Hostname: page-component-89b8bd64d-5bvrz Total loading time: 0 Render date: 2026-05-06T20:36:09.003Z Has data issue: false hasContentIssue false
  • English
  • Français

Reconsideration for conservation units of wild Primula sieboldii in Japan based on adaptive diversity and molecular genetic diversity

Published online by Cambridge University Press:  30 July 2009

YASUKO YOSHIDA ,
MASANORI HONJO ,
NAOKO KITAMOTO  and
RYO OHSAWA
YASUKO YOSHIDA
Affiliation:
Graduate School of Life and Environmental Sciences, University of Tsukuba, 1-1-1 Tennoudai, Tsukuba, Ibaraki 305-8572, Japan
MASANORI HONJO
Affiliation:
National Agricultural Research Center for Tohoku Region, 4 Akahira, Shimo-kuriyagawa, Morioka, Iwate 020-0198, Japan
NAOKO KITAMOTO
Affiliation:
Graduate School of Life and Environmental Sciences, University of Tsukuba, 1-1-1 Tennoudai, Tsukuba, Ibaraki 305-8572, Japan
RYO OHSAWA*
Affiliation:
Graduate School of Life and Environmental Sciences, University of Tsukuba, 1-1-1 Tennoudai, Tsukuba, Ibaraki 305-8572, Japan
*
*Corresponding author: Graduate School of Life and Environmental Sciences, University of Tsukuba, 1-1-1 Tennoudai, Tsukuba, Ibaraki 305-8572, Japan. Tel.: +81-298-53-6674. Fax: +81-298-53-6674. e-mail: osawaryo@sakura.cc.tsukuba.ac.jp
Rights & Permissions [Opens in a new window]

Summary

Primula sieboldii E. Morren is a perennial clonal herb that is widely distributed in Japan, but in danger of extinction in the wild. In a previous study, we revealed the genetic diversity of the species using chloroplast and nuclear DNA and used this information to define conservation units. However, we lacked information on adaptive genetic diversity, which is important for long-term survival and, thus, for the definition of conservation units. In order to identify adaptive traits that showed adaptive differentiation among populations, we studied the genetic variation in six quantitative traits within and among populations for 3 years in a common garden using 110 genets from five natural populations from three regions of Japan. The number of days to bud initiation was adaptive quantitative trait for which the degree of genetic differentiation among populations (QST) was considerably larger than that in eight microsatellite markers (FST). The relationship between this trait and environmental factors revealed that the number of days to bud initiation was negatively correlated, with the mean temperature during the growing period at each habitat. This suggests that adaptive differentiation in the delay before bud initiation was caused by selective pressure resulting from temperature differences among habitats. Our results suggest that based on adaptive diversity and neutral genetic diversity, the Saitama population represents a new conservation unit.

Information

Type
Paper
Information
Genetics Research , Volume 91 , Issue 4 , August 2009 , pp. 225 - 235
Copyright
Copyright © Cambridge University Press 2009

1. Introduction

The extent of genetic diversity that is maintained among and within remnant wild populations plays a key role in understanding the adaptive evolution of a species in response to fluctuating environments. Such knowledge can therefore support the conservation of endangered species (Frankham et al., Reference Frankham, Ballou and Briscoe2002). Similarly adaptive genetic diversity within and among local populations is particularly important when conservation procedures such as translocation of plant materials between sites or the restoration of local populations is promoted (McKay & Latta, Reference McKay and Latta2002). The term ‘Restoration’ is commonly used for the maintenance of the genetic structure, i.e. the hierarchical structure of genetic variation, in a species and many researchers used different approaches for the implication of restoration efforts in endangered species (e.g. Gordon & Rice, Reference Gordon and Rice1998). In this context, a ‘conservation unit’ is a group of populations of a species whose genetic structure differs from that of other groups of populations (i.e. differs from other conservation units).

Genetic variation is mainly evaluated using quantitative traits and molecular markers. The evaluation of quantitative traits in common-garden studies began more than 60 years ago (Clausen et al., Reference Clausen, Keck and Hiesey1940). The utilization of molecular markers in such studies has become increasingly common during the last several decades (Haig, Reference Haig1998), because it is easier to evaluate genetic variation using molecular markers. The relationships between quantitative and molecular genetic variation within and among populations have been tested by many researchers (Karhu et al., Reference Karhu, Hurme, Karjalainen, Karvonen, Kärkkäinen, Neale and Savolainen1996; Merilä & Crnokrak, Reference Merilä and Crnokrak2001; Reed & Frankham, Reference Reed and Frankham2001; Bekessy et al., Reference Bekessy, Ennos, Burgman, Newton and Ades2003; Carvajal-Rodriguez et al., Reference Carvajal-Rodriguez, Rolan-Alvarez and Caballero2005), but no significant relationship was found between the two parameters in many cases. As a result, because molecular markers do not reflect the differentiation among populations in their adaptive traits, it has become necessary to directly evaluate the genetic variation of these populations based on quantitative traits in conservation efforts (Storfer, Reference Storfer1996; Mckay et al., Reference Mckay, Bishop, Lin, Richards, Sala and Mitchell-Olds2001; Holderegger et al., Reference Holderegger, Kamm and Gugerli2006).

