Botanical Studies (2006) 47: 13-21.
*
Corresponding author: E-mail: mgchung@nongae.gsnu.
ac.kr
Restriction fragment length polymorphisms in the USDA
soybean germplasm from central China
Myong Gi CHUNG
1,
*, Mi Yoon CHUNG
1
, April D. Clikeman JOHNSON
2
, and Reid G. PALMER
3
1
Department of Biology, Gyeongsang National University, Jinju 660-701, Republic of Korea
2
Department of Agronomy, Iowa State University, Ames, IA 50011, USA
3
USDA-ARS, Corn Insects and Crop Genetics Research Unit, Project 3769, Department of Agronomy, Iowa State
University, Ames, IA 50011, USA
(Received April 06, 2005; Accepted October 14, 2005)
ABSTRACT.
To evaluate levels of genetic diversity in USDA soybean germplasm from central China, 107
accessions were examined at 46 RFLP loci. We compared genetic diversity in randomly selected accessions
with pre-selected accessions based upon root tip fluorescence, pubescence morphology, and isozyme pat-
terns at ten enzyme systems. We also evaluated levels of genetic diversity of the central Chinese accessions
(n = 107) by comparing previously studied ancestors and milestone cultivars (NAC, n = 64) in the USA. Fi-
nally, we estimated the degree of genetic differentiation among six Chinese provinces (Anhui, Gansu, Henan,
Jiangsu, Shaanxi, and Shanxi). There was significant difference between pre-selected and random acces-
sions in terms of the mean number of alleles per locus (A, 2.44 vs. 2.13) and allelic richness (2.26 vs. 2.10).
However, the former (H
e
= 0.393) maintained levels of gene diversity or expected heterozygosity (H
e
) similar
to the latter (H
e
= 0.394). This is attributed to the fact that many alleles found in pre-selected accessions
were present at very low frequencies (mean effective number of alleles, A
e
= 1.72). A broader range of alleles
detected in the pre-selected accessions suggests that pre-selection of accessions screened from isozyme data
may be useful for selecting germplasm collections with a greater number of RFLP alleles. The central Chi-
nese accessions maintained a significantly higher level of RFLP genetic diversity than the NAC (H
e
= 0.405,
A = 2.50 for central China vs. H
e
= 0.339, A = 2.08 for the USA). We detected significant genetic differentia-
tion among the six provinces (mean G
ST
= 0.133). These results suggest that Chinese germplasm accessions
from various regions or provinces in the USDA germplasm collection could be used to enhance the genetic
diversity of US. cultivars.
Keywords: Genetic variation; Glycine max; Pre-selection; RFLPs; USDA soybean germplasm collection.
INTRODUCTION
The low level of genetic diversity within US soybean
cultivars has brought more attention to the use of plant
introductions or accessions from other countries, in partic-
ular, the regions where the crop evolved. Plant introduc-
tions, although often agronomically undesirable, can be
used as parents to enhance genetic diversity. Methods to
quickly evaluate plant introductions maintained in the Na-
tional Plant Germplasm System for diversity may increase
the use of the collection by plant breeders as parents to im-
prove modern cultivars. Efforts have been made towards
identifying "core collections" (Brown, 1989). The goal
of core collections is to maximize genetic variability in a
smaller, but representative, group of accessions (Crossa et
al., 1993; Bretting and Weidrlechner, 1995; Wang et al.,
1998).
Soybean [Glycine max (L.) Merr.] is a crop of major
economic importance in China, Japan, Korea, and North
and South America. China is considered to be the soybean
center of origin and the center of diversity (Smartt and
Hymowitz, 1985). Therefore, soybean accessions from
China should include novel genetic diversity and may
be a rich source of alleles from which to identify a core
collection. One concern in evaluating new accessions
of the USDA soybean germplasm collection is how to
most efficiently detect genetic diversity. One approach
is to determine if use of isozyme data can increase the
efficiency of selection of plant introductions. If pre-
selected accessions of Chinese soybean germplasm based
upon root tip fluorescence (Torkelson and Palmer, 1997),
pubescence tip morphology of legumes, and isozyme
data (Liao and Palmer, 1997a, b, c, d) have a wider range
of variation than randomly selected accessions, it would
be predicted that measures at the DNA level maintain
more alleles in pre-selected accessions than those from
randomly selected accessions.
MOLECULAR BIOLOGY
pg_0002
14
Botanical Studies, Vol. 47, 2006
Allozyme markers have been used in soybean to
evaluate genetic diversity in accessions from diverse
geographic regions (Kiang et al., 1987; Perry et al., 1991;
Griffin and Palmer, 1995; Yeeh et al., 1996; Chung et
al., unpublished data). Since many selectively neutral
DNA markers for soybean are currently available, several
studies measuring genetic relationships at the DNA-level
of soybean introductions (accessions from China) with
the ancestors in the U.S. have been conducted to obtain
information useful to breeding programs for selection of
diverse parents (e. g., Keim et al., 1989, 1992; Lorenzen
and Shoemaker, 1996; Brown-Guedira et al., 2000; Li et
al., 2001). Restriction fragment length polymorphisms
(RFLPs) provide a very large number of genetic markers
for detecting and analyzing genetic diversity in plants (e.
g., Helentjaris et al., 1985; Zhang et al., 1993; Dubreuil
and Charcosset, 1998). RFLPs have been used for this
purpose in soybean (e. g., Grabau et al., 1989; Lorenzen
et al., 1995; Kisha et al., 1997). Since RFLPs can detect
variation in both coding and non-coding regions, genetic
variation was greater for RFLPs than for allozymes in
several studies of crop plants (Messmer et al., 1991;
McGrath and Quiros, 1992; Zhang et al., 1993; Dubreuil
and Charcosset, 1998). For example, Dubreuil and
Charcosset (1998) analyzed ten populations of maize (Zea
mays L.) from European and north U.S. germplasm and
detected that the mean number of alleles per locus (A =
6.3) and total genetic diversity (H
e
= 0.60) for RFLPs were
higher than those for allozymes (A= 2.4 and H
e
= 0.23).
