Estimation of the outcrossing rate for annual Asian wild rice under field conditions.

The Asian wild rice, Oryza rufipogon, has partial outcrossing behavior and shows high levels of genetic variation. To estimate an accurate outcrossing rate of annual form of O. rufipogon, two backcross lines (Lines R1 and R2) between Oryza sativa Nipponbare and O. rufipogon W630 were examined under field conditions. A chromosome survey confirmed that these lines had wild chromosomal segments at more than 92% of marker loci. As for the traits of glume, stamen and pistil, Line R1 showed similar floral morphology as that of O. rufipogon W630, whereas Line R2 had larger glumes. In 2005, 2006 and 2008, a total of 22 backcross plants were planted in the middle of wild rice plots. The successive progenies of each plant were examined using microsatellite markers that could clearly detect self-pollination and outcrossing. The outcrossing rates of Line R1 plants ranged from 4.04% to 25.50% with an average of 10.20%. This indicates that cross-pollination of wild rice is a chance event affected by many environmental factors. The outcrossing rates of Line R2 plants also varied, however, no significant difference was observed between the averages of Lines R1 and R2, suggesting that the glume sizes are not critical for outcrossing ability.

Introduction

Rice (Oryza sativa L.) is an important crop and a major source of food for more than one-third of the world’s population. Cultivated rice has evolved from the Asian wild species, Oryza rufipogon Griff. (Oka 1988). O. rufipogon is widely distributed in tropical regions and can be found, for example, in swamps, deep and shallow water and paddy ditches (Vaughan 1994). Because this wild species has an ability to outcross with cultivated rice, many hybrids and intermediate plants are found around paddy fields as a result of mutual gene flow. O. rufipogon contains many useful genes that have the potential to improve rice cultivars, such as insect resistance, abiotic stress tolerance and yield potential (Brar and Khush 1997, Tanksley and McCouch 1997). However, recent industrial development has destroyed some of the wild rice habitats. Gene introgression from cultivars has also disturbed the genetic structure of wild rice. To conserve the genetic variation among natural wild populations, we should understand the reproductive nature of wild rice.

Cultivated rice is predominantly self-pollinated and has lower outcrossing ability than O. rufipogon. According to Messeguer , the natural cross-pollination rates of O. sativa are less than 1%; these data are based on old experimental records of three Asian countries. Recently, the outcrossing rates of rice cultivars were examined with respect to transgene escapes. The rates of gene flow between cultivars were reported to be lower than 0.1% in a normal side-by-side plot design (Endo , Messeguer ). In contrast, several studies that were conducted to estimate the outcrossing rates among natural populations of O. rufipogon revealed outcrossing rates ranging from 4.3% to 55.9% and suggested that the perennial forms of O. rufipogon tend to show higher rates than the annual forms (Barbier 1989, Oka 1988). These estimated values were calculated based on a mixed-mating model with random pollen distribution, i.e., the genotypes were compared between mother plants and their progenies, taking into account allele frequencies and the probabilities of non-discernible outcross events in the populations. Therefore, an accurate outcrossing rate of O. rufipogon still remains to be elucidated.

Chen studied the rates of gene flow from cultivars to wild rice under field conditions. They planted O. rufipogon (perennial form) and Chinese cultivars (Minghui-63) in alternating rows with 50 cm distance and examined the ratio of hybrids among the wild progenies. By using species-specific microsatellite markers, they determined that the direct gene flow rates were 1.21–2.19%. These rates are relatively lower than the outcrossing rates estimated among wild rice populations. This may mainly be attributed to the different heading dates of two accessions. In addition, the spikelet opening time in terms of the pollen emission period may influence the gene flow between wild plants and cultivars. In general, wild rice spikelets open earlier than those of cultivars in the morning, and most of the pollen grains are functional for only three minutes after emission (Oka 1988). Consequently, the flow of genes from cultivars does not correspond to the real outcrossing rates of O. rufipogon.