Population genetic structure is often quantified using Wright's F ST parameter in combination with the analogous measure for quantitative traits, Q ST (Spitze, Reference Spitze1993). When Q ST is compared directly with F ST, there are three possible patterns (Merilä & Crnokrak, Reference Merilä and Crnokrak2001; McKay & Latta, Reference McKay and Latta2002): (1) when Q ST>F ST, populations have become differentiated as a result of directional natural selection that acts on different phenotypes among the populations; (2) when Q ST=F ST, differentiation among populations has arisen mainly through genetic drift; or (3) when Q ST<F ST, stabilizing selection has favoured the maintenance of the same phenotype in different populations. Thus, by comparing Q ST with F ST, it becomes possible to identify adaptive traits that show adaptive differentiation among populations as a result of directional selection.

Understanding adaptive differentiation by comparisons of Q ST with F ST has become increasingly popular in the last 5 years (Leinonen et al., Reference Leinonen, O'Hara, Cano and Merilä2008). Despite wide recognition of the importance of evaluating both adaptive and molecular genetic diversity, few conservation units have been defined for use in conservation efforts that address both adaptive and molecular genetic variation (de Guia & Saitoh, Reference de Guia and Saitoh2007). Many studies using molecular markers have assigned conservation priority to populations that are considered to represent conservation units (e.g. Moritz, Reference Moritz1994), but Crandall et al. (Reference Crandall, Bininda-Emonds, Mace and Wayne2000) and Fraser & Bernatchez (Reference Fraser and Bernatchez2001) emphasized that conservation units should be defined by considering both adaptive and molecular genetic variation. Although increasing numbers of studies have defined conservation units based on both parameters (e.g. Denoël, Reference Denoël2007; Cano et al., Reference Cano, Mäkinen, Leinonen, Freyhof and Merilä2008), few of these studies have focused on plants (see, de Guia & Saitoh, Reference de Guia and Saitoh2007).

Primula sieboldii E. Morren is a perennial clonal herb that occupies a range of moist habitats, from the understory of deciduous forests to well-managed grasslands. The species is distributed in Japan, on the Korean Peninsula, in northern China, and in eastern Siberia (Yamazaki, Reference Yamazaki, Kunio, Takashi, David and Hideaki1993). However, overexploitation and destruction of its habitats and horticultural collection are threatening wild populations with extinction, and the species has become endangered in Japan; originally classified in the lower rank of ‘vulnerable’, the species was reclassified to ‘near threatened’ in 2007 (Iwatsuki, Reference Iwatsuki2008).

The genetic diversity of P. sieboldii has been evaluated using cpDNA (Honjo et al., Reference Honjo, Ueno, Tsumura, Washitani and Ohsawa2004) and microsatellite markers (Honjo et al., Reference Honjo, Kitamoto, Ueno, Tsumura, Washitani and Ohsawa2009) to define conservation units in Japan. These studies revealed that genetic differentiation among populations corresponded to the geographic distance between populations, and that genetic variation within populations correlated with population size. On this basis, we defined four conservation units.

In our previous research on quantitative traits, we investigated the genetic diversity in the petal shape and area of P. sieboldii in a common garden and revealed that genetic variation was mostly maintained within wild populations, i.e. low differentiation among populations (Yoshida et al., Reference Yoshida, Honjo, Kitamoto and Ohsawa2008). These traits were not under directional selection and appeared to be basically selection-neutral. However, there is less information on adaptive traits, and understanding adaptive genetic diversity is required for more precise definition of conservation units for P. sieboldii.

In the present study, we evaluated genetic variation of six quantitative traits within and among wild populations of P. sieboldii to obtain new information that would help us understand the adaptive traits and genetic structure of this species. By comparing Q ST with F ST, calculated using eight microsatellite (simple sequence repeat (SSR)) markers, we estimated the cause of genetic differentiation among the populations and identified potentially adaptive traits of P. sieboldii. In addition, we investigated the relationships between quantitative genetic variation and five ecological factors in order to detect the potential causes of any differentiation among populations. Ultimately, we attempted to use this new knowledge to redefine the conservation units for P. sieboldii by considering both adaptive and molecular genetic variation.

2. Materials and methods

(i) Plant species and sampling populations

Each genet is composed of various numbers of physiologically independent ramets, which propagate clonally by means of short rhizomes. In early spring, these rhizomes bud and subsequently flower from April to June. After July, new rhizomes for next year are generated from a mother ramet below the ground.

In 2000, we collected 110 genets of P. sieboldii from five wild populations in three regions of Japan (Table 1): Hokkaido (a lowland area in northern Japan), Saitama (a lowland area in central Japan) and Nagano (a highland area in central Japan). In the Hokkaido and Nagano 1 populations, we randomly sampled 26 and 27 genets, respectively, from a population that contained more than 100 genets. In the Nagano 2 and Nagano 3 populations, we sampled nearly all genets (26 and 13, respectively). For the Saitama population, we used 18 genets that had been grown from seeds obtained from a natural population in Saitama region in a previous experiment (Washitani & Kabaya, Reference Washitani and Kabaya1988). These 18 genets were from several different mothers and originated from the region surrounding Saitama population from the results of both an assignment test based on nuclear DNA markers and cpDNA (Honjo et al., Reference Honjo, Ueno, Tsumura, Washitani and Ohsawa2004, Reference Honjo, Ueno, Tsumura, Handa, Washitani and Ohsawa2008; Honjo, Reference Honjo2005).

Table 1.

Locations of the five Japanese populations of P. sieboldii, number of genets (total and number sampled) and average monthly temperatures at each site

Average temperatures represent means from 1979 to 2000 obtained from the Japan Meteorological Agency Web site (http://www.jma.go.jp/jma/indexe.html).

