利用核內基因之序列研究青剛櫟與冷杉之親緣地理形式(2/3)

行政院國家科學委員會專題研究計畫 期中進度報告

利用核內基因之序列研究青剛櫟與冷杉之親緣地理形式

(2/3)

期中進度報告(精簡版)

計 畫 類 別 ： 個別型

計 畫 編 號 ： NSC 95-2621-B-002-002-

執 行 期 間 ： 95 年 08 月 01 日至 96 年 07 月 31 日

執 行 單 位 ： 國立臺灣大學植物科學研究所

計 畫 主 持 人 ： 林讚標

處 理 方 式 ： 期中報告不提供公開查詢

中 華 民 國 96 年 06 月 21 日

PARTIAL CONCORDANCE BETWEEN NUCLEAR AND ORGANELLE DNA

IN REVEALING THE GENETIC DIVERGENCE AMONG QUERCUS

GLAUCA (FAGACEAE) POPULATIONS IN TAIWAN

F. L. Shih,1,_{* Y. P. Cheng,}1,_{* S. Y. Hwang,yand T. P. Lin}2,_*

*Institute of Plant Biology, National Taiwan University, Taipei 106, Taiwan, Republic of China; andy_{Graduate Institute of}

Biotechnology, Chinese Culture University, Yangmingshan, Taipei 111, Taiwan, Republic of China

Quercus glauca (Thunb. ex Murray) Oerst (Fagaceae) has a wide distributional range in Taiwan. In this study, the evolutionary history and the most genetically divergent sites of Q. glauca were studied using a nuclear gene marker, glyceraldehyde-3-phosphate dehydrogenase. Also, the consistency of the results obtained from nuclear gene and cytoplasmic loci was investigated. Using a genealogical approach (TCS software), we determined haplotypes and their relationships to one another. We used the level of divergence for each population from the remaining populations (calculated as mean values of pairwise population differentiation, FST, for each population) to locate the most genetically divergent areas in Taiwan. According to the average FST of each population in comparison with the remaining ones, a peak was found in the northern part of central Taiwan, and another was found in the southeastern region. The peak profiles of the mean FSTvalues for all three DNA data sets (nDNA, cpDNA, and mtDNA) showed similar trends on both sides of the Central Mountain Ridge, except for the mtDNA sequence on the western side. This study suggests that two potential refugia existed in Taiwan during the last glaciation: one in the northern part of central Taiwan and another in southern Taiwan.

Keywords: glyceraldehyde-3-phosphate dehydrogenase, phylogeography, refugium, genetic divergence, Taiwan, Quercus glauca.

Introduction

Using molecular markers in combination with paleoecolog-ical studies for the analysis of the late Quaternary history of angiosperms in order to deduce historical information from their present-day geographical distributions has led to the recognition of glacial refugia of many species of Europe, North America, and Asia (Comes and Kadereit 1998). DNA sequences can even provide evidence of refugium sites that have not been detected by geological or fossil data (Rowe et al. 2004). The power of nucleotide sequences comes from the fact that they can be organized into hierarchically or-dered networks of descent and can provide historical infor-mation that nonordered markers are unable to provide (Schaal et al. 1998). The availability of DNA sequence data and the development of coalescent-based analysis of allele ge-nealogies can therefore form the basis of the study of intra-specific processes within a phylogenetic framework (Avise 2000) and can be used to examine the geographic distribu-tion of genetic variadistribu-tions, postglacial recolonizadistribu-tions, and the ways in which recent evolutionary history has shaped pat-terns of intraspecific variations of a wide range of species (Newton et al. 1999).

Gene flow between populations may be achieved through pollen or seeds. Pollen-mediated gene flow is not necessarily faster than gene flow by seeds; however, the rate of the flow through pollen in tree species like oaks with large heavy seeds is faster than that through seeds because pollen grains are smaller and can be carried more easily over longer distances by diverse agents including wind and insects (Ennos 1994). In addition, the chloroplast genome is haploid, whereas nuclear genomes are diploid or polyploid; thus, the effective popula-tion size of cytoplasmic genomes is half that of the diploid nu-clear genome (McCauley 1995; Moore 1995). The deeper coalescence times expected for nuclear haplotypes will obscure spatial genetic patterns revealed by the plastid genomes. To test this idea, we compared the spatial genetic pattern of the plastid genome, including mitochondrial DNA (mtDNA) and chloroplast DNA (cpDNA), with nuclear DNA (nDNA) of a subtropical tree species, Quercus glauca (Fagaceae). In addi-tion, we examined whether the patterns from different gene markers reflect different genetic histories, which may be inte-grated to postulate the possible historical events that have oc-curred in Q. glauca in Taiwan.

Quercus glauca is a member of the subgenus Cyclobala-nopsis. Cyclobalanopsis is monophyletic within the genus Quercus according to molecular data (Manos et al. 2001). The subgenus Cyclobalanopsis contains ca. 122 species in tropical and subtropical eastern Asia (Luo and Zhou 2001). In consequence, Cyclobalanopsis is considered to be of tropi-cal and subtropitropi-cal origin (Luo and Zhou 2001). Among 1_{F. L. Shih and Y. P. Cheng contributed equally to this article.}

2_{Author for correspondence; fax 886-2-2368-9564; e-mail tpl@}

ntu.edu.tw.

Manuscript received October 2005; revised manuscript received March 2006. 863 Int. J. Plant Sci. 167(4):863–872. 2006.

