Kobe University Repository : Kernel

タイトル Extensive Gene Duplications in Diploid () Title 著者 Yahara, Tetsukazu / Kawahara, Takayuki / Crawford, Daniel J. / Ito, Author(s) Motomi / Watanabe, Kuniaki 掲載誌・巻号・ページ American journal of botany,76(8):1247-1253 Citation 刊行日 1989-08 Issue date 資源タイプ Journal Article / 学術雑誌論文 Resource Type 版区分 publisher Resource Version 権利 Rights DOI JaLCDOI URL http://www.lib.kobe-u.ac.jp/handle_kernel/90001248

PDF issue: 2021-10-04 Amer. J. Bot. 76(8): 1247-1253. 1989.

EXTENSIVE GENE DUPLICATIONS IN DIPLOID EUPATORIUM (ASTERACEAE)1

TETSUKAZU YAHARA,2 TAKAYuKI KAWAHARA 2 DANIEL J. CRAWFORD,3 MOTOMI ITO,4 AND KUNIAKI WATANABE5 2BotanicalGardens, Faculty of Science, University of Tokyo, 1842 Hanaishi-cho, Nikko 321-14, Japan;3Department of Botany, The Ohio State University, Columbus,Ohio 43210-1293; 4MakinoHerbarium, Tokyo MetropolitanUniversity, 2-1-1 Fukazawa,Tokyo 158, Japan;and 5Biological Institute,Faculty of GeneralEducation, Kobe University, Kobe 657, Japan

ABSTRACT An electrophoreticstudy of isozyme number for seven soluble enzymes revealed extensive gene duplicationsin eight diploid species of American Eupatoriumbelonging to three mor- phologicalgroups. The enzymes isocitrate dehydrogenase,phosphoglucomutase, phosphoglu- cose isomerase, 6-phosphogluconatedehydrogenase, and shikimate dehydrogenaseoccur as three to six isozymes in all species, whereasthe minimal conserved numbertypical of diploid plantsis two isozymes for each. Fructose1, 6-biphosphatealdolase is expressedas multibanded patternsuggesting fixed heterozygosityin all examined species. It was not possible to document gene duplicationfor triosephosphateisomerase from the electrophoreticpatterns. All species examinedhave a chromosomenumber of 2n = 20, which has been regardedas the basic diploid number for Eupatorium.However, the detection of extensive duplicationssuggests that 2n = 10 may be the originaldiploid chromosomenumber in Eupatoriumand that with 2n = 20 are of polyploidorigin. This hypothesiswould mean that extensive duplicationsat isozyme geneloci have been maintainedsince the originof the genus,despite chromosomal diploidization having occurred.

IT HAS BEEN documented that diploid vascular cannot be employed as the sole criterion. Iso- plants have a minimal highly conserved num- zyme number has been successfullyapplied to ber of isozymes for most of the enzymes rou- systematic questions in several groups. tinely examined by electrophoresis(Gottlieb, Recent studies of ferns and fern-allies have 1981a, 1982, 1983, 1984). This means that an shown that plants traditionallyviewed as poly- increasein isozyme numberfor a given enzyme ploids on the basis of high chromosome num- is the result either of gene duplication at the bers are genetically diploids at isozyme loci diploid level or of polyploidy. If a plant is (Haufler and Soltis, 1986; Haufler, 1987; D. polyploid, then it should have duplications at Soltis and P. Soltis, 1987, 1988). Using the many isozyme loci, whereas a diploid plant same criterion, Soltis et al. (1987) and Chase would exhibit much less extensive increase in and Olmstead (1988) rejected the hypothesis isozymenumber(Gottlieb,1981a, 1983,1984). that species with high chromosome numbers While inferences about whether given plants arepolyploids in the Bromeliaceaeand the sub- are diploids or polyploids have often rested tribe Oncidiinae of Orchidaceae,respectively. primarily on their chromosome numbers, On the other hand, Witter (1988) supported Gottlieb (1981b) employed isozyme number the hypothesis that plants of the Hawaiian sil- to infer the ploidy level of plants when chro- verswordalliance, in which 2n = 28, are poly- mosome data per se were inconclusive. Gott- ploid by demonstratingduplicate gene expres- lieb (198 lb) arguedthat an essential featureof sion for four soluble enzymes. polyploid is the number of genomes being ex- In this paper we reportresults of an electro- pressed, and that chromosome number alone phoretic study of isozyme number for seven enzymesfrom eight speciesof Eupatoriumwith a chromosome number of 2n = 20. This has I Received for publication 1 November 1988; revision been regardedwidely as the originalbase num- accepted2 March 1989. ber for the tribe Eupatoriae (King and Rob- We thank V. I. Sullivan for her kind help in the field. This study was partly supportedby grant no. 63041048 inson, 1970a, 1987;King et al., 1976), although from Ministry of Education, Culture and Scientific Re- a number of n = 5 has been reported for the serch, Japan. tribe (Turnerand Irwin, 1960).

