The sizes of viruses and the methods employed in their estimation

Abstract
In this review we have given an account of the various methods which are available to determine the size of virus particles. In § IV we have endeavoured to bring the ultrafiltration method into agreement with other methods by suggesting a different factor for converting pore size to virus size from the factors commonly used. Throughout we have recognized the probability that most viruses are hydrated in solution and have distinguished between the size and molecular weight in solution and the size and molecular weight when dried.In § VII we have given formulae suitable for interpreting centrifugation and diffusion data when the possibility of hydration is contemplated.It is evident that this complication, added to that of shape, makes it necessary for several measurements by different methods to be made before one can claim to know the size of a virus. For this reason, only in the cases of three viruses have we thought the data sufficiently adequate to enable the size and shape and molecular weight of the Virus, both dry and hydrated, to be stated. These three viruses, tobacco mosaic, tomato bushy stunt and vaccinia respectively, are separately discussed in § X.It will be clear from the preceding sections that, while the position regarding our knowledge of the absolute sizes of viruses is far from satisfactory, there has been amassed a large amount of data bearing on this subject. We should, however, point out that we have found it necessary to select what we consider to be the best experimental data in some cases and that there may be conflicting ideas expressed by various authorities. Frampton (1942) has studied the electron microscope photographs published by Stanley & Anderson (1941) and Anderson & Stanley (1941) and arrives at an entirely different estimate of the length of tobacco mosaic virus. Kausche, Pfankuch & Ruska (1939) reported one value for the length of this virus which is approximately half that given by Stanley & Anderson. Electron photomicrographs published by von Ardenne (1940) and Holmes (1941) for what are probably strains of the same virus, also suggest that the values given should not be taken as absolute. Frampton (1939a,b), on the basis of diffusion and viscosity experiments and the stream birefringence of this virus, has suggested previously that it forms a gel at any concentration and therefore cannot be said to have a size. Lauffer (1940) has given reasons for supposing this argument to be incorrect. Bernal & Fankuchen (1941a) have discussed the possibility of tobacco mosaic virus particles being shorter than the value taken from Kauscheet al.(1939) and conclude that in the plant itself the particle may be as short as 100 mμ.In obtaining values of size arid shape from electron microscope data we have made the assumptions, which may not be correct, that long, thin viruses shrink in width rather than in length on drying and that almost spherical viruses contract approximately evenly in all directions. At the moment there would seem to be no method of proving or disproving the truth of these assumptions, but we believe it unlikely that drying will result in such a gross change in shape that it would invalidate our calculations. For instance, in the case of haemocyanin fromHelix pomatia, it seems improbable that, on drying, an already anhydrous ellipsoidal molecule of 66 × 15·32 mμ would contract in length and expand in-width to form a sphere of some 24 mμ diameter.In our treatment of hydration we have found it necessary to regard the density and volume of ‘bound’ water as being the same as that of water in bulk, which may not be entirely true. However, we regard the total volume occupied by water in cases of great hydration, as shown by tomato bushy stunt virus, as being not markedly smaller than that of the same mass of free water. It is, nevertheless, a well-established fact that in certain cases, gelatin for example (Svedberg, 1924), a small contraction in volume does take place when dry protein is added to water. This phenomenon does not, however, necessitate the assumption that the water of hydration, is denser than ordinary water, and can be explained in other ways.The viscosity of solutions of viruses, especially the rod-shaped plant viruses, has attracted much attention as a method of finding frictional and axial ratios of viruses (Frampton, 1939a,b; Lauffer, 1938; Loring, 1938; Neurath, Cooper, Sharp, Taylor, Beard & Beard, 1941; Kobinson, 1939a,b; Stanley, 1939), but, in addition to the lack of experimental verification of the formulae used, in many cases (Robinson, 1939a,b; Frampton, 1939a,b) the formulae have been applied to experimental results obtained in circumstances which exclude the fundamental postulates on which the formulae are based. For this reason we have omitted a detailed discussion of such methods.It would appear that in order to obtain evidence as to the size of a virus it is desirable to study the virus in as purified a form as possible and also to show that when ‘homo-geneous’ preparations are obtained, they do not consist merely of macromolecular substances contaminated with a small quantity of virus. Furthermore it is desirable to obtain at least sufficient data to enable one to assess both size and shape of the particles rather than to assume some shape or some density value which may be incorrect.

This publication has 103 references indexed in Scilit: