Skip to main content
NCBI home page
Annals of Botany logoLink to Annals of Botany
. 2006 Oct;98(4):741–753. doi: 10.1093/aob/mcl157

Centaurea Revisited: A Molecular Survey of the Jacea Group

N GARCIA-JACAS 1,*, T UYSAL 2, K ROMASHCHENKO 3, V N SUÁREZ-SANTIAGO 4, K ERTUĞRUL 2, A SUSANNA 1
PMCID: PMC2806161  PMID: 16873424

Abstract

• Background and Aims The genus Centaurea has traditionally been considered to be a complicated taxon. No attempt at phylogenetic reconstruction has been made since recent revisions in circumscription, and previous reconstructions did not include a good representation of species. A new molecular survey is thus needed.

• Methods Phylogenetic analyses were conducted using sequences of the internal transcribed spacers (ITS) 1 and 2 and the 5·8S gene. Parsimony and Bayesian approaches were used.

• Key Results A close correlation between geography and the phylogenetic tree based on ITS sequences was found in all the analyses, with three main groups being resolved: (1) comprising the most widely distributed circum-Mediterranean/Eurosiberian sections; (2) the western Mediterranean sections; and (3) the eastern Mediterranean and Irano-Turanian sections. The results show that the sectional classification in current use needs major revision, with many old sections being merged into larger ones. A large polytomy in the eastern Mediterranean clade suggests a rapid and recent speciation in this group. Some inconsistencies between morphology and molecular phylogeny may indicate that hybridization has played a major role in the evolution of the genus.

• Conclusions Phylogenetic analysis of ITS has been useful in identifying the major lineages in the group, and unraveling many inconsistencies in the sectional classification. However, most recent groups in the eastern Mediterranean clade are not resolved and reticulation in the western Mediterranean group of sections makes phylogenetic relationships within these two groups somewhat obscure.

Keywords: Cardueae, Centaureinae, Centaurea, Centaurea jacea group, hybridization, ITS sequences, phylogeny, systematics

INTRODUCTION

The genus Centaurea L., with nearly 300 species (Bremer, 1994; Wagenitz and Hellwig, 1996), has traditionally been considered problematic and none of the early attempts to subdivide it (Cassini, 1829; Boissier, 1875; Löve and Löve, 1961; Dostál, 1969, 1973; Holub, 1973, 1974) was widely accepted. However, more recent molecular analyses of the genus and of subtribe Centaureinae, together with studies of morphology, pollen type and karyology, have enabled the natural limits of Centaurea to be established with greater confidence (Susanna et al., 1995; Wagenitz and Hellwig, 1996; Garcia-Jacas et al., 2000, 2001).

The nomenclatural consequences of establishing the natural limits of Centaurea were significant. The type species of the genus, Centaurea centaurium L. (section Centaurea), occupied an isolated position in the internal transcribed spacer (ITS) and matK phylogenetic trees, distant from the monophyletic group that includes the great majority of species in the genus (Garcia-Jacas et al., 2000, 2001).

In previous studies (Garcia-Jacas et al., 2000, 2001) it was suggested a new type species for the genus Centaurea, selected from the monophyletic core group, not from the current section Centaurea. This solution makes it necessary to change the name of only the 17 species of the section Centaurea, not for the numerous species in the core group. In line with this suggestions, Greuter et al. (2001) proposed conserving the name Centaurea with a new type species for the genus, Centaurea paniculata L., from the group Jacea. Under his proposal, the old section Centaurea would acquire the rank of genus, namely Rhaponticoides Vaillant (Greuter, 2003; Greuter et al., 2005). The old sections Acrocentron and Cyanus are likewise considered by Greuter et al. (2001) to be separate genera, namely Colymbada Hill and Cyanus Mill., respectively. In the authors' opinion, however, this separation of Acrocentron and Cyanus is not fully supported by the available morphological, molecular and karyological evidence (Garcia-Jacas et al., 2001). In view of the evident connections between Acrocentron/Cyanus and the sections of the group Jacea, it is suggested (Susanna and Garcia-Jacas, 2006) that all these sections should be maintained within the genus Centaurea, at least until there is firm evidence to the contrary. Partly coinciding with this circumscription, Greuter (2003) re-included Acrocentron in Centaurea at sectional rank, leaving Cyanus as a distinct genus.

The name ‘Jacea group’ denotes those taxa with the Jacea pollen type (Garcia-Jacas et al., 2000). Within this group there is marked among-section diversity in habit and in the morphology of bract appendages and cypselas, but the species of this group have various characters in common, including (a) the Jacea pollen type, (b) a lateral hilum (considered a derived character by Dittrich, 1968), and (c) a peculiar floral morphology with showy sterile peripheral florets lacking staminodes (Wagenitz and Hellwig, 1996). These latter two characters are shared with the sections Acrocentron and Cyanus. The Jacea group shows descending dysploidy, with chromosome numbers ranging from x = 12 to x = 7 (Garcia-Jacas et al., 1996).

Most species of the Jacea group are distributed in the eastern Mediterranean region and the Irano-Turanian regions, although some species are exclusive to the western Mediterranean region and others have a wider distribution. The widely distributed species are mostly associated with crop cultivation and livestock farming, and in some areas they are economically significant weeds (C. diffusa Lam., C. maculosa Lam. and C. nigra L.).

In a preliminary study of Garcia-Jacas et al. (2000), the Jacea group was found to be monophyletic, in contradiction with other authors who considered some sections to be separate genera (Chartolepis Cass., Cheirolepis Boiss., Cnicus L., Grossheimia Sosn. & Takht., Tomanthea DC.; cf. Klokov et al., 1963). Molecular phylogeny also allowed the possibility to be ruled out that Oligochaeta (DC.) C. Koch and Zoegea L. form part of this group, as already indicated by chromosome numbers (Garcia-Jacas et al., 1998a, b) and pollen types (Martín and Garcia-Jacas, 2000).

Studies have divided the Jacea group into three major clades, largely corresponding to geographical distributions: one clade with widely distributed species (circum-Mediterranean/Eurosiberian), a second clade of species from the western Mediterranean, and a third clade of species from the eastern Mediterranean and adjacent areas of the Irano-Turanian region. In addition, these studies indicate that the sectional classification requires a deep review. Some sections cannot be segregated from the others on the basis of the authors' analyses (specifically the sections Melanoloma and Seridia of the western clade and the sections Cynaroides and Paraphysis of the eastern clade). However, despite these previous findings (Garcia-Jacas et al., 2000), sampling limitations meant that several questions remained to be resolved.

The first question is the division into three major clades (circum-Mediterranean/Eurosiberian, western Mediterranean and eastern Mediterranean). Previous sampling of the circum-Mediterranean/Eurosiberian and eastern Mediterranean clades was limited, and it would be of interest to confirm that the three-clade classification is maintained when a wider range of species is sampled.

The second question to be resolved is that of the sectional classification, notably of the eastern Mediterranean group. We have noted a number of specific problems (e.g. the intercalation of sections Cynaroides–Paraphysis and Cheirolepis–Plumosipappus) in a previous report (Garcia-Jacas et al., 2001), and it remains unclear whether the current sections are natural entities.

The third unresolved question relates to the taxonomic placement of certain species. A clear case is that of C. chrysantha Wagenitz, the chromosome number of which (2n = 18; Romashchenko et al., 2004) is incompatible with its current assignment to the section Acrocentron, whereas it is the commonest number in the Jacea group. There is also a group of eastern species the sectional affiliations of which are unclear (C. antitauri Hayek, C. ensiformis Davis, C. isaurica Hub.-Mor.) or need confirmation, including C. cankiriensis A. Duran & H. Duman.

Following a period of intensive collecting in the Iberian Peninsula and the eastern Mediterranean, a study based on analysis of ITS region sequences, like previous ones, was carried out which had the following main aims:

  1. To verify whether the basically biogeographical groupings proposed in a previous study (Garcia-Jacas et al., 2000) are confirmed when species from a wider geographical range are considered, notably including the most representative species of the eastern Mediterranean and the circum-Mediterranean/Eurosiberian clades.

  2. To verify the existing sectional classification, and to investigate relationships between sections.

  3. To clarify the position of certain species the sectional affiliation of which is unclear.

MATERIALS AND METHODS

Plant material

Sampling focused on the species of Centaurea with the Jacea pollen type which constitute the Jacea group (Wagenitz and Hellwig, 1996; Garcia-Jacas et al., 2000). Included are 98 taxa of the Jacea group, representing 27 sections out of the 29 accepted by Wagenitz and Hellwig (1996); only the monotypic sect. Gymnocyanus from Morocco and the bitypic sect. Corethropsis from the Middle East are missing. The five outgroup species were chosen from Centaurea sect. Cyanus, which is sister to the Jacea group (Susanna et al., 1995; Garcia-Jacas et al., 2001). Voucher data, source and GenBank accession numbers for the ITS sequences are given in Table 1.

