Breeding Science 54 : 389-397 (2004)
Key DNA Markers for Predicting Heterosis in F1 Hybrids of japonica Rice
Young-Il Cho1), Chan-Woong Park1), Soon-Wook Kwon1), Joong-Hyun Chin1), Hyeon-So Ji1), Ki-Jin Park1),
Susan McCouch2) and Hee-Jong Koh*1)
1)
2)
Dept. of Plant Science, College of Agriculture and Life Sciences, Seoul National University, Seoul 151-742, Korea
Dept. of Plant Breeding, Cornell University, Ithaca, NY 14853-1901, USA
Optimizing hybrid vigor (HV) has been a primary objective in hybrid crop breeding programs. The effective
use of DNA markers in this endeavor has been plagued
by inconsistent results among researchers who investigated the relationship between genetic similarity (GS)
among parents and heterosis in their F1’s. The objectives of this study were to evaluate the relationship between GS among parents and heterosis in F1 for grain
yield in japonica rice, and to formulate a strategy for
marker-based prediction of heterosis. Forty-five F1 hybrids were generated from a half-diallel design using 10
japonica rice cultivars. Parents and hybrids were grown
under medium N and 0-N fertilizers and evaluated for
grain yield. The parental varieties were genotyped with
188 SSR markers and/or 129 RAPD primers. Estimates
of GS based on all markers were not highly correlated
with heterosis. After analyzing the association of each
marker to yield heterosis, we were able to identify
markers that were significantly associated with HV in
this collection of japonica rices. The estimates of GS
based on this set of selected markers showed highly significant correlations with HV. In addition, some SSR
and RAPD markers were found to be associated with
hybrid weakness (HW). This suggests that the heterosis
observed in the F1’s could be improved by selectively
combining alleles associated with HV and eliminating
those associated with HW. A new concept for developing “key markers” is proposed and the feasibility of
predicting heterosis through estimation of GS among
parents utilizing key DNA markers is discussed.
Key Words: japonica rice, hybrid vigor, hybrid weakness, molecular marker, SSR, RAPD,
genetic similarity.
Introduction
Heterosis is the genetic expression of the superiority of
a hybrid in relation to its parents, and has been extensively
utilized in breeding programs of crops, starting with hybrid
corn in the 1920s. Predicting hybrid performance has always
Communicated by H. Ikehashi
Received May 17, 2004. Accepted September 2, 2004.
*Corresponding author (e-mail: [email protected])
been a primary objective in hybrid crop breeding programs.
It has been suggested that the genetic distance between parents is positively correlated with heterosis in the resulting F1
hybrids, but until recently, the only data available for testing
this hypothesis were based almost entirely on pedigree analysis of the parents. Recent advances in molecular marker
technology have made it possible to determine whether genetic diversity between parents at the DNA level can be used
as a tool in the analysis of relationship between GS and heterosis. Significant correlations between GS and heterosis
have been reported in rice (Ahn et al. 1998, Liu and Wu
1998), maize (Smith et al. 1990), wheat (Corbellini et al.
2002), sunflower (Cheres et al. 2000) and rapeseed (Diers et
al. 1996). On the other hand, weak or non-significant correlations have also been reported between GS and heterosis in
rice (Liu and Wu 1998, Zhang et al. 1995), maize (Dudley et
al. 1991, Joshi et al. 2001, Melchinger et al. 1990, Smith and
Smith 1992), wheat (Barbosa-Neto et al. 1996, Martin et al.
1995), soybean (Cerna et al. 1997, Majarrez-Sandoval et al.
1997) and in pearl millet (Chowdari et al. 1998). Xiao et al.
(1996) reported a highly positive correlation of yield heterosis with genetic distance based on RAPD and SSR markers
within indica × indica and japonica × japonica crosses but
not for indica × japonica crosses.
The sub-species differentiation of indica and japonica
rice is ancient and has given rise to a sterility barrier that
must be overcome if heterosis for grain yield is to be expressed in indica-japonica hybrids. Furthermore, rice breeders in temperate regions often prefer to work within the
adapted temperate japonica gene pool due to difficulties associated with delayed flowering and cold susceptibility in
indica or tropical japonica germplasm. Therefore studies of
heterosis in japonica × japonica hybrids are highly relevant
to rice improvement programs in many important rice growing areas of the world.
Previous studies have demonstrated that the relationship between molecular marker heterozygosity and hybrid
performance is highly variable in rice and that it depends on
the germplasm used and the complexity of the genetic basis
of heterosis (Saghai Maroof et al. 1997, Xiao et al. 1995,
Zhang et al. 1996). Melchinger (1999) summarized the relationship between parental GS and mid-parent heterosis in a
schematic representation suggesting that the potential application of DNA markers in hybrid breeding depended very
much on whether divergent heterotic groups had been established or not. Despite considerable research effort, hybrid breeding programs utilizing DNA markers to predict
390
Cho, Park, Kwon, Chin, Ji, Park, McCouch and Koh
heterosis have not been well established due to inconsistent
results describing the relationship between GS among parents
and heterosis. Liu and Wu (1998) identified favorable and
unfavorable SSR alleles significantly affecting yield heterosis and suggested their possible use in hybrid rice breeding.
Xiao et al. (1995) identified heterotic effects at quantitative
trait loci (QTL) associated with grain yield and yield components in an experiment with F1 hybrids derived by backcrossing recombinant inbred lines. This work suggested that
complementary dominance (dominant × dominant gene
intgeractions) was responsible for much of the heterosis
observed. More recently, Hua et al. (2003) used a similar
approach and reported heterotic effects associated with four
traits at 33 loci, including 9 loci for grain yield. Consistent
with the work by Xiao et al. (1995), they concluded that
single-locus heterotic effects and pairwise dominance by
dominance interactions could explain the genetic basis of
heterosis in an elite rice hybrid in China, Shanyou 63.
The objectives of the present study were to evaluate the
relationship between GS among O. sativa ssp. japonica parents and heterosis for grain yield in the derived F1’s, and to
formulate a “key marker” concept for predicting heterosis in
crop breeding.
