Freeze-etched pollen walls of Artemisia pycnocephala and Lilium humboldtii

Pollen walls have a unique ultrastructure consisting of 2 major wall layers: the exine composed of a highly polymerized lipid compound, sporopollenin (Zetzsche, 1932; Brooks & Shaw, 1968), and the intine composed of mixed polysaccharides (Bouveng, 1963; Martens & Waterkeyn, 1962). The exine consists of 2 layers: the ektexine, a complex structure often with radiating rods, and the endexine, a smooth layer with some lamellae. Although both layers are composed of sporopollenin (Erdtman, 1943), they differ in ontogeny (Echlin & Godwin, 1969; Rowley & Southworth, 1967; Heslop-Harrison, 1968a) and in ultraviolet absorption spectra (Southworth, 1969a, b). In order to study other differences and ultrastructural details of intine, ektexine and endexine, we have studied pollen walls by freeze-etching.

The freeze-etching procedure (Steere, 1957; Moor, Mühlethaler, Waldner & Frey-Wyssling, 1961) has been used to study a variety of plant cells. Cell walls in particular have been examined in onion root tips (Branton & Moor, 1964), white clover leaves (Hall, 1967), pea root tips (Northcote & Lewis, 1968), Ectocarpus acutus and Elachista fucicola (Bailey & Bisalputra, 1969), Chlorella (Staehelin, 1966; Branton & Southworth, 1967), Cyanidium (Staehelin, 1968), Oscillatoria (Jost, 1965), Sac charomyces (Moor & Mühlethaler, 1963; Branton & Southworth, 1967), Penicillium (Hess, Sassen & Remsen, 1966; Sassen, Remsen & Hess, 1967), and Bacillus (Remsen, 1966; Holt & Leadbetter, 1969). In the majority of electron micrographs contained in these reports, the cell walls are transversely fractured so that neither outer nor inner surfaces are exposed. The freeze-etched walls do show their fibrillar structure, however, and in some cases, e.g. Penicillium, Osdllatoria and Ectocarpus, wall layers differing in morphology and composition can be distinguished. Surface features not exposed by the fracture can be exposed by the etching process (Branton & Southworth, 1967).

In this study no fracture planes within the wall were found. However, granular and fibrillar ektexine structures not seen in standard thin sections were exposed by freeze etching, and fractures were observed at the intine-plasmalemma junction and at the outer ektexine surface.

Fresh pollen of Artemisia pycnocephala DC. (UCBG 60.975) and Lilium humboldtii Roezl & Leichtl. (UCBG 54.667) was suspended in one or more of the following liquids: distilled water; 5%, 10% and 20% glycerol; mineral oil (Squibb, extra heavy); then centrifuged, frozen in Freon-22 on a copper disk and stored in liquid nitrogen. The specimen was fractured at — 100 °C, etched for 0, 1, or 2 min and replicated with platinum and carbon in a Balzers Freeze Etch Device (Moor et al. 1961; Moor & Mühlethaler, 1963). Replicas were hard to clean. The best results were obtained by floating on the following solutions: half-strength hypochlorite bleach (Purex) briefly, full strength bleach 1 h, acidified bleach (Purex and HC1) 2 changes 1 h each, water rinse, 70 % H2SO4 1 h and distilled water 3 changes for a total of several hours. Replicas of material in mineral oil were further rinsed in xylene on the grids. Replicas were picked up on Formvar-coated grids and examined with a Siemens Elmiskop 1 A. Shadows appear white on all freeze-etch electron micrographs and the shadowing direction is indicated by an encircled arrow on each figure.

Pollen for thin sections was fixed for 2 h at room temperature in formaldehyde freshly prepared from paraformaldehyde (Pease, 1964), washed with buffer, and postfixed with 2% OsO4. The loose pollen was centrifuged and mixed with melted 2 % agar and allowed to solidify. The pollen-agar blocks were dehydrated with acetone and embedded in Epon. A second Lilium preparation was made omitting the osmium postfixation. Sections from this material were post stained with 2 % BaMnO4.

In Artemisia pycnocephala the bacula in the ektexine did not direct the fracture in any preferred plane, but wete simply cross-fractured to produce flat uniform surfaces (b in Fig. 3). In preparations frozen in water additional surfaces were exposed by the etching process (Figs. 2, 3) which removes water by sublimation (Branton & Southworth, 1967). Bacula surfaces exposed by etching were smooth (Fig. 3). Both fracture- and etch-faces of the ektexine appeared to be composed of an unstructured homogeneous material in marked contrast to the endexine and intine.

