Diffusion Barriers of Tripartite Sporopollenin Microcapsules Prepared from Pine Pollen

Abstract

Tripartite sporopollenin microcapsules prepared from pine pollen (Pinus sylvestris L. and Pinus nigra Arnold) were analysed with respect to the permeability of the different strata of the exine which surround the gametophyte and form the sacci. The sexine at the surface of the sacci is highly permeable for polymer molecules and latex particles with a diameter of up to 200 nm, whereas the nexine covering the gametophyte is impermeable for dextran molecules, with a Stokes’ radius ≥4 nm (Dextran T 70), and for the tetravalent anionic dye Evans Blue (Stokes’ radius = 1·3 nm). The central capsules obtained by dissolution of the sporoplasts showed strictly membrane‐controlled exchange of non‐electrolytes, with half‐equilibration times in the range of minutes (monosaccharides, oligosaccharides) to hours (dextran molecules with Stokes’ radii up to 2·5 nm). The dependence of the permeability coefficients of the nexine for non‐electrolytes on Stokes’ radius or molecular weight shows that the aqueous pores through the nexine are inhomogeneous with respect to their size, and that most pores are too narrow for free diffusion of sugar molecules. To explain the barrier function of the nexine for Evans Blue, it is assumed that at least the larger pores, which enable slow permeation of dextran molecules, contain negative charges.

INTRODUCTION

Moss and fern spores as well as most pollen grains have an envelope (sporoderm, exine) consisting of sporopollenin (Zetsche and Vicari, 1931), a biopolymer that is highly resistant to enzymatic breakdown and hydrolytic decomposition in strongly acid or alkaline media. Sporopollenin cannot be considered a uniform macromolecule. Infra‐red spectroscopy and ¹³C NMR spectroscopy of sporopollenin derived from pteridophyta and spermatophyta demonstrate the presence of aliphatic, aromatic, hydroxyl, carbonyl/carboxyl and ether functions in various proportions (Wilmesmeier et al., 1993). Sporopollenin is produced largely from acyl lipid and phenylpropanoid precursors (Wiermann and Gubatz, 1992; Piffanelli et al., 1998; Ahlers et al., 2000). The mechanisms of its synthesis and of its consolidation are not yet understood. With respect to its composition and precursors, sporopollenin and suberin show similarities (compare Bernards, 2002), and the biosynthetic capacity of embryophyta to produce suberin and sporopollenin might have a common origin (Kroken et al., 1996). However, in contrast to the suberin lamellae, sporodermata and exines are generally situated outside of the cellulose/pectin wall (intine). Exines can be partly or completely solubilized in hot aminoethanol (Bailey, 1960; Southworth, 1974), and the dissolved polymers re‐aggregate spontaneously when the solvent is exchanged against water (Jungfermann et al., 1997; Thom et al., 1998). After partial degradation of the exine by different methods, including oxidation and aminoethanol treatment, regular substructures of 100–200 nm diameter appear in scanning and transmission electron microscopy (SEM and TEM) images of exines and sporodermata (Southworth, 1986; Rowley, 1996). There is still uncertainty regarding the detailed chemical structure of sporopollenin as well as regarding the conformation and cross‐linking of this polymer.

The multilayered sporopollenin coatings of spores and pollen grains show a fascinating diversity of structural details and an often highly symmetrical pattern of sculptures and (if present) pores or colpi. The shape, pore arrangement, morphology and ultrastructure of sporodermata and exines have proved informative to palaeobotanists and in systematics (e.g. Straka, 1975; Blackmore, 1990; Hesse, 1991; Moore et al., 1991). However, with the exception of UV‐shielding (Rozema et al., 2001), the essential functions of sporopollenin coats are not yet well understood. There are severe deficits in our knowledge on the physical properties of sporopollenin, which might be significant for physiological processes. It is likely that these properties are related to the dispersal of spores and pollen grains, which requires their transient exposure to the dry atmosphere. However, the view that ‘one of the main protective functions’ of the sporoderm and exine ‘must be against excessive desiccation’ (Heslop‐Harrison, 1973), is not convincing, since desiccation and rehydration of spores and pollen are usually rapid processes and, in most cases, the poikilohydric sporopolasts tolerate severe desiccation (e.g. Linskens, 1964).

