Proc. NatI. Acad. Sci. USAVol. 84, pp. 5818-5822, August 1987Evolution

A chloroplast DNA inversion marks an ancient evolutionary split inthe sunflower family (Asteraceae)

(angiosperm evolution/molecular systematics/Mutisieae/Barnadesiinae)

ROBERT K. JANSEN*t AND JEFFREY D. PALMERDepartment of Biology, University of Michigan, Ann Arbor, MI 48109

Communicated by Peter H. Raven, May 7, 1987 (received for review February 10, 1987)

ABSTRACT We determined the distribution of a chloro-plast DNA inversion among 80 species representing 16 tribes ofthe Asteraceae and 10 putatively related families. Filter hy-bridizations using cloned chloroplast DNA restriction frag-ments of lettuce and petunia revealed that this 22-kilobase-pairinversion is shared by 57 genera, representing all tribes of theAsteraceae, but is absent from the subtribe Barnadesiinae ofthe tribe Mutisieae, as well as from all families allied to theAsteraceae. The inversion thus defmes an ancient evolutionarysplit within the family and suggests that the Barnadesiinaerepresents the most primitive lineage in the Asteraceae. Theseresults also indicate that the tribe Mutisieae is not mono-phyletic, since any common ancestor to its four subtribes is alsoshared by other tribes in the family. This is the most extensivesurvey of the systematic distribution of an organelle DNArearrangement and demonstrates the potential of such muta-tions for resolving phylogenetic relationships at higher taxo-nomic levels.

The Asteraceae is one of the largest and economically mostimportant families of flowering plants and consists of 12-17tribes, approximately 1100 genera, and 20,000 species (1). Acombination of several specialized morphological character-istics (e.g., capitula, highly reduced and modified flowers,inferior ovaries, syngenesious anthers) strongly supports thenaturalness of the family. Cronquist (1) emphasized thedistinctness of the Asteraceae by placing it in a monotypicorder at the most advanced position within the subclassAsteridae. In addition to its large size, the family has acosmopolitan distribution and is highly diversified in itshabitat preferences and life forms. This diversity includesaquatics, herbs and shrubby trees in temperate, tropical, andarid environments, and trees in tropical rain forests. Speciesof Asteraceae are ofwide economic importance as vegetables(lettuce, artichokes, endive), sources of oil (sunflower, saf-flower) and insecticides (pyrethrum), and garden ornamen-tals (chrysanthemum, dahlia, marigold, and many others).Although there is some controversy concerning its age (2,

3), fossil evidence (4, 5) and biogeographical considerations(6) suggest that the Asteraceae originated in the middle toupper Oligocene (30 million years ago) and subsequentlyunderwent rapid radiation. This rapid diversification hasposed special problems for understanding phylogenetic rela-tionships at higher taxonomic levels. Previous attempts (4,7-10) at constructing phylogenies have relied on comparativeanatomical, chromosomal, embryological, micromolecular,morphological, and palynological features. These studieshave been largely unsatisfactory because of the repeatedparallel and convergent evolution of these characters. Forexample, three major and highly divergent reformulations of

Cronquist's (1, 4, 7) subfamilial classification for the Aster-aceae have been proposed in the last 12 years (8-10).We are investigating chloroplast DNA (cpDNA) variation

in the Asteraceae to resolve phylogenetic relationships athigher taxonomic levels. Our previous study (11) showed thatthe 151-kilobase (kb) cpDNAs of two species in the family(Lactuca sativa and Barnadesia caryophylla) are colinearthroughout the genome, with the exception of a single 22-kbinversion. The conservative organization of the chloroplastgenome among land plants (12, 13) makes such rearrange-ments potentially valuable characters for phylogenetic stud-ies. Here we report on the evolutionary direction of theinversion in the Asteraceae by comparing the chloroplastgenomes of Lactuca and Barnadesia with that of an out-group, Petunia hybrida (Solanaceae). We also examine thedistribution and phylogenetic significance of this rearrange-ment.