1247 1248 AMERICAN JOURNAL OF BOTANY [Vol. 76

MATERIALS AND METHODS -Eight species of well established for each (Randall and Givan, Eupatoriumwere examined;the species, num- 1981 for IDH; Gottlieb, 1982; Mousdale and ber of plants studied from naturalpopulations Coggins, 1985, for SKDH). In all instances, and localities of the populationsare as follows: one isozyme occursin the plastidsand the other E. capillifolium(Lam.) Small (25, Vernon Par- is cytosolic. ish, Louisiana; 24, Guilford Co., North Car- Chloroplasticisozymes were isolated from olina); E. fistulosum J. Barratt (24, Fairfield a single plant of E. perfoliatum collected in Co., Ohio; 24, Hocking Co., Ohio); E. mac- Columbus, Ohio. The method of extracting ulatum L. (12, Franklin Co., Ohio); E. mika- intact chloroplast followed Gastony and Dar- nioides Chapman (12, Wakulla Co., Florida); row (1983) except that the chloroplasts were E. perfoliatumL. (26, FairfieldCo., Ohio; 25, not centrifugedthrough Percoll. The chloro- Hocking Co., Ohio); E. purpureum L. (24, plast-enrichedpellets were broken with the tris- FairfieldCo., Ohio);E. semiserratumDC. (24, HCIgrinding buffer (above). Activities of IDH, Evans Co., Georgia; 18, Allen Parish, Loui- 6-PGD, and SKDH were stronger in flower siana) and E. serotinum Michx. (12, Franklin buds but very weak in leaves and we could not Co., Ohio; 27, Marion Co., Indiana). For E. obtain sufficient activity of these enzymes in fistulosumand E. perfoliatum,progenies grown the chloroplast fraction from leaves. from seeds were also examined (36 from three maternalparents of E.fistulosum;and 36 from RESULTS-PGI- In all eight species, enzyme two plants of E. perfoliatum; seeds were col- activity appearedin two zones in the gels. In lected from Fairfield Co., Ohio). the chloroplast-enrichedfractions of E. per- Flower buds and a small piece of leaf ma- foliatum, isozymesin the fastermigrating (more terial of individual plants were ground in 1.0 anodal) zone only were expressed. This result ml of cold extraction buffer consisting of 0.1 agrees with all the formerly obtained results M tris-HCl,pH 7.5,20 mm 2-mercaptoethanol, on electrophoreticmobilities of PGI isozymes 1.0 mM EDTA, 10 mm KCI, 10 mM MgCl2, from various plants;i.e., the chloroplasticiso- 0.25% (v/v) triton X-100, and 15 mg PVP zyme migrates faster than the cytosolic one. (modified from Odryzykoski and Gottlieb, In the slowermigrating (more cathodal) zone, 1984). For E. capillifolium and E. mikanioi- three or five bands were detected for this di- des, fresh leaves alone were used. meric enzyme in most individuals of the eight Enzymes were resolved in 12% starch gels species examined. This suggests gene dupli- using three buffer systems. System I had a gel cation for cytosolic PGI in diploid Eupato- bufferof 0.033 M tris, 0.005 M citric acid, 0.004 rium. Among progenies grown from a single M lithium hydroxideand 0.3 M boric acid with maternal parent of E. fistulosum, three phe- the pH adjustedto 7.6 and an electrode buffer notypes segregatedin the more cathodal zone of 0.039 M tris and 0.263 M boric acid with the (1-3 in Fig. IA and Fig. 2); one pattern with pH adjusted to 8.0 (P. Soltis and D. Soltis, a seemingly single heavily staining band (1 in 1987). System II had an electrodebuffer of 0.5 Fig. 1 and 2), one phenotypehaving three bands M tris, 0.016 M EDTA and 0.57 M boric acid consisting of a fast migratingvery faint, a slow with the pH adjustedto 8.0. The gel bufferwas migratingvery dark, and a band intermediate a 1:9 