Cytogenetic, electrophoretic, and root studies of javanica rices

Intercrosses among cultivars of traditional upland varieties and bulu varieties of
Indonesia as well as of the dual-purpose type, the aus varieties, were made to
elucidate their cytogenetic and biochemical relationships as well as similarities in
their root systems. Based on pollen and spikelet fertility, the bulu and upland
varieties showed slightly higher genetic affinity with each other than with the aus
rices; however, the aus group bore similarities to the upland and bulu varieties
based on their root characteristics. Pollen mother cells of the F 1 hybrids showed
slightly lower frequencies of loose pairing at pachynema and of laggards, trivalents,
bridges, and fragments at meiosis than the parents, which also showed

such chromosomal aberrations at low frequencies. Tris-HCI-soluble seed proteins
indicated a more primitive state of the aus varieties than the two other groups,
and the protein bands also supported a close relationship between the bulu and
upland groups. The parents differed in their grouping by peroxidase isozymes.
Moreover, the 13 F 1 hybrids formed 4 distinct clusters based on the combined
esterase-peroxidase bands. Both bulu and upland varieties have more advanced
plant characteristics than does the aus group. The javanica race can be expanded
to include both bulu and upland. The upland varieties appear more diversified
than the bulu varieties because of their broad geographical dispersal.
The great antiquity of the genus Oryza and the broad dispersal of the Asian cultigen
( Oryza sativa L.) following its cultivation and domestication have led to enormous
diversity among its cultivars. The traditional cultivars (land races) probably numbered
about 100,000 before the advent of the modem varieties and the subsequent displacement
of the unimproved germplasm. The major ecogeographic races (variety-groups)
are indica and japonica (the latter also designated sinica), which were known in China
1800 yr ago as “hsien” (nonsticky) and “keng” (sticky) types, respectively. A third
group of tall and 1ongpanicled
Indonesian varieties was designated as javanica by
Morinaga (1954) following the term coined by Körnicke (1885). Javanica corresponds
to Group Ic of Terao and Mizushima (1944), the B plant type of Matsuo (1952), and
the tropical insular group (IIa and IIb) of Oka (1958). Later Oka (1983) changed the
21
designation to “tropical japonica.” Another variety-group comprising the aus varieties
of Bangladesh and eastern India, which also show high genetic affinity with either one
of the two major races, was later combined with the Indonesian cultivars as the
intermediate type by Morinaga and Kuriyama (1958).
We became intrigued by the morphological similarity between the javanica cultivars
and the traditional hill (upland) rices of Southeast Asia—tall stature, low tillering,
rather broad leaves, stout culms, long and well-exserted panicles, large and bold grains,
often awned—although they belong to distinct hydro-edaphic regimes, lowland and
upland, respectively. We were also interested in the aus varieties because they share
some of the drought avoidance mechanism of the hill rices in having moderately deep
and thick roots (Chang et al 1986).
This paper synthesizes our biosystematic studies on the genetic relationship,
karyomorphology and meiotic behavior, root characters, and electrophoretic properties
of selected cultivars in the javanica race, hill rices, and aus varieties and in the F 1
hybrids among the three groups. Our aim is to elucidate their relationships in the
multifaceted process of ecogenetic differentiation and diversification.
Materials and methods
Thirteen varieties selected from three ecotypes were used as the parents of a diallel
cross (Table 1). The pollen fertility of the parents and F 1 hybrids was determined by
staining pollen grains with I-KI. Spikelet fertility was obtained by counting the filled
spikelets of 10 panicles per entry.
For cytological observations, pollen mother cell smears were prepared following a
modified iron-acetocarmine technique (Engle et al 1969). Meiotic chromosomal
aberrations were examined from the mid-pachytene stage through anaphase II.
Table 1. Parental materials, their origin and cultural type.
IRRI accession Country of Ecocultural
Cultivar no. origin type
Baok
Ijo Gading
Hawara Batu
Rodjolele
Khao Lo
Khao Youth
Kinandang Patong
Moroberekan
Aus 61
Aus 3761
Black Gora
Dular
63-83
19248
17741
13524
16544
12904
12901
23364
12048
14725
28924
29571
40275
32561
Indonesia
Indonesia
Indonesia
Indonesia
Laos
Laos
Philippines
Guinea
Ivory Coast
Bangladesh
Bangladesh
India
India and
Bangladesh
Bulu, lowland
Bulu, lowland
Bulu, lowland
Bulu, lowland
Hill or upland
Hill or upland
Upland
Upland
Upland
Aus, dual purpose
Aus, dual purpose
Aus, dual purpose
Aus, dual purpose
22 Chang et al
Forty dehulled seeds from each variety were used in seed protein electrophoresis.
The extraction of seed protein followed the procedure used by Siddiq et al (1972),
except that whole seeds were used in this study.
Isozymes were extracted from 85-h-old germinating embryos in a 2-°C cold room.
For esterases, 300 mg of embryos were ground in 1.5 ml sodium phosphate buffer (0.2
M, pH 7.0) and homogenized. For peroxidases, 400 mg of embryos with 2 ml Tris-HCl
buffer (0.2 M, pH 6.0) was used. The homogenates were centrifuged at 20,000 g for 20
min at 4 °C. Two hundred microliters of the supernatant for esterases or 300 μl for
peroxidases was loaded over each gel. Each sample was loaded over three gels as
replications.
A modified procedure of Davis (1964) for anodic gels was adopted. Electrophoresis
was performed under a constant electric current of 2–3 mA/gel tube in a 2-°C room.
Bromophenol blue was used as the tracking dye. The gels were stained and fixed
following the method standardized by Ng (1977) for seed proteins and esterase
isozymes.
Relative mobility (R f ) was calculated by a formula modified from Weber and
Osborn (1969). Similarity index and average Euclidean distance (d jk ) were computed
by the methods of Sokal and Sneath (1963).
The root systems of 6 parents and 15 F 1 hybrids were studied using the aeroponic
culture technique modified by Armenta-Soto et al (1982). Root length, thickness, and
number were recorded at 45 d after sowing.
Results and discussion
Our findings on the parents and F 1 hybrids are presented under 6 headings: fertility of
F 1 hybrids, karyomorphological study of 9 parents, meiotic behavior of 13 parents and
146 reciprocal F 1 hybrids, Tris-HC1-soluble seed proteins, esterase and peroxidase
isozymes, and root systems under aeroponic culture.
Fertility of F 1 hybrids
The data on pollen and spikelet fertility obtained from the 13 parents and 141 reciprocal
F 1 hybrids are given in Table 2. Pollen fertility values showed more stability within a
variety-group because they are less subject to environmental influences than are
spikelet fertility values. Pollen fertility in the bulu and upland rices averaged 96.0 and
94.9%, respectively, while their F 1 hybrids averaged 85.5%. The aus varieties averaged
93.7%, and their hybrids from bulu/aus and upland/aus crosses averaged 83.2 and
85.3%, respectively. The mean values were not significantly different among crosses,
but individual hybrids ranged from 49.7% in Khao Lo/Ijo Gading to 99.9% in Moroberekan/
ljo Gading. Eight crosses showed significant differences (19–48%) with
their reciprocal crosses.
Mean spikelet fertilities of the three parental groups were lower than pollen
fertilities: 77.9–90.1%. A wide range of spikelet fertilities was obtained in the F 1
hybrids, ranging from 9.7% in Kinandang Patong/Aus 61 to 95.3% in Aus 61/Black
Cytogenetic, electrophoretic, and root studies of javanica rices 23
Table 2. Pollen and spikelet fertility of 13 parents and 141 F 1 hybrids of bulu,
upland, and aus varieties.