Materials and Methods
Plant materials, yield evaluation and DNA preparation
Ten temperate japonica rice cultivars of Korean (K)
and Japanese (J) origin, Chucheongbyeo (K), Dongjinbyeo
(K), Hwacheong waxy (K), Ilmibyeo (K), Ilpumbyeo (K),
Jinmibyeo (K), Nagdongbyeo (K), Nipponbare (J),
Singeumobyeo (K) and Yeongnambyeo (K) were crossed in
a half-diallel design to produce 45 F1 hybrids. Parents and F1
hybrids were grown in 1999 by transplanting one plant per
hill at a distance of 30 × 15 cm in the experimental paddy
field under two levels of nitrogen fertilizer, at Seoul National
University, Korea. Each entry had two replications in randomized blocks, and each plot size was 3.3 m2. Fertilizers
were applied at the rate of N-P2O5-K2O = 100-80-80 kg/ha in
medium nitrogen (N) plots, and N-P2O5-K2O = 0-80-80 kg/ha
in no-nitrogen (0-N) plots. Field management and chemical
inputs for disease and pest control followed the standard
methods of the Experimental Farm. Thirty-six individuals
were harvested from each plot excluding marginal plants
and grain yield in rough rice was recorded at constant moisture content of 14 %. Other characters were collected from
10 plants in each plot.
For each hybrid, mid-parent heterosis (MPH) and
better-parent heterosis (BPH) were calculated as MPH =
{(F1-MP)/MP} × 100 (%) and BPH = {(F1-BP)/BP} × 100
(%), where, F1 = performance of F1 hybrid, MP = mid-parental
value, BP = better-parental value.
Leaves were collected from each parent and used for
DNA extraction according to the method described by
Causse et al. (1994).
SSR and RAPD analysis
A set of 188 rice SSR markers (Table 2) developed by
Chen et al. (1997) was used in this analysis. PCR conditions
were essentially the same as described by Chen et al. (1997).
Each 25 µL reaction contained 50 ng DNA, 5 pmole of each
primer, 0.5 units of Taq. The thermocycler profile was; 5
min 94°C, 35 cycles of 1 min at 94°C, 1 min at 55°C, 2 min
at 72°C and 5 min at 72°C for final extention. PCR products
were run on a 4 % polyacrylamide denaturing gels and marker bands were revealed using the silver-staining protocols as
described by Panaud et al. (1996).
A total of 129 random decamer primers (Operon Tech.)
were used for RAPD analysis of the parents in rice (Table 2).
Amplification reactions were in a final volume of 25 µL containing 20 ng of genomic DNA, 10 pmole primer, 2.5 mM
dNTP 2 µL, 1U of Taq DNA polymerase, 10X buffer 2.5 µL
and 15 mM MgCl2 2 µL. Samples were amplified through 35
cycles of 1 min at 94°C, 2 min at 37°C, 2 min at 72°C, and
followed by a final extension at 72°C for 10 min in a PTC100TM Programmable Thermal Controller (MJ Research).
Amplified products were resolved by electrophorsis in 1.4 %
agarose gels.
Statistical analysis
Genetic similarities (GS) were estimated for all pairwise combinations of parents based on the patterns of amplicon size for SSR and the presence/absence of amplicons
for RAPD markers applying the method described by Nei
and Li (1979). Cluster diagrams of parents were constructed
using an unweighted pair-group method with arithmetical
mean (UPGMA) employing the software package, NTSYSpc, version 2.0 (Rohlf 1997). Allelic similarities and differences were evaluated between each pair of parents for all
188 SSR markers and all scorable amplicons produced by
the 129 RAPD primers. One-way ANOVA was then used to
identify SSR and RAPD markers which were significantly
(P < 0.05) associated with HV or HW by comparison of F1’s
having same alleles (homozygous pattern) with those having
different alleles (heterozygous pattern) at each locus. In addition, epistasis analysis between two loci associated with
hybrid vigor was performed using a software ‘Statistix 7’
(Analytical Software; http://www.statistix.com/).
Results
Mean and ranges of hybrid performance and heterosis
in the 45 F1 hybrids grown under two nitrogen levels are
shown in Table 1. Among the yield components, spikelets
per panicle showed the highest level of heterosis under both
nitrogen levels, while spikelet fertility showed the least.
Parental and F1 yields averaged 5.13~6.95 ton/ha and
5.58~7.64 ton/ha, respectively, in medium nitrogen plots,
and 2.79~3.58 ton/ha and 3.40~4.60 ton/ha in 0-N plots.
Mid-parent heterosis (MPH) for grain yield was 6.1 % in medium nitrogen plots and 29.9 % in 0-N plots. Heterosis for
yield and yield components was much higher in 0-N plots
Key DNA markers for predicting heterosis in japonica rice
than in medium N plots.
Of the SSR’s, 74 (39.4 %) out of 188 markers detected
polymorphism among the japonica parents, and the average
number of alleles was 2.6 per marker. The 129 RAPD markers generated 543 scorable amplicons and 84 (65.1 %) primers produced at least one polymorphic band among the parents (Table 2). GS estimates between parental varieties were
391
0.734~0.930 using the 188 SSR markers (Fig. 1-1),
0.845~0.949 using all 543 amplicons scored for the 129
RAPD primers (Fig. 1-5), and 0.841~940 using both SSR
markers and RAPD amplicons together (Fig. 1-9). The two
types of markers gave different clustering patterns and SSR
analysis provided greater resolution in estimating GS among
cultivars than RAPD analysis. This is largely due to the fact
Table 1. Mean and ranges of hybrid performance and heterosis in the 45 F1 hybrids among 10 japonica varieties grown under medium N
(N-P2O5-K2O = 100-80-80 kg/ha) and 0-N (N-P2O5-K2O = 0-80-80 kg/ha) fertilizers (MPH: mid-parent heterosis, BPH: better-parent
heterosis)
Medium N
Trait
Performance
Days to heading (day)
Culm length
(cm)
Panicle length
(cm)
Panicles per
plant (No.)