Ó2006 by The University of Chicago. All rights reserved. 1058-5893/2006/16704-0010$15.00

these 122 species, Q. glauca has the widest range of distribu-tion and extends to the northern limit of this subgenus. The species Q. glauca is found from the Himalayas to Indochina, via China, and has spread to the coast of the west Pacific in places such as Taiwan, the Ryukyus, Japan, and Korea. In Taiwan, Q. glauca is distributed from sea level up to 1700 m in elevation throughout the entire island. It is a codominant tree in subtropical evergreen forests and grows to 20 m in height. It occurs as a pioneer species that prefers open sites such as areas of landslides and windy ridges.

Taiwan is thought to have been connected to the Asian conti-nent during the glacial maximum in the late Pleistocene (Boggs et al. 1979). A more recent land configuration proposed for the late Pleistocene (0.2–0.02 Ma) indicates that a large land bridge extended from eastern China to Taiwan, to the Ryukyus, and probably to the main islands of Japan (Kimura 1996) and should have provided the opportunity for gene flow among haplotypes. The floristic composition of Taiwan, a continental island, has high levels of endemism and species diversity. Most of the flora was thought to have originated from the Asian mainland during cycles of temperature oscillations. Although the land in Taiwan has never been covered by ice sheets except on the highest peaks, the tremendous temperature and climatic changes should have influenced species distributions and evolu-tion. Palynological data indicate that during the last glacial maximum, most subtropical species disappeared from the low-lands of central Taiwan. At the same time, temperate species expanded their ranges of distribution from high elevations, and lowland forests were dominated by conifers (Tsukada 1966). When the ice retreated, a reverse course of events occurred, with subtropical species recolonizing from south to north and lowland forests retreating to higher elevations. The current geo-graphic distribution of living organisms is the result of both pre-sent and past ecological and historical factors.

We indicated in a previous article that the southeastern part of Taiwan could have been a potential refugium in the last glacial maximum (Huang et al. 2002), according to chlo-roplastic DNA sequences. In addition to higher haplotype and nucleotide diversities, synapomorphic haplotypes were found on the eastern side of Taiwan (Huang et al. 2002). Ac-companied by published palynological records of the last gla-ciation (Tsukada 1966), these observations favor the concept that regards the southeastern part of Taiwan as a potential refugium during the last glaciation.

No climatic/historic or pollen record data are available for the southern part of Taiwan to support this observation. Is there any other parameter that can be used for predicting the potential refugium in the last glaciation? Fortunately, the de-gree of average population differentiation (FST) of each popu-lation in comparison with the remaining popupopu-lations (i.e., genetic divergence) can be used to examine the consequences of historical and contemporary geographical population sub-division on evolutionary processes (Johnson et al. 2000) and is important for reconstructing phylogeographical histories that have evolved during pre- and postcolonization events (Grant and Grant 1997). It was found that in the common ivy (Hedera sp.) of Europe, differentiation of each population from the remaining ones revealed a latitudinal pattern, with populations from the south being significantly more differen-tiated from the pooled remaining populations than were the central or northern populations (Grivet and Petit 2002). Thus, population divergence or genetic differentiation can be a useful criterion for locating regions of glacial refugia. Petit et al. (2003) tested the hypothesis that glacial refuge areas harbor a large fraction of intraspecific diversity. They con-cluded that plant populations in refuge areas have high ge-netic divergence and uniqueness rather than a high number of haplotypes. The concept of genetically highly divergent Table 1

Populations and Haplotypes of Quercus glauca in Taiwan and Japan

Population Location

Elevation (m)

Sample size

(no. alleles) Haplotypes (no. alleles) Taiwan: 1. Yangmingshan 121.51°E, 25.18°N 400 10 A (4), B (1), C (2), D (1), E (2) 2. Wulai 121.54°E, 24.86°N 450 11 A (3), B (1), C (3), D (2), F (1), B3 (1) 3. Baling 121.37°E, 24.68°N 600 11 A (4), C (3), F (1), H (2), I (1) 4. Chilan 121.60°E, 24.58°N 400 9 A (1), B (2), C (5), D (1) 5. Chingchuan 121.09°E, 24.58°N 450 9 A (3), B (1), C (2), F (1), L (2) 6. Kukuan 121.00°E, 24.21°N 700 7 B (3), C (1), D (2), C1 (1) 7. Huisun 121.03°E, 24.09°N 650 8 A (2), B (3), M (2), I1 (1) 8. Wushe 121.13°E, 24.02°N 850 12 A (3) B (2), D (1), K (2), L (1), B5 (1), B2 (1), B6 (1) 9. Liukuei 120.64°E, 23.00°N 550 10 A (4), C (1), D (2), E (1), J (1), B1 (1) 10. Chinshuiying 120.73°E, 22.42°N 1000 10 A (5), B (1), C (2), F (2) 11. Suao 121.84°E, 24.60°N 450 10 A (2), B (4), C (1), D (1), E1 (1), O2 (1) 12. Taroko 121.62°E, 24.15°N 500 9 A (3), B (3), C (1), D (1), E (1) 13. Lungchien 121.43°E, 24.00°N 400 10 A (5), B (1), C (3), E (1) 14. Hungyeh 121.04°E, 22.90°N 400 9 A (3), B (2), C (1), D (1), F (1), B4 (1) 15. Juisui 121.42°E, 23.50°N 200 9 A (3), B (3), C (2), H (1) 16. Fuli 121.27°E, 23.08°N 500 9 A (4), B (2), C (1), N (2) 17. Taimali 121.01°E, 22.67°N 100 3 A (1), E (1), J (1) Japan: 18. Kyushu 130.54°E, 33.67°N 8 B (1), G (4), H1 (1), B7 (1), G1 (1) 19. Kagoshima 130.43°E, 31.55°N 4 I (1), O1 (1), B8 (1), B9 (1)

populations existing in regions of glacial refugia is also supported in the common ash (Fraxinus excelsior; Heuertz et al. 2004).