After these samples were obtained, the ramets were acclimatized in Tsukuba (a lowland area of central Japan) for 3 years. Each genet used in the present study had a unique SSR genotype (Honjo et al., Reference Honjo, Kitamoto, Ueno, Tsumura, Washitani and Ohsawa2009).

In this study, we selected three regions (Hokkaido, Saitama and Nagano) with sample sites distributed over a wide range of altitudes (from 10 to 1120 m asl) and latitudes (from 35°50′N to 42°31′N) (Table 1). We selected three Nagano populations to investigate whether adaptive differentiation has occurred on a small geographical scale.

Meteorological data for each habitat were obtained from the Japan Meteorological Agency Web site (http://www.jma.go.jp/jma/indexe.html) (Table 1). Mean temperatures from February to June (the growing period of P. sieboldii) were obtained using data from 1979 to 2000. Because the Nagano 2 and Nagano 3 sites are so close to each other, they had the same mean temperature.

(ii) Common-garden experiment

From 2003 to 2005, we grew 425 to 450 ramets from the 110 genets in a greenhouse at the University of Tsukuba (140°10′E, 36°11′N) to evaluate genetic variation in quantitative traits. Clonally propagated shoots were planted into 12-cm pots at a depth of 1 cm below the soil surface in Metro-Mix 350 (Hyponex Japan, Corp.) in January. The pots were assigned randomly to their positions in the greenhouse. Four ramets from each genet were used as replications, when available; otherwise, we used two or three ramets.

We measured six quantitative traits: budding date, first flowering date, number of flowers, number of leaves, and the lengths of the peduncle at the start and end of flowering. Because the heights of the plants surrounding P. sieboldii in the habitats increased greatly from early spring to early summer, and the environmental conditions of the habitats between the start and end of the flowering period also changed greatly, we measured the two lengths of the peduncle. We measured all traits for each ramet in each of the 3 years of the study period. We scored the date when a plant attained a height of 1 cm above the soil surface as the budding date. To ensure that we did not miss budding or first flowering, we surveyed the plants daily at the start of each growing season. Before analysis, we converted the budding date and the first flowering into the number of days to bud initiation and flowering from the first budding date observed for the entire population of ramets. We measured the lengths of the peduncle on the first flowering day and at the end of the flowering period, i.e. the date when all flowers had abscised. We recorded the number of flowers and leaves at the end of the flowering period.

(iii) DNA extraction and genotyping of microsatellite markers

We extracted the genomic DNA of each plant from frozen leaves using a modified Cetyl trimethyl ammonium bromide (CTAB) method (Murray & Thompson, Reference Murray and Thompson1980). We have already developed many microsatellite markers for this species (Ueno et al., Reference Ueno, Yoshida, Taguchi, Honjo, Kitamoto, Washitani, Ohsawa and Tsumura2009), but we have no information on linkage between these markers and quantitative traits. We determined the genotypes of the 110 genets using the eight pairs of microsatellite PCR primers (Table 2) used by Honjo et al. (Reference Honjo, Kitamoto, Ueno, Tsumura, Washitani and Ohsawa2009). The PCR conditions for each primer followed the protocol described by Honjo et al. (Reference Honjo, Kitamoto, Ueno, Tsumura, Washitani and Ohsawa2009). The PCR products were analysed using a 3100 Genetic Analyzer and the GeneScan software (Applied Biosystems).

Table 2.

Details of the eight microsatellite markers used in this study, and the gene diversity averaged over loci (He), expected number of alleles observed in 13 genets (n13a) detected for the eight loci in P. sieboldii

3. Data analysis

(i) Evaluation of quantitative traits

We used factorial analysis to examine the effects of population, year, the population×year interaction and genet on the investigated traits using the values for each ramet. We performed restricted maximum likelihood (REML) tests to estimate the contribution of variance among populations, among genets within a population and among ramets within a genet to the total phenotypic variance for each investigated trait. The variation among ramets was assumed to represent environmental variation, because each ramet of a given genet has an identical genotype. Thus, the variation among genets and populations was considered to represent genetic variation, and we calculated broad-sense heritability (h 2) by dividing the genotypic variance by the total variance (Falconer, Reference Falconer1981). We used the Tukey–Kramer multiple-comparison test (with P<0·05 considered significant) to detect significant differences in the average values among pairs of populations for all traits. All statistical tests were performed using the JMP 6.0 statistical software (SAS Institute Inc.).

(ii) Genetic variation within populations

The genetic variation within populations in quantitative traits, i.e. the variation among genets V g was calculated as

(1)
V_{\rm g} \equals {{\lpar {\rm MS \ genets} \minus {\rm MS \ ramets}\rpar } \over r}\comma

where MS represents the mean square. To represent the number of ramets per genet, we used r, the effective number of replicates:

(2)
r \equals \left( {\mathop\sum\limits_{i \equals \setnum{1}}^{a} {r_{i} \minus \mathop\sum\limits_{i \equals \setnum{1}}^{a} {r_{i} ^{\setnum{2}} } } \sol \mathop\sum\limits_{i \equals \setnum{1}}^{a} {r_{i} } } \right)\bigg/ \lpar a \minus 1\rpar \comma

where a is the number of genets and ri is the number of replicates of genet i.

