Crop Science
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 20 May 2008
Published in Crop Sci 48:1071-1079 (2008)
© 2008 Crop Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Kuroda, Y.
Right arrow Articles by Vaughan, D. A.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Kuroda, Y.
Right arrow Articles by Vaughan, D. A.
Agricola
Right arrow Articles by Kuroda, Y.
Right arrow Articles by Vaughan, D. A.
Related Collections
Right arrow Soybean
Right arrow Plant Genetic Resources

Gene Flow and Genetic Structure of Wild Soybean (Glycine soja) in Japan

Yosuke Kuroda, Akito Kaga, Norihiko Tomooka and Duncan A. Vaughan*

Genebank, National Institute of Agrobiological Sciences, 2-1-2 Kannondai, Tsukuba, Ibaraki 305-8602, Japan. The Ministry of Agriculture, Forestry and Fisheries of Japan and Global Environment Research Fund of the Japanese Ministry of the Environment supported this research. The first two authors contributed equally to this research

* Corresponding author (duncan{at}affrc.go.jp).


   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
In many parts of Japan cultivated soybean and wild soybean are sympatric. The objective of this study was to measure gene movement from cultivated soybean to wild soybean and within wild soybean in natural populations in Japan. Seven microsatellite markers that were found to have particularly high ability to discriminate components of the soybean complex in Japan were used to measure the extent of pollen and seed dispersal within and between populations of G. soja at seven localities (14 populations) in northern, central, and southern regions of Japan (1334 seeds/168 individuals) were measured. Each G. soja site consisted of two neighboring populations (100 m–2 km apart), which were adjacent to ( <5 m), or isolated from ( >50 m), cultivated soybean fields. Gene flow from G. max to G. soja was not detected. However, in populations of G. soja, the outcrossing rate ranged from 0–6.3% (average 2.2%), and the dispersal distance of pollen ranged from 5–25 m (average 10.5 m). In addition, 13 individuals of G. soja (7.7%) collected in four populations were assigned to neighboring populations (100–400 m). The results suggest the occurrence of long-distance seed dispersal in G. soja. Spatial autocorrelation analysis revealed a clear isolation by distance model, and a strong positive genetic correlation among individuals was observed within a range of 400 m (r = 0.168–0.551). Gene flow by hybridization among G. soja individuals occurs more frequently than the rare hybridization events between G. max and G. soja in Japan.

Abbreviations: SSR, simple sequence repeats


   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
SOYBEAN [Glycine max (L.) Merrill] belongs to the genus Glycine subgenus Soja (Singh et al., 2007). The subgenus Soja consists of soybean and its presumed direct wild ancestor, G. soja Sieb. & Zucc. (Singh et al., 2007). G. soja and G. max are cross compatible and belong to the primary gene pool of Harlan and deWet (1971). G. soja is distributed in the Russian Far East, eastern China including Taiwan, the Korean peninsula, and Japan. Therefore, gene flow from cultivated soybean to G. soja can occur in these areas when they are sympatric and have synchronous flowering. Analysis of the chloroplast genome of wild and cultivated soybeans has suggested hybridization from cultivated to wild soybean has occurred during the history of soybean cultivation in Japan (Abe et al., 1999).

Two steps are required for cultivated soybean genes to be dispersed to natural populations of G. soja. First, pollen dispersal from G. max to G. soja must occur. The second step is dispersal of soybean genes in G. soja by further pollen and seed flow. In several parts of Japan field surveys indicate that where G. max and G. soja are sympatric their flowering periods overlap (Kuroda et al., 2005). Among the wild soybean populations surveyed by the authors for individual plants with unusually large leaves, pods and seeds, presumed hybrid derivatives between G. max and G. soja have been found in northern and southern parts of Japan (Kaga et al., 2005; Kuroda et al., 2005). The hybrid status of these individuals was confirmed by checking their simple sequence repeats (SSR) profile in comparison with wild and cultivated plants from the same site. Under experimental conditions the outcrossing rate from G. max to G. soja with a plant spacing of 50 cm has been reported as 0.73% (Nakayama and Yamaguchi, 2002). However, the extent of outcrossing in natural habitats has not been reported. Although G. soja is known to be a predominantly inbreeding species, outcrossing rates ranging from 2 to 13% have been reported (Kiang et al., 1992; Fujita et al., 1997; Kuroda et al., 2006). In addition, natural seed shattering results in seeds being scattered up to 4.5 m from mother plants of G. soja (Oka, 1983). Rare long-distance seed dispersal (200 m and 12.4 km) of G. soja has been reported (Kuroda et al., 2006). However, information on the extent of pollen and seed dispersal for G. soja is limited.

