SummaryVacuolar compartments associated with leaf senescence and the subcellular localization of the senescencespecific cysteine-protease SAG12 (senescence-associated gene 12) were studied using specific fluorescent markers, the expression of reporter genes, and the analysis of high-pressure frozen/freeze-substituted samples. Senescence-associated vacuoles (SAVs) with intense proteolytic activity develop in the peripheral cytoplasm of mesophyll and guard cells in Arabidopsis and soybean. The vacuolar identity of these compartments was confirmed by immunolabeling with specific antibody markers. SAVs and the central vacuole differ in their acidity and tonoplast composition: SAVs are more acidic than the central vacuole and, whereas the tonoplast of central vacuoles is highly enriched in c-TIP (tonoplast intrinsic protein), the tonoplast of SAVs lacks this aquaporin. The expression of a SAG12-GFP fusion protein in transgenic Arabidopsis plants shows that SAG12 localizes to SAVs. The analysis of Pro SAG12 :GUS transgenic plants indicates that SAG12 expression in senescing leaves is restricted to SAV-containing cells, for example, mesophyll and guard cells. A homozygous sag12 Arabidopsis mutant develops SAVs and does not show any visually detectable phenotypical alteration during senescence, indicating that SAG12 is not required either for SAV formation or for progression of visual symptoms of senescence. The presence of two types of vacuoles in senescing leaves could provide different lytic compartments for the dismantling of specific cellular components. The possible origin and functions of SAVs during leaf senescence are discussed.

We have investigated the process of somatic-type cytokinesis in Arabidopsis (Arabidopsis thaliana) meristem cells with a three-dimensional resolution of~7 nm by electron tomography of high-pressure frozen/freeze-substituted samples. Our data demonstrate that this process can be divided into four phases: phragmoplast initials, solid phragmoplast, transitional phragmoplast, and ring-shaped phragmoplast. Phragmoplast initials arise from clusters of polar microtubules (MTs) during late anaphase. At their equatorial planes, cell plate assembly sites are formed, consisting of a filamentous ribosomeexcluding cell plate assembly matrix (CPAM) and Golgi-derived vesicles. The CPAM, which is found only around growing cell plate regions, is suggested to be responsible for regulating cell plate growth. Virtually all phragmoplast MTs terminate inside the CPAM. This association directs vesicles to the CPAM and thereby to the growing cell plate. Cell plate formation within the CPAM appears to be initiated by the tethering of vesicles by exocyst-like complexes. After vesicle fusion, hourglass-shaped vesicle intermediates are stretched to dumbbells by a mechanism that appears to involve the expansion of dynamin-like springs. This stretching process reduces vesicle volume by~50%. At the same time, the lateral expansion of the phragmoplast initials and their CPAMs gives rise to the solid phragmoplast. Later arriving vesicles begin to fuse to the bulbous ends of the dumbbells, giving rise to the tubulo-vesicular membrane network (TVN). During the transitional phragmoplast stage, the CPAM and MTs disassemble and then reform in a peripheral ring phragmoplast configuration. This creates the centrifugally expanding peripheral cell plate growth zone, which leads to cell plate fusion with the cell wall. Simultaneously, the central TVN begins to mature into a tubular network, and ultimately into a planar fenestrated sheet (PFS), through the removal of membrane via clathrin-coated vesicles and by callose synthesis. Small secondary CPAMs with attached MTs arise de novo over remaining large fenestrae to focus local growth to these regions. When all of the fenestrae are closed, the new cell wall is complete. Few endoplasmic reticulum (ER) membranes are seen associated with the phragmoplast initials and with the TVN cell plate that is formed within the solid phragmoplast. ER progressively accumulates thereafter, reaching a maximum during the late PFS stage, when most cell plate growth is completed.

Plastoglobules are lipoprotein particles inside chloroplasts. Their numbers have been shown to increase during the upregulation of plastid lipid metabolism in response to oxidative stress and during senescence. In this study, we used stateof-the-art high-pressure freezing/freeze-substitution methods combined with electron tomography as well as freeze-etch electron microscopy to characterize the structure and spatial relationship of plastoglobules to thylakoid membranes in developing, mature, and senescing chloroplasts. We demonstrate that plastoglobules are attached to thylakoids through a half-lipid bilayer that surrounds the globule contents and is continuous with the stroma-side leaflet of the thylakoid membrane. During oxidative stress and senescence, plastoglobules form linkage groups that are attached to each other and remain continuous with the thylakoid membrane by extensions of the half-lipid bilayer. Using three-dimensional tomography combined with immunolabeling techniques, we show that the plastoglobules contain the enzyme tocopherol cyclase (VTE1) and that this enzyme extends across the surface monolayer into the interior of the plastoglobules. These findings demonstrate that plastoglobules function as both lipid biosynthesis and storage subcompartments of thylakoid membranes. The permanent structural coupling between plastoglobules and thylakoid membranes suggests that the lipid molecules contained in the plastoglobule cores (carotenoids, plastoquinone, and tocopherol [vitamin E]) are in a dynamic equilibrium with those located in the thylakoid membranes.

SV buds outside the plane of the cisterna and a 30-35% reduction in cisternal membrane area. Free TGN compartments are defined as cisternae that have detached from the Golgi to become independent organelles. Golgi-associated and free TGN compartments, but not trans Golgi cisternae, bind anti-RabA4b and anti-phosphatidylinositol-4 kinase (PI-4K) antibodies. RabA4b and PI-4Kβ1 localize to budding SVs in the TGN and to SVs en route to the cell surface. SV and CCV release occurs simultaneously via cisternal fragmentation, which typically yields ∼30 vesicles and one to four residual cisternal fragments. Early endosomal markers, VHA-a1-green fluorescent protein (GFP) and SYP61-cyan fluorescent protein (CFP), colocalized with RabA4b in TGN cisternae, suggesting that the secretory and endocytic pathways converge at the TGN. pi4kβ1/pi4kβ2 knockout mutant plants produce SVs with highly variable sizes indicating that PI-4Kβ1/2 regulates SV size.

Most of our current knowledge of cellular ultrastructure is derived from studies of chemically fixed and chemically cryoprotected preparations. In the first part of this review, we document the many artifacts associated with chemical techniques that render them unsuitable for further refinement of our understanding of cellular ultrastructure. The best method currently available for the preservation of cellular ultrastructure is ultrarapid freezing. The second part of this review is a consideration of the physics of ice crystal formation in biological systems, which suggests that ice crystals will be present in any frozen, uncryoprotected specimen. We define an ultrarapidly frozen preparation as one in which the ice crystals are so small as to be invisible at the electron microscopic level. Improvements in the ease of application and reliability of ultrarapid freezing techniques have reached the point that these techniques can be used by anyone requiring the best achievable preservation of cellular ultrastructure. In the third part of this review, we describe and critique the five methods of ultrarapid freezing in current use.