To evaluate the genetic variation for the eight microsatellite markers, we used the FSTAT 2.9.3 software (Goudet, Reference Goudet2001) to calculate unbiased estimates of gene diversity averaged over all loci (H e; Nei & Roychoudhury, Reference Nei and Roychoudhury1974); the expected number of alleles observed in 13 genets from each population, which represents the smallest sample size (n 13a, allelic richness).

(iii) Genetic variation among populations

We estimated the degree of population differentiation for each of the investigated traits using Q ST (Spitze, Reference Spitze1993), which was calculated as

(3)
Q_{{\rm ST}} \equals \sigma _{{\rm gb}}^{\rm \setnum{2}} \sol \lpar \sigma _{{\rm gb}}^{\rm \setnum{2}} \plus 2\sigma _{{\rm gw}}^{\rm \setnum{2}} \rpar

where σ2gb is the between-population variance component and σ2gw is the within-population, i.e. among genets variance component taken from the nested ANOVA using the JMP software.

The degree of population differentiation in molecular markers (F ST; Weir & Cockerham, Reference Weir and Cockerham1984) was estimated using the genotype data for the 110 genets based on the eight pairs of microsatellite primers. Furthermore, we compared the pairwise Q ST and F ST values to provide insights into the potential mechanisms responsible for genetic differentiation among the populations. We tested for correlations between each pair of Q ST and F ST values using the Mantel test (Mantel, Reference Mantel1967). We calculated F ST and the pairwise F ST and carried out the Mantel test using the FSTAT 2.9.3 software.

(iv) Relationships between pairwise Q ST and ecological factors

To explore the potential causes of differentiation among populations, we investigated the relationships between five ecological factors and the traits that revealed to have differentiated among the populations by directional selection from the relationship between Q ST and F ST. We investigated between four environmental factors (temperature, altitude, latitude and longitude) and the average values of these traits per population, respectively. We also investigated the relationship between geographical distance and the pairwise Q ST for these traits. These relationships were quantified using Spearman's rank-correlation coefficient, calculated using the JMP software.

4. Results

Because few genets flowered in 2005, we have only analysed the number of days to flowering, the number of flowers, and the lengths of the peduncle at the start and end of flowering using the 2003 and 2004 data. Table 3 shows the number of genets and years that we actually analysed.

Table 3.

Number of genets and years analysed for each trait in this study

(i) Evaluation of quantitative traits

All traits differed significantly among genets (P<0·01; Table 4). This suggests that all the quantitative traits were determined by genetic factors. All traits differed significantly among years, with large F-values (Table 4), indicating that quantitative traits were greatly influenced by variations in environmental effects among the years. In 2004, the number of days to bud initiation was longer, the length of the peduncle at the start of flowering was longer, the length of the peduncle at the end of flowering was shorter and the number of leaves was smaller than in 2003; this was also true in 2005.

Table 4.

Results of the factorial analysis for the six investigated traits

* P<0·05

** P<0·01.

The number of days to bud initiation, number of days to flowering, number of flowers and number of leaves differed significantly among populations, but the two lengths of the peduncle did not (P<0·01; Table 4). The number of days to bud initiation showed the largest F-value of the six traits (Table 4), and the Saitama population produced buds consistently and significantly earlier than the other populations in all 3 years, whereas Nagano 2 tended to produce buds later than the other four populations, with the difference significant in 2003 (Table 5). The number of days to flowering tended to differ between years in each population (Table 5). For the number of flowers, the differences among populations were smaller in 2004 than in 2003, and there were fewer significant differences. The genets from the Saitama and the Nagano1 populations tended to have a large number of leaves in all 3 years (Table 5).

Table 5.

Mean values and genetic variation (Vg) for each trait within the five populations from 2003 to 2005

Means within a year followed by different letters differ significantly among populations (P<0·05).

We found a significant population×year interaction for three traits (Table 4): the number of days to bud initiation and flowering (both P<0·01) and the length of the peduncle at the end of flowering (P<0·05).

Table 6 presents the broad-sense heritability values for the six traits. The heritabilities of all traits were largest in 2003, with one exception: the value for the number of leaves was largest in 2005. In each year, the number of days to bud initiation consistently showed the smallest variance among the ramets and the largest heritability (61·2–74·6), whereas the number of leaves showed the biggest variance among ramets and the smallest heritability (32·4–40·1).

Table 6.

Proportions (%) of the three variance components (among populations, genets, and ramets) calculated by means of REML, and the broad-sense heritability (h2) values for the investigated traits from 2003 to 2005. QST represents the degree of genetic differentiation among populations

Variance values may not add to 100% due to rounding errors.

(ii) Genetic variation in quantitative traits within and among populations

Of the total variance in the number of days to bud initiation, an average of about 45% was partitioned among the populations across the 3 years, and the variance among populations was larger than that among genets (Table 6). For the other five traits, the variance among genets was larger than that among populations. This reveals a high degree of genetic variation within the populations.

The degree of genetic differentiation among populations (Q ST) ranged widely, from 0·004 to 0·623 (Table 6). The Q ST values for the number of days to bud initiation were particularly large (0·400–0·623), indicating that this trait was considerably differentiated among the populations. Of the other five traits, Q ST ranged from 0·004 for the length of the peduncle at the end of flowering in 2004 to 0·296 for the number of leaves in 2005, and these Q ST values were clearly smaller than those for the number of days to bud initiation.

(iii) Molecular versus quantitative genetic variation

According to our SSR analysis, each locus was polymorphic, with 6–24 alleles per locus (Table 2), and we detected a total of 96 alleles over the eight loci. Mean H e ranged from 0·604 to 0·751, mean n 13a on ranged from 4·03 to 6·63.

