Amer. J. Bot. 75(5): 753-766. 1988. SPECIAL PAPER PHYLOGENETIC IMPLICATIONS OF CHLOROPLAST DNA RESTRICTION SITE VARIATION IN THE MUTISIEAE (ASTERACEAE)l ROBERT K. JANSEN2 AND JEFFREY D. PALMER3 -Department of Ecology and Evolutionary Biology, The University ofConnecticut, Box U-42, Storrs, Connecticut 06268; and 'Department of Biology, University of Michigan, Ann Arbor, Michigan 48109 ABSTRACT Phylogenetic relationships among 13 species in the tribe Mutisieae and a single species from each of three other tribes in the Asteraceae were assessed by chloroplast DNA restriction site mapping. Initially, 211 restriction site mutations were detected among 16 species using 10 restriction enzymes. Examination of 12 of these species using nine more enzymes revealed 179 additional restriction site mutations. Phylogenetic analyses of restriction site mutations were performed using both Dollo and Wagner parsimony, and the resulting monophyletic groups were statistically tested by the bootstrap method. The phylogenetic trees confirm an ancient evolutionary split in the Asteraceae that was previously suggested by the distribution of a chloroplast DNA inversion. The subtribe Bamadesiinae of the tribe Mutisieae is shown to be the ancestral group within the Asteraceae. The molecular phylogenies also confirm the paraphyly of the Mutisieae and provide statistical support for the monophyly ofthree ofits four currently recognized subtribes (Bamadesiinae, Mutisiinae, and Nassauviinae). The fourth subtribe, Gochnatiinae, is shown to be paraphyletic. Within the subtribes, several closely related generic pairs are identified. Chloroplast DNA sequence divergence among genera of the Asteraceae ranges between 0.7 and 5.4%, which is relatively low in comparison to other angiosperm groups. This suggests that the Asteraceae is either a relatively young family or that its chloroplast DNA has evolved at a slower rate than in other families. THEASTERACEAE is one ofthe largest and most position in the class Dicotyledonae. In addition successful flowering plant families, consisting to its large size, the Asteraceae has a cosmo­ of approximately 1,100 genera and 20,000 politan distribution and is highly diversified in species (Cronquist, 1981). Several specialized its habitat preferences and life forms. Fur­ morphological features (capitulum, highly re­ thermore, although there is some controversy duced and modified flowers, inferior ovary, concerning the age ofthe family (Turner, 1977; ovule basal and erect, connate anthers, etc.) Boulter et al., 1978), fossil evidence (Cron­ strongly support the naturalness ofthe family. quist, 1977; Muller, 1981) and biogeographical Recent angiosperm classification systems considerations (Ravenand Axelrod, 1974) sug­ (Thome, 1976; Dahlgren, 1980; Takhtajan, gest that the Asteraceae originated in the mid­ 1980; Cronquist, 1981) emphasize the phy­ dle to upper Oligocene (30 million years ago) logenetic isolation of the family by placing it and subsequently underwent rapid and exten­ in a monotypic order at the most advanced sive morphological, chemical, and biological diversification. This explosive radiation has posed special problems for elucidating phy­ I Received for publication 27 August 1987; revision ac­ cepted 5 January 1988. logenetic relationships at higher taxonomic We thank the following individuals for providing seeds levels. or live plant material: James M. Affolter, Donald G. Hut­ During the past 30 years, six markedly dif­ tleston, Charles Jeffrey, David J. Keil, Guy L. Nesom, H. ferent schemes of phylogenetic relationships James Price, Roger W. Sanders. We also thank H. Tucker for technical assistance, K. Holsinger and S. Pacala for among the subfamilies and tribes ofAsteraceae assistance with the phylogenetic analyses, K. Bremer for have been proposed (Cronquist, 1955, 1977; providing an unpublished cladistic analysis of the Aster­ Carlquist, 1976; Wagenitz, 1976; Jeffrey, 1978; aceae, K. Holsinger for critically reading the manuscript, Thome, 1983; Bremer, 1987). Although there and M. Hommel and the Matthaei Botanical Garden for is general agreement expert care and growth of plants. This study was supported that two distinct subfam­ by a grant from the National Science Foundation (BSR­ ilies (Asteroideae and Cichorioideae) should 8415934). be recognized, there is no consensus concerning 753 754 AMERICAN JOURNAL OF BOTANY [Vol. 75 TABLE 1. Sources ofchloroplast DNA from species ofAsteraceae and outgroup families Species Source- Voucher information" Geographic origin Asteraceae Mutisieae Barnadesiinae 1. Barnadesia caryophylla MBG J907(MICH) Peru 2. Chuquiragua oppositifolia UC 83-303 Chile 3. Dasyphyllum diacanthoides UC 62-28851 unknown Gochnatiinae 4. Ainsliaea dissecta KEW 110-80-00757 Japan 