SummaryIn cultivated rice two wild-type alleles, Wx a and Wx b , predominate at the waxy locus, which encodes granulebound starch synthase. The activity of Wx a is 10-fold higher than that of Wx b at the level of both protein and mRNA.Wx b has a ϩ1G to T mutation at the 5Ј splice site of the first intron. Sequence analysis of Wx b transcripts revealed that splicing occurs at the mutant AG/UU site and at two cryptic sites: the first is A/GUU, one base upstream of the original site and the second is AG/GU found approximately 100 bases upstream of the mutant splice site. We introduced single base mutations to the 5Ј splice sites of both Wx a and Wx b , fused with the gus reporter gene and introduced them into rice protoplasts. Analysis of GUS activities and transcripts indicated that a G to T mutation in Wx a reduced GUS activity and the level of spliced RNA. Conversely, a T to G mutation of Wx b restored GUS activity and the level of spliced RNA to that of wild-type Wx a . These results demonstrated that the low level expression of Wx b results from a single base mutation at the 5Ј splice site of the first intron. It is of interest that the Wx b allele of rice carrying the G to T mutation of intron 1 has been conserved in the history of rice cultivation because there is a low amylose content of the seed caused by this mutation.

The quantitative trait locus controlling the number of primary rachis branches (PRBs) in rice was identified using backcrossed inbred lines of Sasanishiki/Habataki//Sasanishiki///Sasanishiki. The resultant gene was ABERRANT PANICLE ORGANIZATION 1 (APO1). Habataki-genotype segregated reciprocal recombinant lines for the APO1 locus increased both the number of PRB (12-13%) and the number of grains per panicle (9-12%), which increased the grain yield per plant (5-7%). Further recombination dividing this region revealed that different alleles regulated the number of PRB and the number of grains per panicle. The PRB1 allele, which includes the APO1 open reading frame (ORF) and the proximal promoter region, controlled only the number of PRB but not the number of grains per panicle. In contrast, the HI1 allele, which includes only the distal promoter region, increased the grain yield and harvest index in Habataki-genotype plants, nevertheless, the ORF expressed was Sasanishiki type. It also increased the number of large vascular bundles in the peduncle. APO1 expression occurred not only in developing panicles but also in the developing vascular bundle systems. In addition, Habataki plants displayed increased APO1 expression in comparison to Sasanishiki plants. It suggests that APO1 enhances the formation of vascular bundle systems which, consequently, promote carbohydrate translocation to panicles. The HI1 allele is suggested to regulate the amount of APO1 expression, and thereby control the development of vascular bundle systems. These findings may be useful to improve grain yield as well as quality through the improvement of translocation efficiency.

SummaryIn rice, silencing of the aleurone-specific Ltp2-gus transgene, causing easily detectable staining patterns on the grain surface, offers a convenient tool to study quantitative aspects of gene silencing in monocots. In this paper we analyzed phenotypes, occurrence, inheritance and environmental effects on the silencing. We also report on the cloning of transgenes, determination of their structure and analysis of transcripts from the transgene loci. The results show that various patterns of silencing appeared in the R2 generation at which most of the transgenes became homozygous and that they were inherited for five generations. In addition, silencing independently occurred in three generations and reversion to full expression was also found. Cloning of transgenes from a silenced L3.3 line demonstrated that this line carried two transgene loci: one carried an intact Ltp2-gus gene and the other carried a rearranged transgene in which part of the gus gene was in the antisense orientation. Analysis of gus transcripts indicated that partial antisense RNA derived from the rearranged transgene was present in silenced lines and was polyadenylated but that it was absent in non-silenced lines. RNA analyses suggested that the Ltp2-gus silencing in the aleurone layer was post-transcriptional and that it may be caused by interaction of partial antisense gus transcripts with normal sense transcripts. Possible involvement of antisense transcripts in post-transcriptional silencing is discussed.