The F ST value calculated using the data for the eight SSR markers was 0·172 (two-sided 95% confidence interval, 0·115–0·232). The average Q ST values for the number of days to bud initiation and the number of leaves were considerably and slightly larger than F ST, respectively (Fig. 1), whereas the average Q ST values for the lengths of the peduncle at the start and end of flowering were smaller than F ST. The average Q ST for days to flowering and the numbers of flowers were approximately equal to F ST.

Fig. 1.

Q ST values for all traits in 2003, 2004 and 2005, and the F ST value evaluated using the data for the eight microsatellite markers in Table 2. The grey lines represent the two-sided 95% confidence interval for F ST (mean=0·172, range=0·115–0·232). Traits: DB, days to bud initiation; DF, days to flowering; LPS, length of the peduncle at the start of flowering; LPE, length of the peduncle at the end of flowering; NF, number of flowers; NL, number of leaves.

Pairwise Q ST and pairwise F ST were not significantly correlated for any parameter. However, the pairwise Q ST value for the number of days to bud initiation differed greatly between the Saitama population and the other four populations in 2004 and 2005 (Fig. 2). These results suggest that the Saitama population differed considerably from the other populations. While the pairwise Q ST value for the number of leaves ranged widely from 0 to 1 and Saitama population differed from each population expect Nagano 1 population in 2005, this difference is not yet pronounced as the days to bud initiation (Fig. 2).

Fig. 2.

Relationship between the pairwise Q ST and pairwise F ST values for days to bud initiation and the number of leaves. The data points surrounded by a dotted border represent the genetic distance between the Saitama population and the other four populations.

(iv) Relationships between quantitative traits and ecological factors

Because the average Q ST values for the number of days to bud initiation and the number of leaves were larger than F ST (Fig. 1), we examined the relationship between these parameters and the five ecological factors. The mean temperature of each month, from February to June, were significantly negatively correlated with the number of days to bud initiation in 2004 and 2005, and they were significantly positively correlated with the number of leaves in all 3 years (Fig. 3), suggesting that the differences in mean values of these parameters among populations correspond to differences in the mean temperature of each habitat. Latitude, longitude and altitude were not significantly correlated with the mean values of these two parameters, and geographical distance was not significantly correlated with the pairwise Q ST for these two traits.

Fig. 3.

The relationship between the mean temperature and the mean number of days to bud initiation and the mean number of leaves in 2003, 2004 and 2005. Because the relationships at each month were same patterns, only mean February temperature was shown. The data points surrounded by a dotted border represent the data of Saitama population. *P<0·05, **P<0·01.

5. Discussion

H e and n 13a of five populations were compared to those of wild populations evaluated by Honjo et al. (Reference Honjo, Kitamoto, Ueno, Tsumura, Washitani and Ohsawa2009), and genetic variation in the five populations is reflected in the mean variation that P. sieboldii have maintained in Japan.

(i) What traits were adaptive in P. sieboldii?

The lengths of the peduncle at the start and end of flowering did not differ significantly among populations (Table 4), although the length of the peduncle at the end of flowering varied among the years. The mean Q ST was smaller than F ST for both parameters (Fig. 1), suggesting that the length of the peduncle has not undergone directional selection and maintained mostly genetic variation within populations. However, the numbers of days to bud initiation, days to flowering, flowers, and leaves differed significantly among the populations (Table 4). The numbers of days to flowering and flowers appear to have become differentiated among populations mainly as a result of genetic drift because Q ST was nearly equal to F ST for these parameters (Fig. 1). Because high genetic diversity of these traits is maintained within the wild populations, Q ST values were relatively low (Table 6). The mean Q ST values for the number of days to bud initiation and the number of leaves were both larger than the corresponding F ST. However, we do not necessarily conclude that the number of leaves is an adaptive trait from the result that this trait showed the smallest heritability in all years and is susceptible to environmental variation. Thus, the days to bud initiation is likely to be involved in adaptive differentiation among the populations as a result of directional selection.

The mean numbers of days to bud initiation in 2004 and 2005 were negatively correlated with the mean temperatures at each site (Fig. 3; P<0·05), indicating that the mean temperature of the Saitama population, which appears to have differed from the other populations (Fig. 2), is considerably warmer (by more than 5°C) than that of the other populations (Table 1). Thus, the winter temperature appears to be involved in determining the earliness of the budding date in P. sieboldii. In the Nagano region, which is located at the highest altitudes of the five study areas and showed correspondingly low temperatures, sudden late frost is often observed until the middle of May. Because P. sieboldii usually buds in early spring, genets with early budding are susceptible to frost injury and selection would favour genets with late budding in all except the Saitama population, where sudden late frosts are uncommon. In out-crossing forest trees such as Scots pine, Populus spp. and Douglas-fir, it has often been reported that traits related to budding, such as bud set and bud flush dates, are adaptive traits that are closely related to winter temperatures (Hurme et al., Reference Hurme, Repo, Savolainen and Pääkkönen1997; Frewen et al., Reference Frewen, Chen, Howe, Davis, Rohde, Boerjan and Bradshaw2000; Howe et al., Reference Howe, Aitken, Neale, Jermstad, Wheeler and Chen2003).

