[ホーム発表論文>本文]

吉田総合案内へ戻る.

本頁は Zuriat 6(1) : 48〜55 (1995) に掲載したものです.

BREEDING FOR HIGH PALATABILITY RICH RICE BY ANTHER CULTURE

K.F.Oosatol, Y.Hamachi1 and T.Yoshida2
1 Fukuoka Agricultural Research Center, Chikushino, Fukuoka, 818, Japan
2 Kyushu University, Fukuoka, 812-81, Japan

Keywords. palatability, rice, anther culture

Abstract
Breeding procedure by anther culture for high palatability Japanese type rice cultivars was detailed with special emphasis on comparison between means of doubled haploid (DH) and conventional breeding (CB) lines. No difference was found between the mean palatability of doubled haploid and conventional breeding lines. During selections of conventional breeding lines, no selection for palatability was conducted and only selection by observation for grain appearance was conducted, which had little effect on palatability. The mean yield of doubled haploid lines was lower than conventional breeding lines. Population size in conventional breeding lines was much larger than that of doubled haploid lines. During selections of conventional breeding lines, yield was not measured but the selections by observation for good plan type must have resulted into high yield conventional breeding lines. In conclusion, rice lines with high palatability can be obtained by anther culture method as by conventional breeding method. To obtain high yield lines, It is necessary to culture more anthers or increase the efficiency of each process of anther culture and select doubled hoploid lines from a large population size.

Introduction
Anther culture has the potential to product homozygous lines from crosses more rapidly than conventional breeding methods. Several rice cultivars were released by anther culture (Maeda et al., 1990; Kunihiro et al., 1993). Agronomic characters of rice lines derived from anther culture were evaluated (Oono, 1975; Wakasa, 1982; Shinbashi and Aikawa, 1986). However, comparison of the field performance and palatability between doubled haploid (DH) and conventional breeding (CB) lines to determine the relative efficiency of these breeding methods is scarce in rice. In this paper, we will report the details of breeding by anther culture for high palatability Japonica type rice cultivars with special emphasis on comparison between means of DH and CB lines for field performance and palatability to study whether DH method offered the same opportunities as CB method.

Several factors could affect the means and variances of DH lines : (i) reduced opportunities for genetic recombination when linkage is present, (ii) possibility of gametophytic selection in anther culture, and (iii) variation induced by anther culture. Previous studies comparing DH and CB lines of other cereal crops showed that DH lines had lower yield in wheat (Baenziger et al.,1989), but not different in barley (Choo et al., 1982) and wheat (Henry et al.,1988, Mitchell et al., 1992). Lower grain weight of DH lines was reported in triticale (Charmet and Brandlard 1985).

One of the most important aspects of DH lines is homozygosity. If CB lines have segregation for some traits, DH lines must increase the selection efficiency for the traits. Hornogeneity within a field-grown line was also compared between DH and CB lines.

Materials and Methods
DH lines were obtained by anther culture using the procedure of Oono (1975). Generally, F1 plants obtained by previous year's crossing were grown in a greenhouse in early spring, anthers were placed on a modified N6 medium, callus were transferred to a regeneration medium, regenerated plantlets were acclimated, haploid plants were treated with colchicine, they were grown to maturity in a greenhouse and the seeds derived originally from a different anther were harvested individually in winter. Next spring, DH lines consisting of 26 plants were planted in a field for seed increase and preliminary evaluation test, except lines tested in 1991. Promising lines having good plant type and good grain appearance were selected by observation for the next year's field test. Lines tested in 1991 were seed-increased during winter-early spring of 1990-1991 in a greenhouse and without preliminary evaluation test, they were field-tested. DH lines tested in 1991, 1993 and 1994 derived from crosses which were made generally in 1989, 1990 and 1991, respectively.

A small number of clonal variations induced by anther culture were detected. All of them were dwarfing and easily distinguishable. They were discarded during the seed increase and preliminary field test. Other DH lines seemed having no mutation, judging by observation.

