In the present study, nucleotide sequences of the nrDNA ITS and the plastid DNA psbA-trnH region were used to reconstruct the phylogeny of the tribe Astereae in SW Asia using parsimony, Bayesian and likelihood methods. The ITS analysis showed that the SW Asian representatives of the tribe divide into two groups. One group arises just above the African lineage, while the second is part of the large Australasian polytomy at the crown of the tribe. SW Asian Aster, Crinitina, Galatella, Lachnophyllum and Psychrogeton appear to be non-monophyletic, whereas Chamaegeron, Eurasian Erigeron and Myriactis are monophyletic. Dichrocephala integrifolia is allied with S African members of subtribe Grangeinae. Chamaegeron and Lachnophyllum gossypinum are sister taxa and both are allied with the Bellis and Galatella group. Aster bachtiaricus is the earliest diverging branch of a large polytomy at the crown of the tribe. Psychrogeton species form three distinct clades. Heteropappus altaicus is nested in the Aster clade. Conyzanthus squamatus is nested within Symphyotrichum species.
Introduction
The Astereae is the second largest tribe of the Asteraceae, with 222 genera and c. 3100 species (Brouillet & al. 2009a). The tribe is characterized by deitate to triangular or lanceolate style appendages (Nesom & Robinson 2007). It is part of the subfamily Asteroideae and forms a clade with the Calenduleae, Gnaphalieae and Anthemideae, the last of which is often considered its sister tribe (Brouillet & al. 2009a; Funk & al. 2009).
Based on the recent classification of Nesom & Robinson (2007), members of the tribe present in SW Asia can be classified into eight subtribes: Asterinae (including Aster L., Crinitina Soják, Galatella Cass., Heteropappus Less., Psychrogeton Boiss. and Tripolium Nees); Bellidinae (Bellis L.); Conyzinae (Conyza Less. and Erigeron L.); Grangeinae (Dichrocephala L'Her. ex DC.); Homochrominae (Chamaegeron Schrenk and Lachnophyllum Bunge); Lagenophorinae (Myriactis Less.); Solidagininae (Solidago L.); and Symphyotrichinae (Conyzanthus Tamamsch.). However, Neobrachyactis Brouillet, a new genus of three species (type: Neobrachyactis roylei (DC.) Brouillet), was not attributed to a subtribe (Chen & al. 2011). In Flora of China, Chen & al. (2011) recognized Neobrachyactis as distinct from Brachyactis s.str. (the type of which, B. ciliata (Ledeb.) Ledeb., belongs to the North American genus Symphyotrichum Nees) in being glandular and having compressed, 2-ribbed achenes.
The first global phylogenetic analysis of the tribe based on nrDNA ITS sequence data was done by Brouillet & al. (2009a). The presence of two S African genera, Denekia Thunb. and Printzia Cass., at the base of the tree suggested that the tribe originated in Africa. The Chinese genus Nannoglottis Maxim. is the second lineage in the tribe. It is followed by dispersals to S Africa, South America, New Zealand and Australia. Next, an African lineage gave rise to Asian and Mediterranean members. Finally, the crown lineage of the Astereae includes Callistephus chinensis (L.) Nees, Australasian lineages, South American lineages and the North American clade.
Hitherto, many Asian members of the tribe (Chamaegeron, Lachnophyllum, Neobrachyactis and Psychrogeton) as well as the cosmopolitan Conyzanthus have not been phylogenetically studied using DNA sequences. Only few Eurasian representatives of some genera, especially Aster s.str. and segregate genera, were investigated in previous phylogenetic studies (Noyes & Rieseberg 1999; Brouillet & al. 2009a; Li & al. 2012).