dilution of the electrode buffer. System in both mobility and staining intensity (2 in III consisted of an electrode of 0.065 M L- Fig. 1 and 2), and the last pattern having six histidine (free base) and 0.007 M citric acid bands (3 in Fig. 1 and 2). The first phenotype (monohydrate) adjusted to pH 6.5. The gel is interpretedas having three bands that mi- bufferwas a 1:3 dilution of the electrodebuffer grate very closely and the most anodally mi- (Cardy,Sturber, and Goodman, 1981). gratingband is very faint (see dots in Fig. IA System I resolved fructose 1, 6-biphosphate and Fig. 2-1). Also, among progenies grown aldolase (ALD, EC 4.2.1.3), phosphoglucoiso- from a single maternal parent of E. perfolia- merase (PGI, EC 5.3.1.9) and triosephosphate tum, three- and five-banded phenotypes seg- isomerase (TPI, EC 5.3.1.1). System II re- regated in the more cathodal zone (Fig. iB, solved phosphoglucomutase (PGM, EC Fig. 2-4, 2-5). These patternssuggest that fixed 2.7.5.1). System III was employed to resolve heterozygositydue to duplicated gene loci for isocitrate dehydrogenase(IDH, EC 1.1.1.42), cytosolic PGI isozyme occur in E. fistulosum 6-phosphogluconate dehydrogenase (6PGD, and E. perfoliatum. In both species, three- EC 1.1.1.44) and shikimate dehydrogenase banded phenotypes have one faint band, and (SKDH, EC 1.1.1.25). Staining schedules fol- the same pattern was seen in E. mikanioides lowed Soltis et al. (1983). Enzymes for study (Fig. I E), in which one faint band of the three- were selected because the minimal conserved banded phenotype migrated near to a band numberof isozymes in diploid plants has been representinga chloroplasticisozyme. By com- August 1989] YAHARA ET AL. -GENE DUPLICATIONS IN EUPATORIUM 1249

A B~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~A

F G I

.40

Fig. 1. Electrophoreticpatterns of six solubleisozymes in diploidspecies of Eupatorium.(A-E) PGI, A, E.fistulosum; B, E. perfoliatum;C, E. semniserratum,D, E. capillifolium,E, E. mikanioldes.(F-I) PGM; F, E. fistulosum; G, E. perfoliatum;H, E. semiserratum;I, E. capillifolium.(J-M) 6PGD; J, E. fistulosum;K, E. perfoliatum;L, E. semiser- ratum; M, E. capillifolium.(N) ALD in E. capillifolium.(O-Q) IDH; 0, E. perfoliatum;P, E. semiserratum;Q. E. capillifolium.(R-T) SKDH; R, S, E. fistulosum;S, E. perfoliatum;T, E. semiserratum.All isozymes migrateto the anode at the top of the photograph. 1250 AMERICAN JOURNAL OF BOTANY [Vol. 76

ed patterns were seen (Fig. 1H, I), indicating that at least four loci are being expressed. 6PGD-InE.fistulosum (Fig. IJ) andE. pur- pureum, complex multibanded patterns were resolved. In most individuals, more than five bands were seen, and these two species appear to have at least three loci expressed.In E. per- foliatum (Fig. 1K) and E. capillifolium (Fig. I M), electromorphsof 6PGD occurredin two electrophoretic zones. In E. perfoliatum, a three-bandedpattern was always observed in the more cathodal zone. Bands in the anodal 1 23 4 5 regionwere not well resolved but three-banded patternswere frequentlyobserved; we cannot Fig. 2. Diagrammaticscheme of representativepat- tems of PGI isozymes segregatedin progeniesof Eupa- say with certainty,however, that all plantshave toriumfistulosum(1-3) and E. perfoliatum(4, 5). Numbers three bands. These patterns indicate that at are correspondentto those in Fig. IA and lB. least threeloci encode 6PGD in E. perfoliatum. In E. capillifolium, two faint but clearly ob- servablebands were always present in the more paring with these