key: cord-0044577-j2lx4fxt
authors: DOANE, FRANCES W.; ANDERSON, NAN
title: Electron and Immunoelectron Microscopic Procedures for Diagnosis of Viral Infections
date: 2013-11-17
journal: Human and Related Viruses
DOI: 10.1016/b978-0-12-429702-9.50019-3
sha: 2e6a46e3ccaf6c445493109397e3b186f659a6d3
doc_id: 44577
cord_uid: j2lx4fxt

nan

The small size of viruses places them beyond the limit of resolution of the light microscope (-250 nm) . Thus, they can be visualized only with the electron microscope, which provides a resolving power of approximately 0.3 nm. Although the first electron microscope was developed in the 1930's, the complexity of the instrument and the slow development of preparative methods for examining biological material delayed its application in the study of viruses. The first use of the electron microscope as a diagnostic instrument in virology was reported by Nagler and Rake (1948) and van Rooyen and Scott (1948) , who demonstrated that virus particles could be identified in clinical specimens from patients affected with smallpox, vaccinia, and varicella. Virologists were slow to adopt this approach, and it was more than a decade before the electron microscope was finally being used in the differential diagnosis of smallpox and chicken pox (Peters et al., 1962; Nagington, 1964; Cruickshank et al., 1966) .

Only recently has the electron microscope (EM) become generally recognized by virologists as more than a research instrument, but rather as an important tool in the rapid diagnosis of virus infections (Banatvala et al., 1975) . There now exists ample evidence that viruses can be detected by EM in clinical specimens from a wide variety of human and veterinary infections (Doane et al., 1967 (Doane et al., , 1969 Joncas et al., 1969; Spradbrow and Francis, 1969; Pavilanis et al., 1971; McFerran et al., 1971) .

Direct visualization of virus in clinical specimens obviously provides the most rapid diagnostic method. For reasons discussed later, it is not always practical to examine by EM every specimen received in a diagnostic laboratory, and many laboratories continue to carry out virus isolation in a host cell system, but identify the viral isolate by electron microscopy (Doane et al., 1969; Spradbrow and Francis, 1969; Pennington et al., 1975) .

Identification of a virus on the basis of its morphology bypasses the usual problem of loss of viral infectivity in a clinical specimen prior to isolation. Accordingly, by electron microscopy, it has been possible to demonstrate for the first time a number of elusive viruses that are difficult or impossible to culture, notably the viruses of subacute sclerosing panencephalitis (Bouteille et aL, 1965; Tellez-Nagel and Harter, 1966) , progressive multifocal leukoencephalopathy (ZuRhein and Chou, 1965; Howatson et aL, 1965) , rubella (Best et aL, 1967) , hepatitis ( B a y e r s aL, 1968), and acute gastroenteritis (Bishop et aL, 1973; Flewett et aL, 1973) .

By morphology alone, most viruses can be identified only to the level of their taxonomic family or genus. For example, enteroviruses cannot be distinguished from rhinoviruses, or mumps virus from parainfluenza virus ( Fig. 1) . Techniques are gradually being developed, however, within the field of immunoelectron microscopy, that offer a means of specific differentiation by serological typing directly on the EM specimen grid.

Ideally, an active virus laboratory should have its own electron microscopy facilities. In Toronto, three of the largest hospitals, as well as the Central Public Health Laboratories, have an EM Unit as an integral component of the virus diagnostic laboratory. The microscopes are in continuous use, both for direct examination of clinical specimens and as backup for identifying viruses isolated in cell culture. In the pediatric hospital, as many as 30-40 clinical specimens might be examined daily by EM during the height of the infantile gastroenteritis season.

To be useful for identifying viruses, the electron microscope should be capable of resolving better that 10-15 Â. It should also be possible for the operator to discern clearly ultrastructural details at a viewing magnification of 30,000 to 40,000 times, which provides a sufficiently enlarged field of view for detection of even small viruses (Doane et al., 1969; McFerran et al., 1971) .

The electron microscopist should have a basic technical knowledge of the operation and routine maintenance of the instrument. Thus, he should be capable of keeping the microscope optically aligned and at a maximum performance level. He should perform, or supervise, filament changes and routine cleaning of apertures and specimen holders. Preferably the instrument will be covered by a service contract, thereby reducing to a minimum the time lost for repairs and general servicing.

It is, of course, vital that the operator should be able to recognize viral ultrastructure. Most negatively stained specimens contain large quantities of cellular debris and small quantities of virus. An experienced eye is needed to differentiate between the two.

Where negative staining is the principle technique being used, it is advisable to have a simple vacuum evaporation unit for preparing carbon films on Formvar-coated specimen grids. As quantities of these can be prepared 2 to 3 weeks before use, it is not essential to have an evaporator located in the laboratory itself.