Data on the permeability of sporodermata and aporate exines for solutes of different size are informative with respect to the dimensions of the water‐filled channels through the polymer matrix. Such data might be relevant for a better understanding of the physiological role of the sporopollenin membranes and their eventual technical application, too. Advantages of sporopollenin packing particles and microcapsules purified from pollen or spores are their high resistance to extreme pH values, their relatively uniform size and their mechanical stability. The relatively high content of hydrophilic groups in sporopollenin may be used to anchor primers for the solid‐phase synthesis of biopolymers (Mackenzie and Shaw, 1980; Adamson et al., 1983) or the immobilization of ligands for chromatographic purposes (Pehlivan and Yildiz, 1994; Ersöz et al., 1995; Vural et al., 1995; Çengeloglu et al., 1998).

Pine exines are available in large amounts and show excellent mechanical rigidity and filtration properties for fixed bed applications (Woehlecke et al., 2002). A detailed description of the ultrastructure and ontogenesis of the pine exine has been given by Rowley et al. (1999, 2000a, b). In this study, we describe the exchange of sugars, low‐molecular‐weight dyes, dextran molecules of different size and submicrometer latex particles with tripartite microcapsules obtained from pine pollen by solubilization of the sporoplasts.

MATERIALS AND METHODS

Dyes

The following dyes were used: carboxyfluorescein (Mole cular Probes Inc., Eugene, OR, USA); Evans Blue (Reanal, Budapest, Hungary); FITC (fluorescein isocyanate)‐Dextran (FD‐250S, M_R = 282 000 g mol^–1; Sigma‐Aldrich Chemie GmbH, Steinheim, Germany); FITC‐labelled latex beads (Fluoresbrite Plain YG 0·2 micron microspheres, Ø = 0·217 ± 0·015 µm; Polyscience Europe, Eppelheim, Germany) and polyacrylic acid 2100 sodium salt (Fluka Chemie AG, Buchs, Switzerland)

Pine pollen and sporopollenin capsules

Protoplasts were removed from mature pine pollen by the following preparation steps: (1) lipid extraction, (2) treatment with hot diluted acid, (3) treatments with hydrolytic enzymes and (4) washings. Pollen grains were denatured using 50 % ethanol, and lipids were extracted using hot ethanol (60 °C) and dioxan (80 °C). Particles were then incubated in a closed vessel containing diluted sulfuric acid (30 g l^–1) and sodium sulfite (3 g l^–1) for 48 h at 90 °C. The subsequent treatment with cell‐wall lysing enzymes was carried out as described by Jungfermann et al. (1997). Particles were bathed at 37 °C for 1 d in a phosphate buffer (pH 6·5) containing 2 g l^–1 of a bovine pancreas enzyme mixture supplied as ‘trypsin for cell culture’ by Bernd Belger (Klein‐Machnow, Germany). They were subsequently washed on a glass filter in a sodium chloride solution (10 g l^–1) and then incubated for 1 d at room temperature in a large volume of a sodium carbonate solution (10 g l^–1) containing sodium dodecyl sulfate (1 g l^–1) and sodium dithionite (1 g l^–1). Finally, particles were washed using a sodium chloride solution (10 g l^–1) and water until the extract was free of UV‐absorbing substances (280 nm). The sporopollenin microcapsules obtained were stored as a suspension in 50 % ethanol.

