вход по аккаунту


Ruthenium red and violet. II. Fine structural localization in animal tissues

код для вставкиСкачать
Ruthenium Red and Violet
Department of Biological Structure, University of Washington,
Seattle, Washington 981 05
The inorganic dye, ruthenium red, stains extracellular materials
in animal tissues which probably are acidic mucopolysaccharides. It complements
other techniques, its advantages being fine grain, high resolution and good contrast. Localization is shown in mouse and rat muscle, heart, lung and intestine,
frog cartilage and cells scraped from oral epithelium of human beings. Attention
is paid to collagen bundles, the cell/collagen interface and particularly the myotendinal junction, cartilage matrix and agar gel, desmosomes, intestinal microvilli, erythrocytes and vascular endothelium, nerve fibers and the T-system of
striated muscle. Although ruthenium red generally is excluded by plasma membranes, it penetrates giving intracellular density, if the membrane is broken.
Even when the cell membrane is intact, exceptions occur with selective staining
of the T-tubules or the sarcoplasmic sacs depending upon the state of contraction
of the muscle cell, and with intracellular staining of certain nuclei and epithelial cells. Ruthenium red stains intracellular lipid droplets revealing lamelJae,
and stains myelin forms grown from crude egg lecithin but cannot penetrate
deeply. It is localized in extracellular materials which have an important mechanical function. Its exclusion by cell membranes permits tracing tortuous cellular
invaginations and those exceptions to its exclusion invite a comparison of the
localization of the dye with the function of the cell.
Cells advertise themselves to the outside
world and to each other by their surfaces.
Scientists, however, are better acquainted
with cells in terms of their internal organelles and molecules, having violated their
privacy with microscopes, homogenizers,
and separation procedures. The research
directed toward this outer surface, called
the plasma membrane, has been concerned
mainly with the lipids and proteins which
are known to reside there. However, during
the 1960’s evidence accumulated prompting consideration of another component
which is minor in terms of dry mass but
major in terms of function. This third
component is polysaccharide, usually present as a glycoprotein. Its importance was
strongly emphasized from a morphological
standpoint, perhaps f m t by Bennett, who,
in 1963, coined the general term “glycocalyx,” although Chambers by 1940 had
designated similar layers in other cells as
“extraneous coats,” and their history can
ANAT. REC.,171: 369-416.
be traced much earlier. In 1966, Rambourg,
Neutra and Leblond published an important paper after examining a wide variety
of tissues with several histochemical tests
for polysaccharides. They concluded that
“nearly all cells investigated are covered
by a thin but definite band of stained material indicating the existence of a surface
layer.” They also felt that “the layer is not
part of the plasma membrane itself but
is a surface coat.” The biochemical difficulties which hindered work on these substances as well as their general importance
in biology has been elegantly reviewed
recently by G. M. W. Cook (’68).
It was not out of ignorance that the fine
structure of these substances was delayed
so long. After all, at the light microscope
level, the periodic acid-Schiff (PAS) reaction had been used reliably for several
decades, and the Hale technique for acid
polysaccharides using colloidal ferric oxide
Received Dec. 23, ’70. Accepted M a y 17, ‘71.
was disclosed in 1946 (Hale, '46). The
problem lay in the difficulty of applying the
light microscope methods to the electron
microscope, and to the general coarseness
of the colloidal particles which were deposited (Gasic and Berwick, '63; Yardley
and Brown, '65; Benedetti and Emmelot,
'67; Gasic et al., '68; Jones, '69). In 1964
Revel published a new method for electron
microscopical localization of acid mucopolysaccharides in cartilage matrix. Instead of colloidal iron oxide he used colloidal thorium dioxide (Thorotrast) which
was bound selectively to the thin section.
This method is applicable to other tissues
as well (Berlin, '67). Again, granularity
limited the resolution.
Also in 1964, Seligman and his group
(Hanker et al., '64; Seligman et al., '65)
applied their discovery of selective staining by "osmiophilic reagents" to the periodic acid reaction already well established
for polysaccharides. This gave promise of
fine-grained, high-contrast electron microscopy together with a precise histochemical
method selective for those polysaccharides
containing the vic-glycol grouping. ThiCry
('67) has shown that the promise can be
realized and has improved the procedure
with the use of silver proteinate.
In 1965, Doggenweiler and Frenk, while
attempting to replace the calcium of cell
membranes with lanthanum, obtained unusual staining of nerve intercellular gap
substance when their tissue was fixed in
lanthanum permanganate. Revel and Karnovsky ('67) traced this effect to the use of
solutions of ionic lanthanum brought near
neutrality with dilute sodium hydroxide to
the point of faint opalescence. Using this
solution they could demonstrate an hexagonal array of subunits in the narrow
spaces at certain cell junctions. The mechanism or specificity of lanthanum staining
have received little attention, although the
colloidal lanthanum oxide resembles chemically the Thorotrast which Revel had used
earlier (1964). Pictures obtained with Ianthanum (Revel and Karnovsky, '67; Overton, '69; Khan and Overton, '70) are quite
similar to those shown in the present study
using ruthenium red.
In 1967 Rambourg and Leblond repeated
their 1966 work with the electron microscope using the periodic acid-silver meth-
enamine procedure (which had gradually
evolved through the efforts of many
workers [Marinozzi, '611 beginning with
Gomori in 1946). They could demonstrate
the existence of cell coat materials, as well
as intercellular substances in a great variety of rat tissues, with evidence that they
contained polysaccharides of the type reactive with periodic acid as well as the acidic
polysaccharides capable of binding thorium
dioxide. Similar localization of polysaccharides in plant cells was demonstrated by
Pickett-Heaps ('67, '68). Unfortunately,
the granularity of the silver deposit limited
the resolution, but provided excellent contrast at lower magnifications. The periodic
acid techniques have been compared very
nicely by ThiCry ('67).
Although phosphotungstic acid (PTA)
has had a long history in electron microscopy as a rather nonspecific stain for proteins (Zobel and Beer, '65), Pease ('66)
showed in a detailed paper that PTA could
stain substances in tissue sections with a
specificity similar to that of the periodic
acid methods provided that the tissues were
embedded in the water-soluble methacrylates (glycol or hydroxypropyl methacrylate) and that concentrated PTA solutions
were used. The resulting density was nearly
grainless, but lacked the crispness necessary for highest resolution, and the PTA
washed out of the section quickly. Subsequently, Rambourg ('67) and Rambourg
et al. ('69) showed that if the PTA were
further acidified with chromic acid or HCl,
the crispness could be improved and the
sections could be rinsed quickly with water,
giving less contamination. ThiCry ('67) has
shown that PTA can be used under very different conditions to give similar results.
In 1964 the author found that the inorganic dye ruthenium red (which had given
venerable seriice in the hands of botanists
for demonstrating the location of pectins
in plant tissue histology) could be adapted
to electron microscopy (Luft, '64a). This
method gave especially satisfying localization in various tissues without resorting to
the low pH's commonly employed by the
earlier methods. At the same time, it developed densities in tissues with fine
enough granularity to define reliably
plasma membrane structure. This group
of three papers (Luft, '71, I; Szubinska and
Luft, '71 ), summarizes subsequent progress
in defining the mechanism of the ruthenium red (RR) reaction as well as the several impurities present in the commercial
material. A series of progress reports concerning biological results together with the
basic details of the method have been published (Luft, '64a; Szubinska, '64; Luft,
'65a, b, c; Luft, '66a, b, c; Luft, '68) or distributed as preprints (Luft, '64b; Luft,
This present paper reports the appearance of various tissues after the RRosmium tetroxide reaction has occurred.
No attempt has been made to verify the
specificity of the reaction in these various
tissues. The question of specificity is left
at the level of what can be deduced from
the chemistry, which is discussed in the
preceding paper (Luft, '71, I). The selectivity of the method is sufficient to reveal
at high resolution substances or structures
which have been difficult or impossible to
visualize previously, and in several instances confirms the existence of structures
which had been predicted indirectly by
earlier experiments. Hopefully, future work
will be done to elucidate the identity of
those materials which accumulate density
in tissue under the conditions described
herein, i.e., are RR positive. Already papers
upon this theme are appearing (Gustafson
and Pihl, '67b; Pihl et al., '68; Morgan, '68).
It is hoped that the reliability and control
of the reaction resulting from these reports
of chemical mechanisms, purification and
biological localization for both RR and
ruthenium violet (RV) will make easier
the future task of assigning specificity.
cessed as tissue fragments. Agar was examined as a 0.5% Bacto-agar gel and myelin figures were generated from crude egg
lecithin (Nutritional Biochemicals Corporation).
The RR was introduced with the fixatives
and no attempt was made to use it in physiological solutions on living tissue because
of its toxicity, although this was investigated using RV on living amoebae (Szubinska and Luft, '71). For general purposes,
the most intense reaction was also the most
useful and was achieved by exposing the
tissue to RR in both aldehyde and OsOI
fixation stages according to the schedule :
1. Fix one hour (cold or at room temperature) in the following mixture:
a. 3.6% glutaraldehyde (biological
grade) or 2.0% acrolein
b. 0.2 M cacodylate buffer, pH 7.3
c. RR stock solution (1500 ppm
in distilled water)
0.5 ml
0.5 ml
0.5 ml
2. Rinse tissues in two to three changes
of 0.15 M buffer over a ten-minute period.
3. Fix again for three hours at room
temperature in the following mixture :
a. 5% OsOl in distilled water
b. 0.2 M cacodylate buffer, pH 7.3
c. RR stock solution (1500 pprn
in distilled water)
0.5 ml
0.5 ml
0.5 ml
4. Rinse briefly with buffer, dehydrate
as usual through an ethanol series and
embed in an epoxy resin.
The RR stock solution was prepared from
commercial material by weighing out the
predetermined amount of RR (consistent
for any one batch of RR and often about
10 mgm/ml) and crushing i t in a small
mortar with a little water. The suspension
was then transferred to a centrifuge tube
and made up to the requisite volume with
distilled water, heated to about 60°C with
agitation for five minutes, cooled, centriMost of the work was done with mouse fuged, and the supernatant used as the
diaphragm, since the use of naturally thin stock solution. The concentration of RR in
tissues make less objectionable the re- the stock solution was determined by
stricted penetration of RR and avoids the spectrophotometry (Luft, '71, I ) , and it
damage introduced by cutting a larger was diluted to 1500 ppm with dstilled
water. Although there was no evidence
The xiphoid cartilage from frogs was that the concentration-of RR was critical,
also used, again because of its thinness, for this work an attempt was made to use
where cartilage was required. Mouse or the same concentration consistently. In two
rat provided the heart, intestine and lung. experiments oral epithelial cells were fixed
A very useful test tissue with which to eval- according to the schedule above, except
uate the RR reaction was obtained by that two intermediates in RR synthesis
scraping one's own oral mucosa the clumps were substituted for the RR at the same
of superficial epithelial cells being pro- concentration, one for each experiment.
The two intermediates were ruthenium
(111) hexammine trichloride and ruthenium
(111) chloropentammine dichloride, both of
which are possible impurities in commercial RR, and are discussed in the first
paper (Luft, '71, I). The tissues were embedded in Epon, cut with a diamond knife
and examined in a Siemens Elmiskop I at
60 kv, either without additional heavy
metal staining, or after brief staining using
alkaline lead solutions with or without the
uranyl acetate mordant.
cells and capillary endothelium likewise decreases continuously along certain pathways and falls to the low density normally
seen in conventional unstained sections
by electron microscopy. When one examines a new block of tissue treated with
RR, it is advantageous to look first at unstained sections (- 800 A thick) in the
electron microscope, since lead or uranyl
stains on the sections introduce density
which cannot be distinguished from that
resulting from the RR/Os04 reaction alone.
For many purposes additional staining is
unnecessary; nearly half of the electron
micrographs in this paper were taken from
unstained sections.
From the fmt, it was apparent that the
penetration of RR into the tissue was unpredictable, being dependent upon the histological structure of each particular tissue
A. General appearances
block which offered a different diffusion
4 illustrates the appearance of a
pathway into its interior. Therefore, i t
became necessary to examine first thick strong RR effect in mouse diaphragm show(2-3 p ) sections with the light microscope ing two capillaries or venules between
to determine the extent and location of the cross sections of four muscle cells and one
RR reaction and to reject those blocks unmyelinated nerve. Except for the muscle
which were useless. No stain was applied cell in the upper right corner, none of the
to these sections, since the brownish de- cells have been penetrated by RR and the
posit from the RR/Os04 reaction was more density is confined to the extracellular, parobvious without than in competition with ticularly the pericellular, material and to
a dye. This is illustrated in figures 1 and 2, collagen bundles. The muscle cell in the
which are light micrographs of a section upper right has acquired density in its myoof mouse diaphragm after RR exposure. fibrils as well as around and within its
The brownish density illustrates the extent mitochondria, which are darker in their
of RR penetration. The interior of the interior than comparable mitochondria in
block (fig. 1) is unstained, but from the the adjacent muscle cells. The various
outside RR has invaded the interstices be- features in this figure are exemplified by
tween the muscle cells to various depths. the following electron micrographs from
The dense stripes at the bottom of the sec- other sections. However, the same nerve
tion are muscle cells which were damaged is shown later in figures 44 and 45. In
(after fixation) when the block was cut figure 4 the density is concentrated around
from the hemidiaphragm and into which the cells and there is no dense structure
RR entered. Elsewhere, however, the intra- such as a gel in the spaces between the
cellular space apparently is inaccessible to cells. This extracellular space is unusually
RR. Figure 2 is an enlargement of the area large and may be artifactual; the hemiwithin the rectangle of figure 1 and illus- diaphragm was removed from the anesthetrates how the optical density generated by tized mouse and dissected into strips in
RR/Os04 continuously decreases along cer- Locke's solution with glucose for 25 mintain intercellular pathways to disappear al- utes before fixation, a situation conducive
together. Figure 3 is an electron micro- to edema. Conversely, the diaphragm in
graph a few sections away from the area figures 1-3 in which the extracellular spaces
of figure 2 , again unstained with either lead are smaller, was fixed in situ by injecting
or uranium, and showing how the mass glutaraldehyde into both pleural and peridensity as imaged in the electron micro- toneal cavities. Nevertheless, nowhere is
scope corresponds to the optical density in there evidence of a gel in intercellular
the light microscope. The heavy density space other than the coat surrounding the
seen here extracellularly between muscle various structures themselves.
