TEXTE COMPLET :
Current attempts to fit the arboviruses within a universal system of classification have revealed severe deficIencies In the knowledge of the basic physical structural and biochemical properties of these agents, on which proposed systems are chiefly based.
Although virus size is not among the properties used for classification in one of the proposed schemes, it nevertheless seems logically connected with at least one element used in that scheme; for markedly dissimilar sizes within a collection of viruses may reflect differences in number of capsomeres as well as in the triangulation numbers of agents with cubic or icosahedral symmetry. In addition, heterogeneity in size may correspond with heterogeneity in still other properties.
The present observations regarding the behaviour of arboviruses on filtration through "Millipore" membranes were made as part of a continuing investigation in this laboratory. Properties other than size. may affect the filterability of viruses through such memoranes, and the commercially available membranes represent only a limited range of average pore diameter (APD), and so these observations are at best estimates of virus size within rather wide limits rather than accurate size determinations.
The technique usually employed was similar in some respects to that used with simian viruses. Except for group C agents, each virus preparation was made up as a 10-2 suspension of infected newborn mouse brain tissue usIng an electrical blender; with group C viruses (Caraparu, Murutucu. Oriboca in Table ), a similar dilution of infected serum or. whole blood was employed. The diluent was 0.02 M phosphate buffer in physiological saline, pH 7.2., to which was added 0.75 per cent bovine plasma albumin (Armour fraction V). Between 50 and 60 ml of the freshly prepared virus suspension was spun in a Spinco preparatory centrifuge (Model L, Type 40 rotor) at 15,000 r.p.m. for 30 min. or, In the case of some of the larger viruses (Nyamanini , herpes, Newcastle disease), at 5,000 r.p.m. for 30 min. The supernatant fluid was then of APD 450, 220, 100 and 50 mµ ; the diameter of each filter was 47 mm and a micro glass prefilter was used in all cases. At each stage about 5 ml. was reserved for infectivity titration. The least volume emerging through the last filter was 10 to 15 ml. The original crude suspension, the supernatant fluid and each successive filtrate were titrated by tbe intracerebral route of inoculation in 2-4 day old mice, eight mice being used for each dilution.
With a few exceptions, the size of the viruses tested was such that they were neither sedimented by the centrifugation employed nor excluded by the filter, of APD 450 mµ to a significant degree. In nearly all instances, the infective titre of material passing the 450 APD membrane was not more than 0.7-1.0 dex 5 lower than the titre of the original suspension, and this loss was caused as much by centrifugation as by filtration. For our purposes, therefore, the 450 APD filtrate has been adopted as the workIng suspensIon, and reduction of infective titres resulting from subsequent filtrations has been calculated with reference to it rather than to crude original 10-2 brain suspension.
Several. viruses not arboviruses (Reovirus 3, Herpes, N D V) were included because their sizes have been reported, usually as determined by electron microscopy; similarly, for some of the arboviruses included there are similarly, for some of the arboviruses included there are reports of measurements, by the same and other techniques. It was presupposed that the observed degree of filtrability of viruses of known size in the present system would permit certain inferences, by analogy, the relative size of the other arboviruses.
The results presented in Table show, not unexpectedly,that the viruses tested differed considerably in degree of filterability, and presumably therefore in relative size, in our experimental conditions. Considering a loss of 1.5 dex or more in infective titre as significant, viruses (Dengue 2, Ntaya, MVE, SLE, Chikungunya, Nos. 1-5,) represented by group B agents and one from group A (chikungunya), were not affected by final filtration through the smallest APD used, 50 mµ. Viruses (Sindbis, EEE, WEE, Bunyamvera, Nos. 6-9), however, suffered a loss of titre on filtration through the 50 mµ APD membrane, particularly pronounced in the case of WEE and Bunyamwera and less so for Sindbis and EEE. Kemerovo virus (No. 11) has been reported to have a virion size, including its envelope, of between 6O and 100 mµ, and its behaviour on filtration corresponds well with this size. The behaviour of reovirus type 3 with: {No. 27) on filtration similarly corresponded with its reported size of 70-77 mµ.
The greatest number of viruses (Nos. 12-25) experienced a significant loss of infectivity on filtration through the 100 mµ APD membrane but either no loss or only a at moderate loss on filtration through a 220 mµ membrane. These viruses comprised members of group C (Nos. 13, 14 and 15), Bwamba (No. 12) and tick-borne viruses not of group B (Nos. 16-25).
Nyamanini virus (No. 26) has repeatedly shown considerable loss of titre on filtration through a 220 mµ APD membrane; if particle aggregation can eventually be ruled out, this loss in titre may indicate an unusually large size of arbovirus.
These observations seem to provide further evidence the that the group of the animal viruses carried by arthropods is heterogeneous in physical properties; inclusion of the arboviruses in a universal scheme will have to acknowledge such heterogeneity.