Screening and analysis of wide compatibility loci in wide crosses of rice

With indica and japonica testers to screen out wide compatibility types, a number
of varieties seemed to be indicas but differed from them by showing semisterility
in crosses with Ketan Nangka, a donor of the wide compatibility allele (neutral
allele). Another varietal group showed good fertility with indica and japonica
testers, but revealed sterility in crosses with Ketan Nangka. Thus, Ketan Nangka
is suggested as a standard variety, along with the aus varieties, which show semisterility
in crosses with indica and japonica testers but normal fertility with most
aus varieties. A set of four varieties—Achar Bhog, Ketan Nangka, IR36, and a
japonica type—is proposed as standard testers for hybrid sterility. F 1 hybrid
sterility in rice is understood with allelic interactions at the S-5 locus. With the
identification system for S-5, a large number of crosses were made to test the
extent to which the neutral allele at the S-5 locus is effective. Hybrid sterility in
Penuh Baru II and aus varieties, which is not explained by the testers for S-5, was
found to be due to an additional locus rather than to a new allele. The neutral
allele at the S-5 locus can now be effectively used, but a new neutral allele
indicated by Dular would also be important in rice breeding.

Hybrid sterility limits the application of wide crossing in rice breeding and lowers the
productivity of hybrid rices. Its genetic basis was not understood for a long time.
Recently, it was found that the F 1 sterility of hybrids between indica and japonica
varieties is caused by an allelic interaction at a locus at which the indicas have S-5 i , the
japonicas have S-5 j , and some javanicas have a neutral allele, S-5 n . The donor of S-5 n
is termed a wide compatibility variety (WCV). The S-5 i /S-5 j genotype is sterile because
of the abortion of gametes carrying the S-5 j allele, but S-5 n /S-5 i and S-5 n /S-5 j are fertile
(Ikehashi and Araki 1986). Allele S-5 n has been incorporated as a wide compatibility
gene into indica and japonica backgrounds to overcome sterility problems in wide cross
and hybrid rice breeding (Araki et al 1988). This approach has been promising for
breeding commercial hybrids between indicas and japonicas. A record of very high
yield is reported in indica/japonica hybrids using the wide compatibility allele
(Maruyama 1988).
Following initial successes in utilizing S-5 n , a number of problems emerged: the
extent to which the locus is effective, the possibility of an additional locus, and systems
33
to identify the wide compatibility alleles. So far, the S-5 n allele seems to be effective
for most typical indicas and japonicas. However, in the expansion of our screening of
WCVs, many varieties were found for which S-5 n was not effective. Hybrids between
some aus varieties and javanicas including WCVs show clear semisterility, while some
aus varieties such as Dular show fertility when crossed with javanica WCVs as well as
with indicas or japonicas. As hybrid sterility caused at the S-5 locus can be identified
with tester varieties and marker genes, it was possible to know whether the hybrid
sterility between javanica and aus varieties is caused at the S-5 locus or not. Contrasting
results between the javanica WCV and Dular in their crosses with aus varieties
appeared to be caused at a new locus at which Dular has another neutral allele and the
javanicas and aus varieties each have interacting alleles. Genetic analyses suggested
that a new locus is linked with Rc (red pericarp) in linkage group IV. Simultaneous use
of neutral alleles at the two loci may solve the F 1 sterility problem in practically all kinds
of crosses in rice.
A set of indica and japonica varieties has been used for screening WCVs. With our
expanded knowledge of the types of hybrid sterility, it is now necessary to use more
varieties as standards for screening WCVs. Several varieties are suggested.
Male sterility was recorded for individual plants in all the experiments, but no
relationship was found between male sterility and the marker genes. Information is
necessary to identify the genetic basis of male sterility; however, spikelet sterility is
discussed here for simplicity. So far, pollen semisterility does not seem to lower the
seed fertility of indica/japonica hybrids.