Most of the ancestors of U.S. soybean cultivars were
introduced from China in the early part of the 20th
century. In the USA, over 400 publicly released cultivars
were developed from approximately 80 soybean ancestral
lines (Gizlice et al., 1994). Twenty-eight introductions
(ancestors) and seven first progenies (U.S. - developed
cultivars with uncertain pedigrees) have contributed over
95% of the genes in public cultivars released between
1947 and 1988 (Gizlice et al., 1994). This strongly
suggests that the number of ancestors that constitute the
genetic base of soybean breeding programs in the USA is
very limited. Considering the smaller number of the U.S.
ancestors (64) relative to the 348 Chinese ancestral lines
(Cui et al., 1999), we predict that Chinese germplasm
collections should harbor higher levels of genetic variation
than those for the U.S. ancestors. To date, little is known
about quantification of the DNA-level genetic diversity
using standard genetic parameters (Berg and Hamrick,
1997; Hamrick and Godt, 1997) in the Chinese soybean
accessions at the provincial level and in the U.S. ancestors.
This information also would aid soybean breeders in
selecting parents to enhance the performance of future
soybean cultivars in the USA.
In this study, we analyzed levels of RFLP variation
in the random versus pre-selected subsets to test the first
prediction. Next we compared our data with Lorenzen
et al. (1995) who analyzed RFLP diversity in 64 soybean
lines including the ancestors and milestone cultivars in the
USA to test the second prediction.
MATERIALS AND METHODS
Plant materials
Seeds of the accessions used in this study were
obtained from the USDA Soybean Germplasm Collection
courtesy of Dr. R. Nelson (USDA-ARS, University of
Illinois, Urbana). Two subsets of this collection originated
from six provinces in central China (Gansu, Hebei, Henan,
Jiangsu, Shaanxi, and Shanxi): the first was selected
at random and the second was pre-selected based on
allozyme diversity (Liao and Palmer, 1977a, b, c, d), root
tip fluorescence (Torkelson and Palmer, 1997), and the
pubescence tip morphology of the legumes. These subsets
consisted of 43 accessions for random selection (Gansu, 8;
Hebei, 2; Henan, 8; Jiangsu, 5; Shaanxi, 17; and Shanxi,
3) and 53 pre-selected accessions (Gansu, 18; Hebei, 1;
Henan, 3; Jiangsu, 5; Shaanxi, 2; and Shanxi, 24) (Table
1). An additional 11 accessions (all pre-selected) from
three other provinces (Anhui, 7; Ningxi, 1; and Shandong,
3) were included in this study (thus a total of 107; Table
1) to compare our accessions with the ancestral lines and
milestone cultivars (n = 64) studied by Lorenzen et al.
(1995).
DNA preparation and RFLP probe analysis
Seeds were planted in a greenhouse and grown to
the second leaf stage. They were then collected and
freeze-dried. Total DNA was extracted from the leaves
using a chloroform extraction method (Sambrook et
al., 1989). The DNA was digested with five restriction
endonucleases, DraI, EcoRI, EcoR V, HindIII, and Ta q I .
Restriction endonuclease digestions, electrophoresis,
Southern transfer, and DNA hybridizations were
performed as described by Keim et al. (1990). Genomic
DNA probes were screened against these genotypes with
probe-enzyme combinations that were identical to those
used in the preparation of the USDA-ARS public soybean
Glycine max by G. soja Sieb. Et Zucc. (A81-356022 X PI
468916) genetic map (Shoemaker and Specht, 1995). As
controls, G. max breeding line A81-356022 and G. soja
plant introduction PI 468916 were included for RFLP
probe analysis to consistently score banding patterns.
Data analysis
The RFLP banding patterns were scored according to an
allele-locus model as suggested by Bruebaker and Wendel
(1994). The low copy number of RFLP fragments seen, as
often observed in highly homozygous species, facilitates
the use of the allele-locus model. Bands (alleles) at a
given locus that were difficult to score were recorded as a
"no score." A locus was considered to be polymorphic if
at least one genotype had a mapped fragment that differed
from the remaining genotypes. Loci that were difficult to
score or had a high frequency of missing data points were
discarded. A total of 46 probes were used for the analysis
of the 107 accessions.
We estimated three genetic parameters to determine
pg_0003
CHUNG et al. — RFLP variation in USDA soybean germplasm
15
Table 1. Soybean accessions from the central provinces of China assayed for RFLP diversity. The accessions were selected at
random or were pre-selected based upon isozyme diversity.
PI
a
Province SM
b
PI
a
Province SM
b
PI
a
Province SM
b
567287 Gansu
P
567428
Shanxi
P
567394A Shaanxi
R
567292 Gansu
P
567436
Shanxi
P
567396C Shaanxi
R
567294 Gansu
P
567440
Shanxi
P
567403B Shaanxi
R
567295 Gansu
R
567443
Shanxi
R
567406A Shaanxi
R
567297 Gansu
P
567446
Shanxi
P
567410B Shaanxi
R
567304 Gansu
P
567446
Shanxi
P
567660A Henan
R
567305 Gansu
P
567459
Shanxi
P
567682A Henan
R
567312 Gansu
R
567462
Shanxi
P
567684B Henan
R
567314 Gansu
P
567462
Shanxi
P
567765C Jiangsu
R
567322 Gansu
P
567467
Shanxi
P
567299A Gansu
P
567327 Gansu
P
567468
Shanxi
R
567299B Gansu
P
567329 Gansu
R
567478
Shanxi
R
567316A Gansu
R
567330 Gansu
P
567502
Hebei
P
567336A Gansu
P
567332 Gansu
P
567509
Hebei
R
567336B Gansu
P
567340 Gansu
R
567517
Hebei
R
567417A Shanxi
P
567342 Gansu
P
567565
Shandong P
567417B Shanxi
P
567345 Gansu
R
567613
Henan
R
567417C Shanxi
P
567348 Gansu
P
567635
Henan
P
567419A Shanxi
P
567353 Gansu
P
567637
Henan
R
567419B Shanxi
P
567356 Gansu
R
567657
Henan
R
567433A Shanxi
P
567365 Ningxi
P
567663
Henan
R
567433B Shanxi
P
567380 Shaanxi R
567670
Henan
R
567441A Shanxi
P
567383 Shaanxi R
567673
Henan
P
567441B Shanxi
P
567386 Shaanxi
R
567704
Anhui
P
567441C Shanxi
P
567391 Shaanxi
P
567707
Anhui
P
567466A Shanxi
P
567393 Shaanxi
R
567715
Anhui
P
567466B Shanxi
P
567397 Shaanxi
R
567725
Anhui
P
567554A Shandong P
567399 Shaanxi
P
567733
Anhui
P
567554B Shandong P
567400 Shaanxi
R
567750
Jiangsu
R
567630A Henan
P
567402 Shaanxi
R
567763
Jiangsu
R
567706A Anhui
P
567405 Shaanxi
R
567766
Jiangsu
R
567706B Anhui
P
567408 Shaanxi
R
567773
Jiangsu
R
567756A Jiangsu
P
567412 Shaanxi
R
567776
Jiangsu
P
567756B Jiangsu
P
567420 Shanxi
P
567349A Gansu
R
567780A Jiangsu
P
567421 Shanxi
P
567376A Shaanxi R
567780B Jiangsu
P
567423 Shanxi
P
567381B Shaanxi R
　
　
　
a
PI, plant introduction numbers. For comparison between pre-selected and random selection, PIs from Ningxi (n =1), Shandong (3),
and Anhui (7) were excluded because they were all pre-selected accessions. For the genetic diversity analysis at the provincial
level, seven PIs from three provinces [Hebei (3), Ningxi (1), and Shandong (3)] were excluded due to very small sample sizes.
b
Selection method: P, pre-selected and R, randomly selected.
pg_0004
16
Botanical Studies, Vol. 47, 2006
levels of RFLP diversity using the programs POPGENE
(Yeh et al., 1999) and FSTAT (ver.2.9.3 by Goudet, 2002):
mean number of alleles per locus (A), mean effective
number of alleles per locus (A
e
), and Nei’s (Nei, 1978)
unbiased gene diversity (H
e
). Estimates in 43 random
accessions were compared with 53 pre-selected accessions
from six provinces. For the genetic diversity analysis at
the provincial level, seven accessions from three provinces
(Hebei, 3; Ningxi, 1; Shandong, 3) were excluded due to
small sample sizes. Thus, we estimated genetic diversity
(A and H
e
) of soybean accessions in six Chinese provinces
(Anhui, Gansu, Henan, Jiangsu, Shaanxi, and Shanxi).
Nei’s (1973) gene diversity (H
eN
= 1 – Σp
i
2
, whereas p
i
is
the frequency of the ith RFLP allele per locus) also was
estimated to compare total diversity in 107 accessions
with 64 accessions of the U.S. ancestral lines including
milestone cultivars studied by Lorenzen et al. (1995).
Lorenzen et al. (1995) estimated gene diversity per locus
(H
eN
) using the Nei’s (1973) formula.
We calculated Nei’s (1973) G
ST
(proportion of total
genetic diversity partitioned among provinces in China) at
each locus, and then averaged them across 46 RFLP loci to
determine degree of genetic differentiation among the six
provinces. In addition, we tested statistical significance
for genetic differentiation among the six provinces using
the exact test of Raymond and Rousset (1995). This test
is analogous to Fisher’s exact test but uses a Markov
chain to explore all potential states of an r × k contingency
table based on r groups and k genotypes. This test was
conducted using the program ARLEQUIN (Schneider et
al., 2000) and 10000 Markov steps.
RESULTS
The 107 accessions were polymorphic for all 46 RFLP
loci and a total of 115 alleles were detected across the loci,
giving rise to high levels of genetic diversity (mean A = 2.5
and mean H
e
= 0.410) (Table 2). The USDA germplasm
collections from central China harbor significantly higher
levels of genetic diversity than those for 21 ancestors and
derived cultivars in the USA (mean A = 2.1 and mean
H
eN
= 0.339) (Wilcoxon signed rank test statistic: for A,
z = -3.139, one-tail probability, P = 0.001 and for H
e
, z =
-2.305, P = 0.010).
Fifty-three pre-selected accessions (mean A = 2.44) had
significantly more alleles than the 43 randomly selected
accessions (mean A = 2.13) (z = -2.982, P = 0.001), but
H
e
was not significantly different between the two groups
(mean H
e
= 0.393 and 0.394) (Table 2). For pre-selected
accessions, mean effective number of alleles per locus
(A
e
) was only 1.72, which highlights the fact that many of
the alleles were present at very low frequencies. This is a
reason that the pre-selected accessions maintain nearly the
same levels of H
e
and A
e
as randomly selected accessions
(Table 2), though the former had significantly more alleles
than the latter.
At the provincial level, RFLP diversity (A and H
e
)
ranged from 1.64 and 0.263 (Province Anhui) to 2.26
(Province Shanxi) and 0.423 (Province Henan) with means
of 2.06 and 0.368 (Table 3). There was a significant
difference among the six provinces for H
e
(Kruskal-Wallis
test statistic: H = 15.2, P = 0.009), which is primarily due
to the low estimate in Anhui. Spearman rank correlation
analysis revealed that the mean number of alleles per locus
(A) is closely associated with the number of accessions
representing each province (r
S
= 0.886, P = 0.019). This
suggests that if sample size is increased in Province
Anhui, we would expect more alleles per locus. However,
no significant correlation between sample size and H
e
was
detected
(r
S
= 0.086).
