Summaryrsw3 is a temperature-sensitive mutant of Arabidopsis thaliana showing radially swollen roots and a deficiency in cellulose. The rsw3 gene was identified by a map-based strategy, and shows high similarity to the catalytic a-subunits of glucosidase II from mouse, yeast and potato. These enzymes process N-linked glycans in the ER, so that they bind and then release chaperones as part of the quality control pathway, ensuring correct protein folding. Putative b-subunits for the glucosidase II holoenzyme identified in the Arabidopsis and rice genomes share characteristic motifs (including an HDEL ER-retention signal) with bsubunits in mammals and yeast. The genes encoding the putative a-and b-subunits are single copy and, like the rsw3 phenotype, widely expressed. rsw3 reduces cell number more strongly than cell size in stamen filaments and probably stems. Most features of the rsw3 phenotype are shared with other cellulose-deficient mutants, but some -notably, production of multiple rosettes and a lack of secreted seed mucilageare not and may reflect glucosidase II affecting processes other than cellulose synthesis. The rsw3 root phenotype develops more slowly than the rsw1 and rsw2 phenotypes when seedlings are transferred to the restrictive temperature. This is consistent with rsw3 reducing glycoprotein delivery from the ER to the plasma membrane whereas rsw1 and rsw2 act more rapidly by affecting the properties of already delivered enzymes.

Dynamin-related proteins are large GTPases that deform and cause fission of membranes. The DRP1 family of Arabidopsis thaliana has five members of which DRP1A, DRP1C, and DRP1E are widely expressed. Likely functions of DRP1A were identified by studying rsw9, a null mutant of the Columbia ecotype that grows continuously but with altered morphology. Mutant roots and hypocotyls are short and swollen, features plausibly originating in their cellulose-deficient walls. The reduction in cellulose is specific since non-cellulosic polysaccharides in rsw9 have more arabinose, xylose, and galactose than those in wild type. Cell plates in rsw9 roots lack DRP1A but still retain DRP1E. Abnormally placed and often incomplete cell walls are preceded by abnormally curved cell plates. Notwithstanding these division abnormalities, roots and stems add new cells at wild-type rates and organ elongation slows because rsw9 cells do not grow as long as wild-type cells. Absence of DRP1A reduces endocytotic uptake of FM4-64 into the cytoplasm of root cells and the hypersensitivity of elongation and radial swelling in rsw9 to the trafficking inhibitor monensin suggests that impaired endocytosis may contribute to the development of shorter fatter roots, probably by reducing cellulose synthesis.

Arabidopsis thaliana plants were stably transformed with DNA encoding green fluorescent protein and with sequences ensuring retention in the endoplasmic reticulum (ER). Confocal laser scanning microscopy shows fluorescent ER in many cells of seedlings so allowing developmental changes to be documented. The arrangement of the cortical ER changes as cells mature in the hypocotyl and root epidermis. In the root, cells that have completed expansion have reticulate cortical ER resembling the ER described in many previous studies. Expanding cells, however, show extensive perforated sheets of cortical ER which transform quite abruptly into a loose reticulum at the basipetal end of the elongation zone. The reticulum compacts in trichoblasts beginning at sites where root hairs are about to emerge. The compacted form is maintained throughout the hair until growth ceases and the open reticulate form returns. All forms of cortical ER are dynamic and we use a color overlay method to distinguish stable and moving structures in a single composite image. Reticulate ER continuously rearranges its polygonal layout and perforations move and change their shape in the ER sheets of younger cells. ER deeper in the cell (i.e. not close to the plasma membrane) moves more actively so that almost no tubules remain stable even over short periods of less than one minute. The function of the perforated sheets of cortical ER present in growing cells is unknown.

