A combination of little used preparative techniques permits the visualization of macromolecular complexes in peripheral nerve myelin. They can be employed effectively as well for similar particulate recognition in at least some other redundant cytomembrane systems, i.e. retinal disc membranes. The cores of the globular structures are thought to be the integral proteins of the membrane system, in glutaraldehyde-fixed myelin probably associated with phospholipid annuli.Fixation is limited largely to aldehydes, with osmium tetroxide being particularly contraindicated. Cadmium ions, which form water-insoluable salts with lecithins, seem to offer some advantages when included with a glutaraldehyde fixation. Tannic acid as a secondary fixative has proven necessary when embedments are to be avoided.

Under certain preparative conditions the lipid bilayers of glutaraldehyde-fixed, PNS myelin demonstrate a marked compartmentalization, which can be augmented by lipid extraction following sectioning. The results are interpreted as indicating a supramolecular domain pattern of arrangement centered upon the transmembrane protein (P0) molecules. The latter are thought to be surrounded by annuli of substantially immobilized phospholipids. In the lamellar planes particular lipids are considered to have a nonrandom distribution. The visualization of bilayer compartmentalization was seen only in negatively stained sections obtained from unembedded or glutaraldehyde-urea-embedded myelin. Lipids were unextracted in the basic preparations except in so far as some unfixed, amphipathic molecules escaped at the trough-fluid interface at the time of sectioning, an observed phenomenon which probably aided in the visualization of the compartmentalization. Visualization was also augmented by surface tension expanding section fragments as they floated on the trough fluid. All stages of transition between well-ordered myelin and dispersed globular units were commonly to be found. Deliberately delipidated myelin exposed more sharply defined and smaller globular units in bilayer regions, but even these are regarded as being supramolecular aggregates including residual lipid annuli around the transmembrane proteins. The addition of cadmium ions as a "fixative" for lecithin seemed to improve the preservation of glutaraldehyde-urea-embedded myelin but was not strictly necessary to reveal its domain structure. A secondary tannic acid fixation was required to process unembedded myelin so as to reveal the fundamental compartmentalization of its lipid bilayers.

Much ultrastructural detail is retained in tissue fixed only with aldehydes and subsequently air-dried after suspension in a polyvinyl acetate emulsion. The latter provides an external support only, but permits ultrathin sectioning; thus, an exposure of intracellular contents for potential immunocytochemical reactions is achieved. Sections of unembedded frog retina so prepared have been studied with success. The tissue was incubated first with a rabbit antiserum prepared against gradient purified bovine rod outer segments. Following incubation, reacted sites were labeled with ferritin-conjugated goat anti-rabbit IgG and stained with phosphotungstic acid. Intense labeling of the rod outer segments was clearly achieved, whereas the cone outer segments were without label. Other parts of the retina, including the ellipsoid region of both rods and cones, were also without significant label. These regions provided an intrinsic control for the specificity of the antiserum and established the validity of the general technique.

It has been demonstrated that it is feasible within limits to section a wide variety of unembedded, glutaraldehyde-fixed tissues thinly enough for transmission electron microscopy. Much cytological detail is preserved. The success of the method is thought to depend mainly upon glutaraldehyde adequately cross-linking cytosol proteins so that cellular components are not only held in place, but the tissue is sectionable after it is air dried. A polyvinyl-acetate emulsion in the form of carpenter's white glue is used as an external support for tissue blocks. If necessary, this can be rendered insoluble by the addition of a small quantity of serum albumin, subsequently cross-linked with glutaraldehyde. Glass knives are advisable for sectioning, and sections are floated on a layer of water reaching the knife edge with a minimal meniscus. Since alkaline lead stains have proven to be quite destructive to unembedded tissue, and other positive stains have not been found that are particularly effective, partially neutralized phosphotungstic acid, used as a negative stain, is the principal staining technique employed. Because it is thought that the technique ultimately may prove to be most useful for immunocytochemical applications, emphasis has been placed upon studies of tissues fixed only with glutaraldehyde. If tissue is secondarily fixed with osmium tetroxide and/or treated with uranyl salts, however, tissue blocks then have superior sectioning properties, and can be more effectively and positively stained.