5. Gochnatia paucifolia FrG 64-276 Peru 6. Stifftia chrysantha KEW 386-39-3860I Brazil 7. Onoseris hyssopifolia LG J921(MICH) Ecuador Mutisiinae 8. Gerberajamesonii MBG J915(MICH) South Africa 9. Leibnitzia seemannii Nesom N5084(TEX) Mexico 10. Mutisia acuminata UC 64-1510 Chile Nassauviinae II. Acourtia microcephala Keil KI8945(OBI) California 12. Perezia multiflora KEW 280-64-28004 Brazil 13. Trixis californicum Keil KI8528(OBI) California Cichorieae 14. Lactuca sativa GS none unknown Heliantheae 15. Helianthus annuus Price PHa89(TAES) unknown Cardueae 16. Carthamnus tinctorius USDA PIl98990 Israel Dipsacaceae 17. Cephalaria leucantha MBG J930(MICH) Spain 18. Dipsacus sativus MBG J931(MICH) Asia Rubiaceae 19. Pentas lanceolata MBG J914(MICH) Africa 20. Psychotria bacteriophila MBG 66517 Comoros Island a FrG = Fairchild Tropical Garden, GS = Grocery Store, Keil = Dr. David J. Keil, KEW = Kew Botanical Garden, LG = Longwood Garden, MBG = Matthaei Botanical Garden, Nesom = Dr. Guy L. Nesom, Price = Dr. H. James Price, UC = University of California Botanical Garden, USDA = United States Department of Agriculture. b Numbers preceded by a letter indicate collector (J = Jansen, K = Keil, P = Price) and collection number. All other numbers are accession numbers for live plants at botanical gardens. Standard herbarium acronyms (Stafleu, 1981) that follow collection numbers indicate the location of voucher specimens. the circumscription of the subfamilies, the parative systematic investigations. Previous number of monophyletic tribes, and the rela­ studies have clearly demonstratedthe potential tionships among the 12 to 17 recognized tribes. ofcpDNA for resolving evolutionary relation­ These previous intrafamilial classifications re­ ships among species in such genera as Brassica lied on more traditional data, including char­ (Palmer et al., 1983a), Clarkia (Sytsma and acters derived from anatomical, chromosomal, Gottlieb, 1986a, b), Cucumis (Perl-Treves and embryological, morphological, palynological, Galun, 1985), Linum (Coates and Cullis, 1987), phytochemical, and ultrastructural studies. Lycopersicon (Palmer and Zamir, 1982), Lis­ Furthermore, only the most recent reassess­ ianthius (Sytsma and Schaal, 1985), Nicotiana ment of relationships (Bremer, 1987) has (Kung, Zhu, and Chen, 1982; Salts et al., 1984), applied explicit cladistic methodology in deve­ Pisum (Palmer, Jorgensen, and Thompson, loping hypotheses of evolutionary relation­ 1985), Solanum (Hosaka et al., 1984; Hosaka, ships. 1986), and Triticum (Bowman, Bonnard, and We are performingcladistic analyses ofchlo­ Dyer, 1983; Tsunewaki and Ogihara, 1983). roplast DNA (cpDNA) mutations in order to Our studies are the first to use cpDNA to assess clarify phylogenetic relationships among the phylogenetic relationships at higher taxonomic subfamilies and tribes of the Asteraceae. The levels. conservative organization and evolution of Our previous studies (Jansen and Palmer, cpDNA among angiosperms (Palmer, 1985a, 1987a, b) revealed that there are two chloro­ b) makes this molecule well suited for com- plast genome arrangements in the Asteraceae. May 1988] JANSEN AND PALMER-PHYLOGENY IN MUTISIEAE (ASTERACEAE) 755 Chloroplast DNAs from species in the subtribe Barnadesiinae (Mutisieae) are colinear with cpDNAs of almost all land plants, including 10 families putatively related to the Astera­ ceae. All other Asteraceae, including 57 genera from all currently recognized tribes (sensu Jef­ frey, 1978), share a derived 22 kilobase (kb) inversion. Two alternative explanations were proposed for the phylogenetic distribution of the inversion. The most parsimonious inter­ pretation is that the Barnadesiinae primitively lacks the inversion and that this derived mu­ tation groups all otherAsteraceae together. This single character would then position the Bar­ nadesiinae at an ancestral positionin the family and indicate that the tribe Mutisieae is para­ phyletic. Alternatively, the inversion could have occurred in the common ancestor of the Asteraceae and subsequently been reversed in the Barnadesiinae. In this paper, we perform rigorous cladistic Fig. 1. Restriction site map oflettuce cpDNA showing analyses of a large number of restriction site the location ofthe 22 cloned restriction fragments used as mutations to decide between these alternative hybridization probes. All restriction sites are for Sac! ex­ phylogenetic explanations for the distribution cept where indicated. The two, heavy black lines indicate ofthe 22 kb cpDNA inversion. Our results also the extent ofthe inverted repeat. Fragment sizes and scale provide insights into evolutionary relation­ are given in kilobases (kb). The location of the 37 length mutations (Table I) is indicated by open triangles (dele­ ships