Simple and Rapid Methods for SEM Observation and TEM Immunolabeling of Rubber Particles

Abstract

We developed a method involving air-drying of a rubber suspension after fixation in glutaraldehyde-tannic acid and postfixation in osmium tetroxide for SEM observation. For TEM immunolabeling the suspension was air-dried after osmium-only fixation. Whereas conventional methods failed to satisfactorily stabilize rubber particles, the methods described here proved successful in preserving their integrity.

Rubber is a natural renewable substance and a source of valuable products. Although over 2000 plants from various taxa produce rubber latex, the rubber tree (Hevea brasiliensis) is still the most important commercial source of rubber because of the high quantity and quality of rubber from its latex. However, because Hevea plantations are declining and because rubber from this plant can cause severe allergic reactions, there is a pressing need to find alternative sources of natural rubber and to understand the mechanism underlying rubber biosynthesis.

Electron microscopic examination of rubber particles (Wood and Cornish 2000) suggests that they consist of hydrophobic rubber polymers enclosed by a monolayer membrane. This membrane consists of lipids, proteins, and other components and is the site of rubber biosynthesis. In recent years, significant progress has been made in identifying components of rubber biosynthetic machinery, and rubber transferase is regarded as the key enzyme. Loss in the activity of this enzyme after isolation from rubber particles suggests involvement also of other factors, the composition and function of which are poorly understood.

We are presently characterizing rubber particles and the rubber biosynthetic process in a number of rubber-producing plants in search of alternative plant sources to obtain rubber of desirable properties (Singh et al. 2003). With the ultimate goal of engineering the properties of rubber particles, we are studying their physical and chemical characteristics. We had to develop suitable methods for rapidly screening rubber particles from various plant sources for their physical and chemical properties. The method developed for determining the size and shape of rubber particles using SEM involved air-drying of a particle suspension after fixation in glutaraldehyde-tannic acid and post-fixation in osmium tetroxide. A preparation method was also developed to characterize the limiting membrane using TEM immunolabeling, initially to localize a 24-kD protein. This protein is considered to be closely associated with Hevea rubber particles and to play an important role in certain aspects of rubber biosynthesis (Oh et al. 1999). The need to develop these methods arose because our initial attempts to prepare samples using conventional methods provided unsatisfactory results.

Although the work in our laboratory involves characterization of rubber particles from several alternative plant sources, here we compare Hevea with only one other species, Ficus benghalensis. Rubber particles from these two sources were purified according to the procedure of Singh et al. (2003). For SEM observations, the suspension of rubber particles was fixed in 3% glutaraldehyde (prepared in 0.05 M sodium cacodylate buffer with or without 1% tannic acid) for 1 hr at room temperature (RT). The sample was washed in several changes of buffer, postfixed in 1% aqueous osmium tetroxide also for 1 hr at RT, and washed in distilled water. A small drop of the suspension (appropriately diluted) was placed on a membrane filter (0.2-μm pore size), dehydrated in ethanol, and critical point-dried. Alternatively, a small drop of the suspension, fixed as above, was placed on a piece of cover-glass and then air-dried. All samples were mounted on aluminum stubs, coated with gold particles in a sputter coater, and examined with a Hitachi S-2400 SEM. For TEM immunolabeling, samples were prepared as follows: (a) suspension was not fixed; (b) suspension was fixed in a mixture of 2.5% paraformaldehyde and 0.5% glutaraldehyde (prepared in 0.05 M sodium cacodylate buffer) for 1 hr at RT; (c) suspension was fixed in 1% osmium tetroxide (prepared in 0.05 M sodium cacodylate buffer) also for 1 hr at RT. Samples b and c were washed in several changes of buffer. Before immunolabeling, all samples were either air-dried or prepared in the conventional way for ultrathin sectioning. For air-drying a small drop of suspension was placed on formvar-coated nickel grids and air-dried. Alternatively, samples were dehydrated in ethyl alcohol and embedded in LR White resin. Ultrathin sections cut on an RMC MT-X ultramicrotome were collected on formvar-coated nickel grids. Samples from all preparations were immunolabeled according to Kim et al. (2002). Briefly, before treatment with the primary antibody (for 24-kD protein), grids were sequentially floated on glycine (in PBS buffer), distilled water, PBS buffer, and normal goat serum (NGS)-PBS buffer. Grids were treated with 24-kD protein-antibody (1:200 dilution) for 90 min at 37C. Grids were washed sequentially in PBS-non-fat milk-Tween 20, PBS, Tris-HCl, and then treated with gold (10-nm particles)-conjugated anti-rat IgG (1:20 dilution) for 2 hr at RT. Grids were washed sequentially in TBS-Tween 20, TBS, Tris-HCl, and distilled water. For controls, grids were processed as above but were not treated with the primary antibody. Grids from all preparations were examined with a JEOL JEM 1010 TEM.

For SEM, the most effective preparation for retaining the natural form and size of rubber particles for both Ficus and Hevea involved air-drying after fixation in glutaraldehyde (containing tannic acid) and postfixation in osmium. The same fixation failed to retain these features of rubber particles when used in combination with critical point-drying (CPD). CPD caused considerable distortion in the shape of rubber particles from both Ficus and Hevea (illustrated only for Ficus; Figure 1) due to collapse of their surface. Although the extent of collapse was variable, with some particles suffering only slight to no distortion, most particles showed signs of severe collapse. In comparison, in the air-dried preparation particles from both Ficus (Figure 2) and Hevea (Figure 3) retained their natural shape and size. Ficus rubber particles were spherical, occurring predominantly as discrete individual particles in a range of sizes. Particle size measurements undertaken (Singh et al. 2003) showed particles in the range of 1.6–6.0 μm. Hevea particles (Figure 3) were also discrete but were smaller and more highly variable in size compared to Ficus particles. Size measurement for Hevea particles (Singh et al. 2003) placed the particles in the 0.08–0.75-μm range, which agrees with previously reported values for particle size in this species (Gomez and Hamzah 1989; Cornish et al. 1999; Wood and Cornish 2000).

Figure 1

CPD preparation of F. benghalensis rubber particles. Particles show severe collapse. SEM. Bar = 4 μm.

Figure 2

Air-dried preparation of F. benghalensis rubber particles. Particles show no sign of collapse and appear spherical. SEM. Bar = 4 μm.

Figure 3

Air-dried preparation of H. brasiliensis rubber particles. Particles show no sign of collapse and appear spherical. SEM. Bar = 2 μm.

The TEM immunolabeling method proved most effective in retaining the spherical shape of rubber particles and in specific labeling of the 24-kD protein. This involved fixation of rubber particle suspensions directly in osmium before preparing whole mounts and labeling. Fixation in paraformaldehyde-glutaraldehyde, which is normally employed in immunocytochemistry, was largely ineffective in retaining the natural shape of rubber particles. The specificity of gold labeling was also variable. Most unsatisfactory was the method where labeling was carried out on unfixed material. Furthermore, only the rubber particles from Hevea were labeled for 24-kD protein. The particles from Ficus were not labeled in any of the preparation methods used and therefore are not illustrated. Illustrations have been selected to cover the range of preparation methods used. In unfixed preparations, rubber material appeared as highly irregular masses (Figure 4). The background labeling was high compared to the intensity of gold particles associated with the masses of rubber material (Figure 4). Fixation in paraformaldehyde-glutaraldehyde led to some improvement in retention of particle shape and specificity of labeling (Figure 5). Particle shape varied from near-spherical to highly irregular. Labeling intensity was rather poor, although there was noticeable improvement in specificity over unfixed material. Dramatic improvements in retention of particle shape and in specificity and intensity of labeling were obtained when the material was directly fixed in osmium. Rubber particles of all sizes were spherical and labeled intensely with gold particles, closely associated with the surface of particles (Figure 6 and 7). Background labeling was negligible.

It was apparent that CPD, the method of choice for SEM with most biological specimens, was unsuitable for rubber particles because it greatly distorted their shape. Curiously, air-drying, rarely used in SEM because it causes severe shrinkage distortion of biological materials, was most satisfactory in stabilizing the shape and size of rubber particles when used in conjunction with prior fixation in glutaraldehyde-tannic acid and postfixation in osmium. There are indications that plant products, such as latex, are susceptible to extraction by dehydration fluids, such as acetone and ethanol (Condon and Fineran 1989) employed in CPD preparations. Apparently, CPD caused an irreversible change in the properties of the limiting membrane of rubber particles, rendering the membrane “leaky.” This led to extraction of rubber material and consequent collapse of particles. Obviously, fixation of rubber suspensions in glutaraldehyde-tannic acid and postfixation in osmium before CPD drying did not stabilize all of the membrane components sufficiently to prevent extraction of rubber material, although tannic acid is known to prevent extraction of some tissue components (Simionescu and Simionescu 1976). Tannic acid has also been used to enhance osmification of tissue culture cells (Katsumoto et al. 1981). Observations of some collapse in air-dried preparations of rubber suspensions that had not received tannic acid treatment (not illustrated) indicates that tannic acid was an important component of the fixation vehicle. A combination of glutaraldehyde-tannic acid and osmium stabilized and strengthened the limiting membrane sufficiently to resist the forces of surface tension generated during air-drying.

Figure 4

TEM-immunolabeled H. brasiliensis rubber material. Preparation: unfixed, air-dried whole mount. The material is irregular in form and background labeling is high. Bar = 500 nm.

Figure 5

TEM-immunolabeled H. brasiliensis rubber material in an ultrathin section. Preparation: paraformaldehyde-glutaraldehyde fixation. The material is irregular in form. Gold particles are associated with the rubber material but labeling intensity is low. Bar = 400 nm.

Figure 6

TEM-immunolabeled H. brasiliensis rubber particles. Preparation: osmium-fixed, air-dried whole mount. The particle is of the large size class (Singh et al. 2003) and spherical. Labeling is intense, gold particles being associated with the particle surface. Bar = 200 nm.

Figure 7

TEM-immunolabeled H. brasiliensis rubber particle. Preparation: osmium-fixed, air-dried whole mount. Particles are spherical. Labeling is intense, gold particles being associated with the particle surface. Of the two particles present, one is of the medium size class and the other of the small size class (Singh et al. 2003). Bar = 200 nm.

Comparison provided here of particle shape and specificity of immunolabeling using various TEM preparation methods suggests that osmium-only fixation combined with air-drying proved to be the most suitable method. It is apparent from the highly irregular form of the rubber material that there was little or no stabilization of particle membranes in unfixed preparations. Consequent fusion and fragmentation of rubber particles may be the reason for the high background labeling in the unfixed material. Aldehyde fixation destabilizes latex particles, rendering them prone to fusion (Condon and Fineran 1989). Also, in preembedding dehydration steps, acetone appears to have extracted rubber from particles, which was more severe for Ficus (not shown). These may be the reasons why the standard protocol for TEM immunolabeling, using paraformaldehyde-glutaraldehyde fixation, was not effective in preserving the shape of rubber particles. However, labeling specificity was markedly superior to that in unfixed samples, which may be related to better preservation of membrane proteins and to less fragmentation and spreading of the rubber material. The most satisfactory preparation for retention of particle shape and labeling specificity involved air-drying after osmium-only fixation. Although osmium fixation is considered undesirable in immunocytochemistry because of its adverse effect on antigenicity, this fixation was the method of choice in our work.

The methods recommended here for SEM and TEM immunolabeling with rubber particles are simple and rapid. Although they might be considered unconventional and unsuitable for most biological samples, they were most effective in our work because of their compatibility with the unique chemical and structural properties of rubber particles. These technical developments have made it possible to obtain valuable information on the architecture of rubber particles and have also helped clarify the role of 24-kD protein in rubber biosynthesis (Singh et al. 2003).