Arch. Histol. Cytol., Vol. 55, Suppl. (1992) p. 217-224

Ultrastructural Organization of Two Tapetal Types in Angiosperms

Susan H. BARNES and Stephen BLACKMORE

The Natural History Museum, London, United Kingdom

Received February 18, 1992

Summary. The development of preparation techniques As has been the case for other tissues within the that include freeze fracturing provide an ideal method developing anther, previous studies have documented for studying the differentiation of plant tissues in the the organization of the tapetum at the level of optical scanning electron microscope. This is illustrated with microscopy (see for example, SCHNARF,1923; UBISCH, reference to tapetal development in Catananche caer- 1927) and through the application of transmission ulea, which has a plasmodial tapetum, and in Lolium electron microscopy (see for example, DICKINSONand perenne, which has a secretory tapetum. LEWIS, 1973, PACINI and KEIJZER, 1989; EL-GHAZALY and NILSSON, 1991). Scanning electron microscope The tapetum is a specialised tissue concerned with studies had to await the development of techniques the nutrition of the developing spores, and is found in which permitted the examination of the internal sur- the sporangia of lower plants and anthers of higher faces of organs, tissues, cells and organelles. We have plants (for reviews, see PACINI et al.,1985; CHAPMAN, reviewed the historical development of such tech- 1987). Tapetal cells exhibit a variety of develop- niques in relation to plant ultrastructure (BARNES and mental pathways, especially in terms of nuclear BLACKMORE,1986a) and emphasised the importance divisions, the behaviour of the cell walls and the of the pioneering work of Professor Keiichi TANAKA. synthesis of pollen wall precursors. Two major He devised a series of techniques ranging from crack- ultrastructural classes of tapetum are found in the ing resin embedded blocks (TANAKA, 1974) through anthers of angiosperms. The secretory tapetum is to methods involving chemical fixation followed characterised by persistent cell walls and is also by freeze fracturing (TANAKA, 1981; TANAKA and referred to as glandular, parietal, or cellular. In NAGURO, 1981). Advances in specimen handling and contrast, the cells walls of the plasmodial tapetum processing have enabled the basic technique to be break down during microspore development. This applied to free cells suspended in culture medium second kind of tapetum is sometimes referred to as (FUKUDOME and TANAKA, 1986). invasive or amoeboid. The two major classes can Although initially applied to animal tissues, the basic themselves be subdivided on the basis of various technique proved readily adaptable to plant tissues criteria, particularly the sequence of events during by adopting an extended period of extraction in microsporogenesis (PACINI et al., 1985; PACINI, 1990; osmium tetroxide (BARNES and BLACKMORE, 1984a, PACINI and FRANCHI (1991). b). One area of botanical research where the tech- In the course of a programme of comparative nique, which we refer to as freeze-fracture and cyto- studies of pollen ontogeny aimed at determining the plasmic maceration, has proved particularly suitable morphogenetic pathways associated with taxon- has been the study of pollen and spore ontogeny specific features, we have examined representatives (BLACKMORE and BARNES, 1985, 1987, 1990; BARNES of the two major tapetal classes. Here, we describe and BLACKMORE, 1986b; DICKINSON and SHELDON, and illustrate the later, mainly post tetrad, develop- 1986; WILMS et al., 1986). mental stages of the secretory tapetum of Lolium As observations of the tapetum demonstrate, the perenne L. (Gramineae) and the plasmodial tapetum freeze fracture and cytoplasmic maceration tech- of Catananche caerulea L. (Compositae: Lactuceae), nique is particularly suited to the study of the spatial as studied by means of scanning electron microscopy. relations of membrane-bound organelles.

217 218 S. H. BARNES and S. BLACKMORE:

plasmodial tapeta (ALBERTINIet al., 1987), the tapetal MATERIALS AND METHODS cells subsequently intrude into the spaces between developing microspores. In contrast, the tapetum Anthers at various stages of development were taken remains an organized cylindrical tissue in the anthers from plants of Lolium perenne L. (Gramineae) col- of flowering plants with secretory tapeta, recognized lected from wild populations in Sussex and from from 175 families (ALBERTINI et al., 1987). specimens of Catananche caerulea L. (Compositae: In Catananche, as in Cichorium (PACINI and KEIJZER, Lactuceae) cultivated at Chelsea Physic Gardens, 1989), the tapetum does not become invasive until London. after the tetrad stage (Fig. 2). During the tetrad stage The anthers were prepared by the freeze fracture the tapetal cell walls remain intact and distinct. The and cytoplasmic maceration technique as described nuclei of most, if not all, tapetal cells undergo mitosis by BARNES and BLACKMORE (1984a). Catananche during this stage. As in other tapetal cells (see for anthers at later stages of development were dissected example, CHAPMAN, 1987), the nuclear division is not from the bud and individually trimmed at each end followed by cytokinesis. The cells are therefore binu- to assist the penetration of solutions. Early stage cleate, with the two nuclei remaining in close contact. anthers were processed in buds or groups of anthers After the callose special cell wall around the tetrads to facilitate easy handling. Lolium anthers were disperses, the free microspores are released into the dissected and processed singly. Prepared anthers locule (Fig. 3). The developing microspores have a were fixed in 1% osmium tetroxide in M/15 phos- spiny exine by this stage, the outer layer of the exine phate buffer for 2-16h with continuous rotation. (the ectexine) is not fully differentiated (Fig. 4). The They were then washed in buffer and transferred tapetum starts to intrude between the microspores through 15%, 30% and 50% dimethyl sulphoxide for early in the free microspore stage. Each tapetal cell 30 min in each solution. The anthers were freeze- has a continuous plasma membrane and has an fractured on a liquid nitrogen-cooled metal block organelle-rich cytoplasm, with particularly abundant using a razor blade and hammer. The fragments were endoplasmic reticulum (Fig. 5). At this stage of devel- collected and thawed in fresh 50% dimethyl sulphox- opment the tapetum is generally considered to be a ide. After copious washing in buffer the specimens highly active secretory tissue producing sporopol- were transferred to 0.1% osmium tetroxide in M/15 lenin precursors which form the exine. Consistent phosphate buffer and left to macerate at 4C for 14 with this interpretation, numerous dictyosomes are days. During this period the specimens were regularly present in the tapetal cytoplasm (Fig. 6). checked and the solution replenished if it had dis- As the microspores develop (Figs. 7, 8) the exine coloured. To enhance electrical conductivity the spec- differentiates and becomes caveate and the micro- imens were fixed in 1% osmium tetroxide, washed, spore nuclei enlarge. At this stage, the tapetum treated with 2% tannic acid, washed again and intrudes between the microspores but retains a con- refixed in 1% osmium tetroxide. They were then tinuous plasma membrane around its convoluted dehydrated by transfer through an acetone series and inner surface. As the tapetal cells extend towards the critical point dried. The pieces were mounted on to centre of the anther locule their large nuclei maintain specimen stubs using Araldite adhesive and sputter a peripheral position. The tapetal cytoplasm (Fig. 8) coated with approximately 15nm of gold/palladium. continues to be actively synthetic and retains a highly Images were taken using a Hitachi 5800 field emis- organized system of endoplasmic reticulum. sion scanning electron microscope at an accelerating Once the exine reaches its mature morphology, the voltage of 8kV. microspores enter a characteristic vacuolate stage (Fig. 9) in which the mlcrospore cytoplasm is dis- placed to the periphery by a large vacuole. The tapetum, having completed its contribution to exine RESULTS AND DISCUSSION synthesis, begins to degenerate. The most conspicu- ous change, at this stage, is the absence of extensive Plasmodial tapetum sheets of endoplasmic reticulum (Fig. 10). Numerous When the anther tissues first differentiate, the tapetal vesicles are now present in the tapetal cytoplasm and cells form a cylinder surrounded by both the parietal the plasma membranes are no longer continuous. The cells and the epidermal cells and enclosing the micro- freeze fracture and cytoplasmic maceration tech- sporocytes. This configuration (Figs. 1, 2) initially nique reveals fine details of the pollen wall, such as occurs in both major classes of tapetum. In the Com- the minute cavities known as internal f oraminae positae, and in 31 other flowering plant families with (SKVARLA and TURNER, 1966) present within elements Two Tapetal Types in Angiosperms 219

I 4

2 5

3 6

Figs. 1-6. Catananche caerulea. Fig. 1. Premeiotic anther locule. x 960. Fig. 2. Locule at tetrad stage, tapetum binucleate. x 880. Fig. 3. Locule with free microspores. x 880. Fig. 4. Detail of Fig. 3. x 2,400. Fig. 5. Detail of microspore and tapetum from Fig. 3. x 8,800. Fig. 6. Detail of tapetal dictyosomes in Fig. 3. x 48,000 Abbreviations: A aperture, E exine, ER endoplasmic reticulum, D dictyosome, M microspore, MS microsporo- cyte, N nucleus, O orbicule, L lipid, T tapetum, V vacuole. 220 S. H. BARNES and S. BLACKMORF:

of the ectexine. Catananche, in common with other members of the At a later stage (Figs. 11, 12), the nucleus of each Lactuceae, has tricellular pollen grains at maturity. microspore undergoes mitosis, giving rise to a gener- This situation results from a further mitotic division ative and a vegetative cell. Technically this division of the generative cell (Fig. 11), giving rise to the male marks the transition from microspore to pollen grain. germ unit (BARNES and BLACKMORE, 1987). By this

7 10

8 11

9 12

Fig. 7-12. Catananche cc rulea. Fig. 7. Locule with plasmodial tapetum and free microspores. x 800. Fig. 8. Detail of Fig. 7. x 3,200. Fig. 9. Locule with vacuolate microspores and degenerating tapetum. x 640. Fig. 10. Detail of exine and tapetum from Fig. 9. X8,000. Fig. 11. Locule of tricellular pollen grains. x560. Fig. 12. Surface of mature pollen. x 5,600 Two Tapetal Types in Angiosperms 221 stage the tapetum has almost completely degenerat- known as pollenkitt. Entomophilous plants, such as ed. Some discrete organelles, notably mitochondria, Catananche, have pollenkitt which contains a large are still present but lipid droplets are the most abun- number of lipid droplets and is therefore very sticky dant features at this stage (Fig. 12). This process of in nature, thus allowing pollen grains to adhere to degeneration gives rise to a oily surface coating, each other to aid dispersal by insects.

13 16

14 17

15 18 Figs. 13-18. Lolium perenne. Fig. 13. Anther at vacuolate microspore stage. x 180. Fig. 14. Orientation of aperture towards tapetal membrane. x 2,000. Fig. 15. Tapetal membrane and organelles. x 12,000. Fig. 16. Epidermal and tapetal cells. x5,600. Fig. 17. Amyloplast in detail of Fig. 16. x 20,000. Fig. 18. Concentric membranes in tapetal cytoplasm. X28,000 222 S. H. BARNES and S. BLACKMORE:

opment (VITHANAGE and KNOX, 1980). The SEM Secretory tapetum observations presented here represent a selection of Development of the anther of Lolium perenne, has the later developmental stages and supplement the been investigated in detail, with special reference to detailed findings of the transmission electron micro- the tapetum (PACINI et al., 1992) and to pollen devel- scope studies.

19 22

20 23

21 24 Figs. 19-24. Lolium perenne. Fig. 19. Locule with almost mature pollen grains. x 480. Fig. 20. Detail of Fig. 19. x3,200. Fig. 21. Pollen grains and degenerating tapetum. x1,600. Fig. 22. Detail of Fig. 21. showing chloroplasts of parietal cell. x 12,000. Fig. 23. Inner surface of tapetal membrane x 2,800. Fig. 24. Detail of Fig. 23. x 16,000 Two Tapetal Types in Angiosperms 223

The anther of Lolium has four locules with a nar- point where large lipidic droplets are the most con- row central connective (Fig. 13). The anther wall is spicuous feature. relatively thin, consisting of an outer epidermal Lolium like other members of the Gramineae is layer, with a very narrow layer of parietal and anemophilous. The products of tapetal degeneration tapetal cells. The centre of each locule is a large are absorbed by the pollen grains leaving no tapetal space, which is filled with fluid until shortly before remnant except the Ubisch bodies (HESSE, 1980; anther dehiscence. The developing microspores ex- PACINI, 1990), consequently the pollen grains are hibit a special orientation with respect to the devoid of sticky pollenkitt and are readily dispersed tapetum, as in other Gramineae (CHRISTENSEN and by the wind. HORNER, 1974). Each microspore lies with the single The inner surface of the tapetal membrane after distal porate aperture in intimate contact with the anther dehiscence (Figs. 23, 24) shows clearly, the tapetal membrane (Fig. 14). This arragement has boundaries between individual tapetal cells. No spe- been considered by PACINI (1990) and others to facili- cial features are recognizable that would indicate the tate the uptake of nutrients which pass from the points of contact of the pollen apertures. At higher tapetum through the aperture into the microspore. magnification it can be seen that the orbicules are This tapetal membrane is of considerable interest, attached to a sporopollenin membrane with a some- because, in contrast to that of the plasmodial tape- what reticulate organization. tum, it develops special accretions of sporopollenin. These structures, termed Ubisch bodies, were de- scribed by VON UBISCH in 1927. Ubisch bodies (Fig. 15) CONCLUSIONS are characteristically found in seed plants with secre- tory tapeta (PACINI, 1990) and similar structures occur The example of the tapetum demonstrates the value in pteridophytes with secretory tapeta (LUGARDON, of the freeze fracture and cytoplasmic maceration 1981). technique as a method of studying the developmental The organelles within the cytoplasm of the secre- programmes of various tissues within an organ such tory tapetum differ from those of the plasmodial as the anther. tapetum. Rather than the abundant sheets of rough endoplasmic reticulum, numerous plastids can be Acknowledgements. We thank Prof. Keiichi TANAKA for seen (Figs. 15-18). As PACINI et al. (1992) have docu- his advice during a visit to the Natural History Museum mented in Lolium, there is a transition from chromo- in 1983 which inspired us to apply his techniques to pollen plasts to elaioplasts during microspore development. development. We are grateful to Chelsea Physic Garden In the shallow, spreading cells of the tapetum (Fig. for the provision of plant material. 16) the disposition of the organelles and their pro- ducts can be seen. These include numerous amylo- plasts (Fig. 17) and abundant lipid droplets. Such tapetal accumulations of lipid are temporary, and are normally considered to be absorbed by the developing REFERENCES pollen grains (PACINI, 1990). Thus, when degeneration of the taptum takes place, these lipids are absorbed ALBERTINI, L., A. SOUVRE and J. C. AUDRAN: Le tapisde into the pollen grains, rather than contributing to a 1'anthere et les relations avec les microsporocytes et les surface pollenkitt. The tapetal cytoplasm also con- grains du pollen. Rev. Cytol, Biol. Veget. Bot. 10: 211- tains tightly packed concentric membranes that may 242 (1987). themselves be derived from plastids (Fig. 18). BARNES, S. H. and S. BLACKMORE: Freeze fracture and In anthers with almost mature pollen prams (Figs. cytoplasmic maceration in botanical scanning electron 19-24) further details of the relationship between the microscopy. J. Microsc. 136: RP 3-4 (1984a). tapetum and pollen can be observed. As in plasmodial : Scanning electron microscopy of tapeta, the nuclei of the tapetum undergo incomplete chloroplast ultrastructure. Micron Microsc. Acta 15: mitotic division (Figs. 19, 20) after which the chromo- 187-194 (1984b). : Plant ultrastructure in the scanning somes remain partially condensed (Fig. 21). electron microscope. Scan. Electron Microsc./1986/I: The contrasting ultrastructure of parietal cells and 281-289 (1986a). tapetal cells illustrates the differing developmental BARNES, S. H. and S. BLACKMORE: Some functional fea- programme of each tissue. Parietal cells (Figs. 21, 22) tures during pollen development. In: (ed. by) S. BLACK- contain functional chloroplasts of typical form. The MORE and I. K. FERGUSON: Pollen and spores: form and degeneration of the tapetum has progressed to the function. Academic Press, London, 1986b (p. 71-80). 224 S. H. BARNES and S. BLACKMORE

BARNES, S. H. and S. BLACKMORE: Preliminary observa- PACINI, E. and C. J. KEIJER: Ontogeny of intruding tions on the formation of the male germ unit in non-periplasmodial tapetum in the wild chicory, Cichor- Catananche caerulea L. (Compositae: Lactuceae). Proto- ium intybus (Compositae). Plant Syst. Evolut. 67: 149- plasma 138: 190-192 (1987). 164 (1989). BLACKMORE,S. and S. H. BARNES: Cosmos pollen PACINI, E., G. G. FRANCHIand M. HESSE: The tapetum: ontogeny: a scanning electron microscope study. Proto- its form, function and possible phylogeny in Embryo- plasma 126: 91-99 (1985). phyta. Plant Sys. and Evolut. 149: 155-185 (1985). : Pollen wall morphogenesis in PACINI, E., P. E. TAYLOR,M. B. SINGHand R. B. KNOX: Tragopogon porrifolius (Compositae: Lactuceae) and its Development of plastids, including amyloplasts and taxonomic significance. Rev. Palaeobot. Palynol. 52: starch granules, in pollen and tapetum of rye-grass, 233-246 (1987). Lolium perenne L. Ann. Bot. (1992, in press). : Angiosperm pollen wall ontogeny. In SCHNARF,K.: Kleine Beitrage zur Entwicklungsgeschich- (ed. by) S. BLACKMORE and R.B. KNOX: Microspores: te der Angiospermen. Osterr. Bot. Z. 72: 242-245 (1923). ontogeny and evolution, Academic Press, 1990 (p. 173- SKVARLA,J. J. and B. L. TURNER: Systematic implica- 192). tions from electron microscope studies of Compositae CHAPMAN,G. P.: The tapetum. In: (ed. by) K. L. GILES pollen-a review. Ann. Missouri Bot. Garden 53: 230- and J. PRAKESH: Pollen: cytology and development. 256 (1966). Academic Press, London, 1987 (p. 111-125). TANAKA, K.: Frozen resin cracking method and its role in CHRISTENSEN, J. E. and H. T. HORNER: Pollen pore cytology. In: (ed. by) M. A. HAYAT: Principles and development and its spatial orientation during micro- techniques of scanning electron microscopy. Van Nor- sporogenesis in the grass Sorghum bicolor. Amer. J. strand Reinhold, New York, 1974 (Vol. 1, p. 125-134).- Bot. 61: 604-623 (1974). : Demonstration of intracellular structures by DICKINSON,H. G. and D. LEWIS: The formation of high resolution scanning electron microscopy. Scan. tryphine coating the pollen grains of Raphanus and its Electron Microsc./ 1981/ II: 1-8 (1981). properties relating to the self incompatibility system. TANAKA,K. and T. NAGURO:High resolution scanning Proc. Roy. Soc. Lond. B184: 149-165 (1973). electron microscopy of cell organelles by a new speci- DICKINSON, H. G. and J. M. SHELDON: The generation of men preparation method. Biomed. Res. 2: 63-70 (1981). patterning at the plasma membrane of the young UBISCH,G. von.: Zur Entwicklungsgeschichte der Anthe- microspore of Lilium. In: (ed. by) S. BLACKMORE and I. ren. Planta 3: 490-495 (1927). K. FERGUSON:Pollen and spores: form and function. VITHANAGE, H. I. M. V. and R. B. KNOX: Periodicity of Academic Press, London, 1986 (p. 1-17). pollen development and quantitative cytochemistry of EL-GHAZALY, G. and S. NILSSON : Development of tapetum exine and intine enzymes in the grasses Lolium perenne and orbicules of Catharanthus roseus (Apocynaceae). In: L. and Phalaris tuberosa L. Ann. Bot. 45: 131-141(1980). (ed. by) S. BLACKMORE and S. H. BARNES:Pollen and WILMS, H. J., H. B. LEFERINK-TEN KLOOSTER andA. C. spores: Patterns of diversification. Clarendon Press, VANAELST: Isolation of spinach sperms cells: 1. Ultras- Oxford, 1991 (p. 317-329). tructure and three dimensional construction in the FUKUDOME, H. and K. TANAKA: A method for observing mature pollen grain. In: (ed. by) D. L. MULCAHY, G. B, intracellular structures of free cells by scanning elec- MULCAHY and E. OTTAVIANO,: Biologytechnology and tron microscopy. J. Microsc. 141: 171-178 (1986). ecology of pollen. Springer, New York, 1986 (p. 307- HESSE, M.: Entwicklungsgeschichte and Ultrastruktur 312). von Pollenkitt and Exine bei nahe verwandten entomo- philen Angiospermensippen. Plant Syst. Evolut. 134: 229-267 (1980). LUGARDON, B.: Les globules des Filicinees, homologues des corps d'Ubisch des Spermatophytes. Pollen et Spores 23: 93-124 (1981). PACINI,E.: Tapetum and microspore function. In: (ed. by) S. BLACKMORE and R. B. KNOX:Microspores: evolution and ontogeny. Academic Press, London, 1990 (p. 213- 237). Dr. Susan H. BARNES PACINI,E. and G. G. FRANCHI:Diversification and evolu- The Natural History Museum tion of the tapetum. In: (ed. by) S. BLACKMORE and S. H. Cromwell Road BARNES:Pollen and spores: Patterns of diversification. London SW7 5BD Clarendon Press, Oxford, 1991 (p. 301-316). United Kingdom