The nuclear ribosomal internal transcribed spacer (ITS) and external transcribed spacer (ETS) and plastid sequence data have been used to infer phylogenetic relationships in Astereae (Morgan 1997; Noyes & Rieseberg 1999; Noyes 2000; Brouillet & al. 2001; Lowrey & al. 2001; Markos & Baldwin 2001; Cross & al. 2002; Fiz & al. 2002; Liu & al. 2002; Semple & al. 2002; Wagstaff & Breitwieser 2002; Lowell & al. 2003; Morgan 2003; Roberts & Urbatsch 2003, 2004; Urbatsch & al. 2003; Beck & al. 2004; Brouillet & al. 2004; Eastwood & al. 2004; Field & al. 2006; Watanabe & al. 2006; Selliah & Brouillet 2008; Andrus & al. 2009; Brouillet & al. 2009a, 2009b; Karaman-Castro & Urbatsch 2009; Vaezi & Brouillet 2009; Wagstaff & al. 2011; Li & al. 2012; Nakamura & al. 2012; Strijk & al. 2012).
The main goals of this study are (1) to investigate the phylogenetic relationships among SW Asian members of the tribe Astereae, and (2) to evaluate the monophyly and relationships of several species-rich genera of the tribe occurring in SW Asia (Aster, Chamaegeron, Erigeron and Psychrogeton) mainly based on nrDNA ITS sequence data.
Material and methods
Taxon sampling
The ITS sequences included 142 accessions representing 52 genera and 117 species, with 51 accessions newly sequenced (all from Iran except two from Turkey and one from Kazakhstan) and 91 obtained from GenBank. Achillea millefolium L. and Ursinia nana DC. (Anthemideae), Anaphalis margaritacea (L.) Benth. & Hook. f. (Gnaphalieae), Blumea brevipes (Oliv. & Hiern.) Wild (Inuleae) and Calendula officinalis L. (Calenduleae) were used as outgroups (Noyes & Rieseberg 1999; Funk & al. 2009).
The psbA-trnH sequences included 18 accessions representing six genera and 15 species, with 15 accessions newly sequenced (all from Iran) and three obtained from GenBank. Galatella litvinovii Novopokr. (Bellidinae) was used as an outgroup.
Leaf samples were obtained from specimens in the following herbaria in Iran (herbarium codes according to Thiers 2015+): Ferdowsi University, Mashhad (FUMH), Research Institute of Forests and Rangelands, Tehran (TARI), Tehran University (TUH) and Shahid Bahonar University, Kerman. Respective voucher information and GenBank accession numbers are listed in the Appendix.
DNA extraction, amplification and sequencing
Genomic DNA was extracted using the modified CTAB protocol of Doyle & Doyle (1987). Sodium metabisulfite (1 % w/v) was added to the DNA isolation buffer. The entire nrDNA ITS (ITS1—5.8S—ITS2) region was amplified by polymerase chain reaction (PCR) using either the universal primers AB 101 and AB 102 (Douzery & al. 1999) or ITS5m (Sang & al. 1995) and ITS4 (White & al. 1990). The psbA-trnH region was amplified using the psbA (Sang & al. 1997) and trnH (Tate & Simpson 2003) primers.
The PCR was performed in a 25 µL volume, containing 10 µl deionized water, 10× PCR Buffer with 1.5 mmol/L MgCl2 (Roche Diagnostics, Canada), 10 µmol/L of each dNTP, 0.5 µl of each primer (10 pmol/µl), 2.5 %–5 % DMSO, one unit of Taq DNA polymerase and 1 µl template DNA (20–70 ng).
PCR procedures for the nrDNA ITS region were 2:30 min at 94 °C for predenaturation, followed by 35–37 cycles of 94 °C for 1 min, 51–58 °C for 1 min, and 72 °C for 1 min plus a final extension of 72 °C for 7 min. For the psbA-trnH region, the PCR conditions were 2:30 min at 94 °C for predenaturation, followed by 35–37 cycles of 94 °C for 45s, 58 °C for 40s, and 72 °C for 1–1:10 min plus a final extension of 72 °C for 7 min. The quality of PCR products was checked by electrophoresis on a 1 % (w/v) agarose gel (using 1X TBE as the gel buffer) stained with ethidium bromide and then visualized under UV light. Amplified products were cleaned. Cleaned products were sequenced using the Bigdye terminator cycle sequencing ready reaction kit (Applied Biosystems, U.S.A.) with the appropriate primers in an ABI Prism 3730xl DNA sequencer (Applied Biosystems, U.S.A.).
Sequence alignment
Sequences were edited using BioEdit v. 7.0.5.3 (Hall 1999). Matrices were aligned initially using MUSCLE (Edgar 2004) and subsequently edited manually. Insertions and deletions (indels) were observed in all alignments and treated as missing data.
Table 1.
Dataset and tree statistics from analyses of nuclear and chloroplast regions.
Phylogenetic analyses
Maximum parsimony — Maximum parsimony analyses were conducted using PAUP* version 4.0M0 (Swofford 2002). The heuristic search option was employed with the following options: MulTrees, tree bisection reconnection (TBR) branch swapping with 100 random addition sequence replicates, and a maximum of 10000 trees retained.
To evaluate clade support, a bootstrap analysis was performed using a full heuristic search with 1000 replicates (Felsenstein 1985), each with simple addition sequence and TBR branch swapping, and a maximum of 100 trees retained per replicate.
Bayesian method — The most appropriate model and parameter estimates for Bayesian analyses were selected using the program MrModeltest version 2.0 (Nylander 2004) based on the Akaike information criterion (AIC) (Posada & Buckley 2004). On the basis of the Model-test results, datasets were analysed using the GTR+I+G model for nrDNA ITS and F81+G model for psbA-trnH sequences.
Bayesian analysis was performed using MrBayes 3.1.2 (Ronquist & Huelsenbeck 2003). The analysis was carried out with 10 million generations for ITS and 4 million generations for psbA-trnH, using the Markov chain Monte Carlo search. MrBayes performed two simultaneous analyses starting from different random trees (Nruns=2) each with four Markov chains and trees sampled at every 100 generations. The first 25% of trees were discarded as the burn-in. The remaining trees were then used to build a 50 % majority rule consensus tree accompanied with posterior probability values. Tree visualization was carried out using Tree ViewX version 0.5.0 (Page 2005).
Maximum likelihood method — Maximum likelihood analyses were performed for the datasets in raxmlGUI v. 1.3. (Silvestro & Michalak 2012). The model employed for each dataset is the same as that for Bayesian analyses. Parametric bootstrap values for maximum likelihood were calculated in raxmlGUI base on 1000 replicates with one search replicate per bootstrap replicate.
Results
Characterization of nucleotide data
The aligned ITS matrix comprises 142 sequences and 678 characters, including 350 (350/678=52 %) potentially parsimony-informative sites and 328 parsimony-uninformative ones.
For the psbA-trnH region, the matrix of 18 sequences contains 356 characters, of which 11 are potentially parsimony-informative sites and 345 are parsimony-uninformative. More information about the datasets and tree statistics from analyses of the nuclear and chloroplast regions is summarized in Table 1.
Phylogenetic analyses
ITS phylogeny (Fig. 1A & B)
Maximum parsimony, maximum likelihood and Bayesian methods gave very similar results. We here show only Bayesian trees along with posterior probability (PP) and bootstrap values from both maximum likelihood (LBS) and maximum parsimony (PBS) analyses. The Astereae form a monophyletic group (PP=1.00, LBS=93, PBS=93). The S African genus Printzia is sister to all other Astereae, followed by the Chinese genus Nannoglottis, which is sister to the remaining Astereae (PP=1.00, LBS=98, PBS=79). The following clades are successively recovered along the spine of the Bayesian tree: (1) the S African Mairia Nees and the New Zealand-paleo-South American clade (PP=0.79); (2) the African clade, the subtribe Homochrominae including Commidendrum robustum (Roxb.) DC., which is sister to Amellus strigosus (Thunb.) Less, and Felicia aethiopica (Burm. f.) Grau (PP=1.00, LBS=78, PBS=85); and (3) and a clade (PP=1.00, LBS=89, PBS=79) comprising the remaining Astereae. Within (3), the African Conyza gouanii (L.) Willd. and Lachnophyllum noeanum Boiss. are successive sisters to an unresolved polytomy of the African-S Asian subtribe Grangeinae, the Mediterrane - an-Eurasian Bellidinae (sensu Brouillet & al.), C Asian taxa including Chamaegeron and Lachnophyllum gossypinum Bunge, and a large polytomic clade comprising Aster bachtiaricus Mozaff. and the Australasian, South American and North American clades.