B. Extracellular fibers and matrix
Collagen fibrils invariably are coated
with, or embedded in, a substance which
is strongly RR-positive and of variable
thickness. Figure 5 presents some interstitial tissue from mouse lung showing collagen fibers in several bundles embedded
in very thick material which stains strongly
with RR. The effect is that of a negative
stain which reveals the periodicity of the
individual collagen fibrils, particularly
where they are aligned. Brief staining with
alkaline lead solutions enhances the density of this coating substance so that
thinner sections are useful at higher magnification as with the collagen cross sections in figures 6 and 7. In figure 6 the
dense material can be seen to surround
each collagen fibril, thus separating each
from direct contact with others. Also, the
density of the coating material in the center
of the collagen bundle is nearly as high as
it is on the periphery, suggesting a selective
affinity of the RR for the coat material
rather than a nonspecific adsorption of
colloidal material on a surface. In the lower
left of figure 6 are seen several delicate
strands of density extending from the main
collagen bundle to adjacent structures.
Figure 7 at higher magnification shows
that the coat material exhibits a delicate
granularity in the range of 50-100 A. The
collagen fibrils occasionally show one, or
rarely, two dense dots near their centers
such as are visible in figure 6. A similar
coat surrounds elastic fibers but is not illustrated here.
Cartilage matrix provides one of the
most easily recognizable extracellular matrix materials, readily identified both chemically and histologically. Figure 8 is a thin
section of hyaline cartilage from frog's
xiphoid stained with toluidine blue to show
the typical pattern of chondrocyte pairs
embedded in their metachromatic matrix.
Figure 9 shows a control electron micrograph of a similar piece of cartilage stained
with lead. Near the perichondrium is a
pair of chondrocytes surrounded by matrix.
The cells are seen to be much darker than
the matrix. which stains feebly, if at all.
with any combination of lead or uranyl
acetate. Figure 10 is a comparison micrograph of cartilage from a similar region at
the same magnification and section thick-
ness, but which has been exposed to RR in
the fixatives. This section is unstained except for the RR/Os04 during fixation. It is
clear that the matrix is much more dense
than the cells. The RR/Os04 has generated
considerable density in the matrix material
itself as well as staining the coat around
the collagen fibers in the perichondrium.
The use of additional lead staining (similar
to fig. 9 ) only enhances the already high
density of the matrix compared with the
cell and never reverses the relative densities to that seen in figure 9 (i.e., the
matrix density in figure 10 is due to the
RR/OsO, reaction and not to any lack of
lead staining on the section).
The rectangular area in figure 10 is enlarged in figure 11 and shows that the
matrix material is coarsely granular. A
layer of RR-positive material is present at
the external surface of the chondrocyte
and several channels, containing rod-like
densities of about the same diameter as
the matrix globules, invaginate the cell in
the region of the paired centrioles (fig.
l o ) . Figure 12 is a further enlargement of
the matrix globules in figure 11. They appear to be about 200-300 A in diameter
and connected to their nearest neighbors
by delicate tapered threads which sometimes measure only about 20A in diameter
(Luft, '65b). They are similar to those
recently described using lanthanum (Khan
and Overton, '70; but see Pihl et al., '68).
Matukas et al. ('67) clearly demonstrated
these globules by conventional methods
and furthermore showed their affinity for
colloidal ferric oxide. Even in figure 11,
obtained from a section which is three
times the globule diameter in thickness, a
large open area is visible between globules.
Figure 13 is only half as thick, lead stained
and closer to focus, all of which contribute
better detail of the globules and threads.
The globules are dense but structureless
down to the 20 A intrinsic phase granularity in the picture. The globules are distorted from sphericity where the threads
insert into a broadened or conical base.
Since the section thickness here approximates the diameter of the globules, the
nearest neighbor distances are more realistic than with the superimposition of
globules in figure 12; this distance averages about 800 A center-to-center so that
the globules themselves occupy a relatively
small partial volume. However, in neither
figures 12 or 13 is there evidence of the
collagen known to be an abundant constituent of hyaline cartilage. Figure 12 is
unstained and figure 13 is stained strongly
with lead; if an adjacent section of the
same material is lightly stained first with
uranyl acetate and then with lead, many
long, striated fibrils appear between the
globules as in figure 14. These fibrils do
not have the band pattern characteristic of
normal collagen, but instead show a cross
striation, the periodicity of which is variable between 80 and 120 A. In figure 13
none of the threads can be recognized to
clearly span the distance from one globule
to the next; rather many appear to terminate midway. Both figures 13 and 14 suggest that some of the threads join the
globules to the collagen fibrils rather than
to each other.
Since cartilage matrix was a large
enough structure to be recognized unequivocally with both the light and the electron
microscopes, an experiment was devised
to elucidate the mechanism of the RR/
OsO, reaction. Frog's xiphoid cartilage
was exposed to RR in cacodylate-buffered
glutaraldehyde and then directly dehydrated and embedded in Epon with no
exposure to OsO+ Figure 15 is a thick,
unstained section of this cartilage photographed with the light microscope; the
density outlining the pairs of chondrocytes
was bright magenta in the original section,
the color arising solely from the RR bound
by the cartilage matrix. The perichondrium
lies immediately to the right of the dense
zone and although the cartilage matrix
extends off the micrograph to the left, the
RR was unable to penetrate that far. At
the upper arrow is barely visible the ghost
of a chondrocyte just beyond the boundary
of RR penetration. It is apparent that little
else in the cartilage approaches the affinity
of cartilage matrix for RR except for the
unusual staining of the nuclei of several
chondrocytes. Knowing the location of the
RR by its optical absorption in the light
microscope, i t was possible to examine an
adjacent unstained section in the electron
microscope to compare corresponding
areas. Figure 16 shows the area within
the rectangle of figure 15. Without other
stains the contrast in the electron microscope was extraordinarily low, since OsOl
was omitted in processing, so that extreme
photographic methods were used to raise
the contrast. Despite this, and even knowing where the RR must be located within
the section, there is still no recognizable
density in figure 16 corresponding to that
in figure 15. Thus the RR must be much
more effective as an optical tag than as a
mass label, a consequence of its very high
molecular extinction coefficient (Luft, '71,
I ) . Other pieces of the same cartilage
which were transferred through OsO, according to the standard procedure showed
brown deposits in the light microscope
where the matrix had been penetrated by
the RR, and in the electron microscope the
same brown areas appeared similar to figures 10 and 11 (Luft, '65b).
Agar gel was examined after RR treatment as a model substance possessing certain similarities to cartilage matrix, i.e.,
being a classical gel with the agaropectin
component a highly sulfated polysaccharide. Figure 17 shows the branching filaments seen throughout these sections, with
figure 18 being an enlargement of the area
within the rectangle of figure 17. The
branching filaments are thin, about 20 A
in diameter, and interspersed with occasional dense granules about 100 A in
C. Cell walls and extraneous coats
1. Celt-collagen junction. Figures 5-7
illustrate that collagen fibrils are coated
with a material which has a strong affinity
for RR. An interesting region should be
found where collagen fibrils make contact
with cells and particularly at the myotendinal junction. Figure 19 shows an unstained cross section of a striated muscle
cell with an RR-positive region at the cell
surface. Collagen fibrils can be seen in
negative contrast in this region, but this
can hardly be called a myotendinal junction since the collagen fibrils are oriented
circumferentially. Instead this collagen
must be part of the sarcolemmal reinforcing fibrils. Figure 20 shows the region in
the rectangle at high magnification still
unstained except for RR/OsOa. A single
collagen fibril stands out in bold relief
against the density which not only coats
but seems partly to infiltrate it. It appears
that the whiter major bands of the fibrils
have a larger diameter than the interband
regions into which the RR-positive material partially penetrates, a configuration
which should strongly anchor the collagen
to the surrounding dense material. This
dense material has an irregular surface
facing the tissue space toward the upper
left, but it terminates abruptly against the
plasma membrane of the muscle cell at the
lower right. At the interface the trilaminar
structure of the plasma membrane can be
recognized; the thin dark line (arrow) is
the inner or cytoplasmic leaflet of the unit
membrane, with the central light space
likewise being visible. The outer leaflet is
not visible as a discrete layer because the
dense material reaches and obliterates it.
This dense material is not the basement membrane or basal lamina; the
basal lamina is seen only in negative image by displacing the RR-positive
material, if i t is visible at all. In figure 20
there is a zone of somewhat lower density
intermediate between the collagen fibrils
and the plasma membrane which corresponds to the basal lamina. It is seen more
clearly in figure 24. Figure 21 depicts a
circumferential band of collagen fibrils
around another heavily stained muscle cell.
Between individual collagen fibrils are
seen globules of RR-positive material
which are periodically positioned on the
band pattern of the collagen. Because of
this periodicity, adjacent collagen fibrils
must likewise lie with their bands in register. The outer (upper) and inner boundaries of the collagen bundle are encased in
dense RR-positive material, the inner layer
again being continuous with the outer leaflet of the unit membrane which is barely
visible at this low magnification.
2. Myotendinal junction. The myotendinal junction, where the tension developed by the muscle cell is transferred to
the collagen fibers of tendon, is a rewarding
region to search for an extensive development of the interfaces illustrated in figures
19-21. Figure 22 illustrates a myotendinal
junction from mouse diaphragm with conventional methods using heavy metal
(uranyl/lead) staining. The collagen is a
light gray; a darker line is seen around the
cells which at higher magnification (not
illustrated) turns out to be the plasma
membrane and basal lamina of the muscle
cell. Figure 23 is a similar preparation,
but exposed to RR in the fixative and
lightly stained with lead. Here it can be
seen that the collagen fibers are infiltrated
as before with RR-positive material and
that there is a uniform layer of dense material everywhere interspersed between the
muscle cytoplasm and the organized collagen bundles. The area within the rectangle is enlarged in figure 24 to show the
collagen fibrils embedded in RR-positive
material which accumulates adjacent to
and is confluent with the external leaflet
of the unit membrane of the muscle cell.
In this thin section (figs. 23, 24) the contrast is not as high as in figures 19 and 20,
but the thinness does enable the basal lamina to be recognized as an irregular light
zone at the expected distance from the
plasma membrane of the muscle cell. The
basal lamina appears here as negative
density (the light zone) because it itself is
RR-negative and displaces an equivalent
volume of RR-positive material.
Here it is appropriate to elaborate further on the effect of section thickness just
mentioned above. This is illustrated in figure 25, which shows the periodic sinusoidal
variation in section thickness familiarly
known as “chatter” or “wash-boarding.”
The average section thickness is about 500
A, probably varying from about 200-800
A. The apparent variation in amount of
RR-positive material along any particular
extracellular space as well as apparent
variations in the structure of the material
must in reality be due only to variations in
thickness of the section, since the section
itself is unstained and the RR reaction was
completed in the block even before it was
embedded. Very thin sections, on the other
hand, can be usefully enhanced by heavy
staining to attain high resolution. Figure 26
is a 300 A thick section of heart muscle
stained an hour each with aqueous uranyl
acetate and Reynolds’ lead solution. Here
i t can be seen that the outer leaflet of the
plasma membrane is more dense after exposure to RR because of a substance closely
applied to or confluent with it. There are
patches here and there (arrow) where the
outer leaflet is bare and the two leaflets can
be seen to have symmetrical density. This
heart muscle cell showed a similar pattern
over a large area.
3. Desmosomes. The desmosome presents an interesting specialization of cell
coat material with which to test the RR
reaction. Many epithelia have desmosomes
but epithelia lining viscera often have tight
junctions which block effectively the penetration of RR into the intercellular space
from the lumen, so that desmosomes are
rarely labeled. Epithelial cells from the
oral mucosa, however, present an abundance of desmosomes which are easily
penetrated by RR. Their instant and universal availability, coupled with the assortment of cell surfaces which they present
and the reproducibility of their RR reaction, make them nearly ideal test objects
for checking the RR reaction with any
particular batch or fraction of RR. Figure
27 shows three superficial epithelial cells,
with a bacterium encased in a thick RRpositive density which is absent in controls
omitting the RR. The desmosomes have
distinct intercellular density even at this
low magnification. The interior of the external cell is quite pale, whereas the deeper
cell reveals the interior full of gray tonofilaments. The intracellular plaques at the
sites of the desmosomes (arrows), while
roughly symmetrical in size, are unequal
in density, being consistently darker in the
darker cell. Figure 28 illustrates the diversity of cell coats and cell wall materials
found in these oral epithelial cell preparations. The epithelial cell itself shows the
thin layer of RR-positive material already
demonstrated to be attached to the outer
leaflet of the unit membrane along with
shreds and strands of similar material projecting from the surface and toward adjacent cells. Three bacteria are visible, two
of which have a similar compact cell coat
and one having an extended filamentous
coat. The usefulness of RR for study of
bacteria has already been demonstrated
(Pate and Ordal, '67; Jones et al., '69).
Returning to the desmosome, figure 29
shows that the layer of density is confined
to the extracellular material. This extends
from the outer leaflet of each unit membrane to fill the intercellular space at the
desmosome. In some cases the density decreases toward the center between the two
halves of the desmosome leaving a dis-
cernible lighter band or zone. Figure 30
shows a similar desmosome at higher magnification with the benefit of light lead
staining. The central light layer of the
unit membrane is visible in both participating cells but is clearer on the left side.
External to it lies a layer of high density,
about 50-60 A thick which includes and
masks the location of the outer leaflet of
the unit membrane, and which spreads
laterally over the cell surface without a
break from the desmosome proper. A mirror-image of this density is visible on the
other cell of the pair, with the spacing of
about 150 A between them being occupied
by similar material of lower density in
which a suggestion of pillars or strands
can be made out. Figure 31 shows a section near the edge of the desmosome which
illustrates these features more clearly.