If thin sectioning techniques are to be used in the EM Unit, several major items of equipment are necessary. These include an ultramicrotome, a knife breaker, and an embedding oven. It should be recognized that thin sectioning techniques require a considerable amount of technical skill and experience, and it may be advisable to have this phase of the work performed by a service laboratory, if one is available. The only task then remaining for the electron microscopist is to examine the prepared thin sections.

Photography is an essential adjunct to the electron microscope, and photographic facilities should be readily available for the virology EM Unit. The microscopist should personally process his own films, plates, and micrographs, or at least supervise their processing. (Because of the low cost, ease of handling and storing, and high resolution of roll films, this format is well-suited to the diagnostic laboratory). A photographic enlarger should also be available; it can be equipped with contrast filters if the convenient polycontrast photographic paper is used. It is our experience that photographic processors greatly reduce the work load in the EM Unit, and yet these instruments produce micrographs of high quality. In our own laboratory, all micrographs are prepared by processors (15 seconds from enlarger to print!), rather than by the standard tray method of development.

The scanning electron microscope (as opposed to the better known transmission electron microscope) is still in its infancy with respect to its application to diagnostic virology. It appears to be useful in monitoring cell cultures for mycoplasma contamination (Section VI), but at present it is essentially a research instrument in virology.

Negative staining of clinical specimens provides the simplest and most rapid method for virus detection. The main limitation to this approach relates to the low concentration of virus particles in many clinical specimens. Depending on the negative staining method used, between 10 7 and 10 9 particles per milliliter must be present in the original specimen in order to be detected on the grid (Anderson and Doane, 1972b) . Because of the high particle-to-infectivity ratio observed with many viruses, this may represent less than 100 TCID 50 in terms of infectivity (Doane et al., 1967; Valters et al., 1975) .

Some of the more fragile viruses may be structurally altered by the negative staining process. Although this may result in structural artifacts, from a diagnostic point of view this may actually be an advantage; for example, it is usually easier to detect a ruptured paramyxovirus with its uncoiled nucleocapsid than an intact virion with the envelope unpenetrated by stain (see Fig. 15A ).

Phosphotungstic acid (PTA) serves as an excellent general purpose negative stain. It is used as a 2% aqueous solution, raised to pH 6-7 with a few drops of l Ν KOH. The water should be filtered to remove physically any bacteria that might be present. It is convenient to store the PTA solution in a small syringe at 4°C and to dispense it by drop from the syringe needle.

Copper grids of 200 to 400 mesh can be used, with 300-mesh grids providing optimal support and open area. Although bare Formvar support films have been used successfully for negative staining by some workers (Palmer et al., 1975) , we prefer a film of 0.3% Formvar in ethylene dichloride, stabilized by a very thin layer of evaporated carbon.

Specimens that contain little or no salt can be added directly to the grid and negatively stained without preliminary treatment. A fine-bore Pasteur pipette is used to add a drop of the specimen; a drop of negative stain is then added. Excess fluid is removed by bringing it briefly in contact with a torn edge of filter paper, then allowing it to air-dry (1-2 minutes).

Most fluid specimens contain a high concentration of salt which, if allowed to dry on the specimen grid, will crystallize and obliterate viral particles during EM examination. To circumvent this problem, a simple dialysis method can be used for specimens that contain at least 10 9 virus particles per milliliter (Doane et al., 1969) . One microdrop of specimen is placed on a drop of sterile distilled water resting on a waxed surface. A coated grid is touched briefly to the surface of the drop. Negative stain is then added, excess fluid is removed with filter paper, and the grid is airdried and examined.

This is a modification of the method of Kelen et al. (1971) . It is useful for salty specimens, especially when the virus concentration is low (Anderson and Doane, 1972b) . The limit of detectability by the agar diffusion method is approximately 10 7 virus particles per milliliter of original sample. Cups of disposable microtiter plates are approximately threefourths filled with 1% aqueous Noble agar, and a 300-mesh coated specimen grid is placed on the solid agar surface of each cup. Plates so prepared can be used immediately, or they can be covered tightly with strips of adhesive sealing tape and stored at 4°C for several weeks. To prepare a specimen for examination, one to two drops are added to a grid and allowed to air-dry (15-30 minutes). A drop of negative stain is then added. After 1 minute the grid is removed with forceps, aid-dried briefly, and examined by EM.

Small pieces of tissue (1 mm 3 ) are placed in a metal planchet and are frozen and thawed four to six times by alternately touching the planchet to Dry Ice and to the palm of the hand (Doane et al. y 1969) . A drop of filtered distilled water is then added and the tissue is once more frozen and thawed. The resulting lysed tissue suspension is briefly mixed with a fine-bore Pasteur pipette and a small drop is placed on a coated grid. The specimen is then negatively stained and air-dried.