patterns, seemingly four- cathodal area (indicated by dots in Fig. 1M), banded phenotypes in E. semiserratum(Fig. and three or more bands were observed in the 1C), and E. capillifolium(Fig. 1D) can be in- more anodal area. Two-bandedpatterns in the terpretedas having an additional band that is more cathodal area could be due to very faint too faint to be recognized. Eupatorium sero- staining for an additional band, but we could tinum expresseda pattern similar to E. semi- not document this with certainty.This species serratum while E. maculatum and E. purpu- is interpretedas having at least three loci for reum displayed easily interpretablethree- or 6PGD. In E. semiserratum(Fig. 1L) and E. five-bandedpatterns. Clearly, two gene loci are serotinum, five bands were always resolved, expressedfor the more cathodal,cytosolic iso- and faint bands were sometimes detected in zymes. In the more anodal zone, three-banded the more cathodalarea. These species are con- patternswere seen for most plants of E. semi- sidered to have at least three loci for 6PGD. serratum (Fig. IC) and E. capillifolium (Fig. In E. maculatum and E. mikanioides, 6PGD ID). However, these bands were not well re- was not resolved well. solved in many instances and it is not clear whether one or two gene loci are specifying ALD-A multibanded pattern was invari- chloroplasticisozymes of PGI. Thus, a total of ably expressed in all eight species examined. at least three genes encoded PGI in all eight In well-resolvedgels, five bands with two faint species. ones can be recognized;however, a three-band- ed pattern was seen on most gels (Fig. 1N). PGM-Six- to seven-banded patterns were Because ALD is a tetrameric enzyme, three- observedin 36 progeniesgrown from two ma- banded patterncould be due to low activity of ternal parentsof E. perfoliatum.Two of these the two additional bands. Those gels contain- bands, which were the most anodally migrat- ing five well-resolvedbands provide good evi- ing, werealways expressed in 36 progenies(Fig. dence for duplicated loci for ALD in all eight IG). The chloroplast-enrichedfraction from a species. singleplant of E. perfoliatumyielded these two bands. BecausePGM is a monomeric enzyme, IDH-For E. perfoliatum(Fig. 10), E. cap- the data suggest that E. perfoliatum has two illifolium (Fig. I Q), E. purpureum and E. gene loci encodingchloroplastic PGM. The re- serotinum,enzyme activity was found in two maining four to five bands are viewed as cy- distinct zones of the gels. In these species, one- tosolic, and there are at least three gene loci or three-bandedpatterns were seen in the more encoding these isozymes. A total of 26 prog- anodal zone, suggestingsegregation of two al- enies from three maternalparents of E. fistu- leles at one locus. These bands often stained losum always expressed four-bandedpatterns quite faintly and were not always detectable. (Fig. IF) and this species is consideredto have The more cathodal zone of activity consisted a total of at least fourgene loci specifyingPGM. of a fixed four-bandedpattern with the most In the remainingsix species, four- to six-band- cathodal of these bands being much fainter August 1989] YAHARA ET AL. -GENE DUPLICATIONS IN EUPATORIUM 1251 than the other three. It was not possible to some number 2n = 20 has been reported for ascertainwith certaintywhether this represents these seven species. With regardto E. purpu- enzyme activity or is some sort of "ghost." reum, an exceptional tetraploid was reported These four species are considered