II.A.3. Negative- Staining

a. Introduction

Fig. II.39. Schematic representation of a specimen particle completely embedded in a negative stain. (From Hayat and Miller, p.2)

For structural detail to be recorded in a micrograph, it must be faithfully preserved as the specimen is prepared and subsequently exposed to the vacuum and electron beam. The main purpose of negative-staining is to surround or embed the biological object in a suitable electron dense material which provides high contrast and good preservation (Fig. II.39). This method is capable of providing information about structural details often finer than those visible in thin sections, replicas, or shadowed specimens. In addition to the possibility of obtaining a spectacular enhancement of contrast, negative-staining has the advantage of speed and simplicity.

The potential value of negative stain microscopy derives from the relative ease with which specimens can be imaged at the molecular level of resolution. The technique has mainly been used to examine particulate (purified) specimens - e.g.. ribosomes, enzyme molecules, viruses, bacteriophages, microtubules, actin filaments, etc. at a resolution of 1.5-2.5 nm. This technique generally allows the shape, size, and the surface structure of the object to be studied as well as provide information about subunit stoichiometries and symmetry in oligomeric complexes. Any surface of the specimen accessible to water can potentially be stained, and thus, that part of the specimen will be imaged at high contrast.

b. Historical notes

Hall (1955) was the first to accidentally demonstrate the negative staining effect in a study in which particles were being positively-stained with phosphotungstic acid. Imperfectly washed particles were surrounded and embedded in the dried reagent and, instead of appearing dark on a light background, they were seen light on a dark background. Huxley (1957) independently noticed the same effect with tobacco mosaic virus. Brenner et al. (1959) also observed the same phenomenon with T2 bacteriophage and were the first to call it "negative staining".

c. Enhancing contrast of biological specimens

An object, whatever its chemical composition is, under normal conditions of TEM will barely be obvious (i.e. just noticeable contrast) in the electron image if the product of its thickness (in nm) and weight density (in g/cm3 is more than 10. Biological specimens have a density of about 1g/cm3, thus it is difficult in a micrograph to discern specimen features any smaller than 10 nm without some treatment to enhance its contrast. Positive staining involves treatment of the specimen with a chemical that increases the weight density, however, even intense staining that doubles the weight density would still only allow objects larger than 5 nm to be seen. When positively-stained, the object of interest appears in the image as an electron-opaque area against a relatively light background. Alternately, if the electron-transparent object is more or less completely embedded in a matrix of electron dense-material, and appears in reversed contrast as an electron-transparent area against a dark background, the object is said to be negatively-stained. Contrast enhancement by positive staining involves a direct interaction of a stain material with the protein. Negative staining does not require a stain-protein interaction, and denaturation of the molecule by staining is minimized, yet visualization to a limit of resolution of about 2.5 nm is easily realized.

d. Properties of an ideal negative stain

Each of the commonly used stains fails in some respect to meet the "ideal" specifications outlined below. An ideal stain should have:

- High density - Ability to protect specimen against dehydration effects- High solubility - Non-chemically reactive with the specimen- High melting and boiling points - Uniform spreading on the support film- Amorphous structure (i.e. structureless) when dry

High weight density to give high contrast

The density of most biological specimens ranges between 1.0 and 1.7 g/cm3. To increase the contrast of the specimen its mass thickness must be increased. The addition of non-specific stain to the specimen increases the specimen mass thickness, the effect being greater, the greater the mass density of the stain. Ideally the stain should be of uniform thickness and just sufficiently deep to completely engulf the specimen molecules. The "whole specimen" should be of limited thickness (10-50 nm). The resulting contrast of the specimen is much lower than that of the embedding amorphous film. In the photographic negative, the specimen appears dark against a light background. Hence the term "negative staining" which is not staining at all, but is embedding and is "negative" only in the sense that the biological material has less contrast than the surrounding material.