It is worth reexamining the potential refuge site reported using chloroplastic DNA (Huang et al. 2002). Also, this site needs to be tested with independent nuclear markers. In this article, we address whether the data obtained from nuclear genes are able to corroborate previous results based on cyto-plasmic loci and to further characterize the population struc-ture and evolutionary history of Q. glauca in Taiwan. The results of phylogeographical analysis in this study support those obtained from independent DNA sequences from the cytoplasmic genome in Q. glauca.

Material and Methods Sampling of Plant Populations

Every stand used in this study was considered to be geneti-cally original, as low economic interest strongly suggests their indigenous status. Most of the populations used in this study are the same as those used in a previous study of cyto-plasmic data (Huang et al. 2002) but are not necessarily from the same individuals. On average, the number of alleles used in this study was 2.5 times greater than that used in the chloroplast DNA study.

In total, 19 populations, including 17 from Taiwan and two from Japan, were collected (table 1; fig. 1). In general,

six individuals or fewer represented each population. Leaves from fresh or silica-gel-dried collections of each individual were deposited in a freezer at
70°C.

Polymerase Chain Reaction and Sequencing DNA was extracted from leaves following modification of a standard protocol (Murray and Thompson 1980). The primers used by Olsen and Schaal (1999) for glyceraldehyde-3-phosphate dehydrogenase (G3pdh) were tested to screen the same gene in Quercus glauca. A polymerase chain reac-tion (PCR) was performed with newly designed forward (TGG AAT TGT TGA GGG TCT CAT; denoted GPD-CG-F) and reverse (TGC TGT CAC CAA TGA AGT CG; denoted GPD-CG-R) primers so that they best fit our sequence. The PCR solution was prepared as follows: 500 mM KCl, 15 mM MgCl2, 0.01% gelatin, 100 mM Tris-HCl (pH 8.3), 250 mM of each dNTP, 2 mM of each primer, 0.04–20 ng of tem-plate DNA, 1 mg RNase, 0.5 units of Taq polymerase (Amer-sham Pharmacia Biotech, Taipei, Taiwan), and double-distilled water to a final volume of 10 mL. Amplifications were per-formed with an initial denaturing of 2 min at 95°C, followed by 32 cycles of 1 min at 95°C, 90 s at 62°C, and 2 min at 72°C, ending with a 9-min extension at 72°C. We amplified a G3pdh gene sequence of 898 base pairs (bp) (including a par-tial sequence of 59-UTR, exon A [79–176 bp], intron a, exon B [594–736 bp], intron b, and a partial sequence of exon C

Fig. 1 Map of East Asia showing the location of Taiwan and the populations sampled for Quercus glauca. Only elevations above 2000 m of the mountain areas are shown. CMR ¼ Central Mountain Ridge; HR ¼ Hsueshan Range.

865

[ca. 865 bp]) that covers the region highly homologous for the G3pdh gene in cassava (Olsen and Schaal 1999).

Nucleotide sequences were determined by direct sequenc-ing of the purified PCR products on an ABI 3300 genetic an-alyzer with BigDye terminator cycle sequencing reagents (Applied Biosystems, Foster City, CA). This was applied to se-quences of homozygotes or sese-quences containing one polymor-phic site. For cloning of PCR products from heterozygotes, DNA fragments were ligated into a vector (pGEM-T Easy, Promega, Madison, WI) with Ampr. The constructs were transformed to competent cells (ECOS 101, Yeastern Biotech, Taipei, Taiwan) with 45 s of heat shock at 42°C. Bacterial cultures containing X-Gal and IPTG were spread onto LB broth medium agar plates and incubated at 37°C overnight. Three clones, however, were sequenced for each individual when sequences contained two or more polymorphic sites and/or an unpaired indel (indel of different length). The two sequences of a heterozygote were separated by comparing the sequences of the PCR product and cloned sequence. Fortu-nately, no sequence containing more than one indel was found in the length of G3pdh used in this study. Because the Taq po-lymerase error was estimated to be as high as 0.1% (Okuyama et al. 2005) or even higher, the singleton was removed after

the cloned sequences were compared with the sequences from direct sequencing in the forward and/or reverse directions.

Two lines of evidence suggest that the present G3pdh hap-lotypes were derived from a single gene. First, haplotype de-termination was made using direct sequencing of PCR products, and no more than two haplotypes were identified per individual (as would be expected for a diploid nuclear genome). Second, for one trial homozygous individual, the PCR product was cloned and then sequenced; the DNA se-quences of multiple clones from this individual were identical.

Sequence Analysis

G3pdh sequences were aligned by eye. Haplotypes and their relationship were determined using the TCS program (Templeton et al. 1992) by considering gaps as missing data. Nucleotide diversity and haplotype diversity were carried out using the DnaSP program (Rozas and Rozas 1999).