Spikelets per
panicle (No.)
Spikelet fertility (%)
1000-grain wt.
(gr.)
Harvest index
Grain yield
(ton/ha)
MPH (%)
Mean
Range
Mean
Range
116
103~
120
79~
102
20.4~
24.2
9.8~
13.6
112~
180
92~
99
24.4~
30.0
0.44~
0.56
5.58~
7.64
6.2
−22.6~
36.4
1.7~
20.3
0.1~
12.2
−13.7~
3.9
0.7~
26.7
−3.7~
3.1
−5.3~
8.1
−2.2~
16.3
−3.8~
18.7
87
21.9
11.2
138
97
26.8
0.49
6.47
7.4
5.0
−4.8
9.0
0.7
1.9
3.6
6.1
0-N
BPH (%)
Mean
−11.1
2.8
2.5
−11.1
−1.6
−0.3
−1.8
−2.2
−0.4
Performance
MPH (%)
Range
Mean
Range
Mean
Range
−34.7~
3.3
−4.4~
15.1
−5.5~
9.5
−22.2~
1.2
−17.4~
20.1
−5.9~
2.5
−9.2~
6.1
−13.7~
10.4
−10.5~
14.2
117
103~
122
62~
77
18.9~
22.5
6.1~
8.8
104~
158
94~
99
25.1~
29.8
0.46~
0.58
3.40~
4.60
5.5
−12.5~
56.2
4.5~
26.8
0.3~
11.0
−10.3~
21.2
−3.9~
32.2
−1.9~
4.9
−3.3~
6.8
−1.7~
16.0
8.9~
56.1
69
20.6
7.3
128
98
27.2
0.51
4.01
11.0
4.6
7.6
14.1
1.6
2.1
5.4
29.9
BPH (%)
Mean
−10.9
6.1
1.5
−1.1
4.3
0.4
−1.1
0.1
23.6
Table 2. Polymorphism survey of parents using 188 SSR markers and 129 RAPD primers
Chromosome
SSR marker
1
2
3
4
5
6
7
8
9
10
11
12
Total
RAPD primer
OPA~OPL (Operon)
1)
No. of markers
tested
No. of polymorphic
markers
Average1)
heterozygosity of
polymorphic loci
(Mean ± SD)
18
23
18
14
17
12
17
15
13
13
16
12
188 (100 %)
10
8
4
5
4
3
8
8
6
3
10
5
74 (39.4 %)
0.41 ± 0.18
0.45 ± 0.21
0.28 ± 0.12
0.38 ± 0.20
0.48 ± 0.19
0.23 ± 0.08
0.34 ± 0.13
0.35 ± 0.21
0.51 ± 0.20
0.38 ± 0.24
0.41 ± 0.23
0.46 ± 0.09
0.39 ± 0.17
No. of primers tested
No. of bands produced
129 (100 %)
543 (100 %)
No. of alleles of polymorphic
markers
Range
Mean
2~4
2~4
2
2~3
2~4
2
2~3
2~3
2~5
2~5
2~7
2~4
2~7
2.6
3.0
2.0
2.4
2.5
2.0
2.1
2.4
3.3
3.0
3.0
2.6
2.6
No. of primers showed
polymorphism
No.of polymorphic
bands
84 (65.1 %)
157 (28.9 %)
Calculated by method of Nei and Kumar (2000); heterozygosity = 1 −
the i-th allele and q is the number of alleles.
q
2
∑ xi ,
i=1
where xi is the population frequency of
Range
−32.2~
5.0
−0.9~
24.6
−6.3~
9.4
−16.3~
16.4
−15.1~
21.4
−2.9~
3.7
−11.0~
4.7
−10.3~
11.7
4.3~
52.3
392
Cho, Park, Kwon, Chin, Ji, Park, McCouch and Koh
Fig. 1. Cluster diagrams among ten parents in rice based on genetic similarity estimates. Refer
to Table 3 for abbreviations. Variety code; A:Chucheongbyeo, B:Nagdongbyeo, C:
Dongjinbyeo, D:Jinmibyeo, E:Ilpumbyeo, F:Yeongnambyeo, G: Singeumobyeo, H:
Ilmibyeo, I: Hwacheong waxy, J: Nipponbare.
that SSR’s are multi-allelic markers while RAPDs are biallelic; detecting only the presence or absence of any specific amplicon.
In both medium N and 0-N plots, estimates of GS
based on SSR analysis were significantly correlated with
heterosis (r = −0.557 and −0.563, respectively (with MPH),
and −0.536 and −0.542, respectively (with BPH); P < 0.001)
and with hybrid yield (r = −0.732 and −0.548; P < 0.001), re-
spectively (Table 3 (SSR-total)). One-way ANOVA was
performed to determine whether markers were significantly
associated with heterosis. Using the data from the medium N
plots, two groups of F1’s derived from parents having either
the same (homozygous pattern) or different (heterozygous
pattern) alleles at each locus were compared with each other
for yield heterosis and yield. Of 74 polymorphic markers,
25 were associated with a significant (P < 0.05) difference
Table 3. Correlations of estimates of genetic similarity (GS) with yield heterosis and yield
Medium N
GS based on
No. of
markers
SSR-total1)
SSR-poly
SSR-HV
SSR-HW
RAPD-total
RAPD-poly
RAPD-HV
RAPD-HW
Syn-total
Syn-poly
Syn-HV
Syn-HW
188
74
22
3
543
157
13
11
731
231
35
14
1)
2)
MPH2)
−0.557***
−0.552***
−0.674***
0.463**
−0.017
0.009
−0.688***
0.475***
−0.304*
−0.279
−0.689***
0.484***
BPH
−0.536***
−0.531***
−0.643***
0.350*
−0.072
−0.032
−0.620***
0.369*
−0.333*
−0.298*
−0.644***
0.378*
0-N
Yield
−0.732***
−0.730***
−0.600***
0.113
−0.177
−0.203
−0.426**
0.198
−0.513***
−0.525***
−0.546***
0.157
No. of
markers
188
74
22
0
543
157
14
0
731
231
36
0
MPH
−0.563***
−0.564***
−0.665***
—
−0.113
−0.133
−0.662***
—
−0.378*
−0.388**
−0.681***
—
BPH
−0.542***
−0.541***
−0.666***
—
−0.123
−0.148
−0.614***
—
−0.374*
−0.387**
−0.670***
—
Yield
−0.548***
−0.550***
−0.479***
—
−0.096
−0.138
−0.411**
—
−0.357*
−0.385**
−0.473***
—
SSR-total: 188 SSR markers, SSR-poly: 74 SSR markers showing polymorphism among parents; SSR-HV: SSR markers associated with hybrid vigor; SSR-HW: SSR markers associated with hybrid weakness; RAPD-total: 129 RAPD primers; RAPD-poly:
157 polymorphic bands among parents amplified by 84 RAPD primers; RAPD-HV: RAPD bands associated with hybrid vigor;
RAPD-HW: RAPD bands associated with hybrid weakness; Syn-total: both 188 SSR markers and 129 RAPD primers; Syn-poly:
both SSR and RAPD markers showing polymorphism among parents; Syn-HV: both SSR and RAPD markers associated with
hybrid vigor (HV); Syn-HW: both SSR and RAPD markers associated with hybrid weakness.