The density of most stains lies in the range 3.8-5.7g/cm3. The actual chemical composition of the stain has little effect on contrast. Note that the degree of contrast in the immediate vicinity of the object is determined by the density of the contrasting ion, not by the density of the contrasting substance as a whole. Heavy metal stains mostly cause elastic scattering of electrons, thus the main process of contrast formation is that of scattering contrast.

High solubility

The stain must be sufficiently soluble to allow enough to dissolve and come out of solution only in the final stages of drying. A relatively insoluble salt comes out of solution well before the drop finally dries and deposits as crystals without outlining or penetrating the specimen.

High melting and boiling points

The stain must be reasonably stable under the conditions to which the specimen is exposed in the microscope so it does not run or volatilize when heated by the electron beam. Energy transferred by inelastic scattering events produces radiation damage in the specimen. The rupture of chemical bonds with the formation of new structures and the ejection of materials from the irradiated area will eventually degrade the specimen. Transfer of energy to the specimen may also result in thermal degradation of the specimen and produce image drift due to thermal stresses in the specimen. Useful stains are highly resistant to such attack. Since the ultrastructural detail observable by negative staining is replicated by the stain, any degradation of the stain deposit during exposure to the beam will result in the loss of detail in the image.

Amorphous (non-crystalline) when dry

The drying of the stain and the specimen together with any buffer may produce artifacts, or contrast unrelated to the structure under study. Such "noise" degrades the quality of the image and may be of sufficient magnitude to obscure structural detail. To be most effective, the stain must deposit in an amorphous bed of uniform thickness and density to produce an artifact-free background against which the images of the specimen are viewed. A tendency to form microcrystals or other anomalies of structure or density in an otherwise homogeneous deposit will result in excessive "grain" effects in the slightly underfocused micrograph. Contrast from these effects may obscure the ultrastructural detail of the specimen.

If the molecular dimensions of the hydrated stain are the same or larger than the ultrastructure of the specimen, the visualization of that structural feature may be prohibited. Thus, some advantages may accrue from the use of stains of low molecular weight. Stains such as phospho-tungstate and silicotung state have dimensions of the order 0.8-0.9nm whereas uranyl stains are considerably smaller (0.4-0.5nm). This limits the possible resolution of specimen detail to values which are seldom if ever below 0.7 nm. Under normal conditions the resolution is limited to about 2 nm by inherent noise in the electron image. The highest resolution possible is obtained with negative stains such as uranyl formate, which are characterized by smaller size and hence increased penetrability. Less penetrating stains such as phosphotungstate are often used advantageously in systems where extensive penetration might obscure certain surface detail.

Non-reactive with the specimen

A most desirable, sometimes essential property of the stain is that it be chemically inert. An ideal stain is inert to the specimen as well as to the buffers, salts, metal ions, cofactors, or other reagents commonly used to maintain the integrity of the specimen structure. It is soluble and stable over a wide range of pH values. Any one stain may prove unsuitable for use with a particular specimen preparation by virtue of its reactivity with the specimen or the buffer solution. Uranyl salts are unstable at pH values >6 and are generally used below that pH. Uranyl ions form an insoluble compound with phosphate ions, thus, use of phosphate buffers in samples may lead to reaction between buffer and negative stain.

Partial disruption or changes in the tertiary structure of specimen molecules may occur with negative staining. This may prove advantageous or it may be necessary to chemically fix the specimen before staining to prevent disruption. Low concentrations (0.5-2%) of gluteraldehyde are sometimes used to fix the specimen before staining. To prevent inter-particle cross-links, the fixation is carried out for short periods (generally < 30 min) with extremely dilute samples. Another useful cross-linking agent is the bifunctional reagent dimethylsuberimidate which preserves positive charges at lysine residues in proteins, thus reducing conformational changes.