Hybrid sterility within and between varietal groups
To identify WCVs, more than 80 varieties, most of which were aus or javanicas, were
crossed with indica and japonica tester varieties, and the pollen and spikelet fertility of
the F 1 hybrids was examined (Ikehashi and Araki 1984). Along with the screening, a
large number of crosses were made to determine compatibility types in each varietal
group. The compatibility types of a given variety are here defined by hybrid sterilities
shown in a set of crosses between the variety and testers. In the initial tests, several
compatibility types in the javanica as well as in the aus varieties were found with the
use of indica and japonica tester varieties. Additional crosses were made between the
different types within each varietal group, and the F 1 hybrids were tested for fertility
from 1983 to 1985. Many crosses were also made between javanica and aus varieties
in the same period. All the tests were conducted at the Okinawa Branch of the Tropical
Agriculture Research Center, where spikelet fertility is least affected by cold temperature
due to the subtropical climate. For all the crosses, pollen and spikelet fertilities of
the F 1 hybrids were determined by a standard method (Ikehashi and Araki 1984).
Compatibility types in javanicas
Following the identification of WCVs such as Ketan Nangka, Calotoc, and CPSLO-
17, other compatibility types in the javanica group were found. Of 24 javanicas, 15
34 Ikehashi et ai
were classified into 1 group based on their high pollen fertility with both the indica and
japonica testers, normal spikelet fertility with the japonica tester, and clear semisterility
with the indica tester. These 15 varieties were designated Banten types. Six
varieties showed semisterility in their crosses with indicas as well as japonicas. A
representative variety, Penuh Baru II, was selected from this group. Only one variety,
Padi Bujang Pendek, was identified as a WCV. The rest seemed to be exceptional.
Hybrids between types of javanicas
The six javanicas that showed semisterility in their crosses with the indica and japonica
testers showed normal fertility in their cross with Ketan Nangka, a javanica WCV, and
with varieties of the Banten group. It was thus concluded that, with a few exceptions,
F 1 hybrids between different compatibility types in the javanica group show normal
fertility.
Compatibility types in aus varieties
Forty-one aus varieties were tested; all the data are presented by Ikehashi and Araki
(1987). The fertilities of some hybrids between aus varieties and indica or japonica
testers are shown in Table 1. Aus 373 and Dular seemed to be widely compatible,
although the pollen fertility in their crosses with IR varieties was marginal. Next to
these two, five varieties including Panbira showed good fertility with the testers. A
majority of 18 varieties were not classified into any definite category. Of them, five
including Achar Bhog showed semisterility with indica and japonica testers. In many
cases, the hybrid sterility exhibited by these varieties was marginal, leaving some
possibility of reclassification.
Table 1. Spikelet fertility (%) of F 1 hybrids between testers and some aus varieties, 1984-85. a
Spikelet fertility (%)
Variety Source Testers Javanicas
Japonicas IR36 Ketan Banten Penuh
Nangka Baru II
Achar Bhog
Aus 373
CH972
D1123
Dular
lngra
Kaladumai
Kele
Panbira
Prambu Vattan
Satika
Acc. 25826
Acc. 29158
200011
Acc. 8455
200041
Acc. 27552
200040
210013
VT. 64
200049
210003
71.1
88.0
30.5
87.2
83.7
61.4
94.7
69.4
84.8
89.1
81.8
62.0
91.3
70.9
86.7
68.3
81.1
88.2
92.1
34.0
54.4
84.8
59.2
50.1
65.3
57.9
90.9
48.8
97.5
22.6
18.1
92.4
60.1
39.1
24.0
64.4
89.6
85.5
50.0
25.0
81.4
48.6
50.0
58.6
73.1
89.2
44.6
54.7
59.4
a Details in lkehashi and Araki (1987).
Wide compatibility loci in wide crosses of rice 35
–
–
–
–
Hybrids between aus varieties
Because there were different compatibility types in the aus varieties—some were like
indicas and others like japonicas—many varieties from different types were chosen
and crossed with each other to test the fertility of the F 1 hybrids. The fertility of such
hybrids was normal regardless of the compatibility type of the parent variety.
Exceptionally low fertility was found only in the crosses of Prambu Vattan, which
seemed to be a japonica in various aspects.
Crosses between aus varieties and javanica WCVs
Some aus varieties were crossed with WCVs, and the fertility of the F 1 hybrids was
examined. The fertilities of some hybrids between aus varieties and Ketan Nangka are
shown in Table 1. Many aus varieties showed hybrid sterility in their crosses with Ketan
Nangka. Two japonica-like varieties—Kaladumai and Prambu Vattan—and Dular
showed normal fertility in the cross with Ketan Nangka (Table 1).
To determine whether semisterility is a common fact in F 1 hybrids between WCVs
and aus varieties, additional aus varieties were crossed with other WCVs, viz., Calotoc
and CPSLO, and with other javanicas. With some exceptions, the F 1 s between WCVs
and aus varieties showed semisterility. Earlier, it was indicated that the Penuh Baru
group of javanicas and some aus varieties such as Achar Bhog were similar in their
semisterility both with indica and japonica testers; however, F 1 hybrids between
Penuh Baru II and some aus varieties such as Achar Bhog showed clear semisterility
(Table 1).
Tests of South Indian varieties
Varieties from South India or Sri Lanka were tested together with additional varieties
in the aus group (Table 2). Karalath, Pusur, and Eat Samba showed good fertility in
their cross with the indica or japonica testers, but not so with Ketan Nangka. Only
Table 2. Spikelet fertility (%) of F 1 hybrids between testers and aus or South Indian varieties,
1985.
Spikelet fertility (%)
Tester Aus varieties South Indian varieties
Karalath Pusur Surjamukhi Triveni Eat Dahanala
Samba
Japonicas
IR36
Ketan Nangka
Panbira
Achar Bhog
91.7
77.2
58.8
91.6
85.1
91.1
47.2
95.0
96.2 a
41.4 a
91.2 a
83.1 a
38.7 a
85.7 a
37.9
98.6
90.7 a
87.5 a
49.9 a
45.0 a
74.3 a
67.6 a
a Used as pollinator.
36 lkehashi et al
– –
– – –
–
–
–
Surjamukhi showed high fertility with Ketan Nangka. Triveni and Dahanala were
definitely not classified into indicas or japonicas. These varieties from South India or
Sri Lanka seemed to be similar to most aus varieties in that they showed semisterility
when crossed with Ketan Nangka.
Tests of varieties from Bhutan, China, and Korea
Crosses between some testers and Asian varieties gave the results shown in Table 3.
Two improved Korean lines and Nanjing 11 seemed to be typical indica types. Three
native varieties from China were similar to indicas but differed in their lower fertility
with Ketan Nangka. They were similar to Triveni and Dahanala. The Bhutan varieties
Jyakuchem and Kuchem showed good fertility both with indica and japonica testers,
suggesting their wide compatibility. Jyakuchem was found to possess S-5 n .
Standard varieties for identifying compatibility types
The compatibility tests revealed a number of varieties in China and India that can be
identified as indicas with the use of the indica-japonica testers. But they differ from
indica testers such as IR36 and IR50 in their low compatibility with Ketan Nangka.
Examples of such varieties are CH972, Triveni, Dahanala, Pe-Bi-Hun, and Tao-Jen-
Chiao. Another type shows good fertility with both indica and japonica testers but
significantly lower fertility in crosses with Ketan Nangka. To this type belong Aus 373,
Panbira, Pusur, and Eat Samba. Ketan Nangka can thus be considered a standard
variety. Whether or not a hybrid between a given variety and Ketan Nangka shows
semisterility can be a criterion for classifying compatibility types. As most aus varieties
show good fertility in crosses with each other, aus variety Achar Bhog was also selected
as a standard variety. This variety shows semisterility in its cross with indica and
japonica testers, so that any variety showing good fertility when crossed with it may
be a kind of aus. Thus, a set of four varieties—Achar Bhog, Ketan Nangka, IR36, and
japonica variety Taichung 65 or Akihikari—would be useful to identify compatibility
types.