MPH: mid-parent heterosis, BPH: better-parent heterosis
Key DNA markers for predicting heterosis in japonica rice
between the two groups in mid-parent heterosis (MPH) in
medium N plots. Of these, 22 markers were associated with
an increase in MPH when paired in the heterozygous state at
each locus, while, 3 markers decreased MPH when paired in
the heterozygous state. As a consequence, they were regarded as HV and HW markers, respectively (Table 4). The other
49 SSR markers were found to be neutral to MPH in medium
N plots. In 0-N plots, 22 HV SSR markers associated with
HV were selected as in Table 4, and 19 of them were overlapped with markers selected in medium N plots. However,
no markers associated with HW were detected in 0-N plots.
The HV SSR markers were distributed over 11 chromosomes, and 3 HW SSR markers were over 2 chromosomes.
These HV loci stand for the loci where the heterozygous
combination of alleles in F1 showed significantly higher heterosis than the homozygous combination.
GS estimates based on 22 HV markers were found to be
significantly correlated with heterosis (r = −0.674 and −0.665
(with MPH), and −0.643 and −0.666 (with BPH); P < 0.001)
and with hybrid yield (r = −0.600 and −0.479; P < 0.001)
(Table 3) in medium N and 0-N plots, respectively. GS estimates based on 3 HW markers, on the contrary, showed
significantly positive correlation coefficients with heterosis
(r = 0.463 with MPH and 0.350 with BPH; P < 0.05) in medium N plots, suggesting that the smaller the GS between
parents, the less the HV in the F1 hybrids. That is, the heterozygous allele combination might be associated with poor
performance while specific homozygous combinations of
alleles may be beneficial for HV.
In the case of RAPD markers, GS estimates based on
a total of 129 primers showed no correlation with heterosis
(r = −0.017 and −0.113 (with MPH), and −0.072 and −0.123
393
(with BPH)) or with hybrid yield (r = −0.177 and −0.096)
(Table 3), in medium N and 0-N plots, respectively. The
same procedure as above was conducted with RAPD analysis to identify HV or HW markers. Out of 543 bands (amplified by 129 primers) that were polymorphic among parents,
13 bands associated with HV and 11 bands associated with
HW were identified (Table 4). Interestingly, there was no
RAPD band detected as associated with HW in 0-N plots,
just as in SSR analysis.
GS estimates based on 13 HV markers revealed significant negative correlations with heterosis (r = −0.688 and
−0.662 (with MPH), and −0.620 and −0.614 (with BPH);
P < 0.001) and with grain yield (r = −0.426 and −0.411; P < 0.01)
in medium N and 0-N plots (Table 3), respectively. This is a
dramatic change of predictability of GS estimates based on
RAPD markers between before and after selection of key
markers. Just as in SSRs, GSs based on 11 HW markers
were positively correlated with heterosis (r = 0.475 (P < 0.01)
with MPH and 0.369 (P < 0.05) with BPH), respectively.
GS estimated derived through a combination of unselected RAPD and SSR were only weakly correlated with
heterosis and yield in F1 hybrids. However, when the selected marker sets were used together to estimate GS of parental
varieties, there were highly significant associations between
GS estimates and heterosis/yield as in Table 3. Interestingly
however, the use of both markers together did not significantly change the predictive value of SSR or RAPD analysis
alone. This is probably due to the fact that most of the information on GS between parents estimated from the two marker sets might be overlapping, as revealed in Table 5, in which
the relationships between GS estimates determined using the
22 HV SSR markers and the 13 HV RAPD markers, and
Table 4. SSR and RAPD markers associated with hybrid vigor or hybrid weakness in rice F1 hybrids
SSR markers for
Hybrid vigor
Marker
RM220**
RM9
RM5
RM279
RM53*
RM185
RM252
RM169
RM31**
RM253*
RM3
RM70
RM44
RM223
Chr.
1
1
1
2
2
4
4
5
5
6
6
7
8
8
Marker
RM80
RM264
RM105
RM257**
RM205
RM244
RM286
RM120
RM144
RM313*
RM270
RAPD markers for
Hybrid weakness
Chr.
8
8
9
9
9
10
11
11
11
12
12
Marker
RM302*
RM212*
RM224*
Chr.