Dichrocephala integrifolia (L. f.) Kuntze is sister to Grangea maderaspatana (L.) Poir., Nidorella polycephala DC. and N. resedifolia DC. (PP=1.00, LBS=93, PBS=79). Crinitina , Galatella and Tripolium constitute a well-supported clade (PP=1.00, LBS=100, PBS=97), which is sister to Bellis annua L. and B. perennis L., this relationship being moderately supported (PP=1.00, LBS=54). Chamaegeron and Lachnophyllum gossypinum form a moderately supported clade (PP=0.95, PBS=66), which is sister to a clade comprising Bellis and the Galatella group (PP=0.51). Aster bachtiaricus is allied with the large polytomy at the crown of the Astereae (PP=0.96). Myriactis wallichii Less, is sister to M. humilis Merr. (PP=1.00, LBS=100, PBS=100); they form a strongly supported clade with Keysseria maviensis (H. Mann.) Cabrera (PP=1.00, LBS=100, PBS=100). The C Asian clade includes Psychrogeton species (excluding P obovatus (Benth.) Grierson, P. persicus (Boiss.) Grierson and P. pseudoerigeron (Bunge) Novopokr. ex Nevski) and is moderately supported (PP=0.99, LBS=64) by the ITS analyses. The three accessions of P obovatus form a well-supported clade (PP=1.00, LBS=100, PBS=100), which is sister to the Aster clade (PP=0.55). The two accessions of P. pseudoerigeron are closely related to Neobrachyactis roylei (PP=0.73). Aster ageratoides Turcz., A. koraiensis Nakai and P persicus form an unresolved subclade (PP= 0.99, LBS=58, PBS=59), which is sister to the subclade of A. hayatae H. Lév. & Vaniot and Heteropappus altaicus (Willd.) Novopokr. Conyzanthus squamatus is nested within Symphyotrichum. Erigeron acris L. and subspecies, E. caucasicus subsp. venustus (Botsch.) Grierson, E. hyrcanicus Bornm. & Vierh. and E. uniflorus L. and subspecies form a strongly supported clade (PP=1.00, LBS=99, PBS=99).
Fig. 1A.
Fifty percent majority rule consensus tree resulting from Bayesian analysis of nrDNA ITS dataset. Lower half of tree. Numbers above branches are posterior probability (PP) and bootstrap values from maximum likelihood (LBS) and maximum parsimony (PBS) analyses, respectively. Values <50 % are not shown. Region of origin: AF = Africa; ALT = Australia; CA = Central Asia; EA = East Asia; EUA = Eurasia; NA = North America; NAF = North Africa; NZ = New Zealand; SA = South America; SWA = Southwest Asia.
Fig. 1B.
Fifty percent majority rule consensus tree resulting from Bayesian analysis of nrDNA ITS dataset. Upper half of tree. For explanation see Fig. 1A.
psbA-trnH phylogeny (Fig. 2)
Neither the subtribe Asterinae sensu Nesom & Robinson (2007) nor sensu Brouillet & al. (2009a) seems to be monophyletic. Aster, Galatella, Psychrogeton, Solidago and Tripolium form an unresolved clade (PP=0.96, LBS=74, PBS=72). The phylogram includes: (1) Aster bachtiaricus; (2) Psychrogeton nigromontanus (Boiss. & Buhse) Grierson; (3) the Erigeron clade (PP=0.96, LBS=66, PBS=73), comprising E. annuus (L.) Pers. and E. strigosus Muhl, ex Willd.; (4) the clade comprising Solidago gigantea Aiton as sister to Aster alpinus L. and Psychrogeton persicus (PP=0.71, LBS=66 , PBS=61 ); (5) the Psychrogeton clade (PP=0.99, LBS=83, PBS=83), in which the species occupy unresolved positions; and (6) the Tripolium—Psychrogeton obovatus clade (PP=0.69, PBS=66).