Pollen fertility (%) Spikelet fertility (%)
Parent or hybrid
Range Pooled mean Range Pooled mean
Bulu
Upland
Aus
Bulu/upland
Bulu/aus
Upland/aus
Within bulus
Within upland
Within aus
91.5–98.5
86.0–98.7
94.3–100
49.7–99.9
50.9–9.3
52.9–98.7
93.3–99.5
87.5–99.9
84.0–98.0
96.0 ± 2.1
94.9 ± 5.8
93.7 ± 3.9
85.5 ± 17.0
83.2 ± 16.3
85.3 ± 15.1
96.0 ± 2.1
94.9 ± 5.8
93.7 ± 3.9
76.4–81.7
63.3–91.3
79.6–92.2
28.6–92.9
2.7–91.3
30.7–93.0
52.3–85.2
46.5–92.3
82.8–96.5
77.9 ± 7.7
80.3 ± 11.1
90.1 ± 5.4
70.4 ± 19.9
65.3 ± 2.3
77.7 ± 13.8
77.9 ± 7.7
80.3 ± 11.1
90.1 ± 5.4
Gora. However, the three parental array means did not show significant differences.
Thus, the three parental groups indicated a high degree of affinity between any two.
Moreover, the F 1 pollen fertility readings were generally higher than those of earlier
studies involving indica/japonica crosses (Demeterio et al 1965, Jennings 1966,
Morinaga and Kuriyama 1958, Terao and Mizushima 1939), japonica/aus crosses
(Morinaga and Kuriyama 1955), Indian indica/Indonesian indica crosses, and Indian
indica/javanica crosses (Engle et al 1969). The fertility values were also slightly higher
than those of the F 1 hybrids of upland/semidwarf indica crosses (Lin and Chang 1981).
Meiotic behavior of 13 parents and 146 F 1 reciprocal hybrids
The commonly observed chromosomal aberrations in the 146 F 1 reciprocal hybrids
included loose pairings, univalents, chains-of-four, straggling chromosomes and
laggards, bridges and fragments, and deficiency loops (Table 3). These occurred at
rather low frequencies, rarely reaching 6.9% for any type in an F 1 hybrid. Low
frequencies of the same kind of chromosome aberration were also observed in the
parents.
No obvious correlation between the chromosomal aberrations and F 1 sterility could
be found. Only in the F 1 hybrids of Aus 61/Ij0 Gading and Ijo Gading/Kinandang
Patong crosses were significantly higher frequencies of loose pairing (20 and 10%,
respectively) as well as high spikelet sterility (10.1 and 9.7%, respectively) observed.
Fewer chromosomal aberrations were identified than in earlier indica/japonica
(Demeterio et al 1965) and in Indian indica/javanica crosses (Engle et al 1969).
Tris-HCl-soluble seed proteins
The 13 parental varieties had 32 seed protein bands, varying from variety to variety as
well as from group to group. The smallest number of protein bands (22–27) with the
largest variation was found in the aus group. The bulu group had 26–27 bands and the
upland group 26–29. This difference may suggest that the aus rices evolved earlier than
the bulus and upland rices. This finding is also in general agreement with the 12–16 seed
24 Chang et al
Table 3. Chromosome behavior at meiosis in 13 parents and 146 F 1 hybrids.
Parents (%) F 1 hybrids (%) Cells
Range X Range X (no.)
Stage Behavior a examined
Pachynema
Diplonema
and
diakinesis
Metaphase I
and
anaphase I
Metaphase II
and
anaphase II
Normal
Loose pairing
Inversion and
translocation
loops
I (2)
Others
Normal
I (2)
Chain-of-4
Ring-of-4
Others
Normal
Laggards
I (2)
I (3)
Bridges and
fragments
Others
Normal
Laggards
III
Bridges and
fragments
Others
90.9 – 100.7
0 – 3.7
0 – 3.0
0 – 3.7
0 – 2.4
89.8 – 100
0 – 4.3
0 – 4.8
0 – 1.4
0 – 2.6
90.1 – 99.5
0.4 – 3.3
0 – 1.5
0 – 1.2
0 – 2.9
0 – 5.8
97.9 – 100
0 – 1.3
0 – 0.6
0 – 1.5
0 – 0.9
97.2
1.5
0.6
0.5
0.2
95.7
1.9
1 .0
0.7
0.8
95.2
1.7
0.6
0.6
1 .0
0.9
99.2
0.4
0
0.3
0.1
80 – 100
0 – 20
0 – 4.2
0 – 6.9
0 – 6.7
89.9 – 100
0 – 5.7
0 – 4.5
0 – 3.6
0 – 3.2
90.1 – 100
0 – 5.0
0 – 2.0
0 – 1.7
0 – 2.9
0 – 2.5
93.6 – 100
0 – 3.5
0 – 2.3
0 – 3.1
0 – 2.6
96.9
1.6
0.2
0.8
0.4
95.9
1.8
1.1
0.7
0.5
95.3
1.8
0.7
0.5
0.8
0.9
97.1
0.7
0.3
0.6
0.3
509 (P)
3,570 (F 1 )
1,529 (P)
19,813 (F 1 )
2,608 (P)
31,747 (F 1 )
1,348 (P)
13,738 (F 1 )
a Others included (a) trivalent association, deficiency loops, fragments, unequal bivalents at pachynema; (b) 13II,
1 Ill + 1III + 2I + 10II, 12II +nucleolar bodies at diplonema and diakinesis; (c) persistent nucleolus, late disjunction,
2 laggards, fragments without bridges, 1 dyad with early disjunction at metaphase II and anaphase II.
straggling chromosomes, early division chromosomes, and 12II + fragments at metaphase I and anaphase I; (d)
protein bands in indica varieties, 14–15 in japonicas, and 15–17 in javanicas found by
Siddiq et al (1972). The more advanced nature of the upland rices in relation to other
ecotypes was pointed out by Chang (1976).
The mean similarity indices of seed protein profiles for the within-group and
between-group comparisons are shown in Table 4. The mean index within the upland
group (86.9%) was the highest and was significantly higher than the index within the
aus group (75.5%). Moreover, the mean index of the bulu-upland comparison (75.5%)
was significantly higher than the mean index between aus and bulu (66.8%). These
comparisons suggest that there is a close genetic relationship among the three varietygroups
and that the upland rices are closer to the bulu rices than to the aus rices. These
indices are much higher than those among indica, japonica, and javanica races as
reported by Siddiq et al (1972).
Cytogenetic, electrophoretic, and root studies of javanica rices 25
Table 4. Mean similarity indices and standard deviations (%) of electrophoretic patterns
showing within- and between-group comparisons. a
Mean similarity indices ± SD (%)
Seed protein Esterase Peroxidase
profiles zymograms zymograms
Varietal comparison
Within-group
Aus
Bulu
Upland
Between-group
Aus vs bulu
Aus vs upland
Bulu vs upland
75.5 ± 8.7 bc
83.3 ± 7.0 ab
86.9 ± 7.0 a
66.8 ± 4.4 e
71.1 ± 8.7 de
75.5 ± 5.3 cd
93.8 ± 4.0 a
85.4 ± 9.4 ab
85.0 ± 7.9 ab
73.4 ± 4.3 c
82.5 ± 9.2 b
71.3 ± 10.0 c
100 ± 0.0 a
96.4 ± 3.9 abc
92.9 ± 5.8 bc
94.7 ± 3.2 bc
95.7 ± 5.9 abc
92.5 ± 4.3 c
a In a column, mean similarity indices followed by a common letter are not significantly different at the 5% level
by LSD test.
Esterase and peroxidase isozymes
Sixteen bands of esterase isozymes in the 13 varieties were identified. Eight bands (2,
3, 11, 12, 13, 14, 15, and 16) were constantly present, while the other 8 were variable.
There were as many as 10 zymogram patterns in the 13 varieties. No hybrid band was
detected in any of the 138 hybrids studied.
On the other hand, 12 isozymic bands of peroxidase were found in the 13 varieties,
and little variation was detected. It was difficult to identify or pinpoint each varietal
group by its zymogram patterns of both enzymes; authors of previous studies also
failed to do so (Ng 1977, Palanichamy and Siddiq 1977, Shahi et al 1969). Nevertheless,
it is likely that the three groups could be compared by the presence or absence of
certain group-specific bands. For esterases, band 7 at R f 0.39 was limited to the aus and
upland varieties, whereas band 8 at R f 0.41 was present in the bulu rices and in only one
upland variety, Kinandang Patong. The bulu rices could also be separated from the aus
and upland rices by the absence of peroxidase band 5.
High values of the mean similarity indices on zymogram patterns of both enzymes
for within-group and between-group comparisons were observed (Table 4). Unlike
with seed protein electrophoresis, the highest value was observed within the aus
varieties in both cases (93.75% for esterases and 100% for peroxidases). The mean
indices of the aus-upland comparison were significantly higher than the other mean
indices of between-group comparisons, suggesting that the aus and upland varieties
share more similar isozymic banding patterns. This is not surprising, because many
workers generally place the two groups under the indica race. This finding also
coincides with those of Shahi et al (1969) and Second (1982) that the cultivars of
O. sativa could be grouped primarily into two, the indica and japonica clusters, with
a range of intermediate varieties. Nakagahra (1978) classified rice varieties into five
groups on the basis of esterase isozymes in the leaves, and placed the hill and mountain
rices of Southeast Asia under Javanica 1 and 2. Glaszmann (1986) combined Chinese
“keng,” Japanese varieties, upland rices, and bulus under enzymatic group VI.
26 Chang et al
1. Distribution of the 13 hybrid populations based on mean frequencies of all isozymic bands of the 2
enzymes (esterase and peroxidase).
The isozymic bands of esterases and peroxidases found in the 13 parents were also
found in their hybrids, but with different frequencies. When the distribution frequencies
of all esterase or peroxidase bands in the parental arrays of 138 F 1 hybrids were
averaged and arranged in a 2-way diagram (Fig. 1), 4 clusters could be identified:
1) all hybrids of the four aus varieties, 2) hybrids from 3 bulu varieties (Baok,
Rodjolele, and Ijo Gading) and 1 upland variety (63-83), 3) hybrids of 1 bulu variety
(Hawara Batu) and 3 upland varieties (Khao Lo, Khao Youth, and Moroberekan), and
4) hybrid populations derived from the upland variety Kinandang Patong, located in
the upper right comer and away from the three other clusters. Kinandang Patong is an
old cultivar from the Philippines (Bortolini 1844), morphologically different from
most other upland varieties in panicle and grain features.
Root systems under aeroponic culture
Root length, thickness, and number were examined.
Root length. There was no significant difference in root length among groups (Table
5). All the F 1 hybrids produced longer roots than their parents. There was no significant
difference among the F 1 hybrids within each of the three groups or among groups.
However, root lengths of F 1 hybrids among groups were longer than those of the withingroup
F 1 hybrids.
Cytogenetic, electrophoretic, and root studies of javanica rices 27
Root thickness. The upland and aus groups had significantly thicker roots than those
of the bulu varieties (Table 5). The upland group produced slightly thicker roots than
the aus group. The root diameter surpassed that of most irrigated lowland varieties
(0.51 mm and upward), although some varieties grown in drought-prone rainfed
lowland areas also had thick roots (up to 0.88 mm), but not as thick as those of
Moroberekan (1.26 mm) (Loresto et al 1983).
In between-group F 1 hybrids, the thickest roots were observed in the aus/upland
group—significantly thicker than those of the F 1 hybrids of aus/bulu crosses, but not
different from those of the F 1 hybrids of bulu/upland crosses. The within-group F 1
hybrids also showed that the bulu group had the thinnest roots (0.69 mm). Our earlier
findings indicate that both dominant and recessive genes control root thickness in
different crosses (Armenta-Soto et al 1983, Chang et al 1986).
Root number. The bulu varieties had a significantly higher root number than the
upland and aus varieties (Table 5).
The root number of F 1 hybrids in crosses between groups was generally higher than
that of the parents. The aus/upland F 1 hybrids gave the lowest root number, while
crosses with the bulus gave significantly higher root number. Within-group crosses
showed higher root numbers of the bulu group than those of the aus and upland groups.
Our past studies have indicated the dominant nature of high root number (Armenta-
Soto et al 1983; Chang et al 1982, 1986).