Table 1.

Origin of the materials, herbaria where the vouchers are deposited and GenBank accession numbers

Species Range Voucher ITS accession
Centaurea aggregata Fisch. & C. A. Mey. ex DC. Caucasus, Iran, Turkey (weed) Turkey, Adana: Ala Dağ above Dağdibi, 2000 m, Ertuğrul, Garcia-Jacas, Susanna 2305 & Uysal, 3·8·2002 (BC) DQ319077
Centaurea aladaghensis Wagenitz Turkey endemic Turkey, Adana: Ala Dağ above Dağdibi, 2000 m, Ertuğrul, Garcia-Jacas, Susanna 2304 & Uysal, 3·8·2002 (BC) DQ319078
Centaurea alba L. subsp. costae (Willk.) Dostál Iberian Peninsula endemic Spain, Huesca: Peña de Oroel, Fernández-Galiano & Rivas Goday 23733, 15·7·1947 (GDA) AM114325
Centaurea alba L. subsp. latronum (Pau) Dostál Iberian Peninsula endemic Spain, Ávila: La Adrada, Sánchez-Mata & Cantó 24946, 27·7·1982 (GDAC) AM114326
Centaurea albonitens Turrill Caucasus, Iran, Turkey Garcia-Jacas et al. (2000) DQ319081
Centaurea alexandrina Delile Egypt, Libia, Middle East Egypt, Alexandria: 20 km W Alexandria on the coast road, Susanna 1854 & Vilatersana, 8·6·1998 (BC) DQ319082
Centaurea amadanensis Schultz-Bip. Iran endemic Garcia-Jacas et al. (2000) DQ319083
Centaurea antitauri Hayek Turkey endemic Turkey, Adana: Ala Dağ above Dağdibi, 2000 m, Ertuğrul, Garcia-Jacas, Susanna 2306 & Uysal, 3·8·2002 (BC) DQ319084
Centaurea armena Boiss. Caucasus, Turkey Turkey, Van: Van, between Van and Gürpinar, Kurubaş pass, 2237 m, Ertuğrul 2940, 5·7·2003 (KNYA) DQ319085
Centaurea aspera L. W Mediterranean Susanna et al. (1995) DQ319086
Centaurea aucheri (DC.) Wagenitz Caucasus, Iran, Turkey Garcia-Jacas et al. (2000) DQ319087
Centaurea avilae Pau Iberian Peninsula endemic Spain, Ávila: Sierra de Gredos, Blanca 6087, 30·7·1979 (GDAC) AM114309
Centaurea aziziana Rech. f Turkey endemic Iran, Azarbayjan-e-Sharghi: between Tatar and Golfa, 85 km from Golfa, Garcia-Jacas, Mozaffarian, Susanna 1680 & Vallès, 7·8·1996 (BC) DQ319089
Centaurea behen L. Caucasus, Iran, Iraq, Middle East, Turkey Susanna et al. (2006) AY826250
Centaurea behen L. Caucasus, Iran, Iraq, Middle East, Turkey Armenia, Abovian: between Gehart and Garni, Fajvush, Gabrielyan, Garcia-Jacas, Guara, Hovannisyan, Susanna 1568, Tamanian & Vallès, 23·8·1995 (BC) DQ319090
Centaurea benedicta L. Widespread Nancy Botanical Garden (1993) DQ319091
Centaurea boissieri DC. subsp. boissieri Iberian Peninsula endemic Spain, Granada: Sierra de Cázulas, Blanca 6597, 8·6·1979 (GDAC) AM114278
Centaurea bruguierana (DC.) Hand.-Mazz. Caucasus, Iran, Turkey Garcia-Jacas et al. (2000) DQ319093
Centaurea cadmea Boiss. Turkey endemic Turkey, Burdur: 4 km from Burdur on the road to Sparta, mountains above Burdur, 1200 m, Ertuğrul, Garcia-Jacas, Susanna 2249 & Uysal, 28·7·2002 (BC) DQ319094
Centaurea calcitrapa L. Circum-Mediterranean Egypt, Alexandria: Alexandria road to Wadi Natrum, 130 km from Cairo, Susanna 1866 & Vilatersana, 9·6·1998 (BC) AY826252
Centaurea calolepis Boiss. Turkey endemic Turkey, Burdur-Muğla: Dirimli mountain pass, 1600 m, Ertuğrul, Garcia-Jacas, Susanna 2254 & Uysal, 29·7·2002 (BC) DQ319095
Centaurea cankiriense A. Duran & H. Duman Turkey endemic Turkey, Çankiri: Kalfat–Atkaracalar road, 1455 m, 40°41′49N and 33°05′27E, Uysal 524, 26·7·2003 (KNYA) DQ319096
Centaurea cariensis Boiss. Turkey endemic Turkey, Antalya: 40 km from Elmali on the road to Korkuteli, N slopes of the Karamanbeli mountain pass, 1400 m, Ertuğrul, Garcia-Jacas, Susanna 2258B & Uysal, 30·7·2002 (BC) DQ319097
Centaurea carratracensis Lange Iberian Peninsula endemic Spain, Málaga: Carratraca, Sierra de Aguas, Blanca 42802, 4·7·1998 (GDAC). AM114302
Centaurea cataonica Boiss. & Hausskn. Turkey endemic Turkey, Gaziantep: Gaziantep, at the entry of the town, 800 m, Ertuğrul, Garcia-Jacas, Susanna 2319 & Uysal, 4·8·2002 (BC) DQ319099
Centaurea cheirolepidoides Wagenitz Turkey endemic Turkey, Antalya: Elmali, from Tekke to Çiglikara, 1500 m, Uysal 1000, 28·6·2004 (KNYA). DQ319100
Centaurea cheirolopha (Fenzl) Wagenitz Lebanon, Turkey Turkey, Maraş: 40 km from Gaziantep on the Maraş road, 2 km from Narh, 500 m, Ertuğrul, Garcia-Jacas, Susanna 2324 & Uysal, 4·8·2002 (BC) DQ319101
Centaurea chrysantha Wagenitz Turkey endemic Turkey, Niğde: Ala Dağlar, track above the village of Demirkazik, 2000 m, Ertuğrul, Garcia-Jacas, Susanna 2298 & Uysal, 3·8·2002 (BC) DQ319102
Centaurea corymbosa Pourr. S France endemic France, Narbonne: La Clappe, M. Riba, 1995 (BC) DQ319103
Centaurea debdouensis Breitw. & Podlech Morocco endemic Morocco, Debdou: Gaada de Debdou, Pasquier & Ch. Rungs, 18·6·1954 (MPU) AM114317
Centaurea deflexa Wagenitz Turkey endemic Turkey, Konya: between Çukuryurt pass and Gevne valley, 25 km from Taşkent, 1700 m, Ertuğrul, Garcia-Jacas, Susanna 2274 & Uysal, 1·8·2002 (BC) DQ319105
Centaurea derderiifolia Wagenitz Turkey endemic Turkey, Sivas: between Gürün and Divrigi, 18 km from Divrigi, 1500–1600 m, Uysal 530, 39°16′53′N and 37°59′27E 28·6·2003 (KNYA) DQ319106
Centaurea deusta Ten. Italy endemic Italy, Calabria: Crotone, Torrente Matassa near Caccuri, 360 m, Vogt 15531, Berlin Botanical Garden, Index Seminum 1997 DQ319107
Centaurea diffusa Lam. Widespread (weed) Armenia, Talin : between vil. Pokr Arthik and Bagravan, Fajvush, Gabrielyan, Garcia-Jacas, Guara, Hovhannisyan, Susanna 1589, Tamanyan & Vallès, 26·8·1995 (BC) DQ319108
Centaurea diluta Aiton Iberian Peninsula, Morocco (weed) Morocco, Ksar es Souk: 10 km N of Midelt to Meknes, 1700 m, Garnatje, Susanna 1793 & Vilatersana, 17·6·1997 (BC) DQ319109
Centaurea donetzica Klokov Ukraine endemic Ukraine, Donetzkaya: Krasny Liman, Romashchenko, 12·8·2002 (BC) DQ319110
Centaurea drabifolia Sm. Irano-Turanian Garcia-Jacas et al. (2000) DQ319111
Centaurea ensiformis P. H. Davis Turkey endemic Turkey, Muğla: Köyceğiz district, Sandras Dağ range 13 km from Ağla, 1700 m, Ertuğrul, Garcia-Jacas, Susanna 2251 & Uysal, 29·7·2002 (BC) DQ319112
Centaurea exarata Boiss. ex Coss. Iberian Peninsula endemic Spain, Huelva: road A-983, Almonte to Matalascañas km 25, Roché & Susanna 1909, 9·7·1999 (BC) DQ319113
Centaurea fenzlii Reichardt Turkey endemic Turkey, Erzurum: between Erzurum and Varto, 1 km from Varto, 1650 m, Uysal 895, 31·7·2004 (KNYA). DQ319114
Centaurea gadorensis Blanca Iberian Peninsula endemic Spain, Almería: Sierra de Gádor, Pico La Estrella, 1730 m, Martínez Lirola & Salinas 44171, 29·7·1996 (GDAC) AM114298
Centaurea glastifolia L. Caucasus, Middle East, Turkey Garcia-Jacas et al. (2000) DQ319116
Centaurea glomerata Vahl Egypt, Middle East Egypt, Alexandria: Alexandria to Burg-el-Arab, Badr, Guara, Kamel & Vallès, 10·4·1995 (BC) DQ319117
Centaurea hermanni F. Herm. Turkey endemic Turkey: Saray to Istambul, 35 km from Saray, Baytop, 17·6·1972 (E) DQ319118
Centaurea hyssopifolia Vahl Iberian Peninsula endemic Spain, Toledo: near Ontígola, 500 m, Garcia-Jacas, Susanna 1600 & Vilatersana, 22·6·1996 (BC) DQ319119
Centaurea iberica Trev. ex Sprengel Caucasus, Iran, Turkey (weed) Turkey, Ankara: 20 km from Ankara to Eskisehir, 900 m, R. Ilarslan 4308 & H. Ilarslan, 5·8·1994 (ANK) DQ319120
Centaurea imperialis Hausskn. ex Bornm. Iran endemic Garcia-Jacas et al. (2000) DQ319121
Centaurea inexpectata Wagenitz Turkey endemic Turkey, Antalya: Gevne valley, above village Küçüklü, 1750 m, Uysal 598, 30·6·2004 (KNYA). DQ319122
Centaurea involucrata Desf. Algeria, Morocco Susanna et al. (1995) DQ319123
Centaurea isaurica Hub.–Mor. Turkey endemic Turkey, Konya: Bozki, Sorkun town, 2000 m, Uysal 509, 13·7·2003 (BC) DQ319124