Note the asymmetry of the intracellular
desmosomal plaques in the two cells; again
there is a significant difference in the
degree to which the plaques and tonofilaments have taken up the lead with which
the section was stained. This difference in
density is noticeable also in figure 32 which
shows desmosomes connecting three epithelial cells from human oral mucosa. The
left-hand cell is quite light in contrast to
the other two and permits comparison between the asymmetrically stained desmosome in the upper left with the symmetrical
one in the lower right. In the upper right
is a desmosome which appears to have
been pulled apart somewhat. The lower
right desmosome has an intercellular segment of normal density and extent. A different desmosome configuration is shown
in figure 33.
The usefulness of oral epithelial cells as
test objects can be judged from figure 34.
An attempt was made during purification
of RR from commercial sources to identify
certain impurities (Luft, '71, I) and to
test them in pure form instead of RR itself.
One of the intermediates in the formation
of RR from ruthenium trichloride is ruthenium (111) hexammine trichloride and from
this ruthenium (111) chloropentammine dichloride can be synthesized easily (Sidgwick, '50, V. 11, p. 1469). Both the hexammine and chloropentammine convert to
ruthenium red with aging in aqueous ammonia, more rapidly with the pentammine.
Likewise both of these pure compounds
used in the fixatives in place of RR give
the same general density distribution but
more weakly than that expected with genuine RR, and, of the two, the chloropentammine gives the more intense reaction. Figure 34 is a desmosome between oral epithelial cells exposed to the chloropentammine,
the section being lightly stained with lead.
The density distribution is familar although
there is less differentiation of the parts of
the intercellular desmosomal components
compared to figure 30 for example. Occasinally, a quite different pattern is found
following exposure to RR as indicated in
figure 35. The section of frog gut was very
lightly stained with lead, but the density
distribution is mainly due to the RR reaction. The intercellular material is lightly
stained showing multiple filaments extending across the space and most of the
density is developed in the plasma membrane and an adjacent layer which is parallel to the plasma membrane but located
about 300 A deeper in the cytoplasm. There
is no obvious equivalent structure in the
cheek cell desmosomes. The explanation
of this difference is unknown, but a desmosome from newt skin showing in fine
detail a similar density distribution has
been published (Kelly, ’66).
4. Intestinal microvilli. The external
surface of intestinal microvilli has a characteristic thick layer of material picturesquely called a “fuzz” layer because of the
closely packed and matted filaments which
are easily recognized by conventional staining procedures (Ito, ’65). Ruthenium red
also stains the fuzz filaments, and at low
magnification in unstained sections the
density produced by RR in the fuzz layer
contrasts strongly with the virtually unstained cytoplasm (fig. 36). In figure 37
the unit membrane is visible, revealing
that the fuzz layer (which is thin in frog
intestine) is confluent with and extends
from the outer leaflet of the unit. The mucous coat which is secreted by the goblet
cells of the intestine is also RR-positive and
extends as a layer over the tips of the
microvilli. It has a coarse filamentous texture which is quite distinct from the fuzz
of the microvilli. This can be seen in
figure 36 and to better advantage in figure
The rat has a thicker fuzz layer and a
thick section, such as that in figure 38
shows an impressive zone of density in
which the cytoplasmic cores of the microvilli stand out as if negatively stained. The
thick layer of mucus seen here appears
more like an irregular aggregate of flocculent material than as a single layer. Of
particular interest is the density in the
apical cytoplasm which occurs as densely
lined vacuoles, one of which is continuous
with the surface through a tubule originating at the base of the microvilli. It would
appear that the RR has penetrated between
the microvilli to gain access to this tubulevacuole system. It is probable that the
other vacuoles which are likewise labeled
also were connected to the surface through
similar tubules which were not included
in the section, and conversely, that the
other pale vacuoles and vesicles were isolated from the exterior when the tissue
was fixed.
Although not illustrated, it is appropriate
to comment here on the goblet cells found
in the intestinal epithelium. Their mucous
granules do not take up RR in most instances even though these granules may
be very close to the cell surface. An intact
plasma membrane is sufficient to prevent
their reaction with RR, although once released their contents stain avidly. In this
respect, goblet cells behave quite differently from mast cells (see page 379).
5. Red blood cells and vascular endothelium. Erythrocytes bind RR as a thin
layer along their plasma membrane (figs.
39, 40), and again this density is located
against the outer leaflet of the unit membrane (fig. 41). In figure 40 the plasma
membrane of the upper RBC is cut nearly
perpendicular to the section and the layer
is seen to be thin and dense; the lower
RBC is cut obliquely resulting in the density
of the layer projecting over a larger area
with correspondingly lower density. In the
oblique cut the density is not uniform but
reveals an irregular texture of delicate interconnecting filaments resembling tissue
paper. This suggests that the layer is discontinuous or porous in the size range below 200-400 A. In figure 41 the RBC is
close to the vascular endothelium whose
dense layer of RR-positive material is attached to and extends 100-200 A from the
outer leaflet of the unit membrane into
the capillary lumen. Several strands extend 500-1000 A from the endothelial RRpositive cell coat to that of the RBC. These
appear to have been pulled out from the
two cells following momentary contact.
At X is a region where the strand connects
a conical projection of the RBC to a clump
of material on the endothelial coat and the
appearance is compatible with the assumption of tension between the two cells. If
so, the phenomenon is of obvious relevance
to the problem of “sticky” cells or “sticky”
endothelium in microcirculatory physiology
(Grant, ’65).
The appearance of vascular endothelium
after exposure to RR has already been described (Luft, ’66c; but see Copley and
Scheinthal, ’70). The RR-positive layer
which lines the luminal surface of the
endothelial cells was identified (Luft, ’66c)
as the “endocapillary layer.” This layer had
been described and named 25 years earlier
on the basis of physiological and microsurgical methods (Chambers and Zweifach,
’40, ’47). It is illustrated in figure 42
(ECL). The endocapillary layer is confluent
with the outer leaflet of the unit membrane
and extends several hundred A into the
lumen of the capillary. The inner boundary of the layer is the outer leaflet of the
unit membrane itself, but the luminal
boundary is indeterminate since the density becomes ragged and diluted toward
the lumen. The particular termination of
the layer in figure 42 is an accident of
photographic printing since a lighter or
darker print would have shown the layer
apparently narrower or wider respectively.
Instead of a physical boundary, there is a
gradient in the concentration of the endocapillary layer decreasing nearly to zero
somewhere in the lumen. There is a similar layer on the connective tissue side of
the endothelial cell, and it is heavily stained
as shown in figure 42. Both this layer and
the endocapillary layer meet at the lateral
surfaces of the endothelial cells where the
cells join their neighbors to form a simple
squamous epithelium. These lateral contact zones figure importantly in the various
hypotheses of capillary permeability, and
it is here also that a distinctive modification
known as a “tight junction,” or “close junction” is found. This region is illustrated in
figure 43. Here two endothelial cells make
side-to-side contact with RR penetrating
from the direction of the connective tissue.
The RR has labeled the extracellular space
which is normally about 150 A wide ( A )
but which abruptly narrows at two successive points (arrows). At these points, the
extracellular space is reduced to a slit about
30-40 A wide. Some density attributable
to RR has penetrated beyond the first slit,
but little or none is detectable past the second slit or in the lumen. The failure of the
RR to reach the lumen and to label the
endocapillary layer here may be due to
insufficient time or concentration of the
RR to permit full penetration. The relationship of these observations to capillary
physiology is discussed elsewhere (Luft,
’65a; Karnovsky, ’68).
6. Nerve fibers. Neurophysiology is
concerned with ion fluxes between intraand extracellular spaces as well as the
accessibility of ions to the spaces around
and between axons and Schwann cells.
Because RR itself is a polyvalent cation it
m a y serve as a useful tracer for cations.
Figure 44 shows an unmyelinated nerve
(same nerve as in fig. 4 ) after exposure
to RR. Although the nerve is heavily reinforced with collagen, RR has penetrated
around the Schwann cell and has been
bound to a thicker layer of high affinity
material there. It has also penetrated the
spaces between the Schwann cell and the
axons embedded in its periphery. The area
within the rectangle is enlarged in figure 45
where, in some regions at least, the components of the unit membranes of both the
axon and the Schwann cell are resolved.
From previous examples one might expect
that the RR-positive extracellular density
would extend from the outer leaflet of the
unit membrane of the Schwann cell to the
outer leaflet of the unit membrane of the
axon, a distance of 100-150 Angstroms.
Figure 45 indicates that this is, in fact,
the case. The situation is less simple for
myelinated nerves. Figure 46 shows a
branch of the phrenic nerve embedded in
muscle which was exposed to RR. The
collagen bundles in the space around the
muscle cells show the typical reaction,
but RR has been stopped abruptly at a
boundary surrounding the myelinated
axons and has been unable to penetrate
to the Schwann cells or the collagen which
surrounds them. This boundary is the
layer of epithelioid cells associated with the
perineurium and has been discussed elsewhere (Burkel, '67). However, if this protecting layer is damaged and its integrity
breached, then RR is able to invade this
compartment like any other intercellular
space (Luft, '66a). This is illustrated in
figure 47; nerve damage has occurred as
exemplified by the disorganization of the
myelin. Although the RR effect still is
strongest outside the perineural layer, some
RR has entered the endoneural space to
stain the collagen bundles which reinforce
the Schwann cells. The site of entry is
unknown, but the cytoplasm of the perineural fibroblasts is unusually dense, suggesting cellular damage. The RR may, in
fact, have had gradual access everywhere
once the cells no longer functioned as a
barrier. While the penetration of RR is
arrested by the perineurium, and even
when this is breached, RR is unable to
penetrate the myelin lamellae, lanthanum
oxide sols penetrate both with ease and
into the innermost turn of the myelin wrapping (Revel and Hamilton, '69).
7. T-system of striated muscle. A final
example of the interaction of RR with cell
walls and extraneous coats is exhibited by
the T-tubules, which penetrate centripetally
in striated muscle and are known to be
a highly specialized derivative of the
plasma membrane (Franzini-Armstrong
and Porter, '64; Huxley, '64). When the
staining is successful, the density produced
by the RR reaction is found in either the
T-tubules or in the adjacent sarcoplasmic
sacs of the triads, but never in both at the
same time. Figure 48 shows a large region
of diaphragmatic muscle in longitudinal
sections in which density has been generated in nearly every available pair of sarcoplasmic sacs leaving the pale T-tubule in
the center of each triad. Contrast this with
figures 49 and 50 in which the T-tubule is
dense while the adjacent sarcoplasmic sacs
are completely negative. In figure 50 it is
seen clearly that the density within the Ttubule fills the interior out to the equivalent
of the external leaflet of the unit membrane
from which the T-tubule is derived. The
density within the sacs of sarcoplasmic
reticulum is seen in figure 51. Although
the density is not so high and therefore not
so self-evident as in figure 50, the comparison of figure 51 with its control in figure 52
makes it clear that RR has penetrated into
the sarcoplasmic sacs. This has been presented previously in some detail (Luft,
'66b). It is not certain why RR is found
in one or the other location but not both.
Frequently, however, in those micrographs
in which RR is localized to the T-tubule,
the muscle cells are in the extended state,
whereas in micrographs showing RR in the
sarcoplasmic sacs, the muscle is contracted
with little or no I-band visible.
D. Intracellular penetration. So far the
emphasis of this report has been on the
apparent exclusion of RR from the interior
of cells and its affinity for materials in
those extracellular spaces which are accessible without traversing any membranes.
This manifests itself in micrographs as
pale cytoplasm with density confined to
the outside of the cells. Several exceptions
have been mentioned, however, such as
the upper right muscle cell in figure 4, the
nuclei of chondrocytes in figure 15, the
oral epithelial cells in figures 27, 31, 32;
the endothelial cell in figure 42 and the
sheath cells in figure 47. The appearance
of density in the sarcoplasmic sacs of the
triad in figures 48 and 51 may be another
instance. Except for the epithelial cells,
the nuclei, and the triads, mechanical damage to the cells during fixation is the
most probable cause of this intracellular
There is one instance where RR routinely
appears to penetrate intact cell membranes,
this being in the case of the mast cell.
Figures 53 and 54 show low and high
magnification light micrographs of mast
cells on the peritoneal surface of mouse
diaphragm in a whole mount without sectioning. The diaphragm was fked in situ
in glutaraldehyde containing RR and
mounted for photography without exposure
to 0~01. The mast cell granules were
stained deep red and photographed dark,
with nothing else in the tissue taking up
RR to the same degree. In figure 54 one
can see the nucleus of each mast cell
which is outlined as an unstained oval by
the mast cell granules which it displaces.
In an electron micrograph such as figure
55 the mast cell granules are dense but
not as dense as would be anticipated in
view of their high affinity for RR in the
fixative. Compare their density, for example, with the muscle cell in the lower right
corner of figure 55; there the plasma membrane was damaged, admitting RR to the
interior with intense staining of the myofibrils and material around the mitochondria. The cytoplasm in which the mast
cell granules are embedded is also much
less dense than that of the muscle cell, if
it is stained at all.
Another example of cytoplasmic penetration after mechanical damage is shown
in figure 56. This is a myoneural junction
from mouse diaphragm in which the muscle cell again was labeled internally from
a distant inury, and the RR was unable to
penetrate the plasma membrane of the
muscle cell from inside outward to stain
the plasma membrane of the nerve. Note
the absence of stain in the material in the
synaptic clefts between these two cells.
This cleft material normally stains intensely when RR is admitted by the usual
extracellular pathway (Kelly, D., '67).
E. Lipids. Evidence has accumulated
that binding of RR not only is related to
the acidic mucopolysaccharides (as previous pictures have emphasized), but also
to materials or sites known to be rich in
lipids as the following observations illustrate.