Whenever possible, duplicate grids should be prepared of every specimen; one might be prepared by the direct application method, for immediate examination, and the other by the agar diffusion method, to serve as a backup grid. In general, no more than 15 minutes need be spent in examining a grid, unless it is a selected case, e.g., associated with smallpox. Screening is carried out at viewing magnifications of approximately 40,000 times. A single magnification should be selected as a routine, so that the operator can become familiar with the relative size of particles being observed. (On one of the microscopes in our laboratory, a small circle on the viewing screen spans approximately 100 nm at a magnification of 35,000x.)

Virus particles are usually recognized by their ability to trap the negative stain at their outer surface. Thus virus particles usually stand out more clearly than background debris. This enables large viruses, such as myxoviruses, to be detected even at low magnification ( Fig. 2) . Several areas of the grid should be examined, and with experience, one learns to select moderately dense squares under very low magnification, then proceed to examine those squares at high magnification.

Specimens from vesicles produced by herpesviruses or poxviruses provide excellent material for direct examination because of their high content of virus (Fig. 3) . The use of the electron microscope in smallpox diagnosis was first indicated, in 1948, by van Rooyen and Scott and by Nagler and Rake; it is probably in the differential diagnosis of chicken pox and smallpox* that the electron microscope is best known in diagnostic virol- ogy. Thus electron microscopy now provides the method of choice in the rapid detection of poxviruses and herpesviruses (Peters et al., 1962; Almeida et al., 1962; Smith and Melnick, 1962; Nagington, 1964; Cruickshank et al., 1966; Macrae et al., 1969; Bland et al., 1970; Gibbs et al., 1970) , although for specific identification it is still necessary to compliment morphological detection with virus isolation and identification procedures (Long, 1970) .

Coxsackie A16 virus, the etiological agent of "hand, foot and mouth disease" (Robinson et al., 1958; Alsop et al., 1960) can also be detected by direct EM examination of vesicle fluid (P. J. Middleton and M. T. Szymanski, personal communication).

Vesicle fluid is collected in a capillary tube, or into a fine-bore needle attached to a small syringe. If the volume is sufficient, grids are prepared both by direct application and by the agar diffusion method. If the amount of fluid is limited, the agar diffusion method is recommended. Crusts removed from dried vesicles provide an excellent source of viruses (P. J. Middleton and M. T. Szymanski, personal communication) . The excised crust is placed, underside facing down, in a drop of 1% unbuffered aqueous ammonium acetate on a glass slide. A smear is made by grinding the crust against the side by means of a scalpel blade. One or two more drops are added to the smear, and a coated grid is touched briefly to the fluid, then negatively stained and examined. Smears can also be made from vesicle fluid, or from scrapings collected from the base of a vesicle with a 25-gauge needle. Dried smears are then resuspended in two to three drops of filtered distilled water or 1% ammonium acetate.

Nasopharyngeal secretions collected by the suction method of Auger (1939) have been found suitable for direct EM detection of influenza, parainfluenza, and respiratory syncytial viruses ( Fig. 1 ) (Doane et al., 1967; Joncas et al., 1969) . For examination, the aspirated nasopharyngeal secretion is diluted five-to tenfold in balanced salt solution, bacteria are deposited by low-speed centrifugation, and the clarified specimen is prepared for negative staining by the agar diffusion or water drop method. Following similar procedures, herpesvirus particles have been found in tracheobronchial suctions from a child with herpes simplex pneumonia (P. J. Middleton and M. T. Szymanski, personal communication) .

Throat washings or gargles rarely contain sufficient virus to be detected by electron microscopy. Lipman et al. (1975) observed Epstein-Barr virions in a single throat washing from a patient with infectious mononucleosis, but only after a 23 x concentration and purification by density gradient centrifugation. Nasal secretions from acute upper respiratory Evans and Melnick (1949) reported the EM detection of herpesviruses in cerebrospinal fluid (CSF) from a patient with herpes zoster. In our own laboratory, mumps virus was observed in negatively stained CSF from a patient with suspected mumps encephalitis ( Fig. 1) (Doane et aL, 1967) . Our overall results, however, indicate that CSF rarely contains sufficient quantities of virus to be detected by direct examination. The agar diffusion method is recommended for negative staining of CSF.

A variety of viruses and viruslike particles have been found in negatively stained fecal specimens by electron microscopy (Fig. 5) . Included in the rapidly growing list are coronaviruses (Caul et aL, 1975) , adenoviruses (Anderson and Doane, 1972b; Flewett et aL, 1974b) reolike viruses (Doane et aL, 1969; Anderson and Doane, 1972b; Flewett et aL, 1973 Flewett et aL, , 1974b Middleton et aL, 1974) , enteroviruses (P. J. Middleton and M. T. Szymanski, personal communication), hepatitis A virus (Feinstone et aL, 1973) , bacterial viruses (Flewett et aL, 1974b) , and miscellaneous unclassified viruses (Kapikian et aL, 1972b; Paver et aL, 1973; Madeley and Cosgrove, 1975) , as well as "pleomorphic viruslike particles" Mathan et al., 1975) . (For a comprehensive review of agents associated with acute gastroenteritis, see Chapter 11, Volume I.)

Ultracentrifuged deposits (UCD) of stool suspensions (clarified of bacteria by light centrifugation) have been used for many years in this laboratory. The UCD is resuspended in two to three drops of water prior to negative staining. A more satisfactory technique for processing large numbers of specimens, recommended by M. T. Szymanski and P. J. Middleton (personal communication), is to resuspend a small portion of stool in 1% ammonium acetate without any preliminary clarification or concentration. If difficulty is encountered in negative staining the specimen due to, e.g., excess mucus, a smear of the specimen is made on a glass slide. This is allowed to air-dry and is then resuspended in water or ammonium acetate, negatively stained, and examined. Smears of fecal swabs can also be used with considerable success.

It appears that viruses are not readily detected in urine. Efforts to visualize cytomegalovirus (CMV) are generally discouraging. However, Montplaisir et al. (1972) found CMV by centrifuging 30 to 50 ml volumes of urine, at 82,000 g for 90 minutes, then resuspending the pellet in a few drops of water, followed by negative staining. Papovaviruses and adenoviruses have only rarely been observed in urine (P. J. Middleton and M. T. Szymanski, personal communication).

It would appear that blood is not reliable for direct EM detection of viruses. Hepatitis Β antigen has been observed in the serum, but a more sensitive method for visualizing the components of this antigen is provided by immunoelectron microscopy.

Skin biopsies can be used to demonstrate molluscum contagiosum (Fig. 6) , or virus from contagious pustular dermatitis, and papovaviruses from warts (Macrae etal., 1969; Strausshof/., 1949; Williams et al., 1961) . The papovavirus associated with progressive multifocal leukoencephalopathy (PML) has been found by negative staining of brain tissue (Howatson et al., 1965) , and Deutsch and Spence (1972) detected hepatitis Β virus in liver specimens.

Although negative staining provides the most rapid method for examining a clinical specimen, the thin sectioning technique is usually more reliable when the clinical specimen is in the form of autopsy or biopsy tissue. Viruses that have been demonstrated, for the first time, as a result of this technique, include the measles-like virus of subacute sclerosing panencephalitis (SSPE) found in brain cell nuclei (Bouteille et al., 1965; Tellez-Nagel and Harter, 1966) (Fig. 7) , and the reovirus-like particles of acute gastroenteritis, first revealed in thin sections of duodenal mucosa biopsy tissue (Bishop et al., 1973) . Thin sectioning has also been used to demonstrate papovaviruses in brain tissue from progressive multifocal leukoencephalopathy (PML) (Zu Rhein and Chou, 1965; Dolman et al., 1967) and herpesviruses in brain from herpetic encephalitis (Harland et al., 1967) .

Useful retrospective EM studies can be done with autopsy material regardless of fixation (Burns et al., 1975) , and techniques have been described for processing paraffin-embedded sections to permit visualization of intracellular virus particles (Morecki and Becker, 1968; Blank et al, 1970) .

Using standard histological procedures normally employed for electron microscopy, sections of tissue are not generally available until at least 36-48 hours after receipt of the specimen. It has been found that these procedures can be abbreviated considerably, with little loss in ultrastructural detail, to provide sections within 2 to 3 hours  Rowden and Lewis, 1974) (Fig. 8) . The rapid embedding method used in this laboratory is described below.

Tissue specimens not exceeding 1 mm in thickness are collected in cold 2.5% glutaraldehyde in Millonig phosphate buffer. A minimum of 15 minutes is allowed for fixation; however, tissues can safely be stored in glutaraldehyde for several days. Primary fixation is followed by three rinses (1 minute each) in phosphate buffer, and a second fixation of 15 minutes at room temperature in 1% osmium tetroxide in phosphate buffer.

Fixed specimens are dehydrated through acetone as follows: 70% acetone, two changes in 3 minutes; absolute acetone, three changes in 5 minutes. After 10 minutes in a 1:1 mixture of absolute acetone and epoxy embedding plastic (either Spurr's medium, or a mixture of Araldite and Epon), followed by two changes of 100% plastic (5 minutes each), the specimen is placed in fresh plastic in an embedding capsule and heated to 95°C for 60 minutes to achieve polymerization of the plastic. The total processing time is 2 hours. Embedded tissue processed by this method can be trimmed for sectioning within 15 minutes after removal from the oven. Thin sections can be cut with either glass or diamond knives, and are collected on uncoated 300-mesh copper specimen grids prior to staining with lead citrate and uranyl acetate.

