Heredity 63(1989) 171—179 The Genetical Society of Great Britain Received 25 January 1989 Pollination with irradiated pollen in rice Oryza sativa L. II. The second (M2) generation Shin-Foong Chin* and Department of Agriculture, University of Queensland, Geoff H. Gordon St. Lucia, Queensland 4067, Australia. Surviving Ml plants derived from pollination with irradiated pollen were selfed to produce the second (M2) generation. in three quantitative traits, the means and phenotypic frequency distributions of the M2 populations were identical to that of the F2 population. In five major gene traits, all but six of the 244 M2 segregations were in the expected ratios is the F2's. Aberrant segregations produced both excess paternal as well as maternal phenotypes. There was no consistent trend of a "maternal shift" in the M2 generation that may render the technique of irradiated pollination useful by preferentially producing pure breeding maternal materials having a few specific paternal characteristics. Qverall, pollination with irradiated pollen in rice produced mainly normal hybrid progeny with very little heritable variations. These lack of responses in rice may be explained by the fact that rice pollen is relatively insensitive to irradiation on the one hand, and that the pollen genome is less able to accommodate mutational damage on the other hand. Thus, the rice plant is considered less amenable to the application of irradiated pollination as a practical breeding technique. INTRODUCTION all the Ml progeny were either normal hybrid plants indistinguishable from the Fl controls, or Theuse of irradiated pollen to transfer single genes were hemizygous recessive plants in which certain from the pollen parent to the seed parent in cross- dominant paternal alleles were lost. This reduction pollination was first proposed by Pandey (1975). in the transmission of paternal alleles was accom- Subsequently, it was observed that when pollen panied by morphological mutations and sterility. was irradiated, the pollen nuclei were "pulverized" In total, 17,626 rice florets were pollinated with and the resultant nuclear "debris" was discharged irradiated pollen but not a single Ml plant was onto the embryo sac and fused with the egg cells found to contain only one or a few paternal traits Grant et al., 1980). On the basis of these observa- of the pollen parent in a largely maternal genetic tions, Pandey (1980a, b) further proposed that the background. Therefore, "egg transformation" by method of pollinating plants with irradiated pollen irradiated pollen was not shown to have occurred could be used for "egg transformation", defined in the rice plants. is the transfer of limited intact genes rather than It has been suggested that pollination with the total pollen genome to the egg cell. He sug- irradiated pollen may nevertheless be useful in gested that transformation could be the result of practical plant breeding by causing a shift in the diploid parthenogenesis of the egg induced by segregation ratio towards the maternal phenotypes irradiated pollen, followed by the incorporation in the second (M2) generation. This effect was of paternal chromatin during embryogenesis. In the companion paper, we reported the use of irradi- observed in barley, (Powell et al., 1983) and wheat, (Snape et aL, 1983). Compared to backcrossing, ated pollen of the rice cultivar Basmati, having this could be a quicker method to produce pure dominant genetic markers, to pollinate the mater- breeding maternal progeny with a few specific nal cultivar Bellemont, having recessive markers paternal characteristics. This issue of "maternal Chin and Gordon, 1989). In the first generation, shift" will be addressed in this report of the genetic Present address: CSIRO, Division of Plant Industry, GPO consequences of pollination with irradiated pollen Box 1600, Canberra, ACT 2601, Australia. in the M2 generation of rice plants. 172 S -F. CHIN AND G. H. GORDOr' MATERIALS AND METHODS least overnight at 4°C. Hydrolysis and Feulge staining was done as for mitotic studies. M2 and M3 plants Overtwo seasons, a total of 66 M2 families (857) Dataanalysis plants)were field-planted. These M2 families were obtained from selfing the randomly selected Ml Duncan'smultiple comparisons were done on ths plants which were derived from crosses with pollen statistical software SAS (Statistical Analysis Sys given various radiation doses. M2 families were tem, 1982 edition) using the general linear mode planted in randomized single rows consisting of procedure (PROC GLM), Skewness and kurtosi 15 to 20 lants per row. As controls, a total of tests on