Analysis of the Population Substructure

Measures of diversity and population differentiation, GST, were analyzed using the Hapstep program (Pons and Petit 1996), which employed Nei’s (1977) approach to the case of haplotypes under Wright’s model of population structure. Table 2

Haplotypes and Accession Numbers of Quercus glauca according to Mutations of the G3pdh Gene

Mutation site of haplotype

0001222222333333333344444444445555555555556666678888 Haplotype and accession no. 4672003558000134577803356667890023344477890267850245 9467385266189723578122426788312302814668531529274848 B. AY780452 TTTCATAATCAATTGTATGGGTTGTTTAATAATTACCGAGCATTACTGAATT A. AY780451 ---T---C. AY780453 --G-C---C---A---G---D. AY780454 C---E. AY780455 ---A-A---F. AY780456 ---C---T---T---G. AY780457 ---G---H. AY780458 -A-T--T---C---A---G---A----C--T---I. AY780459 ---G---C---A---J. AY780460 ---C-G---A-C---A---K. AY780461 ---T---G----G-C---C-C-A--A---C---AG-T--C--G---C-L. AY780462 ----G---A---C---G---A---M. AY780463 ---G---G---N. AY780464 ---T---G---C---C-C-A---C---AG-T--C--G---B1. AY780465 ---A---B2. AY780466 --G--C---G---A---B3. AY780467 --G--C---C---A---B4. AY780468 --G--C---C---A---AG---B5. AY780469 ---T---C---C-C-A---C---AG-T---G---C-B6. AY780470 ---T---G---C---C-C-A---C---AG-T---G---C-B7. AY780471 -A-T--T---C---G--B8. AY780472 -A-T--T---C---C---C---B9. AY780473 ---A---C---T---C1. AY780474 --G-C---C-C---A---G---E1. AY780475 C---A--A---H1. AY780476 ---C---A---G---A----C--T---G1. AY780477 ---G---T---G---I1. AY780478 C---C-G---C---A---O1. AY780479 ---C---C O2. AY780480

---C---The conventional genetic distance, FST (Wright 1931), ac-cording to DNA sequences for population subdivision, was estimated using the Arlequin program (Schneider et al. 2000), which provides a matrix of pairwise FST values be-tween populations. The Mantel test, implanted in the PAS-SAGE program (Rosenberg 2001), was performed to test whether genetic distances correlated with geographical dis-tances. In addition, the level of divergence for each popula-tion from the remaining populapopula-tions was calculated as mean values of pairwise FST for each population against the re-maining populations. The contribution to total expected het-erozygosity (CT) of each population was calculated using the Contrib software (Petit et al. 1998). This contribution is split into two components: one due to the diversity of the popula-tion (CS) and the other due to its differentiapopula-tion from the re-maining populations (CD). In this case, CS and CD for each population were computed relative to the mean population diversities and mean population differentiation, respectively. These led to either positive or negative contributions to pop-ulations (Petit et al. 1998).

Sequences of cpDNA and mtDNA of the populations from Taiwan were obtained from published data (Huang et al. 2002; Lin et al. 2003) and were used to estimate the popula-tion differentiapopula-tion, FST, using the Arlequin program. Con-ventional FST values and the level of divergence for each population from the remaining populations were obtained as described above.

A test involving three matrices, including two matrices of mean FST values and a matrix of geographic latitude, was performed using ZT, a software tool for simple and partial Mantel tests (Bonnet and Van de Peer 2002). In these tests, we arranged the corresponding mean FST values in the matrices according to latitude.

Results

Sequence Analysis of the G3pdh Gene

In total, 96 individuals were sampled, and 86% of them were heterozygous. However, only 168 alleles of the total 192 G3pdh sequences in the sample could be identified un-ambiguously, and these could be assigned to one of only 30 haplotypes. Thirty haplotypes were detected from 52 substi-tution sites (GenBank accession numbers are given in table 2). Haplotypes, assigned from A to N, were shared by at least by two alleles, while the others were singletons. The ge-nealogical relationship of these haplotypes is shown in figure 2, and the numbers of alleles containing such haplotypes are shown in table 1.

Distribution Patterns of Haplotypes

Mapping the geography onto the haplotype network pro-duced a complicated pattern (fig. 3). Among the 22 haplo-types found in Taiwan, haplohaplo-types A–C accounted for 68.6%

Fig. 2 Relationships of haplotypes of the nuclear gene marker, G3pdh (unrooted), reconstructed by the computer program TCS, version 1.04. Numbers of alleles are in parentheses. Haplotype B, indicated by a circle, is located in the center, from which the Taiwanese and Japanese haplotypes originated. The other haplotypes, indicated by squares of various sizes, reflect the number of alleles found. B4 and B2, which are connected to two internal branches, respectively, probably resulted from homoplasy or recombination. A dot represents a mutation for which a haplotype has not been found within the sample. Numbers along the lines represent the base positions of the mutations that separate the haplotypes.

867

(107) of 156 alleles. These three haplotypes were spread over Taiwan’s population (table 1). Haplotype C occurred mainly in northern Taiwan, while haplotype B occupied a high pro-portion of populations of the northern part of central Tai-wan. Haplotype B is a possible ancestral haplotype and was not found in the populations of Baling (population 3), Liu-kuei (9), or Taimali (17). Ten singleton haplotypes were re-stricted to only one allele, while haplotypes D–F and H–J were found in widely separated populations.