Undoubtedly the most useful application of the electron microscope in routine diagnostic virology is in the identification of viruses isolated in tissue culture. Because of the impracticability of performing direct EM examination on every clinical specimen submitted for virus studies, most virus laboratories with EM facilities continue to rely on cell cultures for virus isolation, but perform a presumptive morphological identification of the viral isolate by negatively staining cultures showing a cytopathic effect (CPE) (Doane et al., 1969; Spradbrow and Francis, 1969; Pennington et al., 1975) . Viral isolates can readily be identified as to group by electron microscopy; subsequent specific identification can then be performed by standard serological methods.

The amount of virus in a clinical specimen may be well below the limit of EM detectability, making direct examination useless. Following passage in tissue culture, however, the virus concentration can be amplified to yield large quantities of virus that can easily be detected using negative staining techniques. As shown in Table I , some viruses isolated in cell cultures can be identified by electron microscopy as much as 2 days before the appearance of a cytopathic effect (Doane and Anderson, 1972) .

EM identification of viruses isolated in organ cultures was used by Almeida and Tyrrell (1967) in their discovery of human coronaviruses. Marsolais et al. (1971) reported that this approach permitted a rapid diagnosis of avian coronaviruses isolated in embryonated hen's eggs.

A single tube culture of infected cells provides ample material for EM examination. Although we have found virus in negatively stained infected cultures showing no cytopathic effect, as well as those showing complete cell destruction, in general, it is preferable to work with cultures showing a well-developed CPE (Figs. 9-12) . Viruses, such as orthomyxoviruses and paramyxoviruses, can often be detected in the culture medium, simply by withdrawing a few drops and negatively staining by the water drop or agar diffusion methods (Sections III,C,2 and 3). As a routine, however, it is more reliable to negatively stain a sample of the cells, especially if the cytopathic effect has not progressed to involve the entire cell sheet. Furthermore, with viruses that are markedly cell-associated, such as adenoviruses and reoviruses, virus particles remain in the cells long after complete destruction of the culture.

To process the cells for negative staining, the culture medium is withdrawn and held temporarily, and the cells are scraped with a Pasteur pipette into two or three drops of filtered distilled water, thereby lysing the cells. A drop of this lysate is then negatively stained by the water drop method. In the majority of cases this procedure is effective in revealing viral particles if they are present in the cells. If no virus is found in the cell lysate, however, the culture medium is processed both by the water drop and agar diffusion methods. Enveloped viruses such as togaviruses and oncornaviruses are often so severely ruptured by negative staining that they are impossible to recognize by electron microscopy. P. J. Middleton and M. T. Szymanski (personal communication) have overcome this problem by adding a few drops of 2.5% buffered glutaraldehyde to the drained culture for 1 to 2 hours. The cells are then scraped into the fixative and added to a grid for negative staining (Fig. 13 ).

Although negative staining obviously provides the fastest method for identifying viruses isolated in a laboratory host system, thin sectioning may enable the cells or tissues to be examined in a more systematic fashion. This is an important consideration when the viral isolate is a togavirus (Fig. 14) or oncornavirus, which may be difficult to observe in negatively stained preparations, due either to a low particle concentration or to disruption of viral ultrastructure during preparation for electron microscopy. For preparing infected cell cultures for thin sectioning, the following method requires only a single tube culture . Cells are gently resuspended in the culture medium, transferred to a small conical-tipped centrifuge tube, and pelleted in a clinical centrifuge at 1500 rpm for 3 minutes. The medium is discarded, leaving only two to three drops in the tube to allow for transfer of the cells to a flat waxed surface (e.g., Parafilm). A fine-bore glass tube, 1.3 x 75 mm, is touched to the drop to draw in the cell suspension by capillary action. One end is sealed with Plasticine and the tube is centrifuged in a hematocrit centrifuge for 3 minutes at 12,500 rpm. The cells now form a compact pellet immediately above the Plasticine plug. The tube is scored and broken at a distance of 6 to 7 mm above the cell pellet. The tube is then inverted, and a blunt wire (paper clip) of diameter slightly narrower than the bore of the tube is used to push against the Plasticine, forcing the cell pellet into a vial of fixative. The cell pellet at this and subsequent stages remains tightly packed, and can be transferred easily in the tip of a Pasteur pipette. It can be processed either by standard EM embedding methods or by the rapid embedding method (Section IV,B).

Cell cultures are now firmly established as the prime host cell system for the isolation and identification of viruses and for the investigation of virus-cell interactions. Concurrent with their widespread use, however, is the increased recognition that cell cultures are often contaminated with adventitious agents. These may be viruses originating from the host, or they may be viruses or mycoplasma acquired in the laboratory during continuous cultivation. In most instances these contaminants are extremely insidious, causing little or no cytopathic effect; consequently, their presence can often be detected only by electron microscopy.