Molecular markers are useful for estimating hybridization events and seed movement (Parker et al., 1998; Cain et al., 2000). Microsatellite SSR markers show high levels of polymorphisms in both intra- and inter-population studies. The microsatellite data have been applied to the statistical analyses of pollen and seed dispersal using paternity assignment (e.g., Sato et al., 2006), individual assignment (e.g., Gustafsson, 2000), and spatial autocorrelation (e.g., Heuertz et al., 2003) analyses. In soybean, microsatellite markers are available for all 20 linkage groups of the soybean genome (Song et al., 2004). A previous study identified seven microsatellite markers on different linkage groups of cultivated soybean that are particularly useful for discriminating between wild and cultivated soybean and within wild soybean (Kuroda et al., 2006). Therefore, these microsatellite markers can be used to assess the extent of pollen and seed dispersal in G. soja.

The population genetic structure of wild soybean, and gene flow and seed dispersal in wild soybean based on samples of wild and cultivated soybean from areas throughout Japan where they have overlapping distribution, has previously been evaluated using SSR markers (Kuroda et al., 2006). The focus of this research is a detailed analysis of wild soybean genetic structure in three localities in Japan, northern (Akita prefecture), central (Ibaraki prefecture), and southern (Saga prefecture). The specific objectives were to measure (i) pollen dispersal from G. max to G. soja, (ii) pollen dispersal in populations of G. soja, (iii) seed dispersal of G. soja, and (iv) spatial genetic structure of G. soja populations, using nuclear microsatellite variation. The results from this work will provide experimental evidence for the role of gene flow in soybean populations under natural conditions in Japan.


   MATERIALS AND METHODS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
Collection of Materials
Prefectures in northern (Akita), central (Ibaraki), and southern (Hiroshima and Saga) Japan were visited one to three times during 2004 during the vegetative, flowering, and maturing stages of wild soybean (G. soja) to explore, collect, and record changes based on previous visits to the same populations of G. soja (Kuroda et al., 2005) (Fig. 1 ). Among the visited sites, a total of 14 populations from Akita (four populations on the Omono river system), Ibaraki (four populations on the Kokai river system), and Saga (six populations on the Chikugo river system) were selected for this study. G. soja populations in these localities were found in the agricultural landscape where soybean is commonly cultivated. In addition, we found natural hybrids between cultivated and wild soybean in the Omono and Chikugo river systems. Spatial distances among populations ranged from approximately 0.1 to 1200 km based on GPS readings (Fig. 1). To evaluate pollen dispersal from G. max, seven G. soja populations adjacent to soybean fields ( <5 m) were selected (‘a’ population). Another seven G. soja populations (‘b’ populations) were selected from the sites 100 m–2 km away from soybean fields. This second set of populations was isolated from soybean fields ( >60 m) a distance greater than reported for pollen flow in soybeans (Caviness, 1966; Jin et al., 2003). Seed samples were collected at the optimum maturing stage of G. soja (October– November) a single transect that for populations adjacent to soybean fields followed the line of the soybean field. In southern parts of Japan G. max and G. soja tend to flower at the same time, whereas in northern and central Japan flowering time of G. max tends to be earlier than that of G. soja. However, there is some overlap in flowering time between G. soja and G. max in northern and central Japan based on field observations. For each population, 96 seeds from 12 individuals (eight seeds one from each of eight pods per individual) were analyzed by microsatellite markers, and those individuals were spaced at 5-m intervals (Fig. 2 ). Pod samples were taken from eight different nodes up the stem that would represent pods from flowers that had flowered over a period of about two weeks. This sampling method was chosen to increase the possibility that flowers in wild and cultivated soybean had overlapping flowering time. At each site one pod from several individuals (1–9) from adjacent soybean fields were collected (Table 1 ).


Figure 1
View larger version (67K):
[in this window]
[in a new window]

  		Figure 1. Distribution of 14 populations from seven localities of G. soja analyzed.