At first glance, it appears that ruthenium
red fails to penetrate the cytoplasm of most
cells and that its penetration into even a
small (?42 mm) block of tissue is restricted
to the outer 100 p or so. This point was
made early in the paper and was illustrated
in figures 1-4. When a cell is damaged
near the periphery of the tissue block, the
intracellular density is so distinct and so
uniformly distributed within the cell that
this exception proves the rule, particularly
in low magnification, large area pictures of
unstained sections. However, peculiar densities have been encountered in lipidcontaining structures within undamaged
cells even near the center of tissue blocks
such as those from which figures 1 and 2
were taken. These densities at first were
attributed to the prolonged exposure to
OsOc ( 3 hours at room temperature) but
controls omitting RR do not show these
densities. They appear as dense margins
in lipid droplets, as localized enhancement
of the dense layers in myelinated nerve,
and as increased density with consequent
great visibility of the unit membrane in
mitochondria, particularly after staining
the sections. Some of these dense deposits
associated with lipid droplets can be seen
in the otherwise pale cytoplasm in figures
3, 1 9 , 2 3 , 4 6 , 5 5 and 56. Particularly good
examples are seen among the mitochondria
in figures 55 and 56 and near the axon in
figure 56. In figure 46, while the myelin
sheath appears uniformly pale grey in
most fibers, there is a knuckle of myelin
intruding toward the axoplasm in the third
nerve fiber from the top. On close inspection unusual densities are visible here,
which, although faint in comparison with
the extracellular densities in this figure,
are none the less absent from the rest of
the myelin.
Figures 57 and 58 illustrate similar regions in myelin, one examined without
uranyl or lead staining (fig. 57), and the
other a thinner section examined after light
lead staining (fig. 58). A particularly significant feature is that in figure 57, the
myelin is stained especially heavily where
there has been obvious disorganization.
There is some stain in the major dense
lines generally, but these are seen to expand into large, densely stained segments
where the disorder is the greatest and converge again to a line of normal width and
density. There is no obvious path of stain
density to the surface either here or in
figure 46. In figure 58 there is little evidence of irregularity in the myelin layers,
but rather local widening of the major
dense line in short segments.
A more obvious association with lipid is
evident in figures 59 and 60 taken from
the pale interior of a block of diaphragmatic muscle possessing, among its mitochondria, lipid droplets similar to those in
the upper right muscle cell in figure 55.
In figure 59 the lipid droplets show a lamination with slit-like densities which are not
seen in the RR-free controls. Even without
additional stain, multiple concentric lamellae with lenticular densities inserted in the
dark layers are visible in figure 60. These
are similar to figures 57 and 58 of nerve
myelin. However, the spacing of these
layers is about 42-45 A, much smaller than
the 165 A spacing measured for the myelin
in figure 58.
Figures 61 and 62 illustrate the enhanced contrast observed in mitochondrial
membranes after using RR. As can be seen
in figure 61, the plasma membrane of this
particular muscle cell was damaged, admitting RR to the interior of the cytoplasm.
The same enhancement, but less intense,
takes place in noninjured cells. It is also
apparent that the external mitochondrial
membrane is at least a partial barrier to
RR penetration because most of the mitochondria are relatively pale in comparison
to the two darker mitochondria, one of
which ( X ) can be seen to have a broken
limiting membrane. The area within the
rectangle is enlarged for figure 62, showing increased density in the membranes of
both the cristae and the external membranes. The leaflet of the unit membrane
which faces the matrix is considerably
lighter than the leaflet which faces the
cristal cavity, and the center-to-center
spacing measures about 50-55 A.
In order to verify the reaction of RR with
lipids an experiment was carried out on a
model system. Crude egg lecithin was exposed for an hour to dilute cacodylate
buffer and permitted to generate myelin
figures. (It had been established in the previous paper (Luft, '71, I ) that these myelin
figures are strongly birefringent in the
polarizing microscope, that they bind RR,
and that the bound RR is faintly dichroic.)
These myelin figures were treated as a
tissue block, fixed in glutaraldehyde followed by OsOl, both containing RR, and
embedded and sectioned in the usual manner. The result is shown in figures 63 and
64. Figure 63 shows cross sections of several myelin figures with the multiple concentric lamellae typical of these figures.
It is obvious that the RR is bound to the
outside with little or no penetration from
one layer to the next, and that appreciable
density is generated with this crude lecithin
as well as with components in extracellular
tissue spaces. Figure 64 is an enlargement
of figure 63 taken from a region at the
right edge near the tip of the arrow. In
figure 64 many leaflets can be seen. Each
displays a pale core and a flocculent or
amorphous dense material around the periphery which serves to outline the core.
38 1
The thickness of these cores is difficult to
establish accurately but they appear to be
about 25-30 A in this micrograph.
A disadvantage to the ruthenium red
(RR) method as employed here is the requirement that it be carried out on a block
of tissue during fixation. Although there
are certain advantages which accrue from
this procedure, for some purposes i t would
have been better if the method could have
been applied to conventionally prepared
sections mounted on grids. Such attempts
on epoxy-embedded tissue have been unsuccessful (however, see Gustafson and
Pihl, '67a). Furthermore, the dye penetrates
the tissue block from the outside following
a tortuous path toward the interior but
producing a useful reaction in the extracellular spaces only in the superficial 100 p
or so, and much less in some tissues. For
reasons discussed previously (Luft, '71, I),
attempts to produce thorough penetration
by vascular perfusion have likewise failed.
Thus the method is unpredictable in that
it depends upon histological variations in
each block which are unknown in detail
until after the reaction has occurred. However, because the reaction product, which
is dense in the electron microscope is likewise dark in the light microscope, it is
possible to survey and discard useless material before valuable electron microscope
time is wasted. The reason for slow penetration is uncertain but i t may result from
the following consideration. If RR penetration is likened to a chromatography experiment, the extracellular spaces become
the column and the fixative mixture becomes the solvent. As solvent plus dye (RR)
diffuse together toward the center of the
block, the dye will be retarded to the degree that it is adsorbed to the column
packing material. If the RR is strongly
bound to the extracellular acidic polysaccharides, then poor penetration is a necessary consequence of the very feature which
produces selective labeling in the first
place. Although poor penetrability seems
apparent at dimensions of the tissue block,
the fact that RR can penetrate the microvilli of the gut to trace out blind channels
(fig. 38) and even the T-tubules in striated
muscle (figs. 49, 50) indicates that its
penetrating ability is surprisingly high at
micron dimensions.
haze. Figure 20 also shows that the dense
material partially infiltrates the collagen
fibril, particularly in the interband regions,
but not at the major bands where the diameter of the fibril appears larger. Thus
Association with mechanical function
the collagen fibril seems to be anchored
One of the generalizations which can be in the dense material in the same way
drawn from the pictures presented here is that the series of flanges on rods of reinthat many of the sites of RR localization forcing steel anchor it into concrete. The
are the same as those at which substantial similarity may extend from the structural
mechanical forces are known to exist or to the functional level if, from consideraare suspected to occur, such as in collagen tions of the mechanism by which strength
fibers generally (and at the myotendinal is conferred in composite materials, the
junction particularly), at desmosomes and reinforcing steel in concrete and the colin cartilage. It is known that the striated lagen fibrils in the dense substance respecmuscle cell develops tension internally tively play the part of high modulus rods
when stimulated. Regardless of how the embedded in a low modulus matrix (Mortension is generated, it is necessary to ley, '66; Kelley, A., '67). Although undeliver it to extracellular components, such proven by the micrographs shown here, i t
as tendon, cartilage or bone in order to is implied that the dense material is the
produce useful movement. Overlooking the element which couples the outer leaflet of
enigma of how tension is transmitted from the plasma membrane of the muscle cell
the myofiaments across the plasma mem- to the collagen fibrils of the sarcolemma
brane, but accepting that this occurs, the and tendon. As a corollary it may be worth
next question is the mode of attachment considering on the basis of symmetry
of the plasma membrane at the end of whether the myofilaments of muscle or the
the muscle cell to the collagen fibrils which tonofdaments of desmosomes might not be
eventually condense to tendon. Previous similarly anchored to the inner leaflet of
work on the fine structure of the myotendi- the unit membrane by some analogous innal junction (Muir, '61) shows little more tracellular cement substance (see Douglas,
than that visible in figure 22; the collagen Ripley and Ellis, '70).
Related to the RR-positive material
fibrils come close to, but do not insert into,
the plasma membrane of the muscle cell. which is found in the myotendinal juncWhatever is between the cell and the col- tion is the density distributed between
lagen fibril does not stain well or is poorly collagen fibrils in tendon. If strength is
preserved by conventional techniques. an important characteristic of the subHowever, after exposure to RR, a layer of stance at the junction, then for consistency
dense material is seen to be interposed be- of argument i t should likewise cement
tween the collagen and the muscle cell. collagen fibrils together in tendon, or at
Furthermore, this dense substance seems to least sustain a shearing force between
be continuous down to, and confluent with them. It was noted earlier that figure 21
or fused to, the outer leaflet of the unit showed globules of RR-positive material
membrane surrounding the muscle cell. In periodically intercalated between adjacent
the opposite direction the same density is collagen fibrils with a period of 610 A.
continuous to and appears to surround or Furthermore, the densities were associated
even embed (fig. 20) the collagen fibril with the interband region, sometimes visitself. The distribution of density around ible as dense slits between two collagen
collagen fibrils and in the intercellular fibrils and at other times giving the apspaces has been clearly described recently pearance of a ring or bracelet around the
using RR by Highton, Myers and Rayns fibril. Regardless of the exact geometry, a
('68). Of course, the mechanical properties direct association of dense material with
of this material which appears dense with the interband region seems assured, and
KR are unknown, but it is more satisfying if two collagen fibrils share between them
to attribute shear strength to a substance one density, and if the dense substance is
having this density than to an ethereal strong, then this is sufficient to promote
registration between the fibrils with quasicrystallinity over many fibrils. Again considering composite materials, the similarities suggest that tendon may derive
distinct advantages and strength from this
arrangement. It would antagonize the
propagation of fracture across the tendon
from a defective segment of a collagen
fibril and thus help to achieve in bulk
tendon nearly the same strength per unit
cross sectional area as in a single defectfree collagen fibril (i.e., a "whisker") (Gordon, '64; Morley, '66; Kelly, A., '67).
Turning to the desmosome, we encounter
a quite different but mechanically important structure which is an epithelial derivative. Nevertheless, figures 27-35 indicate that RR-positive material is involved
because of the density which appears at
the site of the desmosome. This density
occupies part of the intervening space between the two cells, and is delimited from
the cytoplasm of each cell by the middle
(light) leaflet of the unit membrane. The
density thus extends from, and is continuous with, the outer leaflet of the unit
membrane of one cell to that of its partner.
The layer of external density extends no
farther over the outer leaflet of the plasma
membrane than the extent of the desmosomal cytoplasmic plaques along the internal leaflet of the membrane. There is
no sign of the lamellae which have been
shown repeatedly in epidermal desmosomes
(Odland, '58; Farquhar and Palade, '65;
Stern, '65; Kelly, '66; Douglas, Ripley and
Ellis, '70). If there is any correlation in
intercellular density patterns between classical and RR-labeled desmosomes, the density distribution exemplified by figure 30
is the inverse of that in the literature. A
comparison between the conventional and
RR-treated desmosome of newt is available
in the paper of Kelly (Kelly, D., '66) in
his figures 7 and 10. He suggests that the
midplane density may result from overlap
of extracellular elements from each side.
Except for this, there is no obvious explanation of the discrepancy. but in figures
20 and 24 it was noted that the basal
lamina revealed itself as a negative image
in the layer of RR-positive material at the
surface of the muscle cell. These observations imply that materials with different
staining characteristics may occupy inter-
penetrating but mutually exclusive domains at cell surfaces. The structures enhanced by conventional staining methods
(uranyl/lead) appear to form a fabric
50-100 A away from the trilaminar plasma
membrane, but with intermittent anchoring to its outer leaflet. Because of their
staining characteristics, they are suspected
by the author to be proteinaceous and are
represented here by the basal lamina and
the intercellular components of desmosomes described in the literature (Farquhar and Palade, '65). Between this fabric
and the unit membrane lies the space occupied by the RR-positive material. This
material appears to be a gel in intimate
contact with the outer leaflet, and may be
acid-substituted polysaccharide associated
with protein. The mechanical properties of
this substance may be quite variable, from
a slippery lubricant to a tough and rubbery or even brittle material, depending
both on its molecular constitution as well
as the amount of intermolecular crosslinking (Kerr, '63; Krizek, '68; Lake and
Thomas, '67; Oppenlander, '68; Baier,
Shafrin and Zisman, '68). Particularly
with the acid polysaccharides, the crosslinking is easily regulated by calcium
(Bonner, '36; Preston, '52; Chambers and
Chambers, '61, pp. 57-62; Rasmussen, '67;
Armstrong and Jones, '68; Woodward and
Davidson, '68) so that the mechanical
properties of the gel potentially are readily
variable from point-to-point in space along
the cell surface, as well as from momentto-moment in time at the discretion of the
Figures 27-35 suggest that the desmosomal intercellular density is in fact a
thickening of the preexisting layer of RRpositive material which everywhere covers
the cell. Thus the desmosome would be
merely quantitatively but not qualitatively
different from the dense layer which envelops the epidermal cell and which itself
appears as though it were a thickening of
the outer leaflet of the unit membrane
when stained by RR/Os04. The bacteria
which adhere to this layer (figs. 27, 28)
do so with a layer of material which likewise enshrouds the outer leafiet of their
own unit membrane and which is RR positive. The strength of this layer is well docu-
mented in the literature on protoplasts
(Pate and Ordal, ’67; Salton, ’64).
The mutual adherence of these bacteria
and epithelial cells probably is more than
accidental and is indicative of surface
similarity (Gray, ’64). In fact, the bacterial cell wall may be the best model at
the moment for the desmosome layers.
Since it is known that the glycosaminopeptide (mucopeptide) which confers the
strength to the bacterial wall may represent only 10-20% of the wall mass (Salton, ’64, p. 100; Sharon, ’69), the extrapolation of this concept to the desmosome
may harmonize the discrepancy between
its appearance after RR compared with
uranyl/lead staining. Expressed differently, the strengthening component may
leave room in the desmosome for other
materials with different functions.