to have at from Canada(Grant, 1953) but sexual diploids least three isozymes, one in the more cathodal are widespread in the United States (Wata- and two in the more anodal zone of activity. nabe, 1986). In E. semiserratum(Fig. 1P), a seven-banded The second reason for choosing these eight patternwas seen in all 42 individuals collected species of Eupatorium is that they represent from Georgia and Louisiana. It was not pos- the three morphologicallydistinct groups rec- sible to interpretthis very complex patternge- ognized for the American representatives of netically,but it appearslikely that at least four the genus (King and Robinson, 1970a, 1987; gene loci are expressed. In E. fistulosum, E. Sullivan, 1972). Eupatorium fistulosum, E. maculatumand E. mikanioides,IDH was not maculatum, and E. purpureumare from the resolved well. Eutrochium group, E. capillifolium from the Traganthes group, and E. mikanioides, E. per- SKDH- Six- or seven-bandedpatterns were foliatum, E. semiserratum,and E. serotinum observed in certain individuals of E. perfolia- fromthe Uncasiagroup. Thus, these five species tum (Fig. 1S) and E. mikanioides. Because representa cross section of the variation found SKDH is a monomeric enzyme, these species in American Eupatorium. must have at least four gene loci specifying it. Gottlieb (1982) summarized data demon- Eupatoriumfistulosum (Fig. 1R), E. semiser- stratingthat diploid plants have two isozymes ratum (Fig. IT), E. purpureumand E. capil- for enzymes ofglycolysis and the pentose phos- lifolium sometimes had five or more bands, phate pathway, including the enzymes ALD, indicatingthat at least three loci are expressed PGI, PGM, 6PGD and TPI examined in this in these species. In E. maculatumand E. sero- study. In each case, one of these isozymes is tinum, SKDH was not resolved well. cytosolic and the other is localized in the plas- tids. Isocitratedehydrogenase in higher plants TPI-We were unableto document gene du- typicallyhas two isozymes, one being cytosolic plications for this enzyme. In progenies of E. and the other in plastids (Randall and Givan, fistulosum and E. perfoliatum, one- or three- 1981; Gottlieb, 1987). Shikimate dehydroge- banded patterns were segregatedwith an ad- nase in higher plants also has two isozymes, ditional invariable band, indicating that two one in the plastids and the other possibly being loci, one monomorphic and the other poly- cytosolic (Mousdaleand Coggins, 1985). Thus, morphic, are expressed in these species. Re- the minimal conserved number of isozymes in sults from chloroplastisolation for E. perfolia- diploid plants of all the enzymes included in tum suggestedthat the monomorphic isozyme this study is two. This means that for the di- is cytosolic. This isozyme migrated between meric enzymes (IDH, PGI, 6PGD, and TPI) the fast and slow homodimeric allozymes of six is the greatestnumber of bands that would the chloroplasticisozyme. One- or three-band- be expectedwhen two genes areexpressed. This ed patternswith an additional invariableband banding pattern would occur when a plant is were also observed in E. capillifolium. In E. heterozygous at both loci. For the two mo- maculatum,two-, four-, and six-banded pat- nomericenzymes PGM and SKDH, fourbands terns were found, indicating that two poly- wouldbe the maximumnumber observed when morphic loci were expressed. In E. mika- two gene loci are expressed. nioides,E. semiserratum,E. serotinumand E. The complex banding patterns resolved for purpureum,two bands were