Discussion
The nrDNA ITS tree for the Astereae in the present study (Fig. 1A & B) is topologically similar to that of Brouillet & al. (2009a). We sequenced and analysed several Asiatic taxa that were not included in that study. In contrast to that study, the Eurasian subtribe Bellidinae is sister to the Chamaegeron—Lachnophyllum gossypinum clade rather than to the Grangeineae. Also, an assemblage of Asian species of Aster and allies, the so-called Aster clade (Li & al. 2012), is in the large polytomy at the crown of the Astereae, but it does not group with any of the Australian or Asian species of the Australasian lineages. This is in agreement with the nrDNA ITS study of Li & al. (2012).
In the current analysis, the SW Asian representatives of the Astereae are generally divided into two groups: (1) representatives that are outside the crown lineages and are the first lineage to diverge after the Grangeinae, including Chamaegeron, Dichrocephala integrifolia, the Galatella group (or Bellidinae of Brouillet & al. 2009a, b), Lachnophyllum gossypinum and L. noeanum; and (2) representatives that are part of the large polytomy at the crown of the tribe, e.g. Aster, Heteropappus, Myriactis, Neobrachyactis and Psychrogeton occurring in the Australasian polytomy, and the other representatives, e.g. Conyzanthus and Erigeron (including two species of Conyza) nested in the North American clade. Below we examine the status of these genera within the Astereae in further detail.
The first genera derived after the African lineage
Lachnophyllum — This is a small C and SW Asian genus of two species (Nesom & Robinson 2007). The genus is characterized by recurved ray flowers and often lanate, long hairs and stipitate glands. The nrDNA ITS data did not support a close relationship between them. Lachnophyllum gossypinum is closely allied with Chamaegeron (see below), whereas L. noeanum appears in an isolated position above Conyza gouanii, as sister to an unresolved polytomy of the African subtribe Grangeinae, Mediterranean-Eurasian Bellidinae (sensu Brouillet & al. 2009a, 2009b), C Asian taxa including Chamaegeron and Lachnophyllum gossypinum and the large polytomic clade comprising Aster bachtiaricus and the Australasian-Asian, South American and North American clades (PP=0.66).
Chamaegeron — This includes four species distributed in C Asia, Iran, Afghanistan and Pakistan. With the exception of Chamaegeron oligocephalus Schrenk, the species are endemic to Iran, i.e. C. asterellus (Bornm.) Botsch., C. bungei (Boiss.) Botsch. and C. keredjensis (Bornm. & Gauba) Grierson. Chamaegeron has distinctive characters, is glabrous or pubescent with branching habit and basally connate pappus bristles falling as a unit. The Chamaegeron clade is strongly supported in the ITS tree (PP=1.00, LBS=100, PBS=97). Chamagaeron bungei is the basal species and is characterized by one series of ray florets and obovate leaves that are covered by hirsute and glandular hairs. It is followed by the clade consisting of C. asterellus, C. keredjensis and C. oligocephalus (PP=1.00, LBS=96, PBS=99). This clade is, in turn, divided into two subclades: (1) a subclade of two accessions of C. oligocephalus (PP=1.00, LBS=100, PBS=100); and (2) a subclade including C. asterellus as sister to two accessions of C. keredjensis (PP=1.00, LBS=79, PBS=86). Chamaegeron asterellus and C. keredjensis are covered by glandular, villous and hirsute hairs, whereas C. oligocephalus is glabrous.
Galatella group — According to Nesom (1994), the Galatella group includes Crinitina, Galatella and Tripolium. They constitute a well-supported clade in the ITS tree (PP=1.00, LBS=100, PBS=97). The present analysis is in agreement with the previous studies (Fiz & al. 2002; Brouillet & al. 2009a, 2009b; Li & al. 2012) regarding the sister-group relationship of the Galatella group with the Euro-Mediterranean genus Bellis (B. annua and B. perennis). Brouillet & al. (2009a) pointed out that Crinitina, Galatella and Tripolium should be included in Bellidinae rather than in Asterinae, which is also consistent with our analyses. The Galatella group comprises two well-supported subclades: (1) a subclade including G. scoparia (Kar. & Kir.) Novopokr. from Kazakhstan and two sister taxa, C. linosyris (L.) Soják and G. coriacea Novopokr.; and (2) a subclade comprising C. linosyris, C. villosa (L.) Soják, Galatella litvinovii, G. punctata (Waldst. & Kit.) Nees and T. pannonicum (Jacq.) Dobrocz.