Protect against dehydration effects

In the final stages of drying, the dense stain ideally forms a uniformly thin, amorphous film in which the specimen is supported and preserved (Fig. II.39). The stain theoretically replaces the water in the interstices of the object until, in the dried specimen, all hydrated volumes are ideally filled with it. The enveloping stain then supports the specimen structure, if the deposit of stain is thicker than the embedded specimen. The specimen thus escapes exposure to surface tension forces and hopefully, retains its native morphology. A material of limited solubility may begin to precipitate before the last stages of drying and fail to engulf and protect the specimen. In general, those materials which make the best stains are very soluble.

Considerable changes in salt concentration and pH accompany drying of the specimens, but the stain helps protect the specimen from serious dehydration damage. When the specimen molecules are dried within a matrix of negative stain, surface tension forces are dissipated against the stain bed surface, thereby minimizing distortion of the specimen ultrastructure.

Uniform spreading on the support film

Excellent and uniform wetting of the support film surface is necessary to consistently achieve micrographs of good quality. A hydrophilic film surface enhances the tendency of the stain solution to spread and deposit the stain in an acceptably thin bed, free of artifacts (Fig. II.40). The support film may be made sufficiently hydrophilic either by glow discharging the grids or by treating the surface of the film first with a dilute (0.1 mg/ml) solution of cytochrome c or serum albumin before applying the specimen sample.

e. Additional notes about negative staining

1) Specimen purity

Negative stain microscopy is an excellent technique for assessing the purity of preparations and to appraise the efficiency of preparative procedures. Negative staining is of value in purification procedures, both for establishing a criterion of purity and for locating small amounts of a particular specimen particle in a fractionation scheme.

2) One- verses two-sided staining

With negative staining, sometimes only the top or bottom of a particle is contrasted with the stain. This usually occurs if the stain layer is very thin. In thicker layers of stain, particles will generally stain on both sides. There seems to be no way to reproducibly obtain one- or two-sided images for most specimens. Also, it is often difficult to ascertain the degree to which stain embeds both sides of a particle. The use of holey films (see Sec. II.A.3.g.1) provides some control over achieving more uniform staining of particulate specimens.

Fig. II.40. Schematic representation of the effect of the support film and the specimen surface on the meniscus of the stain before drying. (a) Hydrophilic specimen on hydrophilic support film. The adhesiveness of the stain to both the specimen surface and the support film is almost equal; it is greater than the cohesiveness of the stain. (b) Hydrophilic specimen on hydrophobic film. The adhesiveness of the stain to the specimen surface is stronger than the adhesiveness to the film, but the adhesiveness of the stain to the film is less than the cohesiveness of the stain. (c) Hydrophobic specimen on hydrophilic film. The adhesiveness of the stain to the film is stronger than the adhesive-ness to the specimen surface. The adhesiveness of the stain to the specimen surface is less than the cohesiveness of the stain. (d) Hydrophobic specimen on hydrophobic film. The adhesiveness of the stain to the specimen surface is greater than stain cohesiveness, but the stain cohesiveness is less than the adhesiveness of the stain to the film. Alternatively, the adhesiveness of the stain to the specimen surface is equal to the adhesiveness to the film, but the adhesiveness to the film is stronger than the cohesiveness of the stain. (From Hayat and Miller, p.23)

3) Positive vs. negative staining

Since most biological specimens have charged groups exposed to the surrounding aqueous solvent, and the stain molecules are ions, there will almost always be some degree of positive staining (Fig. II.41). For example, phosphotungstate ions are negatively charged and they can be used most effectively as a negative stain only for specimens raised above their isoelectric point so they do not attract the contrasting ions. Uranyl acetate at pH 4.5 is used as a negative stain with specimens that are positively charged below pH 5. With the anionic stains such as the phosphotungstates, positive staining is expected in the region of low (acidic) pH values because of the powerful interaction between positive charges in the specimen and the negatively charged tungstate ions.