MATERIALS AND METHODScpDNAs were isolated by the sucrose gradient technique(14). Where tissue amounts were limited, total DNA wasisolated (15) and further purified by centrifugation in CsCl/ethidium bromide gradients. Restriction endonuclease diges-tions, electrophoresis, transfer of DNA fragments fromagarose gels to Zetabind filters (AMF Cuono), and hybrid-izations were performed as described (11, 14). Recombinantplasmids containing cpDNA fragments from Lactuca andPetunia were described previously (11, 16).

RESULTS

Filter hybridizations using cloned restriction fragments (16)from petunia (Petunia hybrida, Solanaceae) were performedto assess cpDNA genome arrangement in the Asteraceae.The petunia genome appears to have the ancestral cpDNAarrangement for angiosperms, since it is colinear with thegenomes of a fern, a gymnosperm, and several diverseangiosperms (17-21). Barnadesia cpDNA is colinear with thepetunia genome (Fig. 1) and therefore has the same geneorder as the ancestral angiosperm type. In contrast, lettuce(Lactuca sativa) cpDNA has a derived inversion in the largesingle copy region, as evidenced by the hybridization ofnonadjacent petunia Pst I fragments of 9.0 and 15.3 kb to thesame two regions of the lettuce genome. For example, bothof these petunia probes hybridize to 7.5-kb Sac I-Sal I and6.7-kb Sac I lettuce restriction fragments (Fig. 1). Further-more, the atpA through rpoB genes have an inverted orderand are transcribed in the opposite direction in lettucerelative to Barnadesia (Fig. 1; ref. 11).

Abbreviation: cpDNA, chloroplast DNA.*Present address: Department of Ecology and Evolutionary Biology,University of Connecticut, Storrs, CT 06268.

tTo whom reprint requests should be addressed.

5818

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

May

19,

202

0

Proc. Nati. Acad. Sci. USA 84 (1987) 5819

E E Bg S E E

t " r9

.G '4 'd' Z, CI IQaSmm * rn~~~~~ ~~~~~~~~~~~ ~~~~ ~~~~~~~~~~~~~~~~com~m u

mmTT m mu tM T1 Iftt NwIf I Lettuce

-Dt1 X :--.-L, +lk+ Petunia7.6 '4.6 19 2 4. 7614 9. 9. 15.3 13.1 8.0 1411 8.9 11.4

.IA I'-m m

Q00)IMc "p 4'4IR

I IN a IN on ------

*9*4-ZCU;vUe QUrn Q

U~~~4 ~ g-

Barnadesia

Petunia

FIG. 1. Physical maps showing the arrangement of homologous sequences in the petunia and either lettuce or Barnadesia chloroplastgenomes. Numbers indicate fragment sizes in kb. Each of 15 petunia fragments was hybridized to filter blots containing Nsi I and Sac I fragmentsof cpDNA from lettuce and Barnadesia. The lettuce or Barnadesia fragments to which the probes hybridize are indicated by lines leading fromthe petunia fragments to the lettuce or Barnadesia fragments. The heavy black lines on each map indicate the inverted repeat and the arrows

at far right show the orientation (i.e., direction of transcription) of mapped genes. The enlargements of the 7.5-kb Sac I-Sal I and 6.7-kb SacI restriction fragments show the four inversion endpoint fragments used as probes. Arrows pointing at the EcoRl sites indicate the approximatelocations of the inversion endpoints. Lettuce and Barnadesia restriction site and gene mapping data are from ref. 11 and petunia data are fromref. 16. Restriction sites shown: m, Nsi I; A, Pst I; *, Sac I; *, Sal I; Bg, Bgl 1I; E, EcoRI; S, Sal 1.

Many additional taxa were surveyed for the inversion byperforming filter hybridizations using cloned lettuce cpDNAfragments that contain the inversion endpoints. The 7.5-kbSac I-Sal I and 6.7-kb Sac I lettuce fragments were used as

hybridization probes against filter blots containing 12 restric-tion enzyme digests of DNA from one species of each of 80genera representing 10 putatively allied families and 16 tribesof Asteraceae (Table 1). The 7.5-kb Sac I-Sal I and 6.7-kbSac I probes will hybridize to different restriction fragments