Measuring sugar and dextran efflux

A previously described system enabling the continuous polarimetric analysis of the particle‐free medium (Ehwald et al., 1973; Fleischer and Ehwald, 1995) was used to record the efflux of sugars and Dextran T 70 (Pharmacia, Uppsala, Sweden) from sporopollenin capsules that had been presaturated with the sugar or dextran solution. A flow of water‐saturated air (approx. 40 ml min^–1) was used to agitate the suspension and to drive circulation of the particle‐free liquid through a 2 ml bypass including the polarimeter cell. The whole volume in the system was 8–9 ml, and the flow through the detecting bypass was 20–30 ml min^–1. Using the recording polarimeter 141 M (Perkin Elmer, Ueberlingen, Germany) and an analogue/digital converter, the angle of rotation was registered at a wavelength of 366 nm with an absolute accuracy of ± 0·001°.

Sporopollenin capsules were saturated with the respective sugar or dextran solution for 3 d, vacuum‐filtered on a nylon sieve (20 µm pore size) for 30 s, and stored in a closed vessel. Sampling was carried out by filling an acrylic‐glass tube (diameter 8 mm) with the filtered mass, which was then weighed on an electronic balance. To start an efflux course, the mass was pushed into the measuring system by means of a fitting cylinder.

Partitioning of Dextran T 70

Sporopollenin capsules were saturated with a standard buffer solution (10 mm sodium phosphate, pH 7, 100 mm NaCl, and 1 g l^–1 NaN₃) and collected on a nylon filter as described above. One‐gram samples of the filtered mass were mixed with 1 ml standard buffer containing α‐methylglucoside (16 g l^–1) and Dextran T 70 (32 g l^–1). After the respective diffusion time, sporopollenin capsules were sedimented by centrifugation. The original solution and the supernatant were chromatographed on a Superdex 75 HR column (30 × 1 cm; Pharmacia) with a polarimetric detector (Chiralyser; IBZ Messtechnik, Hanover, Germany) to determine the ratio between sugar and dextran concentrations from the peak areas. Since the sugar partition space in the filtered mass is 0·94 ml g^–1, the partition space of dextran in 1 g of the filtered mass of sporopollenin capsules is equal to (1.94   – 1) ml, where r is the sugar : dextran ratio obtained in the supernatant and r′ is the sugar : dextran ratio of the original solution.

Exchange of the smaller dextran size fractions

Dextran exchange with the central capsule was analysed using a previously described method based on size exclusion chromatography of a polydispersed dextran‐probing solution (DPS) in combination with the diffusion experiment (Woehlecke and Ehwald, 1995). The composition of the DPS used in this study is given in Table 1. Sporopollenin capsules were shaken in DPS for 10 d, collected by vacuum‐filtration on a 20 µm nylon sieve, and drained for 30 s. Samples of the filtered mass (150 mg) were shaken with 450 µl standard buffer (see above) for different periods. Subsequently, the liquid was separated from the capsules by centrifugation and fractionated by SEC on a size‐calibrated Superdex 75HR (30/1) column (Pharmacia). The elugraphs were compared with that of the original DPS using a computer program, as described previously (Woehlecke and Ehwald, 1995; Dautzenberg et al., 1999).

Table 1.

Composition of the dextran‐probing solution (DPS)

Component	Manufacturer	Stokes’ radius ofthe most concentratedsize fraction * (nm)	Concentration in theprobing solution (mg ml–1)
Dextran T 70	Pharmacia (Uppsala, Sweden)	4·56	0·900
Dextran T 20	Pharmacia (Uppsala, Sweden)	3·99	0·975
Dextran 15	Serva (Heidelberg,Germany)	3·61	1·125
Dextran 8	Serva (Heidelberg,Germany)	2·46	3·000
Dextran 4	Serva (Heidelberg,Germany)	1·99	0·750
Dextran 4, hydrolysed for 24 hours†		1·50	0·750
Dextran 4, hydrolysed for 48 h†		1·24	1·050
α‐Methyl‐glucoside	Fluka, Buchs, Switzerland	0·38	0·525

* Obtained by SEC on a calibrated Superdex‐75‐R‐column at the peak maximum.

