植物研究雑誌 J. J. Jpn. Bo t. Originals Originals 73: 241-247 (1998)

Allozyme Variation in Six Flowering Species Characterizing Characterizing Ullung Island ，Korea

Hye Gi CHUNG ， Jae l¥ lin CHUNG and Myong Gi CHUNG

Department Department ofBiology ，Gyeongsang National University ，Chi 吋 u，660-701 KOREA (Received (Received on February 18 ，1998)

A prelimin 訂 y allozyme study on six populations of Fagus multinervis (endemic to the the Ullung Island) ， Pinus parviflora ， Hepatica maxima (endmic to the Island) ， takesimana takesimana (endemic to the Island) ，Tim ・ella polyphylla ， and Thymus quinquecostatus va r. japonicus on the Ullung Island from Korea was made. Populations of F. multinervis and and P. parviflora maintain moderate levels of allozyme variation (mean expected heterozygosity heterozygosity ofO.253 in F. multinervis and 0.149 in P. parviflora) ， whereas those of H. maxima (0.032) ，仁 takesimana (0.021) ，T. polyphylla (0.016) ， and T. quinquecostatus var.japonicus var.japonicus (0.066) maintain low levels of allozyme diversity within populations. For the the latter four taxa ，factors such as founder effect ，genetic drift associated with small population population size ， and probable inbreeding contributed to their situation.

The Ullung Island ，75 .4 km 2 area ， is 10- (Yim 1993). catedat catedat 130 0 E and 37 0 N ， ca.141 kmeastofthe Estimates of levels and distribution of ge- Korean Peninsula. The Island had been formed netic diversity within and among rare 'and by volcanic activities from the late Tertiary to endangered species could be useful to provide the the Quaternary. The flora on the Island has a basis on which to build sound programs for been ofbotanical interest because several sub- the conservation of genetic diversity of these tropical tropical grow on the Island owing to a plants (Soule 1986). In addition ，genetic diver- warm current ，though the Island is located in sity routinely from allozyme markers can be

the the northern temperate region. About 500 vas- used as a yardstick to measure the effective 同 cular cular plants grow on the Island (Oh 1978) ，and ness of in situ and ex situ conservation pro- among them ca. 40 species are endemics. Re- grams (Hamrick et al. 1991). Despite the im- cently ， the number ofvisitors to the Island has portance of know ledge concerning genetic been rapidly increased (several thousands in variation for providing information for con- summer season) because several express pas- ser 、ration purposes ，little is known of the ge- senger senger boats are available. In addition ，several netic variation in plants native to the Island. roads roads on hillsides have been recently enlarged Island plantpopulations generally harborfewer and newly constructed. These activities could polymorphic loci and fewer alleles per locus severel severel y affect the flora and natural habi tats on than mainland populations because they are the the Island. For these reasons ， the Korean gov- probably descended from one or a few found- ernment has designated six sites of natural ers (Hamrick et al. 1979). forests forests as N atural Monuments to protect rare As a part of studies to preserve plant diver- and endemic plants from human disturbances sity on the Island ， the objective ofthis prelimi-

一 241- 242 植物研究雑誌第73 巻第5号 平成10 年10 月 nary nary study is to explore patterns of genetic subshrub ， grows on limestone and andesite variations variations in six plant taxa on the Island. F agus regions in Japan (Hokkaido ，Honshu and multinervis multinervis N akai (Fagaceae) ， an endemic plant Kyushu) ， the northern China ，Mongolia ， and to to the Island ，is an element of cool-temperate China (Ki tamura and Murata 1986). Only one 2 forests forests in East Asia. It is a large ，long-lived ， population (3 ，226 m ) at the Nari Basin is monoecious monoecious tree ， and its wind-bome pollen known on the Ullung Island and has been and and potentially animal-dispersed seeds may designated as the N atural Monument in Korea allow allow fairly long-distance gene dispersal on and is now surrounded by a wire fence because the the Island. Pinus parviflora Siebold & Zucc. of its rarity in South Korea. Flowering indi- (Pinaceae) (Pinaceae) is distributed in J apan (westem viduals of T. quinquecostatus var. japonicus Honshu ， Shikoku and Kyushu) and Ullung are visited by bees (M. Chung pers. obs よ Island. Island. Pinus parviflora is also a component of cool-temperate cool-temperate forests in the region. On the Materials and Methods