It is difficult to detect the accurate outcrossing rates of wild rice under natural conditions, because the high levels of genetic variation in wild populations can cause non-discernible outcrossing events. In addition, the survey of genotype comparison between mother plants and their progenies (or seeds) is limited in scale because wild rice has strong seed-shattering behavior. Therefore, in this study the outcrossing rates of O. rufipogon were examined under field conditions. To detect the direct outcrossing rates, back-crossed plants with similar morphology as that of the wild recurrent plant of O. rufipogon were used. They were planted in the middle of plots with the wild recurrent plants, and their outcrossed progenies were identified by microsatellite markers. The objective of the present study was to estimate the outcrossing rates between wild rice plants under field conditions.

Materials and Methods

Plant materials

An annual form of Asian wild rice (O. rufipogon W630 from Myanmar), a cultivated rice variety (O. sativa Japonica Nipponbare) and their backcrossed progenies were used in this study. The wild accession was obtained from the National Institute of Genetics, Japan.

Previously, about 200 BC2F1 plants were produced from a backcross between O. sativa Nipponbare (donor parent) and O. rufipogon W630 (recurrent parent). These plants were further self-pollinated and the BC2F2 lines were obtained. In theory, these backcross lines have 87.5% of the wild genome. Among them, two lines (Lines R1 and R2), which had similar morphology and heading dates as those of the wild parents, were selected and their selfed seeds at BC2F3 generation were used for further experiments.

Measurement of floral traits

Floral traits are considered to play an important role in outcrossing ability. The following 13 floral traits were measured for O. rufipogon W630, O. sativa Nipponbare and the two backcrossed lines (Lines R1 and R2): length, thickness and width of lemma and palea, awn length, anther length, stigma length and breadth, style length, stigma exertion, spikelet opening angle. The seven glume traits (length, thickness and width of the lemma and palea, awn length) were evaluated with 20 spikelets per plant and other six traits were examined with five spikelets. An average of these spikelets was calculated for six plants from each accession or line. Stigma exsertion was examined as percentage of spikelets showing exserted stigma on one or both sides, and spikelet opening angle was measured as an angle between lemma and palea of flowering spikelet. The rest traits were evaluated after Uga .

Field experimental design

In the summer of 2005, 2006 and 2008, the outcross experiments were independently conducted in the university paddy field in Kobe, Japan (34°43′N, 135°14′E). Each year, in the second week of June, about four-week-old seedlings (BC2F3 generation) of the two backcross lines and O. rufipogon W630 were transplanted at intervals of 20 cm in a plot as shown in Fig. 1. Each backcross plant was surrounded by two rows of O. rufipogon W630. Three of each backcross line were planted in the first year of 2005 and four of each line were used in the second and third years (2006 and 2008).

Fig. 1

Field experimental design for the estimation of the outcrossing rate of wild rice. (A) Schematic layout of the plot. White and black circles represent the positions of the plants of O. rufipogon and the backcross line, respectively. (B) Overview of the plot one month after transplanting. (C) Flowering peak in the plot (two and half months after transplanting).

Seed collection

In 2005, 2006 and 2008, the heading date of O. rufipogon W630 was around August 15 and the peak of the flowering period was from August 20 to 30 (Fig. 1). The backcross plants of the two lines also flowered at the same time as the surrounding wild plants. About two weeks after the heading of the wild rice, the backcross plants were transferred from the paddy field to the pots. Mature seeds were collected by hand from each plant every 2–4 days. The seeds were collected from August 29 to September 8 in 2005, from August 31 to September 11 in 2006 and from August 31 to September 12 in 2008.

Chromosomal survey of the backcross lines

The BC2F2 plants of two backcross lines were selected for the field experiment. To examine their chromosomal composition, total DNA was extracted from the leaves according to the method described by Ikeda . PCR was carried out with microsatellite markers almost covering the genome (Chen , Temnykh ). In total, 89 microsatellite loci that showed polymorphisms between O. rufipogon W630 and O. sativa Nipponbare were surveyed. PCR was performed in a 25 μl reaction volume containing 0.2 μM of each primer, 100 μM of each dNTP, 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl2, 0.1% Triton X-100, 2.5 μl of template DNA, and 1 unit of Taq DNA polymerase. The PCR profile was as follows: initial denaturation at 94°C for 5 min, followed by 35 cycles of 94°C for 1 min, 55°C for 1 min, 72°C for 2 min, and extension for 5 min at 72°C. The amplified products were analyzed by electrophoresis in a 4% polyacrylamide denaturing gel, and the banding patterns were visualized by the non-radioactive silver staining method described by Panaud .