1
1
11
Hybrid vigor
Marker1)
OPC08-2
OPF01-3
OPG01-1
OPG02-1
OPH03-10
OPH04-3
OPI07-2
OPK03-1
OPK03-5
OPL03-2
OPL03-4
OPL03-5
OPL19-2
OPF01-1**
Hybrid weakness
(bp)2)
Marker
800b)
OPC13-1*
OPD03-2*
OPD03-4*
OPD13-1*
OPJ01-3*
OPJ01-4*
OPJ07-3*
OPJ08-2*
OPJ08-3*
OPK01-1*
OPK14-1*
Size
700
950
1,500
1,400
1,350
1,500
1,700
700
850
500
400
900
900
Size (bp)
850
1,500
1,100
1,000
1,300
1,000
500
500
400
450
1,000
*,**: Of HV SSR markers, RM53, RM253 and RM313 were identified only in medium N plots, and RM220, RM31, RM257 were
only in 0-N plots. Of HV RAPD markers, OPF01-1 was identified only in 0 N plots. * and ** markers represent those identified
only in medium N and 0-N plots, respectively.
1) RAPD band; for example, OPC08-2 indicates the 2nd band from the top amplified by OPC08, an Operon primer.
2) Approximate size of the RAPD band
394
Cho, Park, Kwon, Chin, Ji, Park, McCouch and Koh
Table 5. A correlation matrix among GS estimates through RAPD and SSR analysis in medium N plots
SSR-total
(1) SSR-total1)
(2) SSR-poly
(3) SSR-HV
(4) SSR-HW
(5) RAPD-total
(6) RAPD-poly
(7) RAPD-HV
(8) RAPD-HW
1)
SSR-poly
1.00
1.00***
0.84***
−0.15
0.24
0.26
0.69***
−0.18
1.00
0.84***
−0.13
0.25
0.27
0.69***
−0.17
SSR-HV
SSR-HW
1.00
−0.20
0.06
0.08
0.94***
−0.25
Discussion
There have been several reports on the feasibility of applying DNA marker-based GS among parents to predict heterosis of F1 hybrid crops. So far, however, the results have
been inconsistent due to the diversity of parental varieties,
the different types of DNA markers used (Bernardo 1992,
Table 6. Significant two-loci epistatic interactions between loci associated with hybrid vigor in medium N plot
MPH (%)
F value
RM53-SG2)
RM5-RM3
RM213-SG
RM253-RM313
RM253-SG
RM21-NP
OPF01-3-SG
OPF01-3-OPH04-3
2)
RAPD-poly
1.00
0.99***
0.00
0.48***
1.00
0.01
0.48***
RAPD-HV
RAPD-HW
1.00
−0.21
1.00
Refer to Table 3 for abbreviations.
between the 3 HW SSR markers and the 11 HW RAPD markers were highly significant.
Two-loci epistatic interactions between loci associated
with HV were analyzed as shown in Table 6. Five significant
interactions in SSR analysis and two significant interactions
in RAPD analysis were found for each of MPH and grain
yield of hybrids in medium N plots. As for hybrid yield, as a
whole, hybrids of homozygous-heterozygous pairs in two
loci out-yielded those of heterozygous-heterozygous pairs,
however the interactions for MPH varied depending on the
locus. It is likely that two-loci epistatic interactions of
heterozygous-heterozygous pairs might not always positively
affect the heterosis level in this study.
1)
1.00
0.39**
0.40**
−0.17
0.94***
RAPD-total
28.88**
12.79*
19.98**
24.29**
18.22**
12.91*
18.70**
20.29**
Allelic combination of two loci1)
0-0
0-1
1-0
11.63
13.44
13.44
6.54
11.63
9.29
12.22
9.27
6.01
5.57
4.41
10.33
5.24
6.30
13.90
12.23
18.70
4.20
4.20
9.07
18.70
13.60
9.27
10.77
1-1
2.51
3.82
4.66
1.43
4.36
2.74
3.61
3.21
Allelic combination of two loci; 0: homozygous, 1: heterozygous.
NP: Two marker loci, RM264 and RM286, where only
Nipponbare had a different allele from the other parental varieties.
SG: Thirteen marker loci, RM279, RM185, RM252, RM169,
RM217, RM3, RM70, RM223, RM80, RM105, RM244,
RM120, and RM144, where only Singeumobyeo had different
alleles from the other parental varieties.
Charcosset et al. 1991, Liu and Wu 1998), and the complex
nature of heterosis itself (Melchinger 1999, Saghai Maroof
et al. 1997, Zhang et al. 1996, Zhao et al. 1999). The array
of conflicting reports on the subject slowed the general utilization of DNA marker technology in the breeding of hybrid
crops.
Heterosis level in this study (Table 1) was relatively
low compared to the indica × indica hybrids, possibly due to
the narrow range of genetic diversity in japonica rice
(Glaszman 1987, Koh et al. 1991, Virmani 1994). In addition, heterosis for yield and yield components was much
higher in 0-N plots than in medium N plots, which is consistent with a review by Virmani (1994) that heterosis was decreased with the increase of nitrogen fertility level.
Polymorphism rate of SSR and RAPD markers and
average heterozygosity of polymorphic loci among parents
(Table 2) was as low as in previous reports on japonica rice
(Ji et al. 1998, Mackill 1995, Ni et al. 2002, Yang et al.
1994, Zhang et al. 1992).
GS estimates based on random arrays of DNA markers
proved to show no or weak correlation with heterosis and
thus were not an effective measure for predicting heterosis.
However, GS estimates based on selected markers revealed
highly significant correlations with heterosis in this data set.
We identified several SSR and RAPD markers in rice that
were predictive of HV. When the parents possessed different
alleles at these marker loci, higher HV could be expected.
