The complete nucleotide sequence (155 844 bp) of tobacco (Nicotiana tabacum var. Bright Yellow 4) chloroplast DNA has been determined. It contains two copies of an identical 25 339 bp inverted repeat, which are separated by a 86 684 bp and a 18 482 bp single‐copy region. The genes for 4 different rRNAs, 30 different tRNAs, 39 different proteins and 11 other predicted protein coding genes have been located. Among them, 15 genes contain introns. Blot hybridization revealed that all rRNA and tRNA genes and 27 protein genes so far analysed are transcribed in the chloroplast and that primary transcripts of the split genes hitherto examined are spliced. Five sequences coding for proteins homologous to components of the respiratory‐chain NADH dehydrogenase from human mitochondria have been found. The 30 tRNAs predicted from their genes are sufficient to read all codons if the ‘two out of three’ and ‘U:N wobble’ mechanisms operate in the chloroplast. Two sequences which autonomously replicate in yeast have also been mapped. The sequence and expression analyses indicate both prokaryotic and eukaryotic features of the chloroplast genes.

The entire chloroplast genome of the monocot rice (Oryza sativa) has been sequenced and comprises 134525 bp. Predicted genes have been identified along with open reading frames (ORFs) conserved between rice and the previously sequenced chloroplast genomes, a dicot, tobacco (Nicotiana tabacum), and a liverwort (Marchantia polymorpha). The same complement of 30 tRNA and 4 rRNA genes has been conserved between rice and tobacco. Most ORFs extensively conserved between N. tabacum and M. polymorpha are also conserved intact in rice. However, several such ORFs are entirely absent in rice, or present only in severely truncated form. Structural changes are also apparent in the genome relative to tobacco. The inverted repeats, characteristic of chloroplast genome structure, have expanded outward to include several genes present only once per genome in tobacco and liverwort and the large single copy region has undergone a series of inversions which predate the divergence of the cereals. A chimeric tRNA pseudogene overlaps an apparent endpoint of the largest inversion, and a model invoking illegitimate recombination between tRNA genes is proposed which accounts simultaneously for the origin of this pseudogene, the large inversion and the creation of repeated sequences near the inversion endpoints.

A major component of the large genomes of higher plants and vertebrates comprises transposable elements and their derivatives, which potentially reduce the stability of the genome. It has been proposed that methylation of cytosine residues may suppress transposition, but experimental evidence for this has been limited. Reduced methylation of repeat sequences results from mutations in the Arabidopsis gene DDM1 (decrease in DNA methylation), which encodes a protein similar to the chromatin-remodelling factor SWI2/SNF2 (ref. 7). In the ddm1-induced hypomethylation background, silent repeat sequences are often reactivated transcriptionally, but no transposition of endogenous elements has been observed. A striking feature of the ddm1 mutation is that it induces developmental abnormalities by causing heritable changes in other loci. Here we report that one of the ddm1-induced abnormalities is caused by insertion of CAC1, an endogenous CACTA family transposon. This class of Arabidopsis elements transposes and increases in copy number at high frequencies specifically in the ddm1 hypomethylation background. Thus the DDM1 gene not only epigenetically ensures proper gene expression, but also stabilizes transposon behaviour, possibly through chromatin remodelling or DNA methylation.

Summary Different colors, such as purple, brown, red and white, occur in the pericarp of rice. Here, two genes affecting proanthocyanidin synthesis in red‐ and brown‐colored rice were elucidated. Genetic segregation analysis suggested that the Rd and A loci are identical, and both encode dihydroflavonol‐4‐reductase (DFR). The introduction of the DFR gene into an Rcrd mutant resulted in red‐colored rice, which was brown in the original mutant, demonstrating that the Rd locus encodes the DFR protein. Accumulation of proanthocyanidins was observed in the transformants by the introduction of the Rd gene into the rice Rcrd line. Protein blot analysis showed that the DFR gene was translated in seeds with alternative translation initiation. A search for the Rc gene, which encodes a transacting regulatory factor, was conducted using available DNA markers and the Rice Genome Automated Annotation System program. Three candidate genes were identified and cloned from a rice RcRd line and subsequently introduced into a rice rcrd line. Brown‐colored seeds were obtained from transgenic plants by the introduction of a gene containing the basic helix–loop–helix (bHLH) motif, demonstrating that the Rc gene encodes a bHLH protein. Comparison of the Rc locus among rice accessions showed that a 14‐bp deletion occurred only in the rc locus.

Phosphatidylinositol 4,5-bisphosphate [PtdIns(4,5)P 2 ] functions as a site-specific signal on membranes to promote cytoskeletal reorganization and membrane trafficking. Localization of PtdIns(4,5)P 2 to apices of growing root hairs and pollen tubes suggests that it plays an important role in tip growth. However, its regulation and mode of action remain unclear. We found that Arabidopsis thaliana PIP5K3 (for Phosphatidylinositol Phosphate 5-Kinase 3) encodes a phosphatidylinositol 4-phosphate 5-kinase, a key enzyme producing PtdIns(4,5)P 2 , that is preferentially expressed in growing root hairs. T-DNA insertion mutations that substantially reduced the expression of PIP5K3 caused significantly shorter root hairs than in the wild type. By contrast, overexpression caused longer root hairs and multiple protruding sites on a single trichoblast. A yellow fluorescent protein (YFP) fusion of PIP5K3, driven by the PIP5K3 promoter, complemented the short-root-hair phenotype. PIP5K3-YFP localized to the plasma membrane and cytoplasmic space of elongating root hair apices, to growing root hair bulges, and, notably, to sites about to form root hair bulges. The signal was greatest in rapidly growing root hairs and quickly disappeared when elongation ceased. These results provide evidence that PIP5K3 is involved in localizing PtdIns(4,5)P 2 to the elongating root hair apex and is a key regulator of the machinery that initiates and promotes root hair tip growth.