation in the distribution
of forward, reverse, complement, and palindromic repeats. (B) Number of different repeat lengths.

FIGURE 8 | Maximum likelihood tree of Dendroseris and closely related Sonchus based on the complete plastome sequences (A) and 81 concatenated
protein-coding genes (B). Numbers above nodes are bootstrap values with 1000 replicates. The double slash (“//”) indicates that the branch lengths between
outgroup and ingroup are shortened by one third size to improve readability of trees. Newly sequenced eight plastid genomes in this study are marked with an
asterisk (*).

trees of Dendrosonchus (152,071 ∼ 152,194 bp) and Dendroseris individual plants differing only in the relative orientation of their
(152,199 ∼ 152,620 bp), the differences due mainly to the SSC single copy sequences (Palmer, 1983). Since the orientation of the
and LSC sizes (Cho et al., 2019a,b). Generally, the length of the SSC region is not standardized in GenBank (Walker et al., 2014,
plasid genome and its quadripartite regions varies among plant 2015), we reported one particular orientation of the SSC region
lineages due to the contraction and expansion of the inverted for the comparative analysis in Dendroseris, Dendrosonchus, and
repeat regions. Evaluating their contraction and expansion by Sonchus. All of them contained the functional protein coding
comparing the location of the boundaries among the four regions gene of ycf 1ψ at IR/SSC with its pseudogene copy, ycf 1ψ at
can provide some insights into plastid evolution (Menezes SSC/IR, and functional rps19 at LSC/IR with pseudogene copy
et al., 2018). The lengths of inverted repeats in the plastids rps199 at IR/LSC endpoints (Figure 3). Such a high conservation
of Dendroseris, Dendrosonchus, and Sonchus s. str. were very of plastid organization further corroborates that they share the
similar as a result of the complete duplication of: the genes most recent common ancestors, which has been previously
rpl2, rpl23, ycf 2, ycf 15, ndhB, rps7; exons 1 and 2 of rps12; all suggested (Kim et al., 1996a,b, 2007; Kilian et al., 2009).
three rRNA genes (5S, 16S, and 23S); and seven tRNA genes The frequency of codon usage vary by factors in species-
(trnLCAA , trnMCAU , trnVGAC , trnAUGC , trnEUUC , trnRACG , and specific ways, showing different preferences for the codons used
trnNGUU ). Furthermore, the boundaries among the four regions to encode specific amino acids, probably as the result of evolution
IRs, LSC and SSC were fairly conserved, sharing exactly the in the presence of mutational biases, selection for translation
same genes and similar gene contents at all junctions. The SSC rate and accuracy, and possibly other factors (Orešiè and
region has been shown to exist in two equimolar states within Shalloway, 1998). The patterns of frequently used codons were

Frontiers in Plant Science | www.frontiersin.org 10 November 2020 | Volume 11 | Article 594272