^† 0·06 m sulfuric acid, 80 °C.

RESULTS AND DISCUSSION

Structure and size of the sporopollenin microcapsules

At the site of a saccus, two outer strata of the ectexine (tectum and columellae) are separated from two inner layers (endexine and foot layer of the ectexine), as illustrated schematically in Fig. 1. The external membrane of the sacci represents the part of the exine termed the sexine. The two inner layers covering the lumen of the central capsule have been termed nexine (Straka, 1975; Rowley et al., 2000b). SEM images of the purified microcapsules are shown in Fig. 2. The external wall of the sacci consists of a thin sporopollenin membrane (tectum), which covers a thick but macroporous layer formed by honeycomb‐shaped support structures (columellae), which are open to the lumen (Fig. 2A and B). The tectum is perforated by scattered pores of submicrometer size (Fig. 2C). The lumen of the central capsule is bordered by a continuous dense sporopollenin envelope. Apart from the sacci, the two layers representing the sexine are present on the surface of the central capsule (Fig. 2D).

Fig. 1. Scheme of the sporopollenin strata of the pine exine. The two outer layers of the ectexine are detached from the foot layer, thus forming a saccus.

Fig. 2. Scanning electron micrographs of dehydrated capsules and cryo‐sections of a pine exine (Pinus sylvestris). A, Residue of the central capsule envelope with lateral sacci. B, Close‐up of a saccus, showing the surface layer and the honeycomb‐shaped support structures. C, Image of the surface of the sexine covering a saccus. D, Close‐up of a central capsule, showing the dense nexine at the luminal face and the alveolar sexine at the external face. Capsules or 20‐µm‐thick cryo‐slices were air‐dried from tetramethyl‐silane after dehydration in ethanol and sputter‐coated with 10 nm gold. Imaged using a Leica S360 scanning electron microscope (Leica, Cambridge, UK).

Since the ratio of volume (V) to surface area (A) of the central capsule was required for the calculation of permeability coefficients, the approximately ellipsoid central capsules were analysed morphometrically using a slide calliper and printed light microscope images of dense suspensions of the sporopollenin microcapsules in water. The two short perimeters were obtained either from the top view or from the lateral view (Table 2).

Table 2.

The ratio between volume V and surface area A of the central capsules as determined from morphometric estimates of their perimeters

2

* Introduced species.

Mean values with standard deviation and number of measured perimeters (n).

Permeation of dyes, fluorescent polymers and particles

The diffusion of dyes, fluorescently labelled dextran, and latex particles into the central capsule and the lateral sacci was studied by fluorescence and transmission light microscopy (Fig. 3). Carboxyfluorescein was excluded from the empty central capsules for the first minutes (not shown) but labelled the central capsule lumen within some hours (Fig. 3A). Evans Blue, a water‐soluble and protein‐staining dye containing four sulfonic acid residues (960 Da, Stokes’ radius = 1·3 nm) did not enter the empty central capsule or the denatured pollen grain within 24 h, but accumulated rapidly in the denatured sporoplasts when the exine was injured (Fig. 3B). FITC‐labelled serum albumin (image not shown) and FITC‐Dextran with a mean molecular weight of 282 kDa (Fig. 3C) did not permeate into the empty central capsule. The fluorescing dextran and even fluorescing latex beads of a preparation with a mean diameter of 218 nm entered the lumen of sacci rapidly (Fig. 3D), whereas 1 µm beads did not (image not shown).

Fig. 3. Partitioning of low‐molecular‐weight dyes, stained polymers and particles within the tripartite sporopollenin microcapsules, and denatured pollen grains (Pinus sylvestris). Images of the capsules or pollen grains in the fluorescent or stained medium were produced using Leica CLSM (Leica Laser‐Technik GmbH, Heidelberg, Germany); the wavelength of excitation was 488 nm, and of emission approx. 535 ± 15 nm. A, Carboxyfluorescein, 0·01 g l^–1, 3 h, CLSM‐image. B, Evans Blue, 10 g l^–1, 3 h, transmission image. C, FITC‐Dextran, mean molecular weight 282 kDa, 0·4 g l^–1, 3 h, CLSM image. D, FITC‐latex particles, nominal size 0·2 µm, 0·26 mg l^–1, 15 min, CLSM‐image. Capsules (A, C, D) or denatured pollen grains (B) were washed with de‐ionized water and dispersed in the stained solutions.