Fractures across the endexine produced a rough surface different from that of the ektexine (Figs. 5–7). The endexine lamellae appeared as a double row of particles (double arrows, Fig. 6) with a central depression between the rows or as a single row of particles (single arrows). The particles measured 15–20 × 10 nm. In some regions particle spacing was quite regular, but in others it was not (Fig. 6). In regions of the endexine in which the lamellae are quite compressed, the depression between 2 rows of particles and the regular spacing of particles were lost (Figs. 7, 9). However, the endexine always retained its rough particulate appearance and never formed the smooth homogeneous appearance of the ektexine (Fig. 3). Thin sections of Artemisia showed no globular subunits in the endexine but did show an osmium staining discontinuity which was central in some regions of the lamellae (Fig. 4, inset).

In Lilium humboldtii the ektexine structures are not homogeneous as in Artemisia. The ektexine fractured along radiating fibrils in cross-fracture and over protuberances arranged in a striated pattern in tangential fractures near the surface (Figs. 12–15). The etch-face appeared smooth as in Artemisia (Figs. 12, 13). The radiating structures appeared as interconnected units the ends of which may produce the striated pattern seen in the tangential fractures (Fig. 13). Both the convex face of the bacula (Fig. 16) and the concave face of the shells remaining when the bacula have been fractured away (Fig. 17) showed the striated pattern with a periodicity of 10 nm. The baculum face appeared rougher than the concave shell face. Exine fracture planes were smeared by knife marks when pollen was suspended in mineral oil.

In thin sections of Lilium fixed with formaldehyde-osmium and post-stained with lead, the outer surface of the ektexine stained slightly less densely than the core (Fig. 10). In thin sections fixed with formaldehyde only and post-stained with permanganate the surface region stained darker than the core (Fig. 11). These staining differences with osmium-lead and manganese indicate that the surface of the ektexine has different chemical properties from the core.

The endexine of Lilium is thin (Fig. 10) and was not detected in the freeze-etch replicas.

The intine of Artemisia most frequently fractured transversely producing a coarsely pitted surface on a fairly homogeneous background (Fig. 8). In the outer part of the intine were scattered fibrils at least 300 nm long (Fig. 9). These did not occur deeper in the intine. The intine of Lilium showed no fibrils and was generally smoother than the intine of Artemisia (Fig. 19).

The plasmalemma fracture face had numerous particles 8–14 nm in diameter which were irregularly spaced (Figs. 19, 22). In preparations of Artemisia, especially those freeze-etched in water or 5% glycerol, there were also slightly smaller particles (7·5 nm) which were arranged in regular arrays with a centre-to-centre spacing of 22 nm (Fig. 21). In Lilium a less particulate regular pattern also appeared (Fig. 20). The composition and function of the plasmalemma particles are not known.

Vesicles were seen adjacent and attached to the plasmalemma of Artemisia (Figs. 2, 8, 22). The vesicle faces were smooth in contrast to the particulate faces of the plasma-lemma (Fig. 22). Some of the vesicles appeared to lie outside the plasmalemma in the intine (Figs. 2, 22).

In Lilium the vesicles appeared to connect with the plasmalemma (Figs. 18, 19), but in no case were they observed outside it. As in Artemisia, the fracture face of the vesicles near the plasmalemma was smoother than the plasmalemma face. The vesicles in Lilium appeared to be produced by dictyosomes located near the plasma-lemma (Fig. 18). No recognizable dictyosomes were observed in Artemisia.

The ektexine of Artemisia appears as a homogeneous material without fracture planes or other substructure visible by freeze-etching. On the other hand the ektexine of Lilium shows a fibrillar substructure. However, fibrils are not visible in standard thin sections of mature exines of either Lilium or Artemisia after osmium or per manganate staining (Rosen, Gawlik, Dashek & Siegesmund, 1964; Heslop-Harrison, 1968a; Skvarla & Larson, 1965). Fibrillar exines have been observed, but only in young stages of wall formation (Rowley, 1962; Dickinson & Heslop-Harrison, 1968). All pollen used in this study had been discharged by the mature anther. Fibrils may be visualized after freeze-etching of mature pollen because this technique depends on subtle differences in long-range molecular order and tensile strength while the thin sectioned image depends primarily on the location of chemically reactive sites.

Near the apertures the endexine of Artemisia contains osmiophilic lamellae. The lamellae average 50 nm in thickness and have a central electron-transparent ‘white line’ of 5 nm. They have been observed in many other genera and form lamellae or tapes rather than tubes (Rowley & Southworth, 1967; Southworth, 1966; Godwin, Echlin & Chapman, 1967). The composition of the ‘white lines’ is unknown. In some cases they seem to originate from the cytoplasmic membrane (Rowley & Dunbar, 1967; Echlin & Godwin, 1969), but nothing in the fracture of mature endexine indicates structural similarity to membranes. However, the observation of a central depression in the freeze-etched endexine lamellae represents the first time the ‘white line’ has been observed by a method other than positive staining and confirms that it is a real discontinuity. The width of the freeze-etched lamellae with 2 rows of particles is 20 nm whereas the width of the osmium-stained lamellae averages about 50 nm. Hence the 2 do not fully correspond. Some of the osmiophilic material may not be visible in the freeze-etch image.