There were significant differences in allele frequencies
among the six provinces (mean G
ST
= 0.133). Overall,
about 87% of the total variation in the samples was
common to the six provinces in China. In addition,
the exact test of genetic differentiation among the six
provinces was highly significant (P < 0.0001).
DISCUSSION
Genetic diversity between random accessions
vs. pre-selected accessions
Our first prediction was that pre-selected accessions
of Chinese soybean germplasm collections would have
significantly more alleles than those from randomly
selected accessions. This suggests that pre-selection of
accessions based on allozyme data may be an effective
approach for selecting germplasm collections with more
RFLP alleles. However, we failed to detect a significant
difference for gene diversity (expected heterozygosity,
H
e
) between pre-selected and randomly chosen accessions
owing to the very low frequencies of several alleles in the
pre-selected accessions. Gene diversity (H
e
) calculated
in this study is a composite measure that summarizes
genetic variation at the locus level. The maganitude of
H
e
is a function of the proportion of polymorphic loci, the
number of alleles per polymorphic locus, and the evenness
of allele frequencies within populations or accessions. It
is the most commonly used index of genetic diversity for
codominant data, because it summarizes the fundamental
genetic variation of a population in a single statistic (Berg
and Hamrick, 1997).
Some propose that decisions for gene conservation of
crops and their wild relatives should be based to a greater
degree on "allelic richness" (AR) (number of alleles)
(e.g., Marshall and Brown, 1975; Lee, 1998; Brown and
Brubaker, 2000). In complementary analysis, thus, we
calculated RFLP allelic richness between random and pre-
selected accessions through rarefaction that accounts for
sample size effects (Leberg, 2002) to produce unbiased
estimates using the FSTAT (Goudet, 2002). A very similar
to the mean number of alleles per locus, estimates of
allelic richness revealed that 53 pre-selected accessions
(mean AR, allelic richness = 2.26) had significantly more
alleles than the 43 randomly selected accessions (mean AR
pg_0005
CHUNG et al. — RFLP variation in USDA soybean germplasm
17
Table 2. Summary of RFLP diversity for 46 probes (loci) in central Chinese accessions and ancestors and milestone cultivars (n =
64) in the USA.
Pre-selected
Random
Total sample
NAC
e
A
a
A
e
b
H
e
c
A
a
A
e
b
H
e
c
A
a
H
e
c
H
eN
d
A
a
H
eN
d
A023
3 1.06 0.081 2 1.06 0.057 3 0.070 0.070 2 0.47
A063-1 2 1.90 0.483 2 2.00 0.511 2 0.500 0.494 2 0.28
A063-2 3 2.05 0.519 3 1.97 0.500 3 0.507 0.504 n.a. n.a.
A085
3 2.39 0.593 2 1.96 0.505 3 0.573 0.567 2 0.50
A086
2 2.00 0.510 2 1.76 0.444 2 0.493 0.488 2 0.36
A095
2 1.90 0.485 2 1.85 0.472 2 0.474 0.469 2 0.40
A186
3 1.46 0.324 2 1.29 0.232 3 0.286 0.283 2 0.20
A257
3 1.98 0.504 2 1.78 0.450 3 0.476 0.471 2 0.45
A333
3 2.15 0.545 2 1.98 0.511 3 0.526 0.520 2 0.13
A374
3 2.04 0.522 2 2.00 0.512 3 0.531 0.526 2 0.10
A381
3 1.13 0.121 2 1.44 0.315 3 0.205 0.201 2 0.38
A398
2 1.97 0.503 2 2.00 0.515 2 0.504 0.498 2 0.31
A401-1 2 1.95 0.495 2 1.94 0.497 2 0.505 0.500 2 0.48
A401-2 2 1.93 0.493 2 1.98 0.508 2 0.504 0.499 n.a. n.a.
A461-1 2 1.89 0.482 2 1.72 0.431 2 0.458 0.453 2 0.48
A461-2 2 1.08 0.074 2 1.05 0.049 2 0.062 0.062 n.a. n.a.
A481
3 2.15 0.543 2 1.82 0.463 3 0.512 0.505 2 0.06
A505
2 1.76 0.439 2 1.86 0.477 2 0.499 0.493 2 0.45
A520
2 1.04 0.041 2 1.39 0.286 2 0.153 0.151 2 0.06
A567
2 2.00 0.510 2 1.66 0.409 2 0.490 0.484 2 0.28
A586-1 2 1.19 0.164 2 1.67 0.414 2 0.267 0.263 2 0.47
A586-2 2 1.72 0.429 2 1.85 0.476 2 0.441 0.435 n.a. n.a.
A668
2 1.59 0.378 2 1.67 0.413 2 0.388 0.384 2 0.33
A681
3 2.05 0.523 3 1.71 0.427 3 0.482 0.476 2 0.43
A691-1 2 1.76 0.439 2 1.40 0.292 2 0.385 0.381 3 0.47
A691-2 2 1.96 0.500 2 1.88 0.481 2 0.488 0.483 n.a. n.a.