Centaurea jacea L. Eurosiberian Spain, Barcelona: La Garrotxa, Pass of Bracons, 1100 m, Garcia-Jacas & Susanna 1593 (BC) DQ319125
Centaurea jaennensis Degen & Debeaux Iberian Peninsula endemic Spain, Jaén: Pozo Alcón, La Bolera dam, Blanca & Varo 6724, 19·6·1978 (GDAC) AM114287
Centaurea kotschyi (Boiss. & Heldr.) Hayek Caucasus, Iran, Turkey Garcia-Jacas et al. (2000) DQ319127
Centaurea kurdica Reichardt Turkey endemic Turkey, Elâziğ: road Elâziğ to Bingöl, 1 km from the cross to Alacakaya, 700 m, Ertuğrul, Garcia-Jacas, Susanna 2360 & Uysal, 6·8·2002 (BC) DQ319128
Centaurea linifolia L. Eurosiberian Garcia-Jacas et al. (2000) DQ319129
Centaurea lycopifolia Boiss. & Kotschy Turkey endemic Turkey, Adana: Osmaniye–Yarpuz road, 800 m, Uysal 534, 30·6·2003 (KNYA) DQ319130
Centaurea macrocephala Muss. Puschk. ex Willd. Caucasus, Turkey Armenia, Idjevan: N side of Sevan Pass, 1900 m, Fajvush, Gabrielyan, Garcia-Jacas, Guara, Hovhannisyan, Susanna 1519, Tamanyan & Vallès, 17·8·1995 (BC) DQ319131
Centaurea melitensis L. Widespread (weed) France, Montpellier: in front of museum ‘Agropolis’, Roché & Susanna 2013, 2·7·1999 (BC) DQ319132
Centaurea monticola Boiss. ex DC. Iberian Peninsula endemic Spain, Granada: Pantano del Cubillas, Blanca 6750, 6·6·1977 (GDAC) AM114313
Centaurea napifolia L. Italy, Sardinia, Sicilia Italy, Calabria: Crotone, La Castella, 80 m, Vogt 15542, Berlin Botanical Garden, Index Seminum 1997 DQ319135
Centaurea nemecii Náb. Iran, Turkey Garcia-Jacas et al. (2000) DQ319137
Centaurea nigra L. Eurosiberian Garcia-Jacas et al. (2000) DQ319138
Centaurea nivea (Bornm.) Wagenitz Turkey endemic Turkey, Eskişehir: Mihalıçık–Alpu road 19–20 km, 900–950 m, Uysal 519, 15·7·2003 (KNYA) DQ319139
Centaurea odyssei Wagenitz Greece, Turkey Turkey, Balikesir: Kaz mountains, Sarikiz hill, 1650–1675 m, Uysal 520,16·7·2003 (KNYA) DQ319140
Centaurea pallescens Delile Egypt, Libya Egypt, Alexandria: Alexandria between Amiriya and Burg-el-Arab, 20 km south of Burg-el-Arab, Susanna 1838 & Vilatersana, 7·6·1998 (BC). DQ319141
Centaurea paphlagonica (Bornm.) Wagenitz Turkey endemic Garcia-Jacas et al. (2000) DQ319142
Centaurea patula Boiss. Iran, Turkey Garcia-Jacas et al. (2000) DQ319143
Centaurea pectinata L. W Mediterranean Spain, Barcelona: Montseny, Santa Fe to Sant Marçal, 1300–1400 m, Garcia-Jacas & Susanna 1469, 6·7·94 (BC) DQ319144
Centaurea pinae Pau var. pinae Iberian Peninsula endemic Spain, Teruel: Puerto Ragudo, 900 m, Blanca, Socorro & Valle 6768, 15·7·1978 (GDAC) AM114310
Centaurea polyacantha Willd. Iberian Peninsula, Morocco Susanna et al. (1995) DQ319147
Centaurea polypodiifolia Boiss. Caucasus, Iran, Iraq, Syria, Turkey Garcia-Jacas et al. (2000) DQ319148
Centaurea proto-gerberi Klokov Ukraine endemic Ukraine, Luganskaya: Stanichno-Lugansk, Romashchenko, 5·9·2002 (BC) DQ319149
Centaurea pseudoleucolepis Kleopow Ukraine endemic Ukraine, Donetzkaya: Kamennye Mogily national reservation, Romashchenko, 1·8·2002 (BC) DQ319150
Centaurea pterocaula Trautv. Caucasus Garcia-Jacas et al. (2000) DQ319151
Centaurea ptosimopappa Hayek Turkey endemic Turkey, Seyhan-Hatay: Nur Dag, Gavur Dag, near Osmaniye 1000–1350 m, Wagenitz & H. J. Beug, 11·10·1957 (B) DQ319152
Centaurea ptosimopappoides Wagenitz Turkey endemic Turkey, Adana: Adana-Karsanti, Pos Ormonlari, Ormanalti, A. Savran, 25·7·02 (BC) DQ319153
Centaurea pullata L. Iberian Peninsula, Morocco Morocco, Ouarzazate: Dj. Mgoun, Amesker, Standinger & Finckh, 2·04·2001 (BC) DQ319154
Centaurea pungens Pomel Algeria, Morocco Morocco, Bouarfa: Bouarfa-Figuig, 65 km from Figuig, Garnatje, Susanna 1781 & Vilatersana, 16·6·1997 (BC) DQ319155
Centaurea resupinata Coss. subsp. resupinata Iberian Peninsula endemic Spain, Albacete: between Elche de la Sierra and Hellín, Cenajo dam, Blanca & Varo 6714, 6·7·1977 (GDAC) AM114288
Centaurea rhizantha C. A. Mey. Caucasus, Iran, Turkey Garcia-Jacas et al. (2000) DQ319157
Centaurea rigida Banks & Sol. Turkey endemic Turkey, Erzurum: 62 km from Erzurum on the road to Bingöl, 10 km from Çat, Ertuğrul, Garcia-Jacas, Susanna 2377 & Uysal, 6·8·2002 (BC) DQ319158
Centaurea saligna (C. Koch) Wagenitz Turkey endemic Turkey, Erzurum: between Karayazi and Göksu, 2200 m, Uysal 891, 31·7·2004 (KNYA) DQ319159
Centaurea sarandinakiae N. B. Illar Ukraine endemic Ukraine, Crimea: Planerskoe, Kara-Dağ mountain, Romashchenko, 16·8·2002 (BC) DQ319160
Centaurea sclerolepis Boiss. Turkey endemic Turkey, Diyarbakir: between Diyarbakir and Silvan, 38 km from Diyarbakir, 751 m, Uysal 898, 1·8·2004 (KNYA) DQ319161
Centaurea semijusta Juz. Ukraine endemic Ukraine, Crimea: Simferopol, Chatyr-Dağ mountain, Romashchenko, 1·9·2002 (BC) DQ319162
Centaurea solstitialis L. Widespread (weed) Spain, Tarragona: road from Xerta to Port del Compte, 5 km from cross-road, Garnatje & Susanna 1875, 6·7·1998 (BC) DQ319163
Centaurea spectabilis (Fisch. & C. A. Mey.) Sch. Bip. Caucasus, Turkey Turkey, Van: Van, between Van and Gürpinar, Kurubaş pass, 2237 m, Ertuğrul 2941, 5·7·2003 (KNYA) DQ319164
Centaurea spinosa L. Aegean Greece, Thrakia: Nomos Evrou, Samothraki, 2 m, Raus/Sch 18942, Berlin Botanical Garden, Index Seminum 1997 DQ319165
Centaurea stapfiana (Hand.-Mazz.) Wagenitz Turkey endemic Turkey, Diyarbakir: between Diyarbakir and Silvan, Yeşilköy village, hills of Kendal, 627 m, Uysal 900, 1·8·2004 (KNYA) DQ319166
Centaurea sterilis Stev. Ukraine endemic Ukraine, Crimea: Planerskoe, Kara-Dağ mountain, Romashchenko, 16·8·2002 (BC) DQ319167
Centaurea steveni M. Bieb. Caucasus, Iran Garcia-Jacas et al. (2000). DQ319168
Centaurea sulphurea Willd. W Mediterranean Morocco, Khenifra: Aguelmame Azigza, Standinger & Finckh, 6·6·2001 (BC) DQ319169
Centaurea thracica (Janka) Hayek Aegean, Balkans, Greece, Turkey Turkey, Balikasir: between Balikesir and Bursa, near Harmancık, 1005 m, Uysal 564, 30·6·2004 (KNYA) DQ319171
Centaurea tomentella Hand.-Mazz. Turkey endemic Turkey, Malatya: 4 km NW Doğansehir to Polat, 1000 m, Ertuğrul, Garcia-Jacas, Susanna 2353 & Uysal, 6·8·2002 (BC) DQ319172
Centaurea vankovii Klokov Ukraine endemic Ukraine, Crimea: Alupka, Ai-Petri mountain, Romashchenko, 30·8·2002 (BC) DQ319173
Centaurea virgata Lam. Turkey endemic Turkey, Muğla: Köyceğiz district, Sandras Dağ range 13 km from Ağla, 1700 m, Ertuğrul, Garcia-Jacas, Susanna 2252 & Uysal, 29·7·2002 (BC) DQ319174
Centaurea wiedemanniana Fisch. & C. A. Mey. Turkey endemic Turkey, Bilecik: Selimiye, between Osmaneli and Bilecik, 100 m, Davis & Coode, 1·7·1962 (E) DQ319175
Centaurea xylobasis Rech. f. Greece (Islands) Greece, Samos: Mt. Kerki, 1000–1400 m, Runemark & Nordenstam, 2·8·1960 (E) DQ319176
Outgroup
Centaurea cyanus L. Cosmopolitan Susanna et al. (2006). AY826254
Centaurea mollis Waldst. & Kit. E Europe Ukraine, Podolia: Lysa Hora, 2 km E of Vilshanitsa near Zolochiv, Boratyñski & Romo 0506D, 5·6·2000 (BC) DQ319133
Centaurea napulifera Rochel subsp. thirkei (Sch. Bip.) Dostál E Europe Romania, Constanta: Dobrogea, N from Cheia village, Cheia gorges of the Casimcea river, 130 m, Badarau 2·4·1996 (BC) DQ319136
Centaurea pinardii Boiss. Balkans, Turkey Turkey, Burdur: 4 km from Burdur on the road to Sparta, Askeriye outskirts, 950 m, Ertuğrul, Garcia-Jacas, Susanna 2244 & Uysal, 28·7·2002 (BC) DQ319146
Centaurea tchihatcheffi Fisch. & C. A. Mey. Turkey endemic Turkey, Ankara: Gölbasi, lake shore, 950 m, M. Vural, 30-5-2002 (BC) DQ319170
Open in a new tab