 

Figure 2
View larger version (62K):
[in this window]
[in a new window]

  		Figure 2. Map of inferred pollen dispersal for all 14 populations of G. soja. Arrows with solid and dotted lines indicate pollen dispersal from particular and unknown individuals, respectively.

 

View this table:
[in this window]
[in a new window]

  		Table 1. Passport data for G. soja populations used in this study.

 
DNA Extraction, PCR, and Genotyping
Total DNA was extracted from 1344 (168 individuals) and 32 (32 individuals) seeds of G. soja and G. max, respectively (Table 1). The extraction method was based on Kamiya and Kiguchi (2003) with modifications with respect to the buffer used. 250 mM of sodium chloride was added to precipitate DNA efficiently. The seed coat was removed before DNA extraction to prevent contamination with maternal DNA. Seven microsatellite primer pairs located on different linkage groups of G. max were used in this study (Table 2 ). These primers showed a particularly high ability to detect variation between cultivated and wild soybeans and within Japanese wild soybeans (Kuroda et al., 2006). Forward primers were labeled with different dyes, 6-FAM, VIC, NED, or PET dyes (Applied Biosystems, Tokyo), and PCR amplification was performed using a thermal cycler (iCycler, BIORAD, Tokyo) under the conditions of one cycle of 2 min at 94°C, followed by 40 cycles for 30 s at 94°C, 30 s at about 50°C and 30 s at 68°C and finally maintained at 4°C in a volume of 10 µL [1 x buffer, 0.2 mM of dNTP, 1 mM of MgSO4, 0.3 µM of primer pairs, 0.01 unit of polymerase (KOD-plus-, Toyobo, Tokyo) and 1–5 ng of genomic DNA]. 2 µL of PCR product were denatured in 13 µL of Hi-Di formamide with 0.2 µL of GeneScan-500LIZ size standard. PCR products were separated on an AB3100 capillary sequencer (Applied Biosystems) for allele detection. Allelic data were scored using GeneMapper 3.0 software and the genotype of each sample was determined using this software.


View this table:
[in this window]
[in a new window]

  		Table 2. Variation at seven microsatellite loci in wild and cultivated soybeans.

 
Genotype Data Analysis
Intra-Population Variation
Maternal genotypes were inferred from progeny arrays (eight seeds per individual) from each population based on the method of moments using the mLTR program (Ritland and Jain, 1981). Using these genotypes, parameters of intra-population variation, which were the total number of alleles (A), expected and observed heterozygosity (HE and HO), and genetic differentiation (FIS) were estimated using the FSTAT program (Goudet, 2001). Genetic differentiation (Wright, 1965) shows deviation from the Hardy-Weinberg equilibrium (HWE). The test to evaluate either excess or deficit of heterozygotes was computed using the FSTAT program.

Outcrossing Rate and Paternity Assignment
Multilocus and single locus outcrossing rates (tm and ts) were inferred based on correlated-mating model using the mLTR program (Ritland, 2002) for the eight progeny from 12 individuals in 14 populations sampled. Although multilocus methods have lower statistical variance and higher robustness than single locus methods, the level of biparental inbreeding can be estimated from (tm– ts) (Ritland and Jain, 1981). Biparental inbreeding occurs when mating is between relatives and this is most common in plants with limited seed dispersal that are insect- and animal-pollinated. Standard deviations were based on 1000 bootstraps where family was the unit of resampling for outcrossing rate of the population. In addition, the paternity of each outcross seed was assigned to a particular pollen parent using CERVUS 2.0 program (Slate et al., 2000). All the G. soja and G. max samples were candidate pollen sources for outcrossing seed. The most likely parents were assigned when genotypes of inferred maternal parents, offspring seeds, and the candidate parents were compared and their LOD scores were larger than 3.0 (Slate et al., 2000). To reduce the effect of mismatching caused by null alleles and mutation the typing error rate was set at 0.001.

Individual Assignment
An assignment test was performed on the inferred maternal genotypes to determine whether it is possible to assign individuals to their original population or another Glycine soja population using GENECLASS 2 software (Piry et al., 2004). Each individual was assigned to the closest population to which it had the highest likelihood value of belonging when genetic distances of an individual and reference populations were compared. All the G. soja populations analyzed in this study deviated from Hardy-Weinberg equilibrium. Therefore the DA (Nei et al., 1983) distance-based method, which does not require Hardy-Weinberg equilibrium (Cornuet et al., 1999), was used to calculate the likelihood values for each individual belonging to the sampled population. An individual was rejected from the population if the probability was below a standard threshold of 5%.