  DH lines tested in 1991 derived from crosses of Akitakomachi/ Hayahikari, Aichi 80/Koshihikari and Aichi 80/Hinohikari. Lines tested in 1993 derived from Saikai 176/Mineasahi, Kinuhikari/Saikai 190 and Mineasahi/Saikai 190. Lines tested in 1994 derived from Nankai 109/Kinuhikari, Saikai 196/Chikushi 6, Kinuhikari/Saikai 190 and Chubu 56/ Hinohikari. All cross parents are well adapted to Northern Kyushu and relatively high yield and high palatability cultivars. Population size of breeding procedure are summarized in Table 1. Seventeen, 14 and 13 DH lines in 1991, 1993 and 1994, respectively were field-grown. They were transplanted on 9-12 of June at Fukuoka Agricultural Research Center with one or two replicates. Four row plots were planted with rows 30 cm apart, hills 15 cm apart and 3.5-6 m long. One hill consisted of 34 plants. Nitrogen of 5-7 g m-2 basal dressing and 3.5-4.5 g m-2 top dressing were applied. Sixty hills were harvested for yield, 1000-grain weight, and
 
Table 1. Population sizes of breeding doubled haploid (DH) lines
----------------------------------------------------------------
Total number of  Lines tested in 1991 1993  1994
---------------------------------------------------------------- 
Crosses              3   3   4  
Anthers cultured(A)      23715  12100  29376 
Callus obtained (a)      6820  2762  4212 
Green plantlets obtained (C) 218   67l   194 
   (B/C %)          28.8   22.8  14.3
   (C/A %)          0.9   5.5   0.7 
Lines for preliminary test   0   211  159 
Lines for yield test      17   14   13  
---------------------------------------------------------------- 
palatability evaluation. Ten plants per plot were measured for culm length, panicle length and number of panicles and averaged.
From the materials selected in the field, palatability of six lines in 1991 and 12 lines in 1994 were evaluated by a sensory evaluation test. The check cultivar of the sensory evaluation test was Koshihikari. Ten entries including the check cultivar were evaluated at the same time and overall eating quality relative to the check cultivar was scored by 15-20 panel members and averaged (Matsue et al., 1991).

In the same field of DH lines, 49, 69, and 73 CB lines in 1991, 1993, and 1994, respectively, were grown and palatability of 26 and 37 CB lines in 1991 and 1994, were evaluated. CB lines were, generally, obtained by planting Fl and F2 plants in a green house, growing F3 plants in a field followed by individual selection. F4 and F5 generations were pedigree-selected in a field. During F3-F5 generations, plants and lines with good plant type and good grain appearance were selected by observation. F6 lines were field-grown with DH lines. The seeds from a few F5 plants which were derived from a single F4 plant were mixed and used for the test of F6 lines. CB lines tested in 1991, 1993, and 1994 derived from crosses which were made generally in 1988, 1989, and 1990 and derived from 8, 23, and 24 crosses, respectively. There is no large difference of a genetic background among cross parents of DH and CB lines. Population size of breeding procedure for CB lines are summarized in Table 2.

To compare homogeneity of DH and CB lines, coefficient of variation within a line was calculated for culm lengh and panicle length.

Results and Discussion
The frequency of DH line production is shown in Table 1. Total green plantlets regenerated by anther culture in three years were 1,083 by culturing 6,5191 anthers. Overall average frequency of callus and green plantlet production per 100 anthers was 21.1 and 1.7. Total of 380 DH lines were tested for the preliminary fields test.

Table 3 shows the mean values of traits measured for DH lines along with CB lines. Student's t-test was applied to test the difference between the mean of DH and CB lines. No differences were found between the mean heading date in 1991, 1993, and 1994, showing that the comparison between DH and CB lines was based on the materials having the same mean heading date. Mean palatability of DH lines was -0.10 and -0.53 in 1991 and 1994. No difference was found between the mean of
 
Table 2. Population sizes of breeding conventional breeding lines 
---------------------------------------------------------- 
Total number of   Lines tested in 1991 1993 1994 
---------------------------------------------------------- 
Crosses                     8   23  24 
Lines for first pedigree selection  8958 16520 19490 
Lines for preliminary test       898  1722 2115 
Lines for yield test           49  69  73 
---------------------------------------------------------- 
DH and CB lines. Many factors concern palatability (Chikubu et al., 1983). Therefore, palatability must be governed by many genes, requiring selection from a large population size to obtain high palatability lines. Table 1 shows that 380 DH lines were tested for the preliminary field test, which was the first field test. For CB lines, on the other hand, population size were much larger than DH lines. Total of 44,968 plants were pedigree-selected in a field. Total of 4,735 lines were grown in a field for the second pedigree-selection or preliminary test. During selections of CB lines, no selection for palatability was conducted and only selection by observation for grain appearance was conducted. It shows that selections by observation had little effect on palatability.

The mean yield of DH lines was 411, 418, and 629 g m-2 in 1991, 1993, and 1994, respectively. lt was significantly lower than CB lines in 1991 (at the 0.1 % level). The mean of 1000-grain weight of DH lines was also lower than CB lines in 1991 (at the 10 % level). The average yield % of three years was 0.95. Low yield of DH lines of two years supports the findings ot Baenziger et al. (1989) in wheat. There must be many genes concerning the crop yield, which requires selection from a large size population to obtain high yield lines. During the selection of CB lines, yield was not measured but the selection by observation for good plant type must have resulted into developing CB lines with higher mean yield.
 
Table 3. Means of DH and CB lines and the results of t-test 
---------------------------------------------------------------------------- 
Trait    Year  DH mean (σ) CB mean (σ)  t-value  df    Probability 
---------------------------------------------------------------------------- 
Heading date  1991  8.19(0.05)  8.20(0.07)  0.85  64  0.4< P <0.5 
(Month.Day)   1993  8.28(0.07)  8.26(0.06)  0.87  81  0.4< P <0.5 
_____________  1994  8.15(0.07)  8. l7(0.08)  0.68  84  0.4< P <0.5 
  
Culm length  1991  67.6(4.9)  69.4(7.1)   1.13  64  0.2< P <0.4 
 (cm) ______ 1993  64.6(4.5)  69.0(3.2)   3.49  81  P <0.001 
_____________  1994  83.2(4.8)  80.5(5.4 )   1.85  84  0.05< P <0.1 
  
Panicle length 1991   19.9(1.2)  19.8(1.3)   0. 44  64  0.5< P 
 (cm ) ____   1993   l9.0(0.8)  l9.9(0.7)    3.78  81  P <0.001 
____________  1994   19.5(1.3) 20.4(0.9)    2.41  84  0.025< P <0.01 
  
Number of    1991   283(35)  283(36)    0.0l  64  0.5< P 
panicles    1993   254 (30)  265(24)    1.33  81  0.1< P <0.2
  (m-2)_______ 1994   424(45)  417(33)    0.51  84  0.5< P 
  
Grain yield   1991   411(35)  46l (74)   3.66  64   P< 0.001 
  (g m-2) ____ 1993   418(49)  438(36)   1.42  81   0.1< P <0.2 
_____________  1994   629(46)  623(42)   0.46  84   0.5< P 
  
1000-grain   1991   22.1 (0.9)  22.5(0.9)  1.70  64   0.05< P <0.10 
weight (g)   1993   20.9(1.8)  21.5(1.8)  1.11  79   0.20< P <0.40 
____________   1994   22.7(0.3)  22.6(0.3)  0.79  84   0.40< P < 0.50 
  
Palatabilily  1991   -0.10(0.38) -0.37(0.18)  1.64  30   0.1< P <0.2 
(check;0)    1994    -0.53(0.42) -0.46(031)  0.79  47   0.5< P 
-------------------------------------------------------------------------------- 
Riggs and Snape (1977), Yonezawa (1986), and Ukai (1987) simulated the relative efficiency of DH and CB methods and showed that the frequency of recombinants in a DH population was lower than in an equal CB population size in the presence of linkage. Genes controlling yield in the crosses of this study might be linked. Technical difficulties in anther culture limited the population size of DH lines. These factors could have also caused the lower mean yield of DH lines.