All the sequences that had been used in previous studies (Susanna et al., 1995; Garcia-Jacas et al., 2000) have been redone, with the addition of the 5·8S gene and the elucidation of ambiguities that characterized the old manually generated sequences.

DNA extraction, amplification and sequencing

Total genomic DNA was extracted following the 2xCTAB method of Doyle and Doyle (1987) as modified by Soltis et al. (1991) and Cullings (1992) from silica gel-dried leaves collected in the field or fresh leaves of plants cultivated in the Botanic Institute of Barcelona. In some cases, herbarium material was used.

Double-stranded DNA of the ITS region was mostly amplified using 17SE as the forward primer and 26SE as the reverse primer (Sun et al., 1994). The profile used for amplification included a hot start at 94 °C for 2 min, followed by 80 °C for 5 min, during which the polymerase (Ecotaq, Ecogen S. R. L., Barcelona, Spain) was added. Then 30 cycles of amplification were carried out under the following conditions: 94 °C for 1 min 30 s, 57 °C for 2 min and 72 °C for 3 min, with an additional extension step of 15 min at 72 °C. Some species were amplified using 1046F (Nickrent et al., 1994) as the forward primer and ITS4 (White et al., 1990) as the reverse primer. The profile used for amplification was the same as above, apart from the annealing temperature (55 °C). The PCR products were purified with the QIAquick PCR Purification Kit (Qiagen Inc., Valencia, CA, USA). Both strands were sequenced with 17SE and 1046F as forward primers and 26SE and ITS4 as reverse primers. Direct sequencing of the amplified DNA segments was performed using a BigDye Terminator Cycle Sequencing v3·1 (PE Biosystems, Foster City, CA, USA), following the manufacturer's protocol and analysed on an ABI PRISM 3730 DNA analyser (PE Biosystems). The sequences were edited using Chromas 1·56 (Technelysium Pty, Tewantin, Australia).

Phylogenetic analysis

Sequences were aligned visually by sequential pairwise comparison (Swofford and Olsen, 1990). The data matrix is available on request from the corresponding author.

Parsimony analysis involved heuristic searches conducted with PAUP version 4·0b4a (Swofford, 1999) using tree-bisection-reconnection (TBR) branch swapping with character states specified as unordered and unweighted. The indels were treated as missing data. All most-parsimonious trees were saved. To locate other potential islands of most-parsimonious trees (Maddison, 1991), 100 replications were performed with random taxon addition, also with TBR branch swapping. Bootstrap analysis (BS) was performed (Felsenstein, 1985) with 100 replicates with simple taxon addition and TBR branch swapping.

Bayesian inference estimation was calculated using MrBayes 3·01 (Ronquist and Huelsenbeck, 2003). The best-available model of molecular evolution required for Bayesian estimations of phylogeny was selected using hierarchical likelihood ratio tests and Akaike information criteria as implemented in the software MrModeltest 1·1b (Nylander, 2002), which considers only nucleotide substitution models that are currently implemented in PAUP and MrBayes 3·01 (Huelsenbeck and Ronquist, 2001; Ronquist and Huelsenbeck, 2003). The symmetrical model, with equal base frequencies with some sites assumed to be invariable and variable sites assumed to follow a discrete gamma distribution (SYM+I+G; Zharkikh, 1994), was selected as the best-fit model of nucleotide substitution for the dataset using both hierarchical likelihood ratio tests and Akaike information criteria. Bayesian inference analyses were initiated with random starting trees and were run for 2 × 106 generations. Four Metropolis-coupled Monte Carlo Markov chains were sampled every 100 generations, resulting in 20 000 sample trees. A critical aspect of the Bayesian analysis is to ensure that the Markov chain has reached stationarity. All sample points prior to stationarity are essentially random and are discarded as ‘burn-in’ 2000 sample trees because, usually, they do not contain useful parameter estimates. Internodes with posterior probabilities ≥95 % were considered statistically significant.