Cho et al. Plastid Genomes of Dendroseris

the same among Dendroseris, Dendrosonchus, and Sonchus, and Rea) as well as their relationships to other Sonchus groups
only showed slightly different RSCU (the relative frequencies of (Supplementary Figure 4).
occurrence of the synonymous codon usages for a specific amino
acid) values. RNA editing alters plastid transcripts by converting
specific cytidines to uridines, which results in a change in the Phylogenetic Relationships Within
amino acid sequence of the translated protein (Mower, 2009). Dendroseris
In this study, several protein-coding genes showed variances in Considering the overall high conservation among the plastomes
RNA editing prediction. The herbaceous weedy S. asper displayed of Sonchus and closely related species, it is not surprising that
less RNA editing at a total of 98 sites, especially for the genes previous studies based on relatively few sequences from the
rpl20 and matK, compared to the woody Dendrosonchus species, plastid genomes provided limited resolution of relationships.
S. canariensis (104 sites), and Dendroseris species (ranging from Although Dendroseris was confirmed to be monophyletic in a
99 to 104 sites). The dissimilarity in the RNA editing of the matK phylogeny, its relationships to other Sonchus groups and
cosmopolitan weedy S. asper may originate from its different inter-subgeneric relationships within Dendroseris were not well-
growth habit, but this needs further confirmation based on wider resolved (Kim et al., 2007). Here, the phylogenomic analyses of
sampling because the woody species of Dendroseris on the Juan entire plastid sequences revealed better resolution among the
Fernández Islands exhibited quite diverse RNA editing prediction three lineages sampled (i.e., Dendrosonchus, Dendroseris, and
including D. berteroana (99 sites) with very similar RNA editing Sonchus). In agreement with earlier results, in the present study
as herbaceous S. asper. As for SSRs and large repeat sequences, Dendroseris was nested deeply within the genus Sonchus sensu
Dendroseris showed patterns comparable to Dendrosonchus and lato on both ML trees (Figures 8A,B). Therefore, Dendroseris
Sonchus in the numbers and frequencies of repeat type (Cho was most likely derived from the Sonchus group, but its closest
et al., 2019a,b). In addition, all showed an abundance of tri- relatives or progenitors remain to be determined based on
nucleotide SSRs (over 80%), which is also consistent with other broader sampling of plastid sequences. Especially, to be desired
weedy Asteraceae species such as Ambrosia trifida (Sablok et al., are other Pacific Island endemics, and additional taxa of subg.
2019), and the highest frequencies of SSRs from the LSC region. Sonchus sections Maritimi and Arvenses because results of a
Furthermore, the similar distribution pattern of large repeats was prior study showed that species from these groups share the
also observed for majority in forward (F) and palindromic (P) most recent common ancestor with Dendroseris (Kim et al.,
matches from Dendroseris (F = 19∼22, P = 19∼23 out of a total 2007). Since all Dendroseris have a diploid chromosome number
of 50 pairs) and herbaceous weedy Sonchus (F = 21, P = 21 out of of 36 (Sanders et al., 1983) and there is no evidence of
total 49 pairs) (Cho et al., 2019b). multivalent formation (Sanders et al., 1983), suggestive of an
The nucleotide diversity (Pi value) was quite low within autopolyploid (Jackson, 1982; Ramsey and Schemske, 2002), it
Dendroseris (average 0.00061) and Dendrosonchus (0.00090), and can be considered allotetraploid. Given the higher frequency
increased to 0.00117 for Dendrosonchus and Sonchus s. str., of polyploid Sonchus species in the Pacific Ocean compared to
and to 0.00283 for Dendroseris, Dendrosonchus, and Sonchus typically diploid ones in the Atlantic Ocean and Old World, the
(Cho et al., 2019a,b). Given the highly conserved nature of geographical origin of Dendroseris is likely somewhere in the
the plastid genome in most angiosperms and the relatively Pacific Ocean (e.g., Australia, New Zealand, etc.) and adjacent
recent origin of Dendroseris species on the Juan Fernández regions (Kim et al., 2007). However, its origin is still enigmatic; it
Islands (estimate ranging from 800,000 to 2.6 million years) remains to be determined which lineages of Sonchus contributed
(Crawford et al., 1992; Sang et al., 1994; Daniell et al., 2016), to allotetraploid origin of Dendroseris.
the low nucleotide diversity within Dendroseris is not surprising. Unlike the uncertain position of Dendroseris relative to
Further, the overall patterns for highly variable regions were other Sonchus groups, this study provides strong evidence
similar among the taxa. The most divergent hot spot was of relationships among the three subgenera Dendroseris,
the ycf 1 region followed by rpl32-ndhF, trnT-trnL-trnF, and Rea, and Phoenicoseris. Previous plastid phylogenies based
psbE-petL, which were suggested as potential cp markers for on the matK gene, the psbA-trnH intergenic spacer and
phylogenetic studies of Sonchus and closely related groups restriction site mutations showed unresolved inter-subgeneric
(Figure 4 and Supplementary Figure 3). Of 10 highly variable relationships, presumably the result of rapid radiation and
regions based on complete plastid genomes across diverse speciation following the arrival of the common ancestor
angiosperm lineages (Shaw et al., 2014), only three regions, (Crawford et al., 1992; Kim et al., 1999, 2007). Other types
i.e., rpl32-trnL, psbE-petL, and ndhF-rpl32, were found to be of molecular markers, such as RAPDs (Esselman et al.,
also highly variable in Dendroseris. In this study, we identified 2000), ITS sequences (Sang et al., 1994), and allozymes
variable molecular markers including SSRs and highly variable (Crawford et al., 1998) failed to resolve relationships among
regions from plastid genomes, which will increase the efficiency the three subgenera. The initial divergence of subg. Dendroseris
and feasibility for species identification and phylogenetic (D. litoralis, D. marginata, and D. macrantha) in the maximum
reconstruction within Sonchus. The ML tree constructed from likelihood tree based on ITS sequences is not significant,
the concatenated sequences of 10 mutation hotspot regions leaving inter-subgeneric relationships unresolved. However,
(12,879 bp in length) demonstrated their effectiveness as potential species relationships within each subgenus have been strongly
molecular markers, and for resolving with high support values and consistently recognized regardless of different marker
inter-subgeneric relationships within Dendroseris (except subg. types. For example, a sister relationship between D. litoralis