No fractures were observed between ektexine and endexine or between endexine and intine, suggesting that these layers are held together by relatively strong bonds, e.g. covalent or electrostatic, not easily fractured during freeze-etching.

The intine of Lilium humboldtii shows no evidence of microfibrillar organization. Heslop-Harrison (1968b) reported a microfibrillar organization in the intine of Lilium henryi. However, microfibrils were not revealed by our methods in either thin sections or freeze-etched preparations of Lilium humboldtii. In Artemisia elongate fibrils were

Freeze-etched pollen walls 197

seen by freeze-etching but they do not appear to be fibres such as cellulose and collagen which are characteristically pulled out of the matrix by the fracture process (Northcote & Lewis, 1968; Clark & Branton, 1969).

Vesicles and particles associated with the plasmalemma are frequently reported in freeze-etch preparations. Moor (1967) observed vesicles apparently fusing with the plasmalemma in yeast in regions where the wall was extending to form a bud. In this case he postulated that the vesicles contained wall-softening enzymes. Staehelin (1966) observed vesicles near the wall of Chlorella. These vesicles seemed to pass through the plasmalemma and to discharge their contents between plasmalemma and wall. In Lilium the vesicles appear to be fusing with the plasmalemma. In Artemisia vesicles appear to fuse with the plasmalemma and to pass through the plasmalemma into the wall region. This may represent 2 functions for the vesicles: the continued deposition of intine wall materials and the discharge of wall-softening enzymes prior to pollen tube formation.

Dry pollen grains of Lilium can be suspended in both water and mineral oil, indi cating that the exine surface may be either hydrophilic or hydrophobic. The pollen grains suspend immediately in mineral oil. A yellow-orange substance from the pollen dissolves in the oil and remains suspended when the pollen grains settle out. When dry pollen is placed in water or 20% glycerol the pollen tends to clump around air bubbles or at the surface. After a few minutes, however, the grains disperse throughout the medium and eventually settle out. No coloured substances are dissolved in the water. McDonald (1964) examined wettability properties ol pollen of 15 species. He found all of them to be hydrophilic rather than hydrophobic, but some reacted slowly to water.

An explanation for these observations is proposed in Fig. 1. It explains the bacula surface fracture by proposing that a sheet of surface-active material covers the bacula of pollen in aqueous suspension. If this material had a hydrophilic outer surface interacting with the aqueous medium and a hydrophobic inner surface interacting with the partly hydrophobic fibrillar baculum, it could produce an easily fractured plane of weak bonding (Fig. 1 A). The location of fracture planes in the frozen state has been attributed to the presence of weak hydrophobic bonds (Branton & Southworth, 1967). The existence of a fracture plane over the surface of the bacula could be correlated with the differentiation in osmium-lead and permanganate staining near the bacula surface (Figs. 10, 11). The differentially stained surface could represent the surface active material.

In mineral oil, in which no surface fracture occurred, the surface-active material would be dissolved away (Fig. 1 c). The pigmented oily substance has not been analysed in Lilium humboldtii. Carotenoids associated with the exine of several other species of Lilium include β-carotene, a-carotene-5,6-epoxide, and a number of xanthophylls (Linskens, 1967). These may be dissolved in other fatty compounds. There may in addition be other unpigmented carotenoids as well as proteins that form a part of the pollen-coat substances which are in the interstices of the exine as well as on the surface (Heslop-Harrison, 1968c, d).

The actual state of the dry pollen surface after discharge might differ from that in water or oil. The surface would have both hydrophobic and hydrophilic regions with associated lipid droplets (Fig. IB). In water the lipid droplets would flow to cover the surface hydrophobic regions with the hydrophobic parts of the droplet lipids (Fig. 1 A). The hydrophilic ends of the droplet lipids would enable the pollen to disperse in water. In oil the lipid droplets would be detached from the pollen surface (Fig. 1 c).

Research was supported by the following grants: National Science Foundation Grant GB-13646 (D. Branton) and National Institutes of Health Grant GM-13943 (R. B. Park), and Predoctoral Fellowship 1-F1-GM-36, 330–01 (D. Southworth) from the National Institute of General Medical Sciences.

All figure descriptions list the pollen source (Artemisia pycnocephala or Lilium humboldtii), the suspending medium or fixative, the etching time and the magnification.

Figs. 2–7. Artemisia pycnocephala.

Figs. 8, 9. Artemisia pycnocepiiala.

Figs. 10, 11. Lilium humboldtii.

Figs. 12–17. Lilium humboldtii.

Figs. 18–20. Liliuvi humboldtii.

Figs. 21, 22. Artemisia pycnocephala.