Determination of the outcrossing rate

The mature seeds collected from the backcross plants in 2005 and 2006 were soaked in water and the germinated seeds were planted in seedling cases in the following year. Total DNA was then extracted from the young leaves. In 2008, total DNA was extracted directly from seed embryos to avoid the germination step, which was strongly influenced by seed dormancy.

A chromosome survey of the backcross lines revealed that Line R1 and Line R2 were homozygous with the O. sativa Nipponbare alleles at the RM3 and RM229 micro-satellite loci, respectively. These markers were used to determine the genotypes of their progenies (or seeds). Because self-pollinated and outcrossed seeds can be distinguished based on their marker genotypes, i.e., Nipponbare homozygote and heterozygote, respectively, the outcrossing rate of each backcross plant could be determined by calculating the number of seeds with heterozygous alleles out of the total number of seeds examined.

Results

Chromosomal composition of the two backcross lines

Two BC2F2 backcross lines (Lines R1 and R2) were selected for this study. To examine the chromosomal composition of these lines, 89 polymorphic microsatellite loci were surveyed between both parents. Three marker loci (RM3, RM18, RM234) were found to be homozygous with the Nipponbare allele in Line R1, and four marker loci (RM21, RM206, RM209, RM229) were homozygous in Line R2 (Table 1). Some loci were also heterozygous in these two lines. When the heterozygous locus was defined as the half-value of the genetic composition compared to the homozygous locus, the percentages of loci with wild alleles in Lines R1 and R2 were found to be 92.7% and 94.4%, respectively. These results indicate that these two backcross lines had almost the same genetic composition as that of the wild rice.

Table 1

Marker genotypes of two backcrossed lines, Line R1 and Line R2, at 89 microsatellite loci

Chr.	Locusa	Genotypeb
		Line R1	Line R2
1	RM84	3	3
	RM220	3	3
	RM1	3	3
	RM243	3	3
	RM23	3	3
	RM35	3	3
	RM9	3	3
	RM5	3	3
	RM237	3	3
	RM246	3	3
	RM212	3	3
	RM265	3	3
	RM315	3	3
	RM14	3	3
2	RM211	3	3
	RM53	3	3
	RM8	3	3
	RM29	3	3
	RM263	3	3
	RM221	3	3
	RM6	3	3
	RM240	3	3
	RM250	3	3
	RM208	3	3
	RM213	3	3
	RM207	3	3
	RM48	3	3
3	RM60	3	3
	RM22	3	3
	RM231	3	3
	RM7	3	3
	RM232	3	3
	RM251	3	3
	RM16	3	3
	RM49	3	3
	RM55	3	3
	RM227	3	3
4	RM261	3	3
	RM252	2	2
	RM241	2	2
	RM131	3	3
	RM349	3	3
5	RM13	3	3
	RM164	3	3
	RM26	3	3
6	RM204	3	3
	RM225	3	3
	RM253	3	3
	RM50	3	3
	RM3	1	3
	RM30	3	3
7	RM82	2	3
	RM214	3	3
	RM2	3	3
	RM11	3	3
	RM10	2	3
	RM234	1	3
	RM18	1	3
	RM248	3	3
8	RM25	3	3
	RM42	3	3
	RM223	3	3
	RM210	3	3
	RM256	3	3
	RM230	3	3
9	RM219	3	3
	RM242	3	3
	RM201	3	3
	RM215	3	3
	RM245	3	3
	RM205	3	3
10	RM222	3	3
	RM216	3	3
	RM258	3	3
11	RM4B	2	3
	RM20B	2	3
	RM167	2	3
	RM202	3	3
	RM209	3	1
	RM229	3	1
	RM21	3	1
	RM206	3	1
	RM224	3	3
12	RM20A	3	3
	RM4A	3	3
	RM19	3	3
	RM247	3	3
	RM260	3	3
	RM17	3	3

Microsatellite markers were listed according to the map positions by Chen and Temnykh .