Many lineages were found on both sides of the Central Mountain Ridge (CMR) and had sporadic distributions; only a few haplotype lineages clearly showed a continuous geo-graphic distribution (fig. 3). The lineage containing B1 and E occurred on the eastern side of the CMR, whereas the lineage containing B2, B3, and C1 was located in the center and north on the western side of the CMR. Haplotype D was found in many locations except southern and southeastern Taiwan.

The Japanese haplotypes were all traced to ancestral hap-lotype B, and some were closely related to those in Taiwan (fig. 2). For example, haplotype O1 of Japan is related to O2 of Suao (11) of northeastern Taiwan, and haplotype H1 of Japan was derived from type H, which occurs in Juisui (15), eastern Taiwan, while haplotype I was found in both Japan and northern Taiwan. This indicates that a relic connection of Quercus glauca existed between Japan and northern and northeastern Taiwan.

Analysis of the Population Genetic Structure Genetic distances among populations did not correlate with geographical distances because the Mantel test was not significant (data not shown). For the G3pdh gene, the value of GST was 0.005, intrapopulational diversity (HS) was 0:825 6 0:0184, and total diversity (HT) was 0:829 6 0:0158. Thus, the spatial genetic pattern was very weak.

Fig. 3 Map showing the geographical distribution of G3pdh haplotypes and the proportion of each haplotype in each population of Quercus glauca in Taiwan. Twenty-two haplotypes were detected for the G3pdh fragments, and only haplotypes occurring in two or more alleles are presented. Ten haplotypes were singleton, and these are indicated in black.

On the western side of the CMR, peak nucleotide diversity was found in the Baling (3; 0.0070) and Wushe (8; 0.0097) populations (table 3), while on the eastern side, the nucleo-tide diversity gradually increased toward the south and peaked in the Fuli (16; 0.0070) and Taimali (17; 0.0060) populations. If the total diversity consists of two components (i.e., genetic differentiation and genetic diversity), then the populations of Chingchuan (5), Kukuan (6), and Huisun (7) contributed most of the differentiation component to the to-tal diversity (fig. 4A). The Taimali (17) and Wushe (8) popu-lations contributed most to the diversity component of the total diversity (fig. 4A).

Pairwise F Statistics according to Haplotype Frequency and Population Divergence

In general, genetic differentiation between pairwise popu-lation comparisons indicated that the genetic distance be-tween Taiwanese and Japanese populations was greater than that between any of the Taiwanese populations (data not shown). On the western side of the CMR, a major peak of average FSTwas found in the Kukuan (6; 0.06393) and Hui-sun (7; 0.05991) populations (fig. 5A). The peak decreased to low FSTvalues for the populations of Wushe (8; 0.02935) and Wulai (2; 0.02611) to the south and north, respectively. On the eastern side of the CMR, the major peak of FSTwas found in southeastern Taiwan at Taimali (17; 0.0551), Chin-shuiying (10; 0.04915), and Hungyeh (14; 0.04829), and an-other peak was found at Taroko (12; 0.0526) of east-central Taiwan, which coincides with the latitude of the major peak on the western side of the CMR.

Pairwise F Statistics and Population Divergence Revealed by cpDNA and mtDNA Sequences

The results showed that a major peak of average FST according to cpDNA data was found in the Huisun

popula-tion (7; 0.80) on the western side of the CMR (fig. 5B). The peak decreased to low FST values for the populations at Wushe (8; 0.47) and Kukuan (6; 0.14) to the south and north, respectively. A second peak was found in the popula-tion at Laiye (18; 0.62), close to the populapopula-tion at Chinshui-ying (10). On the eastern side of the CMR, a major peak of FSTwas found at Lungchien (13; 0.98), a population close to Taroko (12) and conciding with the latitude of the popula-tion at Huisun (7). A small minor peak was found at Taimali (17; 0.29) that coincides with the latitude of the peak of Laiye on the western side of the CMR.

According to mtDNA data, Liukuei (9; 0.800) and Mutan (0.815) in the south had the highest FSTvalues (fig. 5C), but there was no detectable peak in the region covering the Ku-kuan and Huisun populations. On the eastern side of the CMR, a major peak of FST was found at Lungchien (13; 0.867), as well as a smaller peak to the south at Yuli (0.577) and Taimali (17; 0.468).

Correlations between the two matrices of mean FSTderived from the two different DNA marker sequence data sets against latitude were tested using the partial Mantel test. Pairwise comparisons, i.e., cpDNA-mtDNA (R ¼ 0:895111, P ¼ 0:02778), cpDNA-G3pdh (R ¼ 0:483809, P ¼ 0:036411), and mtDNA-G3pdh (R ¼ 0:43721, P ¼ 0:040278), showed that average FSTvalues for each population on the eastern side of the CMR were significantly correlated. However, correlations between mean FST values for each population of any two genome markers on the western side of the CMR showed no correla-tions. Subsequently, only the relationship between cpDNA and G3pdh was further tested after moving populations of cpDNA northward slightly so that population 7 of cpDNA met popula-tion 6 (Kukuan) of G3pdh. The Mantel test showed that the cpDNA-G3pdh pair was now significantly correlated with lati-tude (R ¼ 0:486740, P ¼ 0:027797). Populations at Tayuan-shan (very close to population 4), Huisun (7), Lungchien (13), Table 3

