227
The effect of certain fixatives on particles of a globular
protein: an electron-microscopical study
By K. D E U T S C H , E. F I S C H E R , and W. KRAUSE
(From the Department of Electron Microscopy, University of Greifswald, East Germany)
With one plate (fig. i)
Summary
Particles (regarded as molecules) of bovine serum albumin have been studied by
electron microscopy both in the untreated state and after fixation with osmium
tetroxide, formaldehyde, and potassium permanganate. All three fixatives, especially
potassium permanganate, increased the density of the particles considerably.
Introduction
A BRIEF report has already been made of a preliminary study of the effect
of osmium tetroxide on minute protein particles (Deutsch and others, 1962).
The purpose of the work described in the present paper was to extend the
study, especially by working with fixatives other than osmium tetroxide and
by making measurements of the changes in density caused by fixation.
Materials and methods
Specimens of bovine serum albumin were obtained from Sera of Heidelberg
and from Professor J. Segal, Department of General Biology, University of
Berlin, who himself prepared some of the material used in the investigation.
The protein was dissolved in a phosphate/sodium hydroxide buffer solution
at pH 7-2, in the proportion of 1 mg of protein to 1 ml of buffer. 100 ml
of twice-distilled water was added to 1 ml of the protein solution.
For the investigation of the untreated protein a drop of the diluted solution
was deposited on a specimen grid coated with a collodion film. The solution
was allowed to dry.
The following methods were applied to study the effect of fixatives on
the particles. (1) A drop of the diluted solution was deposited on the specimen
grid, exposed to the vapour of formalin, and allowed to dry in air. (2) Solutions
of osmium tetroxide or of potassium permanganate were added to the diluted
protein solution, to make the final concentration of the fixative substance
0-2% w/v in the case of osmium tetroxide, or 0-1% w/v in that of potassium
permanganate. The solutions were then dialysed to remove all fixative substance that had not attached itself to the protein particles. Small drops of the
dialysed solutions were deposited on specimen grids and allowed to dry in
air. Collodionfilmswere used throughout, because agglomeration of the protein
occurred on carbon films.
[Quart. J. micr. Sci., Vol. 105, pt. 2, pp. 227-30, 1964.]
228
Deutsch, Fischer, and Krause
Micrographs were taken with a Zeiss D 2 electron microscope working
at 48 kV, with instrumental magnification of 18650. A photocell was used
as an exposure photometer. The photocell was attached to the screen and connected to a sensitive galvanometer (von Ardenne, 1944). The micrographs
were enlarged and evaluated densitometrically. Great care was taken to do
all electron microscopical and photographic work under the same conditions
to enable a quantitative comparison of the staining effect of the various
fixatives to be made. For comparison, thin layers of various thicknesses were
prepared by evaporating gold on a specimen grid. The mass density of the
protein molecules (treated and untreated) was correlated with the thickness
of gold films of corresponding mass density.
Results and discussion
The particles are represented on the electron micrographs as small dots
(fig. 1, A to D). The facts are consistent with the belief that each particle
represents a protein molecule. The diameter (about 5 m^.) and shape of the
particles are in agreement with the results of an electron microscopical
investigation by Valentine (i960), who used negative staining for specimen
preparation. The danger of artifacts was avoided by examining in the electron
microscope blank supporting films and films on which a drop of the diluted
buffered solution had been allowed to dry. The instrumental resolution
was about 2-0 m/x; hence accurate measurements of the size and shape of the
particles were not possible. Working near the limit of resolution of the instrument also decreased the accuracy of the densitometric measurements.
TABLE I
d (in myx) is defined as the thickness of a gold layer which has the same
mass density as a monomolecular layer of protein.
untreated
fixed with formaldehyde
fixed with osmium tetroxide
fixed with potassium permanganate
about
about z-od
about z-od
about 3mod
Inspection of the electron micrographs shows that the particles treated
with formaldehyde, osmium tetroxide, or potassium permanganate are much
more electron-dense than untreated ones. This is confirmed by the results
of the densitometric evaluation (table 1). There are several possible reasons
for the increased contrast. (1) It is well known that the mass density of some
specimens is reduced in the electron beam, as a result of chemical changes
FIG. I (plate).