Fig. 2.
Fifty percent majority rule consensus tree resulting from Bayesian analysis of plastid psbA-trnH dataset. Numbers above branches are posterior probability (PP) and bootstrap values from maximum likelihood (LBS) and maximum parsimony (PBS) analyses, respectively. Values <50 % are not shown.
Galatella is composed of 30 species (Nesom & Robinson 2007) and, as currently understood, the genus is not monophyletic. Crinitina is a small genus of 13 species (Nesom & Robinson 2007). The nrDNA ITS data show that Crinitina is also not monophyletic. Two accessions of C. linosyris were placed in two different subclades within the group, and were not allied with the congeneric C. villosa. It is worth noting that two ITS accessions of C. linosyris obtained from GenBank (DQ478987 and AF046949), determined by Karaman-Castro & Urbatsch (2009) and Noyes & Rieseberg (1999), respectively, are completely different from each other and placed in different subclades. Therefore, it appears that one of these accessions may be misidentified. Crinitina villosa is similar to Galatella scoparia in having yellow discoid capitula. The two accessions of C. villosa were not united, however, with G. scoparia.
Tripolium is a unispecific genus growing in salt-marshes, salt-marsh meadows and moist meadows across Europe and N Asia to America (Tamamschjan 1959; Nesom 1994). Nesom (1994) and Brouillet & al. (2009a, 2009b) noted the morphological similarities of Tripolium (including its distinctly corymboid capitulescence, herbaceous, broadly rounded, multinervate phyllaries, and tendency for raylessness) to Crinitina and Galatella and suggested that these taxa may be closely related. Our results support their hypothesis.
Dichrocephala—This contains ten species occurring in Africa and tropical countries according to Nesom & Robinson (2007), who defined it as a member of the Grangeinae. Dichrocephala integrifolia is characterized by lyrate-pinnatifid leaves and achenes without a pappus. It is allied with Grangea maderaspatana and the two species of Nidorella in the Grangeinae. In the present study, the inclusion of Chamaegeron and Lachnophyllum showed that the Bellidinae (including the Euro-Mediterranean genus Bellis and the Eurasian Galatella group) are not sister to the African Grangeinae, which disagrees with a previous study (Brouillet & al. 2009b).
Genera belonging to the large crown polytomy of the tribe
Psychrogeton
— This includes 20 species from Asia (Grierson 1967; Grierson & Rechinger 1982; Nesom & Robinson 2007). Grierson (1967) recognized Psychrogeton as an isolated genus due to the sterile (functionally male) disc achenes. However, two species (P. chionophilus (Boiss.) Krasch. and P. obovatus represent exceptions in that the disc florets appear to be fertile. Grierson indicated Pamir-Hindu Kush as the likely biodiversity centre of Psychrogeton, with dispersal occurring from there to C Asia, Afghanistan, Iran, NE Iraq, E Turkey and NE Syria. Both the ITS and psbA-trnH analyses do not support the monophyly of Psychrogeton. In the ITS tree, all species except P. obovatus, P. persicus and P. pseudoerigeron formed a clade (Psychrogeton s.str.). Psychrogeton obovatus is sister to the Aster clade. This species is one of the most isolated species of the genus (Grierson & Rechinger 1982). It is characterized by leaflike outer involucral bracts, obovate disc achenes (fertile) and coarsely toothed leaves, which are similar to those of Aster species. Psychrogeton persicus, represented by two accessions, along with Aster ageratoides and A. koraiensis, is deeply nested within the Aster clade. Psychrogeron pseudoerigeron is closely related to Neobrachyactis roylei. These species have similar geographical distributions and habitats. However, Neobrachyactis differs from P. pseudoerigeron in being an annual and having fertile disc florets. In addition, the numerous highly reduced ray florets of Neobrachyactis contrast with those of P. pseudoerigeron, which are much longer than the pappus. Psychrogeton s.str. is composed of three lineages. First, P. nigromontanus is sister to the remaining species (PP=0.99, LBS=64). It differs from the other taller species (e.g. P. aucheri (DC.) Grierson and P. pseudoerigeron) in its numerous obliquely cut tubular female ray florets with the corolla 
as long as the style. The second lineage is P. aucheri, represented here by five accessions. It is characterized by a 2-seriate pappus and ray-floret corollas as long as the style. The third lineage comprises five species: P. aellenii (Rech. f.) Grierson, P. alexeenkoi Krasch., P. amorphoglossus (Boiss.) Novopokr., P. cabulicus Boiss. and P. chionophilus.
Different accessions of P. aellenii, P. amorphoglossus and P. cabulicus did not cluster together. There were some single-nucleotide polymorphisms between the accessions of these species, maybe caused by hybridization. We analysed two accessions of P. amorphoglossus and P. cabulicus that had morphological differences (e.g. shape of leaves and density of hairs). The separation of P. amorphoglossus accessions in different clades can be evidence of hybridization. Considering the great diversity in chromosome numbers and ploidy levels in the Astereae (Ito & al. 1994; Brouillet & al. 2009a), it is possible that there are both diploid and polyploid populations for P. amorphoglossus and P. cabulicus. One accession of P. cabulicus is sister to P. amorphoglossus and an unresolved subclade of P. alexeenkoi, P. amorphoglossus and P. chionophilus. The latter three species are caespitose and monocephalous and restricted to C to S Iran.
The psbA-trnH analysis (Fig. 2) showed that P. aellenii, P. amorphoglossus, P. aucheri, P. cabulicus and P. pseudoerigeron form a well-supported clade (PP=0.99, LBS=83, PBS=83). However, P. nigromontanus occupies an unresolved position, and P. obovatus, as in the ITS tree, is distinct from other Psychrogeton species. Chloroplast data show a relationship between P. persicus and Aster alpinus (PP=0.78, LBS=60, PBS=55).
Asian Aster and allies — The ITS tree showed that the Aster clade did not group with any of the Australian (e.g. Brachyscome Cass., Calotis R. Br., Isoetopsis Turcz., Keysseria Lauterb., Kippistia F. Muell. and Olearia Moench) or Asian (Myriactis) species of the Australasian lineages, in agreement with the study of Li & al. (2012). Of 16 sampled species of Aster s.str., A. asteroides (DC.) Kuntze, A. bachtiaricus, A. diplostephioides (DC.) Benth. ex C. B. Clarke and A. flaccidus Bunge place in the lower half of the ITS tree (Fig. 1A) while the others place in the upper half of that tree (Fig. 1B). Considering the position of Kalimeris (Cass.) Cass., Heteropappus altaicus and Psychrogeton persicus, which are nested within the Eurasian Aster clade, our results confirm the hypothesis of Li & al. (2012) that Eurasian Aster is paraphyletic and polyphyletic.
Aster bachtiaricus, a woody perennial plant, is distinct from other Aster s.str. in the ITS tree and is isolated within the large polytomy at the crown of the Astereae (PP=0.96). It appears to be sister to the vast radiations that occurred in Australasia-Asia and in South and North America. On the psbA-trnH tree, A. bachtiaricus also has a distinct position and does not ally with A. alpinus.
Heteropappus — Molecular data support a close relationship between Aster s.str., Heteropappus, Kalimeris, Rhinactinidia Novopokr., Rhynchospermum Reinw. and Sheareria S. Moore, as has been found in previous studies (Ito & al. 1995, 1998; Noyes & Rieseberg 1999; Fiz & al. 2002; Brouillet & al. 2009a; Gao & al. 2009; Li & al. 2012). Previous studies (Ito & al. 1998; Li & al. 2012; Chen & al. 2011) showed that Heteropappus is embedded within Aster. Our ITS tree also supports the placement of Heteropappus in Aster. Heteropappus altaicus and A. hayatae form a subclade (PP=1.00, LBS=91, PBS=83) that is sister to the unresolved subclade including A. ageratoides, A. koraiensis and Psychrogeton persicus (PP=1.00, LBS=91, PBS=79).