5.8 _ t 14.9 tBarnadesia

9Lettuce

14.7 Vernonia

7.5 6.7 JLettuceFIG. 2. Physical maps showing the arrangement of homologous

sequences in the 22-kb inversion region of the lettuce and eitherBarnadesia or Vernonia chloroplast genomes. The Barnadesia andVernonia Sac I fragments to which the 7.5-kb Sac I-Sal I and 6.7-kbSac I probes hybridize are indicated by lines leading from the lettucefragments to the Barnadesia and Vernonia fragments. Numbers indi-cate fragment sizes in kb. Restriction sites shown: e, Sac I; *, Sal I.

in those genomes that contain the lettuce inversion. Thissituation is illustrated in Figs. 2 and 3, in which the twoinversion endpoint fragments from lettuce are hybridizing toSac I fragments of 14.7 and 17.0 kb in Vernonia. Similarhybridization results are evident for Helianthus and Trixis(Fig. 3), which are both members of the Asteraceae. Incontrast, in those genomes that are not rearranged, the twolettuce probes will hybridize to two of the same restrictionfragments. For example, the 7.5-kb Sac I-Sal I and 6.7-kbSac I lettuce probes both hybridize to Sac I fragments of 5.8and 14.9 kb in Barnadesia (Figs. 2 and 3). The autoradi-ograms (Fig. 3) reveal that representatives of three relatedfamilies, Cephalaria (Dipsacaceae), Pentas (Rubiaceae), andScaevola (Goodeniaceae), also lack the 22-kb inversion.The results of the inversion survey for all 80 examined taxa

are summarized in Table 1. The genome arrangements for 69of these taxa have been confirmed by constructing completerestriction maps (R.K.J., H. Michaels, and J.D.P., unpub-lished data). The inversion is absent from all putatively alliedfamilies and, within the Asteraceae, from the subtribeBarnadesiinae of the tribe Mutisieae. All other examinedmembers of the Asteraceae, including the three other sub-tribes in the Mutisieae, were found to have the inversion. The80 genera surveyed represent the major evolutionary lineageswithin the 16 tribes of Asteraceae and 10 related families. Weare confident that the selection of only one species from eachgenus is an adequate sampling because more extensivestudies of 60 species in Carthamus (R. Johnson and J.D.P.,unpublished data), Coreopsis (D. Crawford and J.D.P.,unpublished data), Hieracium (R.K.J. and J.D.P., unpub-lished data), and Lactuca (E. Jandourek and J.D.P., unpub-lished data) have revealed no intrageneric variation in chlo-roplast genome arrangement in the Asteraceae.

rn1

Evolution: Jansen and Palmer

-

Dow

nloa

ded

by g

uest

on

May

19,

202

0

5820 Evolution: Jansen and Palmer

Table 1. Distribution of a 22-kb cpDNA inversion

Inversion absent

RosidaeApiaceae

Angelica archangelicaHydrocotyle verticillata

AraliaceaeHedera helix

AsteridaeBrunoniaceae

Brunonia australisCampanulaceae

Campanula ramosaJasione perennisLobelia ramosa

CaprifoliaceaeViburnum acerifolium

DipsacaceaeCephalaria leucanthaDipsacus sativusScabiosa atropurpurea

GoodeniaceaeDampiera strictaGoodenia hederaGoodenia lanataScaevola frutescens

RubiaceaePentas lanceolataPsychotria bacteriophila

StylidiaceaeStylidium adnatum

ValerianaceaeValeriana officinalis

Asteraceae*CichorioideaeMutisieaeBarnadegiinaeBarnadesia caryophyllaChuquiragua oppositifoliaDasyphyllum diacanthoides

Mutisieae (continued)GochnatiinaeAinsliaea dissectaGochnatia paucifoliaOnoseris hyssopifoliaStifftia chrysantha

MutisiinaeChaptalia tomentosaGerbera jamesondiLeibnitzia seemanniiMutisia acuminataPiloselloides hirsutaNassauviinaeAcourtia microcephalaPerezia multifloraTrixis californicum

ArctotideaeArctotis stoechadifoliaGazania splendensHaplocarpha scaposa

CardueaeCentaurea montanaCirsium sp.Echinops exaltatusSilybum marianum

CichorieaeHieracium pratenseLactuca sativaTragopogon porrifolius

EupatorieaeChromolaena sp.Eupatorium atrorubensLiatris spicata

LiabeaeCocosmia rugosaLiabum glabrum

VernonieaeLycnophora tomentosaPiptocarpha axillarisStokesia laevisVernonia mespilifolia

Inversion present

AsteriodeaeAnthemideae

Achillea millefoliumChrysanthemum maximumSantolina chamaecyparissus

AstereaeAster cordifoliusBellis perennisErigeron hybridusFelicia bergerianaSolidago sp.