Fig. II.41. The effects produced by (a) positive staining, (b) negative staining, and (c) metal shadowing on a fragment of collagen lying on a supporting membrane. The overall effect of banding is clearly shown by all three methods, but each supplies certain information which is lacking in the others. Shadowing shows clearly the granular nature of the carbon support film. (From Meek 1st ed., p.466)

4) Maintenance of biological structure/function

It has been found that some viruses are still infectious after mixing and spraying with phosphotungstate and certain enzymes remain active even after they have dried in droplets of stain and are subsequently rehydrated.

5) Reliability of images

It is always important, especially with unfamiliar specimens, to try a variety of conditions (stain, pH, temperature, concentration of specimen, stain and buffer, etc.) when preparing specimens for microscopy using negative staining techniques. Quite often, under varied conditions, different features of a specimen will be enhanced, and either complementary or perhaps even contradictory information may be obtained. If, for example, a variety of stains are used and similar staining patterns are obtained, then it is likely that the features revealed are consistent with genuine specimen morphology.

f. Common negative stains

The following table lists some chemicals used as negative stains along with some of their properties:

STAIN FORMULA Solubility
( g/100ml H2O )
Anhydrous Density
( g/cc )
ANIONIC STAINS
Ammonium molybdate' NH4Mo7O24•4H2O 44 2.5
Sodium phosphotungstate Na3PO4•12WO3 ? 3.8
Sodium tungstate Na2WO4 90 4.2
CATIONIC STAINS
Silver nitrate AgNO3 220 4.4
Cadmium iodide CdI2 85 5.7
Uranium nitrate UO2(NO3)2•H2O 150 3.7
Uranyl acetate UO2(C2H3O2)2•2H2O 8 2.9
Uranyl formate UO2 (CHO2)2•H2O 7 3.7

Oliver (1973) reports that methyl phosphotungstates and methylamine tungstate may be used as negative stains. They have the primary advantage of a greater tendency to wet the support film surface, spreading over sufficiently large areas of relatively thin stain bed of good uniformity. They appear to be stable over a wide range of pH values (ranging from pH 4 to 9.5).

g. Negative Staining procedures

Techniques for the preparation of negative stain specimens are simple and direct. The essential aim of the procedure is to embed the specimen in a uniformly thin deposit of stain. Resolution of molecular features is only accomplished at the stain-specimen boundary where there is maximum contrast. This result is only achieved if the deposition of buffer salts or other materials with densities less than the stain at that boundary are severely limited; otherwise the specimen molecules will be imaged at low resolution and will appear as nondescript blobs.

The specimen sample is usually applied directly to the surface of the support film where a population of specimen particles becomes adsorbed. Attachment of the molecules is usually secure enough that they are not removed by subsequent rinsing and staining operations which do remove most of the buffer salts. Since different specimens often have different affinities for the particular support film being used, some measure of control over how much specimen attaches to the film may be effected by adjustment of the specimen and buffer concentrations, adsorption time, etc. Appropriate conditions must be established by experiment for each new specimen. For protein solutions, typical concentrations range between 50-500 µg/ml and adsorption times from as little as 1-5 seconds to several minutes. Stains are usually applied in a range of concentration from 0.25-4.0%. Adjustment of stain concentration provides some control over the thickness of the deposit.

It does NOT follow that a procedure that is successful with one type of specimen is also suitable for another, so various modifications should always be tried until good contrast and spreading conditions are achieved.

There are two common procedures for preparing negatively-stained specimens on EM grids:

Adhesion (drop) method (Figs. II.42 and II.43)

A droplet of specimen is placed on the surface of the grid support film, making sure it sufficiently wets the surface. After an appropriate time interval, excess specimen is wicked away by touching a piece of filter paper to the edge of the grid surface. Without letting the grid dry, a droplet of rinse or stain solution is applied to the grid. Rinsing is necessary if the specimen preparation contains high concentrations of buffer salts or other solutes which may interfere with deposition of stain. The nature of an appropriate rinse depends on the conditions that the specimen can tolerate. Many viruses, for example, can withstand rinsing with distilled water. In some instances the stain solution can itself act as a suitable rinse. After rinsing and staining, excess fluid is wicked from the grid, leaving a thin aqueous film on the surface which is left to dry, usually in air.