One of the most notorious groups of contaminants are the simian viruses, which may occur in over 50% of "normal" monkey kidney cell cultures (Hull, 1968; Hsiung, 1968; Anderson and Doane, 1972a) . Commonly encountered simian viruses include paramyxoviruses, the papovavirus SV40, foamy virus, and cytomegalovirus. Mixed infections in a single batch of monkey kidney cells are not uncommon (Hsiung and Atoynatan, 1966) .

A possible source of contamination is bovine serum, commonly incorporated as a nutrient supplement into most tissue culture media. Fong et al. (1975) used electron microscopy techniques to screen 25 lots of bovine serum, and found viruslike particles in 17 lots (68%). Sera were concen- trated 100-fold or more by centrifugation. It was difficult to identify viruses in negatively stained serum; however, when sera were pelleted by ultracentrifugation into Beem capsules, and the pellets were processed for thin sectioning, viruses were found in all 17 lots. Routine monitoring of cell cultures by negative staining, and preferably also by thin sectioning, serves to keep in check the spread of adventitious agents in stock cultures and also in virus pools prepared in contaminated cultures (Anderson and Doane, 1972a; de Harven, 1973) (Figs. 15-18) .

Detection of mycoplasma is often difficult by transmission electron microscopy. Although negative staining provides the most rapid technique, it requires considerable experience to be able to distinguish the pleomorphic microorganisms from normal cellular material (Wolanski and Maramorosch, 1970) . Mycoplasma can also be recognized in thin sectioned cell cultures (de Harven, 1973) (Fig. 18) . Recent reports suggest that the scanning electron microscope offers a sensitive method for detection of mycoplasma contamination of cell cultures (Brown et aL, 1974; Doane and Anderson, 1975) (Fig. 19) .

Negative staining methods used for screening cell cultures are similar to those described in Section V. For thin sectioning, cells are fixed in 2.5% glutaraldehyde in phosphate buffer, followed by 1% osmium tetroxide in phosphate buffer, dehydrated either with acetone or alcohol, and embedded in epoxy resin, either by standard EM histological procedures or by the rapid embedding method (Section IV,B).

Cultures to be checked by scanning electron microscopy (SEM) for mycoplasma contamination are grown on glass cover slips, preferably for several days, to allow the mycoplasma (if present) to become well established. They are then processed by the following method: cultures are fixed at 4°C in 2.5% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.2) for at least 18 hours. They are then rinsed in three changes of 0.1 M cacodylate buffer, and immersed in 1% osmium tetroxide in 0.1 M cacodylate buffer (pH 7.2) for 30 minutes at room temperature. Fixed cultures are then rinsed in cacodylate buffer, passed through 5-minute changes of 50, 70, and 95% alcohol, three changes of absolute alcohol, then quickly transferred to a critical point dryer, where they are dried by passing through the critical point of C0 2 . In preparation for SEM examination, dried cover slips are mounted on SEM stubs by means of silver paint, and coated with gold-palladium.

Immunoelectron microscopy (IEM) is the direct visualization by electron microscopy of an antigen-antibody complex, and was first reported in 1941 by Anderson and Stanley and by von Ardenne et aL Its usefulness in the differentiation of viral antigens owes much to the work of Almeida and her colleagues who recognized the diagnostic potential of the technique as a means of serotyping viruses directly on the EM grid (Almeida et aL, 1963; Almeida and Waterson, 1969) .

IEM has proved invaluable in the detection and identification of elusive viruses such as rubella virus (Best et aL, 1967) , coronaviruses , and rhinoviruses (Kapikian et aL, 1972a) , and of viruses that are difficult or impossible to culture, such as wart viruses (Almeida and Goffe, 1965) , hepatitis virus (Bayer et aL, 1968; Feinstone et aL, 1973) , and the reovirus-like particles associated with gastroenteritis (Flewett et aL, 1974a) . In the serotyping of viruses within a major virus group, IEM has been successfully applied to the differentiation of papovaviruses (Gardners aL, 1971; Penney et aL, 1972; Penney and Narayan, 1973) , enteroviruses (Anderson and Doane, 1973) , myxoviruses and paramyxoviruses (Kelen and McLeod, 1974) , and adenoviruses (Luton, 1973; Vassall and Ray, 1974; Edwards et aL, 1975) . For a more detailed survey of the several applications of IEM, the reader is directed to reviews by Almeida and Waterson (1969) , and Doane (1974) .

Viruses in clinical specimens can be detected by IEM by mixing with homologous antibody-in standard typing serum or in the patient's own convalescent serum (Fig. 20) . This approach has been valuable in demonstrating viruses that are present in specimens at too low a concentration to be detected by direct EM examination (Paver et aL, 1973) . No preliminary clarification or concentration of the specimen is needed for IEM if the virus is present at detectable levels. In general, however, a clearer preparation is obtained if a suspension of the specimen is clarified of bacteria and debris by centrifugation for 30 minutes at approximately 8000 g prior to IEM and negative staining.

Greater sensitivity is obtained by more time-consuming methods such as ultracentrifugation or density gradient centrifugation (Flewett et al., 1974b) . Devine and Lee (1975) detected polio virus particles in stool at a concentration of 104/ml when they combined IEM with virus concentration by the polyelectrolyte PE60. This compared with a limit of EM detection of 106 when PE60 was used alone.

Crude infected cell lysates can be used in identifying viral isolates by IEM; no concentration or purification of virus is necessary (Fig. 21) . Sensitivity tests by Valters et al. (1975) on IEM typing of adenovirus showed that as few as 16 to 32 TCID50 per 0.5 ml could be detected. Anderson and Doane (1973) were able to serotype as few as 103·5 TCID50 per milliliter of poliovirus from infected culture lysate, representing a detection sensitivity of approximately 100-fold over that required for EM detection in the absence of antibody. 

Whether working with standard viral antisera or patients' convalescent sera, unfractionated serum is satisfactory for most IEM tests in diagnostic virology. Where problems are encountered with cloudy serum, or where fine resolution of the immune complex is required, the serum should be clarified by centrifugation at 40,000 rpm for 1 hour (Almeida and Waterson, 1969) . Even greater definition can be achieved by working with the globulin fraction (Lafferty and Oertelis, 1963; Bayer and Mannweiler, 1963; Mandel, 1971) . Almeida and Waterson (1969) recommend working with heat-inactivated sera to avoid complications arising from the presence of complement.

In serotyping enteroviruses and adenoviruses, antisera have been used either individually or in pools, with similar results (Anderson and Doane, 1973; Luton, 1973; Vassall and Ray, 1974) . The concentration of antibody used, relative to the concentration of virus, may affect the specificity of the test. Enteroviruses can be distinctly differentiated by IEM when typing sera are used diluted (generally at 50 or greater) (Anderson and Doane, 1973) but cross-reactions occur when sera are used undiluted, or at very low dilutions (Doane, 1974) . Similar problems of cross-reactivity with myxoviruses have been reported by Kelen and McLeod (1974) .

Antiserum concentration also affects the appearance of the immune complex. When there is little antibody present in relation to viral antigen, aggregates are small, consisting of only a few clearly outlined virus particles. As the concentration of antibody increases, aggregates become larger and more numerous, the interconnecting antibody layers become thicker, and the surface details of the virus particles become more obscure (Fig. 21) . At high concentrations of antibody, individual virus particles are surrounded by a halo of antibody, and the incidence of aggregates is greatly reduced.

C. Methods

The most commonly used method for serotyping by IEM is the direct mixing of virus and serum, recommended by Almeida and Waterson (1969) . The viral antigen is reacted at 37°C for 1 hour with an equal volume of antiserum, then left overnight at 4°C. The following morning the mixture is spun at 10,000 rpm fori hour to sediment immune complexes. For smaller viruses, e.g., picornaviruses, centrifugation of at least 15,000 rpm is recommended, whereas vaccinia-antibody complexes will sediment using a clinical-type bench centrifuge. The resultant pellet is usually resuspended in a small volume of distilled water prior to negative staining and examination in the electron microscope. Some workers have shortened the entire procedure to 1 hour incubation at 37°C (Lutton, 1973; Vassal and Ray, 1974; Doane, 1974) . Edwards et al. (1975) and Valters et al. (1975) described a more sensitive indirect IEM method, whereby the virus-antibody complexes were further aggregated by anti-immune globulin sera (anti-IgG). When applied to serotyping of adenoviruses, it was from 4 to over 32 times more sensitive than the direct method, indicating its value not only in differentiation of viruses, but also as a sensitive method for revealing new viruses (e.g., when used with patients' serum).