A702
3 1.93 0.491 2 1.83 0.468 4 0.537 0.531 2 0.50
A708
2 1.97 0.503 2 1.69 0.440 2 0.508 0.500 2 0.46
A806
2 1.40 0.294 3 1.71 0.427 3 0.352 0.349 2 0.33
A816
3 2.09 0.531 2 1.36 0.272 3 0.449 0.444 2 0.26
A847
1 1.00 0.000 1 1.00 0.000 2 0.390 0.384 2 0.50
A890
4 2.41 0.597 3 1.80 0.456 4 0.547 0.541 2 0.20
A946-1 3 1.65 0.402 2 1.63 0.398 3 0.396 0.391 2 0.04
A946-2 3 1.60 0.382 3 1.97 0.504 3 0.430 0.426 2 0.15
A963
2 1.32 0.350 2 1.58 0.378 2 0.358 0.354 2 0.40
B039
2 1.99 0.507 2 1.97 0.505 2 0.506 0.500 2 0.40
B122
2 1.32 0.250 2 1.44 0.315 2 0.273 0.269 2 0.50
B164
3 1.09 0.081 2 1.14 0.129 3 0.098 0.098 2 0.44
B166
6 1.63 0.393 4 1.86 0.475 6 0.429 0.424 3 0.33
K002
2 1.63 0.394 2 1.58 0.378 2 0.382 0.378 2 0.44
K003
2 1.89 0.472 2 1.33 0.258 2 0.499 0.493 2 0.40
K007
2 1.48 0.332 2 1.65 0.405 2 0.361 0.357 3 0.57
K069-1 2 1.60 0.384 2 1.98 0.505 2 0.483 0.478 2 0.24
K070-1 2 1.41 0.296 2 1.67 0.412 2 0.351 0.347 2 0.10
K070-2 2 1.26 0.212 2 1.32 0.246 2 0.225 0.223 n.a. n.a.
R017
2 1.97 0.502 2 1.98 0.508 2 0.560 0.500 2 0.30
Mean 2.44 1.72 0.393 2.13 1.68 0.394 2.50 0.410 0.405 2.08 0.34
1±SE 0.12 0.06 0.024 0.07 0.04 0.020 0.12 0.019 0.019 0.04 0.02
a
Mean number of alleles per locus or probe. n.a. represents data not available.
b
Mean effective number of alleles per locus.
c
Nei’s (1978) unbiased gene diversity.
d
Nei’s (1973) biased gene diversity.
e
NAC, ancestors and milestone cultivars in the USA; H
e N
estimates from Lorenzen et al. (1995).
pg_0006
18
Botanical Studies, Vol. 47, 2006
= 2.10) (z = -2.812, P = 0.002). If several low frequency
alleles that are not observed in randomly selected
accessions are closely linked with other agronomically
important traits, our approach to pre-selection would be
a useful tool for developing core collections. In other
words, low frequency alleles may not effect allelic
evenness measures (e.g., expected heterozygosity or H
e
),
but could provide significant affects when incorporated
into a plant breeding program.
The 107 soybean accessions from central China harbor
comparable levels of genetic diversity (mean A = 2.5 and
mean H
e
= 0.410) relative to other crops. For example,
similarly high levels of RFLP genetic diversity for 35
probe-restriction enzyme combinations were observed
in ten maize populations of European and northern U.S.
germplasm. Dubreuil and Charcosset (1998) estimated
3.5 of the mean number of alleles per locus and 0.470 of
Nei’s (1978) unbiased genetic diversity in the ten maize
populations.
Genetic diversity between central China vs.
ancestors and derived cultivars in the USA
Our second prediction was that the accessions from
central China harbor significantly higher levels of genetic
diversity than those for 21 ancestors and derived cultivars
in the USA. Considering the fact that the 43 non-ancestral
U.S. cultivars studied by Lorenzen et al. (1995) trace
back to just 19 of the earliest plant introductions used in
breeding programs in the USA, and that among the 19
ancestral lines, only 12 account for 88% of the germplasm
in the derived 43 cultivars, our results are not surprising.
Keim et al. (1992) analyzed 38 soybean accessions in the
USA (18 ancestral lines and 20 adapted lines) with 132
RFLP probes and found 69% polymorphism in the probes
and a very similar estimate (H
eN
= 0.30) to Lorenzen et al.
(1995). Skorupska et al. (1993) identified 29 RFLP probes
with gene diversity estimate . 0.30 in elite southern U.S.
cultivars and ancestral lines of those cultivars. All these
studies suggest that the USDA Soybean Germplasm
Collection from China harbors significantly higher levels
of genetic diversity than those for the U.S. ancestral lines
and their derived cultivars.
More recently, three of the authors analyzed allozyme
variation in 1777 accessions in the USDA Soybean
Germplasm Collection taken from 19 provinces in China
and estimated high levels of genetic diversity (H
e
= 0.240,
Chung et al., unpublished data). There was no significant
difference for H
e
between the 83 soybean ancestors both in
the U.S. and China and the 1777 accessions from China,
suggesting that the ancestral lines in the U.S. and China
also harbor high levels of allozyme diversity. However,
our RFLP data revealed a significant difference between
accessions from central China and the U.S. ancestors
and their derived cultivars, suggesting that data from the
RFLPs are more informative than those from allozyme
data for evaluation of the USDA Chinese Germplasm
Collection. Again, previous studies by simple sequence
repeats (SSRs) revealed high levels of genetic diversity in
Asian soybean accessions. Abe et al. (2003) analyzed 20
SSR loci in 131 Asian soybean accessions from 14 Asian
countries and found high levels of genetic diversity (mean
H
e
= 0.782) in the accessions. To determine if the Asian
accessions harbor significantly higher levels of genetic
diversity than those of the U.S. ancestors (Diwan and
Cregan, 1997) and their derived elite lines (Narvel et al.,
2000), we selected the same SSR loci (five loci for each
comparison) for pairwise comparisons between studies
of Abe et al. (2003) and Diwan and Cregan (1977) and
Narvel et al. (2000). The 131 Asian accessions harbored
significantly larger number of alleles per locus and higher
levels of genetic diversity than those of the 35 ancestors in
the U.S. (Diwan and Cregan, 1997) (mean A = 12.6 vs. 8.8,
Wilcoxon signed rank test: z = -2.023, P = 0.022, one tail
probability; H
e
= 0.842 vs. 0.772, z = -1.753, P = 0.040).
