Multiple vacuoles in plant cells.

After one retires, it is a pleasant surprise to learn that one’s prior publications are being mentioned in the current literature. In this regard, two articles in the December 2007 issue of Plant Physiology pay me the compliment of challenging some of our published work, and it is my honor to respond. Hunter et al. (2007) and Olbrich et al. (2007) separately challenge the hypothesis that plant cells contain functionally distinct types of vacuoles, in which tonoplast intrinsic protein (TIP) isoforms serve as markers for specific functional types. It is my view that the challenges are unsuccessful and that the plant cell biology community would benefit from a brief discussion of the problems with those articles. Hunter et al. (2007) argue that results in Arabidopsis (Arabidopsis thaliana) contradict the multiple vacuole hypothesis. They present results indicating that, by expression of fusions of yellow fluorescent protein (YFP) with a-, g-, and d-TIPs in transgenic Arabidopsis, driven either by the cauliflower mosaic virus 35S promoter or the respective endogenous TIP promoter, they cannot identify separate vacuolar compartments. Thus ‘‘TIP isoform distribution is tissue and development specific, rather than organelle specific’’ (p. 1371). They further claim, based on their fusion protein results and on publicly available microarray data, that ‘‘only a-TIP is present during embryo maturation and germination. This is in apparent contrast with results obtained by immunofluorescence in Arabidopsis seeds, using peptide antibodies against the d-TIP’’ (p. 1379). Results from western blots of extracts from mature seeds and germinating seeds probed with antiGFP antiserum were stated to support this claim. Thus, the authors appear to argue that Arabidopsis is different from other plants in which traffic to vacuoles has been studied. However, there is strong literature support for the concept that sorting of proteins to vacuoles in Arabidopsis is not different from that described in other species. Using quantitative immuno-electron microscopy (EM) analyses of the distributions of storage proteins, the Cys protease aleurain, the BP80/VSR (vacuolar sorting receptor), and the RMR protein (Park et al., 2007), Hinz et al. (2007) demonstrated in developing Arabidopsis embryos that storage proteins partitioned into dense vesicles with the RMR protein, whereas the BP80/VSR protein was largely associated with clathrincoated vesicles. These results, and the finding that storage protein sorting initiated by aggregation early in the Golgi, were fully consistent with previous studies of developing pea (Pisum sativum) embryos (Hinz et al., 1999; Hillmer et al., 2001), and with data strongly supporting the concept that RMR proteins are sorting receptors for proteins carrying C-terminal vacuolar sorting signals (Park et al., 2005, 2007). The claim by Hunter et al. (2007) that it was not possible to visualize organelles in vegetative Arabidopsis tissues, in which separate sorting of soluble proteins for the lytic and storage protein pathways could be identified, is at odds with results presented in another article in the same December 2007 edition (Poustka et al., 2007). Additionally, the interpretation of Hunter et al. (2007) is at odds with published data demonstrating a key role for a conserved sequence motif in the cytoplasmic C terminus of a-TIP in targeting it in its unique pathway to a vacuole destination (Oufattole et al., 2005). In those experiments, a fusion of GFP with the N terminus of a-TIP was transiently expressed in tobacco (Nicotiana tabacum) suspension culture protoplasts. The wild-type protein clearly localized to small vacuoles that were separate from the central vacuole, whereas a mutant protein lacking the conserved cytoplasmic sequence underwent altered proteolytic cleavage and was additionally present on the central vacuole tonoplast (Oufattole et al., 2005). These considerations are relevant because the transgenic Arabidopsis plants studied by Hunter et al. (2007) expressed C-terminal fusions of YFP to the TIPs. Although the authors studied the localization of both the N-terminal and C-terminal fusions, the possibility cannot be eliminated that C-terminal fusions may have altered the targeting. Overexpression of our a-TIP fusion proteins appeared to be toxic to the protoplasts (Oufattole et al., 2005), a fact that would not be surprising considering that it presumably actively functioned as an aquaporin and might well have altered vacuolar volume or composition adversely. Although a portion of the results of Hunter et al. (2007) used expression from the native TIP promoters, in effect this would still result in overexpression of the respective proteins because the transgene expression would be added on top of the endogenous gene expression. Their western-blot results strongly indicate that this probably was detrimental to the plants (supplemental figure S5B from Hunter et al. [2007]). There, the anti-GFP antibodies detect only a protein the expected size of GFP itself, not of a GFP-TIP fusion protein. This would be explained if the altered vacuole membrane containing an excess of TIP protein were targeted for degradation. Indeed, internalization of vacuole membrane is associated with the protein aggregation/membrane internalization pathway followed by RMR proteins (Jiang et al., 2000; Oufattole et al., 2005); when a fusion protein in which the RMR cytoplasmic tail tagged with GFP was expressed, the cytoplasmic portion was internalized such that it was proteolytically cleaved to release intact GFP that was visualized in the central vacuole lumen (Park et al., 2007). www.plantphysiol.org/cgi/doi/10.1104/pp.107.900248