No yield difference was found between DH and CB lines in 1994, when the yield was very high under favorable growing conditions. Mitchell et al.(1992) also found the lack of a yield difference of wheat under less stressed conditions. Yield difference might have been masked due to the unusually favorable growing condition in 1994.

For culm length, DH lines were shorter than CB lines in 1993 (significant at the 0.1 % level) but longer in 1994 (at the 10 % level). lt is reported that DH lines had shorter culrn length in rice (Oono, 1975) and in wheat (Inagaki 1987, Winzeler et al., 1987) but in this study, no consistent tendency was found. For panicle length, DH lines were shorter than CB lines in 1993 and 1994. No significant differences for number of panicles were found between DH and CB lines.

Albino and clonal variation were observed during developing DH lines They were obviously inferior in agronomic characters. All of them were discarded. Standard deviations among DH lines did not differ from those of CB lines for all traits studied (Table 3), showing that increased deviation by anther culture could not be expected for agronomically desirable traits.

Occurrence of a gametophytic selection by anther culture could affect the means of DH lines for a quantitative character. But Choo et al. (1982) and Charmet and Branlard (1985) suggested that no gametophytic selection occurred in DH lines. In this study, no consistent differences except yield were found between means of DH and CB linea after discarding apparently mutated variants. lt suggests that gametophytic selection in anther culture might not have occurred for the non-mutated DH lines.

Fig.1(19KB) shows the distribution of coefficient of variation within a line for the traits calculated. Student's t-tests in Table 4 revealed no significant differences between DH and CB lines except panicle length in 1991. CB lines in 1991
Table 4. Means of coefficient of variation (A) within doubled haploid (DH) and conventional breeding(CB) line
--------------------------------------------------------------------- 
Trait        Year  DH mean(σ) CB mean(σ) t-value  df  Probability
--------------------------------------------------------------------- 
Culm length  1991   3.6(0.95)    3.4(0.90)    0.46    64  0.5< P
            1993   4.0(1.07)    3.4(1.23)    1.70    81  0.05< P <0. 1
            1994   2.6(0.70)    2.7(0.72)    0.63    84  0.5< P

Panicle     1991   4.4(0.91)    5.2(1.S3)    2.10    64  0.025< P <0.05
  length     1993   4.7(1.09)    5.1(1.64)    1.09    81  0.4< P <0.5
            1994   5.6(1.05)    5.6(1.63)    0.01    84  0.5< P
---------------------------------------------------------------------- 
were F5 lines derived from a single F3 plant, which might have segregation in panicle length, resulting into larger mean coefficient of variation of CB lines. CB lines in 1993 and 1994 were in F6 generation derived from a single F4 plant. No difference between DH and CB lines of F6 generation were found, showing that F6 CB lines had no serious segregation for agronomic traits. The homogeneity in very early generations shows that cross parents of CB lines are closely related.

DH and CB lines tested in 1993 and 1994 were obtained in three and four years, respectively, after crossings. DH lines were obtained in one year less than CB lines, showing that anther culture method could produce homogeneous lines more rapidly than pedigree selection method.

DH and CB lines in this study were derived from many crosses and no DH and CB lines derived from a common cross. It may be said that DH and CB lines having no common parentage might not allow direct comparison of them. But, on the other hand, the results obtained by using materials of an identical genetic background might be limited only to the genetic background. In this study, many crosses were included in both of DH and CB lines and their cross parents were an array of well adapted high yield and high palatability cultivars. There was no large difference of a genetic background between DH and CB cross parents. No difference of average heading date between DH and CB lines supports it. Therefore, it may be possible to conclude that the results obtained in this study justify the comparison of DH and CB lines.

Conclusion
Rice lines with high palatability and agronomicaly desirable traits can be obtained by anther culture method as by conventional breeding method except yield. To obtain high yield lines, it is necessary to culture more anthers or increase the efficiency of each process of anther culture and select DH lines from a large population size.
References   
Baenziger, P.S., D.M. Wasenberg, V.M.
 Smail, W.L. Alexander, and C.W.
 Schaeffer. l989. Agronomic performance of wheat doubled-haploid lines
 derived from cultivars by anther culture. Plant Breeding 103:101-109.
Charmet, G., and C. Branlard. 1985. A comparison ot androgenetic
 doubled-haploid, and single seed descent lines in triticale. Theor. Appl. 
 Genet. 71:193-200.
Chikubu, S., S. Watanabe, T. Sugimoto, F.Sakai, and Y. Taniguchi. 1983. 
 Relation between palatability evaluations of cooked rice and physicochemical
 propeties of rice. J.Jpn. Sot. Starch Sci. 30:334-341.
Choo, T.M., E. Reinbers, and S.I. Park. 1982. Comparison of frequency distributions
 of doubled haploid and single seed descent lines in barley. Theor.
 Appl. Genet. 61:215-218.
Henry, Y., I. De Buyscr, S.Agache, RB.Parker, and I.W. Snare. 1988. 
 Comparisons of methods of haploid production and performance of wheat lines 
 produced by doubled haploidy and single seed descent. In 'Proc. Int. Wheat Genetics
 Symp., 7th, Cambridge, England. 13-19 July 1988' Miller, T.E. and A.M.D.
 Koebner (eds.) Inst. Plant Sci. Res., Cambridge, l087-1092.
Inagaki, M. 1987. Variation in plant height of doubled haploid lines of wheat
 derived from intergeneric crosses with Hordium bulbosum L. Japan. J. Breed.
 37:275-282.
Kunihiro, Y., Y. Ebe, N. Shinbashi, H. Kikuchi, H. Tanno, and K. Sugawara. l993.
 A new paddy rice variety 'Aya' with good eating quality due to low arnylose content
 developed by anther culture breeding. Japan. J. Breed. 43:155-163.
Maeda, H., T. Tsuchiya, N. Hiraoka, M. Maeshige, S. Uernoto, and M. 
 Nakayabu. 1990. New cold resistant rice varieties 'Hirohikari' and 'Hirohonami',.
 Bull. Hiroshima Agric. Exp. Std. 53:9-20.
Matsue, Y., K.Mizuta, K.Furuno, and T.Yoshida. l99l. Studies on palatability of
 rice grown in Northern Kyushu I. Effects of transplanting time and lodging time on 
 palatability and physicochemical properties of milled rice. Jpn.J.Crop Sci. 60:490-496
Mitchell, M.I., R.H. Busch, and H.W. Rines. 1992. Comparison of lines derived 
 by anther culture and single-seed descent in a spring wheat cross. Crop Sci.32:1446-1451
Oono, K. 1975. Production ot haploid plants of rice (Oryza sativa) by anther culture
 and their use for breeding. Bull. Natl. Inst. Agric. Sci. Ser. D 26:139-222.
Riggs, T.I., and J.W. Snare. 1977. Effects ot linkage and interaction in a 
 comparison of theoretical populations derived by diploidized haploid and single
 seed descent methods. Theor. Appl. Genet. 49:111-115.
Shinbashi, N., and M.Aikawa. 1986. Use of anther culture in rice breeding.
 Ikushugaku saikinnoshinpo 27:13-18.
Ukai, Y. 1987. Fixation of genes and breakup of linkage blocks in selfed generations. 
 Japan. J. Breed. 37:40-53.
Wakasa, K. 1982. Application of tissue culture to plant breeding - Method
 improvement and mutant production -. Bull. Natl. Inst. Agric. Sci. Ser. D. 33:121-200.
Winzeler, H., I. Schmid, and P.M. Fried. 1987. Field performance of 
 androgenetic doubled haploid spring wheat lines in comparison with lines selected
 by the pedigree system. Plant Breeding 99:41-48.
Yonezawa, K. 1986. Evaluation of the doubled haploid breeding method by computer simulation. 
 Ikushugaku saikinno shinpo 27:44-49.


以上