Pollen-pistil interactions are crucial for controlling plant mating. For example, S-RNase-based self-incompatibility prevents inbreeding in diverse angiosperm species. S-RNases are thought to function as specific cytotoxins that inhibit pollen that has an S-haplotype that matches one of those in the pistil. Thus, pollen and pistil factors interact to prevent mating between closely related individuals. Other pistil factors, such as HT-B, 4936-factor and the 120 kDa glycoprotein, are also required for pollen rejection but do not contribute to S-haplotype-specificity per se. Here we show that S-RNase is taken up and sorted to a vacuolar compartment in the pollen tubes. Antibodies to the 120 kDa glycoprotein label the compartment membrane. When the pistil does not express HT-B or 4936-factor, S-RNase remains sequestered, unable to cause rejection. Similarly, in wild-type pistils, compatible pollen tubes degrade HT-B and sequester S-RNase. We suggest that S-RNase trafficking and the stability of HT-B are central to S-specific pollen rejection.

SUMMARYWild tomato species in Solanum Section Lycopersicon often exhibit two types of reproductive barriers: selfincompatibility (SI) and unilateral incompatibility or incongruity (UI), wherein the success of an inter-specific cross depends on the direction of the cross. UI pollen rejection often follows the 'SI · SC' rule, i.e. pistils of SI species reject the pollen of SC (self-compatible) species but not vice versa, suggesting that the SI and UI pollen rejection mechanisms may overlap. In order to address this question, pollen tube growth was measured after inter-specific crosses using wild tomato species as the female parents and pollen from cultivated tomato (Solanum lycopersicum). Two modes of UI pollen rejection, early and late, were observed, and both differed from SI pollen rejection. The structure and expression of known stylar SI genes were evaluated. We found that S-RNase expression is not required for either the early or late mode of UI pollen rejection. However, two HT family genes, HT-A and HT-B, map to a UI QTL. Surprisingly, we found that a gene previously implicated in SI, HT-B, is mutated in both SI and SC S. habrochaites accessions, and no HT-B protein could be detected. HT-A genes were detected and expressed in all species examined, and may therefore function in both SI and UI. We conclude that there are significant differences between SI and UI in the tomato clade, in that pollen tube growth differs between these two rejection systems, and some stylar SI factors, including S-RNase and HT-B, are not required for UI.

Improvement of leaf photosynthesis is an important strategy for greater crop productivity. Here we show that the quantitative trait locus GPS (GREEN FOR PHOTOSYNTHESIS) in rice (Oryza sativa L.) controls photosynthesis rate by regulating carboxylation efficiency. Map-based cloning revealed that GPS is identical to NAL1 (NARROW LEAF1), a gene previously reported to control lateral leaf growth. The high-photosynthesis allele of GPS was found to be a partial loss-of-function allele of NAL1. This allele increased mesophyll cell number between vascular bundles, which led to thickened leaves, and it pleiotropically enhanced photosynthesis rate without the detrimental side effects observed in previously identified nal1 mutants, such as dwarf plant stature. Furthermore, pedigree analysis suggested that rice breeders have repeatedly selected the high-photosynthesis allele in high-yield breeding programs. The identification and utilization of NAL1 (GPS) can enhance future high-yield breeding and provides a new strategy for increasing rice productivity.

The sundew genus Drosera consists of carnivorous plants with active flypaper traps and includes nearly 150 species distributed mainly in Australia, Africa, and South America, with some Northern Hemisphere species. In addition to confused intrageneric classification of Drosera, the intergeneric relationships among the Drosera and two other genera in the Droseraceae with snap traps, Dionaea and Aldrovanda, are problematic. We conducted phylogenetic analyses of DNA sequences of the chloroplast rbcL gene for 59 species of Drosera, covering all sections except one. These analyses revealed that five of 11 sections, including three monotypic sections, are polyphyletic. Combined rbcL and 18S rDNA sequence data were used to infer phylogenetic relationships among Drosera, Dionaea, and Aldrovanda. This analysis revealed that all Drosera species form a clade sister to a clade including Dionaea and Aldrovanda, suggesting that the snap traps of Aldrovanda and Dionaea are homologous despite their morphological differences. MacClade reconstructions indicated that multiple episodes of aneuploidy occurred in a clade that includes mainly Australian species, while the chromosome numbers in the other clades are not as variable. Drosera regia, which is native to South Africa, and most species native to Australia, were clustered basally, suggesting that Drosera originated in Africa or Australia. The rbcL tree indicates that Australian species expanded their distribution to South America and then to Africa. Expansion of distribution to the Northern Hemisphere from the Southern Hemispere occurred in a few different lineages.

SummaryTo elucidate the molecular basis of loss of self-incompatibility in Lycopersicon, S-RNases and HTproteins were analysed in seven self-compatible (SC) and three self-incompatible (SI) taxa. No or low stylar RNase activity was a common feature in most SC taxa examined, in contrast to the uniformly high levels of activity found in all SI species. The S-RNase gene is most likely deleted in the four red-fruited SC taxa (L. esculentum, L. esculentum var. cerasiforme, L. pimpinellifolium and L. cheesmanii) because S-RNase genes could not be ampli®ed from genomic DNA. S-RNase genes could, however, be ampli®ed from the genomes of the three green-fruited SC taxa examined. L. chmielewskii and L. hirsutum f. glabratum show a decreased accumulation of transcripts, possibly re¯ecting changes in the 5¢¯anking regions of the S-RNase genes. The remaining green-fruited SC species, L. parvi¯orum, has a functional S-RNase gene in its genome that is expressed at high levels in the style, suggesting a genetic factor responsible for the low S-RNase activity. Together these results argue for several independent mutations in the S-RNase gene over the course of Lycopersicon diversi®cation, and that loss of S-RNase function is unlikely to the primary cause of the loss of self-incompatibility. We also examined the HT-B genes that play a role in self-incompatibility. HT-B transcripts were markedly reduced in the styles of all the SC taxa examined. A scenario is described where a mutation causing reduced transcription of HT-B in an ancestral SI species was central to the loss of self-incompatibility in Lycopersicon.