Table 3. Spikelet fertility (%) of F 1 hybrids between testers and some Asian varieties
showing atypical performance.
Spikelet fertility (%)
Tester Korea (1984) China (1984) Bhutan (1985)
Milyang Suweon Pi-bi-hun Tuan-ku- Tao-jen- Nanjing Jyaku- Kuchem
23 258 chao chiao 11 chem
Japonicas
IR36
Ketan Nangka
Achar Bhog
Aus 373
38.5
85.5
96.8
64.6
44.8
86.5
92.8
87.8
93.6
41.7
89.2
44.8
74.2
86.5
37.9
94.9
83.3
98.4
12.5
82.1
44.8
57.8
89.5
48.5
94.3
94.8
89.7
93.6
90.2
90.5
84.1
89.7
71.4
Wide compatibility loci in wide crosses of rice 37
–
–
–
– –
– –
Genetic analysis of hybrid sterility between varietal groups
A large number of crosses were made to test compatibility types of varieties, and Penuh
Baru II was found to show clear semisterility in its cross with indica as well as japonica
testers, suggesting that the tester system for S-5 is not applicable to this variety. Also,
many aus varieties showed clear semisterility with javanica WCVs such as Ketan
Nangka and Calotoc. Some aus varieties were crossed for genetic analysis.
New locus suggested by a javanica variety
Among the compatibility types in the javanica group, Penuh Baru II was found to
produce semisterile F 1 hybrids when crossed with japonica and indica testers, while
producing fertile F 1 s in its cross with Ketan Nangka. In the three-variety cross Ketan
Nangka/Penuh Baru II//IR50, close linkages were found between spikelet fertility and
the wx or C gene from Ketan Nangka (Table 4). In the complementary cross, Ketan
Nangka/IR36//Penuh Baru II, the same linkage relationship was shown (Table 4).
Therefore, Ketan Nangka’s S-5 n allele, which is closely linked with the wx and C genes,
must be allelic to the sterility-causing allele in the F 1 between the indicas and Penuh
Baru II. Since the F 1 hybrids from Penuh Baru II and indicas are semisterile, the allele
possessed by Penuh Baru II must be different from S-5 i . On the other hand, the sterility
of the F 1 of Penuh Baru II and japonica variety Akihikari was not affected by the locus
near C and wx in the related cross of Ketan Nangka/Akihikari//Penuh Baru II (Table
4). The compatibility relation among indica, javanica, and japonica varieties can thus
be ascribed to an allelic interaction at the S-5 locus. However, the sterility of the F 1 s of
Table 4. Distribution among 8 fertility classes (0–30 to 91–100%) of spikelet fertility in 3-
variety crosses with Ketan Nangka and Penuh Baru II.
Plants (no.) in each fertility class
Marker Total Mean
0-30 31-40 41-50 51-60 61-70 71-80 81-90 91-100 (no.) (%)
C/C +
C + /C +
wx/+
+/+
C/C +
C + /C +
wx/+
+/+
C/C +
C + /C +
wx/+
+I+
1
1
3
3
2
2
3
3
Ketan Nangka/Penuh Baru ll/lR50 (1984)
1 7 16
7 14 1
2 3 6 13
6 11 1 4
Ketan Nangkal/lR36//Penuh Baru II (1985)
1
6
2
5
26
2
24
2
21
9
14
2
1
1
12
9
3
1
5
3
3
3
2
3
2
Ketan Nangka/Akihikari//Penuh Baru II (1985)
14 15 23
9 11 15
15 16 21
8 10 17
3
1
2
42
29
13
4
6
5
5
27
25
25
27
57
58
52
63
63
48
66
45
82.1
50.9
76.5
58.5
91.7
52.1
83.3
61.4
73.7
74.2
73.4
76.2
38 lkehashi et al
Penuh Baru II and the japonicas must be caused at a locus other than S-5, where Ketan
Nangka and indicas may have a neutral allele, since the sterility of the two varieties is
caused only at the S-5 locus (Fig. 1).