Fig. II.42. Preparation of a specimen from particles in aqueous suspension. (From Hall, p.290)
Fig. II.43. Washing a specimen. (From Hall, p.290)

The specimen can also be applied to the support film by floating the grid on top of a droplet of the specimen solution. The grid is then transferred to droplets of rinse and stain solutions and then dried as before.

An additional variation of the usual adhesion method is to apply the sample to a holey support film in the same way as is done on regular films. The sample dries in a thin layer of stain stretched out over the holes, thereby giving maximum contrast since there is no plastic and/or carbon support. Also, the stain tends to be more evenly distributed around the particle although the particle often undergoes distortions (shrinkage and flattening) due to the surface tension forces created as the layer of stain dries. The stain layer also has a tendency to break either before or after it is exposed to the electron beam. The layer of stain can be stabilized with a thin layer of evaporated carbon. Another advantage of this technique over the usual method is that, if small enough, the specimen particles will be randomly oriented in the stain layer. On regular support films, particles often settle on the surface of the film in one or a few preferred orientations, thus limiting the possible views of the specimen.

Spray droplet technique(Fig. II.44)

The normal adhesion method of preparing a particulate suspension may lead to erroneous conclusions about the relative proportion of particles since different particles are likely to have different affinities for the substrate. Also the microscopist may select fields attractive to the eye but which are not representative. The only reliable way of preparing specimens without introducing a bias is to dry a drop of the original sample in its entirety. Non-volatile salts and buffers must be removed by centrifugation and washing or by dialysis so they don't obscure the particles under study or alter the structure when the salt concentrates in the last stages of drying. The entire residue from the drop must be examined, thus it is necessary to obtain very small drops. The suspension is atomized as a fine mist and the droplets are allowed to impinge on the substrate.

Fig. II.44 A simple hand-held nebulizer from which the sample (s) is sprayed onto mica (m). (From Willison and Rowe, p.69)

The spray droplet technique is particularly useful for examining specimens that adsorb so poorly to the support film that application and removal of rinsing and stain solutions also removes the specimen. Appropriate volumes of the specimen and stain solutions are mixed and sprayed in small droplets onto a wetable support surface. If the solution itself has the propensity to wet and spread over the surface, uniformly thin deposits of negatively-stained specimen result. The resultant aqueous film will be of uniform depth, and the mass of stain deposited per unit area of support film tends to be constant. When the specimen appears to have dried, some water may still be present in the stain bed, and a rearrangement of the stain deposit could result from its rapid vaporization if the specimen is suddenly placed in the vacuum of the microscope. Thus, the specimen is usually allowed to dry (sometimes over a desiccant) for at least 10 minutes.

Drying of the aqueous film proceeds from the edges, the central area covered by the droplet being last to dry. Minor solutes tend to be held in solution until the last stages of drying and are deposited in highest concentrations in the central area. As a result, the specimen in this area ordinarily is of inferior quality.

Fig. II.45. Drop pattern. The small particles are tomato bushy stunt virus, and the large spheres are polystyrene latex particles 2,600 Å in diameter. The wedge-shaped sector has been magnified and superimposed. (From Hall, p.359)

Note that, using the adhesion drop method, there may be preferential adherence of particles so relative particle distribution counts cannot be made. A major advantage of the spray technique over the adhesion method is that preferential adherence to the grid of one type of particle over another type of particle in a mixture cannot occur. Thus, this is the method of choice in quantitative studies where relative concentrations of particles in the sample need to be determined. By knowing the volume of the original drop (from adding known concentrations of polystyrene spheres) a count of the number of particles in a drop pattern provides immediately the number of particles per unit volume (Fig. II.45: Note that this figure shows a metal-shadowed specimen). This, together with the mass per unit volume obtained by weighing the dried residue from a measured volume can be used to calculate a value for the mean molecular weight of the particles.