CalenduleaeCalendula officinalisDimorphotheca pluvialisOsteospermum muricatum

CotuleaeCotula barbata

HeliantheaeCoreopsis grandifloraDahlia pinnataGeraea canescensHelianthus annuusPerityle emoryiiWedelia trilobata

InuleaeAntennaria neodioicaGnaphalium luteoalbumInula helenium

SenecioneaeBlennosperma nanaEuryops pectinatusSenecio mikanioides

TageteaeDyssodia pentachaetaTagetes erecta

UrsinieaeUrsinia nana

Voucher specimens deposited at MICH (Ann Arbor, Ml).*The list of subfamilies and tribes follows Jeffrey (10).tThe list of subtribes follows Cabrera (22).

We also performed filter hybridizations to a single enzymedigest of DNA from the 80 species, using four restrictionfragments (whose sizes and locations are shown in theenlargement at the top of Fig. 1) subcloned from the 6.7-kbSac I and 7.5-kb Sac I-Sal I fragments. We previouslyshowed (11) that the lettuce inversion endpoints are locatedvery close to the EcoRI sites separating these two pairs ofadjacent fragments (Fig. 1, arrows). These smaller, moreprecise probes hybridize to those genomes containing theinversion in exactly the same manner as to the parentallettuce genome (data not shown). This gives us greaterconfidence that these taxa have the same inversion as lettuce,rather than a similar but different inversion in the same regionof the chloroplast genome.

DISCUSSIONThe 22-kb inversion must be derived within the Asteraceaesince all putative outgroup families lack this rearrangement.Furthermore, more inclusive outgroups, including 30 addi-tional families of angiosperms, a gymnosperm, and a fern,also lack the inversion (12, 13, 17-20). There are two

alternative explanations for the phylogenetic distribution ofthe inversion (Fig. 4). The most parsimonious interpretationis that the three genera in the Barnadesiinae primitively lackthe inversion and that this derived mutation groups all otherAsteraceae together. Alternatively, the inversion occurred inthe common ancestor of the entire family and subsequentlyreverted in the Barnadesiinae. The former explanation seemsmore likely, both on a parsimony basis (23) and, morecompellingly, because cpDNA inversions are rare amongland plants (12, 13). This particular inversion appears to haveoccurred only once in some 400 million years of land plantevolution. Furthermore, independent cladistic studies usingdata from restriction site mapping (unpublished) and mor-phology (24) support the phylogeny shown in Fig. 4 and placethe Barnadesiinae as an ancestral lineage within the Aster-aceae.The distribution of the cpDNA inversion within the Aster-

aceae (Table 1, Fig. 4) defines the primary evolutionary splitwithin this large and important family offlowering plants, andthus it has significant phylogenetic implications. Indeed, oneof the most controversial systematic issues within the familyhas been the identification of the most primitive lineage. A

Proc. Natl. Acad. Sci. USA 84 (1987)

Dow

nloa

ded

by g

uest

on

May

19,

202

0

Proc. Natl. Acad. Sci. USA 84 (1987) 5821

Cep Pen Sca6.7 7.5 6.7 7.5 6.7 7 5

Bar Tri6.7 7.5 6.7 7.5

Lac Ver6.7 7.5 6.7 7.5

qw_ O_

4,da

4,

3-

1-

FIG. 3. Hybridization of cloned (11) lettuce restriction fragments to Sac I digests of DNA from eight representative species from theAsteraceae and related families. Bar, Barnadesia; Cep, Cephalaria; Hel, Helianthus; Lac, Lactuca; Pen, Pentas; Sca, Scuevola; Tri, Trixis;and Ver, Vernonia. Numbers above the filters refer to hybridization probes. Numbers alongside the filters indicate fragment sizes in kb.