Furthermore, some SSR and RAPD markers were associated
with HW. That is, when the parents were heterozygous at
these loci, the heterosis observed in the F1 was significantly
lowered. The presence of HW markers explains why GS estimates based on many unselected markers, which represent
a mixture of HV, HW and neutral (N) markers, were not
highly correlated with heterosis. For example using the data
set from medium N plots in this study, the number of HV,
HW and N bands in RAPD analysis was 13 (8.3 %), 11 (7.0
%) and 133 (84.7 %), respectively, out of 157 (100 %)
polymorphic RAPD bands. Since 13 HV and 11 HW markers were simultaneously involved in estimating GSs when
all markers were used together, the HV and HW markers
effectively nullified the predicting potential of GS in relation
to heterosis (Fig. 2 and Table 3). In SSR analysis, on the
other hand, only 3 markers were associated with HW, and
Key DNA markers for predicting heterosis in japonica rice
395
Fig. 2. A schematic diagram of the key marker concept for predicting heterosis through correlation analysis between heterosis and estimates of GS based on markers and relative
proportion of markers to each group in this study (HV: associated with hybrid vigor, N
(neutral): not associated with heterosis, and HW: associated with hybrid weakness).
therefore estimates of GS based on all 188 or on the 74 polymorphic markers were significantly correlated with heterosis, even though the correlation was less significant than the
GS estimates based on 22 HV markers alone. The reason
why HW markers in both SSR and RAPD analysis were not
detected in 0-N plots might be due to the fact that the heterosis level was much higher in 0-N plots than in medium N
plots and all F1’s out-yielded their mid-parents.
There arise an argument with RAPD analysis, because
the unamplified parents using certain random primers might
consist of more than one allele even though they can be
grouped into one phenotype; absence or the unamplified.
However, due to the fact that parental pairs as “homozygous
(presence/presence and absence/absence) or heterozygous
(presence/absence)” by the phenotype were compared in
analyzing the relationship between GS estimates based on
RAPD analysis and heterosis, that might be no critical defect
of using RAPD analysis allowing for the low precision
caused by the possible subdivision of the ‘absence’ genotype. In fact, the results in selecting key markers were almost
similar to SSR analysis (Table 3). Anyhow, unselected
RAPD markers are unlikely to be useful in genotyping varieties with narrow genetic diversity because of their low resolution.
We suggest that the markers associated with HV and
HW are called “key markers” (Fig. 2 and Table 4), and they
are available for future experiments aimed at determining
how widely applicable they are for predicting heterosis.
Most of the HV markers selected in medium N plots were
also detected as HV markers in 0-N plots indicating that the
markers might be applicable under different environments.
In addition, as shown in a GS diagram (Fig. 1), GS estimates among parents differ depending on the data set used to
estimate it. The use of selected markers allows parents to be
grouped into heterotic or non-heterotic clusters. For exam-
ple, using Syn-HV markers (Fig. 1-11), F1 hybrids of parental pairs of Chucheongbyeo, Nagdongbyeo, Dongjinbyeo,
Jinmibyeo, Ilpumbyeo and Yeongnambyeo, will be less heterotic than other parental pairs.
It is assumed that there must be additional key markers
besides those identified in this study. Here we evaluated
only 45 japonica × japonica hybrids, and it seems that different subspecies and ecotypes of rice are likely to have their
own sets of key markers capable of predicting pairwise heterosis. Given the fact that the 22 HV SSR markers identified
in rice were located on 11 of the 12 chromosomes, it is reasonable to expect that key markers might be dispersed
throughout the genome. If HV allele combinations are assembled into parental lines and poor alleles at HW markers
are removed by selection, we predict that F1 hybrid vigor
will be much increased. This interpretation is consistent with
previous reports (Dudley et al. 1991, Godshalk et al. 1990,
Joshi et al. 2001, Lee et al. 1989, Melchinger et al. 1990,
Smith and Smith 1992, Zhang et al. 1995) where weak or no
correlation was documented between GS estimates based on
a random array of DNA markers and heterosis/hybrid performance. Liu and Wu (1998) experimented with SSR markers in rice and reported that some favorable alleles and
heterogenic patterns observed in parental lines significantly
contributed to heterosis for grain yield of their hybrids,
while some unfavorable alleles and heterogenic patterns
significantly reduced heterosis. Their findings on favorable
and unfavorable alleles and heterogenetic patterns for yield
heterosis using SSR analysis are in agreement with our key
marker concept. There have been several other reports documenting the use of DNA markers for measuring GS or diversities and their effectiveness in plants (Karp et al. 1997,
Mackill et al. 1996, Parsons et al. 1997, Virk et al. 2000).
From this study, we suggest that estimates of GS should be
based on key markers that are appropriately selected. In
396
Cho, Park, Kwon, Chin, Ji, Park, McCouch and Koh
addition, utilization of a few key markers to select parents
for heterotic hybrid combinations may enable the breeders to
save time and cost for genotyping varieties. In future experiments, we are evaluating the potential of these markers for
predicting heterosis in indica × indica hybrids as well as in
additional japonica × japonica hybrids.
What would be the nature of key markers or loci affecting the amount of heterosis? This question is directly related
to the mechanism of heterosis which, by genetic interpretation, has been attributed to three aspects; the accumulation
of dominance alleles, overdominance effects at particular
loci (Crow 1999, Hua et al. 2003), and/or epistatic interactions among loci in the F1 (Goodnight 1999, Hua et al. 2003,
Luo et al. 2001, Yu et al. 1997). At a glance, it could be inferred that HV markers should be related to overdominance
alone because HV was greater when the F1’s were heterozygous for marker alleles at HV marker-loci. However, as
reported by Fu and Dooner (2002), this situation may be
equally consistent with the dominance hypothesis, given that
a corresponding ‘null allele’ may represent the overdominance counterpart at any given locus. It is unlikely that epistatic interactions between pairs of loci would significantly
affect heterosis for yield in this study because the directions
of epistasis were not consistent depending on locus. Hua et
al. (2003) reported that single-locus heterotic effects and
dominance by dominance interactions at two-locus level
could explain the genetic basis of heterosis in an elite rice
hybrid. The key marker loci in this study, in conceptual viewpoint, may correspond to the heterotic loci which mainly
represented overdominance effects in the report by Hua et al.
(2003). Further studies are needed to elucidate the nature of
key marker loci in relation to the mechanism of heterosis and
to investigate the relationship of key markers with QTLs
conditioning yield and its heterosis.