Positive relationships between bud flush and cold hardiness have been reported, and their quantitative trait loci (QTLs) were detected at the same position as a result of pleiotropy or strong linkage (Howe et al., Reference Howe, Aitken, Neale, Jermstad, Wheeler and Chen2003). Winter chilling, spring temperatures and frost are factors capable of controlling the timing of bud flush (Jermstad et al., Reference Jermstad, Bassoni, Jech, Ritchie, Wheeler and Neale2003). Thus, low temperatures in early spring appear to act as a selective pressure that selects genets with late budding in the three Nagano populations and the Hokkaido population. In contrast, genets with early budding appear to have been selected in the Saitama population. Frewen et al. (Reference Frewen, Chen, Howe, Davis, Rohde, Boerjan and Bradshaw2000) noted that trees in which bud flush occurs too late in the spring have a shortened growing season that reduces their competitive ability and growth potential. Thus, temperature appears to be a selective factor involved in local adaptation, and phenotypic fixation of the budding date has occurred as a result of the selection of optimal genotypes in each habitat.

(ii) Defining conservation units for P. sieboldii based on adaptive and neutral genetic markers

Honjo et al. (Reference Honjo, Kitamoto, Ueno, Tsumura, Washitani and Ohsawa2009) defined four P. sieboldii conservation units (Hokkaido, northern Honshu, central Honshu and western Japan) using eight microsatellite markers and cpDNA. Based on those results, the Saitama population and the three Nagano populations in the present study belonged to the same conservation unit, i.e. central Honshu, and only the Hokkaido population belonged to a different conservation unit. However, our budding date results suggest that the Saitama population differed considerably from the other populations. Therefore, the Saitama population should represent different conservation units. Among these Nagano populations, Nagano 2 differed significantly in the budding date from Nagano 1 and Nagano 3 in 2003 (Table 5). Therefore, further study is required to exploit the genetic basis of such differences that may be because of difference in the conservation unit.

As a result, the five populations in the present study could be divided into three conservation units previously proposed by Honjo et al. (Reference Honjo, Kitamoto, Ueno, Tsumura, Washitani and Ohsawa2009) as well as at least one new conservation unit based on our analysis of both their adaptive traits and molecular markers. These results clearly demonstrate the importance of defining conservation units based on adaptive diversity, not just molecular genetic diversity.

Hypothetically speaking, if transplantation of genets is necessary to maintain remnant populations of P. sieboldii, it should be guided by a careful consideration of the genetic composition of the genets in terms of the budding date, which represents an important adaptation to environmental conditions. The introduction of foreign genotypes that have adapted to a different environment can potentially disrupt the genetic architecture of native populations, which is optimal for long-term existence of these populations in their native habitat. This phenomenon has been called ‘outbreeding depression’ and has frequently been reported for some species (Keller et al., Reference Keller, Kollmann and Edwards2000; Montalvo & Ellstrand, Reference Montalvo and Ellstrand2001). For P. sieboldii, the temperature during the budding stage appears to be a strong selection force for the budding date. Therefore, it should be possible to roughly estimate the timing of budding using the mean February temperature at a genet's native habitat. Wild genets or populations used for transplantation between regions should be selected based on the temperature during the budding stage, not based solely on geographical distance.

We thank Professor Kazutoshi Okuno (University of Tsukuba) for his helpful comments for improving this paper. This study was partly supported by the Japanese Society for the Promotion of Science via a Grant in aid for Scientific Research (A) (20241055).