Similarly, the Asian accessions maintained significantly
higher levels of genetic diversity than those of 39 U.S.
elite soybeans (Narvel et al., 2000) (mean A = 12.8 vs.
3.0, Wilcoxon signed rank test: z = -2.023, P = 0.026,
one tail probability; H
e
= 0.841 vs. 0.426, z = -2.023, P =
0.022). Thus, the present RFLP and previous SSR studies
consistently revealed that Chinese soybean accessions and
other Asian accessions harbor significantly higher levels
of genetic diversity than those of the U.S. ancestors and
their derived cultivars.
We noted four points resulting from this study that
might be interesting to soybean breeders in both the USA
and China. First, our results are consistent with previous
studies that found levels of the genetic variation were
greater for RFLPs than for allozymes in a number of
crop plants (Messmer et al., 1991; McGrath and Quiros,
1992; Zhang et al., 1993; Dubreuil and Charcosset, 1998).
Second, comparing our results with Lorenzen et al. (1995),
the probes used in this study were clearly effective at
detecting overall diversity in the Chinese germplasm
Table 3. Sum m ary of R FLP dive rsi ty for 46 loci in s ix
provinces in central China.
Province
a
N
b
A (SE)
c
H
e
(SE)
d
Anhui
7 1.64 (0.08) 0.263 (0.035)
Gansu
26 2.15 (0.09) 0.405 (0.026)
Henan
11 2.04 (0.07) 0.423 (0.029)
Jiangsu
10 2.07 (0.07) 0.409 (0.026)
Shaanxi 19 2.21(0.07) 0.353 (0.025)
Shanxi
27 2.26 (0.09) 0.355 (0.028)
Mean
16.7 2.06 (0.09) 0.368 (0.024)
a
Seven accessions from three provinces (Hebei, 3, Ningxi, 1,
and Shandong, 3) were not included due to small sample sizes.
b
Sample size.
c
Mean number of alleles per locus or probe. SE, standard error.
d
Nei’s (1978) unbiased gene diversity.
pg_0007
CHUNG et al. — RFLP variation in USDA soybean germplasm
19
collections. Third, as pre-selected accessions of Chinese
soybean germplasm collections have significantly more
alleles than those from randomly selected accessions,
allozymes have proven to be an effective tool for
identifying an RFLP diverse (in terms of the number of
alleles across loci) germplasm collection, though less
diversity was detected with allozymes. Finally, we found
a significant difference in levels of genetic diversity
among the six provinces in central China and also detected
significant differences in allele frequencies among the
six provinces. Collectively, all these results suggest that
Chinese germplasm collections from various regions or
provinces are important to enhancing the genetic diversity
of the current soybean cultivars in the US.
In summary, pre-selected accessions based on allozyme
diversity have significantly more alleles than randomly
selected accessions. Our results suggest that the use of
prior data to select genetically diverse germplasm in terms
of the number of alleles would be effective. This may
be a helpful tool for breeders interested in working with
germplasm collections to select for unique traits. It is our
hope that methods to identify small but diverse sets of the
USDA Soybean Germplasm Collection will encourage
breeders in the future to use these accessions in their
breeding programs.
Acknowledgements. This research was in part supported
by a grant (USB Project # 9211: Identification and
utilization of exotic germplasm to improve soybean
productivity) from the United Soybean Board to the
USDA. Eric Myers read earlier versions of the manuscript
and made helpful suggestions. Mention of a trademark,
proprietary product, or vendor does not constitute a
guarantee or warranty of the product by the USDA or by
Iowa State University and does not imply its approval to
the exclusion of other products or vendors that also may
be suitable. M. G. Chung gratefully acknowledges the
exchange program sponsored by the USDA/FAS/ICD/
Research and Scientific Exchange Division from June
2001 to September 2002.
LITERATURE CITED
Abe, J.D., H. Xu, Y. Suzuki, and A. Kanazawa. 2003. Soybean
germplasm pools in Asia revealed by nuclear SSRs. Theor.
Appl. Genet. 106: 445-453.
Berg, E.E. and J.L. Hamrick. 1997. Quantification of genetic
diversity at allozyme loci. Can. J. For. Res. 27: 415-424.
Bretting, P.K. and M.P. Widrlechner. 1995. Genetic m arkers
and plant genetic resource management. In J. Janick (ed.),
Plant Breeding Reviews, Vol. 13, John Wiley and Sons,
Indianapolis, pp. 11-86.
Brown, A.H.D. 1989. Core collections: a practical approach to
genetic resources management. Genome 31: 818-824.
Brown, A.H.D. and C.L . Brubaker. 2000. Genetics and the
conservation and use of Australian wild relatives of crops.
Aust. J. Bot. 48: 297-303.
Brown-Guedira, G.L., J.A. Thompson, R.L. Nelson, and M.L.
Warburton. 2000. Evaluation of genetic diversity of soy-
bean introductions and North American ancestors using
RAPD and SSR markers. Crop Sci. 40: 815-823.
Bruebaker, C.L. and J.F. Wendel. 1994. Reevaluation of the
origin of domesticated cotton (Gossypium hirsutum: Mal-
vaceae) using nuclear restriction fragment length polymor-
phisms (RFLPs). Amer. J. Bot. 81: 1309-1326.
Crossa, Z., C.M. Hernandez, P. Bretting, S.A. Eberhart, and S.
Taba. 1993. Statistical genetic considerations for main-
taining germplasm collections. Theor. Appl. Genet. 86:
673-678.
Cui, Z., T.E. Carter Jr., J. Cai, J. Qui, and R.L. Nelson. 1999.
Origin, description, and pedigree of Chinese soybean culti-
vars released from 1923 to 1995. Washington: U.S. Depart-
ment of Agriculture, Agricultural Research Service, Tech.
Bull. No. 1871.
Diwan, N. and P.B. Cregan. 1997. Automated sizing of fluo-
rescent-labeled simple sequence repeat (SSR) markers to
assay genetic variation in soybean. Theor. Appl. Genet. 95:
723-733.
Dubreuil, P. and A. Charcosset. 1998. Genetic diversity within
and among ma ize populations: a comparis on between
is ozyme and nuclear RFLP loci. Theor. Appl. Genet. 96:
577-587.
Gizlice, Z., T.E. Carter Jr., and J.W. Burton. 1994. Genetic base
for North American public soybean cultivars released be-
tween 1947 and 1988. Crop Sci. 34: 1143-1147.
Goudet, J. 2002. FSTAT, a program to estimate and test gene di-
versities and fixation indices (ver. 2.9.3.2). Dorigny, Swit-
zerland: Universite de Lausanne. Available at http://www.
unil.ch/izea/softwares/fstat.html. Updated from Goudet
(1995).
Graba u, E.A., W.H. Davi s, a nd B.G. Ge ngenb ach. 198 9.
Restriction fragment length polymorphism in a subclass
of the "Manda rin" s oybean cytopl asm . Crop Sc i. 29:
1554-1559.
Griffin, J.D. and R.G. Palmer. 1995. Variability of thirteen iso-
zyme loci in the USDA s oybean germplasm collections.
Crop Sci. 35: 897-904.
Hamrick, J.L. and M.J.W. Godt. 1997. Allozyme diversity in
cultivated crops. Crop Sci. 37: 26-30.
Helentjaris, T., G. King, M. S locum, C. Siedenstrang, and S.
Wegman. 1985. Res triction fragment polymorphis ms as
probes for plant diversity and their development as tools for
applied plant breeding. Plant Mol. Biol. 5: 109-118.
Keim, P., R.C. Shoemaker, and R.G. Palmer. 1989. Restriction
fragment length polymorphism diversity in soybean. Theor.
Appl. Genet. 77: 786-792.
Keim, P., B.W. Diers, T.C. Olson, and R.C. Shoemaker. 1990.
RFLP mapping in soybean: association between marker loci
and variation in quantitative traits. Genetics 126: 735-742.
Keim, P., W. Beavis, J. Schupp, and R. Freestone. 1992. Evalu-
ation of soybean RFLP marker diversity in adapted germ-
plasm. Theor. Appl. Genet. 85: 205-212.
pg_0008
20
Botanical Studies, Vol. 47, 2006
Kiang, Y.T., Y.C. Chiang, J .H.Y. Doong, and M.B. Gorman.
1987. Genetic variation of soybean germplasm. In S. C.
Hsieh (ed.), Crop exploration and utilization of genetic
resources, Taichung District Agricultural Improvement Sta-
tion Publisher, Changhua, Taiwan, pp. 93-99.
Kisha, T., C.H. Sneller, and B.W. Diers. 1997. Relationship be-
tween genetic distance among parents and genetic variance
in populations of soybean. Crop Sci. 37: 1317-1325.
Leberg, P. L. 2002. Estimating allelic richness: Effects of sample
size and bottlenecks. Mol. Ecol. 11: 2445-2449.
L e e, M. 19 98 . G en om e pr oj e ct s a nd g en e po ol s : Ne w
germplasm for plant breeding. Proc. Natl. Acad. Sci. USA.
95: 2001-2004.
Li, Z., L. Qiu, J.A. Thompson, M.M. Welsh, and R.L. Nelson.
2001. Molecular genetic analysis of U.S. and Chinese soy-
bean ancestral lines. Crop Sci. 41: 1330-1336.
Liao, W. and R.G. Palmer. 1997a. Genetic study of a diaphorase
2 null mutant. Soybean Genet. Newsl. 24: 157-159.
Liao, W. and R.G. Palmer. 1997b. Inheritance and linkage stud-
ies of an aconitase 1 null in soybean. S oybean Genet.
Newsl. 24: 160-163.
Liao, W. and R.G. Palmer. 1997c. Inheritance and linkage stud-
ies of 6-phosphogluconate dehydrogenase in soybean. Soy-
bean Genet. Newsl. 24: 164-167.
Liao, W. and R.G. Palmer. 1997d. A new variant of phospho-
glucose isomerase in soybean. Soybean Genet. Newsl. 24:
179-182.
Lorenzen, L.L., S. Boutin, N. Young, J .E. Specht, and R.C.
Shoemaker. 1995. Soybean pedigree analysis using map-
based molecular markers: I. tracking RFLP markers in cul-
tivars. Crop Sci. 35: 1326-1336.
Lorenzen, L.L. and R.C. Shoemaker. 1996. Genetic relationships
within soybean cultivar groups. Crop Sci. 36: 743-752.
Marshall, D.R. and A.H.D. Brown. 1975. Optimum sampling
strategies in genetic conservation. In O. H. Frankel and J.
G. Hawkes (eds.), Crop genetic resources for today and
tomorrow. Cambridge University Press, London, pp. 53-80.
McGrath, J.M. and C.F. Quiros. 1992. Genetic diversity in iso-
zyme and RFLP loci in Brassica campestris as related to
crop type and geographical origin. Theor. Appl. Genet. 83:
783-790.
Messmer, M.M., A.E. Melchinger, M. Lee, W.L. Woodman, E.A.
Lee, and K.R. Lamkey. 1991. Genetic diversity among pro-
genitors and elite lines from the Iowa Stiff Stalk Synthetic
(BSS S) maize population: comparis on of allozyme and
RFLP data. Theor. Appl. Genet. 83: 97-107.
Narvel, J.M., W.R. Fehr, W.C. Chu, D. Grant, and R.C. Shoe-
maker. 2000. Simple sequence repeat diversity among soy-
bean plant introductions and elite genotypes. Crop Sci. 40:
1452-1458.