number of putatively primitive morphological and anatomicalcharacters have been used to hypothesize an ancestralposition for the Heliantheae (2, 4, 7, 9, 25). However, fourother tribes, the Cardueae, Mutisieae, Senecioneae, andVernonieae, have also been suggested as being most primi-tive (7-9, 26). The primary reasons for this lack of agreementare the uncertainty about which family or families constitutethe best outgroup and the high incidence of parallel andconvergent evolution in the characters that have been used.Identification of primitive character states has thus beendifficult, whereas the cpDNA inversion is unambiguouslyrooted and appears free of parallelism and convergence. Thedata presented here, together with the two recent morpho-logical and restriction site studies cited above, clearly indi-cate that the Barnadesiinae is the primitive group within theAsteraceae. This conclusion agrees with recent suggestions

OsmundaGinkgo

40 Families, 26 Orders,and 9 Subclasses Inversion

of Angiosperms Absent

SubtribeBarnadesiinaeof the Mutisieae

Asteraceae: 16 Tribes,Inversion

- 57 Genera, 3 other Present

Subtribes of the Mutisieae

FIG. 4. Evolutionary tree based on the 22-kb chloroplast DNAinversion. Data are summarized from Table 1 and refs. 12, 13, and 17.

(8, 9, 26) that the Mutisieae contains the most primitive taxain the family.Our identification of the Barnadesiinae as the most prim-

itive lineage in the Asteraceae provides support for sugges-tions that the Asteraceae originated in montane South Amer-ica (2, 6, 27, 28), as the eight genera of this subtribe arecentered in the northern Andes (22). Our results are consist-ent with suggestions (8, 26, 29) that bilabiate (two-lipped)flowers and woody habit, which are common features in theBarnadesiinae, are primitive within the Asteraceae. Thissuggests affinities between the Asteraceae and the familieswith some bilabiate or woody members, including theCampanulaceae, Lobeliaceae, Goodeniaceae, and Stylidi-aceae.The distribution of the cpDNA inversion also provides

insights into phylogenetic relationships within the Mutisieae.Our data confirm previous suggestions (22, 30, 31) that theMutisieae is not a monophyletic group (i.e., one derived froma common ancestor not shared by any other tribes in theAsteraceae), since three of its four subtribes are more closelyrelated cladistically to 15 other tribes than they are to theBarnadesiinae (Table 1, Fig. 4). The uniqueness of thesubtribe Barnadesiinae is evident in its lack of the 22-kbcpDNA inversion and its distinctive pollen (31-33) and floral(30) morphology.

This is the most extensive survey of the systematicdistribution of a structural mutation in an organelle genomeand clearly demonstrates the potential of rearrangements forresolving phylogenetic relationships at higher taxonomiclevels. Detailed studies ofcpDNA inversions (12, 13) in otherflowering plant families, including the economically impor-tant grasses and legumes, should be equally valuable inmaking major phylogenetic groupings among and withinthese families.

We thank the following individuals for providing seeds or live plantmaterial: James M. Affolter, James A. Armstrong, Randall J. Bayer,Roger C. Carolin, Nancy C. Coile, Daniel J. Crawford, Michael 0.

Dillon, Charles Jeffrey, Samuel B. Jones, David J. Keil, Timothy K.Lowrey, Guy L. Nesom, Tycho Norlindh, Robert Ornduff, JamesPrice, Roger W. Sanders, John L. Strother. We also thank K. Bremer

25-

15- -_ _

9-

6-

Hel6.7 7.5

4,

Evolution: Jansen and Palmer

Dow

nloa

ded

by g

uest

on

May

19,

202

0

5822 Evolution: Jansen and Palmer

for providing an unpublished cladistic analysis of the Asteraceae, M.Hommel and the Matthaei Botanical Garden for expert care andgrowth of plants, E. Clark and M. Hanson for petunia Sal I clones,and W. Brown, T. Bruns, M. Chase, D. Crawford, J. Manhart, H.Michaels, B. Milligan, C. Moritz, W. Wagner, Jr., and M. Zolan forcritical reading of the manuscript. This study was supported by GrantBSR-8415934 from the National Science Foundation.