Spatial Genetic Structure
Analysis of molecular variance (AMOVA) was performed to partition the observed genetic variation between individuals among regions, between individuals within regions, and between individuals within populations using the ARLEQUIN software (Schneider et al., 2000). AMOVA creates a genetic distance matrix between samples to measure the genetic structure of the population from which the samples are drawn. F-statistics were tested by 1000 permutations, and significant differences between groups declared if measured variance was lower than 95% of the variance in the null distribution (Excoffier et al., 1992).

Spatial autocorrelation analysis for multiallelic, codominant loci was used to assess spatial genetic structure of the 14 Glycine soja populations (Smouse and Peakall, 1999). A correlation coefficient (r) was calculated from pairwise geographic and squared genetic distance matrix using the GenAlex V5 program (Peakall and Smouse, 2001). The coefficient r has a mean value of ‘0’ when there is no correlation between geographical and genetic distances, and is bounded by ( –1, +1) (Smouse and Peakall, 1999). The test for statistical significance was performed based on 999 random permutations. After permutations of shuffling the individual genotypes among geographical locations, recomputing the 25th and 975th r values were taken to define the upper and lower bounds of the 95% confidence interval.


   RESULTS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
Microsatellite Variation
In total, 101 and 12 alleles over seven microsatellite loci were detected in G. soja and G. max, respectively (Table 2). All alleles detected in soybean were also detected in G. soja except one allele (Satt126, 149bp). The number of alleles at each locus in G. soja (range 6–24) was much higher than those in G. max (only one or two). Therefore, a high level of genetic differentiation was observed between G. soja and G. max (FST = 0.185–0.389, P < 1.0 x 10–5), and identical genotypes to G. max were not found in G. soja using these seven microsatellite loci.

Intra-Population Variation
Intra-population variation of G. soja varied in terms of number of alleles (mean 23, range 11–40), number of genotypes (mean 6.5, range 2–11) and expected heterozygosity (mean 0.563, range 0.130–0.818) (Table 3 ). Genetic differentiation among all the G. soja populations was significantly higher than Hardy-Weinberg equilibrium (mean 0.932, range 0.712–1.000, P < 5.1 x 10–5) (Table 3). The high level genetic differentiation indicates the predominantly inbreeding nature of G. soja.


View this table:
[in this window]
[in a new window]

  		Table 3. Microsatellite variation within populations of G. soja

 
Outcrossing Rate and Paternity Assignment
Among 1344 G. soja seeds tested, a total of 29 seeds were determined to be outcross seeds (mean outcrossing rate 2.3%, Table 4 ). Although seven G. soja populations were adjacent to G. max fields, pollen dispersal from G. max to G. soja was not detected when genotypes of the outcross seeds from these seven G. soja populations were compared with G. max genotypes (Fig. 2). For each wild population, outcrossing rate ranged from 0.0 to 6.3%. There was geographic variation in the level of outcrossing, with outcrossing higher in southern Japan (3.3%) than other regions [central (2.2%) and northern (0.8%)]. The mean level of biparental inbreeding (tm–ts) was 0.006, indicating that 26.1% of cross-pollination was biparental inbreeding (Table 4). In the analysis of paternity assignment, exclusion probability for second (pollen) parent analysis with the genotype of the first (recipient) parent was high (0.996152, 0.998601, and 0.999110 in Akita, Ibaraki, and Saga, respectively) indicating that the results are robust (unpublished data, 2005). The most likely pollen parents were assigned to 10 outcross seeds, but parents of the remaining 19 seeds were not assigned because they had more than two candidates, or no candidate was found in the sample (Fig. 2). All the assigned pollen parents belong to the same population, and the distances were 5 m (6 seeds), 10 m (1 seed), 15 m (1 seed), and 25 m (2 seeds). In spite of the low outcross rate, the 29 outcrossed seeds were found in only 19 individuals, indicating that some specific individuals tended to have more outcross seeds than others (Fig. 2).


View this table:
[in this window]
[in a new window]

  		Table 4. Estimates of multilocus outcrossing rates (tm) and single locus outcrossing rate (ts), biparental inbreeding (tm– ts) for 14 populations of G. soja.