always seen. These all enzymes except TPI indicate that three or species are regardedas having two monomor- more isozymes occur in all species examined. phic loci. The multiplicity in these enzymes is also ex- pressed in some Asian species of Eupatorium DISCUSSION-The eight species included in (Kawaharaet al., unpublished data). While it this study were selected with two criteria in would be highly desirableto have more exten- mind. First, seven species (E. purpureumis the sive genetic data for the banding patterns(this exception) are all strictly diploid and sexual, work is in progress),in certain instances vari- and polyploidy and agamospermy are not ation in bandingpatterns among the progenies known. Cytology of American species of Eu- of individual plants, variation among individ- patoriumhave been studied extensively (Sul- ual plants within populations (together with livan, 1972, 1976; Watanabe, 1986; Watanabe knowledge of the active subunit composition et al., in preparation)and only the chromo- of the enzymes), and subcellular localization 1252 AMERICAN JOURNAL OF BOTANY [Vol. 76 of isozymes of several enzymes combine to and Huziwara, 1982; Watanabe, unpublished allow for rather strong inferences about the data). Thus, cytological features seem to in- genetic bases of the banding patterns. In all dicate an amphidiploid or diploid condition in cases, one must hypothesize gene duplication these "diploid"species of Eupatorium.As stat- as the basis of the banding patterns. ed earlier, American species of Eupatorium The extensive isozyme multiplicity in dip- have diversified into three morphologically loid Eupatoriumcontrasts sharplywith many distinct groups. It seems unlikely, therefore, other members of the Asteraceae,such as Aster that evolution and diversification have oc- and Machaeranthera of Astereae (Gottlieb, curredrecently. Eupatorium has a disjunctdis- 198 ib), Lasthenia (Crawford, Ornduff, and tribution in eastern North American and East Vasey, 1985) of ,and Crepis(Roose Asia, and it is hypothesized to be of Tertiary and Gottlieb, 1978) and Stephanomeria(Gal- origin (King and Robinson, 1970b; Watanabe, lez and Gottlieb, 1982) of Lactuceae,in which 1986).If this hypothesisis correct,then it means diploid species with haploid chromosome that extensive gene duplication and fixed het- number less than 10 typically have the con- erozygosityhave been maintained for a rather servedisozyme numberfor most enzymes. The long geologicalperiod subsequentto the origin reason or reasons for the extent of gene du- of the genus and through extensive chromo- plicationin these taxonomicallydiverse species somal-diploidization. of Eupatorium are not known. Although x = If, by contrast, 2n = 20 does representthe 10 has been regardedas the originalbasic chro- originaldiploid number for Eupatorium,then mosome numberfor the tribeEupatorieae (King extensivegene duplicationshave occurredboth et al., 1976; King and Robinson, 1987), a chro- in eight species belonging to the three mor- mosome number of n = 5 was reported for phological groups in North America and in Adenostemma brasiliense (Pers.) Cass. (Turner certain species in Asia. If this were the case, and Irwin, 1960), now correctlydetermined as then it would mean that the mechanism (or A. involucratumR. King et H. Robinson (King mechanisms) producing these gene duplica- et al., 1976). King et al. (1976) and King and tions in Eupatorium is rare in diploid plants Robinson (1987) suggestedthat the n = 5 num- because Clarkia is the