We expect that the key marker-loci provide a starting
point for addressing questions related to the mechanism of
heterosis as well as for establishing an efficient method for
predicting and improving heterosis in the breeding of hybrid
crop varieties.
Acknowledgements
This research was supported by a grant (code
#CG3111) from Crop Functional Genomics Center of the
21st Century Frontier Research Program funded by the
Ministry of Science and Technology, Republic of Korea.
Authors are greatful to Prof. Hiroshi Ikehashi for his kind
translation of summary in Japanese.
Literature Cited
Ahn, S.N., T.S. Kwak, K.H. Kang, Y.H. Jeon, H.C. Choi and H.P. Moon
(1998) Relationship between heterosis and genetic distance as
measured by RAPDs analysis in rice. Korean J. Breed. 30: 1623.
Barbosa-Neto, J.F., M.E. Sorrells and G. Cisar (1996) Prediction of het-
erosis in wheat using coefficient of parentage and RFLP-based
estimates of genetic relationship. Genome 39: 1142-1149.
Bernardo, R. (1992) Relationship between single-cross performance
and molecular markers in maize. Theor. Appl. Genet. 85: 10551062.
Causse, M.A., T.M. Fulton, Y.G. Cho, S.N. Ahn, J. Chunwongse, K. Wu,
J. Xiao, Z. Yu, P.C. Ronald, S.E. Harrington, G. Second, S.R.
McCouch and S.D. Tanksley (1994) Saturated molecular map
of the rice genome based on an interspecific backcross population. Genetics 138: 1251-1274.
Cerna, F.J., S.R. Cianzio, A. Rafalski, S. Tingey and D. Dyer (1997) Relationship between seed yield heterosis and molecular marker
heterozygosity in soybean. Theor. Appl. Genet. 95: 460-467.
Charcosset, A.M., M. Lefort-Buson and A. Gallais (1991) Relationship
between heterosis and heterozygosity at marker loci: a theoretical computation. Theor. Appl. Genet. 81: 571-575.
Chen, X., S. Temnykh, Y. Xu, Y.G. Cho and S.R. McCouch (1997)
Development of a microsatellite framework map providing
genome-wide coverage in rice (Oryza sativa L.). Theor. Appl.
Genet. 95: 553-567.
Cheres, M.T., J.F. Miller, J.M. Crane and S.J. Knapp (2000) Genetic distance as a predictor of heterosis and hybrid performance within
and between heterotic groups in sunflower. Theor. Appl. Genet.
100: 889-894.
Chowdari, K.V., S.R. Venkatachalam, A.P. Davierwala, V.S. Gupta,
P.K. Ranjekar and O.P. Govila (1998) Hybrid performance and
genetic distance as revealed by the (GATA)4 microsatellite and
RAPD markers in pearl millet. Theor. Appl. Genet. 97: 163169.
Corbellini, M., M. Perenzin1, M. Accerbi1, P. Vaccino1 and B. Borghi
(2002) Genetic diversity in bread wheat, as revealed by coefficient of parentage and molecular markers, and its relationship
to hybrid performance. Euphytica 123: 273-285.
Crow, J.F. (1999) Dominance and overdominance. In “The Genetics
and Exploitation of Heterosis in Crops” Coors, J.G. and S. Pandy
(eds.), ASA-CSSA-SSSA, Madison, Wisconsin. p. 49-58.
Diers, B.W., P.B.E. McVetty and T.C. Osborn (1996) Relationship between heterosis and genetic distance based on restriction fragment length polymorphism markers in oilseed rape (Brassica
napus L.). Crop Sci. 36: 79-83.
Dudley, J.W., M.A. Saghai Maroof and G.K. Rufener (1991) Molecular
markers and grouping of parents in maize breeding programs.
Crop Sci. 31: 718-723.
Fu, H. and H.K. Dooner (2002) Intraspecific violation of genetic colinearity and its implications in maize. Proc. Natl. Acad. Sci. 99:
9573-9578.
Glaszmann, J.C. (1987) Isozymes and the classification of Asian rice
varieties. Theor. Appl. Genet. 74: 21-30.
Godshalk, E.B., M. Lee and K.R. Lamkey (1990) Relationship of restriction fragment length polymorphisms to single-cross hybrid
performance in maize. Theor. Appl. Genet. 80: 273-280.
Goodnight, C.J. (1999) Epistasis and heterosis. In “The Genetics and
Exploitation of Heterosis in Crops” Coors, J.G. and S. Pandy
(eds.), ASA-CSSA-SSSA, Madison, Wisconsin. p. 59-68.
Hua, J., Y. Xing, W. Wu, C. Xu, X. Sun, S. Yu and Z. Qifa (2003) Singlelocus heterotic effects and dominance interactions can adequately explain the genetic basis of heterosis in an elite rice
hybrid. Proc. Natl. Acad. Sci. 100: 2574-2579.
Ji, H.S., H.J. Koh, S.U. Park and S.R. McCouch (1998) Varietal identification in japonica rice using microsatellite DNA markers.
Korean J. Breed. 30: 350-360.
Key DNA markers for predicting heterosis in japonica rice
Joshi, S.P., S.G. Bhave, K.V. Chowdari, G.S. Apte, B.L. Dhonukshe,
K. Lalitha, P.K. Ranjekar and V.S. Gupta (2001) Use of DNA
markers in prediction of hybrid performance and heterosis for a
three-line hybrid system in rice. Biochem. Genetics 39: 179200.
Karp, A., K.J. Edwards, M. Bruford, S. Funk and B. Vosman (1997)
Molecular technologies for biodiversity evaluation: opportunities
and challenges. Nature Biotechnol. 15: 625-628.
Koh, H.J., H.Y. Kim and M.H. Heu (1991) Current status and prospects
in research and development of hybrid rice. Coll. Agric. Bull.,
Seoul Natl. Univ. 16: 11-23.
Lee, M., E.B. Godshalk, K.R. Lamkey and W.W. Woodman (1989) Association of restriction fragment length polymorphisms among
maize inbreds with agronomic performance of their crosses.