It is commonplace to recognize that the depth to which we explore ourselves and our environment is frequently determined by the development of new instruments and the creation of techniques for their use . Usually in such developments, one can recognize a time when fragments of information, acquired previously, are ready to be used to satisfy a concept or an urge to do or see what had not seemed possible before . So it was in the early 1930s that a group ofphysicists and engineers, mostly in Berlin, found conditions right to create an electron microscope . Max Knoll and his students, Ernst Ruska and Bodo von Borries, had available the knowledge that electrons would move through a vacuum and be deflected in their motion so as to be focused by solenoid lenses . It was mostly engineering skills that were needed to generate a microscope . Interest in the applications of the first microscopes naturally followed and, by the late 1930s, electron micrographs of recognizable value to biologists were being published . Any consideration of the pace at which biological electron microscopy then developed must take into account worldwide events and constraints related to the outbreak and prosecution of World War II . Hitler invaded Poland on September 1, 1939 . The very first Siemens & Halske AG electron microscope made for commercial sale was delivered in that year, only a few months before the War actually started (2) . However, since the political alignments of Axis and Allied countries had been substantially established the year before, after the annexation of Czechoslovakia's Sudetenland, it is not suprising that not a single Siemens & Halske microscope was ever delivered to countries outside of Axis control . None was in Allied hands until one microscope was captured intact and brought to England after the 1944 Normandy invasion . In the United States, the Radio Corporation of American was not ready to deliver its first commercially available electron

The preservation for electron microscopy of saturated phospholipids in general, and phosphatidyl choline (PC)in particular, remains and unsolved problem since OsO(4) and glutaraldehyde are incapable of interacting with PC directly. However, by introducing tannic acid preceding osmication, we were able to demonstrate highly ordered, preserved lamellar structures in model experiments with saturated PC, and in vivo experiments type II pneumocytes of lung tissue. The secretory bodies of the latter are known to contain a high proportion of these saturated phospholipids. In both cases, the repeating periodicity approximated 45 A. It was determined that tannic acid interacts with the choline component of PC to form a "complex," which then could be stabilized by treatment with OsO(4). In the absence of osmication, the PC-tannic acid complex acid did not survive conventional dehydration techniques, but osmication permitted conventional Epon embedment. Sphingomyelin (SPH), which contains choline, behaved similarly in model experiments. But there was no evidence of a comparable reaction with tannic acid using phosphatidyl ethanolamine (PEA), phosphatidyl serine (PS), or phosphstidy inositol (PI). Chemical studies indicted a high pH dependency for the formation of the PC- tannic acid complex. Also, experiments demonstrated its dissociation in various organic solvents. Sharp delineation and great contrast of the polar zones in the ordered lamellar structures was achieved by additional staining with lead citrate thus leading to the conclusion that tannic acid serves as a multivalent agent, capable of simultaneous interaction with saturated PC, OsO(4), and lead citrate stains.

Unsaturated natural and synthetic phosphatidyl cholines (PCs), when treated with tannic acid and OsO4, demonstrated a substantial increase in contrast as compared to PC treated only with OsO4. This was not observed when phosphatidyl ethanolamine (PEA) was similarly exposed to tannic acid. The increased electron density observed in the lamellar organization of the PC phospholipids was limited to the hydrophilic layers corresponding to the polar regions of the phospholipid molecules. The repeating periods of lamellae were identical in PC, treated with both tannic acid and OsO4, and when treated only with OsO4. In each case, this approximated 45 A. The enhancement of membrane contrast by tannic acid in the presence of OsO4 is interpreted as being at least in part due to its multivalent capacity, binding to reactive sites on choline, as well as with OsO4.

Rat lung alveolar surfaces and contents were studied after using concanavalin A as a bifunctional agent to link exposed sugars to horseradish peroxidase, in accordance with a technique developed by Bernhard and Avrameas ('71). The Graham and Karnovsky ('69) diaminobenzidine procedure then was used to provide and electron-dense reaction product so as to define the distribution of complex carbohydrates in alveoli. A layer of very dense raction product was intimately associated with the outer leaflets of the luminal plasma membranes of type I and II pneumocytes. Masses of generally less dense reaction product extended irregularly into the alveolar coated by reaction product. When highly ordered tubular myelin bodies were seen, the reaction product filled all the "gutters" created by the intersentions of the membranes. Reasons are presented for believing that the intrinsic periodicity of the tubular myelin may be created and maintained by the domains ol technique. The distribution of autologous albumin, demonstrated by antibody staining by Bignon et al. ('75), apparently coincides with the carbohydrate pattern, suggesting both may be associated as a glycoprotein. The ultimate relationships between carbohydrate moieties and phospholipid membrane systems proved to be of such complexity that we believe it justifiable to think that they also may be truly complexed together.

