Ubisch bodies and pollen ontogeny in Oxalis articulata Savigny

The correlation between the ontogeny of Ubisch bodies and pollen development in Oxalis articulata was studied with Transmission Electron Microscopy (TEM). The ultrastructural changes observed during the different stages of development in the tapetal cells are related to Ubisch bodies, sporopollenin and pollen-kitt formation. The pro-orbicules have the appearance of lipid globuli and their formation is related to the endoplasmic reticulum of rough type (ERr). The lipid globules or pro-orbicules disappear in the mature Ubisch bodies, and the places that they occupied remain free of contents or with pollen-kitt. BIOCELL 2005, 29(3): 271-278 ISSN 0327 9545 PRINTED IN ARGENTINA


Introduction
The genus Oxalis L. comprises about 800 cosmopolitan species. Previous studies on microsporogenesis and microgametogenesis in this family are meagre (Johri et al., 1992). The majority of them have been carried out in the genus Averrhoa and very few ones in the genus Oxalis (Herr, 1972;Guth and Weller 1986). Ubisch (1927) observed granular bodies along the inner walls of tapetal cells in Oxalis and she pointed out the staining properties similarities of these bodies to pollen exine. These granular bodies were described by Ubisch (op. cit.) as very showy small plates due their size and form. In spite of this fact, ultrastructural studies of these bodies have been performed in only one species of this genus until this moment (Carniel, 1967). The aim of this paper is to describe the ultrastructure and development of Ubisch bodies and pollen grains in Oxalis articulata Savigny in order to advance the understanding of the orbicules morphology and function.

Material and Methods
The material of Oxalis articulata Savigny was collected in FCEyN, Ciudad Universitaria, Nuñez, Buenos Aires, Argentina.
Material was fixed in FAA (formalin, alcohol, acetic-acid) for light microscopy studies. Sections were made by means of standard microtome techniques. The slides were stained in a safranin-fast green combination, cresyl violet or toluidine blue and mounted in synthetic resin (D'Ambrogio, 1986). The material was observed with a Wild M20 microscope.
Material fixed in FAA was transferred to 100° acetone and then air dried for scanning electron microscopy (SEM) studies. The sputtering treatment was made with gold-palladium for 3 min.
For transmission electron microscopy (TEM) studies, the material was pre-fixed in 2.5% glutaraldehyde in phosphate buffer (pH 7.2) for 2 h and post-fixed in OsO 4 at 2 o C in the same buffer for 3 h. Then the material was dehydrated in ascending ethanol series and embedded in Spurr's resin. Fine sections were made on a Sorvall ultramicrotome, stained with uranyl acetate and lead citrate (O'Brien and Mc Cully, 1981), observed and photographed in a JEOL 100c TEM.

Results
We identified five stages of Ubisch bodies and pollen development in Oxalis articulata.
The cytoplasm of the tapetal cells is very dense, with a similar ultrastructure to that of the MMC (Fig. 4).

Stage 2: Tetrad formation
Cytokinesis is simultaneous resulting in tetrahedral microspore tetrads (Fig. 5). The middle lamella of the MMC is disintegrated but the primary wall is still present (Fig. 6). When the tetrad is formed, an incipient primexine is seen between the callosic wall and the plasmalemma (Fig. 6).
The microspore cytoplasm shows high activity and has many mitochondria. Plastids and large osmiophilic globules are scattered in the cytoplasm (Fig. 6).
In this stage, the tapetal cells are binucleate and more vacuolated. Their cytoplasm shows an ERr increase. The endoplasmic reticulum cisternae are arranged in dense and uniformly thick rows, parallel to the cell wall (Fig. 7). There are also many proplastids, mitochondria and small lipidic globules (Fig. 7).
The tapetal cell wall begins to dissolve and presents a lax structure. Many cytoplasmic connections go through the radial walls of these cells (Fig. 7).

Stage 3: Free young microspores
After the dissolution of the callose wall, the sporopollenin wall begins to form. Three different electron-dense layers form the pollen grain wall: ectexine with interrupted tectum, columellae and foot layer; endexine and intine still in formation (Figs. 8,11).
The microspores have a conspicuous nucleus and the structure of their cytoplasm is similar to the previous stage (Fig. 11).
The tapetal cells show a dense cytoplasm with abundant ERr, numerous free ribosomes and well development plastids. The ERr lost the parallel configuration and its cisternae are extensive (Fig. 9).
Moderate electron-dense globules or pro-orbicules appear outside the plasmalemma at the inner tangential face of the tapetal cells (Figs. 8,9). These globules are observed on depressions or crypts of the tapetal cells plasmalemma.
A fibrillar electron-dense material is observed inside the anther locule . This material is deposited on the developing exine and on several globules or prorbicules (Fig. 10). One to six globules are assembled for sporopollenin deposition and they form the young Ubisch bodies (Figs. 8-10).

Stage 4: Free microspores
The microspores have a conspicuous nucleus and their cytoplasm is limited to a parietal position due to the presence of a large vacuole (Fig. 12).
Many mitochondria and large lipidic globules are observed in the microspores cytoplasm (Fig. 14).
Ubisch bodies and exine have the same high electron-density (Fig. 14). The globules present in the Ubisch bodies at the previous stage are no longer observed (Figs. 14,16).

Stage 5: Mature pollen grain
Pollen grains are two-celled when shed. The vegetative cell encloses the generative cell (Fig. 20).
The generative cell acquires a fusiform aspect with a lobulate outline (Fig. 20).
This cell has an electron-translucent wall with some plasmodesmata and membranous inclusions (Fig. 20). The generative cell cytoplasm is highly reduced when compared to the large nucleus (Fig. 20).
The pollen grain wall and the Ubisch bodies are fully developed (Figs. 13, 15, 17). The ectexine and the Ubisch bodies sporopollenin is compressed and shows less electron-density that in the previous stage (Figs. 15, 17).
At this stage, abundant pollenkitt is present among pollen wall columellae and connects different grains in the anther locule (Figs. 15, 21).
The places occupied by the globules during Ubisch bodies formation are reduced in size and are observed as small perforations in fine sections. These places are free of contents or occupied by pollen-kitt (Figs. 16, 17).
These Ubisch bodies are seen as multiperforated plates in paradermal view under light microscopy (Fig.  18) and as plates with dome shape in lateral view and smooth surface under scanning electron microspoy (Fig. 24).
The ectexine is reduced to small granular fragments in the colpus (Fig. 23).

Discussion
This is the first report about Ubisch bodies ontogeny in relation with pollen development in the genus Oxalis.
The microsporogenesis and microgametogenesis in O. articulata are similar to previous descriptions in other species of the genus (Johri et al. 1992).
In this species the pollen grains are two-celled when shed. According to Pargney (1978), O. corniculata pollen grains are three celled at the shedding stage.
The ultraestructural changes observed during the different stages of development in the tapetal cells are related to Ubisch Bodies, sporopollenin and pollen-kitt formation.
At the free young microspores stage, numerous globuli appear outside the plasmalemma at the inner tangencial face of the tapetal cells. Due to their electron-density, these globuli seem to be constituted by lipidic substances and may be consider as pro-orbicules. Similar observations and interpretations were made by Clement and Audran (1993 a,b) and Heslop-Harrison and Dickinson (1969) in Lilium; Christensen et al. (1972) in Sorgum bicolor and Steer (1977) in Avena.
Pro-orbicules are observed on depressions or crypts of the tapetal cells plasmalemma. Rowley and Walles (1987) and Steer (1977) found similar crypts and thought that they constituted a site of lipid deposition during pro-orbicule formation.
A fibrillar substance with the same electron density as sporopollenin accumulates on the globuli after these bodies are outside the plasmalemma of the tapetal cells. For this reason, the accretion of sporopollenin on them appears extracellular. These observations are in concordance with the opinion of Amela Garcia et al. (2002), Galati and Strittmatter (1999), Galati and Rosenfeldt (1998), Clement and Audran (1993 a,b), Christensen et al. (1972), and Echlin (1971).
According to Carniel (1967), the Ubisch bodies of O. pubescens are developed before tapetal cell wall total desintegration. On the contrary, in O. corniculata the tapetal cells wall begins to dissolve at the tetrade stage, and these walls are totally degraded when the sporopollenin accumulates on the globuli at the young microspore stage.
The lipid globules or pro-orbicules disappear in the mature Ubisch bodies, and the places that they occupied remain free of contains or with pollen-kitt.
In the young pollen grain, the vegetative cell goes through a growth period, characterized by the synthesis of new cytoplasm and an increase in the number of organelles, as it has been reported for the majority of the ultrastructurally studied species (Raghavan, 1997). Many small vesicles are scattered in the cytoplasm of this cell. According to Fisher et al. (1968) these vesicles derive from dictyosomes and according to Noguchi (1990), vesicles of this type could fuse to form vacuoles in the emerging pollen tube. Amela García et al. (2002) thought that the small vesicles could constitute a membrane reservoir for the initial fast growth of the pollen tube, assuring its anchoring on the stigma.
The generative cell of Oxalis articulata presents some evaginations that protrude into the cytoplasm of the vegetative cell. Tiwari (1994) found that the wall processes arise from all faces of the generative cell and protrude into the cytoplasm of vegetative cell. According to Theunis et al. (1985) the ends of Rhododendron laetum generative cell narrow into long evaginations.
We can speculate that O.articulata generative cell morphology could support the transport of metabolites from the vegetative cell.
The pollen morphology coincides with the general description for the genus (Erdtman, 1952).
According to Wang and Chen (2001) O. acetocella pollen grains are tetracolpate and rarely tri or penta colpate. However, we have not observed pollen grains 4-or 5-colpate in O articulata, only 3-colpate.