References

Bekessy, S. A., Ennos, R. A., Burgman, M. A., Newton, A. C. & Ades, P. K. (2003). Neutral DNA markers fail to detect genetic divergence in an ecologically important trait. Biological Conservation 110, 267–275.CrossRefGoogle Scholar
Cano, J. M., Mäkinen, H. S., Leinonen, T., Freyhof, J. & Merilä, J. (2008). Extreme neutral genetic and morphological divergence supports classification of Adriatic three-spined stickleback (Gasterosteus aculeatus) populations as distinct conservation units. Biological Conservation 141, 1055–1066.CrossRefGoogle Scholar
Carvajal-Rodriguez, A., Rolan-Alvarez, E. & Caballero, A. (2005). Quantitative variation as a tool for detecting human-induced impacts on genetic diversity. Biological Conservation 124, 1–13.CrossRefGoogle Scholar
Clausen, D. J., Keck, D. D. & Hiesey, W. M. (1940). Experimental studies on the nature of species. I. Effect of Varied Environments on Western North American Plants. Washington, DC: Carnegie Institute of Washington Publications. No. 520.Google Scholar
Crandall, K. A., Bininda-Emonds, O. R. P., Mace, G. M. & Wayne, R. K. (2000). Considering evolutionary processes in conservation biology. Trends in Ecology and Evolution 15, 290–295.CrossRefGoogle ScholarPubMed
de Guia, A. P. O. & Saitoh, T. (2007). The gap between the concept and definitions in the evolutionarily significant unit: the need to integrate neutral genetic variation and adaptive variation. Ecological Research 22, 604–612.CrossRefGoogle Scholar
Denoël, M. (2007). Priority areas of intraspecific diversity: Larzac, a global hotspot for facultative paedomorphosis in amphibians. Animal Conservation 10, 110–116.CrossRefGoogle Scholar
Falconer, D. S. (1981). Introduction to Quantitative Genetics, 2nd edn. London, UK: Longman.Google Scholar
Frankham, R., Ballou, J. D. & Briscoe, D. A. (2002). Introduction to Conservation Genetics. New York, NY: Cambridge University Press.CrossRefGoogle Scholar
Fraser, R. & Bernatchez, L. (2001). Adaptive evolutionary conservation: towards a unified concept for defining conservation units. Molecular Ecology 10, 2741–2752.CrossRefGoogle ScholarPubMed
Frewen, B. E., Chen, T. H. H., Howe, G. T., Davis, J., Rohde, A., Boerjan, W. & Bradshaw, H. D. Jr (2000). Quantitative trait loci and candidate gene mapping of bud set and bud flush in Populus. Genetics 154, 837–845.CrossRefGoogle ScholarPubMed
Gordon, D. R. & Rice, K. J. (1998). Patterns of differentiation in Wiregrass (Aristida beyrichiana): Implications for restoration effects. Restoration Ecology 6, 166–174.CrossRefGoogle Scholar
Goudet, J. (2001) FSTAT, A Program to Estimate and Test Gene Diversities and Fixation Indices (version 2.9.3). Available from http://www2.unil.ch/popgen/softwares/fstat.htmGoogle Scholar
Haig, S. M. (1998). Molecular contributions to conservation. Ecology 79, 413–425.CrossRefGoogle Scholar
Holderegger, R., Kamm, U. & Gugerli, F. (2006). Adaptive vs. neutral genetic diversity: implications for landscape genetics. Landscape Ecology 21, 797–807.CrossRefGoogle Scholar
Honjo, M. (2005). Molecular ecological genetics for the conservation of genetic diversity in Primula sieboldii. Doctoral dissertation, University of Tsukuba (in Japanese).Google Scholar
Honjo, M., Kitamoto, N., Ueno, S., Tsumura, Y., Washitani, I. & Ohsawa, R. (2009). Management units of the endangered herb Primula sieboldii based on microsatellite variation among and within populations throughout Japan. Conservation Genetics 10, 257–267.CrossRefGoogle Scholar
Honjo, M., Ueno, S., Tsumura, Y., Handa, T., Washitani, I. & Ohsawa, R. (2008). Tracing the origins of stocks of the endangered species Primula sieboldii using nuclear microsatellites and chloroplast DNA. Conservation Genetics 9, 1139–1147.CrossRefGoogle Scholar
Honjo, M., Ueno, S., Tsumura, Y., Washitani, I. & Ohsawa, R. (2004). Phylogeographic study based on intraspecific sequence variation of chloroplast DNA for the conservation of genetic diversity in the Japanese endangered species Primula sieboldii. Biological Conservation 120, 215–224.CrossRefGoogle Scholar
Howe, G. T., Aitken, S. N., Neale, D. B., Jermstad, K. D., Wheeler, N. C. & Chen, T. H. H. (2003). From genotype to phenotype: unraveling the complexities of cold adaptation in forest trees. Canadian Journal of Botany 81, 1247–1266.CrossRefGoogle Scholar
Hurme, P., Repo, T., Savolainen, O. & Pääkkönen, T. (1997). Climatic adaptation of bud set and frost hardiness in Scots pine (Pinus sylvestris). Canadian Journal of Forest Research 27, 716–723.CrossRefGoogle Scholar
Isagi, Y., Honjo, M. & Washitani, I. (2001). Development of microsatellite markers for Primula sieboldii using degenerate oligonucleotide-primed PCR-amplified DNA. Molecular Ecology Notes 1, 22–24.CrossRefGoogle Scholar
Iwatsuki, K. (2008). Endangered vascular plants in Japan – present status and a proposal for conservation. Proceedings of the Japan Academy, Series B, Physical and Biological Sciences 84, 275–286.CrossRefGoogle Scholar
Jermstad, K. D., Bassoni, D. L., Jech, K. S., Ritchie, G. A., Wheeler, N. C. & Neale, D. B. (2003). Mapping of quantitative trait loci controlling adaptive traits in coastal Douglas Fir. III. Quantitative trait loci-by-environmental interactions. Genetics 165, 1489–1506.CrossRefGoogle ScholarPubMed
Karhu, A., Hurme, P., Karjalainen, M., Karvonen, P., Kärkkäinen, K., Neale, D. & Savolainen, O. (1996). Do molecular markers reflect patterns of differentiation in adaptive traits of conifers? Theoretical and Applied Genetics 96, 215–221.CrossRefGoogle Scholar
Keller, M., Kollmann, J. & Edwards, P. J. (2000). Genetic introgression from distant provenances reduces fitness in local weed populations. Journal of Applied Ecology 37, 647–659.CrossRefGoogle Scholar
Kitamoto, N., Honjo, M., Ueno, S., Takenaka, A., Tsumura, Y., Washitani, I. & Ohsawa, R. (2005). Spatial genetic structure among and within populations of Primula sieboldii growing beside separate streams. Molecular Ecology 14, 149–157.CrossRefGoogle ScholarPubMed
Leinonen, T., O'Hara, R. B., Cano, J. M. & Merilä, J. (2008). Comparative studies of quantitative trait and neutral marker divergence: a meta-analysis. Journal of Evolutionary Biology 21, 1–17.CrossRefGoogle ScholarPubMed
Mantel, N. (1967). The detection of disease clustering and a generalized regression approach. Cancer Research 27, 209–220.Google Scholar
Mckay, J. K., Bishop, J. G., Lin, J.-Z., Richards, J. H., Sala, A. & Mitchell-Olds, T. (2001). Local adaptation across a climatic gradient despite small effective population size in the rare sapphire rockcress. Proceedings of the Royal Society London B 268, 1715–1721.CrossRefGoogle ScholarPubMed
McKay, J. K. & Latta, R. J. (2002). Adaptive population divergence: markers, QTL and traits. Trends in Ecology and Evolution 17, 285–291.CrossRefGoogle Scholar
Merilä, J. & Crnokrak, P. (2001). Comparison of genetic differentiation at marker loci and quantitative traits. Journal of Evolutionary Biology 14, 892–903.CrossRefGoogle Scholar
Montalvo, A. M. & Ellstrand, N. C. (2001). Nonlocal transplantation and outbreeding depression in the subshrub Lotus scoparius (Fabaceae). American Journal of Botany 88, 258–269.CrossRefGoogle ScholarPubMed
Moritz, C. S. (1994). Defining ‘evolutionarily significant units’ for conservation. Trends in Ecology and Evolution 9, 373–375.CrossRefGoogle ScholarPubMed
Murray, M. G. & Thompson, W. F. (1980). Rapid isolation of high molecular weight plant DNA. Nucleic Acids Research 8, 4321–4325.CrossRefGoogle ScholarPubMed
Nei, M. & Roychoudhury, A. K. (1974). Sampling variance of heterozygosity and genetic distance. Genetics 76, 379–390.CrossRefGoogle ScholarPubMed
Reed, D. H. & Frankham, R. (2001). How closely correlated are molecular and quantitative measures of genetic variation? A meta-analysis. Evolution 55, 1095–1103.Google ScholarPubMed
Spitze, K. (1993). Population structure in Daphnia obtusa: quantitative genetic and allozymic variation. Genetics 135, 367–374.CrossRefGoogle ScholarPubMed
Storfer, A. (1996). Quantitative genetics: a promising approach for the assessment of genetic variation in endangered species. Trends in Ecology and Evolution 11, 343–348.CrossRefGoogle ScholarPubMed
Ueno, S., Tsumura, Y. & Washitani, I. (2003). Development of microsatellite markers in Primula sieboldii E. Morren, a threatened Japanese perennial herb. Conservation Genetics 4, 809–811.CrossRefGoogle Scholar
Ueno, S., Kitamoto, N., Ohsawa, R., Tsumura, Y. & Washitani, I. (2005). Nine additional microsatellite markers for Primula sieboldii E. Morren. Conservation Genetics 6, 1063–1064.CrossRefGoogle Scholar
Ueno, S., Yoshida, Y., Taguchi, Y., Honjo, M., Kitamoto, N., Washitani, I., Ohsawa, R. & Tsumura, Y. (2009). Development of 120 microsatellite markers for Primula sieboldii E. Morren for linkage mapping. Conservation Genetics DOI 10.1007/S10592-009-9861-9.CrossRefGoogle Scholar
Washitani, I. & Kabaya, H. (1988). Germination responses to temperature responsible for the seedling emergence seasonality of Primula sieboldii E. Morren in its natural habitat. Ecological Research 3, 9–20.CrossRefGoogle Scholar
Weir, B. S. & Cockerham, C. C. (1984). Estimating F-statistics for the analysis of population structure. Evolution 38, 1358–1370.Google ScholarPubMed
Yamazaki, T. (1993). Primula. In Flora of Japan, vol. IIIa (ed. Kunio, I., Takashi, Y., David, B. E. & Hideaki, O.), pp. 87–94. Tokyo: Kodansha. (in Japanese).Google Scholar
Yoshida, Y., Honjo, M., Kitamoto, N. & Ohsawa, R. (2008). Genetic variation and differentiation of floral morphology in wild Primula sieboldii evaluated by image analysis and SSR markers. Breeding Science 58, 301–307.CrossRefGoogle Scholar
Figure 0