The surfaces of cells known to be involved in the uptake of dissolved material,
as typified by intestinal epithelial cells
and cells of kidney tubules, also are reactive toward RR. The microvilli of intestinal
epithelial cells possess a “fuzz” which
stains selectively with RR (figs. 36-38),
and the microvilli of the proximal convoluted tubule of the kidney show a similar
phenomenon (Groniowski et al., ’69). The
function of this elaboration of the outer
leaflet of the plasma membrane is not
known. The affinity of RR in OsO, for this
“fuzz” may be related in some way to its
staining characteristics with OsOl in the
presence of phosphate buffer, as demonstrated by Pratt and Napolitano (’69), in
which a polar environment was shown to
be necessary for osmium binding to the
fuzz. Wissig and Graney (’68) have clearly
shown the remarkable pattern which replaces the fuzz in the “apical endocytic
complex” of suckling rats. Again this surface specialization seems to be related to
selective uptake of certain substances (proteins). There is no doubt that this complex
also binds RR as shown by the increased
density of the tubules in figure 38. Nevertheless, there still is no obvious explanation for the association of a specialized
absorbing surface and RR binding.
Several recent reports have documented
the value of RR in visualizing the changes
in the surface layers of cultured cells before and after their transformation by on-
cogenic viruses (Martinez-Palomo and
Brailovsky, ’68; Morgan, ’68). In the first
report, an association was found between
loss of contact inhibition shown by the
transformed cells and a thickening of their
surface layers which showed an affinity for
RR. The second paper, which dealt with
different cells and viruses, again described
an increased thickness of the surface layers
after viral infection, but the thickness
could not be correlated with loss of contact inhibition. Nevertheless, these reports
show clearly that the structure of the surface coat of cultured cells as stained by RR
is sensitive to changes in cell function
which are associated with cancer.
Cartilage is a material having unique
mechanical properties well known to biologists, if not all of us carnivores, and
which recently have been determined quantitatively (Edwards, ’67). It is the matrix
which has these properties, and it is the
matrix which stains strongly with RR
(figs. 8-16). Cartilage is another example
of a substance of great mechanical importance in biology which is poorly visualized by conventional EM preparation
methods. The collagen fibrils within hyaline cartilage have been known for many
years; but although the collagen is necessary for the over-all properties of the cartilage, it is not sufficient itsel€ to provide
such properties, and the chondroitin-sulfate-protein complex is also necessary
(Tsaltas and Greenawald, ’66). Figures
12-14 indicate that cartilage matrix is a
complex gel with at least three structurally
distinct components (globules, threads and
collagen filaments) in addition to the fluid
phase. This porous gel with dimensions
indicated earlier (page 373) is quite consistent with the models of cartilage which
are emerging from chemical, physicochemical and bioengineering studies (Partridge,
’68; Maroudas, ’68; Edwards, ’67).
To return to the generalization which
introduced the section, it seems that some
of the RR-positive materials in the extracellular compartment have substantial mechanical strength. An attempt was made
to justify this contention in the case of
the myotendinal junction, tendons, desmosomes and cartilage. Mucus is another RRpositive substance which is weak rather
than strong, but again, in a mechanically
important way, namely that of providing
lubrication (Florey, ' 5 5 ) . It is difficult to
provide a mechanical argument for the
heparin which presumably binds RR in the
mast cell (Gustafson and Pihl, '67b), for
the substance filling the T-tubules or for
the material between axon and Schwann
cell. Still the correlation between RR-positive material and mechanical properties
such as strength is sufficiently frequent
to warrant a tentative association of one
with the other where such material appears. Since many of the cells described
in this paper, as well as others,' have a
RR-postive coat surrounding them, it is
conceivable that this layer is of mechanical
importance to the cells. Whether it be relatively thick as in muscle or epidermal cells
or thin as in the RBC, so far it has been
always continuous with, or a component
of, the external leaflet of the unit membrane of the cell. It may be this layer, perhaps in conjunction with a mirror image
of itself on the inner or cytoplasmic leaflet
of the unit membrane, which provides the
tensile strength which most cells display
but which is not accounted for by the continuous lipid bilayer of the Davson-DanielliRobertson model of the plasma membrane
(Davson, '62; OBrien, '67; Bangham and
Haydon, '68). On a strong, felted gel of
polysaccharides supporting the lipid bilayer may be hung the various substituents
and charged groups, such as sialic acid or
hyaluronic acid which generate at least
some of the positive RR reaction (Morgan,
Intracellular penetration
Early in this paper attention was directed toward the RR reaction occurring
almost exclusively in the extracellular compartment with the cytoplasmic space acquiring the dye only when the plasma
membrane was itself damaged, an infrequent event. This observation holds in general, but the rare exceptions to it are reproducible and eventually may become
valuable as a morphological probe for
various functional states. These exceptions
in which there is no reason to suspect
mechanical damage are: mast cell granules, lipids within some cells, some nuclei
as in chondrocytes, certain cheek epithelial cells and the sarcoplasmic sacs in the
T-system. A full explanation of these findings cannot be made, but an hypothesis is
presented here, which, at least, has the
merit of providing a framework in which
to discuss these observations. It is proposed
that the "normal," intact, but fixed plasma
membrane is not absolutely impermeable
to RR but rather that it has a low and
finite permeability to the dye. Also, the
amount of RR bound to any particle is
assumed to be related both to the concentration of RR to which the particle is exposed and to the "afEnity" of the particle
for RR (the affinity is the descriptive term
for an association constant between the
dye and the particle which so far has not
been measured, but which probably is a
function of the charge density of the particle as well as steric effects [Sterling,
'701 ). Accordingly, when the plasma membrane of cells is intact, a tissue block exposed to RR (according to the schedule
used in this paper) experiences a high
concentration of RR in the peripheral extracellular spaces and a low but finite concentration of RR in the cytoplasm of the
peripheral cells. Also, over a period of
time, the concentration of RR in the interior of the block will rise to a low value.
This concentration can become nearly
equal in the cytoplasm and extracellular
spaces depending on the rate of RR diffusion inward and the quantities of RR
being bound by small amounts of high
affinity materials along the path. It is postulated that the large amount of RR-positive density normally found in the extracellular spaces is due to a large amount of
extracellular acidic polysaccharide which
has relatively low affinity for RR but which
is exposed to a high concentration of the
dye. The concentrations of RR deep in the
block are below the level at which appreciable RR can be bound to the extracellular material. However, within cells
there are substances, such as the mast cell
granules and certain lipids, which are postulated to have very high affinity for RR
and which could bind significant amounts
1 Andersqn, '68; Behnke, ,'68a,b. '69. gehnke and
Zelander 70. Bondareff, 67. Bkoki, 69. Faur.5Fremiet 'and 'And&, '68; Fa&-Remiet, Aid& and
Ganier, '68; Fowler, '70; Groniowski et al.. '69; Leak,
'67; Mohos and Wagner, '69; Monis ayd Zambrano,
'68; Monis et 51.. 69; Reimann et al., 65; ScarpeF
and Kanczak, 65; Siew and Wagner, '68; Szollosl,
'67, '70.
of dye from even low concentrations. This
hypothesis follows closely the discussion
of staining in electron microscopy presented by Zobel and Beer (’65). It is possible to test these requirements of the hypothesis quantitatively.
Continuing the argument, a puzzling observation arises in the case of myelin.
There are large amounts of myelin in nerve
tissue, such as in figure 46; yet on the
whole, very little RR is bound to it. However, for any RR to accumulate deep in the
lamellae in the knuckle of myelin in this
figure, the entire region must have been
exposed to at least this concentration if
not higher. Why should density be localized
to a small region? It is possible to rationalize such localized density and still retain
the hypothesis by an additional assumption: that materials which have high affinity for RR are frequently found in biology to be structurally incorporated in
such a way that their affinity is “masked”
and that injury (probably mechanical in
the case of myelin) can “unmask the
groups which then exhibit a high affinity
for RR. The masking results from their
charged groups being used to form the
structure in the first place; the unmasking
results from the mechanochemical disruption of these bonds (Watson, ’61; Katchalsky and Oplatka, ’ 6 3 ) in the structure to
invite combination with RR. This may account for the unusual density of myelin in
the region of structural disorganization
(figs. 57, 58).
This argument with slight modification
may be invoked to account for the stained
nuclei of the chondrocytes in figure 15.
It is highly unusual for a nucleus to stain
with RR even when it can be shown that
the plasma membrane has admitted RR
to the cytoplasm. For example, in figure
54, although the granules of the two mast
cells have taken up much RR, the nucleus
has not and is visible only as a negative
image where the volume of the nucleus
displaces stained granules. In figure 15 the
RR has stained only the two nuclei of the
chondrocytes and the scores of other nuclei
which certainly exist within the field of the
picture are essentially unstained. It is conceivable in light of the senility known to
occur in chondrocytes (Quintarelli and
Dellovo, ’66) that biochemical damage
concomitant with aging in the life cycle
of the chondrocytes has unmasked the nucleic acids from their previously masked
(and functional) state and that they now
are free to bind RR. It would be even more
exciting if aging and unmasking could be
demonstrated to be causally related.
The hypothesis requires in general, that
the cytoplasmic materials of cells (except
for lipids) have only a moderate affinity
for RR since they do not stain if the plasma
membrane is intact (excluding RR at high
concentration) but stain intensely if the
plasma membrane is ruptured (Tani and
Ametani, ’70). It is also conceivable that
the permeability of the plasma membrane
of certain cells to RR varies slowly in time
as cells mature, or perhaps rapidly and reversibly in parallel with some physiological change. The former circumstance might
be illustrated by the oral epithelial cells,
with cells of different physiological age
admitting more or less RR to the intracellular space. If the concentration rose
high enough within a cell to permit binding to intracellular constituents, a cell
would appear darker than its neighbor. In
figures 21, 31 and 32 the dark cells may
in this way have signified their increased
permeability to RR which may likewise be
related to their functional state as they
ascend in time through the strata of their
epithelial environment eventually to be discarded. A similar instance of graded density in cytoplasmic staining by RR has
been noted in granulosa cells surrounding
ova (Szollosi, personal communication) in
which, again, the cells are known to be
destined for early death and rapidly are
declining in function. It must be kept in
mind, however, that very unusual assumptions are implicit in this argument. An
appeal is made to physiological conditions
existing in life, whereas the permeability
demonstrated by RR density is not “physiological.” The RR is used during fixation,
not in vivo, and RR is not a “physiological”
It is suggested, finally, that the circumstances of a rapid and a reversible change
in plasma membrane permeability may be
responsible for the peculiar distribution of
RR in the tubules and the sacs of the Tsystem. In figures 48-51 in muscle fixed
with glutaraldehyde containing RR, it was
noted that RR appeared to localize in the
T-tubules but not in the sarcoplasmic sacs
when the muscle was found in the extended state, whereas the density was observed in the sacs but not in the T-tubule
when the muscle was contracted. This differential selectivity does not seem to occur
in Os04-fixedmuscle followed by exposure
to OsO, containing RR (Kelly, '69) but this
requires further examination. An abundance of work has now made it clear that
the T-system together with calcium ion
are intimately involved in the contractionrelaxation cycle in muscle (Ebashi, Endo
and Ohtsuki, '69; Winegrad, '68; Hoyle,
'70; Ford and Podolsky, '70). It was previously suggested (Luft, '66b) that it might
be more than coincidence that the polyvalent cation RR appears to move selectively along a pathway postulated for
another polyvalent cation, namely calcium.
The physiology of membrane coupling to
muscle excitation and the mechanism of
tissue fixation are both too poorly understood to permit firm predictions as to
where RR necessarily must be found in
muscle. However, it is possible that the
variations in calcium permeability which
are thought to occur in various compartments during muscle contraction may be
clamped in one state or another during
fixation, during which time the highly
charged and slow moving RR would have
the opportunity to trace out and label a
pathway taken by calcium under more
physiological and reversible conditions.
This suggestion sounds less improbable
when considered together with the calciuminduced, calcium-release positive feedback
loop discussed by Ebashi, Endo and Ohtsuki ('69, p. 372) and by Ford and Podolsky ('70). A recent report strengthens
further the association between RR and
calcium movement. Using rat liver mitochondria, Moore ('71) found specific inhibition of active calcium transport by RR
at concentrations of about lo-' M. Additional encouragement is found in the work
of Elbers ('66) which indicates that cell
membranes retain their relative impermeability for ten minutes or more during
glutaraldehyde fixation, whereas Os04 depolarizes the cell within seconds. The potential usefulness of the ability to trace
visually an ionic pathway makes this pos-
sibility worth considering and exploring
further. The localization by Zacks and
Sheff ('68) of tetanus toxin in the Ttubules and the sarcoplasmic sacs is further
evidence of unusual permeability in the
Binding of RR by lipids
Although the majority of the observations reported in this paper, as well as
most of the literature on RR, indicates its
affinity for acidic polysaccharides, there
is considerable evidence that RR interacts
with and binds to certain lipids, particularly to the acidic phospholipids, such as
phosphatidylethanolaine (cephalin). The
chemical basis for this interaction has been
discussed in the previous paper (Luft, '71,
I), and there is confirmatory evidence of
its localization near or in lipids in biological material as figures 57-64 illustrate. In
the case of myelin (figs. 57, SS), and the
mitochondria (figs. 61, 62), it is not certain that the RR is being bound to acidic
lipids since other materials are present as
well, but the images are compatible with
this interpretation. There is less doubt concerning the fat droplets in figures 59 and
60 and the myelin figures in figures 63
and 64. For a variety of reasons, mostly
indirect, the electron microscopists over
the years have become convinced that
droplets similar to those in figure 59 are
lipid in nature (Fawcett, '66). The crude
lecithin used in figures 63 and 64 was
shown by chromatography to contain lecithin, lysolecithin and phosphatidylethanolamine. Lamellae occasionally are seen
faintly in fat droplets under conventional
conditions but only after intense staining.
Images with a contrast and density such
as depicted in figures 59 and 60 are found
only in tissues exposed to RR and are not
seen in the controls. Similar lamellae in
megakaryocytes were reported by Nakao
and Angrist ('68) and in platelets by
Shirasawa and Chandler ('69) using RR.