frequency distributions was done usin 15 F2 families (212 plants) were also grown. M2 the frequency procedure (PROC FREQ) on th and F2 plants were assessed for final plant height, SAS. Chi-square and t-tests were calculated on total tiller number, number of days to anthesis, program written in FORTRAN. Cluster analysi leaf texture, anthocyanin pigmentation on seed was carried out on the computer progran apiculus, seed awning and seed phenol staining 'HACLUS" written by Mr I. H. De Lacy of th reaction. All assessments were done as reported in Department of Agriculture, University of Queens the previous paper of this series. One M2 family land, St. Lucia, Australia. with apparent paternal shift was grown on to M3 generation. All the 15 M2 plants of this family were selfed to produce M3 lines and 20 plants of each RESULTS of the 15 M3 lines were field-grown and scored for leaf texture. These 15 M2 plants were also Quantitative traits examined for meiosis. For meiotic studies, panicles The means and standard error of means of plan were collected when the auricle of the flag leaf had height, tiller number and number of days t reached the same height as the auricle of the pre- anthesis obtained from two seasons of field plant ceding leaf (after De Datta, 1981); Cotlman and ing are shown in table 1. Results of Duncan' Herrera, 1980). Florets with immature pollen were multiple range tests on the means are also shown fixed with Carnoy's fixative (six parts ethanol, Phenotypic frequency distributions of these thre three parts chloroform, one part acetic acid) at phenotypic traits are shown in fig. 1(a) and 1(b) Table I Means and standard error of means of three quantitative traits of F2 and M2 progenies from two seasons of field trial Height (cm) Tiller number Anthesis (days) No. of No. of — Treatment family plant Mean S. E. Mean S. E. Mean S. E. First season Bellemont 4 51 909c 074 53d 042 1192d 113 Basmati 4 56 1517a 300 364a 194 1444a 052 F2 10 113 1352bc 207 225b 110 1359c 106 05 krad 10 121 1383b 241 233b 142 1353c 094 lkrad 10 111 1344bc 193 213bc 100 1317c 095 2 krad 8 102 138th 203 176c 082 1342c 098 4krad 10 121 130Oc 218 225b 174 133Oc 102 6krad 4 37 1382b 312 233b 240 141Ob 159 Second season Bellemont 2 40 967c 125 33c 068 115Oc 078 Basmati 2 40 161la 117 222a 104 140Oa 089 F2 5 99 1436b 213 13'7b 062 1315b 139 5krad 9 179 141Sb 153 ilIb 036 134lb 085 6 krad 3 60 1376b 237 126b 036 1373b 149 8krad 3 46 141Ib 270 116b 077 1336b 180 l0krad 9 80 144Ob 207 137b 071 1326b 149 Means with the same letters are not different at the 5 per cent level of Duncan's multiple range test. IRRADIATED POLLEN IN RICE. II. M2 173 a Kr Height (cm) Tiller(no) Anthesis (days) b BMT BAS BMT BAS BMT BAS 30 80 10 Kr 110 100 SO 180 5 30 Height (cm) Tiller (no) Anthesis (days) Figure 1 (a) Frequency distributions of three quantitative traits of F2 and M2 progeny from the first season of field trial. All means are indicated by arrows and the number of plants are shown on the vertical axes. Bellemont (BMT) is the female parent and Basmati (BAS) the male parent. (b) Frequency distributions of three quantitative traits of F2 and M2 progeny from the second season of field trial. All means are indicated by arrows and the number of plants are shown on the vertical axes. Bellemont (BMT) is the female parent and Basmati (BAS) the male parent. These distributions are tested for departure from gi and g2 are after Sokal and Rohif (1981) as normality by two sample statistics (gi) and (g2) follows:t. =(gl-P1)/Sgiand t. =(g2-P2)/S2, which measure skewness (asymmetry) and kurtosis where P1 and P2, the population parameters, are (peakedness) respectively. Test of significance of equal to zero for a normally distributed population. 174 S-F. CHIN AND G. H. GORDON Table 2 Test of departure from normality on the frequency distributions of F2 and M2 phenotypes. Data were obtained from two seasons of field trial. The two sample statistics (gi) and (g2) test for skewness and kurtosis respectively Height Tiller Anthesis Treatment gi I g2 I gi I g2 I gi t g2 I First season F2 03 ns 00 ns 71R hip —02 ns 14 ns 05 krad 06 ns 10 ns 156R 39Sp 06 ns —01 ns 1 krad 13 ns 06 ns 44R 40p —11 ns 07 ns 2krad 03 ns 02 ns 33R 26p 04 ns —01 ns 4krad —4OL 39p 188R 57•4p —03 ns —02 ns 6krad 07 ns 08 ns 42R Sip 10 ns —1-2 ns Second season F2 —20 ns 12 ns 26R IOip —14 ns 20 ns 5krad —19 ns 12 ns 6OR 49p —1.2 ns —04 ns 6 krad —05 ns 09 ns 47R 53p —10 ns —02 ns 8krad 03 ns —13 ns 23R 22p —13 ns 05 ns l0krad —01 ns —07 ns 38R 24p —12 ns —01 ns A negative significant gi indicates left-skew (L) and a positive significant gi indicates right-skew (R).