Frontiers in Plant Science | www.frontiersin.org 11 November 2020 | Volume 11 | Article 594272

Cho et al. Plastid Genomes of Dendroseris

and D. marginata of subg. Dendroseris has been consistently characteristics of compound pinnatifid leaves, single-stemmed
recognized. Despite morphological and molecular similarities habit and being monocarpic (Sanders et al., 1987). In addition,
due to common ancestry, the species are ecologically and subg. Rea was regarded as the most primitive subgenus based
altitudinally well-differentiated: D. litoralis occurs in coastal on morphological features and occurrence in middle elevation
lower elevations and D. marginata is found on higher elevation forests, and more open, arid habitats (Sanders et al., 1987).
exposed cliffs (Stuessy et al., 1998). The second species pair The current study, however, does not necessarily support the
of subg. Rea, D. pruinata and D. micrantha, also shows hypothesis that Phoenicoseris is highly derived, which was
marked ecological and altitudinal differentiation: the former based on morphological features. Rather, the complete plastid
species commonly in middle elevation forests and the latter sequences strongly suggest that after initial divergence of the
one in open lower coastal areas or open higher elevation two major lineages, the subg. Phoenicoseris lineage speciated at
windy cliffs (Sanders et al., 1987). Lastly, the species pair higher elevations with modification in some life history traits
of subg. Phoenicoseris, D. beteroana and D. pinnata, are (e.g., monocarpic). A sister relationship between Dendroseris
morphologically and ecologically quite divergent despite sharing and Rea has never been postulated based on various molecular
the most recent common ancestor: the former in high altitude markers and, given their ecological preferences (Dendroseris in
tree/fern forests and the latter in open wind-swept ridges at lower elevations along drier seacoasts and cliffs in full sun versus
higher elevations. Therefore, all these species pairs suggest Rea in middle elevation in the edges of the cooler forests), it is
that spatial/ecological/altitudinal factors likely promoted the plausible that the evolution of these two subgenera progressed
divergence and speciation of this largest and fascinating group from lower to middle elevations of the island.
on the Juan Fernández Islands.
The current whole plastid phylogenomic study shows the
unresolved relationship of D. marginata within subg. Dendroseris.
The phylogenetic tree suggests that D. marginata is either CONCLUSION
sister to the clade containing the subgenera Dendroseris and
Rea (complete plastome sequence tree; Figure 8A) or is We characterized the first complete plastid sequences of seven
sister to the clade of D. litoralis and D. macrantha of subg. species of Dendroseris, the largest endemic genus on the Juan
Dendroseris (protein coding genes only tree; Figure 8B). Given Fernández Islands. As in most angiosperms, we found highly
that the subg. Dendroseris is monophyletic based on various conserved plastomes at the generic level, including gene order
molecular markers and morphology, the tree based on protein and content. Despite the recent origin of Dendroseris species
coding genes only most likely represents the most plausible on the islands and the low rate of plastome evolution, we
hypothesis of species relationships. In addition, our current achieved the first fully resolved phylogeny for the genus.
study suggested that D. litoralis is more closely related to Especially, noteworthy was the complete resolution with strong
D. macrantha than to D. marginata. It has been demonstrated support of relationships among the three subgenera. The plastid