As for the myelin figures in figures 63 and
64, i t is clear that RR has reacted with
at least some component of the crude mixture; and from the chemistry summarized
in the first paper (Luft, '71, I), phosphatidylethanolamine is the most likely candidate, although there is interaction with
lysolecithin. From figure 64 it can be seen
that lamellae persist through the reaction, (Luft, ’71, I), the probability of encounterbut that a fuzzy or flocculent density has ing them in tissue spaces is very low unaccumulated on the periphery of these less they were introduced deliberately.
lamellae. The nature of this density is not However, it is important to keep in mind
known, and this phase of the work needs the possibility of accidental leakage of
to be repeated with mixtures of purified densely staining cytoplasmic materials
phospholipids. Mixtures will be required from damaged cells, particularly in insince myelin figures of pure phosphatidyl- stances of unusual or localized extracelluethanolamine are rapidly destroyed on lar density. It would also be prudent to
contact with RR (Luft, ’71, I). It should withhold judgment in the interpretation
be possible to determine whether the fuzzy of density resulting from the use of RR or
density is due to an impurity of some sort, similar substances until inconsistencies,
perhaps a glycolipid, or is the reaction such as those raised in the report of Behnke
product resulting from the presence of and Zelander (’70) using Alcian blue, can
phosphatidylethanolamine (or other acidic be explained. Nevertheless, the general exphospholipids).
clusion of RR by cell membranes can be
These examples imply that the biological useful for tracing within cells, long or
situation concerning localization in lipids tortuous channels which had continuity
is not as straightforward as simple ionic with the surface at the time of fixation,
attraction of the RR polycation to existing and serves to reveal boundaries and comionizable groups of certain phospholipid partments within tissues which otherwise
molecules or (rarely) free fatty acids. Per- are not obvious. The several exceptions to
haps the presence of ionizable acid groups the exclusion rule invite exploration of the
in tissue lipids is a necessary but not suf- possibility that RR may serve as an ultraficient condition for RR binding. The fur- structural tracer at certain sites during
ther requirements may be an “unmasking” periods of membrane permeability to polyor displacing of those moieties already valent ions. There is also evidence to sugfunctionally associated with the charged gest that RR may localize at regions of
groups of the lipid and with which the mechanochemical damage to cells, by bindRR must compete for successful staining ing to polyvalent cations or to cationic arto occur, and that the lipids be so organized rays which were torn loose by the damthat the acidic groups offer an extended aging event. Finally, there is evidence that
RR can bind to and generate density at
This paper has attempted to demonstrate the site of fatty acids and acidic phosphothe diversity of images which can be ob- lipids. Considerably more work is required
tained reproducibly when ruthenium red to establish the specificity of the RR reac(RR) is applied to tissue blocks during tion for acidic polysaccharides and the mefixation and in the presence of osmium chanisms of other reactions described
tetroxide. Even if there were no suggestion here. Although the work may be tedious,
of a mechanism for the reaction, the the experiments are not difficult to carry
method would be useful in certain circum- out, so that some years in the future perstances for the staining selectivity which haps, this paper can be rewritten in a less
it can produce in tissue. If one accepts the speculative vein.
botanists’ experience that RR reacts with
pectins, and if this is extended to include
acidic mucopolysaccharides (acidic proteinThe author wishes to express his gratipolysaccharides or glycosaminoglycans) of tude to Drs. N. B. Everett and D. E. Kelly
animal origin as well, then the results of for reading the manuscript, and to Dr.
staining in the extracellular compartment Kelly for his critical comments and helpor at cell surfaces may be interpreted ful suggestions during its preparation.
reasonably as indicating sites of these maThis study was supported by USPHS
terials. Although there are many sub- grants NB-00401 and GM-16598 from the
stances other than acidic mucopolysac- National Institutes of Health and by a
charides which react vigorously with RR, grant-in-aid from the American Heart Asas listed in the fkst part of this series sociation, 67-665.
1947 Intercellular cement and capillary permeability. Physiol. Rev., 27: 436-463.
Anderson, W. A. 1968 Cytochemistry of sea Cook, G. M. W. 1968 Glycoproteins in memurchin gametes. 11. Ruthenium red staining of
branes. Biol. Rev., 43: 363-391.
gamete membranes of sea urchins. J. UltraCopley, A. L., and B. M. Scheinthal 1970
struct. Res., 24: 322-333.
Nature of the endoendothelial layer as demoArmstrong, P. B., and D. P. Jones 1968 On the
strated by ruthenium red. Exptl. Cell Res., 59:
role of metal cations i n cellular adhesion: cation
specificity. J. Exp. Zool., 167: 275-282.
Davson, H. 1962 Growth of the concept of the
Baier, R. E., E. G. Shafrin and W. A. Zisman
paucimolecular membrane. Circulation, 26:
1968 Adhesion: mechanisms that assist or im1022-1037.
pede it. Science, 162: 1360-1368.
Doggenweiler, C. F., and S. Frenk 1965 StainBangham, A. D., and D. A. Haydon 1968 Ultraing properties of lanthanum on cell membranes.
structure of membranes: bimolecular organizaProc. Nat. Acad. Sci., 53: 425-430.
tion. Brit. Med. Bull., 24: 124-126.
Douglas, W. H. J., R. C. Ripley and R. A. Ellis
Behnke, 0. 1968a Electron microscopical ob1970 Enzymatic digestion of desmosome and
servations on the surface coating of human
hemidesmosome plaques performed on ultrathin
blood platelets. J. Ultrastruct. Res., 24: 51-69.
sections. J. Cell Biol., 44: 211-215.
1968b A n electron microscope study of Ebashi, S., M. Endo and I. Ohtsuki 1969 Conthe megacaryocyte of the rat bone marrow. I.
trol of muscle contraction. Quart. Rev. Biophys.,
The development of the demarcation membrane
2: 351-384.
system and the platelet surface coat. J. Ultra- Edwards, J. 1967 Physical characteristics of
struct. Res., 24: 412-433.
articular cartilage. In: Lubrication and Wear
1969 A n electron microscope study of
in Living and Artificial Human Joints. Proc.
the rat megacaryocyte. 11. Some aspects of
Instn. Mech. Engrs., 181, Pt. 3J: 16-24.
platelet release and microtubules. J. Ultrastruct. Elbers, P. F. 1966 Ion permeability of the egg
Res., 26: 111-129.
of Limnaea stagnalis L. on fixation for electron
microscopy. Biochim. Biophys. Acta, 112: 318Behnke, O., and T. Zelander 1970 Preservation
of intercellular substances by the cationic dye
Alcian Blue in preparative procedures for elec- Farquhar, M. G., and G. E. Palade 1965 Cell
tron microscopy. J. Ultrastruct. Res., 31: 424junctions in amphibian skin. J. Cell Biol., 26:
263-29 1.
Benedetti, E. L.,and P. Emmelot 1967 Studies Faure-Fremiet, E., and J. Andr.6 1968 Structure
corticale d'une amibe edaphique. Protistologica,
on plasma membranes. IV. The ultrastructural
4: 195-207.
localization and content of sialic acid in plasma
membranes isolated from rat liver and hepa- Faure-Fremiet, E., J. AndrC and M.-C. Ganier
1968 Calcification Ggumentaire chez les cili6s
toma. J. Cell Sci., 2: 499-512.
due genre Coleps Nitzsch. J. Microscopie, 7:
Bennett, H. S. 1963 Morphological aspects of
extracellular polysaccharides. J. Histochem.
Fawcett, D. W. 1966 A n Atlas of Fine StrucCytochem., 11: 14-23.
ture. The Cell. Its Organelles and Inclusions.
Berlin, J. D. 1967 The localization of acid
W. B. Saunders Co., Philadelphia, pp. 307-318.
mucopolysaccharides i n the Golgi complex of
intestinal goblet cells. J. Cell Biol., 32: Florey, H. 1955 Mucin and the protection of
the body. Proc. Roy. SOC.B, 143: 147-158.
Bondareff, W. 1967 A n intercellular substance Ford, L. E., and R. J. Podolsky 1970 Regeni n rat cerebral cortex: submicroscopic distribuerative calcium release within muscle cells.
Science, 167: 58-59.
tion of ruthenium red. Anat. Rec., 157: 527535.
Fowler, B. A. 1970 Ruthenium red staining of
Bonner, J. 1936 The chemistry and physiology
rat glomerulus. Perfusion of ruthenium red
into normal and nephrotic rat kidney. Histoof the pectins. Botan. Rev., 2: 475497.
Brooks, R. E. 1969 Ruthenium red stainable
chemie, 22: 155-162.
surface layer on lung alveolar cells: electron Franzini-Armstrong, C., and K. R. Porter 1964
microscopic interpretation. Stain Technol., 44:
Sarcolemmal invaginations constituting the T
system in fish muscle fibers. J. Cell Biol., 22:
Burkel, W. E. 1967 The histological fine structure of perineurium. Anat. Rec., 158: 177-189. Gasic, G., and L. Berwick 1963 Hale stain for
Chambers, R. 1940 The relation of extraneous
sialic acid-containing mucins. Adaptation to
coats to the organization and permeability of
electron microscopy. J. Cell Biol., 19: 223-228.
cellular membranes. Cold Spring Harbor Symp., Gasic, G. J., L. Benvick and M. Sorrentino 1968
8: 144-153.
Positive and negative colloidal iron as cell surChambers, R., and E. L. Chambers 1961 Exface electron stains. Lab. Invest., 18: 63-71.
plorations Into the Nature of the Living Cell. Gomori, G. 1946 A new histochemical test for
Harvard University Press, Cambridge, Mass.,
glycogen and mucin. Am. J. Clin. Pathol., 10
(Tech. Section): 177-179.
Chambers, R., and B. W. Zweifach 1940 Capil- Gordon, J. E. 1964 Whiskers. Endeavour, 23:
lary endothelial cement in relation to perme8-12.
ability. J. Cell. and Comp. Physiol., 15: 255- Grant, L. 1965 The sticking and emigration of
white blood cells in inflammation. In: The
Inflammatory Process. B. W. Zweifach, L.
Grant and R. T. McCluskey, eds., Academic
Press, New York, pp. 197-244.
Gray, V. R. 1964 Contact angles, surface tensions and adhesion. In: Aspects of Adhesion.
Vol. 2, D. J. Alner, ed., Univ. of London Press,
London, pp. 4 2 4 8 .
Groniowski, J., W. Biczyskowa and M. Walski
1969 Electron microscope studies on the surface coat of the nephron. J. Cell Biol., 40:
Gustafson, G. T., and E. Pihl 1967a Staining
of mast cell acid glycosaminoglycans in ultrathin sections by ruthenium red. Nature, 216:
19671, Histochemical application of ruthenium red i n the study of mast cell ultrastructure. Acta pathol. et. microbiol. scand.,
69: 393-403.
Hale, C. W. 1946 Histochemical demonstration
of acid polysaccharides in animal tissues. Nature, 157: 802.
Hanker, J. S., A. R. Seaman, L. P. Weiss, H. Ueno,
R. A. Bergman and A. M. Seligman 1964
Osmiophilic reagents: new cytochemical principle for light and electron microscopy. Science,
146: 1039-1043.
Highton, T. C., D. B. Myers and D. G. Rayns
1968 The intercellular spaces of synovial tissue. New Zealand Med. J., 67: 315-325.
Hoyle, G. 1970 How is muscle turned on and
off? Sci. Am., 22214): 85-93.
Huxley, H. E. 19W Evidence for continuity between the central elements of the triads and
extracellular space in frog sartorius muscle.
Nature, 202: 1067-1071.
Ito, S. 1965 The enteric surface coat on cat intestinal microvilli. J. Cell Biol., 27: 475-491.
Jones, D. B. 1969 Mucosubstances of the glomerulus. Lab. Invest., 21: 119-125.
Jones, H. C., I. L. Roth and W. M. Sanders, I11
1969 Electron microscopic study of a slime
layer. J. Bact., 99: 316-325.
Karnovsky, M. J. 1968 The ultrastructural basis
of transcapillary exchanges. In: Biological Interfaces: Flows and Exchanges. Little, Brown
and Co., Boston, pp. 64-95.
Katchalsky, A., and A. Oplatka 1963 Mechanochemistry. In: Proceedings, Fourth International
Congress Rheology, Part 1. E. Lee and A. Copley, eds., Interscience, New York, 1965, pp. 7397.
Kelly, A. 1967 The nature of composite materials. Sci. Am., 217(3): 160-176.
Kelly, D. E. 1966 Fine s t r u c t u e of desmosomes,
hemidesmosomes, and a n adepidermal globular
layer in developing newt epidermis. J. Cell Biol.,
28: 51-72.
1967 Fine structure of cell contact and
the synapse. Anesthesiology, 28: 6-30.
1969 The fine structure of skeletal
muscle triad junctions. J. Ultrastruct. Res., 29:
Kerr, P. F. 1963 Quick clay. Sci. Am., 209(5):
Khan, T. A., and J. Overton 1970 Lanthanum
staining of developing chick cartilage and re-
aggregating cartilage cells. J. Cell Biol., 44:
Krizek, R. J. 1968 Phenomenological soil-polymer parallels. Am. Sci., 56: 279-287.
Lake, G. J., and A. G. Thomas 1967 The
strength of highly elastic materials. Proc. Roy.
SOC.A, 300: 108-119.
Leak, L. V. 1967 Fine structure of the mucilaginous sheath of Anabuena sp. J. Ultrastruct.
Res., 21: 61-74.
Luft, J. H. 1964a Electron microscopy of cell
extraneous coats as revealed by ruthenium red
staining. J. Cell Biol., 23: 54A-55A.
1964b The methods of preparation and
use of ruthenium red and ruthenium violet in
staining extracellular biological substances.
Preprint, distributed on request.
1965a The ultrastructural basis of capillary permeability. In: The Inflammatory Process. B. w. Zweifach, L. Grant and R. T. McCluskey, eds. Academic Press, New York, pp.
1965b The fine structure of hyaline
cartilage matrix following ruthenium red fixation and staining. J. Cell Biol., 27: 61A.
1965c Fine structure of capillaries: the
endocapillary layer. Anat. Rec., 151: 380.
1966a Fine structure of nerve and
muscle cell membrane permeability to ruthnium red. Anat. Rec., 154: 379-380.
1966b Ruthenium red staining of the
striated muscle cell membrane and the myotendinal junction. In: Electron Microscopy,
1966. Proceedings Sixth International Congress,
Kyoto, Vol. 11, Biology. R. Uyeda, ed. Maruzen
CO., Ltd., Tokyo, pp. 65-66.