Access of solutes to the central capsule is obviously controlled by the nexine, since the sexine is a highly permeable microfilter, at least where it builds the sacci. The rapid equilibration of large polymer molecules and even latex beads with the sacci requires comment, since SEM images of the sexine (Fig. 2C) suggest a relatively low density of submicrometer pores in the surface layer of the sexine. In this respect, it has to be stressed that the sporopollenin membrane is relatively thin at the surface compared with the diameter of the sacci. Owing to the diffusion resistance of the unstirred liquid within the capsules, a small fraction of pores in the membrane area will be sufficient to prevent membrane control of diffusional exchange, provided that the pores are distributed over the whole surface.

A relatively high maximum pore size and the mechanically rigid structure of the sexine (Fig. 2B) might be a necessary presupposition for the entrance of air into the sacci at desiccation, since purified sporopollenin has an hydrophilic surface with a contact angle close to zero for water (not shown here). When desiccation of the water‐saturated cell wall capsules was observed under the light microscope, no shrinkage of the sacci was found before air entrance, whereas central capsules collapsed completely. Since the product of r (pore radius) and Δp (the pressure difference required to remove water from a hydrophilic pore) is known to be 0·144 µm MPa, the critical cohesive tension for the entrance of air into the sacci should be about 1 MPa for a maximum pore size of 200–300 nm.

Efflux of Dextran T 70 and sugars from pre‐equilibrated sporopollenin microcapsules

The efflux of Dextran T 70 from the filtered mass of the sporopollenin capsules was too rapid to be resolved kinetically using the applied method (Fig. 4). The final value was obtained within the 30 s that were needed for convective mixing in the air‐lift system. The corresponding dextran partition space was 0·6–0·7 ml g^–1 of the filtered mass. Dextran T 70 did not equilibrate with the whole liquid space of the filtered mass of sporopollenin capsules within 30 d (Table 3).

Fig. 4. Efflux of Dextran T 70 from liquid‐saturated sporopollenin capsules into water. At time zero, 304 mg filtered sporopollenin capsules (Pinus sylvestris) that had previously been equilibrated in a solution of Dextran T 70 (30 g l^–1) were added to 8 ml water in the measuring system. Arrow: calibration with 200 µl of the dextran solution after a shift of registration. In this experiment the partition space of the dextran molecules in the filtered mass was 0·66 ml g^–1.

Table 3.

Partition space of Dextran T 70 in the filtered mass of sporopollenin capsules (Pinus nigra) in a long‐term experiment

Diffusion time (h)	24	72	240	720
Dextran partition space (ml g–1 fresh weight)	0·58	0·65	0·62	0·68

Sugars equilibrated with approximately the whole liquid volume in the filtered mass (0·94 ml g^–1), and leaked with a biphasic kinetics (Fig. 5). The first rapid phase (efflux from the sacci and interparticle space) could not be resolved, and the slower phase (efflux from the central capsule) showed first order kinetics. Extrapolation of the semi‐logarithmic plot (inset of Fig. 5) to zero time allowed the separation of the slowly exchanging volume, represented by the central capsules (approx. 0·3 ml g^–1 of the filtered mass), from the rapidly exchanging volumes, represented by the sacci and interparticle spaces (0·6–0·7 ml g^–1). The half‐life times of sugar exchange with the central capsule are three orders of magnitude larger than half‐life times expected for their diffusional exchange with a spherical equivalent volume of unstirred water (Table 4).