Quantities of grids can be easily and reliably prepared with supporting nets of carbonized Parlodion if just a few principles and procedures are understood and practiced 1. Film nets are best cast on properly prepared, hydrophobic glass slides. Highly consistent results are achieved if slides receive a protracted soaking in a saturated solution of ferric stearate in benzene. Excess then is washed away, presumably leaving little but the adsorbed layer. 2. Nets are most reliably formed by exposing wet films of 0.4–0.6% Parlodion in amyl acetate to steam. 3. The same Parlodion solutions then can be painted on the slides to outline selected areas of nets. This serves the additional purpose of simplifying the floatation of the nets. 4. Grids with dried nets then require baking at 170–180°C to eliminate attenuated films which otherwise span the meshwork. 5. Finally the nets are heavily coated with carbon to achieve stability, and a wetting agent then can be used to render them hydrophilic.

Frog retinae, fixed only in buffered glutaraldehyde, were embedded for sectioning in glutaraldehyde polymerized with urea. In suitably thin sections globular substructures were seen in negative contrast after ionic staining with uranyl acetate and lead citrate, or after staining with neutralized phosphotungstic acid. Efforts to extract at least some of the lipid from sections before ionic staining enhanced the visualization of the “globules.” Exposure to KMnO 4 solution, used as an oxidative section stain, also outlined globular substructure in negative contrast, but with the additional feature that positively stained surface “leaflets” associated with the aqueous compartment were well defined. Staining sections with OsO 4 vapor resulted in positively stained membranes, but without any evident substructure. However, when sections which previously had been exposed to OsO 4 vapor were secondarily stained with uranyl acetate and/or lead citrate, positively stained globular substructures then were revealed. The globular substructures always were centered in the hydrophobic core region of the disc membranes, and symmetrically spanned the full thickness of this layer. The diameter of individual particles approximated 50–55 Å. Reasons are presented for the supposition that the evident globules incorporate at least hydrophobic components of rhodopsin molecules. Findings are discussed in relation to various models of disc membrane organization that have been proposed in recent years.

Glycol dehydration followed by rehydration prior to conventional fixation appears to demonstrate the essential identity of the thick filaments observed in unfixed, glycol dehydrated and conventionally fixed smooth muscle. Observed differences in the solubilities of actin and myosin filaments also suggest that the thick filaments of smooth muscle are not formed by the apposition of actin filaments or by the deposition of myosin upon actin filaments. Evidence that the thick filaments of smooth muscle are not formed by an unnatural aggregation of smaller myosin aggregates or by the dissociation of myosin “ribbons” during tissue preparation is also reported. Examinations of smooth muscle contracted or relaxed by pharmacological agents appear to indicate that the myosin content of smooth muscle is aggregated into filaments in both the contracted and relaxed cell.

Kidney tissue, incubated in a phosphate-sucrose buffer with diaminobenzidine (DAB), subsequently was embedded in polymerized glutaraldehyde-urea (Pease and Peterson, 1972). The highly polar character of this embedment retains lipids in ultrathin sections and thus permits a precise localization of reaction products in relation to cytomembranes. Furthermore, since conventional organic solvents are not used during processing, it is thought that oxidized DAB polymers certainly remain in place. Their density can be enhanced by exposing mounted sections to OsO4 vapor, rather than by en bloc staining. DAB oxidation takes place only in the compartment between the inner and outer mitochondrial membranes. When aldehyde-fixed tissue is incubated, the deposits are largely limited to the intracristal spaces, whereas when fresh tissue is incubated, the entire compartment is uniformly filled. Morphologic features of fresh, unfixed tissue are stabilized by ethylene glycol and so survive incubation best when about 30% of this substance is added to the medium.

A previous study demonstrated that tissue could be successfully infiltrated with 50% glutaraldehyde, and then subsequently polymerized with urea to create an embedment which retained cytomembrane lipids in sectioned material. As a result, the 180-190 Å periodicity characteristic of fresh, mammalian myelin was preserved in sections, as was a brilliant birefringence, and the capacity to bind OsO4 vapor in the hydrophobic bilayers. An associated (unpublished) study, carried out in co-operation with Drs. C.K. Akers and D.F. Parsons, demonstrated that the high concentration of glutaraldehyde (and urea) did not significantly alter the X-ray diffraction pattern of aldehyde-fixed, myelin. Thus, by itself, 50% glutaraldehyde has little effect upon cytomembrane systems and can be used with confidence for the first stages of dehydration.