Detection of S-5 n in wide compatible aus varieties
It has been indicated that aus varieties include various compatibility types in terms of
F 1 sterility with indica and japonica testers, and that the types of compatibility are likely
to differ from those of other groups. Therefore, the same method that applied to the
analysis of WCVs was attempted for hybrids between aus varieties and the other types.
Aus 373 seemed to be a WCV and was tested in a three-variety cross using indica
and japonica testers (Table 5). Spikelet fertility in the cross Aus 373/IR50//Akihikari
was related to that of the genotype of C/C + , suggesting that an allelic interaction
between an allele from IR50 and another from Akihikari was responsible for spikelet
sterility. Therefore, allelic interaction at the S-5 locus was indicated. Similarly, Dular
and Pusur were found to show normal fertility in their crosses with indica and japonica
1. Loci for hybrid sterility in Penuh Baru II. In its cross with IR36, Penuh Baru II reveals sterility due to
allelic interaction at the S-5 locus. Its sterility in crosses with japonicas is due to another locus, where both
Ketan Nangka and IR36 have a neutral allele.
Wide compatibility loci in wide crosses of rice 39
Table 5. Detection of wide compatibility allele near C locus by differentiating spikelet
sterility in crosses to indica and japonica testers.
C/C +
C + /C +
C/C +
C + /C +
C/C +
C + /C +
alk/ +
+I+
1
5
1
4
2
3
1
2
1
4
6
4
2
8
3
7
2
10
4
7
2
8
Total
(no.)
Plants (no.) in each fertility class
Marker
genes 0-20 21-30 31-40 41-50 51-60 61-70 71-80 81-90 91-100
Aus 373/lR50//Akihikari (1983)
8 6 9 4 2 33
5 3 7 1 26
Pusur/lR36//Akihikari (1985)
1 5 4 22
12 11 3 1
Akihikari/Dular//Tae baekb (1988)
6 8 12 30 25
13 15 18 18 5
7 6 10 30 25
12 17 20 18 4
12
7
7
46
46
99
84
91
90
** = significant at the 1% level. Wigh-yielding indica variety from Korea.
Mean
(%) test a
t-
57.5
47.5
80.9
51.4
68.7
57.0
71.4
55.4
**
**
**
**
testers. Three-variety crosses with the two varieties indicated the existence of an S-5 n
allele (Table 5). But Pusur is different from Dular in respect to its semisterility when
crossed with Ketan Nangka.
Locus suggested by javanica WCVs and aus varieties
Many crosses between aus varieties and indica or japonica varieties were tested, but no
clear relationship was found between fertility level and marker genotype. Allelic
interactions can probably not be detected with the limited number of marker genotypes.
After such tests, a different kind of three-variety cross was made, in which one aus
variety and two WCVs were crossed to detect allelic interaction between the aus
varieties and the javanica WCVs. An aus variety with red pericarp, Ingra, was used in
Ingra/Ketan Nangka//CPSLO 17, where the cross between Ketan Nangka and CPSLO
17 did not show hybrid sterility and the sterility could be due only to Ingra and CPSLO
17. In this cross, the level of spikelet fertility was related to the Rc locus in linkage group
IV (Table 6). Therefore, the pronounced hybrid sterility between javanica WCVs and
aus varieties may be caused at this locus near Rc. To analyze the new locus, several
crosses were tested. But in similar crosses using aus varieties with red pericarp, such
as Kele and Chakila, the effect of the locus near Rc was not found (Table 6). It is likely
that the locus near Rc can function in only some aus varieties.
Dular and Ketan Nangka have the same neutral allele at the S-5 locus but differ in
their cross with aus varieties. Dular also has a neutral allele at the newly suggested
locus, where Ketan Nangka shows allelic interaction with some aus varieties (Fig. 2).