only other genus that ber may be derived from n = 10 because the shows such extensive duplication of isozyme latter number is prevalent in many genera of loci among plants viewed as unquestionably the tribe Eupatorieae,including Adenostemma diploid (Gottlieb, 1986). Even in Clarkia,du- itself. It is also notable that n = 4 has been plication has been confirmed on at most four reported for Eupatorium microstemon Cass. enzymes, ADH, PGI, PGM and 6PGD. We (Baker, 1967), now treated as Fleischmannia considerit more likely that plants with a chro- microstemon(Cass.) R. King et C. Robinson mosome number of 2n = 20 in Eupatorium (King et al., 1976). The same chromosome are of polyploid origin. number was also determined for Fleischman- nia hymenophylla(Klatt) R. King et C. Rob- LITERATURE CITED inson (Graschoff, Bierner, and Northington, BAKER,H. G. 1967. The evolution of weedy taxa in the 1972). Baker (1967) considered n = 4 to be a Eupatorium microstemon species aggregate. Taxon 16: derived numberbecause the chromosomes are 293-300. much largerthan those found in plants with n CARDY,B. J., C. W. STURBER,ANDM. N. GOODMAN.1981. = 10. Additional evidence for this hypothesis Techniques for starch gel electrophoresis of enzyme from maize (Zea mays), revised. Institute of Statistics came from meiotic behavior of hybrids be- Mimeograph Series No. 1317. North Carolina State tween plants with the chromosome numbers n University, Raleigh. = 4 and n = 10. King et al. (1976) cited Baker's CHASE, M. W., AND R. G. OLMSTEAD. 1988. Isozyme statement with a note on the other species of number in subtribe Oncidiinae (Orchidaceae): an Fleischmannia having n = 10 and regardedn evaluation of polyploidy. Amer. J. Bot. 75: 1080- = 4 as the derived state. Our results, however, 1085. CRAwFoRD, D. J., R. ORNDuFF,AND M. C. VASEY. 1985. cast some doubt on the hypothesis that x = 10 Allozyme variation within and between Lasthenia mi- is the original base number of the tribe Eu- nor and its derivative species, L. maritima (Astera- patorieae, and it appears that more extensive ceae). Amer. J. Bot. 72: 1177-1184. cytological surveys of Adenostemma and GALLEZ,G. P., AND L. D. GOTTLIEB. 1982. Genetic evi- Fleischmannia are needed. dence for the hybrid origin of the diploid plant Steph- Species of Eupatoriumwith a chromosome anomeria diegensis. Evolution 36: 1158-1167. GASTONY,G. J., ANDD. C. DARRow. 1983. Chloroplastic = number of 2n 20 are not autotetraploid and cytosolic isozymes of the homosporous fern karyotypicallyif they were polyploids (Wata- Athyriumfilix-femina L. Amer. J. Bot. 70:1409-1415. nabe et al., in preparation),and they form nor- GOTTLIEB, L. D. 1981a. Electrophoretic evidence and mal bivalentsat meiosis (Watanabe,Fukuhara, plant populations. Progr. Phytochem. 7: 1-46. August 1989] YAHARA ET AL. -GENE DUPLICATIONS IN EUPATORIUM 1253

* 1981b. Gene numberin species of Astereaethat genasein Clarkia(Onagraceae) and theirphylogenetic havedifferent chromosome numbers. Proc. Natl. Acad. implications. Syst. Bot. 9: 479-489. USA 78: 3726-3729. RANDALL, D. D., AND C. V. GIVAN. 1981. Subcellular * 1982. Conservationand duplicationof isozymes locationofNADP+-isocitrate dehydrogenase in Pisum in plants. Science 216: 373-380. sativumleaves. P1.Physiol. 