Fig. 5. Efflux of raffinose from liquid‐saturated exine capsules (Pinus sylvestris) into water. At time zero, 215 mg sporopollenin capsules filtered from a 150 mm raffinose solution were added to 8 ml of water in the measuring system. Arrow, Calibration with 0·2 ml of the 0·15 m raffinose solution. The partition space of raffinose in the filtered mass was 0·94 ml g^–1. Inset: Plot of the logarithm of the difference between the final and actual angles of rotation on diffusion time. In this experiment the volume of the slowly exchanging compartment (central capsule) derived by extrapolation of the first‐order line to time zero was 0·28 ml g^–1.

Table 4.

Sugar efflux from the preloaded central capsule

Species	Sugar	Partition space,Vp(ml g–1)*	Rate coefficient,k (s–1)*	Permeability coefficient,P (µm s–1)*	P × Mr0·5(nm s–1 g0·5 mol0·5)	Half‐life timemeasured τ (s)*	Half‐life timetheor.τ′ (ms)†
Pinus sylvestris	α‐Methyl‐glucoside	0·29 ± 0·02	0·0056 ± 0·0021	0·028 ± 0·0104	393	138	10·3
	Raffinose	0·30 ± 0·01	0·0021 ± 0·0001	0·010 ± 0·0004	231	338	15·9
Pinus nigra	α‐Methyl‐glucoside	0·34 ± 0·01	0·0112 ± 0·0021	0·062 ± 0·0116	870	63	12·8

* Obtained by extrapolation of the first‐order line (Fig. 5) to zero time.

^† Calculated for diffusional exchange with a spherical unstirred water zone of the central capsules size. M_r, Sugar molecular weight.

The strict membrane control and low permeability of the sporopollenin membrane for sugars found here contrasts with our previous findings for primary plant cell walls. Exchange of salts, sugars and even small colloids (Stokes’ radius up to 2 nm) with the empty plant cell wall capsules (obtained from a plant cell suspension culture) is rapid enough for efficient permeation chromatography (Ehwald et al., 1992), which requires equilibration within less than 1 s. Although the nexine is a significant diffusion barrier for sugars, its permeability coefficient for a monosaccharide (Table 4) is high enough to enable a rate of sugar uptake by the sporoplasts that satisfies the metabolic need. The permeability coefficients determined for α‐methylglucoside (Table 4) and the volume : surface area ratios of the central capsule (Table 2) were used to calculate the stationary concentration difference across the exine at an assumed high rate of monosaccharide uptake (50 µmol per g fresh weight and hour), which is twice the value found for maize root tips at a glucose concentration of 200 mm (Ehwald et al., 1974). The values obtained are in the millimolar range (see the Appendix). On the other hand, the limitation of solute diffusion by the exine found here is strong enough to reduce hypo‐osmotic shock and solute leakage when the air‐dry pollen grain is rehydrated. Furthermore, the diffusion resistance of the sporopollenin membrane can facilitate the maintenance of a specific pH and ion milieu close to the plasma membrane of the male gametophyte. In its functional role as a relative, not absolute, diffusion barrier for small solutes, the pine nexine resembles suberized and/or cutinized cell walls and Casparian strips (Brauner, 1956; Zimmermann et al., 2000; Hose et al., 2001). The permeability of the nexine for water, alcohols, salts and ampholytes, as well as its barrier function for cation and anion exchange, needs further study.

Exchange of smaller dextran size fractions with the central capsule

Compared with plant cell wall capsules, which reach a constant partition spectrum of the polydisperse dextran‐probing solution within some minutes (Woehlecke and Ehwald, 1995), the central capsule showed an extremely slow exchange of the polymers. When sporopollenin capsules were equilibrated with the polydisperse DPS and then transferred to the buffer for fixed diffusion periods, the exchange ratio (q) and the exchange rate (γ) of permeable size fractions were obtained by the SEC method, as explained in Fig. 6. Dextran molecules with Stokes’ radii between 1 and 2·5 nm showed intermediate values of these parameters after some hours of diffusion (Table 4) but reached the equilibrium values after 3 d (curves not shown).