 
Individual Assignment
One hundred thirty-five out of 168 G. soja individuals (80.3%) were assigned to their original populations (Table 5 ), suggesting that each population was generally distinct. Another 19 individuals (11.3%) could not be assigned to any population due to low probability (<0.05) based on allelic profile. The remaining 14 individuals (8.3%) were assigned to neighboring populations: two A2a individuals assigned to A2b (100m from A2a), five A2b to A2a, five S1b to S1a (400m from S1a), one S2a to S2b (400m from S2a), and one S3a to S3b (200m from S3a). One individual (S2a) was assigned with a low level of probability (<0.07) based on allelic profile, whereas the remaining 13 individuals are assigned with a relatively high probability (>0.22). Those 13 individuals (7.7%) have identical genotypes to individuals in the neighboring populations (100m–400m) for the seven microsatellite markers evaluated.


View this table:
[in this window]
[in a new window]

  		Table 5. Results of the individual assignment test for 168 individuals from 14 populations of G. soja.

 
Spatial Genetic Structure
AMOVA showed significant genetic differentiation at different partition levels, between G. soja individuals from different regions (400 km–1200 km, FCT = 0.07, P < 1.0 x 10–5), between individuals in the same region (100 m–20 km, FSC = 0.35, P < 1.0 x 10–5), and between individuals within populations (5 m– 55 m, FST = 0.39, P < 1.0 x 10–5) (Table 6 ). However, the percentages of total variation for individuals from the same region (32.7%) and within populations (60.7%) were much higher than those for between individuals from different regions (6.6%). Further aspects of spatial genetic structure were revealed by spatial autocorrelation analysis (Fig. 3 ). In this analysis, distance classes were set from fine scale to large scale according to spatial distribution of individuals used in this study (Fig. 1). The correlogram coefficient (r) was highest (0.551) at the first distance class (5m), and it rapidly decreased to nearly zero (0.042) at 14th distance class (2km). The positive coefficient continued until the 5.4 km intercept, and the coefficient beyond 20 km was slightly negative (–0.011 to –0.065). These results suggest that genetic variation in G. soja populations has a strong genetic correlation within a range of about 400 m.


View this table:
[in this window]
[in a new window]

  		Table 6. Partition of microsatellite variation in 14 populations of G. soja analyzed.

 

Figure 3
View larger version (14K):
[in this window]
[in a new window]

  		Figure 3. Correlograms of spatial pattern for G. soja analyzed based on seven microsatellite primers. The r indicates correlation coefficient. U and L indicate the upper and lower limits by 95% confidence interval for the null hypothesis of no spatial structure. Distance classes corresponding to three different levels [5 m, 10 m, 15 m, 20 m, 25 m, 30 m, 35 m, 40 m, 45 m, 50 m, and 55 m (intra-pop)], [0.1 km, 0.4 km, 2.0 km, 10 km, and 20 km (inter-pop/intra-regions)], and [400 km, 1000 km, 1200 km (inter-region)] are shown along logarithmic x-axis.

 

   DISCUSSION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
The objective of this study was to assess the potential for genes to spread from cultivated soybean to wild soybeans and subsequently spread in wild soybean populations. This is the first population genetic analysis based on nuclear microsatellites in relation to gene dispersal found within G. soja. Based on the objectives, we discuss (i) pollen dispersal from G. max to G. soja, (ii) pollen dispersal in G. soja, (iii) seed dispersal in G. soja, and (iv) spatial genetic structure of G. soja.

Pollen Dispersal from Glycine max to G. soja
Seven populations (84 individuals, 672 seeds) were adjacent to G. max in this study, but no outcrossing from G. max to G. soja was detected. Therefore the level of outcrossing between G. max and G. soja is less than 1% in their natural habitat. This very low outcrossing rate is in agreement with field experiments and field observations (Nakayama and Yamaguchi, 2002). Nakayama and Yamaguchi (2002) reported that the outcrossing rate from G. max to G. soja was 0.73% even when they were planted alternately at 0.5-m intervals. An intensive field survey specifically to find hybrids between cultivated and wild soybean involving examination of 1000's of individuals in 58 populations of G. soja across Japan revealed only 11 hybrid derivatives between G. max and G. soja from the three sites (Kuroda et al., 2005).