that, based on morphological traits, D. macrantha is most closely phylogenomics strongly suggest early divergence of two major
related to D. macrophylla on the geologically younger island lineages, one consisting of subg. Phoenicoseris and the other the
Masafuera (Alejandro Selkirk), while flavonoid profiles provided clade comprised of subgenera Dendroseris and Rea. Although we
no specific insights (Pacheco et al., 1991). The current study achieved full resolution within the genus, questions remain such
strongly suggests a sister relationship between D. litoralis and as the geographical origin within several lineages of Sonchus in
D. macrantha (100% BS). However, this relationship contrasts the Pacific Ocean, the monophyly of subg. Rea, and the timing
with morphological traits: D. litoralis has shorter morphological of the origin and radiation of the lineage in the archipelago.
distance with D. marginata (two features) compared to the Our thorough characterization and comparative analyses among
distance with D. macrantha (four features) (Sanders et al., 1987; the plastid genomes have led us to discover several informative
Pacheco et al., 1991). It is probable that the close relationship mutation hot spots and variable SSR regions, which can be
of D. litoralis and D. macrantha based on the plastid genomes used to identify and characterize each individual in these highly
could be the result of past gene flow since D. macrantha is threatened and nearly extinct species of Dendroseris.
from cultivated material in the village and also morphology,
especially leaf margin, is suggestive of gene flow event (TF
Stuessy, personal observation).
Based on fully resolved inter-subgeneric relationships inferred DATA AVAILABILITY STATEMENT
in this study, we can hypothesize some processes with regard
to the divergence and speciation of Dendroseris on the Juan The datasets presented in this study can be found in online
Fernández Islands (Figure 8B). Upon arrival of the common repositories. The names of the repository/repositories
ancestor on the older Robinson Crusoe island, two major and accession number(s) can be found below: https:
lineages might have diverged; one lineage (subg. Phoenicoseris) //www.ncbi.nlm.nih.gov/genbank/, MN893255; https://
toward higher elevation and the other lineage containing the www.ncbi.nlm.nih.gov/genbank/, MK371011; https://www.ncbi.
two subgenera Dendroseris and Rea toward middle and lower nlm.nih.gov/genbank/, MT157217; https://www.ncbi.nlm.
elevations. Subgenus Phoenicoseris was considered to be highly nih.gov/genbank/, MT157218; https://www.ncbi.nlm.nih.gov/
derived within the genus based on morphology, with the genbank/, MK371005; https://www.ncbi.nlm.nih.gov/genbank/,

Frontiers in Plant Science | www.frontiersin.org 12 November 2020 | Volume 11 | Article 594272

Cho et al. Plastid Genomes of Dendroseris

MK371008; https://www.ncbi.nlm.nih.gov/genbank/, MK37 Foundation of Korea (NRF) (NRF-2019R1A2C2009841)

1014; https://www.ncbi.nlm.nih.gov/genbank/, MT157219. to S-CK.

AUTHOR CONTRIBUTIONS
ACKNOWLEDGMENTS
M-SC and S-CK designed the experiment. DC, TS, and PL-
S collected the samples. M-SC, S-HK, and JY performed the We thank the herbarium of the Ohio State University (OS) for
experiments and analyzed the data. M-SC drafted the manuscript. the permission to take leaf samples from specimens.
DC, TS, and S-CK revised the manuscript. All authors approved
the final version of the manuscript.
SUPPLEMENTARY MATERIAL
FUNDING
The Supplementary Material for this article can be found
This project was supported by the Basic Science online at: https://www.frontiersin.org/articles/10.3389/fpls.2020.
Research Program through the National Research 594272/full#supplementary-material

REFERENCES Frazer, K. A., Pachter, L., Poliakov, A., Rubin, E. M., and Dubchak, I. (2004).
VISTA: computational tools for comparative genomics. Nucleic Acids Res. 32,
Anderson, G. J., Bernardello, G., Stuessy, T. F., and Crawford, D. J. (2001). Breeding W273–W279.
system and pollination of selected plants endemic to Juan Fernández Islands. Jackson, R. C. (1982). Polyploidy and diploidy: new perspectives on chromosome
Am. J. Bot. 88, 220–233. doi: 10.2307/2657013 pairing and its evolutionary implications. Am. J. Bot. 69, 1512–1523. doi:
Bernardello, G., Anderson, G. J., Stuessy, T. F., and Crawford, D. J. (2001). A 10.1002/j.1537-2197.1982.tb13400.x
survey of floral traits, breeding systems, floral visitors, and pollination systems Kalyaanamoorthy, S., Minh, B. Q., Wong, T. K., von Haeseler, A., and Jermiin, L. S.
of the angiosperms of the Juan Fernández Islands (Chile). Bot. Rev. 67, 255–308. (2017). ModelFinder: fast model selection for accurate phylogenetic estimates.
doi: 10.1007/bf02858097 Nat. Methods 14, 587–589. doi: 10.1038/nmeth.4285
Bremer, K. (1994). Plants and Islands. Portland: Timber Press. Katoh, K., and Standley, D. M. (2013). MAFFT multiple sequence alignment
Brudno, M., Do, C. B., Cooper, G. M., Kim, M. F., Davydov, E., Green, E. D., et al. software version 7: improvements in performance and usability. Mol. Biol. Evol.
(2003). NISC comparative sequencing program. LAGAN and multi-LAGAN: 30, 772–780. doi: 10.1093/molbev/mst010
efficient tools for large-scale multiple alignment of genomic DNA. Genome Res. Kilian, N., Gemeinholzer, B., and Lack, W. L. (2009). “Cichorieae,” in Systematics,
13, 721–731. doi: 10.1101/gr.926603 Evolution and Biogeography of Compositae, eds V. A. Funk, A. Susanna,
Carlquist, S. J. (1967). Anatomy and systematics of Dendroseris (sensu lato). T. F. Stuessy, and R. J. Bayer (Vienna: International Association for Plant
Brittonia 19, 99–121. doi: 10.2307/2805268 Taxonomy), 343–383.
Cho, M. S., Kim, J. H., Kim, C. S., Mejías, J. A., and Kim, S.-C. (2019b). Sow thistle Kim, S.-C., Crawford, D. J., Francisco-Ortega, J., and Santos-Guerra, A. (1996a). A
chloroplast genomes: insights into the plastome evolution and relationship of common origin for woody Sonchus and five related genera in the Macaronesian
two weedy species, Sonchus asper and Sonchus oleraceus (Asteraceae). Genes islands: molecular evidence for extensive radiation. Proc. Natl. Acad. Sci. U.S.A.
10:881. doi: 10.3390/genes10110881 93, 7743–7748. doi: 10.1073/pnas.93.15.7743
Cho, M. S., Yang, J. Y., Yang, T. J., and Kim, S.-C. (2019a). Evolutionary comparison Kim, S.-C., Crawford, D. J., and Jansen, R. K. (1996b). Phylogenetic relationships
of the chloroplast genome in the woody Sonchus Alliance (Asteraceae) on the among the genera of the subtribe Sonchinae (Asteraceae): evidence from ITS