1: Homozygote of O. sativa Nipponbare alleles. 2: Heterozygote. 3: Homozygote of wild alleles.

Floral traits of the two backcross lines

A total of 13 floral traits of the two backcross lines were compared with those of O. rufipogon W630 (Fig. 2). As a result, Line R1 was found to have similar floral traits as the recurrent parent of O. rufipogon W630 (Table 2). Line R2 also showed similar traits of reproductive organ and awn, however, significant differences were observed for length, thickness and width of lemma and palea. Although Line R2 had larger glumes than O. rufipogon W630, both backcross lines were further used for the field experiment.

Fig. 2

Floral morphology of O. rufipogon W630 and two backcross lines (Line R1 and Line R2). Glumes before flowering (A) and after flowering (B). Npb: O. sativa Nipponbare, W630: O. rufipogon W630.

Table 2

Floral traits measured for O. rufipogon W630, O. sativa Nipponbare and two backcross lines. Value indicates mean ± s.d. The mean values of two lines are compared with that of O. rufipogon W630 by t-test

Accession	Lemma length (mm)	Palea length (mm)	Lemma thickness (mm)	Palea thickness (mm)	Lemma width (mm)	Palea width (mm)	Awn length (cm)	Anther length (mm)	Stigma length (mm)	Stigma breadth (mm)	Style length (mm)	Stigma exertion (%)	Opening angle (degree)
O. rufipogon W630	7.23 ± 0.01	7.30 ± 0.02	1.06 ± 0.01	0.90 ± 0.04	2.06 ± 0.01	1.21 ± 0.01	8.92 ± 0.09	2.14 ± 0.01	1.17 ± 0.05	0.47 ± 0.02	0.71 ± 0.03	100 ± 0.0	29.9 ± 1.2
Line R1	7.20 ± 0.04	7.32 ± 0.04	1.06 ± 0.01	0.91 ± 0.02	2.05 ± 0.03	1.21 ± 0.02	9.16 ± 0.43	2.14 ± 0.01	1.16 ± 0.03	0.46 ± 0.02	0.69 ± 0.02	100 ± 0.0	30.7 ± 1.1
Line R2	7.32 ± 0.04**	7.41 ± 0.06**	1.13 ± 0.04**	1.00 ± 0.05**	2.13 ± 0.03**	1.32 ± 0.06**	9.15 ± 0.26	2.14 ± 0.02	1.16 ± 0.02	0.46 ± 0.02	0.70 ± 0.02	100 ± 0.0	29.6 ± 0.6
O. sativa Nipponbare	6.57 ± 0.07	6.79 ± 0.10	1.25 ± 0.02	1.00 ± 0.01	2.45 ± 0.02	1.60 ± 0.02	n.a.	1.94 ± 0.03	0.94 ± 0.03	0.52 ± 0.01	0.59 ± 0.02	0.0 ± 0.0	23.9 ± 1.1

Significant at 1% level.

Determination of outcross events

The two backcross lines had the same chromosomal composition as that of the wild rice except for a few chromosomal segments with Nipponbare alleles. Marker loci in these regions are good indicators that can be used to monitor outcross events. Of the homozygous marker loci with Nipponbare alleles, RM3 and RM229 were used to examine the genotypes of the progenies of Line R1 and Line R2, respectively. As shown in Fig. 3, the self-pollinated and out-crossed progenies could be clearly distinguished based on the marker genotypes of the electrophoretic bands patterns.

Fig. 3

An example of the determination of outcross events among the progenies of the backcross line. (A) Backcross Line R1 has homozygous Nipponbare alleles at the RM3 microsatellite locus on chromosome 6 in the genetic background of O. rufipogon. Two different genotypes at RM3 were generated among the successive progenies by self-pollination and outcrossing. (B) Marker genotypes of the progenies (nos. 1–14) are clearly distinguished by electrophoretic band patterns. Nos. 2, 7 and 8: Outcrossed progenies showing heterozygous band patterns at RM3.