Ullung Ullung Island ， only a few natural stands in 田 A total of 300 leaf samples was randomly cluding cluding one N atural Monument stand are collected from six populations of the six taxa known (Ahn 1971) ， and the number of indi- on the Ullung Island. Mature leaves (or nee- viduals viduals in each stand are less than hundreds dles) were collected from 50 individuals per (M. (M. Chung pers. obs.). Like other conifers ， population. For Fagus multinervis and Pinus Pinus Pinus parviflora is a large ，long-lived ，wind- parviflora ，individuals (30-60 years old) were 2 pollinated pollinated monoecious tree. The seeds are collected from area covered ca. 5，000 m . Leaf 1.5 1.5 cm long ， 0.8 cm wide with 4 mm wing samples of Hepatica maxima ，Campanula (Ahn 197 1). Hepatica maxima Nakai takesimana ， and Tiarella polyphylla were col- 2 (Ranunculaceae) ，a herbaceous perennial ，is lected from areas covered ca. 500 m . Leaf endemic endemic to the Island. White flowers (1.5 cm samples ，wrapped with a wet paper towel ， wide) wide) are visited by bees (Chung pers. obs.). placed in plastic bags and stored on ice during The achene is 5 mm long. Local populations of transport from the Island to the laboratory. H.m αxima are found in patches ， and isolated Samples were then stored at 4 0 C until proteins from each other under hillside forests on the were extracted. Island. Island. Camp αnula takesimana N akai Leaf samples were cut finely and crushed

() ， an endemic of the Island ， with a mortar and pestle. A phosphate-polyvi 国 grows grows on the forest margin near coastal areas. nylpy 町 olidone extraction buffer (Mitton et al. Whitish Whitish pu 中 le flowers (3-5 cm long) are 1979) was added to leaf samples to facilitate visited visited by bees and each capsule contains crushing and to aid enzyme stabilization. En- hundreds of seeds. In Korea ， Tiarella zyme extracts were absorbed onto 4x6 mm polyphylla polyphylla D.Don (Saxifragaceae) only oc- wicks cut from Whatman 3 mm chromatogra- curs curs under oak forests near the summit of the phy paper ，which were stored at -70 o C until Ullung Ullung Island ， but the species widely distrib- needed. Electrophoresis was performed using utes utes inJapan (Hokkaido ，Honshu and Shikoku) ， 10.5% starch gels to resolve scorable loci for Taiwan ，China and Himalaya (Kitamura and 13 enzyme systems: acid phosphatase (ACP) ， Murata 1987). Small white flowers are visited triosephosphate isomerase (TPI) ，phosphoglu- by bees (M. Chung pers. obs.) ， and each cap- comutase (PGM) ，fluorescent esterase (FE) ， sule sule contains several small seeds (1 .2-1 .4 mm peroxidase (PER) ， alcohol dehydrogenase long). long). Thymus quinquecostatus Cela k. var. (ADH) ，menadione reductase (MNR) ， japonicus japonicus H.Hara (Lamiaceae) ，a deciduous phosphoglucoisomerase (PGI) ， diaphorase October October 1998 Journal of Japanese Botany Vo l. 73 No. 5 243

(DIA) (DIA) ，isocitrate dehydrogenase (IDH) ，rnalate dures for DIA followed the rnethod described dehydrogenase dehydrogenase (MDH) ，6-phosphogluconate by Cheliak and Pitel (1 984). All other stain dehydrogenase dehydrogenase (PGD) ， and skirnate dehydro- recipes were identical to those described by genase genase (SKDH). Buffer systerns and putative Soltis et al. (1 983). Putative loci were desig- loci loci are listed in Table 1. The staining proce- nated sequentially ， with the rnost anodally

Table Table 1. List ofbuffer systerns and resolvable enzyrne loci. FM ，Fagus multinervis; PP ， Pinus