Canary Islands. Genes 10:217. doi: 10.3390/genes10030217 sequences. Syst. Bot. 21, 417–432. doi: 10.2307/2419668
Crawford, D. J., Sang, T., Stuessy, T. F., Kim, S.-C., and Silva, M. O. (1998). Kim, S.-C., Crawford, D. J., Jansen, R. K., and Santos-Guerra, A. (1999). The use of
“Dendroseris (Asteraceae: Lactuceae) and Robinsonia (Asteraceae: Senecioneae) a non-coding region of chloroplast DNA in phylogenetic studies of the subtribe
on the Juan Fernandez islands: similarities and differences in biology and Sonchinae (Asteraceae: Lactuceae). Plant Syst. Evol. 215, 85–99.
phylogeny,” in Evolution and Speciation of Island Plants, eds T. Stuessy and Kim, S.-C., Lee, C., and Mejías, J. A. (2007). Phylogenetic analysis of chloroplast
M. Ono (New York, NY: Cambridge University Press), 97–119. doi: 10.1017/ DNA matK gene and ITS of nrDNA sequences reveals polyphyly of the genus
cbo9780511721823.007 Sonchus and new relationships among the subtribe Sonchinae (Asteraceae:
Crawford, D. J., Stuessy, T. F., Cosner, M. B., Haines, D. W., Silva, M. O., and Baeza, Cichorieae). Mol. Phylogenet. Evol. 44, 578–597. doi: 10.1016/j.ympev.2007.03.
M. (1992). Evolution of the genus Dendroseris (Asteraceae: Lactuceae) on the 014
Juan-Fernandez Islands: evidence from chloroplast and ribosomal DNA. Syst. Kumar, S., Stecher, G., and Tamura, K. (2016). MEGA7: molecular evolutionary
Bot. 17, 676–682. doi: 10.2307/2419735 genetics analysis version 7.0 for bigger datasets. Mol. Biol. Evol. 33, 1870–1874.
Crawford, D. J., Stuessy, T. F., and Silva, M. O. (1987). Allozyme divergence and doi: 10.1093/molbev/msw054
the evolution of Dendroseris (Compositae: Lactuceae) on the Juan Fernandez Kurtz, S., Choudhuri, J. V., Ohlebusch, E., Schleiermacher, C., Stoye, J., and
Islands. Syst. Bot. 12, 435–443. doi: 10.2307/2419268 Giegerich, R. (2001). REPuter: the manifold applications of repeat analysis on
Daniell, H., Lin, C. S., Yu, M., and Chang, W. J. (2016). Chloroplast genomes: a genomic scale. Nucleic Acids Res. 29, 4633–4642. doi: 10.1093/nar/29.22.
diversity, evolution, and applications in genetic engineering. Genome Biol. 4633
17:134. Lack, H. W. (2007). “Cichorieae,” in The Families and Genera of Vascular Plants,
Doyle, J. J., and Doyle, J. L. (1987). A rapid DNA isolation procedure for small vol. 8, Flowering Plants, Eudicots, Asterales, eds J. W. Kadereit and C. Jeffrey
quantities of fresh leaf tissue. Phytochem. Bull. 19, 11–15. (Berlin: Springer), 180–199.
Esselman, E. J., Crawford, D. J., Brauner, S., Stuessy, T. F., Anderson, G. J., and Lagesen, K., Hallin, P., Rødland, E. A., Stærfeldt, H. H., Rognes, T., and
Silva, M. O. (2000). RAPD marker diversity within and divergence among Ussery, D. W. (2007). RNammer: consistent annotation of rRNA genes in
species of Dendroseris (Asteraceae: Lactuceae). Am. J. Bot. 87, 591–596. doi: genomic sequences. Nucleic Acids Res. 35, 3100–3108. doi: 10.1093/nar/gk
10.2307/2656603 m160

Frontiers in Plant Science | www.frontiersin.org 13 November 2020 | Volume 11 | Article 594272

Cho et al. Plastid Genomes of Dendroseris

Laslett, D., and Canback, B. (2004). ARAGORN, a program to detect tRNA genes Shaw, J., Shafer, H. L., Leonard, O. R., Kovach, M. J., Schorr, M., and
and tmRNA genes in nucleotide sequences. Nucleic Acids Res. 32, 11–16. doi: Morris, A. B. (2014). Chloroplast DNA sequence utility for the lowest
10.1093/nar/gkh152 phylogenetic and phylogeographic inferences in angiosperms: the tortoise
Lee, C., Kim, S.-C., Lundy, K., and Santos−Guerra, A. (2005). Chloroplast DNA and the hare IV. Am. J. Bot. 101, 1987–2004. doi: 10.3732/ajb.14
phylogeny of the woody Sonchus alliance (Asteraceae: Sonchinae) in the 00398
Macaronesian Islands. Am. J. Bot. 92, 2072–2085. doi: 10.3732/ajb.92.12.2072 Skottsberg, C. (1953). “The vegetation of the Juan Fernandez Islands,”
Librado, P., and Rozas, J. (2009). DnaSP v5: a software for comprehensive analysis in The Natural History of Juan Fernandez and Easter Island V.2,
of DNA polymorphism data. Bioinformatics 25, 1451–1452. doi: 10.1093/ ed. C. Skottsberg (Uppsala: Almqvist and Wiksells boktryckeri),
bioinformatics/btp187 793–960.
Lohse, M., Drechsel, O., Kahlau, S., and Bock, R. (2013). Skottsberg, C. (1956). “Derivation of the Flora and Fauna of Juan Fernandez and
OrganellarGenomeDRAW—A suite of tools for generating physical maps Easter Island,” in The Natural History of Juan Fernandez and Easter Island V.1,
of plastid and mitochondrial genomes and visualizing expression data sets. ed. C. Skottsberg (Uppsala: Almqvist and Wiksells boktryckeri), 193–427.
Nucleic Acids Res. 41, W575–W581. Spooner, D. M., Stuessy, T. F., Crawford, D. J., and Silva, M. O. (1987).
Mejías, J. A., and Kim, S.-C. (2012). Taxonomic treatment of Cichorieae Chromosome numbers from the flora of the Juan Fernandez Islands. II.
(Asteraceae) endemic to the Juan Fernández and Desventuradas Islands (SE Rhodora 89, 351–356.
Pacific). Ann. Bot. Fenn. 49, 171–178. doi: 10.5735/085.049.0303 Stebbins, G. L. (1953). A new classification of the tribe Cichorieae, family
Menezes, A. P. A., Resende-Moreira, L. C., Buzatti, R. S. O., Nazareno, A. G., Compositae. Madroño 12, 65–81.
Carlsen, M., Lobo, F. P., et al. (2018). Chloroplast genomes of Byrsonima Stuessy, T. F., Sanders, R. W., and Silva, M. O. (1984). “Phytogeography and
species (Malpighiaceae): comparative analysis and screening of high divergence evolution of the Juan Fernandez Islands: a progress report,” in Pacific basin
sequences. Sci. Rep. 8, 1–12. biogeography, eds F. J. Radovsky, P. H. Raven, and S. H. Sohmer (Lawrence,
Mower, J. P. (2009). The PREP suite: predictive RNA editors for plant KS: B. P. Bishop Museum), 55–69.
mitochondrial genes, chloroplast genes and user-defined alignments. Nucleic Stuessy, T. F., Swenson, U., Crawford, D. J., and Anderson, G. (1998). Plant
Acids Res. 37, W253–W259. doi: 10.1016/0168-9452(90)90180-v conservation in the Juan Fernandez archipelago, Chile. Aliso 16, 89–101. doi:
Nguyen, L. T., Schmidt, H. A., von Haeseler, A., and Minh, B. Q. (2014). 10.5642/aliso.19971602.04
IQ-TREE: a fast and effective stochastic algorithm for estimating maximum- Walker, J. F., Zanis, M. J., and Emery, N. C. (2014). Comparative analysis of
likelihood phylogenies. Mol. Biol. Evol. 32, 268–274. doi: 10.1093/molbev/ms complete chloroplast genome sequence and inversion variation in Lasthenia
u300 burkei (Madieae, Asteraceae). Am. J. Bot. 101, 722–729. doi: 10.3732/ajb.