68: 70-73. * 1983. Isozyme numberand plant phylogeny.In ROOSE,M. L., AND L. D. GOTTLIEB.1978. Stability of U. Jensenand D. E. Fairbrothers[eds.], Proteinsand structuralgene numberin diploid species with differ- nucleic acid in plant systematics,209-221. Springer- ent amounts of nuclearDNA and differentchromo- Verlag,Berlin. some numbers.Heredity 40: 159-163. * 1984. Isozyme evidence and problemsolving in SOLTIS,D. E., A. J. GILMARTIN, L. RIESEBERG, AND S. plant systematics.In W. F. Grant [ed.], Plant biosys- GARDNER. 1987. Genetic variationin the epiphytes tematics, 343-357. Academic Press, Orlando,FL. Tillandsia ionantha and T. recurvata (Bromeliaceae). . 1986. Genetic differentiation,speciation and Amer. J. Bot. 74: 531-537. phylogenyin Clarkia (Onagraceae).In K. Iwatsuki, , C. H. HAUFLER,D. C. DARRow,AND G. J. GASTONY. P. H. Raven, and W. J. Bock [eds.], Modem aspects 1983. Starch gel electrophoresisof ferns: a compi- of species, 145-166. University of Tokyo Press, To- lation of grindingbuffers, gel and electrode buffers, kyo. and stainingschedules. Amer. Fern J. 73: 9-27. . 1987. Phosphoglucomutaseand isocitrate de- , AND P. S. SoLTIs. 1987. Polyploidyand breeding hydrogenasegene duplications in (Compositae). systemsin homosporouspteridophyta: a reevaluation. Amer. J. Bot. 74: 9-15. Amer. Naturalist 130: 219-232. GRANT,W. F. 1953. A cytotaxonomicstudy in the genus , ArN . 1988. Are lycopods with high chro- Eupatorium.Amer. J. Bot. 40: 729-742. mosome numbers ancient polyploid? Amer. J. Bot. GRAscHoFF,J. L., M. W. BIERNER, AND D. K. NORTHING- 75: 238-247. TON. 1972. Chromosome numbers in North and SoLTIs,P. S., ANsmD. E. SoLTIs. 1987. Populationstruc- CentralAmerican Compositae. Brittonia 24: 379-394. ture and estimates of gene flow in the homosporous HAUFLER, C. H. 1987. Electrophoresisis modifyingour fern Polystichum munitum. Evolution 41: 620-629. conceptsof evolution in homosporouspteridophytes. SuLLIvAN,V. I. 1972. Investigationsof the breedingsys- Amer. J. Bot. 74: 953-966. tems, formation of auto- and alloploids and the re- AND D. E. SoLTIs. 1986. Genetic evidence sug- ticulate patternof hybridizationin North American gests that homosporousferns with high chromosome Eupatorium(Compositae). Ph.D. dissertation,Flor- numbersare diploid. Proc. Natl. Acad.USA 83:4389- ida State University, Tallahassee. 4393. 1976. Diploidy, polyploidy and agamospermy KING,R. M., D. W. KYHOs,A. M. POWELL,P. H. RAVEN, among species of Eupatorium(Compositae). Canad. AND H. ROBINSON. 1976. Chromosomenumbers in J. Bot. 54: 2907-2917. Compositae XIII. Eupatorieae.Ann. Missouri Bot. TURNER,B. L., ANDrsH. S. IRwIN. 1960. Chromosome Gard. 64: 161-180. numbers in the Compositae. II. Meiotic counts for , AND H. ROBINSON. 1970a. Studies in the Eu- fourteen species of Brazilian Compositae. Rhodora patoriae (Compositae)XXV. A new genus Eupato- 62: 122-126. riadelphus.Phytologia 19: 431-432. WATANABE,K. 1986. The cytogeographyof the genus , AND . . 1970b. Eupatorium,a composite Eupatorium(Compositae)-a review.P1. Species Biol. genus of Arcto-Tertiarydistribution. Taxon 19: 769- 1: 99-116. 774. , T. FuKuHARA,AND Y. HuzIwARA.1982. Studies AND . 1987. The generaof the Eupatorieae on the Asian Eupatorias I. Eupatorium chinese var. (Astraceae).Monogr. Syst. Bot. Missouri Bot. Gard. simplicifoliumfrom the Rokko Mountains.Bot. Mag. 22: 1-581. (Tokyo) 95: 261-280. MOUSDALE,D. H., AND J. R. COGGINS.1985. Subcellular WITTER,M. S. 1988. Duplicateexpression ofbiochemical localization of the common shikimate-pathwayen- gene markersin the Hawaiian silversword alliance zymes in Pisum sativumL. Planta 163: 241-244. (Madiinae;Compositae). Biochem. Syst. Ecol. 16:381- ODRZYKOSKI, I. J., AND L. D. GOTTLIEB. 1984. Dupli- 399. cationsof genes coding 6-phosphogluconatedehydro-