Crop Sci. 29: 1067-1071.
Liu, X.C. and J.L. Wu (1998) SSR heterogenic patterns of parents for
marking and predicting heterosis in rice breeding. Mol. Breed.
4: 263-268.
Luo, L.J., Z.K.Li, H.W. Mei, Q.Y.Shu, R. Tabien, D.B. Zhong, C.S. Ying,
J.W. Stansel, G.S. Khush and A.H. Paterson (2001) Overdominant epistatic loci are the primary genetic basis of inbreeding
depression and heterosis in rice. II. Grain yield components.
Genetics 158: 1755-1771.
Mackill, D.J. (1995) Classifying japonica rice cultivars with RAPD
markers. Crop Sci. 35: 889-894.
Mackill, D.J., Z. Zhang, E.D. Redona and P.M. Colowit (1996) Level of
polymorphism and genetic mapping of AFLP markers in rice.
Genome 39: 969-977.
Majarrez-Sandoval, P., T.E. Carter Jr., D.M. Webb and J.W. Burton
(1997) Heterosis in soybean and its prediction by genetic similarity measures. Crop Sci. 37: 1443-1452.
Martin, J.M., L.E. Talbert, S.P. Lanning and N.K. Blake (1995) Hybrid
performance in wheat as related to parental diversity. Crop Sci.
35: 104-108.
Melchinger, A.E. (1999) Genetic diversity and heterosis. In “The
Genetics and Exploitation of Heterosis in Crops” Coors, J.G.
and S. Pandy (eds.), ASA-CSSA-SSSA, Madison, Wisconsin.
p. 31-48.
Melchinger, A.E., M. Lee, K.R. Lamkey and W.W. Woodman (1990)
Genetic diversity for restriction fragment length polymorphisms: relation to genetic effects in maize inbreds. Crop Sci.
30: 1033-1040.
Nei, M. and W. Li (1979) Mathematical model for studying genetic
variation in terms of restriction endonucleases. Proc. Natl.
Acad. Sci. USA 76: 5256-5273.
Ni, J., P.M. Colowit and D.J. Mackill (2002) Evaluation of genetic
diversity in rice subspecies using microsatellite markers. Crop
Sci. 42: 601-607.
Panaud, O., X. Chen and S.R. McCouch (1996) Development of microsatellite markers and characterization of simple sequence
length polymorphism (SSLP) in rice (Oryza sativa L.). Mol.
Gen. Genet. 252: 597-607.
397
Parsons, B.J., H.J. Newbury, M.T. Jackson and B.V. Ford-Lloyd (1997)
Contrasting genetic diversity relationships are revealed in rice
(Oryza sativa L.) using different marker types. Mol. Breed. 3:
115-125.
Rohlf, F.J. (1997) NTSYS-pc. Numerical taxonomy and multivariate
analysis system: version 2.0. Applied Biostatistics, New York.
Saghai Maroof, M.A., G.P. Yang, Z. Qifa and K.A. Gravois (1997) Correlation between molecular marker distance and hybrid performance in U.S. southern long grain rice. Crop Sci. 37: 145-150.
Smith, O.S., J.S.C. Smith, S.L. Bowen, R.A. Tegborg and S.J. Wall
(1990) Similarities among a group of elite maize inbreds as
measured by pedigree, F1 heterosis, and RFLPs. Theor. Appl.
Genet. 80: 833-840.
Smith, O.S. and J.S.C. Smith (1992) Measurement of genetic diversity
among maize inbreds; a comparison of isozymic, RFLP, pedigree, and heterosis data. Maydica 37: 53-60.
Virk, P.S., H.J. Newbury, M.T. Jackson and B.V. Ford-Lloyd (2000)
Are mapped markers more useful for assessing genetic diversity? Theor. Appl. Genet. 100: 607-613.
Virmani, S.S. (1994) Heterosis and hybrid rice breeding. SpringerBerlag, 189 p.
Xiao, J., J. Li, L. Yuan, S.R. McCouch and S.D. Tanksley (1995) Dominance is the major genetic basis of heterosis in rice as revealed
by QTL analysis using molecular markers. Genetics 140: 745754.
Xiao, J., J. Li, L. Yuan, S.R. McCouch and S.D. Tanksley (1996) Genetic diversity and its relationship to hybrid performance and heterosis in rice as revealed by PCR-based markers. Theor. Appl.
Genet. 92: 637-643.
Yang, G.P., M.A.S. Maroof, C.G. Xu, Q. Zhang and R.W. Biyashev
(1994) Comparative analysis of microsatellite DNA polymorphism in landraces and cultivars of rice. Mol. Gen. Genet. 245:
187-194.
Yu, S.B., J.X. Li, C.G. Xu, Y.F. Tan, Y.J. Gao, X.H. LI, Q.F. Zhang and
M.A. Saghai-Maroof (1997) Importance of epistasis as the genetic basis of heterosis in an elite rice hybrid. Proc. Natl. Acad.
Sci. 94: 9226-9231.
Zhang, Q., M.A. Saghai Maroof, T.Y. Lu and B.Z. Shen (1992) Genetic
diversity and differentiation of indica and japonica rice detected by RFLP analysis. Theor. Appl. Genet. 83: 495-499.
Zhang, Q., Y.J. Gao, M.A. Saghai Maroof, S.H. Yang and J.X. Li (1995)
Molecular divergence and hybrid performance in rice. Mol.
Breed. 1: 133-142.
Zhang, Q., Z.Q. Zhou, G.P. Yang, C.G. Xu, K.D. Liu and M.A. Saghai
Maroof (1996) Molecular marker heterozygosity and hybrid
performance in indica and japonica rice. Theor. Appl. Genet.
93: 1218-1224.
Zhao, M.F., X.H. Li, J.B. Yang, C.G. Xu, R.Y. Hu, D.J. Liu and Q. Zhang
(1999) Relationship between molecular marker heterozygosity
and hybrid performance in intra- and inter-subspecific crosses
of rice. Plant Breed. 118: 139-144.