Estimates of Haplotype Diversity (h) and Nucleotide Diversity (p) for Populations of Quercus glauca according to Mutations of the G3pdh Gene Population (no.) No. sequences No. variable sites No. haplotypes Haplotype diversity (h) Nucleotide diversity (p) Yangmingshan (1) 10 9 5 0.78 6 0.10 0.0034 6 0.0008 Wulai (2) 11 9 6 0.87 6 0.07 0.0041 6 0.0006 Baling (3) 11 18 5 0.82 6 0.08 0.0070 6 0.0013 Chilan (4) 9 7 4 0.69 6 0.15 0.0037 6 0.0006 Chingchuan (5) 9 13 5 0.86 6 0.10 0.0055 6 0.0011 Kukuan (6) 7 7 4 0.81 6 0.13 0.0036 6 0.0011 Huisun (7) 8 8 4 0.82 6 0.13 0.0029 6 0.0011 Wushe (8) 12 24 8 0.92 6 0.06 0.0097 6 0.0016 Taroko (12) 9 9 5 0.83 6 0.10 0.0026 6 0.0009 Liukuei (9) 10 13 6 0.84 6 0.10 0.0039 6 0.0010 Chingshuiying (10) 10 8 4 0.73 6 0.12 0.0033 6 0.0001 Suao (11) 10 10 6 0.84 6 0.10 0.0026 6 0.0009 Lungchien (13) 10 8 4 0.71 6 0.12 0.0037 6 0.0008 Hungyeh (14) 9 11 6 0.89 6 0.91 0.0038 6 0.0011 Juisui (15) 9 14 4 0.81 6 0.09 0.0049 6 0.0016 Fuli (16) 9 17 4 0.78 6 0.11 0.0070 6 0.0022 Taimali (17) 3 8 3 1.00 6 0.27 0.0060 6 0.0020 Taiwan 157 43 23 0.82 6 0.02 0.0049 6 0.0005 Japan 12 21 9 0.91 6 0.08 0.0055 6 0.0011 Total 169 52 30 0.84 6 0.02 0.0050 6 0.0004 869

and Laiye (very close to population 17) contributed most of the differentiation component to the total diversity of cpDNA (fig. 4B) because of their unique haplotypes.

Discussion

Genetic Spatial Structure of Populations

In a previous study, by sequencing three cpDNA intergenic spacer fragments, it was found that the level of differentia-tion among populadifferentia-tions of Quercus glauca was relatively high (GST¼ 0:612; Huang et al. 2002). In this study, we

found that some haplotypes were distributed in widely sepa-rated locations or on both sides of the CMR. This pattern is consistent with the extremely low GSTvalue. The shared hap-lotypes between distant areas could be the result of either recent interpopulational gene exchange or shared ancestral polymorphisms (Hare 2001).

The ancestral haplotypes, e.g., haplotype B, were widely spread out, while the derived haplotypes (fig. 2) also oc-curred in most of the populations; thus, a clear migration

route cannot be elucidated. So the temporal resolution of-fered by genealogies, such as chloroplast DNA markers, was unavailable through nuclear genes. The gene tree for the ge-nealogical relationships of G3pdh among these nuclear hap-lotypes reveals nothing about the history of population divergence almost certainly because they predate the history of population divergence. However, the power of nuclear

Fig. 5 Plot of the mean FSTvalues of each population compared

with every other population against population latitude in Quercus glauca using G3pdh sequence (A), chloroplastic DNA sequence (B; data taken from Huang et al. 2002), and mitochondrial DNA data (C; data taken from Lin et al. 2003). The solid line indicates populations on the western side of the Central Mountain Ridge (CMR); the dashed line indicates populations on the eastern side of the CMR. Arabic numbers refer to populations having high FSTvalues. Population codes

are labeled according to table 1, except 18, which is designated as Laiye. Pinglin (121.70°E, 24.92°N), Wutai (120.72°E, 22.74°N), Mutan (120.78°E, 22.13°N), and Yuli (121.25°E, 23.31°N) are indicated.

Fig. 4 Contribution to the total diversity (CT) of each population of Quercus glauca using G3pdh haplotypes (A) and chloroplastic sequences (B). Gray bars and black bars represent contributions of differentiation (CD) and diversity (CS), respectively. A, Populations 5–7, representing Chingchuan, Kukuan, and Huisun, respectively, contributed most to the differentiation component. See table 1 for population numbers. B, Populations 4, 7, 13, and 18, representing Chilan, Huisun, Lungchien, and Laiye, respectively, contributed most to the differentiation component. See table 1 for population numbers, except 18, which is designated as Laiye (120.72°E, 22.52°N).

genes in providing resolution of historical events, i.e., pre-dicting potential refugia, can still be appreciated.

Differentiation between populations (GST) can be used to calculate the pollen-seed flow ratio, which was as high as 291 (estimated by the equation f23½ð1=GSTCÞ
1

½ð1=GSTNÞ
1g=½1
ð1=GSTCÞ, cited in Oddou-Muratorio

et al. 2001, where C is the chloroplast and N is the nucleus), and was within the range of species of Quercus of 190–500 (Squirrell et al. 2001). The lack of clear spatial genetic pat-terns is far more likely to reflect incomplete lineage sorting, with nuclear genes having a greater expected coalescence time than differential gene flow by pollen.