To investigate the role of N-terminal domains of plant disease resistance proteins in membrane targeting, the N termini of a number of Arabidopsis and flax disease resistance proteins were fused to green fluorescent protein (GFP) and the fusion proteins localized in planta using confocal microscopy. The N termini of the Arabidopsis RPP1-WsB and RPS5 resistance proteins and the PBS1 protein, which is required for RPS5 resistance, targeted GFP to the plasma membrane, and mutation of predicted myristoylation and potential palmitoylation sites resulted in a shift to nucleocytosolic localization. The N-terminal domain of the membrane-attached Arabidopsis RPS2 resistance protein was targeted incompletely to the plasma membrane. In contrast, the N-terminal domains of the Arabidopsis RPP1-WsA and flax L6 and M resistance proteins, which carry predicted signal anchors, were targeted to the endomembrane system, RPP1-WsA to the endoplasmic reticulum and the Golgi apparatus, L6 to the Golgi apparatus, and M to the tonoplast. Full-length L6 was also targeted to the Golgi apparatus. Site-directed mutagenesis of six nonconserved amino acid residues in the signal anchor domains of L6 and M was used to change the localization of the L6 N-terminal fusion protein to that of M and vice versa, showing that these residues control the targeting specificity of the signal anchor. Replacement of the signal anchor domain of L6 by that of M did not affect L6 protein accumulation or resistance against flax rust expressing AvrL567 but removal of the signal anchor domain reduced L6 protein accumulation and L6 resistance, suggesting that membrane attachment is required to stabilize the L6 protein.

SummaryThe Arabidopsis radial swelling mutant rsw10 showed ballooning of root trichoblasts, a lower than wild-type level of cellulose and altered levels of some monosaccharides in non-cellulosic polysaccharides. Map-based cloning showed that the mutated gene (At1g71100) encodes a ribose 5-phosphate isomerase (RPI) and that the rsw10 mutation replaces a conserved glutamic acid residue with lysine. Although RPI is intimately involved with many biochemical pathways, media supplementation experiments suggest that the visible phenotype results from a defect in the production of pyrimidine-based sugar-nucleotide compounds, most likely uridine 5¢-diphosphate-glucose, the presumed substrate of cellulose synthase. Two of three RPI sequences in the nuclear genome are cytoplasmic, while the third has a putative chloroplast transit sequence. The sequence encoding both cytoplasmic enzymes could complement the mutation when expressed behind the CaMV 35S promoter, while fusion of the RSW10 promoter region to the GUS reporter gene established that the gene is expressed in many aerial tissues as well as the roots. The prominence of the rsw10 phenotype in roots probably reflects RSW10 being the only cytosolic RPI in this tissue and the gene encoding the plastid RPI being relatively weakly expressed. We could not, however, detect a decrease in total RPI activity in root extracts. The rsw10 phenotype is prominent near the root tip where cells undergo division, endoreduplication and cell expansion and so are susceptible to a restriction in de novo pyrimidine production. The two cytosolic RPIs probably arose in an ancient duplication event, their present expression patterns representing subfunctionalization of the expression of the original ancestral gene.

Summary.Myosins providing the motors for the actin-based motility that occurs in diverse plants have proved difficult to study. To facilitate those studies, we describe polymerase chain reaction primers that reliably amplify part of the myosin head from diverse plants, consensus sequences that characterise the amplified product as encoding a class V or class VIII myosin, and a monoclonal antibody that recognises an epitope conserved in the head of most plant, fungal, and animal myosins. A pair of stringent oligonucleotide primers was designed that, when used in the polymerase chain reaction, amplified at least eleven different myosins from five species of angiosperms and one sequence from each of the fern Azolla and the algae Nitella and Phaeodactylum. The amplified products, comprising 126 to 135 nucleotides encoding part of the myosin head domain, can be used as myosin-specific probes to screen genomic and cDNA libraries. To identify the products of plant myosin genes, we raised a monoclonal antibody (anti-CHE) to a nine amino acid peptide matching a conserved head epitope showing not more than single amino acid substitutions in most published myosin genes. This antibody recogniscs rabbit skeletal myosin and multiple polypeptides of >100 kDa in four angiosperms and in the alga Nitella. Relating the Mr values of immunoreactive bands in Arabidopsis extracts to the predicted Mr values of the products of five myosin genes supports the view that the antibody recognises both myosins V and VIII together with the products of some as yet unsequenced genes. The previously described MB 170 antibodies may, in contrast, be specific for one or more type V myosins. Together, the polymerase chain reaction primers and the antibody represent versatile tools for identifying and categorising myosins in diverse plants.

The sub-cortical actin bundles of the alga Chara corallina can be selectively extracted with a low salt solution except when cytochalasin B is present. Proteins with molecular weights of 160000, 43000 and 37000 share this extraction behaviour. While chemical cleavage of the 43000 band indicates that it is actin, the nature of the other proteins is unknown. Although the 37000 protein resembles tropomyosin in molecular weight it lacks tropomyosin's distinctively large change in electrophoretic mobility in the presence of urea.

No abstract