1. Cronquist, A. (1981) An Integrated System of Classification ofFlowering Plants (Columbia Univ. Press, New York), pp.1020-1028.

2. Turner, B. L. (1977) in The Biology and Chemistry of theCompositae, eds. Heywood, V. H., Harborne, J. B. & Turner,B. L. (Academic, London), Vol. 1, pp. 21-39.

3. Boulter, D., Gleaves, J. T., Haslett, B. G., Peacock, D. &Jensen, U. (1978) Phytochemistry 17, 1585-1589.

4. Cronquist, A. (1977) Brittonia 29, 137-153.5. Muller, J. (1981) Bot. Rev. 47, 1-142.6. Raven, P. H. & Axelrod, D. 1. (1974) Ann. Mo. Bot. Gard. 61,

539-673.7. Cronquist, A. (1955) Am. Midi. Nat. 53, 478-511.8. Carlquist, S. (1976) Aliso 8, 465-492.9. Wagenitz, G. (1976) Plant Syst. Evol. 125, 29-46.

10. Jeffrey, C. (1978) in Flowering Plants of the World, ed.Heywood, V. H. (Mayflower, New York), pp. 263-268.

11. Jansen, R. K. & Palmer, J. D. (1987) Curr. Genet. 11, 553-564.12. Palmer, J. D. (1985) Annu. Rev. Genet. 19, 325-354.13. Palmer, J. D. (1985) in Molecular Evolutionary Genetics, ed.

MacIntyre, R. J. (Plenum, New York), pp. 131-240.14. Palmer, J. D. (1986) Methods Enzymol. 118, 167-186.15. Saghai-Maroof, M. A., Soliman, K. M., Jorgensen, R. A. &

Allard, R. W. (1984) Proc. Natl. Acad. Sci. USA 81, 8014-8018.

16. Palmer, J. D., Shields, C. R., Cohen, D. B. & Orton,, T. J.(1983) Theor. Appl. Genet. 65, 181-189.

17. Palmer, J. D. & Stein, D. (1986) Curr. Genet. 10, 823-833.18. Palmer, J. D. & Thompson, W. F. (1982) Cell 29, 537-550.19. Fluhr, R. & Edelman, M. (1981) Nucleic Acids Res. 9,

6841-6853.20. De Heij, H. T., Lustig, H., Moeskops, D. J. M., Bovenberg,

W. A., Bisanz, C. & Groot, G. S. (1983) Curr. Genet. 7, 1-6.21. Sytsma, K. J. & Gottlieb, L. D. (1986) Evolution 40, 1248-

1201.22. Cabrera, A. L. (1977) in The Biology and Chemistry of the

Compositae, eds. Heywood, V. H., Harborne, J. B. & Turner,B. L. (Academic, London), Vol. 2, pp. 1039-1066.

23. Crisci, J. V. (1982) J. Theor. Biol. 97, 35-41.24. Bremer, K. (1987) Cladistics, in press.25. Koch, M. F. (1930) Am. J. Bot. 17, 938-952.26. Jeffrey, C. (1977) in The Biology and Chemistry of the

Compositae, eds. Heywood, V. H., Harborne, J. B. & Turner,B. L. (Academic, London), Vol. 1, pp. 111-118.

27. Bentham, G. (1873) J. Linn. Soc. Lond. Bot. 13, 335-577.28. Small, J. (1919) New Phytol. 18, 1-29.29. Carlquist, S. (1966) Aliso 6, 25-44.30. Small, J. (1918) New Phytol. 17, 13-40.31. Wodehouse, R. P. (1928) Bull. Torrey Bot. Club 55, 449-462.32. Wodehouse, R. P. (1929) Am. J. Bot. 16, 297-313.33. Skvarla, J. J., Turner, B. L., Patel, V. C. & Tomb, A. S.

(1977) in The Biology and Chemistry of the Compositae, eds.Heywood, V. H., Harborne, J. B. & Turner, B. L. (Aca-demic, London), Vol. 1, pp. 141-248.

Proc. Natl. Acad. Sci. USA 84 (19.87)

Dow

nloa

ded

by g

uest

on

May

19,

202

0