Nei, M. 1973. Analysis of gene diversity in subdivided popula-
tions. Proc. Natl. Acad. Sci. USA 70: 3321-3323.
Nei, M. 1978. Estimation of average heterozygosity and genetic
distance from a small number of individuals. Genetics 89:
583-590.
Perry, M.C., M.S. McIntosh, and A.K. Stoner. 1991. Geographi-
cal patterns of variation in the USDA soybean germplasm
collection: II. Allozyme frequencies. Crop Sci. 31:
1356-1360.
Raymond, M. and F. Rousset. 1995. Genepop (V. 1.2): Popula-
tion genetics software for exact tests and ecumenicism. J.
Hered. 86: 248-249.
S ambrook, J., E.F. Fritsch, and T. Maniatis. 1989. Molecular
cloning: a laboratory manual. Cold Spring Harbor Press,
Cold Spring Harbor.
Schneider, S., D. Roessli, and L. Excoffier. 2000. ARLEQUIN
ver. 2000: A software for population genetic data analysis.
Geneva, Switzerland: Genetics and Biometry Laboratory,
University of Geneva. Available at http://anthro.unige.ch/
arlequin.
Shoemaker, R.C. and J.F. Specht. 1995. Integration of the soy-
bean molecular and classical genetic linkage groups. Crop
Sci. 35: 436-446.
Skorupska, H.T., R.C. Shoemaker, A. Warner, E.R. Shipe, and
W.E. Bridges. 1993. Restriction fragment length polymor-
phism in soybean germplasm of the southern U.S.A. Crop
Sci. 33: 1169-1176.
Smartt, J. and T. Hymowitz. 1985. Domestication and evolution
of grain legumes. In R. J. Summerfield and E. H. Roberts
(eds.), Grain Legumes, Collins, London, pp. 37-72.
Torkelson, J. and R.G. Palmer. 1997. Genetic analysis of root
fluorescence of central Chinese germplasm. Soybean Genet.
Newsl. 24: 147-149.
Wang, R.W., Z.L. Yang, H.S. Kim, and C. Yen. 1998. Compara-
tive analysis of RFLP diversity in landraces of Triticum
aestivum and collections of T. tauschii from China and
southwest Asia. Theor. Appl. Genet. 96: 312-318.
Yeh, F.C., R.C. Yang, and T.B.J. Boyle. 1999. POPGENE Ver-
sion 1.31, Microsoft window-based free ware for population
genetic analysis. Alberta, Canada: Centre for International
F ores try Res earc h, University of Alberta. Available a t
http://www.ualberta.ca/~fyeh/index.htm.
Yeeh, Y. S.S. Kang, and M.G. Chung. 1996. Evaluations of
the natural monument populations of Camellia japonica
(Theaceae) in Korea based on allozyme studies. Bot. Bull.
Acad. Sin. 37: 141-146.
Zhang, Q., M.A. Saghai Maroof, and A. Kleinhofs. 1993. Com-
parative diversity analysis of RFLPs and isozymes within
and among populations of Hordeum vulgare ssp. sponta-
neum. Genetics 134: 909-916.
pg_0009
CHUNG et al. — RFLP variation in USDA soybean germplasm
21
源自華中地區之美國農業部大豆種源庫之核酸限制.切割長
度多形性分析
Myong Gi CHUNG
1
, Mi Yoon CHUNG
1
, April D. Clikeman JOHNSON
2
,
and Reid G. PALMER
3
1
Department of Biology, Gyeongsang National University
Jinju 660-701, Republic of Korea
2
Department of Agronomy, Iowa State University, Ames, IA 50011, USA
3
USDA-ARS, Corn Insects and Crop Genetics Research Unit, Project 3769
Department of Agronomy, Iowa State University, Ames, IA 50011, USA
為評估源自中國中部之美國農業部大豆種源庫之基因變異度，共 107 登錄樣品之 46 基因座 (loci) 以
核酸限制.切割長度多形性 (RFLP) 分析之。我們比較逢機取樣和預先選拔兩組試驗材料之基因變異度；
比較項目有：根尖螢光，成熟期之形態，以及共 10 種酵素之同功.型。我們也同時評估此 107 源自華
中之登錄樣品之基因變異度，其方法為：和在美國先前已研究過的且具劃時代意義之栽培品種 (NAC, n
= 64)。最後，我們估算六個中國省份（安徽、甘肅、河南、江蘇、陝西、和山西）之品種的基因歧異
分化度。如以每一基因座 (locus) 所含之平均的相對基因 (allele) 數 (A, 2.44 vs. 2.13)，及相對基因之豐富
度 (2.23 vs. 2.10) 兩參數為憑，則上述預先選拔組和逢機取樣組兩者之間有顯著差異。但是，前者 (He =
0.393) 維持基因差異度（或稱期望之歧異度，He）相似於後者 (He = 0.394)。此乃因：事實上預先選拔
組所發現之很多相對基因是以低頻存在的（平均有效相對基因數，Ae = 1.72）。在預先選拔組所發現之
分佈較廣之相對基因暗示：以同功.為根據所選擇出之預先選拔組可能適用於篩選具較多數目之 RFLP
相對基因的基因庫收集樣品。源自華中之登 錄樣品 (n = 107) 維持較高水平之基因變異度（He = 0.405,
A= 2.50）當以 NAC (He = 0.339, A = 2.08) 為比較基準時。我們檢測出上述六個省份樣品間有顯著的基
因分化（平均 G
ST
= 0.133）。這些結果暗示：來自中國各地區或省份之美國農業部種源庫之登錄種源可
用來強化美國大豆品種之基因變異度。
關鍵詞：基因變異；Glycine max；預選；核酸限制.切割長度多形性；美國農業部大豆種源庫。
pg_0010