RESULTS

The numerical results of the analysis of the ITS region are detailed on Table 2. The topology of the trees produced by parsimony and Bayesian approaches was coincident, and Fig. 1 shows the parsimony Bayesian majority-rule consensus tree with the addition of the Bayesian and the bootstrap support values (PP, BS). One of the most-parsimonious trees is shown in Fig. 2.

Table 2.

Results from the ITS region

Data set ITS region
Number of taxa 103
Length range (bp) 629–635
Total characters 644
Number of informative characters 148
Number of most-parsimonious trees 2 31 208
Number of steps 422
Consistency index (CI) 0·4521
Retention index (RI) 0·8471
Range of divergence, ingroup (%) 0–9
Range of divergence, ingroup–outgroup (%) 5–10
Open in a new tab

Consistency and retention indices and divergence ranges were calculated excluding non-informative characters.

Fig. 1.

Fig. 1.

Open in a new tab

Majority-rule consensus tree based on Bayesian Monte Carlo Markov chains. Numbers above branches are Bayesian posterior probability percentages; numbers below branches are bootstrap percentages. P1, Cheirolepis-Pseudoseridia polytomy.

Fig. 2.

Fig. 2.

Open in a new tab

One of the most-parsimonious trees resulting from the ITS analysis.

The strict Bayesian majority-rule consensus tree produced from the ITS analysis (Fig. 1) supports the monophyly of the Jacea group (BS = 85 %, PP = 1·00). Within this clade, there are two main evolutionary branches: the first branch, which we will name ‘the circum-Mediterranean/Eurosiberian clade’, encompasses the sections with a wide distribution in Europe and western Asia (BS = 83 %, PP = 0·99); the second branch is composed of two clades: one clade, which will be named the ‘western Mediterranean clade’, is formed by the representatives of sections of western Mediterranean distribution (BS = 75 %, PP = 1·00); and the other, which will be named the ‘eastern Mediterranean clade’, comprises the sections which have an eastern Mediterranean and Irano-Turanian distribution (BS = 100 %, PP = 1·00).

DISCUSSION

In view of the confirmed close relationship between geographical distribution and phylogenetic patterns as inferred from nuclear DNA, considered here are the three major clades: the widely distributed clade, the western clade and the eastern clade.

The circum-Mediterranean/Eurosiberian clade

In all analyses this clade was placed as sister to the eastern and western Mediterranean clades (Fig. 1), without Bayesian support but with high bootstrap support (BS = 94 %). This group has a wide distribution from central–western Asia to the Iberian Peninsula, with many narrow endemics at the eastern and western extremes of the distribution.

As has been pointed out in a previous report (Garcia-Jacas et al., 2000), some sections are consistently associated in well-supported clades. The most important is the complex made up of the sections Acrolophus, Maculosae, Phalolepis and Pseudophalolepis (BS = 100 %, PP = 1·0), named Acrolophus-Phalolepis in Fig. 1. These four sections are morphologically similar, and, as noted by Wagenitz (1989), the separation of Acrolophus and Phalolepis is clearly artificial. The species of this group are either effective colonizers, e.g. C. diffusa, or narrow endemics, e.g. C. sarandinakiae N. B. Illar and C. pseudoleucolepis Kleopow.

The section Pseudophalolepis was described by Klokov et al. (1963) to include certain species from the Ukraine, on rather weak morphological grounds. The representative species of this section, C. donetzica Klokov, C. proto-gerberi Klokov and C. pseudoleucolepis, do not form a natural group and appear in different clades in the present analysis. For example, C. donetzica is included in a group with three other Ukrainian and Crimean microendemics of the section Phalolepis, namely C. sarandanakiae, C. semijusta Juz. and C. vankovii Klokov, with moderate bootstrap support (BS = 71 %) and high Bayesian support (PP = 1·00). These results argue strongly against the maintenance of Pseudophalolepis as a separate section, as is evident from morphology.

Similarly, there do not appear to be good grounds for maintaining section Maculosae (Dostál, 1973), since in the present analysis the two representatives of this section were deeply nested within other sections. Centaurea corymbosa Pourr. was placed in a polytomy with species of Acrolophus and Phalolepis, with which it shares a similar morphology. Centaurea exarata Boiss. ex Coss. fell in section Jacea (clade Jacea-Lepteranthus in the Fig. 1), although in this latter case the inclusion is not supported by morphology, since the morphology of C. exarata is unlike that of all other members of the genus. However, the chromosome number of C. exarata, x = 11 (Valdés-Bermejo, 1980), coincides with the basic number found in sect. Jacea, a fact that favours its inclusion in this section.

Section Willkommia, composed of Iberian and North African endemics and morphologically similar to section Acrolophus, forms another robust clade with strong bootstrap (100 %) and Bayesian (PP = 1·00) support. It is placed with the Acrolophus-Phalolepis group, but with weak statistical support (Fig. 1). The present results indicate that section Willkommia should be maintained, especially in view of the lack of support for its inclusion within the Acrolophus-Phalolepis group.

A surprising result of the present analysis is the placement of the two Iberian representatives of Phalolepis, C. alba L. subsp. alba and C. alba subsp. latronum (Pau) Dostál. Both are placed within the same clade as section Willkommia. Suárez-Santiago (2005) recently demonstrated the existence of ancient hybridizations between the sections Phalolepis and Acrolophus and section Willkommia: the nuclear genome of the Iberian species of the sections Acrolophus and Phalolepis does not bear any relation to that of other species of these sections, but shows considerable similarities with the nuclear genome of species of section Willkommia.

Another complex is made up of sections Jacea and Lepteranthus, all sampled species of which (C. hyssopifolia Vahl, C. jacea L., C. inexpectata Wagenitz, C. linifolia L., C. nigra L. and C. pectinata L.) are placed in a single clade with moderate bootstrap support (71 %) and strong Bayesian support (PP = 1·00). This clade also included C. exarata Coss., which according to Cosson (1850) belongs to section Acrolophus and according to Dostál (1973) to section Maculosae together with C. corymbosa (as has been seen, C. corymbosa was placed in the present analyses in the Acrolophus-Phalolepis group). The present results do not appear to support recognition of Jacea (represented in the present analyses by C. inexpectata, C. jacea and C. nigra) and Lepteranthus (represented in the present analyses by C. hyssopifolia, C. linifolia and C. pectinata) as separate sections. Likewise, there are no important morphological differences between Jacea and Lepteranthus, which were originally considered as separate sections only because of the slightly more lacerate bract appendices of species of Lepteranthus (Hayek, 1901). Centaurea inexpectata, the easternmost representative of this group studied here (it is a Turkish endemic), is placed as sister to the remaining species with moderate support (BS = 71 %, PP = 1·00).

The position of C. benedicta (L.) L. reflects a conflict between macromorphology and phylogenetic relationships as revealed by pollen type and DNA sequence analyses. It shows some striking differences (especially in achene morphology) that until recently were considered adequate grounds for consideration as a monotypic genus (as Cnicus benedictus L.). However, the Jacea pollen type (Wagenitz, 1955) found in this species was the first clue to its true relationships, which were later confirmed with DNA sequencing (Garcia-Jacas et al., 2000, 2001). Centaurea patula DC. shows a unique characteristic within the genus Centaurea: its chromosome number is x = 7, the lowest found to date in the genus (Garcia-Jacas et al., 1996).

The western Mediterranean clade

This group comprises the only sections found exclusively in the western Mediterranean, Melanoloma and Seridia (C. aspera L., C. involucrata Desf., C. polyacantha Willd. and C. pullata L.), together with other sections made up of widely distributed colonizing species (i.e. the Hymenocentron-Mesocentron group, which includes some well-known weeds, e.g. C. diluta Aiton, C. melitensis L., C. solstitialis L. and C. sulphurea Willd.). This clade has 75 % bootstrap support and high Bayesian support (PP = 1·00). It is the clade in which natural classification of sections is most difficult, no doubt due to frequent intersectional hybridizations (C. aspera × C. pullata is a common hybrid in the southern Iberian peninsula).

This clade also includes C. pungens Pomel, a North African species that has traditionally been placed in section Calcitrapa. The present results suggest that this species forms part of the Hymenocentron-Melanoloma-Seridia group, with which it shares a basic chromosome number of x = 11, unlike the species of the section Calcitrapa with x = 10 (Romashchenko et al., 2004).

The eastern Mediterranean clade

This group is by far the largest in the genus Centaurea, comprising about 200 species in the Caucasus, Greece, Iran and Turkey. As has been pointed out previously (Garcia-Jacas et al., 2000), section Calcitrapa, widely distributed in the western Mediterranean region, falls in this group (clade Calcitrapa in Fig. 1) as sister to the remaining taxa with strong support (bootstrap 100 %, PP = 1·00).

As in the western Mediterranean clade, the present results suggest a series of section groupings. The first comprises the sections Calcitrapa, Seridioides and Tetramorphaea (the Calcitrapa clade in Fig. 1). The species of these three sections make up a clade with strong support (bootstrap 100 %, PP = 1·00). Within this clade, C. bruguierana Hand.-Mazz., the only species described to date in section Tetramorphaea, is sister to the species of the section Calcitrapa. In the analyses of Garcia-Jacas et al. (2000), the species of sections Calcitrapa and Tetramorphaea were grouped together with strong bootstrap support. In the present study, with consideration of more species of the section Calcitrapa, the association of these two sections is confirmed. Also placed within this clade was C. glomerata Vahl, included by Wagenitz and Hellwig (1996) in the section Seridioides, and in the present study placed with C. pallescens Del. with weak bootstrap support (64 %) and moderate Bayesian support (PP = 0·97). The present results thus do not provide any support for the section Seridioides.

Sections Cynaroides and Paraphysis

These are the only two sections of the genus Centaurea with capitula grouped in a spike-like inflorescence; both sections also have highly characteristic entire hastate leaves with velvety indumentum. The species of these two sections fall into two distinct clades: one comprised of a polytomy involving various species of section Cynaroides (C. kurdica Reichardt, C. fenzlii Reichardt, C. imperialis Hausskn. and C. sclerolepis Boiss.) and the two species of section Paraphysis (C. amadanensis Sch. Bip. and C. nemecii Nábělek) with bootstrap support of 71 % and PP = 0·99. The other clade grouped the remaining species of section Cynaroides (C. aladaghensis Wagenitz, C. cataonica Boiss. and C. tomentella Hand.-Mazz.) with strong support (BS = 94 %, PP = 1·00). The two clades fall together (named Cynaroides in Fig. 1), but with weak support (BS = 71 %, PP = 0·96).

On morphological and karyological grounds it seems clear that the species of section Cynaroides (including the species of section Paraphysis) are closely related. It is thus surprising that the present analysis separates the species of this group into two different clades. The nine ITS sequences studied only show ten differences between the two clades, of which four are autoapomorphies, one is shared by two species and the remaining five are in line with the separation of these species into two independent clades.

Sections Pseudoseridia and Rhizocalathium

The two species of section Rhizocalathium included in the study, C. armena Boiss. and C. rhizantha C.A. Mey., are linked to C. steveni M. Bieb. from section Pseudoseridia with strong support (BS = 100 %, PP = 1·00). The species of section Rhizocalathium have an unusual basic chromosome number for the sections in the Jacea group, x = 8, shared with C. steveni. The present results indicate that section Rhizocalathium should be maintained, but with modified limits. However, the status of section Pseudoseridia is highly dubious. Centaurea cheirolepidoides, C. hermanii and C. lycopifolia are included in a major polytomy with the species of section Cheirolepis, C. steveni falls with the species of section Rhizocalathium and, finally, another representative of Pseudoseridia, C. cheirolopha (Fenzl) Wagenitz, forms a clade with C. ptosimopappoides Wagenitz (section Ptosimopappa) and C. chrysantha.

Section Microlophus

The present analysis included four species of section Microlophus: C. behen L., C. polypodiifolia Boiss., C. rigida Willd. and C. thracica (Janka) Hayek. Only two of these species, C. polypodiifolia and C. rigida, both with the same basic chromosome number x = 8, form a group with strong support (BS = 100 %, PP = 1·00). Centaurea behen L., for which two Armenian populations were sampled, was placed with this group, but without statistical support (clade named Microlophus in Fig. 1). Centaurea behen has an unusual chromosome number for the Jacea group, 2n = 34 (Romashchenko et al., 2004) and for subtribe Centaureineae. This chromosome number could be explained if this species were a allopolyploid hybrid between a species with x = 8 and another with x = 9. According to ITS phylogeny, the lack of relationship between the samples of C. behen and the remaining species of the eastern clade suggests that it may have a hybrid origin. Centaurea thracica, with x = 9, is placed in a clade with the species of the complex Cheirolepis-Pseudoseridia-Pteracantha-Plumosipappus, which likewise have a chromosome number of x = 9. These species are grouped with C. odyssei Wagenitz and C. xylobasis Rech. f. of section Pteracantha. This entanglement of sections confirms the difficulties of establishing a sectional classification based on bract appendices, as previously suggested by Ertuğrul et al. (2004). More studies are needed in section Microlophus, especially on the correlation between karyology and systematics.

The Cheirolepis-Pseudoseridia-Pteracantha-Plumosipappus complex

These four sections were placed in a single clade within a polytomy (marked P1 in Fig. 1) with a bootstrap value of 88 % and PP = 1·00. The species are well characterized morphologically, but not at the molecular level: the sequences show practically no differences, and the differences that do exist are mostly autoapomorphies. This probably reflects a recent and rapid speciation, so that there are morphological differences but practically no genetic differences. The same occurs in other groups of the Centaureineae in which speciation is thought to have been recent and rapid, such as Cheirolophus Cass. (Susanna et al., 1999) or Centaurea section Willkommia (Suárez-Santiago, 2005).

Data from ITS thus support merging all these sections into a single section, and morphology supports this option (Ertuğrul et al., 2004). Although there are differences in pappus and bract morphology, these differences lie within the range of variability that can be found in other sections such as Melanoloma and Seridia (Ertuğrul et al., 2004).

In this clade are also found C. odyssei and C. xylobasis (section Pteracantha) and C. thracica (section Microlophus), which are grouped together with low bootstrap support but high Bayesian support (PP = 0·99). In contrast, other species that have previously been included in section Pseudoseridia were placed in other clades, as noted above. Centaurea saligna, traditionally placed in section Cheirolepis, is not grouped with the rest of taxa of this section. According to other studies (T. Uysal, pers. comm.), C. saligna may be located as one of the first branches in the clade Cheirolepis-Pseudoseridia-Pteracantha-Plumosipappus.

Section Ptosimopappus

The present analysis includes the only two species described in this section, C. ptosimopappa Hayek and C. ptosimopappoides. These two species are not grouped together in the analysis, indicating that this is not a natural section. Species in section Ptosimopappus have an unusual shrubby habit, and on the basis of this were at one time considered as a separate genus, Ptosimopappus Boiss. Wagenitz (1975) had already pointed out that C. ptosimopappoides is morphologically intermediate between Ptosimopappus and Microlophus. The present results rule out any relationship between C. ptosimopappa and C. ptosimopappoides, and confirm the isolation of a monotypic section Ptosimopappus (inluding only C. ptosimopappa), which in the present analysis is not exclusively related to any other.

Section Phaeopappus

Four representatives of this section were included, namely C. albonitens Turr., C. aucheri (DC.) Wagenitz, C. spectabilis and C. stapfiana (Hand.-Mazz.) Wagenitz. This section has often been recognized as a distinct genus, Tomanthea, especially by the Russian school. Gabrieljan and Aghababian (1990) considered that it should be divided into two different genera, Chrysopappus Takht. [C. stapfianus (Hand.-Mazz.) Takht.] and Tomanthea (remaining species). The generic distinction of Tomanthea was rejected on molecular grounds by Garcia-Jacas et al. (2000). Chrysopappus also belongs in Centaurea (Fig. 1). In the present results, the section appears to be fragmented (Fig. 1): only C. albonitens and C. aucheri were linked with strong support (BS = 86 %, PP = 1·00). On the one hand, this fragmentation could support the views of Gabrieljan and Aghababian (1990) that Chrysopappus is distinct, if only at the sectional rank. On the other hand, morphological connections between all the species of section Phaeopappus are rather clear (Wagenitz, 1963) and, as in the above-mentioned cases of inconsistencies between morphology and molecular data, a possible explanation for this is past hybridization.

Sections Chartolepis, Grossheimia and Pseudophaeopappoides

Apart from the fact that these sections belong to the eastern clade, the present analysis does not reveal any relationships among these three sections. In the case of section Pseudophaeopappoides, this is unfortunate, because the only species described in this section is the puzzling C. antitauri Hayek, the affinities and relationships of which are unclear. Sections Chartolepis (Centaurea glastifolia L. and C. pterocaula Trautv.) and Grossheimia (C. macrocephala Puschk.) have no close relatives according to the present analyses. Centaurea glastifolia and C. macrocephala have given rise to a fertile hybrid in Armenia, C. caroli-henrici Gabrieljan & Dittrich (Dittrich and Gabrieljan, 1993).

Position of certain species in the eastern clade

One of the goals of this study was to clarify the doubtful position of C. chrysantha within section Acrocentron and the relationships of four species of difficult sectional assignment (C. antitauri, C. cankiriensis, C. ensiformis and C. isaurica). All of them were placed in the eastern clade, as expected, and their affinities seem clear except in the case of the isolated C. antitauri.

Centaurea chrysantha

The present results confirm the conclusion based on chromosome studies, indicating a basic chromosome number of 2n = 18, that this species is related to species of the eastern sections of the Jacea group (Romashchenko et al., 2004). Centaurea chrysantha clearly falls into the Jacea group, and specifically into the clade of eastern species. Within this clade, it is sister to two species usually classified in two different sections (C. ptosimopappoides from section Ptosimopappus and C. cheirolopha from section Pseudoseridia), with good support (BS = 88 %, PP = 1·00; Fig. 1). As has been seen, C. ptosimopappoides is unrelated to section Ptosimopappus, and section Pseudoseridia cannot be mantained. Thus, C. cheirolopha, C. chrysantha and C. ptosimopappoides, upon close morphological examination, could together constitute a new section.

Centaurea cankiriensis

This species was described indicating that it should be placed in section Cheirolepis. The present results confirm that it forms part of the Cheirolepis-Pseudoseridia complex (Fig. 1) together with C. drabifolia and C. kotschyi, with which it shares bract appendix morphology (T. Uysal, pers. comm.).

Centaurea ensiformis

This is a Turkish endemic that occurs on serpentine soils, and its morphological characters do not allow clear assignment to any of the existing sections (Davis, 1956; Wagenitz, 1975). The present results indicate a relationship with the Cheirolepis-Pseudoseridia complex (Fig. 1). A recent morphological study (T. Uysal, pers. comm.) indicates relationships with C. cankiriensis, which is compatible with the present results.

Centaurea isaurica

This species was placed by the present analysis in the Cheirolepis-Pseudoseridia complex (Fig. 1), as suggested by Wagenitz (1975). On morphological grounds, its relationships with C. drabifolia and C. kotschyi are obvious, as noted by Ertuğrul et al. (2004).

CONCLUDING REMARKS

The present study confirms, with a much more comprehensive sample, the existence of three major clades (circum-Mediterranean/Eurosiberian, western Mediterranean and eastern Mediterranean) in the Jacea group of Centaurea, and for the first time indicates that the first of these is sister to the other two. In the eastern Mediterranean clade, it is now clear that section Calcitrapa (with most of the species in the Middle East) is sister to all other sections in this group.

Besides confirming the three natural geographic groups, the present results indicate the need for profound changes to the sectional delimitation of this genus in order to reconcile classification and molecular systematics. Sections Cheirolepis, Microlophus and Ptosimopappus must be redefined. Section Pseudoseridia must disappear and the species formerly classified in this section must be reclassified in other sections. The molecular phylogeny also indicates a need to merge many sections together: Phalolepis and Pseudophalolepis into Acrolophus; Plumosipappus into Cheirolepis; Paraphysis into Cynaroides; Lepteranthus and Maculosae (partly) into Jacea; and Seridioides and Tetramorphaea into Calcitrapa. A morphological reinvestigation of C. cheirolopha, C. chrysantha and C. ptosimopappoides is needed to verify whether they should constitute a new section.

Some questions remain unresolved. First, phylogenetic trees based on nuclear DNA do not support certain groups, the relationships of which seem quite clear on morphological grounds. Sections Cynaroides, Microlophus and Phaeopappus are three good examples of this. Second, phylogenetic trees based on nuclear DNA place certain species in positions different from those expected on the basis of morphology, notably C. alba in the widely distributed clade and C. saligna in the eastern clade.

In the Jacea group it has been demonstrated that the difficulties for phylogenetic reconstruction in the sections Phalolepis and Willkommia using ribosomal DNA sequences arise from periods of gene flow followed by concerted evolution (Suárez-Santiago, 2005). The apparent contradictions between morphology and DNA sequences may reflect periods of intense gene flow between species. Indeed, the addition of new molecular data does not improve phylogenetic definition in this group, as experienced in section Willkommia by Suárez-Santiago (2005) and in the Cheirolepis-Pseudoseridia complex by T. Uysal et al. (unpubl. res.). In conclusion, and as suggested on the basis of similar results for the related section Acrocentron (Font et al., 2002), hybridization may have played, and may still play, a major role in the evolution of this group.

Acknowledgments

Financial support from the Spanish Ministery of Education and Science (Projects BOS2001-3041-C02-02 and CGL2004-04563-C02-01/BOS) and the Generalitat de Catalunya (‘Ajuts a Grups de Recerca Consolidats’ GREB 2001SGR00125) is gratefully acknowledged. Kostyantyn Romashchenko benefited from a grant from the Spanish Agency for International Cooperation, Ministry of Foreign Affairs. Tuna Uysal's stay at the Botanic Institute of Barcelona was funded by Tukish Tubitak. Víctor N. Suárez-Santiago benefited from a grant from the Spanish Ministery of Education and Science (associated with Project PB98-1288). We thank Dr S. Badarau, Romania, and M. Dr Vural, Turkey, for providing material of C. napulifera and C. tchihatcheffii. The nucleotide sequencing was performed at the ‘Serveis Científico-Tècnics’ of the University of Barcelona.

LITERATURE CITED

  1. Boissier E. 1875. Flora orientalis, sive enumeratio plantarum in Oriente a Graecia et Aegypto ad Indiae fines hucusque observatarum. Vol. 3. Geneva & Basel: H. Georg.
  2. Bremer K. 1994. Asteraceae. Cladistics and classification. Portland, OR: Timber Press.
  3. Cassini H. 1829. Tableau synoptique des synanthéreés. Annales de Sciences Naturelles 17: 387–423. Paris. Cited from the reprint by King R, Janaske PC, Lellinger B. 1995. Cassini on Compositae 3: 445–481. Missouri: Missouri Botanical Garden. [Google Scholar]
  4. Cosson ME. 1850. Notes sur quelques plantes nouvelles, critiques ou rares du Midi de l'Espagne, 2. Paris.
  5. Cullings KW. 1992. Design and testing of a plant-specific PCR primer for ecological and evolutionary studies. Molecular Ecology 1: 233–240. [Google Scholar]
  6. Davis PH. 1956. Fourteen new species from Turkey. Notes from the Royal Botanic Garden Edinburgh 22: 65–84. [Google Scholar]
  7. Dittrich M. 1968. Karpologische Untersuchungen zur Systematik von Centaurea un verwandten Gattungen. Teil I. Botanischer Jahrbücher für Systematik, Pflanzengeschichte und Pflanzengeographie 88: 70–122, 123–162. [Google Scholar]
  8. Dittrich M, Gabrieljan E. 1993.Centaurea caroli-henrici Gabrieljan & Dittrich, a new species from Armenia (Compositae–Cardueae). Candollea 48: 245–251. [Google Scholar]
  9. Dostál J. 1969. Taxonomická studie o celedích Lycopodiaceae a Asteraceae a o rodech Lycopodium a Centaurea (Studia Taxonomica I). Acta Universitatis Palackianae Olomucensis, Facultas Rerum Naturalium, ser. 2, Biol. 31: 5–39. [Google Scholar]
  10. Dostál J. 1973. Preliminary notes on the subtribe Centaureinae. Acta Botanica Academiae Scientiarum Hungaricae 19: 73–79. [Google Scholar]
  11. Doyle JJ, Doyle JL. 1987. A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochemical Bulletin, Botanical Society of America 19: 11–15. [Google Scholar]
  12. Ertuğrul K, Uysal T, Garcia-Jacas N, Susanna A, Garnatje T. 2004. The systematic position of Centaurea ensiformis and Centaurea isaurica from Turkey and the evolution of some characters in Centaurea. Israel Journal of Plant Sciences 52: 257–263. [Google Scholar]
  13. Felsenstein J. 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39: 783–791. [DOI] [PubMed] [Google Scholar]
  14. Font M, Garnatje T, Garcia-Jacas N, Susanna A. 2002. Delineation and phylogeny of Centaurea sect. Acrocentron based on DNA sequences: a revindication of the genus Crocodylium and indirect evidence of introgression. Plant Systematics and Evolution 234: 15–26. [Google Scholar]
  15. Gabrieljan ET, Aghababian MV. 1990. On the forgotten genus Chrysopappus (Asteraceae) from Asia Minor. Botaničeskij Žurnal 75: 990–994 [in Russian]. [Google Scholar]
  16. Garcia-Jacas N, Susanna A, Ilarslan R. 1996. Aneuploidy in the Centaureinae (Compositae): is n = 7 the end of the series? Taxon 45: 39–42. [Google Scholar]
  17. Garcia-Jacas N, Susanna A, Mozaffarian V. 1998a. New chromosome counts in the subtribe Centaureinae (Asteraceae, Cardueae) from West Asia. III. Botanical Journal of the Linnean Society 128: 413–422. [Google Scholar]
  18. Garcia-Jacas N, Susanna A, Vilatersana R, Guara M. 1998b. New chromosome counts in the subtribe Centaureinae (Asteraceae, Cardueae) from West Asia. II. Botanical Journal of the Linnean Society 128: 403–412. [Google Scholar]
  19. Garcia-Jacas N, Susanna A, Mozaffarian R, Ilarslan R. 2000. The natural delimitation of Centaurea (Asteraceae: Cardueae): ITS sequence analysis of the Centaurea jacea group. Plant Systematics and Evolution 223: 185–199. [Google Scholar]
  20. Garcia-Jacas N, Susanna A, Garnatje T, Vilatersana R. 2001. Generic delimitation and phylogeny of the subtribe Centaureinae (Asteraceae): a combined nuclear and chloroplast DNA analysis. Annals of Botany 87: 503–515. [Google Scholar]
  21. Greuter W. 2003. The Euro+Med treatment of Cardueae (Compositae)—generic concepts and required new names. Willdenowia 33: 49–61. [Google Scholar]
  22. Greuter W, Wagenitz G, Aghababian M, Hellwig FH. 2001. Proposal to conserve the name Centaurea (Compositae) with a conserved type. Taxon 50: 1201–1205. [Google Scholar]
  23. Greuter W, Aghababian M, Wagenitz G. 2005. Vaillant on Compositae—systematic concepts and nomenclatural impact. Taxon 54: 149–174. [Google Scholar]
  24. Hayek A von. 1901. Die Centaurea Arten Oesterreich-Ungarns. Denkschriften der Kaiserlichen Akademie der Wissenschaften, Mathematisch-Naturwissenschaftliche Klasse 70: 585–773. [Google Scholar]
  25. Holub J. 1973. Some new nomenclatural combinations in Centaureinae (Asteraceae). Preslia 45: 142–146. [Google Scholar]
  26. Holub J. 1974. Some new nomenclatural combinations in Centaureinae (Asteraceae) II. Preslia 46: 225–229. [Google Scholar]
  27. Huelsenbeck JP, Ronquist FR. 2001. MrBAYES: Bayesian inference of phylogenetic trees. Bioinformatics 17: 754–755. [DOI] [PubMed] [Google Scholar]
  28. Klokov MV, Tzvelev NN, Czerepanov SK, Sosnovsky DI. 1963. Centaurea. In: Schischkin K, Bobrov E, eds. Flora USSR, Vol. 28. Moscow/Leningrad: Academy of Sciences of the USSR, 468–729.
  29. Löve A, Löve D. 1961. Some nomenclatural changes in the European flora. I. Botaniska Notiser 114: 33–47. [Google Scholar]
  30. Maddison DR. 1991. The discovery and importance of multiple islands of most-parsimonious trees. Systematic Zoology 40: 315–328. [Google Scholar]
  31. Martín J, Garcia-Jacas N. 2000. Pollen studies in subtribe Centaureinae (Asteraceae): the Jacea group analysed with electron microscopy. Botanical Journal of the Linnean Society 133: 473–484. [Google Scholar]
  32. Nickrent DL, Schuette KP, Starr EM. 1994. A molecular phylogeny of Arceuthobium (Viscaceae) based on nuclear ribosomal DNA internal transcribed spacer sequences. American Journal of Botany 81: 1149–1160. [Google Scholar]
  33. Nylander JA. 2002. MrModeltest v.1·1b. Program distributed by the author. Department of Systematic Zoology, Uppsala University, Uppsala. http://www.ebc.uu.se/systzoo/staff/nylander.html (9 December 2005).
  34. Romashchenko K, Ertuğrul K, Susanna A, Garcia-Jacas N, Uysal T, Arslan E. 2004. New chromosome counts in the Centaurea Jacea group (Asteraceae, Cardueae) and some related taxa. Botanical Journal of the Linnean Society 145: 345–352. [Google Scholar]
  35. Ronquist F, Huelsenbeck JP. 2003. MrBAYES 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19: 1572–1574. [DOI] [PubMed] [Google Scholar]
  36. Soltis DE, Soltis PS, Collier TG, Edgerton ML. 1991. The Heuchera group (Saxifragaceae): evidence for chloroplast transfer and paraphyly. American Journal of Botany 78: 1091–1112. [Google Scholar]
  37. Suárez-Santiago VN. 2005. Utilización de marcadores moleculares en la resolución de problemas sistemáticos y evolutivos: la sección Willkommia Blanca (Género Centaurea L., Compositae) y el género Muscari Mill. (Hyacinthaceae). PhD Thesis, University of Granada, Spain.
  38. Sun Y, Skinner DZ, Liang GH, Hulbert SH. 1994. Phylogenetic analysis of Sorghum and related taxa using internal transcribed spacers of nuclear ribosomal DNA. Theoretical and Applied Genetics 89: 26–32. [DOI] [PubMed] [Google Scholar]
  39. Susanna A, Garcia-Jacas N. 2006. Tribus Cardueae. In: Kubitzki K, ed. The families and genera of vascular plants. Compositae. New York, NY: Springer-Verlag (in press).
  40. Susanna A, Garcia-Jacas N, Soltis DE, Soltis PS. 1995. Phylogenetic relationships in tribe Cardueae (Asteraceae) based on ITS sequences. American Journal of Botany 82: 1056–1068. [Google Scholar]
  41. Susanna A, Garnatje T, Garcia-Jacas N. 1999. Molecular phylogeny of Cheirolophus (Asteraceae: Cardueae-Centaureinae) based on ITS sequences of nuclear ribosomal DNA. Plant Systematics and Evolution 214: 147–160. [Google Scholar]
  42. Susanna A, Garcia-Jacas N, Hidalgo O, Vilatersana R, Garnatje T. 2006. The Cardueae (Compositae) revisited: insights from a ITS, trnL-trnF and matK nuclear and chloroplast DNA analysis. Annals of the Missouri Botanical Garden 93: 153–174. [Google Scholar]
  43. Swofford DL. 1999. PAUP*. Phylogenetic analysis using parsimony (*and other methods), version 4. Sunderland, MA: Sinauer Associates.
  44. Swofford DL, Olsen GJ. 1990. Phylogeny reconstruction. In: Hillis DM, Moritz C, eds. Molecular systematics. Sunderland, MA: Sinauer Associates, 411–501.
  45. Valdés-Bermejo E. 1980. Números cromosomáticos de plantas occidentales 1–34. Anales del Jardin Botánico de Madrid 36: 373–384. [Google Scholar]
  46. Wagenitz G. 1955. Pollenmorphologie und Systematik in der Gattung Centaurea L. s. l. Flora 142: 213–279. [Google Scholar]
  47. Wagenitz G. 1963. Die Eingliederung der ‘Phaeopappus’-Arten in das System von Centaurea. Botanischer Jahrbücher für Systematik, Pflanzengeschichte und Pflanzengeographie 82: 137–215. [Google Scholar]
  48. Wagenitz G. 1975. Centaurea. In: Davis PH, ed. Flora of Turkey. Edinburgh: Edinburgh University Press, 465–585.
  49. Wagenitz G. 1989. Nahe Verwandtschaft zwischen Arten der Centaurea Sektionen Acrolophus and Phalolepis. Flora 182: 341–351. [Google Scholar]
  50. Wagenitz G, Hellwig FH. 1996. Evolution of characters and phylogeny of the Centaureinae. In: Hind DJN, Beentje HG, eds. Compositae: Systematics. Proceedings of the International Compositae Conference, Kew, 1994. Kew: Royal Botanic Gardens, 491–510.
  51. White TJ, Bruns T, Lee S, Taylor J. 1990. Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics. In: Innis M, Gelfand D, Sninsky J, White T, eds. PCR protocols: a guide to methods and applications. San Diego, CA: Academic Press, 315–322.
  52. Zharkikh A. 1994. Estimation of evolutionary distances between nucleotide sequences. Journal of Molecular Evolution 39: 315–329. [DOI] [PubMed] [Google Scholar]

Articles from Annals of Botany are provided here courtesy of Oxford University Press

RESOURCES