Fig. 6. Size‐dependence of the exchange quotients of dextran size fractions with the central capsule at two different efflux times. Curves were obtained using the preparation obtained from P. sylvestris pollen. The exchange quotient q represents the quotient between the concentration of a dextran size fraction in the original DPS and the concentration of the same dextran size fraction in the medium of the capsules. The Stokes’ radii and the respective concentration values were obtained by size exclusion chromatography on a column calibrated with protein standards. q′, Maximum value (impermeable fractions) obtained from the peaks of Dextran T 70 at the void volume of the column; q′′, minimum value (completely equilibrated fractions), obtained from the α‐methylglucoside peaks. The exchange rate represents the fractional equilibration between the central capsule and the medium. The three given levels of γ were used for the calculation of the permeability coefficients and corresponding Stokes’ radii (Table 5).

Size dependence of permeability coefficients

The Stokes’ radii related to certain values of the exchange rate γ (0·33, 0·5 and 0·66) were read out from graphs of the type shown in Fig. 6, and the permeability coefficients of the central capsule envelope for these size fractions were calculated as explained in the Appendix. Permeability coefficients (P) were obtained using the three values of γ fitted to the same curve (Fig. 7). Hence, in the size range studied, exchange kinetics follows the first‐order model. An inhomogeneity of the capsules with respect to the permeability coefficients would lead to significantly higher values of P with the lowest value of γ, but this is not evident. Sporopollenin capsules prepared from Pinus sylvestris showed a higher permeability for the larger dextran size fractions and a lower permeability for the monosaccharide than those prepared from Pinus nigra. Although this difference was reproducible in the three parallel batches analysed, it has to be stressed that all parallels were from only one preparation each.

Fig. 7. Dependence of the permeability coefficient (P) of sugars and dextran size fractions on the Stokes’ radius (r_s). P is given on a logarithmic scale to illustrate the range of values. Points with the same symbol comprise data obtained at certain levels of the exchange rate γ (closed circles, 0·33; open circles, 0·5; closed squares, 0·66) after different diffusion periods (cf. Fig. 6). Closed triangles, Permeability values of α‐methyl‐d‐glucose and raffinose based on efflux kinetics (cf. Fig. 5).

Since the diffusion coefficient of small molecules in water is inversely proportional to the square root of the molecular weight , and the diffusion coefficient of polymers correlates with the reciprocal of the Stokes’ radius (r_s^–1), it is useful to consider the products of the permea bility coefficient (P) of the central capsule membrane (nexine) with the respective size parameters. P × r_s rises sharply as the molecular size of the dextran molecules decreases (Fig. 8) and the product of P and of the trisaccharide raffinose was only half the value found for α‐methylglucose (Table 4). Hence, size‐exclusion effects are relevant for permeability in the whole range of permeand sizes studied, and sugar permeation was not rate‐controlled by unhindered diffusion within the unstirred liquid of the capsule lumen. The volume fraction of the large pores, which enable dextran permeation up to a size of more than 2 nm, seems to be too small to control the sugar permeability. It is obvious that the pore size distribution in the nexine is not homogeneous. At present, it is unclear whether the polymer‐conducting diffusion pathways are open in the physiological system. Their existence might be physiologically significant in pollen ontogenesis. However, in the ripe pollen grain these pathways might be filled with carbohydrate or lipids. In this case, even small digestive enzymes such as trypsin (Stokes’ radius = 2·4 nm) might be excluded efficiently. The finding that dextrans with a Stokes’ radius of up to 2·5 nm equilibrated slowly with the capsules, whereas Evans Blue (Stokes’ radius = 1·3 nm) was excluded, points to the presence of fixed negative charges in the pores. Due to electrical barrier effects, the size exclusion limit of cation exchanger membranes can be much smaller for polyanions than for non‐electrolytes (e.g. Jäschke et al., 1992).