P卯α門川ifl μlor α，

and and TQ ，Thymus quinquecost α， tus v訂 . japonicus: M indicates a rnonornorphic locus ，P indicates indicates a polyrnorphic locus ， X indicates that the locus was poorly resolved could not be scored. scored. For TQ ，diaphorase (DIA) and 6-phosphogluconate dehydrogenase (PGD) were resolved resolved on a rnorpholine citrate buffer systern

Buffer Buffer systern Loci FM PP HM CT TP TQ PMMPMM Acph-1 Acph-1 X P M M 乱4 Acph-2 Acph-2

Acph ・3 Acph-4 Acph-4 MMMX MMMM Tpi-1 Tpi-1 M M M Tpi-2 Tpi-2 M

Tpi ・3 ppp MPM Pgm-1 M P Pgm-2 MMMM Fe-1 Fe-1 P M M Fe-2 Fe-2 MMM MMMMPM Per M M 島f A rnodification Adh-1 孔f M 乱f (Hau f1 er 1985) Adh-2 of of Soltis et al. Adh-3 MMMMPMM (1 983) systern 6 Mnr ・-1 p P X M Mnr-2 p M Mnr-3 MPMM MMP MMMMMMMM Pgi-1 Pgi-1 M

Pgi ・2 p Dia-1 Dia-1 Dia-2 Dia-2 Dia-3 Dia-3 MMPM MMMMMMM MMMMMMM Idh-1 Idh-1 P M M A rnodification Idh-2 M (Chung and Kang Mdh-1 恥f M 1994) 1994) of Soltis Mdh-2 p eta l. (1 983) Mdh-3 Systern Systern 11 Pgd-1 p M M M Pgd-2 Pgd-2 p 乱4

Ab(attu hvdl H &・・も nfmnYRut-eO mmCch9tvdT ψi30、)戸し vst Skdh-1 x 孔f 恥f X M 町 剖旬 Skdh-2 Skdh-2 x M

が日 6.uba ゐ vdkB rl ノ 也、 d パい n .uFら 244 244 植物研究雑誌 第73 巻 第5号 平成10 年 10 月 migrating migrating isozyme designated ‘1' ， the next 0.253) (Table 2) ， though the species is only ‘2' ，and so on. Likewise ，alleles were desig- restricted to the Ullung Island. A very similar nated nated sequentially with the most anodally mi- result was reported within 14 populations of F. grating grating allele designated ‘a'. All isozymes crenata in the Hokkaido and Honshu regions resolved resolved by the methods ， expressed pheno- of northeastern J apan: mean percentage of types types that were consistent in subunit structure polymorphic loci ， the number of alleles per and genetic interpretation with most isozyme locus ，mean effective number of alleles per studies studies in plants ， as documented by Weeden locus ，and the expected heterozygosity were and Wendel (1989). 58% ，2.62 ，1.34 ，and 0.202 ，respectively Levels of allozyme variation within (Takahashi et al. 1994). When we compare populations populations were estimated for individual levels of allozyme variation observed in F. populations populations with a computer program devel- crenata with those of F. multinervis ，it could oped by M D. Loveless and A. Schnabel (pers. be said that populations of F. multinervis on comm.): comm.): percent polymo 叩hic loci (P) ，aver- the Ullung Island are not genetically age age number of alleles per locus (A) ，effective depauperate. number of alleles per locus (Ae) ， observed Pinus parviflora at the Ullung Island popu- heterozygosity heterozygosity (Ho) and expected heterozy- lation also maintains moderate levels of gosity gosity (He). allozyme variation (P = 33 %， A = 1.4 8; Ae = Wrigh t' s (1 922) fixation indices (め were 1.30; andHe = 0.149) (Table 2) ， comparable to calculated calculated for each polymorphic locus and for mean values of gymnosperms ，averaging 53 % each each population to reveal significant devia- of P ，1. 83 of A ，1. 20 of Ae ，and 0.1 51 of He tions tions in genotype frequencies from expecta- (Hamrick et al. 1992). Like other conifers ， tions tions under Hardy-Weinberg (H-W) equilib- factors such as longevity and outcrossing- rium (Li and Horvitz 1953). These indices wind pollination might contribute to harbor a were tested for deviations from 0 by aχ2 moderate level of allozyme variation within a statistic statistic following Li and Horvitz (1953). population of P. parv 伊ora. Although Although 1 calculated at each of one popu- Results Results and Discussion lation of Fagus multinervis and Pinus F agus multinervis maintains higher levels parviflora ，inbreeding coefficients showed that of of genetic variation within a population (He = significant deficiencies of heterozygotes exist