Conformance of the Trend of Population Divergence of Three Genes (Loci)

The most significant discovery in this study is the confor-mance of different genes/loci having similar trends of genetic divergence for each population from the remaining popula-tions. In terms of the mean values of pairwise FST for each population against the remaining populations, the most di-vergent populations were situated in two places. The first re-gion was found in Kukuan (6) and Huisun (7) on the western side of the CMR and Taroko (12) at a comparable latitude on the eastern side of the CMR (fig. 5A). This region is lo-cated in the northern part of central Taiwan between 24.00°N and 24.25°N and is in proximity to the most geneti-cally divergent areas determined for Trochodendron aral-ioides (Huang et al. 2004) and Cunninghamia konishii (Chung et al. 2004). Although the FSTvalues for nDNA were fairly small (around 6%), a peak was evident. In fact, the re-gion of the Huisun population has major peaks of FST for both nDNA and cpDNA. mtDNA data also showed that Lungchien (13) in the northern part of central Taiwan had high degrees of mean FST. An exception was found on the western side of the CMR, where a high FST value from mtDNA analysis was not detected in the northern part of central Taiwan. A second peak of FSTin nDNA was found for the Chinshuiying (10), Hungyeh (14), and Taimali (17) sites, which are similar to sites with peak FST values in cpDNA of Chinshuiying (10), Hungyeh, Taimali, and Laiye (close to population 17) (fig. 5B). This region is located in southeastern Taiwan between 22.40°N and 22.90°N; mtDNA data also showed that the Taimali (17) and Mutan popula-tions of southern Taiwan had high values of mean FST, rein-forcing the impression of similar regions having significant differentiation from other regions. High mean FSTof the popu-lation Liukuei (9) is an exception that could not be detected in cpDNA and G3pdh. The profile on the western side might im-ply that several smaller shelters existed in southern Taiwan

during the last glaciation. The correlations between mean FST values for each population of cpDNA and G3pdh against lati-tude were tested using the partial Mantel test, and the results support the average FSTprofiles being correlated well but not being the result of chance alone.

The average FSTvalues for each population of mtDNA are only partially in accord with those of cpDNA, and values of nDNA are unclear at present. We found that the gene genea-logical tree of mtDNA was partially congruent with the cpDNA tree (Lin et al. 2003). The average FSTvalues of mtDNA for each population against latitude differed from those of cpDNA in Machilus kusanoi Hay. (S. Y. Hwang, unpublished data), in-dicating that mtDNA has a different evolutional history from cpDNA.

The contribution of a differentiation component to total di-versity (CT) of the G3pdh gene and chloroplastic gene of each population also supports the hypothesis that the northern part of central Taiwan contributed most of the differentiation com-ponent to the total diversity of Q. glauca. This was due to the specific rare alleles in each population, which resulted in high contributions to differentiation. The contribution of the differ-entiation component of the G3pdh gene of each population to CT was undetectable in the southeastern part. In the chloro-plastic gene, however, the south (Laiye; close to population 17) contributed most to the differentiation component.

It is interesting to note that the Wushe (8) population has a very low average FSTamong all populations but the highest nucleotide diversity in both nDNA and cpDNA, indicating that the high diversity may have been derived from a mixing of different colonization routes. Wushe (8) also has a large number of derived haplotypes (fig. 2), indicating a mixed na-ture of its composition. Yangmingshan (1), at the northern tip of the island, also had the highest haplotype richness of cpDNA and a high average FSTof nDNA and could also have been a cryptic shelter.

In this article, we predicted potential refugia using high av-erage FSTestimates. This idea is based on a study of numer-ous publications on European plant phylogeography and on a theoretical study of island biogeography. It may be that these cases of European plants are unique and not necessarily applicable to our study. This cannot be resolved until fossil pollen data are available in the future.

Acknowledgments

We thank Dr. P.-F. Lee, associate professor in the Institute of Ecology and Evolutionary Biology, National Taiwan Univer-sity, for invaluable assistance with the preparation of figure 1. This investigation was funded by the National Science Coun-cil (grant NSC93-2313-B-002-037), Executive Yuan, Taiwan. Literature Cited

Avise JC 2000 Phylogeography: the history and formation of species. Harvard University Press, Cambridge, MA.

Boggs S, WC Wang, FS Lewis, JC Chen 1979 Sediment properties and water characteristics of the Taiwan shelf and slope. Acta Oceanogr Taiwan 10:10–49.

Bonnet E, Y Van de Peer 2002 ZT: a software tool for simple and partial Mantel tests. J Stat Software 7:1–12.

Chung JD, TP Lin, YC Tan, MY Lin, SY Hwang 2004 Genetic diversity and biogeography of Cunninghamia konishii (Cupressa-ceae), an island species in Taiwan: a comparison with Cunning-hamia lanceolata, a mainland species in China. Mol Phylogenet Evol 33:791–801.

Comes HP, JW Kadereit 1998 The effect of Quaternary climatic changes on plant distribution and evolution. Trends Plant Sci 3:432–438.

871

Ennos RA 1994 Estimating the relative rates of pollen and seed migration among plant populations. Heredity 72:250–259. Grant PR, BR Grant 1997 Genetics and origin of bird species. Proc

Natl Acad Sci USA 94:7768–7775.

Grivet D, RJ Petit 2002 Phylogeography of the common ivy (Hedera sp.) in Europe: genetic differentiation through space and time. Mol Ecol 11:1351–1362.

Hare MP 2001 Prospects for nuclear gene phylogeography. Trends Ecol Evol 16:700–706.

