19. Mutations affecting embryo size in rice Soon-Kwan Hong1 , Hidemi Kitano2 Hikaru Satoh3 and Yasuo Naoato1
1) Faculty of Agriculture, University of Tokoy, Tokoy 113, Japan
2) Aichi University of Hducation, Kariya 448, Japan
3) Faculty of Agriculture, Kyushu University, Fukuoka 812, Japan     It is important to elucidate how the size of developing embryo is genetically regulated, because the final volume of endosperm as a storage organ of starch and proteins is affected by embryo size in cereal crops. The developing embryo and endosperm would interact both physiologically and physically in a closed, spatially-limited embryo sac. However, this regulatory mechanisms is nuclear, because mutations available for such study have not been obtained with the exception of giant embryo (ge) in rice (Satoh and Omura 1981). Recently, in maize and Arabidopsis, many embryo mutants have been identified and used for genetic study of plant embryogenesis (dark and Sheridan 1991; Mayer et al. 1991; Castle and Meinke 1993), although no embryo size mutants have been identified in the two species.
    To date, we have isolated 188 single-gene recessive mutants affecting embryo development by screening ca. 5600 M2 progenies of rice Taichung 65 and Kinmaze, which were produced after a treatment of spikelets containing zygotes or two-cell embryo with 1.0 mM MNU (N-methyl-N-nitrosourea) (Kitano et al. 1993; Hong et al. 1995). Through characterizing embryonic phenotypes, these mutants could be categorized into the following six groups: (1) lethality at the early globular stage, (2) deletion of embryonic organ(s), (3) abnormal position of embryonic organs, (4) modified embryo size, (5) defect in morphogenesis, and (6) variable abnormal phenotypes.
    The fourth group contained embryo size mutations in different directions, e.g., enlargement (giant embryo) and reduction (reduced embryo) as shown in Fig. 1. In this group, five giant embryo mutants, odm30, odm44, odm90, odm93, and odml32, and four reduced embryo mutants, odml6, odm48, odm49 and odm62, were identified. All of these germinate and grow normally into fertile mature plants similar to wild type plants in size. This suggests that the mutant genes have seed-specific function.
    Allelism tests were carried out among the six giant embryo mutants including ge originally reported by Satoh and Omura (1981), and among the four reduced embryo mutants (Table 1 ). In giant embryo mutants, F1 seeds showed giant embryo phenotype when recessive homozygous plants were crossed, and a half of F1 seeds expressed giant embryos in the crosses between recessive homozygous and heterozygous plants. Accordingly, all the mutant genes are allelic to ge on chromosome 7 (Satoh and Iwata 1990). We designated the six alleles at the Ge locus as ge-2 (odm30), ge-3 (odm 44), ge-4 (odm 90), ge-5 (odm 93) and ge-6 (odm 132), and the original allele as ge-l. On the other hand, in reduced embryo mutants, only the F1 of a cross (odm49/odml6) showed reduced embryo phenotype. The other crosses gave wild type F1 seeds and the segregation in F2 fitted a non-allelic inheritance (9:7 segregation)(Table 1). Therefore, the four reduced embryo mutations represent three independent loci, REI with two alleles, rel-l (odm 16) and re 1-2 (odm49), RE2 with one allele, re2 (odm48), and RE3 with one allele, re3 (odm62). Double mutant analysis indicates that rel-l and re2 are epistatic to ge alleles (Table 1). Some embryonic characteristics of the mutants were examined on median
 
Table 1. F1 phenotype and F2 segregation in the crosses among embyro size mutants
Combination F1 phenotype* F2 segregation X2
wt                 ge
odm30/ge ge no segregation
odm30/odm44 ge no segregation
odm90/odm44 ge no segregation
odm90/odm93 ge no segregation
odm90/odm 132(hetero) ge+wt (1:1) 65               26 0.61 (3:1)
wt              re
odm48/odml6 wt 51              49 1.12(9:7)
odm49/odm48 wt 47              33 0.20 (9:7)
odm49/odml6 re no segregation
odm49/odm62 wt 12               8 0.11 (9:7)
odm62/odm48 wt (in progress)
wt        ge     re
odml6/odm30 wt 1763   532    770 4.07 (9:3:4)
odml6/ge wt 465     134    183 3.32(9:3:4)
103 34 43 wt 103      34      43 0.12(9:3:4)
*ge: giant embryo, re: reduced embryo, and wt: wild type embryo. longitudinal sections of mature embryos (Table 2). Giant embryo mutants have about 1.25 time longer embryos than the wild type. However, all the mutants differentiate shoot and radicle of normal sizes. Only the scutellum is enlarged. As the number of cells in embryo of each mutant is comparable to that of wild type embryo, the enlargement of embryo is exclusively due to the enlargement of scutellum cells. Embryonic phenotypes are very similar among the mutants except ge-6, in which additional radicle or additional set of shoot and radicle with anormal shape is frequently produced. On the contrary, in reduced embryo mutants, embryos are less than half in length of those of wild type, and every  
Table 2. Characteristics of mature embryo in embryo size mutants
Mutation Embryo (microm) 