Eurasian Erigeron — Grierson & Rechinger (1982) divided Erigeron into E. subg. Erigeron and E. subg. Trimorpha (Cass.) Popov. Erigeron subg. Trimorpha (including E. acris) with an intermediate series of eligulate female florets differs from E. subg. Erigeron (including E. caucasicus Steven, E. hyrcanicus and E. uniflorus) without this kind of intermediate florets. In the ITS tree, Erigeron forms a strongly supported clade. Erigeron acris subsp. lalehzaricus Rech. f. is sister to E. acris. Erigeron caucasicus subsp. venustus and E. uniflorus subsp. daenensis (Vierh.) Rech. f. form a moderately supported clade, but E. acris subsp. asadbarensis (Vierh.) Rech. f., E. acris subsp. pycnotrichus (Vierh.) Grierson, E. hyrcanicus and E. uniflorus subsp. elbursensis (Boiss.) Rech. f. occupy unresolved positions. Erigeron uniflorus (AF046988) and E. acris (AF118496), obtained from GenBank are successive sisters to the Eurasian Erigeron clade. It is noteworthy that some differences exist between the E. acris and E. uniflorus accessions obtained from GenBank and those from Iran newly sequenced here. Despite the existence of morphological differences, there are no marked differences between the sequences of species that belong to E. subg. Erigeron and E. subg. Trimorpha. It has been suggested that a northern migration of Erigeron species from North America to Eurasia occurred during the Pleistocene when land bridges may have facilitated intercontinental migration (Huber & Leuchtmann 1992; Noyes 2000). The present study appears to support this hypothesis.
Conyzanthus — Nesom (1994) treated Conyzanthus squamatus as Symphyotrichum squamatum (Spreng.) G. L. Nesom. Nesom & Robinson (2007) included Conyzanthus within Symphyotrichum subg. Symphyotrichum. Our results confirm Nesom's view. In the ITS tree, C. squamatus is well nested with Symphyotrichum species. Symphyotrichum squamatum is sister to the other species introduced in Asia: S. subulatum (Michx.) G. L. Nesom. Considering the position of S. ciliatum and that the clade includes S. squamatum and S. subulatum, it seems that the species of Symphyotrichum migrated to Asia at least two distinct times.
Myriactis — The ITS tree shows a strong relationship between the currently analysed Myriactis wallichii and the Australian representative, M. humilis (PP=1.00, LBS=100, PBS=100). Morphological characters, including campanulate heads and absent pappus, seem to confirm the relationships.
Conclusions
It appears that, due to the level of homoplasy and many parallel mutations in the ITS dataset, we observe polytomies and low resolution in the respective tree. The SW Asian representatives of the Astereae used in this study form two main groups in the ITS tree. One is composed of Lachnophylum, Chamaegeron and the Galatella group near the base of the tree; the second comprises Myriactis, Neobrachyactis, Psychrogeton, Aster, Heteropappus and Erigeron as representatives of the large polytomy at the crown of the tree. The current study does not support the monophyly of Aster, Galatella, Lachnophyllum and Psychrogeton in their current status. Aster bachtiaricus seems to be excluded from Aster. More taxon sampling and DNA sequence data, especially of rapidly evolving genes, are definitely needed to get a clear-cut picture of phylogenetic relationships among some members of the tribe, in particular Psychrogeton.
Acknowledgements
This research was supported by a research fund of Tarbiat Modares University. We are grateful to the curators of the Herbarium of the Research Institute of Forests and Rangelands (TARI), Herbarium of Ferdowsi University of Mashhad (FUMH), Central Herbarium of University of Tehran (TUH) and Shahid Bahonar University of Kerman for the loan of materials and permission to extract DNA from selected specimens. We thank Prof. Luc Brouillet (MT) and an anonymous reviewer for offering useful comments that helped improve an earlier draft of this paper.
References
Appendices
Appendix
The taxa included in the nrDNA ITS and psbA-trnH analyses are listed in this Appendix. For each sequence, a voucher or the source together with its GenBank accession number are cited. Herbarium codes are according to Thiers (2015+). A dash “-” indicates that a sequence was not available in GenBank.
Continued.
Continued.
Continued.