Heuertz M, JF Hausman, OJ Hardy, GG Vendramin, N Frascaria-Lacoste, X Vekemans 2004 Nuclear microsatellites reveal con-trasting patterns of genetic structure between western and southern European populations of the common ash (Fraxinus excelsior L.). Evolution 58:976–988.

Huang SF, SY Hwang, JC Wang, TP Lin 2004 Phylogeography of Trochodendron aralioides (Trochodendraceae) in Taiwan and its adjacent areas. J Biogeogr 31:1251–1259.

Huang SSF, SY Hwang, TP Lin 2002 Spatial pattern of chloroplast DNA variation of Cyclobalanopsis glauca in Taiwan and East Asia. Mol Ecol 11:2349–2358.

Johnson KP, FR Adler, JL Cherry 2000 Genetic and phylogenetic consequences of island biogeography. Evolution 54:387–396. Kimura M 1996 Quaternary paleogeography of the Ryukyu arc.

J Geogr 105:259–285. (In Japanese with English abstract.) Lin TP, WZ Chuang, SSF Huang, SY Hwang 2003 Evidence for the

existence of some dissociation in an otherwise strong linkage disequilibrium between mitochondrial and chloroplastic genomes in Cyclobalanopsis glauca. Mol Ecol 12:2661–2668.

Luo Y, ZK Zhou 2001 Phytogeography of Quercus subg. Cyclo-balanopsis. Acta Bot Yunnanica 23:1–16. (In Chinese with English summary.)

Manos PS, Z-K Zhou, CH Cannon 2001 Systematics of Fagaceae: phylogenetic tests of reproductive trait evolution. Int J Plant Sci 162:1361–1379.

McCauley DE 1995 The use of chloroplast DNA polymorphism in studies of gene flow in plants. Trends Ecol Evol 10:198–202. Moore WS 1995 Inferring phylogenies from mtDNA variation:

mitochondrial-gene trees versus nuclear-gene trees. Evolution 49: 718–726.

Murray MG, WF Thompson 1980 Rapid isolation of high molecular weight plant DNA. Nucleic Acids Res 8:4321–4325.

Nei M 1977 F-statistics and analysis of gene diversity in subdivided populations. Ann Hum Genet 41:225–233.

Newton AC, TR Allnutt, ACM Gillies, AJ Lowe, RA Ennos 1999 Molecular phylogeography, intraspecific variation and the conser-vation of tree species. Trends Ecol Evol 14:140–145.

Oddou-Muratorio S, RJ Petit, B Le Guerroue, D Guesnet, B Demesure 2001 Pollen-versus seed-mediated gene flow in a scat-tered forest tree species. Evolution 55:1123–1135.

Okuyama Y, N Fuji, M Wakabayashi, A Kawakita, M Ito, M Watanabe, N Murakami, M Kato 2005 Nonuniform concerted evolution and chloroplast capture: heterogeneity of observed introgression patterns in three molecular data partition phylogenies of Asian Mitella (Saxifragaceae). Mol Biol Evol 22:285–296. Olsen KM, BA Schaal 1999 Evidence on the origin of cassava:

phylogeography of Manihot esculenta. Proc Natl Acad Sci USA 96: 5586–5591.

Petit RJ, I Aguinagalde, JL de Beaulieu, C Bittkau, S Brewer, R Cheddadi, R Ennos, et al 2003 Glacial refugia: hotspots but not melting pots of genetic diversity. Science 300:1563–1565. Petit RJ, A El Mousadik, O Pons 1998 Identifying populations for

conservation on the basis of genetic marker. Conserv Biol 12:844–855. Pons O, RJ Petit 1996 Measuring and testing genetic differentiation with ordered versus unordered alleles. Genetics 144:1237–1245. Rosenberg MS 2001 PASSAGE: attern analysis, spatial statistics, and

geographic exegesis. Version 1.0. Department of Biology, Arizona State University, Tempe.

Rowe KC, EJ Heske, PW Brown, KN Paige 2004 Surviving the ice: northern refugia and postglacial colonization. Proc Natl Acad Sci USA 101:10355–10359.

Rozas J, R Rozas 1999 DnaSP, version 3: an integrated program for molecular population genetics and molecular evolution analysis. Bioinformatics 15:174–175.

Schaal BA, DA Hayworth, KM Olsen, JT Rauscher, WA Smith 1998 Phylogeographic studies in plants: problems and prospects. Mol Ecol 7:465–474.

Schneider S, D Roessli, L Excoffier 2000 Arlequin version 2.000: a soft-ware for population genetic data analysis. Genetics and Biometry Laboratory, University of Geneva.

Squirrell J, PM Hollingsworth, RM Bateman, JH Dickson, MH Light, M MacConaill, MC Tebbitt 2001 Partitioning and diversity of nuclear and organelle markers in native and introduced popula-tions of Epipactis helleborine (Orchidaceae). Am J Bot 88: 1409–1418.

Templeton AR, KA Crandall, CF Sing 1992 A cladistic analysis of phenotypic associations with haplotypes inferred from restriction endonuclease mapping and DNA sequence data. III. Cladogram estimation. Genetics 132:619–633.

Tsukada M 1966 Late Pleistocene vegetation and climate in Taiwan (Formosa). Proc Natl Acad Sci USA 55:543–548.

Wright S 1931 Evolution in Mendelian populations. Genetics 16: 97–159.