Length       Thickness

 Shoot apex (microm) 
Length    Width		 Radicle(microm) 

Length        Width

 No. of cells*
ge-2 2536 ± 163 1219±121 64± 4 63± 9 489 ± 84 384± 61 6108
ge-3 2704 ± 134 1247± 11 60± 7 64± 4 479 ± 76 404 ± 37 6277
ge-4 2497 ± 89 1167± 76 61± 4 62± 9 487 ± 59 383 ± 55 6043
ge-5 2530± 149 1189± 97 65± 9 64± 6 427 ± 58 366 ± 36 5829
ge-6 2436 ± 184 1241 ± 167 61±10 64± 14 424±119 378 ± 69 6587
rel-l 793 ± 89 529± 59 34± 3 36± 5 137± 28 151± 14 3037
rel-2 889±141 599±119 36± 9 37± 8 156± 29 164± 22 3533
re2 738 ± 128 436±101 34± 10 36± 8 122± 49 156± 50 2458
re3 864±123 576 ± 98 35± 2 41 ± 3 163± 38 145± 23 3388
wild type 2014± 94 1029± 85 59± 9 54± 9 432± 21 373± 12 6326
*Number of cells in the median longitudinal section.
Fig. 1 Median longitudinal sections of mature wild type (A), ge-2 (B) and rel-1 (C) embryo.
            Bar=500 µm

embryonic organ is reduced due to a small number of cells. Embryonic phenotypes are similar among the four mutants.
It is noted that in both giant and reduced embryo mutants, the sizes of embryo and endosperm are negatively correlated, suggesting that the giant embryo and reduced embryo mutations may be primarily associated with endosperm development. Then we made double mutants combining ge-2 and clel-I which produced small undifferentiated club-shaped embryo with normal endosperm, and combining rel-l with shpl which modified shoot position and reduced scutellum. In ge-2/clel-l seed, embryo is clel-l type but the endosperm is reduced as in gel seed. This indicates that ge-2 causes the reduction of endosperm size. resulting in the large space for embryo to develop. On the other hand, rel-llshpl seed has shpl embryo and an enlarged endosperm of rel-l type, suggesting that rel-l enlarges endosperm. Primary involvement of ge-2 and rel-l in endosperm development is further supported by the developmental analysis of ge-2 and rel-l seeds. In ge-2 seed, degradation of endosperm cells located near embryo is recognized at the early stage. In rel-l, vigorous endosperm development is detected, resulting in direct contact of endosperm with embryo, whereas small space is observed between embryo and endosperm in wild type seed. These results suggest that embryo size may be regulated by genes controlling endosperm development.

References

Castle, L. A. and D. W. Meinke, 1993. Embryo-defective mutants as tools to study essential functions and
        regulatory processes in plant embryo development. Seminars in Developmental Biology 4: 31-39.

Clark, J. K. and W. F. Sheridan, 1991. Isolation and characterization of 51 embryo-specific mutations of maize.
        Plant Cell 3: 935-951.
Hong, S. K., T. Aoki. H. Kitano, H. Sato and Y. Nagato, 1995. Phenotypic diversity of 188 rice embryo
        mutants. Dev. Genet, (in press).
Kitano, H., Y. Tamura, H. Satoh and Y. Nagato, 1993. Hierarchical regulation of organ differentiation during
        embryogenesis in rice. Plant J. 3: 607-610.
Mayer, U., R. A. T. Ruiz, T. Berleth, S. Misera and G. Jurgens, 1991. Mutations affecting body organization in
        the Arabidopsis embryo. Nature 353: 402-407.
Satoh, H. and T. Omura, 1981. New endosperm mutations induced by chemical mutagens in rice (Oryza sativa
        L.). Japan. J. Breed. 31: 316-326.
Satoh, H. and N. Iwata, 1990. Linkage analysis in rice. On three mutant loci for endosperm properties, ge
        (giant embryo), du-4 (dull endosperm-4) and flo-I (floury endosperm-l). Japan. J. Breed. 40 (Suppl. 2):
        268-269. (in Japanese)