1 9 6 6 ~ Fine structure of capillary and
endocapillary layer as revealed by ruthenium
red. Fed. Proc., 25: 1773-1783.
-1966d Ruthenium red and violet. I.
Chemistry, purification, methods of use and
mechanism of action. Preprint, distributed by
request. (This reference has been attributed
erroneously in the literature to the Univ. of
Wash. Press.)
1968 Selective staining of acid mucopolysacchatides by ruthenium red. In: Proceedings 26th Annual Meeting EMSA, Claitor's
Publ. Div. Baton Rouge, La., pp. 38-39.
1971 Ruthenium red and violet. I.
Chemistry, purification, methods of use for
electron microscopy and mechanism of action.
Anat. Rec., 171: 347-368.
Marinozzi, V. 1961 Silver impregnation of ultrathin sections for electron microscopy. J.
Biophys. Biochem. Cytol., 9: 121-133.
Maroudas, A. 1968 Physicochemical properties
of cartilage in the light of ion exchange theory.
Biophys. J., 8: 575-595.
Martinez-Palomo, A., and C. Brailovsky 1963
Surface layer in tumor cells transformed by
Adeno-12 and SV 40 viruses. Virology, 34: 379382.
Matukas, V. J., B. J. Panner and J. L. Orbison
1967 Studies on ultrastructural identification
and distribution of protein-polysaccharide in
cartilage matrix. J. Cell Biol., 32: 365-377.
Mohos, S. C., and B. M. Wagner 1969 Damage
to collagen in corneal immune injury. Observation of connective tissue structure. Arch. Path.,
88: 3-20.
Monis, B., A. Candiotti and J. E. Fabro 1969
On the glycocalyx, the external coat of the
plasma membrane, of some secretory cells. Z.
Zellforsch., 99: 64-73.
Monis, B., and D. Zambrano 1968 Ultrastructure of transitional epithelium of man. Z.
Zellforsch., 87: 101-117.
Moore, C. L. 1971 Specific inhibition of mitochondrial C a t + transport by ruthenium red.
Biochem. Biophys. Res. Comm., 42: 298-305.
Morgan, H. R. 1968 Ultrastructure of the surfaces of cells infected with avian leukosissarcoma viruses. J. Virolonv. 2: 1133-1146.
Morley, J. G. 1966- Fibre reinforced metals. Sci.
J., 2(11): 42-47.
Muir, A. R. 1961 Observations on the attachment of myofibrils to the sarcolemma at the
muscle-tendon junction. In: Electron Microscopy i n Anatomy. J. D. Boyd, F. R. Johnson
and J. D. Lever, eds. Edward Arnold, Ltd.,
London, pp. 267-277.
Myers, D. B., T. C. Highton and D. G. Rayns
1969 Acid mucopolysaccharides closely associated with collagen fibri3s in normal human
synovium. J. Ultrastruct. Res., 28: 203-213.
Nakao, K., and A. A. Angrist 1968 Membrane
surface specialization of blood platelet and
megakaryocyte. Nature, 21 7 : 960-961.
O’Brien, J. S. 1967 Cell membranes-composition: structure: function. J. Theoret. Biol., 15:
Odland, G. F. 1958 The fine structure of the
interrelationship of cells in the human epidermis. J. Biophys. Biochem. Cytol., 4: 529538.
OppenIander, G. C. 1968 Structure and properties of crystalline polymers. Science, 159: 13111319.
Overton, J. 1969 A fibrillar intercellular material between reaggregating embryonic chick
cells. J. Cell Biol., 40: 136143.
Partridge, S. M. 1968 The chondroitin sulfateprotein complex from bovine cartilage. In: The
Chemical Physiology of Mucopolysaccharides.
G. Quintarelli, ed. Little, Brown and Co., Boston,
Mass., pp. 51-62.
Pate, J. L., and E. J. Ordal 1967 The fine
structure of Chondrococcus columnaris. 111. The
surface layers of Chondrococcus columnaris.
J. Cell Biol., 35: 37-51.
Pease, D. C. 1966 Polysaccharides associated
with the exterior surface of epithelial cells:
kidney, intestine, brain. J. Ultrastruct. Res., 15:
Pickett-Heaps, J. D. 1967 Preliminary attempts
at ultrastructural polysaccharide localization i n
root tip cells. J. Histochem. Cytochem., 15:
1968 Further ultrastructural observations on polysaccharide localization in plant
cells. J. Cell Sci., 3: 55-64.
Pihl, E., G. T. Gustafson and S. Falkmer 1968
Ultrastructural demonstration of cartilage acid
glycosaminoglycans. Histochem. J., 1 : 26-33.
Pratt, S. A., and L. Napolitano 1969 Osmium
binding to the surface coat of intestinal microvilli in the cat under various conditions. Anat.
Rec., 165: 197-209.
Preston, R. D. 1952 The Molecular Architecture of Plant Cell Walls. John Wiley and Sons,
New York.
Quintarelli, G., and M. C. Dellovo 1966 Age
changes in the localization and distribution of
glycosaminoglycans in human hyaline cartilage.
Histochemie, 7 : 141-167.
Rambourg, A. 1967 DBtection des glycoproteines en microscopie Blectronique: coloration
de la surface cellulaire et de l’appareil de Golgi
par un mblange acide chromique-phosphotungstique. C. R. Acad. Sci., Ser. D., 265:
Rambourg, A., W. Hernandez and C. P. Leblond
1969 Detection of complex carbohydrates in
the Golgi apparatus of rat cells. J. Cell Biol.,
40: 3 9 5 4 1 4 .
Rambourg, A., and C. P. Leblond 1967 Electron microscope observations on the carbohydrate-rich cell coat present a t the surface of
cells in the rat. J. Cell Biol., 32: 27-53.
Rambourg, A., M. Neutra and C. P. Leblond 1966
Presence of a ‘cell coat’ rich in carbohydrate
at the surface of cells i n the rat. Anat. Rec.,
154: 41-71.
Rasmussen, H. P. 1967 Calcium and strength
of leaves. I. Anatomy and histochemistry.
Botan. Gaz., 128: 219-223.
Reimann, B. E. F., J. C. Lewin and B. E. Volcani
1965 Studies on the biochemistry and fine
structure of silica shell formation in diatoms.
I. The structure of the cell wall of Cylindrotheca fusiformis Reimann and Lewin. J. Cell
Biol., 24: 39-55.
Revel, J.-P. 1964 A stain for the ultrastructural
localization of acid mucopolysaccharides. J. Microscopie, 3: 535-544.
Revel, J.-P., and D. W. Hamilton 1969 The
double nature of the intermediate dense line
i n peripheral nerve myelin. Anat. Rec., 163:
Revel, J.-P., and M. J. Karnovsky 1967 Hexagonal array of subunits in intercellular junctions of the mouse heart and liver. J. Cell Biol.,
33: C7-Cl2.
Salton, M. R. J. 1964 The Bacterial Cell Wall.
Elsevier Publ. Co., New York, pp. 92-100.
Scarpelli, D. G., and N. M. Kanczak 1965 U1trastructural cytochemistry: principles, limitations, and applications. Int. Rev. Exptl. Pathol.,
4: 55-126.
Seligman, A. M., J. S. Hanker, H. Wasserkrug,
H. Dmochowski and L. Katzoff 1965 Histochemical demonstration of some oxidized macromolecules with thiocarbohydrazide (TCH )
or thiosemicarbazide (TSC) and osmium
tetroxide. J. Histochem. Cytochem., 13: 629-639.
Sharon, N. 1969 The bacterial cell wall. Sci.
Am., 220(5): 92-98.
Shirasawa, K., and A. B. Chandler 1969 Fine
structure of the bond between platelets in artificial thrombi and in platelet aggregates induced by adenosine diphosphate. Am. J. Path.,
57: 127-152.
Sidgwick, N. V. 1950 The Chemical Elements Tani, E., and T. Ametani 1970 Substructure of
microtubules in brain nerve cells as revealed by
and Their Compounds. Vol. 2, Oxford, London,
ruthenium red. J. Cell Biol., 46: 159-165.
pp. 1459-1489.
Siew, S., and B. Wagner 1968 Application of ThiBry, J.-P. 1967 Mise e n bvidence des polysaccharides sur coupes fines en microscopie
the ruthenium red technique in the electron
Blectronique. J. Microscopie, 6: 987-1018.
microscopy of the cardiovascular tissue. Elec- Tsaltas, T.T., and K. A. Greenawald 1966 Comtron Microscopy, 1968. Proceedings Fourth
parison of changes in the chemical composition
European Reg. Conference, Rome, Vol. 11,
of rabbit cartilage matrix with age and folD. S. Bocciarelli, Ed., Tipografia Poliglotta
lowing intravenous papain. Nature, 21 0:
Vaticana, Rome, pp. 68-69.
Sterling, C. 1970 Crystal-structure of ruthenium Watson, W. F. 1961 Mechanochemistry. New
red and stereochemistry of its pectic stain. Am.
Scientist, 9: 548-550.
J. Bot., 57: 172-175.
Winegrad, S. 1968 Intracellular calcium movements of frog skeletal muscle during recovery
Stern, I. B. 1965 Electron microscopic observafrom tetanus. J. Gen. Physiol., 51: 65-83.
tions of oral epithelium. I. Basal cells and the
basement membrane. Periodontics, 3: 224-238. Wissig, S. L., and D. 0. Graney 1968 Membrane modifications in the apical endocytic
Szollosi, D. 1967 Development of cortical
complex of ileal epithelial cells. J. Cell Biol.,
granules and the cortical reaction in rat and
39: 564-579.
hamster eggs. Anat. Rec., 159: 431-446.
-1970 Cortical cytoplasmic filaments of Woodward, C., and E. A. Davidson 1968 Structure-function relationships of protein polysaccleaving eggs: a structural element correspondcharide complexes: specific ion-binding propering to the contractile ring. J. Cell Biol., 44:
ties. Proc. Nat. Acad. Sci., 60: 201-205.
Yardley, J. H., and G. D. Brown 1965 FibroSzubinska, B. 1964 Electron microscopy of the
blasts in tissue culture. Use of colloidal iron
interaction of ruthenium violet with the cell
for ultrastructural localization of acid mucomembrane complex of Amoeba proteus. J. Cell
polysaccharides. Lab. Invest., 14: 501-513.
Biol., 23: 92A.
Zacks, S. I., and M. F. Sheff 1968 Tetanus
Szubinska, B., and J. H. Luft 1971 Ruthenium
toxin: fine structure localization of binding
red and violet. 111. Fine structure of the plasma
sites in striated muscle. Science, 159: 643-644.
membrane and extraneous coats in amoebae Zobel, C. R., and M. Beer 1965 The use of
(A. proteus and Chaos chaos). Anat. Rec.,
heavy metal salts as electron stains. Int. Rev.
171: 417-442.
Cytol., 18: 363-400.
In the following legends, “section unstained” indicates that no optical
dyes, nor heavy metals such as uranyl acetate or lead solutions, were
applied to the section; so that the density in the micrograph is derived
only from the RR or RRiOs04 introduced during fixation, and the native
mass of the tissue components themselves.
Light micrograph of unstained thick section of mouse diaphragm
exposed to ruthenium red (RR) in fixatives. Density in extracellular
spaces in periphery of the block shows poor penetration of the RR as
indicated by the zonation of the RR/Os04 reaction product. Dense
muscle fibers (arrow) show intracellular penetration into damaged
cells. Rectangle indicates area enlarged in figure 2. x 65.
Enlargement of rectangular area in figure 1. Rectangle indicates
region from which electron micrograph, figure 3, was obtained several
sections away. x 260.
Electron micrograph of mouse diaphragm from section near that
depicted in figures 1 and 2, from area within rectangle in figure 2.
The density within the extracellular space fades away at the bottom
of the picture corresponding to that seen in optical microscopy in
figure 2. Striated muscle cells ( M ) and capillary lumens ( C ) are
unstained. Briefly stained with lead. X 3700.
Electron micrograph of mouse diaphragm showing RR effect among
several cell types. Two capillaries ( C ) are seen in longitudinal and
cross section between four muscle cell cross sections (M). Bundles
of collagen ( B ) traverse the edematous intercellular space. Nerve
( N ) is armored with collagen, and is also pictured in figures 44 and
45. Partial intracellular penetration of RR into MI. Section
unstained. x 3300.
John H. Luft
Electron micrograph from mouse lung showing collagen fibers coated
with variable amounts of RR-positive material, some oQ which lie in
register in bundles to accentuate the collagen periodicity. A fragment
of an unidentified cell is present at X. Section unstained. x 3000.
Collagen bundle adjacent to a capillary (C) from mouse diaphragm.
The density of the RR reaction is greater between the bundle and the
capillary, but uniform density is seen throughout the bundle separating each collagen fiber from its neighbors. Many of the collagen
fibers have one, and rarely two, dense dots at their centers. Briefly
stained with uranyi acetate and lead. X 61,000.
7 Higher magnification of collagen fibers in cross section, similar to
figure 6. The 50-10 A granularity characteristic of the RR-positive
material between the fibers is evident, as well as the central dot.
Briefly stained with lead. X 160,000.
John H. Luft
Light micrograph of 3 p thick Epon section of hyaline (xiphoid:
cartilage of frog stained with toluidine blue. Pairs of chondrocytes lie
in their capsules within the matrix which was strongly metachromatic. X 650.
Electron micrograph from another piece of frog xiphoid cartilage
together with perichondrium, prepared by conventional aldehyde/
OsO, fixation, without RR as a “control.” The cartilage matrix 1s
much lighter than the chondrocytes or the fibroblasts ( F ) within
the perichondrium. Lightly stained with lead. x 3000.
10 Electron micrograph of frog xiphoid cartilage and perichondrium
fixed in the presence of RR for comparison with figure 9. The cartilage
matrix has acquired considerable density and naw is much darker
than the chondrocytes. A pair of centrioles (arrow) lie centered in
the Golgi complex just above the rectangle which indicates the area
enlarged for figure 11. There is also RR-positive density around collagen bundles in the perichondrium. Section unstained x 3000.