Fig. 8. Product of the permeability coefficient and the Stokes’ radius as dependent on molecular size. Symbols as in Fig. 7.

CONCLUSIONS

Results show that the permeability properties of the pine nexine and sexine are very different. The nexine is an ultrafilter membrane, which is impermeable to most proteins. Its permeability to sugars is low compared with that of a primary cell wall, but is nevertheless sufficient for the nutrition of the protoplast. The purified sexine at the external surface of the sacci is a microfilter that is highly permeable for even large polymer molecules and diffusible sub‐micrometer particles. With respect to applications for chromatography and immobilization, the most interesting properties of the sacci are rapid polymer exchange, a large and well accessible inner sporopollenin surface, and a high mechanical rigidity. The slow exchange of the central capsule for polymers and polyvalent anions might be of interest for controlled release applications.

ACKNOWLEDGEMENTS

We are grateful to Dr S. Rogaschewski for enabling and supporting the work with the Leica SEM. The study was supported by German Federal Ministry Economics (project ‘Chromatographieträger’ No 1279/00).

APPENDIX

Permeability coefficients

Permeability coefficients, P, of the central capsule envelope for the sugars and the dextran size fractions were calculated for the case of membrane‐controlled exchange (first‐order kinetics), where P can be derived from the rate constant k, the capsule volume V, and capsule surface area A.

P = k V/A

Volume/surface ratio of the central capsules

An estimate of V/A of the central capsule was obtained from mean values of the three perimeters of the ellipsoid capsule a, b and c (Table 2) as:

Rate constants

Rate constants, k, for sugar exchange were read directly from semi‐logarithmic plots shown in Fig. 5. Rate constants of dextran exchange were calculated from the exchange rate γ (Fig. 6) and the exchange time, t, as:

Half time of sugar exchange without membrane control (Table 4)

A sphere with radius has approximately the same volume as the ellipsoid central capsules with perimeters a, b and c. The half‐saturation time, t*, of sugar exchange with a spherical zone of unstirred liquid with this radius (Table 3) was determined from a solution of the second Ficks’ law given by Crank (1957) as t* = 0.0305 , where D is the diffusion coefficient of sugar in water.

Concentration difference across the nexine required for a relatively high rate of monosaccharide uptake by the male gametophyte

The stationary concentration difference, ΔC across the external diffusion barrier (nexine) of the gametophyte caused by a constant rate of monosaccharide uptake can be obtained as:

where P is the permeability coefficient of the nexine for a monosaccharide (Table 3), V/A is the volume/surface ratio of the gametophyte (Table 2), and u is the volume‐based uptake rate. For an uptake rate of 50 µmol cm^–3 h^–1, the concentration difference ΔC over the nexine is 2·48 mm for the preparation of P. sylvestris and 1·25 mm for the preparation of P. nigra.

Table 5.

Stokes’ radii of dextran molecules coordinated to defined exchange rates after different diffusion times

Diffusion time (s)

1200

3600

10 800

32 400

Exchange rate

0·33

0·50

0·66

0·33

0·50

0·66

0·33

0·50

0·66

0·33

0·50

0·66

Rate constant (s–1 × 10–3)*

0·916

0·578

0·338

0·305

0·193

0·113

0·102

0·064

0·038

0·034

0·021

0·013

Permeability coefficient (nm s–1)*

4·59

2·90

1·69

1·53

0·97

0·56

0·51

0·32

0·19

0·17

0·11

0·06

Stokes’ radius (nm)†

1·30

1·15

1·05

1·83

1·58

1·35

2·20

1·95

1·60

2·95

2·50

2·13

Data obtained for central capsules of Pinus sylvestris L.

* Calculated from exchange rate and diffusion time, as explained in the Appendix.

^† Read from graphs of the type shown in Fig. 6.

Received: 7 February 2003; Returned for revision: 8 April 2003; Accepted: 7 May 2003