40 lkehashi et al
Table 6. Three-variety crosses with aus, javanica, or japonica varieties.
Plants (no.) in each fertility class
Marker Total Mean tgenes
0-20 21-30 31-40 41-50 51-60 61-70 71-80 81-90 91-100 (no.) (%) test a
+/+
wx/ +
Rc/Rc +
Rc + /Rc +
Ingra/Ketan nangka//CPSLO 17
3 10 3 5 12
9 9 5 1 14
11 14 5 2 2
1 5 3 3 19
Rc/Rc +
Rc + /Rc +
Rc/Rc +
Rc + /Rc +
1
1
0
2
2
3
1
1
1
1
2
a ** = significant at the 1% level.
Tatsumimochi/Chakila//Banten (1989)
1 2 6 12 23
0 7 13 4 17
Tatsumimochi/Kele//Banten (1989)
4 5 6 8 11
3 4 7 7 20
11 46 73.8
12 52 71.8
1 38 55.9
28 60 83.7 **
22 68 82.2
21 63 79.0
20 56 81.1
26 69 77.9
2. Loci for hybrid sterility between aus varieties and javanica WCVs. Dular and Ketan Nangka have neutral
alleles at the S-5 locus. But Dular has another neutral allele at a locus near Rc, where an aus variety and
javanica WCVs have two interacting alleles for gamete abortion.
Wide compatibility loci in wide crosses of rice 41
Discussion
The system of F 1 hybrid sterility in rice is now better understood in the light of allelic
interactions at a locus. The basic structure of allelic interactions can be shown as a
triangular relationship between three alleles, i.e., one neutral and two interacting
alleles. Gametes possessing one of the interacting alleles are eliminated in heterozygotes
of such alleles, while in heterozygotes of a neutral allele and another allele no
gamete abortion occurs.
The identification system for S-5 is well constructed, with standard testers from
indica and japonica types as well as such marker genes as C and alk in linkage group
I. Thus, the neutral allele known as the wide compatibility gene has been used in
breeding hybrid varieties.
Since this gene mechanism has been understood, a large number of crosses have
been made to test the extent to which the neutral allele is effective. Then two varietal
groups were found for which the identification system for S-5 is not adequate. First, an
exception was demonstrated by Penuh Baru II, which showed clear semisterility in its
crosses with indica as well as with japonica testers, suggesting that the testers for S-5
are not applicable. Second, many aus varieties showed clear semisterility in their
crosses with javanica WCVs such as Ketan Nangka and Calotoc.
The limitation of the initial identification system for alleles at the S-5 locus implies
that there are more alleles at the S-5 locus or that additional loci function independently
of the S-5. One of the clues for determining the genetic basis was obtained by marker
genes. Because any allelic action at the S-5 locus can be definitely traced by such
markers as C or alk, a sterility reaction without any relation to the markers may be due
to another locus.
Hybrid sterility in Penuh Baru II and aus varieties, which is not explained by the
standard system for S-5 alleles, is found to be caused at an additional locus rather than
by a new allele. In the case of Penuh Baru II, a new locus is suggested, where the indica
tester and Ketan Nangka have a neutral allele, while japonicas and Penuh Baru II have
interacting alleles. In the hybrid sterility between WCVs and some aus varieties, they
are assumed to possess interacting alleles, with Dular possessing a neutral allele at this
locus. Dular showed an exceptionally good compatibility with indica, javanica, and
japonica varieties.
Although the neutral allele at the S-5 locus is effectively incorporated into indica and
japonica varieties, the use of a new neutral allele indicated by Dular would be important
to breeding work on the Indian Subcontinent. Further progress in the study of hybrid
sterility will depend on the availability of marker genes. The use of isozymes provides
a partial solution. But biochemical markers may be more useful. Standardization of
tester varieties is also necessary. Such a system would be useful in classifying varieties,
inasmuch as such terms as indicas and japonicas are as confusing in studies of hybrid
sterility as in other research areas.
42 lkehashi et al