John H. Luft
Successively greater magnification of figure 10. The matrix can
be seen to be composed of approximately spherical globules about
200-300 A in diameter which appear to be connected to their
nearest neighbors by a delicate thread. 800 A thick section,
unstained. Figure 11, >( 34,000; figure 12, x 160,000.
13 Adjacent section to that shown in figures 10-12, but thinner and
strongly stained with lead. The globules assume various densities
from dark to light gray and the threads branch in a complex
manner. There i s no distinct image of collagenous fibrils. 400 A
section stained with lead. x 160,000.
14 Adjacent section to that shown i n figure 13, stained with uranyl
acetate followed bv lead showing several striated fibrils which
probably are collagen. 400 A section, stained with uranyl and lead.
X 160,000.
15 Light micrograph of frog xiphoid cartilage fixed in buffered glutaraldehyde plus RR and then dehydrated and embedded directly
without exposure to OsOJ and with no further staining. Density
here equivalent to magenta in the original color film, showing
localization of RR to the cartilage matrix. Nuclei of several
chondrocytes also stained with RR although other nuclei did not.
The arrow indicates the ghost of a chondrocyte just beyond the
zone of RR penetration, but visible in figure 16. Cartilage matrix,
M; perichondrium, P. Area within rectangle shown from adjacent
section by EM i n figure 16. x 650.
16 Electron micrograph of section adjacent to the region shown in
figure 15, from area within rectangle. Despite the knowledge of
the location of RR within the section from figure 15, there is not
enough additional mass to recognize by EM. 800 A section,
unstained, high contrast printing. x 1200.
Agar gel, 0.5% i n water, fixed and processed as a tissue block
with RR in botb fixatives. Pattern of branching and interconnecting filaments is consistent with gel structure. Area within
rectangle in figure 17 enlarged in figure 18. Section 700 A thick,
stained with lead. Figure 17, X 44,000; figure 18, x 220,000.
John H. Luft
19 Electron micrograph of cross section of a striated muscle cell from
mouse diaphragm exposed to RR. Collagen fibrils in the sarcoleinma
are closely applied to the cell surface and are embedded in a layer
of RR-positive material. The area within the rectangle is enlarged in
figure 20. Section unstained. x 5,200.
Enlargement of figure 19 showing a collagen fibril in negative contrast embedded in, and partly infiltrated by, dense RRpositive material which is continuous to the outer leaflet of the unit membrane
of the muscle cell. The inner (cytoplasmic) leaflet of the plasma
membrane is visible as a thin, dense line (IL) which is separated
by the light, middle leaflet from the RR-positive substances. Section
unstained. x 160,000.
Electron micrograph of transverse collagen fibrils around muscle
from mouse diaphragm. RR-positive material between collagen fibrils
is periodic instead of continuous, coincident with the collagen banding period, and with registration between bands of adjacent collagen
fibrils. Cross sectioned inyofilaments at M, plasmalemina a t arrow.
Section stained with uranyl and lead. X 508,000.
Electron micrograph of myotendinal junction from mouse diaphragm
by conventional preparation methods ( glutaraldehyde-0s04 and
uranyl-lead) but without RR, as a “control” for figure 23. Muscle
cytoplasm a t M, collagen of tendon at T. Saggital section stained
with uranyl and lead. x 5,000.
Electron micrograph of same tissue as that shown in figure 22, but
fixed in the presence of RR which generates density around the collagen fibrils of the tendon ( T ) and at the interface with the muscle
cell ( M ) . Area within rectangle is enlarged in figure 24. Saggital
section 400 A thick, lightly stained with lead. X 5,000.
Enlargement of figure 23, showing muscle-tendon interface, RR-positive material between collagen fibrils is continuous to outer leaflet of
plasma membrane, except for vacancies produced by basal lamina
(BL). Inner leaflet of plasmaleinma faintly visible ( arrow). Section
400 A thick, lightly stained with lead. x 120,000.
John H. Luft
Electron micrograph of section of mouse diaphragm with variations i n thickness produced by stick-slip friction during sectioning (“chatter”). The apparent variation in amount of RR-positive
material in the intercellular spaces depends only upon sectioii
thickness. Section unstained. x 38,000.
Electron micrograph of heavily stained, very thin section of mouse
heart showing details of external coat material applied to the outer
leaflet of the plasma membrane. At arrow the coat substance is
missing and the inner and outer leaflets have equal density. The
coat material has a thin, dense inner layer about 30-40 A thick
and possibly globular, which is confluent with the external leaflet
to which is loosely adherent a low density “fuzzy” or flocculent material about 500 A thick. Cell cytoplasm, C; extracellular space, E;
section 300 A thick, stained one hour each with uranyl acetate
and lead citrate. x 220,000.
Electron micrograph of human oral epithelial cells ( E ) with ad.
herent bacterium (B), after exposure to RR. Bacterial cell wall,
cell surfaces and the interdesmosomal material all are RR positive. Arrows indicate desmosomes. Section unstained. x 23,000.
Same material as that in figure 27, showing several types of cell
wall structure found in resident bacteria. Section stained with
lead. X 23,000.
Electron micrograph of three desinosomes between two oral
epithelial cells. The density due to RR remains extracellular even
a t the desmosomes where the pale central leaflet of the unit membrane faintly can be seen to separate the pale intracellular plaque
(P) from the very dense intercellular desmosomal material. Section unstained. x 75,000.
Electron micrographs of selected desmosomes between oral epithelial cells to illustrate the distributison of RR-positive density. The
dense material a t the desmosome appears to be quantatively
greater but qualitatively similar to that extending over the unspecialized surf aces of the epithelial cells. The density distribution within the desmosome is unlike, or even the inverse of that
reported in the literature by conventional methods. All sections
stained with lead. All x 160,000 except figure 32, x 46,000.
Oral epithelial cell desmosome tested
the synthesis of RR (ruthenium [III]
ride). The density distribution typical
reaction is weak. Section stained with
with an intermediate i n
chloropentammine dichloof RH. is found, but the
lead. x 160,000.
35 Electron micrograph of desmosome from frog gut. The distribution of density is very different from that seen i n figures 27-31;
the reason is unknown. Section lightly stained with lead. x 20,000
John H. Luft
Electron micrograph of frog intestinal epithelium exposed to RR
which produced density i n the fuzz filaments of the microvilli as
well as in thc mucus at their tips. Cytoplasm of the cells ( C )
unstained; gut lumen, L. Section unstained. x 22,000.
Same material as i n figure 36, but showing trilaminar membrane
(arrow) on microvillus with RR-positive material at the external
surface. Section briefly stained with lead. x 100,000.
38 Electron micrograph of 11-day-old rat intestinal epithelium labeled
with RR. Much flocculent mucus is stained in the lumen and the
microvilli appear to be negatively stained due to their strongly
RR-positive surface coat. A tubule extends from the base of the
microvilli to open into a vacuole, all of which stains densely i n
contrast to most of the other vesicles in the apical cytoplasm.
Several other apparently isolated vesicles stain strongly; presumably their connections to the surface lay out of the plane of section. Micrograph courtesy of Dr. Daniel Graney. Section unstained.
s 10,000.
Electron micrograph of mouse erythrocytes exposed to RR. A thin
layer of RR-positive material surrounds each erythrocyte. The area
within the rectangle is enlarged in figure 40, where the layer can
be seen both edge-on and in oblique projection. Section unstained.
x 5,000 and x 50,000 respectively.
Electron micrograph of a n erythrocyte (Er) close to a capillarv
endothelial cell ( C ) after exposure to RR. A thick layer of RR-positive material on the endothelial cell (endocapillary layer, see fig.
42) faces a much thinner layer on the erythrocyte. Several threads
of RR-positive material appear to connect the two layers, especially
at X. Section stained with lead. x 160,000.
Electron micrograph of segment of capillary wall exposed on both
luminal and tissue surfaces to RR. The trilaminar plasma membrane of endothelial cell ( C ) is clearly identified by the central
pale leaflet and is deeply invaginated by several pits or vesicles
( V ) which have open mouths. The outer leaflet is everywhere
continuous with a thick layer of RR-positive material; on the
luminal surface is it distinguished as the endocapillary layer
(ECL). Section lightly stained with uranyl acetate and lead.
x 160,000.
Electron micrograph of a lateral junction between two capillary
endothelial cells ( C ) labeled with RR from the tissue surface. The
lumen is at L where the plasma membranes become lost in
obliquity. The RR has entered the junction from the right in the
usual (Z 150-200 A ) gap between cells ( A ) , but near the lumen
the spacing twice abruptly narrows to a 30-40 A often called
a "close junction" (arrows). Section stained with lead. x 160,000.
Electron micrographs of the same unmyelinated nerve depicted in
figure 4. A number of axons are nestled a t the periphery of the
Schwann cell ( S ) , the whole being armored with parallel collagen fibrils. RR ha5 uniformly penetrated the collagen to label
the space between axon (A) and Schwann cell. The area within
the rectangle is enlarged in figure 45. Section unstained. x 17,000
and 160,OO respectively.
John H. Luft
Electron micrograph of branch of phrenic nerve from mouse diaphragm. The RR has given the typical pattern of density along the
collagen of the sarcolemma (left margin) and around the collagen
at the outer surface of the perineurium, but has not penetrated the
epithelial-like cells of the perineurium ( P ) to gain access to the
endoneural space (ES). Slight invasion, however, occurs at lateral
junctions of the perineural cells (arrows). Portions of five myelinated
axons are visible in cross section as well as the nucleus of a perjneural cell ( N ) . Section unstained. X 6,800.
Electron micrograph similar to that shown in figure 46 except that
the perineurium has been damaged and RR has gained access to the
endoneural space (ES) to label collagen fibrils as well as the surface
of the Schwann cells. There is obvious disorder in the myelin sheath
as well as within the axon ( A ) , and the cytoplasm of the perineural
cells is abnormally dense from RR penetration. Section unstained
X 6,800.
John H. Luft
Electron micrograph of muscle cell from mouse diaphragm (exposed to KR in glutaraldehyde as well as in OsOd) i n which
nearly every pair of sarcoplasmic sacs in the triads has been
labeled leaving the central T-tubule unstained. The muscle is contracted with n o I-band visible. Section unstained. x 6,800.
Electron micrograph of tissue similar to that shown in figure 48,
but showing a n :iccumulation of RR in the T-tubules. The Z-band
( Z ) runs vertically through the center of figure 49 with the
T-tubules marking the A-I junction. The muscle is in the extended
(relaxed) state. The area within the rectangle is enlarged for
figure 50, to permit resolution of the density within the T-tubule
as related to the trilaininar structure of the T-tubule membrane.
Section lightly stained with uranyl acetate. x 35,000 and 125,000
Electron micrographs, both from the same section of mouse
diaphragm exposed to RR. In figure 51 the muscle cell was near
the periphery of the block permitting the sarcoplasmic sacs of
triads to label. The density produced in the sarcoplasmic sacs is
pale, but is real compared with the “control” (fig. 52) obtained
deeper in the block to which RR had not obviously penetrated.
Both sarcomeres contracted. Section lightly stained with lead.
Both x 29,000.
John H. Luft
John H. Luft
Light micrographs of mast cells from a whole mount of mouse
diaphragm fixed with buffered glutaraldehyde plus RR, and
mounted without contact with osmium. The mast cells are revealed
by their content of deeply staining granules, but little else is
visible of the muscle tissue in the background. At higher magnification (fig. 54) the nuclei of two mast cells are revealed as
oval vacancies by displacement of the intensely stained granules.
x 65 and 650 respectively.
Electron micrograph of a mast cell between two capillaries ( C )
of mouse diaphragm prepared with both glutaraldehyde/RR and
Os04/RR. The mast cell granules have taken up RR as evidenced
by their density, but have not developed the density which might
be expected from their affinity for RR as indicated by figures 53
and 54, nor as much density as the cytoplasmic components of
the muscle cell in the lower right. Section unstained x 4,500.
John H. Luft
Electron micrograph of a myoneural junction from mouse diaphragm i n which the muscle cell was injured permitting the RR
to penetrate by the intracellular route instead of the more common extracellular pathway. Nevertheless, the plasma membrane
still is sufficiently impermeable to outline the subsynaptic apparatus and the axon ( A ) i n a reversal of the usual contrast. Section
unstained. x 21,000.
Electron micrograph of myelin from mouse diaphragm showing
localization of RR-positive material particularly at defects in the
lamellae. In figure 57 there is shown substantial damage to the
myelin and the density is seen to localize at the major dense lines,
particularly where they expand. In a thinner and lightly stained
section (fig. 5 8 ) , smaller but otherwise similar stained defects
are seen. Axon and Schwann cytoplasm at A and S. Figure 57,
section unstained; figure 58, lightly stained with lead. Both
x 70,000.
Electron micrographs of lipid droplets i n a muscle cell of mouse
diaphragm. Although RR apparently had not penetrated this far,
lipid droplets still showed density attributable to RR. Area within
rectangle in figure 59 i s enlarged i n figure 60 showing lamellae
with about 45 A spacing. Section unstained. x 29,000 and 160,000
Electron micrographs of mitochondria within a damaged muscle
cell from mouse diaphragm showing various patterns of penetration and localization. The majority of mitochondria are pale, but
two are dark due to RR penetration with one showing the site
of damage ( X ) . The pale mitochondria have dense cristae despite
their pale matrix, a portion of one within the rectangle being
enlarged i n figure 62. Here it can be seen that the internal leaflet
of the trilaminar membrane of the crista is denser than its
counterpart which faces the matrix. Section lightly stained with
lead. x 23,000 and 160,000 respectively.
Electron micrographs of myelin figures prepared from crude egg
lecithin, fixed and exposed to RR in the usual manner. The RR
has reacted at the surfaces of several figures to generate considerable density, but has not penetrated deeper. Near the tip of the
arrow (area too small to outline) is the region enlarged for figure
64. Pale lines, presumptive lipid lamellae, are seen set off by
fiuocculent surface density. Section unstained. x 2,500 and 160,000
John H. Luft
Без категории
Размер файла
3 991 Кб
structure, animals, red, localization, tissue, ruthenium, fine, violet
Пожаловаться на содержимое документа