A, untreated albumin particles.
B, albumin particles treated with osmium tetroxide.
c, albumin particles treated with formaldehyde.
D, albumin particles treated with potassium permanganate.
A few of the protein particles (regarded as molecules) have been indicated by arrows.
I .*.
FIG. I
K. DEUTSCH, E. FISCHER, and W. KRAUSE
Effect of fixatives on protein molecules
229
brought about by the effect of the beam on the specimen. It is quite likely
that this effect is smaller on a 'fixed' specimen; hence the reduction of the
mass density in the beam may be smaller or negligible. (2) It is also likely
that treatment with some fixatives increases the mass density considerably.
This holds for osmium tetroxide and potassium permanganate, and the
possibility cannot be excluded that formaldehyde also reacts with the protein
to produce this effect. The particles of protein treated with potassium
permanganate have a dark core (diameter about 5 m/i) and are surrounded
by a 'halo', presumably a 'deposit' of some sort. This deposit surrounding
the molecule also contributes to the mass density. (3) The molecules may
contract somewhat as a result of the treatment with fixatives and this effect
would also increase the mass density. We have the impression that molecules
of protein treated with formaldehyde have a smaller diameter than untreated
ones.
However this may be, the decisive factor for electron microscopy is the
increase of contrast, and it may be said that these fixatives also act as 'electron
stains'. It is, of course, tempting to generalize from our results. All globular
proteins have a rather similar structure, and the fixatives we have studied
probably have a similar effect on many globular proteins, though there will
be quantitative differences.
The contrast in cells fixed with osmium tetroxide is usually good, and,
as our investigations show, this result must be due in part to the effect of this
fixative on proteins. In specimens fixed with potassium permanganate,
certain structures are very electron-dense. This might indicate that they contain a large amount of protein, as it is conceivable that this fixative has a
specific staining effect on proteins. Reference should be made here to a
comprehensive study of fixation by potassium permanganate for electronmicroscopy, by Bradbury and Meek (i960). These authors found that potassium permanganate is a medium for revealing membrane structures, which
are supposed to contain a considerable amount of proteins. This is in agreement with the 'staining' effect that we have observed. But according to
Bradbury and Meek, potassium permanganate is not a true fixative: the
actual fixation is effected by the treatment with alcohol. In spite of this,
potassium permanganate is usually regarded as a fixative, though basic
proteins are supposed to be leached out by the action of potassium permanganate and subsequent treatment. These findings show that one has to be
very careful in generalizing from the results obtained by studying only one
type of protein. Bradbury and Meek also found that a very fine precipitate
of manganese dioxide is formed after treatment with potassium permanganate.
This deposit is presumably identical with the one found on protein particles
treated with potassium permanganate.
In specimens fixed by formaldehyde, certain structures are fairly electrondense. This may be due to a specific 'staining' effect of formaldehyde on
proteins; but to decide this question, an investigation would be necessary
to find out whether potassium permanganate and formaldehyde 'stain' cellular
230
Deutsch, Fischer, and Krause—Protein molecules
constituents other than protein. However this may be, formaldehyde and
potassium permanganate do increase contrast by 'staining' protein. It should
be pointed out that even untreated protein itself has enough intrinsic contrast
to be visible in the electron microscope.
References
ARDENNE, M. von, 1944. Kolloid-Z., 108, 195.
BRADBURY, S. and MEEK, G. A., i960. Quart. J. micr. Sci., 101, 241.
DEUTSCH, K., 1962. Wissenschaftliche Zeitschrift, Ernst-Moritz-Arndt-Universitat Greifswald, 11, 71.
SEGAL, J., and KALAIDJEV, A., 1963. Nature, 19s,
177.
VALENTINE, R. C, i960. Proceedings of the Regional Conference on Electron Microscopy,
p. 708.