Orešiè, M., and Shalloway, D. (1998). Specific correlations between relative 1400049
synonymous codon usage and protein secondary structure. J. Mol. Biol. 281, Walker, J. F., Zanis, M. J., and Emery, N. C. (2015). Correction to “Comparative
31–48. doi: 10.1006/jmbi.1998.1921 analysis of complete chloroplast genome sequence and inversion variation in
Pacheco, P., Crawford, D. J., Stuessy, T. F., and Silva, M. O. (1991). Flavonoid Lasthenia burkei (Madieae, Asteraceae)”. Am. J. Bot. 102, 1008–1008. doi:
evolution in Dendroseris (Compositae, Lactuceae) from the Juan Fernandez 10.3732/ajb.1500990
Islands, Chile. Am. J. Bot. 78, 534–543. doi: 10.1002/j.1537-2197.1991.tb15220. Walter, K. S., and Gillett, H. J. (1998). 1997 IUCN Red List of Threatened Plants.
x Gland: IUCN-The World Conservation Union.
Palmer, J. D. (1983). Chloroplast DNA exists in two orientations. Nature 301, Wodehouse, R. P. (1935). Pollen Grains, Their Structure, Identification and
92–93. doi: 10.1038/301092a0 Significance in Science and Medicine. London: McGraw-Hill Book Company,
Parks, M., Cronn, R., and Liston, A. (2009). Increasing phylogenetic resolution Inc.
at low taxonomic levels using massively parallel sequencing of chloroplast Wyman, S. K., Jansen, R. K., and Boore, J. L. (2004). Automatic annotation of
genomes. BMC Biol. 7:84. doi: 10.1186/1741-7007-7-84 organellar genomes with DOGMA. Bioinformatics 20, 3252–3255. doi: 10.1093/
Ramsey, J., and Schemske, D. W. (2002). Neopolyploidy in flowering plants. Annu. bioinformatics/bth352
Rev. Ecol. Evol. Syst. 33, 589–639. doi: 10.1146/annurev.ecolsys.33.010802. Zerbino, D. R., and Birney, E. (2008). Velvet: algorithms for de novo short read
150437 assembly using de Bruijn graphs. Genome Res. 18, 821–829. doi: 10.1101/gr.
Sablok, G., Amiryousefi, A., He, X., Hyvönen, J., and Poczai, P. (2019). Sequencing 074492.107
the plastid genome of giant ragweed (Ambrosia trifida, Asteraceae) from a
herbarium specimen. Front. Plant Sci. 10:218. doi: 10.3389/fpls.2019.00218 Conflict of Interest: The authors declare that the research was conducted in the
Sanders, R. W., Stuessy, T. F., Marticorena, C., and Silva, M. O. (1987). absence of any commercial or financial relationships that could be construed as a
Phytogeography and evolution of Dendroseris and Robinsonia, tree-Compositae potential conflict of interest.
of the Juan Fernandez Islands. Opera Bot. 92, 195–215.
Sanders, R. W., Stuessy, T. F., and Rodriguez, R. (1983). Chromosome numbers Copyright © 2020 Cho, Kim, Yang, Crawford, Stuessy, López-Sepúlveda and Kim.
from the flora of the Juan Fernandez Islands. Am. J. Bot. 70, 799–810. doi: This is an open-access article distributed under the terms of the Creative Commons
10.1002/j.1537-2197.1983.tb06415.x Attribution License (CC BY). The use, distribution or reproduction in other forums
Sang, T., Crawford, D. J., Kim, S.-C., and Stuessy, T. F. (1994). Radiation of is permitted, provided the original author(s) and the copyright owner(s) are credited
the endemic genus Dendroseris (Asteraceae) on the Juan Fernandez Islands: and that the original publication in this journal is cited, in accordance with accepted
evidence from sequences of the ITS regions of nuclear ribosomal DNA. Am. academic practice. No use, distribution or reproduction is permitted which does not
J. Bot. 81, 1494–1501. doi: 10.1002/j.1537-2197.1994.tb15635.x comply with these terms.

Frontiers in Plant Science | www.frontiersin.org 14 November 2020 | Volume 11 | Article 594272