Table Table 2. Summary of allozyme variation within populations of FM ，Fagus multinervis; PP ， Pinus Pinus parviflora; HM ， Hepatica m α xima; CT ，Campanula takesimana; TP ， Tiarella polyphyll α; and TQ ，Thymus quinquecostatus va r. japonicus: NL (PL) ，Number of loci scored scored (number of polymorphic loci); P ， percentage of polymorphic loci; A ，mean number of of alleles per locus; Ae ，mean effective number of alleles per locus; Ho ，mean observed heterozygosity; heterozygosity; and He ，mean expected heterozygosity

Taxon NL (PL) P A Ae Ho (SE) He (SE) 1122121413458 FM (8) 57.14 1. 93 1. 57 0.211 (0.048) 0.253 (0.068) PP (7) 33.33 1. 48 1. 30 0.107 (0.036) 0.149 (0.050) HM (2) 8.70 1. 08 1. 05 0.021 (0.017) 0.032 (0.023) CT (1) 7.14 1. 07 1. 03 0.014 (0.004) 0.021 (0.021) TP (1) 4.00 1. 04 1. 03 0.011 (0.013) 0.016 (0.016) TQ (3) 16.67 1. 17 1. 11 0.044 (0.025) 0.066 (0.036) October October 1998 Journal of Japanese Botany Vo 1. 73 No. 5 245 for for two (Mnr-l ，0.736 ，P < 0.001 and Pgd-l ， stricted ecological distribution ， growing only 0.4 46 ，P < 0.01) ofthe eight loci ， with a mean in deciduous and conifer forests on the Ullung F of 0.168 ，and five (Pgm-l ，0.598 ，P < 0.001; Island (Nevo et al. 1984). Although Tiarella Pgm-2 ，0.326 ，P < 0.01; Fe ，0.507 ，P < 0.001; polyphylla has a broader distributional range Pg ι2，0.374 ，P < 0.01; andMdh-l ，0.340 ，P < in East Asia ， the population on the Ullung 0.01) 0.01) ofthe seven loci ， with ameanF ofO.336. Island is allozymically depauperate. Only one Although Although F. multinervis is monoecious ， the locus (Mnr-l) of 25 loci investigated was significant significant heterozygote deficiencies at two polymorphic. Again ， the population of Thy- loci loci indicate that consanguineous matings mus quinquecostatus var. japonicus on the might might occur within a population. We collected Ullung Island also reveals the paucity of samples samples within a 100x50 m area ， several allozymic variability ， though this taxon also patches patches might be included in the electrophoretic distributed in East Asia. Only three (Acph-l ， analysis. analysis. This may create a Wahlund effect Acph-4 ， and Pgm- 2) of 18 loci examined were causing causing heterozygote deficiencies. Higher lev- polymorphic. A very similar result was ob- els els ofheterozygote deficiencies compared with served in Lactoris fernandeziana Phi l. most conifers were observed at one population (Lactoridaceae) on the Juan Fernandez Island. of of Pinus parviflora. This in part may be due to The species revealed no variation at 22 loci logging logging activities during the past several hun- (Crawford et al. 1993). The authors suggested dreds dreds and a small number of individuals per that factors such as founder effect ， genetic extant extant stand (l ess than hundreds). drift associated with small population sizes ， Each of one population of Hepatica maxima and a selfing breeding system contributed to and and Campanula takesimana maintains low the lack of allozyme variation in Lactoris levels levels of allozyme variation: P of 8.7 % and fernandeziana. As Tiarella polyphylla and 7.1 7.1 %;A of 1. 08 and 1. 07;Ae of 1. 05 and 1. 03 ， Thymus quinquecostatus var.japonicus do not andHeofO.032andO.021 ，respectively. These occur in mainland Korea ， but grow in the estimates estimates were lower than those of the other J apanese Archipelago and the Ullung Island mainland mainland Hepatica and Campanula species in had connected to the J apanese Archipelago Korea and J apan. For example ，Korean (Oh 1978) ，it is highly probable that the pro- populations populations of H. asiatica N akai (four genitor populations of these two taxa had mi- populations) populations) and H. insularis Nakai (six grated from the northern J apan into the Ullung populations) populations) revealed mean percentage of poly- Island during the Ic e Age (the glacial “Wurm"). morphic morphic loci (P) of 79 % and 54 %， mean The population size of thse two taxa on the number of alleles per locus (A) of 2.76 and Ullung Island (Chung pers. obs.) is small. 2.31 ， and mean expected heterozygosity (He) After the last Ic e Age ， the remnants of these ofO.248 ofO.248 and 0.212 ，respectively (Ki m andLee two taxa have been isolated to the Island with 1994). 1994). Seventeen Japanese populations of little opportunity forreceiving propagules (e.g. ，

Campanula punctat αLam.， harbor moderate pollen and seeds) from the Japanese Archi- levels levels of allozyme variation (He = 0.123 from pelago. No special seed dispersal mechanism seven seven mainland populations and 0.088 from is known to these two taxa. In addition ， the two ten ten populations of the Izu Islands) (l noue and taxa on the Ullung Island grow only a few Kawahara 1990). The low levels in populations location (Tiarella polyphylla) and only one of of H. maxima and C. takesimana may be location (Thymus quinquecost α tus va r. attributable attributable to relatively small ， isolated japonicus) ，indicating restricted ecological

populations ，endemism to an island ， and re 回 distribution. These information and allozyme 246 246 植物研究雑誌第73 巻第5号 平成10 年10 月 data data suggest that founder effect ，especially 一一一ーヲ Linhart Y. B. and Mitton J. B. 1979. Relation- genetic genetic drift ，and probable inbreeding may be ships between life history characteristics and electrophoretically electrophoretically detectable genetic variation in responsible responsible for the considerably low levels of plants. Annu. Rev. Eco l. Sys t. 10: 173-200. allozyme allozyme variation within populations of these Haufler C. H. 1985. Enzyme variability and modes of two taxa. For a future study ，it would be good evolution in Bommeria (Pteridaceae). Sys t. Bo t. 10: to to examine diversity in mainland populations 92-104. Kim J. H. and Lee N. S. 1994. Allozyme variation in of of the non-endemic taxa (e.g. ， Tiarella Hepatica αsiatica and H. insularis. Kor. J. Plant Tax. polyphylla polyphylla and Thymus quinquescost α tus va r. 24: 79-93 (in Korean). japonicus) japonicus) to compare with populations on the Kitamura S. and Murata G. 1986. Colored Ill ustrations of Island. Island. This would indicate how much diver- Herbaceous Plants of Japan Vo l. 1 (Sympetalae). sity sity has been lost in establishment on the Hoikusha Pub l. Co. ， Osaka. 一一一一- and 一一一一 1987. Colored Ill ustrations of Island. Island. Herbaceous Plants of Japan 2 (Choripetalae). Hoikusha Hoikusha Pub l. Co.. Osaka. 1 thank S. S. Kang ， N. W. Lee ，1. K. Hong Li C. C. and Horvitz D. G. 1953. Some methods of and Pro f. S. Kawano for assistance. This re- estimating the inbreeding coefficien t. Amer. J. Hum. Gene t. 5: 107-117. search search was supported by a grant from the Mitton Mitton J. B.， Linhart Y. B. ， Sturgeon K. B. and Hamrick Korea Science and Engineering Foundation J. L. 1979. Allozyme polymorphisms detected in (96-0500-006-2) (96-0500-006-2) to MGC. mature needle tissue of ponderosa pine. J. Hered. 70: 86-89. 86-89. References References NevoE. ，Beiles A. andBen-Shlomo R. 1984. Theevolu- Ahn K. Y. 197 1. Principal characteristics of Pinus tionary significance of genetic diversity: ecological ， parv{ β ora S. etZ. native to the DageletIsland. J. Kor. demographic and life history correlates. In: Many G. Fo r. Soc. 12: 31-43 (in Korean). S. (ed.) ，Evolutionary Dynamics of Genetic Diver- Cheliak Cheliak W. M. and Pitel J. L. 1984. Techniques for starch sity: Lectures Notes in Biomathematics. pp. 13-2 1. gel gel electrophoresis of enzymes forforest tree species. Springer Verlag ，Berlin. Petawawa National Forestry Ins t.， Information Re- Oh S.- Y. 1978. Fl oral studies on the vascularplants ofthe port port PI-X-42. Canadian Forestry Service ，Agricul- DageletIsland. Res. Rev. KyungbookNat l. Univ. 25: ture ， Chalk River ，Ontario. 131-201 (in Korean). Chung M. G. and Kang S. 1994. Genetic variation and Soltis D. E. ，Haufler C. H. ，Darrow D. C. and Gastony G. population population structure in Korean populations of Eurya J. 1983. Starch gel electrophoresis offerns: a compi- japonica japonica (Theaceae). Ame r. J. Bo t. 81: 1077-1082. lation of grinding buffers ， gel and electrode buffers ， Crawford Crawford D. J. ， Stuessy T. F.， Cosner M. B. and Haines and staining schedules. Ame r. Fern J. 7: 9-27. D. D. W. 1993. Lactorisfernandezi αna (Lactoridaceae) Soule M. E. 1986. Conservation Biology: the Science of on the J apan Fernandez Islands: allozyme uniformity Scarify and Diversity. Sinauer ，Sunderland. and and field observations. Conse r. Bio l. 8: 277-280. Takahashi M. ，Tsumura Y. ，Nakamura T. ， Uchida K. and Hamrick J. L.， Godt M. J. W. ，Murawski D. A. and Ohba K. 1994. Allozyme variation of F agus crenata Loveless Loveless M. D. 199 1. Correlations between species in northeastern Japan. Can. J. Fo r. Res. 24: 1071- traits traits and allozyme diversity: implications for con- 1074. servation servation biology.ln: F a1 k D. A. and Holsinger K. E. Weeden N. F. and Wendel J. F. 1989. Genetics ofplant (eds.) ，Genetics and Conservation ofRare Plants. pp. isozymes. In: Soltis D. E. and Soltis P. S. (eds.) ， 75-86. 75-86. Oxford University Press ，New York. Isozymes in Plant Biology. pp. 46-72. Dioscorides ， 一一一一一，一一一一- and Sherman-Broyles. 1992. Factors Portland. influencing influencing levels of genetic diversity in woody plant Yim G.-B. 1993. Botanical Treasures (Natural Monu- species. species. New Fo r. 6: 95-124. ments) of Korea. Daewonsa ， Seoul (in Korean).

鄭孝基，鄭載E民，鄭明基:酵素多型を用い た韓国欝陵島産植物 6 種の遺伝的多様性に関する 研究 ハマソウ，タケシマホタルブクロ) ，韓国では穆陵 韓国欝陵島固有の 3 種(タケシマブナ，オオス 島のみに産する 2種(ヒメコマツ，ズダヤクシュ) ， October October 1998 Journal of Japanese Botany Vo l. 73 No. 5 247

および韓国内の産地が限られているイブキジャコ 度の比較的高い遺伝的多様度を有することがわ ウソウの計 6 種の植物について，保全生態学的見 かったが，他の 4 種では集団内の遺伝的多様度は 地から島内の集団の遺伝的多様性を酵素多型によ 低かった.後者の理由としては，創始者効果，個体 り調査した.その結果 タケシマブナとヒメコマ 群サイズが小さいことによる遺伝的浮動，内交配 ツでは他地域で報告されている同属の植物と同程 等が考えられる(慶尚大学校生物学科)