The indirect method, as used by Valters et al. (1975) for detecting adenovirus type 7 in throat swab specimens (and which identified 19 out of 25 adenovirus-positive specimens), is as follows: ί ml of diluted adenovirus antiserum was mixed with an equal volume of throat swab fluid and incubated for 1 hour at 37°C and then for 3 hours at 6°C. The mixture was centrifuged at 12,000 g for 30 minutes, and the pellet was suspended in 0.5 ml of phosphate-buffered saline. An equal volume of anti-IgG serum, diluted to the optimal concentration (as determined by a chess board titration of adenovirus type 7 antiserum and anti-IgG) was added, and this mixture was incubated at 6°C for 16 to 20 hours. The mixture was then centrifuged at 12,000 g for 30 minutes, the pellet was suspended in 0.1 ml of distilled water, negatively stained, and examined by EM. Kelen et al. (1971; Kelen and McLeod, 1974) have described a practical microtechnique for demonstrating hepatitis Β antigen-antibody complexes and for serotyping myxoviruses and paramyxoviruses. Virus and antiserum are incubated together at 37°C for i hour. A microdrop of the mixture is then deposited on the surface of a standard microscope slide that has previously been covered with 5 ml of 0.8% agar. Immediately thereafter, a Formvar-carbon coated grid is placed upside down and left floating on top of the drop. When diffusion of the fluid phase into the agar is complete (a matter of a few minutes), the grid is removed for negative staining and EM examination. Specimens prepared by the ADF method are both partially purified and concentrated, the agar acting as a molecular sieve and effectively removing interfering salts, maeromolecules, and tissue debris of 15 nm in diameter, while retaining the virus particles on the surface.

A further modification of the ADF method has been developed by Anderson and Doane (1973) for serotyping of enteroviruses by IEM. In this method, the antisera are incorporated in the agar itself.

Dilutions of single or pooled typing sera are added to a cooled (56°C) molten solution of 1% aqueous agar, which is then pipetted into the cups of disposable microtiter plates. Once the agar has solidified at room temperature, Formvar-carbon coated grids are placed on the agar surface. Viral specimens to be typed by IEM are added to the grids in volumes of 1 to 2 microdrops, and allowed to air-dry (approximately 30 minutes). Preparations are then negatively stained and examined by electron microscopy.

The SIA method is rapid and simple to perform, and requires only small quantities of virus and serum. Although the antisera are contained within the agar itself, homologous antibodies apparently diffuse rapidly into the viral specimen, resulting in the formation of immune complexes. The sensitivity and specificity of this method for enterovirus typing has been found to be comparable to that obtained using the direct method. From a practical point of view, however, the SIA method is more useful in that cups can be stored at 4°C for several months prior to use, and cups containing diffeïent types of antisera can be color-coded for reference.

All too often, little attention is paid to safety precautions that should be taken while examining pathogenic viruses. A report on "Classification of Etiologic Agents on the Basis of Hazard," compiled by the United States Department of Health, Education and Welfare (1974) , lists the following viruses as Class 4 agents, viz., "that require the most stringent conditions for their containment because they are extremely hazardous to laboratory personnel or may cause serious epidemic disease": Alastrim, smallpox, monkey pox, and whitepox, when used for transmission or animal inoculation experiments Hemorrhagic fever agents, including Crimean hemorrhagic fever (Congo), Junin, and Machupo viruses, and others as yet undefined

Herpesvirus simiae (Monkey Β virus) Lassa virus Marburg virus Tick-borne encephalitis virus complex, including Russian spring-summer encephalitis, Kyasanur forest disease, Omsk hemorrhagic fever, and Central European encephalitis viruses Venezuelan equine encephalitis virus, epidemic strains, when used for transmission or animal inoculation experiments Yellow fever virus-wild, when used for transmission or animal inoculation experiments A similar report on "Laboratory Use of Dangerous Pathogens" has been published by the Department of Health and Social Security of Great Britain (1975) . Microscopists working with any of these pathogens are advised to avail themselves of these reports.

Maximum precautions should be taken when preparing specimen grids for electron microscopy. The production of aerosols, through operations such as grinding or sonication, should be avoided, and all stages of specimen preparation should be carried out in a safety cabinet. Potentially hazardous specimens, such as those from suspected smallpox, should be irradiated by ultraviolet light prior to EM examination. There is a paucity of information in the published literature concerning optimum conditions necessary for inactivation of viral infectivity. Personnel working with these specimens should be vaccinated against smallpox at least every 3 years. In the event of smallpox virus being detected in specimens, all associated personnel should immediately be revaccinated.

Because of the highly infectious nature of hepatitis Β antigen, serum suspected of containing this agent should be treated with ß-propiolactone prior to EM examination. The method recommended by L. Spence and M. Fauvel (personal communication) is as follows: to 1 ml of serum add 0.1 ml of 3% saline solution of /3-propiolactone. Incubate at 37°C for 4 hours. Add 1 drop of 5 Ν NaOH to bring the pH to neutrality.

Forceps used to handle specimen grids should be flamed thoroughly after use, cooled, dipped in alcohol, and flamed again. Grids to be discarded should be placed in specially marked containers and autoclaved immediately, care being taken to ensure that the steam can penetrate the container. 

Proc. Natl. Acad. Sei. U.S.A. 71

Contamination in Tissue Culture

Laboratory Use of Dangerous Pathogens

Viral Immunodiagnosis

Viral Immunodiagnosis

Lancet 2, 182

Classification of Etiologic Agents on the Basis of Hazard

Arch. Gesamte Virusforsch