Table 1. Locations of the five Japanese populations of P. sieboldii, number of genets (total and number sampled) and average monthly temperatures at each site

Figure 1

Table 2. Details of the eight microsatellite markers used in this study, and the gene diversity averaged over loci (He), expected number of alleles observed in 13 genets (n13a) detected for the eight loci in P. sieboldii

Figure 2

Table 3. Number of genets and years analysed for each trait in this study

Figure 3

Table 4. Results of the factorial analysis for the six investigated traits

Figure 4

Table 5. Mean values and genetic variation (Vg) for each trait within the five populations from 2003 to 2005

Figure 5

Table 6. Proportions (%) of the three variance components (among populations, genets, and ramets) calculated by means of REML, and the broad-sense heritability (h2) values for the investigated traits from 2003 to 2005. QST represents the degree of genetic differentiation among populations

Figure 6

Fig. 1. QST values for all traits in 2003, 2004 and 2005, and the FST value evaluated using the data for the eight microsatellite markers in Table 2. The grey lines represent the two-sided 95% confidence interval for FST (mean=0·172, range=0·115–0·232). Traits: DB, days to bud initiation; DF, days to flowering; LPS, length of the peduncle at the start of flowering; LPE, length of the peduncle at the end of flowering; NF, number of flowers; NL, number of leaves.

Figure 7

Fig. 2. Relationship between the pairwise QST and pairwise FST values for days to bud initiation and the number of leaves. The data points surrounded by a dotted border represent the genetic distance between the Saitama population and the other four populations.

Figure 8

Fig. 3. The relationship between the mean temperature and the mean number of days to bud initiation and the mean number of leaves in 2003, 2004 and 2005. Because the relationships at each month were same patterns, only mean February temperature was shown. The data points surrounded by a dotted border represent the data of Saitama population. *P<0·05, **P<0·01.