Estimation of the outcrossing rate of wild rice

In 2005 and 2006, total DNA was extracted from the leaves of the young seedlings. However, about half of the progenies could not be examined because they had strong seed dormancy. Therefore, in 2008, total DNA was extracted directly from the seed embryos to avoid the germination step. Thus, more progenies could be analyzed in 2008 than in 2005 or 2006.

The outcrossing rate of each backcross plant was determined by calculating the number of outcrossed progenies out of the total number of progenies examined. Table 3 shows the outcrossing rates observed in 2005, 2006 and 2008. The outcrossing rates of Line R1 backcross plants ranged from 4.04% (plant no. 08-R1-2) to 25.50% (05-R1-2). The average value was 10.20% with a standard deviation of 7.94%. Plants from Line R2 showed outcrossing rates ranging from 1.63% (06-R2-2) to 10.53% (05-R2-3) with an average of 5.55%. To examine the difference between the outcrossing rates of Line R1 and Line R2 plants, a t-test was performed using these values. No significant difference (p = 0.090) was observed between these averages.

Table 3

Outcrossing rates of two backcross lines based on the number of the outcrossed progenies against total number of progenies in 2005, 2006 and 2008

(A) Results in 2005
Plant no.	Number of progenies examineda								Outcrossing rate (%)
	Collection date							Total
	30-Aug	2-Sep	6-Sep	9-Sep
Line R1
05-R1-1	44 (9)	31 (11)	35 (8)	4 (1)				114 (29)	25.44
05-R1-2	56 (7)	40 (9)	51 (19)	6 (4)				153 (39)	25.50
05-R1-3	36 (10)	58 (6)	45 (3)	9 (0)				148 (19)	12.84
Line R2
05-R2-1	19 (1)	32 (5)	27 (0)	0 (0)				78 (6)	7.69
05-R2-2	23 (1)	34 (0)	21 (2)	0 (0)				78 (3)	3.85
05-R2-3	50 (5)	46 (4)	44 (4)	12 (3)				152 (16)	10.53
(B) Results in 2006
Plant no.	Number of progenies examineda								Outcrossing rate (%)
	Collection date							Total
	1-Sep	5-Sep	9-Sep	12-Sep
Line R1
06-R1-1	17 (2)	23 (1)	17 (2)	9 (0)				66 (5)	7.58
06-R1-2	23 (1)	40 (2)	8 (0)	3 (0)				74 (3)	4.05
06-R1-3	29 (5)	18 (0)	6 (0)	5 (0)				58 (5)	8.62
06-R1-4	29 (1)	17 (3)	4 (0)	6 (0)				56 (4)	7.14
Line R2
06-R2-1	25 (0)	95 (7)	129 (5)	33 (1)				282 (13)	4.61
06-R2-2	72 (0)	141 (1)	128 (3)	28 (2)				369 (6)	1.63
06-R2-3	53 (0)	118 (1)	62 (3)	20 (1)				253 (5)	1.98
06-R2-4	48 (7)	140 (4)	77 (2)	40 (0)				305 (13)	4.26
(C) Results in 2008
Plant no.	Number of progenies examineda								Outcrossing rate (%)
	Collection date							Total
	1-Sep	3-Sep	5-Sep	7-Sep	9-Sep	11-Sep	13-Sep
Line R1
08-R1-1	53 (0)	23 (0)	26 (1)	46 (2)	46 (5)	30 (1)	37 (4)	261 (13)	4.98
08-R1-2	55 (4)	46 (3)	39 (0)	52 (2)	47 (2)	39 (1)	19 (0)	297 (12)	4.04
08-R1-3	49 (9)	47 (2)	46 (1)	38 (1)	26 (1)	39 (0)	17 (2)	262 (16)	6.11
08-R1-4	68 (2)	18 (3)	42 (1)	37 (1)	33 (4)	23 (2)	18 (1)	239 (14)	5.86
Line R2
08-R2-1	43 (9)	34 (1)	90 (2)	68 (1)	37 (0)	35 (1)	33 (2)	340 (16)	4.71
08-R2-2	27 (1)	48 (2)	55 (5)	64 (10)	40 (8)	48 (2)	33 (5)	315 (33)	10.48
08-R2-3	50 (3)	84 (2)	81 (2)	27 (1)	38 (1)	47 (0)	29 (2)	356 (11)	3.09
08-R2-4	10 (0)	25 (1)	33 (1)	34 (2)	21 (3)	27 (5)	9 (1)	159 (13)	8.18

Number of the outcrossed progenies is given in parenthesis.