Pollen dispersal from G. max to G. soja was not detected in this study, although the four conditions necessary for pollen dispersal to occur were satisfied, existence of genetic markers to detect gene flow, close proximity ( <5 m), overlapping flowering time, and the presence of pollinators. In addition, pollen sources appear to be sufficient for dispersal because most of the G. max fields adjacent to G. soja analyzed were large (more than 1 ha). There are five possible reasons why pollen dispersal was restricted.

Population fragmentation. Roads separated G. soja populations and G. max fields in the study sites. This fragmentation may restrict the movement of pollinators. Boerma and Moradshahi (1975) showed that pollinators carried most G. max pollen within rows rather than between rows. Chang and Kiang (1987) monitored foraging behavior of honeybee (Apis mellifera L.) in G. max fields, and found that the bees visited the nearest flower or neighboring plants within rows.

Frequency of cleistogamous flowers. In soybean the number of cleistogamous flowers is affected by various ecological factors such as moisture, temperature and light (Uphof, 1938). In this study we were unable to measure the ratio of cleistogamous flowers. However, the frequency of cleistogamous flower is a factor affecting the outcrossing rate.

Effect of herbicides on pollinators. Japanese farmers frequently spray herbicides on and around soybean fields. Herbicides can reduce the number of pollinators in G. max fields.

Morphological differences between G. soja and G. max. The number of flowers per individual is fewer in G. max than G. soja, and large leaves cover flowers of G. max. Sih and Balthus (1987) reported that more pollinators visit plants with higher flower density. From this viewpoint, G. max is less likely to attract pollinators than G. soja when they grow at the same location.

Other environmental factors. General prevailing wind direction, wind strength, and weather at flowering time may also influence the chances of cross-pollination between cultivated and wild soybean.

Pollen Dispersal in Glycine soja
The average outcrossing rate of G. soja found in this study is 2.2% and varied among locations. This outcrossing rate is similar to the 2.3% (Kiang et al., 1992) and 3.3% (Kuroda et al., 2006) previously reported. One reason for the low outcrossing rate is the characteristics of the G. soja flower. The flower of G. soja has a pistil surrounded by 10 anthers, and pollen grains are shed onto the stigma before flowering (Free, 1970). G. soja also produces cleistogamous flowers, which do not produce outcross seeds. The period of producing chasmogamous flowers is 12 d in northern Japan (Miyashita et al., 1999) and more than 1 mo in southern Japan (Nakayama and Yamaguchi, 2002). In this study, the outcrossing rate in northern Japan (0.8%) is lower than central and southern Japan (2.3 and 3.3%), respectively. However, high pollinator populations can affect the outcrossing rate. Fujita et al. (1997) reported a relatively high outcrossing rate (13%) in four G. soja populations in Akita prefecture, northern Japan. They concluded that many visits by pollinators, particularly honeybees and carpenter bees (Xylocopa appendiculata circumovolans Smith), at the study sites caused the high outcrossing rate. The factors that influence the number of pollinators may reflect many local influences such as microclimate of population sites, proximity to urban centers and extent to which farmers apply insecticides in an area. Many studies have reported the short distances over which honeybees forage (Levin and Kerster, 1974; Chang and Kiang, 1987). Such characteristics of pollinator activity might explain the results here of the short-distance pollen is dispersed (5–25 m), that more than one outcross seed is detected within single individuals (42%), and that biparental inbreeding is common (26%).

Glycine soja Seed Dispersal
The individual assignment test is a genetic estimate of long-distance seed dispersal, and the advantage of the test is that populations do not have to be sampled exhaustively (Cain et al., 2000). In spite of the relatively small number of G. soja individuals analyzed in this study, 13 G. soja individuals (7.7%) were assigned to neighboring populations with a high level of probability. These individuals shared the same genotypes with individuals in the neighboring populations (100–400 m): populations of [A2a and A2b], [S1a and S1b], and [S3a and S3b] shared three, one, and one identical genotypes, respectively. These individuals with identical genotypes are most likely the result of long-distance seed dispersal. Previous studies have also detected rare long-distance seed dispersal of G. soja (200 m and 12.4 km) (Kuroda et al., 2006). The present results suggest that seed dispersal up to 400 m appears to be common. In addition, founder genotypes were identified in five G. soja populations (A2a, A2b, I1a, S1b, S2a). Such intra-population variation is commonly observed in Japanese G. soja (Kuroda et al., 2006). The founder effect occurs when plants in new habitats produce large numbers of seeds resulting in populations with few genotypes. The inbreeding nature of G. soja would produce many seeds with the same genotype that can be dispersed. For example, one individual growing in the S2b population produced 3154 seeds (author's unpublished data, 2005). However, the mechanism by which G. soja are dispersed long-distances is unknown. Pod dehiscence results in G. soja seeds being dispersed within 4.5 m (Oka, 1983), and the spread of G. soja branches is less than 3 m (author's unpublished data, 2005). Since G. soja is often found alongside rivers and irrigation channels, on reclaimed land, or recently abandoned land (Kuroda et al., 2005), long-distance seed dispersal via water, humans, or animals is likely (Cain et al., 2000).

Spatial Genetic Structure in Glycine soja
The correlogram (Fig. 3) clearly shows the sharp clinal spatial structure of genetic variation in G. soja, suggesting that geographic distance is the main determinant of spatial genetic structure in Japanese G. soja. The spatial structure obtained in this study can be classified into three patterns. The first pattern is characterized by the high positive correlation found in the distance classes of 5 m (r = 0.551) and 10 m (r = 0.420). Jin et al. (2003) analyzed fine-scale genetic structure of one G. soja population in China based on inter-simple sequence repeat (ISSR) markers, and reported that 81.4% of the loci were positively correlated within 10 m. The rapid decline in correlation indicates that most seeds drop near the mother plant and only a few are dispersed longer distances. Levin and Kerster (1974) reported that seed dispersal is more frequent than pollen dispersal in species that are pioneers of secondary succession. The second pattern is characterized by lower positive correlation observed between 15 and 400 m (r = 0.168–0.341). This result agrees with the existence of founder genotypes within a population and long-distance (100–400 m) seed dispersal among populations described above. The third pattern is characterized by the low negative correlation between 400 and 1200 km (r = – 0.014 to 0.042). The large-scale genetic differentiation of Japanese G. soja has previously been reported based on nuclear microsatellites (Kuroda et al., 2006), chloroplast microsatellites (Xu et al., 2002), isozyme, and nonenzyme protein (Ti) (Kiang et al., 1992), and mitochondrial RFLP (Tozuka et al., 1998) analyses. Spatial genetic structure of Japanese G. soja can be explained by both fine-scale similarity (mainly caused by seed dispersal) and large-scale differentiation (isolation by distance).

Gene Dispersal in Glycine soja
The present study did not detect gene flow from G. max to G. soja. There is a possibility that more intense sampling might have detected hybrids, but the results presented here agree with general observations and large-scale germplasm collecting of wild soybean in Japan that only rarely report hybrids between cultivated and wild soybeans. Thus gene flow from cultivated to wild soybeans is infrequent across Japan. Hybrid individuals are rarely found (Kuroda et al., 2005) compared to other crop complexes in Japan, such as the azuki bean crop complex (Vaughan et al., 2005). However, hybrid derivatives between G. max and G. soja can survive without any intervention for at least three years in semi-natural conditions (Oka, 1983) but stable hybrid populations of weedy soybeans are not reported from Japan. Gene flow in Japanese G. soja is more frequent than that from cultivated soybean to G. soja, and long-distance seed dispersal of G. soja also occurs.

Fitness in natural habitats is generally different between crops and their wild relatives. Kuroda et al. (2006) compared 616 typical G. soja individuals with 53 widely grown G. max varieties based on 20 microsatellite loci, and found that 6.8% of G. soja individuals from northern, central, and southern parts of Japan appeared to have introgression from G. max. We are currently monitoring hybrid derivatives found in natural habitats of G. soja (Kaga et al., 2005; Kuroda et al., 2005), and are studying the persistence of G. max alleles in semi-natural habitats.


   NOTES
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Permission for printing and for reprinting the material contained herein has been obtained by the publisher.

Received for publication November 20, 2007.


   REFERENCES
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Kuroda, Y.
Right arrow Articles by Vaughan, D. A.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Kuroda, Y.
Right arrow Articles by Vaughan, D. A.
Agricola
Right arrow Articles by Kuroda, Y.
Right arrow Articles by Vaughan, D. A.
Related Collections
Right arrow Soybean
Right arrow Plant Genetic Resources

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Vadose Zone Journal
Journal of Natural Resources
and Life Sciences Education		 Soil Science Society of America Journal
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome