Actin during pollen germination

The presence of numerous crystalline-fibrillar bodies in the cytoplasm of the vegetative cell of Nicotiana tabacum pollen has been described earlier (Cresti et al. 1986). These structures were shown to be composed of aggregates of fibrils with an individual width of 4–7 nm, and the electron-microscopic evidence was taken to indicate that during the activation of the pollen prior to germination they decline in number through the dispersal of the contents. The possibility was considered that the bodies might in fact be a storage form of F-actin, the dissociation of which produces clusters or aggregates of microfilaments that will later be involved in protoplasmic streaming in the pollen tube. Fibrillar elements have also been described from disrupted vegetative-cell protoplasts derived from mature, hydrated pollen of another dicotyledon, Helleborus foetidus (Heslop-Harrison et al. 1986), and again these have been tentatively identified as aggregates of actin microfilaments.

In the present paper we report that the fibrillar elements from the pollen of Helleborus and the corresponding structures in the pollens of other species have a strong affinity for phalloidin, a phallotoxin regarded as binding specifically to F-actin (Barak et al. 1980). Furthermore, we show that conspicuous fusiform phalloidin-binding bodies, probably homologous with the fibrillar bodies observed in Nicotiana pollen, are present in ungerminated pollen of Endymion non-scriptus, and that these are dispersed during germination, concomitantly with the appearance of numerous more attenuated phalloidin-binding fibrils.

The observations were made on the pollens of Helleborus foetidus L., Endymion non-scriptus (L.) Garcke (Scilla non-scripta (L.) Hoffmg. & Link; Hyacinthoides non-scripta (L.) Fabr.), Narcissus poeticus (L.) (cultivar), andHordeum bulbosum L. Inflorescences were brought into the laboratory as required, and fresh pollen collected directly from mature anthers.

Pollen for pre-hydration was dispersed on glass slides and retained for the required periods in boxes with an ambient relative humidity (RH) of 90–95 %. Pollen of E. non-scriptus was germinated in roller tubes at room temperature (18–20°C) in aqueous medium containing 10⁻³M-Ca(NC₃)₂, 0·5×10⁻³ M-H₃BO₃ and 15% sucrose (germination medium).

Coomassie Blue (0·1 % in 7 % acetic acid and 30 % methanol) and the fluorochrome, 1-anilino-naphthalene-8-sulphonic acid (1-ANS; approx. 0·0002% in the germination medium) were used as general stains for isolated cytoskeletal elements (Heslop-Harrison et al. 1974, 1986). Phalloidin labelled with FITC (Sigma) and with rhodamine (Molecular Probes Inc.) was used for actin identification. The FITC-labelled product was taken up in sterile 0·05 M-phosphate buffer at pH 6·8 and diluted with germination medium to give concentrations in the range 0·5×10⁻⁶ to 10⁻⁶M for use. Rhodamine-labelled phalloidin was received in methanolic solution; after the removal of the methanol, the product was dissolved in standard phosphate-buffered saline and diluted with germination medium to give a concentration judged by empirical observation to provide optimum coupling; because of the method of use, the final concentration could not be estimated.

Intact, ungerminated pollen and germinated pollen with tubes in the length range 100–250 pm was dispersed on microscope slides coated with poly-L-lysine and fixed for 10—20 min at room temperature in 4 % paraformaldehyde in germination medium. The fixative was then drawn off and the sample washed with the medium before infiltration with the phalloidin solution. In some instances unfixed samples were exposed for 10 min to 0-5 % Triton X-100 in germination medium with the object of facilitating the penetration of the coupling agent.

After the required periods the phalloidin solution was withdrawn, and the samples flooded with fresh medium for examination. Parallel preparations were made without treatment with the labelled phalloidin to check on the possibility of autofluorescence; in no case was this significant. Slight background and exine staining was observed in preparations treated with the FITC-labelled product, possibly attributable to the presence of traces of free dye. This did not obscure the specific staining of the putative actin.

For the observation of dispersed protoplast contents with or without prior fixation, pollen and pollen tubes were fractured in germination medium containing 15–25 % sucrose on poly-L-lysine-coated microscope slides by gentle pressure through a coverslip or another slide. After staining with the general protein stain or exposure to phalloidin-containing media as above, the extruded cytoplasm was carefully rinsed and mounted in fresh medium.

Observations were made with Vickers M17 and Zeiss microscopes with the appropriate filter combinations for 1-ANS, FITC and rhodamine fluorescence. Micrographs were made on Kodak Tri-X and Ilford FP4 film.

Conspicuous phalloidin-binding bodies were observed scattered throughout the cytoplasm of the vegetative cells of ungerminated pollen. These were consistently fusiform in shape, but very variable in size, as may be seen from comparison of Figs 1–3. The bodies were absent from fully hydrated and incipiently germinating pollen. In such grains, phalloidin binding tended to be more generalized, often associated with shorter, more attenuated filamentous inclusions, which sometimes appeared to form a reticulum (Fig. 4).

In germinating grains the fluorescence attributable to the binding of the labelled phalloidin was usually concentrated in the region of the emerging tube. Although it was not always possible to resolve structure at these sites, in occasional favourable preparations fluorescence was clearly seen to be associated with attenuated fibrillar inclusions converging towards the aperture (Fig. 5). In cytoplasm released from germinating pollen by gentle pressure this fibrillar constituent fanned out in a spectacular manner (Fig. 6).

In cytoplasm released into germination medium from mature, hydrated but ungerminated pollen of Helleborus, a tenuous fibrillar framework was revealed by 1-ANS staining, and elements of this formed distinctive Coomassie Blue-staining strands when drawn out (Heslop-Harrison et al. 1986). The provisional identification of this material as F-actin has now received support from the observation that these strands bind labelled phalloidin (Fig. 7). Fig. 8 illustrates a cable extracted from the pollen of Hordeum bulbosum, spread out by pressure and stained with Coomassie Blue; individual fibrils ranging in width down to the resolution limit of the light microscope have been revealed. These extended fibrils could not be obtained from pollen chemically fixed in glutaraldehyde or paraformaldehyde. The significance of this is considered further below.

Accepting that the binding of phalloidin is indeed specific for F-actin, the findings reported here seem adequately to confirm the earlier suggestion based on electron microscopy of Ni. tabacum pollen that F-actin is abundantly present in a storage form in ungerminated pollens. The profiles of the fibrillar bodies illustrated in the electron micrographs of mature, inactivated pollen of N. tabacum (figs 1–4 of Cresti et al. 1986) are precisely of the character one might have expected from the sectioning of pollen grains comparable with those of Figs 2 and 3 in the present paper, and the dimensions fall comfortably within the range observed in E. non-scriptus. Furthermore, it is apparent that the putative actin-storage bodies undergo quite rapid dissociation during pollen activation and germination, since they are no longer present when the tube emerges. In their place the entangled complex of more attenuated filaments appears, converging towards the germination aperture at the time when cytoplasm is leaving the grain and active cytoplasmic streaming has begun (Fig. 5).

The first unequivocal demonstration of actin in pollen tubes was given by Condeelis (1974), who showed that fibrils derived from disrupted pollen tubes and pollen-tube protoplasts from Amaryllis belladonna bound rabbit-muscle heavy meromyosin. Recently, excellent demonstrations of actin, identified by phalloidin binding, in the pollen tubes of Lilium longiflorum and Petunia hybrida have been given by Perdue & Parthasarathy (1985). Fibrillar material, sometimes tentatively identified as actin, has been noted in electron micrographs of pollen and pollen tubes prepared by standard procedures on various occasions (e.g. see Franke et al. 1972; Cresti et al. 1976; Miki-Hirosige & Nakamura, 1982). Collectively the evidence now strongly indicates that actin is an abundant and universal constituent of pollen, where it is stored, at least in some species, in relatively massive, quasi-crystalline bodies. The most likely function of pollen actin is, of course, that it forms one of the components of an actomyosin system concerned with cytoplasmic streaming in the pollen tube. Weighty support for this view comes from the fact that, while microtubule inhibitors do not affect streaming in the tube, the movement of organelles and other cytoplasmic components is rapidly arrested by cytochalasin B, a known actin inhibitor (Mascarenhas & Lafountain, 1972; Franke et al. 1972).

A vital aspect of pollen-tube function is the conveyance of the vegetative nucleus and the generative cell, or the gametes derived from it, towards the ovule. This is accomplished through sustained acropetal migration in the confines of an extending tube in which cytoplasmic streaming and organelle movement is bidirectional (Iwanami, 1959). While the means by which this movement is achieved remain obscure, we shall give evidence in a further paper that the surfaces of the vegetative nucleus and generative-cell nucleus are intimately associated with actin, suggesting a unique relationship with the motility system of the cytoplasm.

A final point concerns the origin of the extended, phalloidin-binding fibrils freely released into osmotically balancing medium from the unfixed, ungerminated mature pollen of all the species examined. Those from H. foetidus pollen (Fig. 7) are strikingly similar to the cytoplasmic fibrils ‘under tension’ illustrated from Amaryllis pollen protoplasts by Condeelis (1974). In the present study, no corresponding extended fibrillar structures could be detected within the vegetative cell of intact pollen in the same state of development, nor could the attenuated fibrils be obtained from chemically fixed grains. The possibility that the fibrils are preparation artefacts has been considered previously (Heslop-Harrison et al. 1986). A possible interpretation of the present observations is that the extended fibrils are actually derived by the rapid dissociation of the compact actin-storage bodies during the fracturing of the grain and the rupturing of the vegetative-cell protoplast. Consistent with this view is the fact that they are not released from fixed protoplasts, in which the storage bodies are likely to have been stabilized by the cross-linking of the constituent microfilaments.

The work was supported in part by the AFRC (UK) as part of the Cell Recognition and Signalling Programme, and by the Cytomorphology Group of CNR (Italy). We thank Dr P. K. Hepler for the gift of rhodamine-labelled phalloidin, and the Director of the Welsh Plant Breeding Station for facilities afforded J.H.H. and Y.H.H. during part of the investigation.