Ultrastructural studies can be considered under three areas: physical methods, chemical methods, and electron microscopy. Physical measurements of virus particles began in the 1930s with the earliest determinations of their proportions by filtration through colloidal membranes of various pore sizes. Experiments of this sort led to the first (rather inaccurate) estimates of the size of virus particles. The accuracy of these estimates was improved greatly by studies of the sedimentation properties of viruses in ultracentrifuges in the 1960s (Figure 1.4). Differential centrifugation proved to be of great use in obtaining purified and highly concentrated preparations of many different viruses, free of contamination from host cell components, that can be subjected to chemical analysis. The relative density of virus particles, measured in solutions of sucrose or CsCl, is also a characteristic feature, revealing information about the proportions of nucleic acid and protein in the particles.
Figure 1.4 A number of different sedimentation techniques can be used to study viruses. In rate-zonal centrifugation (shown here), virus particles are applied to the top of a preformed density gradient, i.e., a sucrose or salt solution of increasing density from the top to the bottom of the tube (top of figure). After a period of time in an ultracentrifuge, the gradient is separated into a number of fractions, which are analysed for the presence of virus particles. In the figure, the nucleic acid of the virus genome is detected by its absorption of ultraviolet light (below). This method can be used both to purify virus particles or nucleic acids or to determine their sedimentation characteristics. In equilibrium or isopycnic centrifugation, the sample is present in a homologous mixture containing a dense salt such as caesium chloride. A density gradient forms in the tube during centrifugation, and the sample forms a band at a position in the tube equivalent to its own density. This method can thus be used to determine the density of virus particles and is commonly used to purify plasmid DNA.
The physical properties of viruses can be determined by spectroscopy, using either ultraviolet light to examine the nucleic acid content of the particle or visible light to determine its light-scattering properties. Electrophoresis of intact virus particles has yielded some limited information, but electrophoretic analysis of individual virion proteins by gel electrophoresis, and particularly also of nucleic acid genomes (Chapter 3), has been far more valuable. However, by far the most important method for the elucidation of virus structures has been the use of x-ray diffraction by crystalline forms of purified virus.This technique permits determination of the structure of virions at an atomic level.
The complete structures of many viruses, representative of many of the major groups, have now been determined at a resolution of a few angstroms (Å) (see Chapter 2 ).This advancement has improved our understanding of the functions of the virus particle considerably; however, a number of viruses have proven to be resistant to this type of investigation, a fact that highlights some of the problems inherent in this otherwise powerful technique. One problem is that the virus must first be purified to a high degree; otherwise, specific information on the virus cannot be gathered. This presupposes that adequate quantities of the virus can be propagated in culture or obtained from infected tissues or patients and that a method is available to purify virus particles without loss of structural integrity. In a number of important cases, this requirement rules out further study (e.g., hepatitis C virus). The purified virus must also be able to form paracrystalline arrays large enough to cause significant diffraction of the radiation source. For some viruses, this is relatively straightforward, and crystals big enough to see with the naked eye and which diffract strongly are easily formed. This is particularly true for a number of plant viruses, such as tobacco mosaic virus (which was first crystallized by Wendell Stanley in 1935) and turnip yellow mosaic virus (TYMV), the structures of which were among the first to be determined during the 1950s. It is significant that these two viruses represent the two fundamental types of virus particle: helical in the case of TMV and icosahedral for TYMV (see Chapter 2). In many cases, however, only microscopic crystals can be prepared. A partial answer to this problem is to use ever more powerful radiation sources that allow good data to be collected from small crystals. Powerful synchotron sources that generate intense beams of radiation have been built during the last few decades and are now used extensively for this purpose; however, there is a limit beyond which this brute force approach fails to yield further benefit. A number of important viruses steadfastly refuse to crystallize; this is a particularly common problem with irregularly shaped viruses—for example, those which have an outer lipid envelope—and to date no complete high-resolution atomic structure has yet been determined for many viruses of this type (e.g., HIV). Modifications of the basic diffraction technique (such as electron scattering by membrane-associated protein arrays and cryo-electron microscopy) may help to provide more information in the future, but it is unlikely that these variations will solve this problem completely. One further limitation is that some of the largest virus particles, such as poxviruses, contain hundreds of different proteins and are at present too complex to be analysed using these techniques.
Nuclear magnetic resonance (NMR) is increasingly being used to determine the atomic structure of all kinds of molecules, including proteins and nucleic acids. The limitation of this method is that only relatively small molecules can be analysed before the signals obtained become so confusing that they are impossible to decipher with current technology. At present, the upper size limit for this technique restricts its use to molecules with a molecular weight of less than about 30,000 to 40,000, considerably less than even the smallest virus particles. Nevertheless, this method may well prove to be of value in the future, certainly for examining isolated virus proteins if not for intact virions.
Chemical investigation can be used to determine not only the overall composition of viruses and the nature of the nucleic acid that comprises the virus genome but also the construction of the particle and the way in which individual components relate to each other in the capsid. Many classic studies of virus structure have been based on the gradual, stepwise disruption of particles by slow alteration of pH or the gradual addition of protein-denaturing agents such as urea, phenol, or detergents. Under these conditions, valuable information can sometimes be obtained from relatively simple experiments. For example, as urea is gradually added to preparations of purified adenovirus particles, they break down in an ordered, stepwise fashion which releases subvirus protein assemblies, revealing the composition of the particles. In the case of TMV, similar studies of capsid organization have been performed by renaturation of the capsid protein under various conditions (Figure 1.5). In simple terms, the reagents used to denature virus capsids can indicate the basis of the stable interactions between its components. Proteins bound together by electrostatic interactions can be eluted by addition of ionic salts or alteration of pH; those bound by nonionic, hydrophobic interactions can be eluted by reagents such as urea; and proteins that interact with lipid components can be eluted by nonionic detergents or organic solvents.
In addition to revealing fundamental structure, progressive denaturation can also be used to observe alteration or loss of antigenic sites on the surface of particles, and in this way a picture of the physical state of the particle can be developed. Proteins exposed on the surface of viruses can be labelled with various compounds (e.g., iodine) to indicate which parts of the protein are exposed and which are protected inside the particle or by lipid membranes. Cross-linking reagents such as psoralens or newer synthetic reagents with side-arms of specific lengths are used to determine the spatial relationship of proteins and nucleic acids in intact viruses.
Figure 1.5 The structure and stability of virus particles can be examined by progressive denaturation or renaturation studies. At any particular ionic strength, the purified capsid protein of tobacco mosaic virus (TMV) spontaneously assembles into different structures, dependent on the pH of the solution. At a pH of around 6.0, the particles formed have a helical structure very similar to infectious virus particles. As the pH is increased to about 7.0, disk-like structures are formed. At even higher pH values, individual capsid monomers fail to assemble into more complex structures.
Since the 1930s, electron microscopes have overcome the fundamental limitation of light microscopes: the inability to resolve individual virus particles owing to physical constraints caused by the wavelength of visible light illumination and the optics of the instruments. The first electron micrograph of a virus (TMV) was published in 1939. Over subsequent years, techniques were developed that allowed the direct examination of viruses at magnifications of over 100,000 times.The two fundamental types of electron microscope are the transmission electron microscope (TEM) and the scanning electron microscope (SEM) (Figure 1.6). Although beautiful images with the appearance of three dimensions are produced by the SEM, for practical investigations of virus structure the higher magnifications achievable with the TEM have proved to be of most value. Two fundamental types of information can be obtained by electron microscopy of viruses: the absolute number of virus particles present in any preparation (total count) and the appearance and structure of the virions (see below). Electron microscopy can provide a rapid method of virus detection and diagnosis but in itself may give misleading information. Many cellular components (for example, ribosomes) can resemble ‘virus-like particles,’ particularly in crude preparations. This difficulty can be overcome by using antisera specific for particular virus antigens conjugated to electron-dense markers such as the iron-containing protein ferritin or colloidal gold suspensions. This highly specific technique, known as immunoelectron microscopy, is gaining ground as a rapid method for diagnosis.
Figure 1.6 Working principles of transmission and scanning electron microscopes.
Developments in electron microscopy have allowed investigation of the structure of fragile viruses that cannot be determined by x-ray crystallography. These include cryo-electron microscopy, in which the virus particles are maintained at very low temperatures on cooled specimen stages; examination of particles embedded in vitreous ice, which does not disrupt the particles by the formation of ice crystals; low-irradiation electron microscopy, which reduces the destructive bombardment of the specimen with electrons; and sophisticated image-analysis and image-reconstruction techniques that permit accurate, three-dimensional images to be formed from multiple images that individually would appear as very poor quality. Conventional electron microscopy can resolve structures down to 50 to 70 Åin size (a typical atomic diameter is 2–3 Å; a protein a-helix, 10 Å; a DNA double helix, 20 Å). Using these newer techniques it is possible to resolve structures of 25 to 30 Å.
In the late 1950s, Sydney Brenner and Robert Horne (among others) developed sophisticated techniques that enabled them to use electron microscopy to reveal many of the fine details of the structure of virus particles. One of the most valuable techniques proved to be the use of electron-dense dyes such as phosphotungstic acid or uranyl acetate to examine virus particles by negative staining. The small metal ions in such dyes are able to penetrate the minute crevices between the protein subunits in a virus capsid to reveal the fine structure of the particle. Using such data, Francis Crick and James Watson (1956) were the first to suggest that virus capsids are composed of numerous identical protein subunits arranged either in helical or cubic (icosahedral) symmetry. In 1962, Donald Caspar and Aaron Klug extended these observations and elucidated the fundamental principles of symmetry, which allow repeated protomers to form virus capsids, based on the principle of quasi-equivalence (see Chapter 2). This combined theoretical and practical approach has resulted in our current understanding of the structure of virus particles.