Discussion

Outcross behavior of wild rice

The common cultivated rice, O. sativa, is estimated to have been domesticated from Asian wild rice, O. rufipogon, about 10,000 years ago (Mannion 1999). During domestication, cultivated rice obtained various morphological and physiological traits such as non-shattering seeds, weak seed dormancy, compact floral organs, uniform seed maturity and high yielding potential. Of these, the size changes of stamen and pistil may directly relate to the ability to self-pollinate. In contrast, wild rice still keeps the outcross behavior and maintains a high level of genetic variation in the natural populations. Compared with cultivated rice, wild rice species have wide genetic variation, and they are rich resources of various genes that can be utilized for future rice improvement (Brar and Khush 1997). In order to understand the outcross behavior of wild rice, several studies have tried to estimate outcrossing rates of O. rufipogon. Oka (1988) summarized the outcrossing rates to be ranging from 4.3% to 55.9%.

In the present study, the outcrossing rates of O. rufipogon W630 (annual form) were estimated under field conditions using two backcross lines (Lines R1 and R2). They had wild chromosomal segments at more than 92% of marker loci, however, only Line R1 showed similar floral morphology as that of wild rice. The outcrossing rates of Line R1 backcross plants ranged from 4.04% to 25.50% with an average of 10.20%. These values are similar to those estimated for the annual form of O. rufipogon (Barbier 1989, Oka 1988).

Temporal variation in the outcrossing rate

The backcross plants of Line R1 were used for the determination of the outcrossing rates over three years. Although the plants had similar genetic backgrounds to the recurrent wild rice, the outcrossing rates varied considerably. The average values observed were 21.26%, 6.85% and 5.25% in 2005, 2006 and 2008, respectively. A high significant difference (p < 0.01) was detected among them by the means of the ANOVA test. These results indicate that cross-pollination is a chance event affected by many environmental factors such as weather, temperature and humidity. Previously, Morishima and Barbier (1990) estimated the out-crossing rates on a single-plant basis in a wild rice population using isozyme data. The reported values ranged from 0% to 90% with an average of 45.5% and a standard deviation of 21.8%. These authors analyzed more than 20 plants with different genotypes, so the variation in the out-crossing rate was mainly due to different plant and flower morphologies, pollen viabilities, pollen distribution and plant location.

Floral traits and outcrossing rates

Based on the marker genotypes, Line R2 was estimated to have almost 95% of wild chromosomal segments in the genome. Among 13 floral traits examined, Line R2 showed larger sizes of lemma and palea than those of O. rufipogon. The average outcrossing rate of Line R2 backcross plants was 5.55% with a standard deviation of 3.18%. Although the average value was lower than that of Line R1, no significant difference was observed between them by t-test. This suggests that glume sizes of Line R2 are not critical for out-crossing ability. Probably, stamen and pistil characters may strongly be responsible for pollen dispersal and reception.

Gene flow from cultivar to wild rice

O. rufipogon has a high cross-affinity with rice cultivars, and gene flow easily occurs between them. In tropical Asian countries, the hybrids or intermediate plants between wild and cultivated rice are often present around paddy fields. Under field conditions, several studies have been conducted to evaluate the gene flow rates from cultivars to weedy or wild rice (Chen , Zhang ). According to the results of these studies, the rates of gene flow to weedy and wild rice were <1% and 1.21–2.19%, respectively. These values are much lower than the wild outcrossing rates we obtained in this study (average 10.20%; maximum 25.50%). However, once genes are introduced from cultivars, they have the potential to spread quickly and widely in wild and weedy populations. Since O. rufipogon is one of the useful genetic resources for future breeding programs, we should be aware of the ecological risks of gene flow to maintain the genetic variation among natural wild populations.