Among the three root characters in our earlier experiments, thickness had higher
narrow sense heritability estimates (62%) than length (60%) and number (44%)
(Armenta-Soto et al 1983), and, therefore, was a more important criterion for comparison.
Based on a matrix of correlation coefficients, Moroberekan in the upland group and
Aus 61 in the aus group ( r = 0.54 *) showed similarity in both root length and number.
On the other hand, similarities in root thickness between Kinandang Patong and Aus
Table 5. Mean root length, thickness, and number of 6 parents and 15 F 1
hybrids grown in aeroponic culture.
Parent or Root length Root thickness Root
hybrid (cm) (mm) number
Parental
Bulu
Upland
Aus
Between-group
Bulu/upland
Bulu/aus
Upland/aus
Within-group
Bulu/bulu
Upland/upland
Aus/aus
92.8
86.4
84.7
104.1
1203.9
102.8
91.0
95.6
91.1
0.58
0.82
0.80
0.84
0.79
0.90
0.69
0.88
0.81
61
25
28
55
57
47
65
35
35
28 Chang et al
61 were shown ( r = 0.69**), and also between Dular (an aus) and Rodjolele (a bulu)
( r = 0.61**). The similarity could arise from the fact that aus and upland varieties are
sown in dry soil and grown under rainfed culture.
Conclusions
A high level of genetic affinity among the three variety-groups as shown by fertility in
the F 1 hybrids and by essentially normal meiotic behavior has been confirmed. It
appears logical to expand the javanica race to include the bulus and the upland (hill)
rices, although they differ in hydro-edaphic regime. In most cases, both types produce
thicker roots than conventional irrigated lowland varieties, an advanced feature in
varietal diversification. Other plant and grain characteristics such as organ size, low
photoperiodicity, and weak grain dormancy of the two groups are also more advanced
features (Chang 1976). Many hill rices of Southeast Asia have glabrous leaves.
Another distinctive feature is their intermediate amylose content (18–25%), lying
between those of the typical indica and japonica types (Chang 1988), a crucial factor
overlooked by rice workers in the past.
Table 6. Ecogeographic races of Oryza sativa: comparison of their morphological
and physiological characteristics (adapted from Chang 1988).
Indica Sinica (or japonica) Javanica
Broad to narrow, light
green leaves
Long to short, slender,
somewhat flat grains
Profuse tillering
Tall to intermediate
plant stature
Mostly awnless
Thin, short hairs on
lemma and palea
Easy shattering
Soft plant tissues
Varying sensitivity to
photoperiod
23–31 % amylose
Variable gelatinization
temperatures (low or
intermediate)
Narrow, dark green leaves
Short, roundish grains
Medium tillering
Short to intermediate
plant stature
Awnless to long-awned
Dense, long hairs on
lemma and palea
Low shattering
Hard plant tissues
Zero to low sensitivity
to photoperiod
10–24% amylose
Low gelatinization
temperature
Broad, stiff, light green
leaves
Long, broad, thick grains
Low tillering
Tall plant stature
Long-awned or awnless
Long hairs on lemma and
palea
Low shattering
Hard plant tissues
Low sensitivity to photoperiod
20–25% amylose
Low gelatinization
temperature
Cytogenetic, electrophoretic, and root studies of javanica rices 29
The aus group is also closely related to the bulu and upland groups but appears to
have differentiated earlier. Whether the aus varieties, which are also called “upland’
varieties by some Indian workers, are the progenitors of the typical hill rices or not
could be the subject of further study. The aus group generally has intermediate to high
(21–30%) amylose content.
Since variation among rice cultivars is continuous, and many cultivars have been
transported by rice growers from one geographic area to another, critical studies should
include an analysis of crop ecosystem in relation to varietal differentiation and
diversification. The necessity of having multiple lines of evidence and an integrated
research approach in interpreting evidence for the origin and dispersal of cultivated
plants has been well expounded by Harlan and de Wet (1973).
The morphological and physiological characteristics of the three ecogeographic
races are contrasted in Table 6.