APPLIED MICROBIOLOGY, May, 1965 Copyright ©) 1965 American Society for Microbiology
Vol. 13, No. 3 Printed in U.S.A.
Factors in the Membrane Filtration of Enteroviruses D. 0. CLIVER Food Research Institute and Department of Microbiology, University of Chicago, Chicago, Illinois
Received for publication 28 December 1964
ABSTRACT CLIVER, D. 0. (University of Chicago, Chicago, Ill.). Factors in the membrane filtration of enteroviruses. Appl. Microbiol. 13:417-425. 1965.-The filtration of two species of enteroviruses through membranes of porosity ranging from 50 to 220 my
studied. It was shown that extensive or total losses of virus may attend filtration at these porosities, apparently owing to adsorption of the virus to the membrane matrix.
This could be minimized by the incorporation of serum into the virus suspension at the time of filtration, or by pretreating the membrane with serum or with a gelatin solution. It was also found that the first few drops of filtrate, even under optimal conditions, were likely to be virus-free, so that the filtration of too small a volume of virus suspension would result in a relatively great loss of titer. The degree to which these factors were critical was found to decrease with increasing pore diameter.
In the course of experiments relating to viruses in foods, it became necessary to devise means of extracting enteroviruses from foods into a fluid suspension and of concentrating this suspension to insure the detection of such viruses as might be present. Two applications of membrane filtration to this work were envisioned. The first related to the clarification of the food extract and the removal of coincident microbial contaminants while still retaining the virus in the filtrate. The second related to the development of methods for concentrating enteroviruses from food extracts. It had become evident in preliminary studies that the virus preparations in use must contain a certain proportion of aggregates. As these aggregates were broken up in the course of manipulations, it was not unusual for the post-treatment suspensions to contain a greater number of infectious units than had originally been present. For this reason, means were sought of filtering virus at limiting porosity immediately before use, thus eliminating the bulk of the aggregates while losing little of the input virus titer. In the present studies, the virus preparations employed were produced in tissue culture and were diluted in a variety of fluids, but were not extracted from foods. The findings reported here are general and will be applied subsequently to enteroviruses in food extracts. Black (1958) summarized reports relating the diameters of virus particles to the limiting pore diameters for gradocol filtration. He proposed a working figure of 0.64 for the ratio of particle diameter to limiting pore diameter. Atoynatan and Hsiung (1964) recently reported the use of Millipore and gradocol membranes in determining
the sizes of simian viruses of several groups, and suggested that their techniques might be used for preliminary typing of virus isolates. The viruses filtered were contained in undiluted tissue culture fluid and were found to lose 0.3 to 1.5 logio units of infectivity during clarifying filtrations with Seitz pads or Millipore GS (220-m, porosity) membranes. Their data indicated that recovery of virus in Millipore filtrates was more nearly quantitative than that in gradocol filtrates at comparable porosity. MATERIALS AND METHODS Viruses. Poliovirus type 1 (Po-1), strain CHAT, was obtained from the Viral and Rickettsial Registry of the American Type Culture Collection. It had been through 15 tissue culture passages, including four subcultures from isolated plaques, in HeLa, primary rhesus monkey kidney (PMK), WISH, KB, and WI-26 cells. The preparations were stored at -20 C as an undiluted harvest (cells and fluid) of infection PMK titering 108 3 plaque-forming units (PFU)/ml, or as a diluted suspension of the above in a maintenance medium titering 104-7 PFU/ml. Coxsackievirus B-2 (CB-2) was obtained from Dorothy Hamre (Department of Medicine, University of Chicago) as a field isolate which had been passed nine times in HEp-2 cells. It was passed once in WI-26 and eight times in PMK, including three subcultures from isolated plaques. The preparations employed were stored at -20 C as an undiluted harvest of infected PMK titering 107-6 PFU/ml, or as a diluted suspension of the above in a maintenance medium titering 105.4 PFU/ml. In early passages, this virus was inhibited by the overlay medium described below, and had to be plaqued under the medium described by Wallis, Melnick, and Bianchi (1962). The final 417
preparations were found to yield higher plaque titers under the overlay medium used with Po-1. Tissue cultures. PMK cultures were prepared from the kidneys of rhesus monkeys (Macaca mulatta) by a method described previously (Gibbs and Cliver, Health Lab. Sci., in press). They were maintained until use in a mixture of equal parts of medium 199 and Eagle's minimal essential medium at room temperature. The volume of inoculum was 0.5 ml per flask, and adsorption was allowed to proceed at room temperature for 2 hr with manual rocking at intervals. Agar overlays consisted of 5 ml per flask of Earle's balanced salt solution (EBSS) containing 0.5 g of lactalbumin hydrolysate, 1.25 g of Noble agar, 0.51 g of MgCl2, 1.5 mg of neutral red, 10,000 units of penicillin G, and 10 mg of dihydrostreptomycin sulfate, per 100 ml. Filtration. The membranes employed and their porosities were: Gelman GM-6 (450 m,), GM-8 (200 mis), GM-9 (100 m,u), and GM-10 (50 m,u); Millipore GS (220 m,), VC (100 m,u), VM (50 m,u), and VF (10 m,u); and Schleicher and Schuell B17 (20 to 35 m,u). All Gelman membranes and the Millipore GS membranes could be sterilized either by autoclaving at 121 C for 15 min or by ultraviolet light (UV). The remaining types of membranes were sterilized by UV, with the few exceptions indicated. The holders employed were the Millipore microsyringe filter holder (MH) and the Gelman centrifugal filter holder (CH). These holders accept membranes 25 mm in diameter, and have effective filtering areas of approximately 3.9 cm2. Because of the pressures generated in filtrations using the MH with Multifit (Becton, Dickinson and Co., Rutherford, N.J.) syringes and membranes of 100 m, or smaller porosity, there was back-flow of virus suspension along the syringe plunger. This was prevented by a very light application of high-vacuum silicone stopcock grease to the plunger. Filtration at these porosities was also facilitated by means of a needle and a 20-ml Vacutainer tube (16 X 200 mm). The CH were found to be of no value when used with 100-m,i or smaller porosity membranes in a centrifuge. Three modifications were performed. First, the long (8.7 cm) funnel barrel was fitted with a no. 3 rubber stopper which had been bored and threaded on a length of polypropylene tube with a glass-wool packed bulb blown in it. Second, two flats were filed on the rim of the lower section of the holder to permit application of an open-end wrench to this part while tightening the top collar with a pipe wrench. Finally, the apparatus was mounted on a 50-ml round-bottomed polypropylene centrifuge tube, and the stopper was clamped in place. These modifications permitted filtration with air pressure at 35 psi. The diluents employed in filtration and recovery of viruses included deionized water; EBSS; EBSS plus 0.5% lactalbumin hydrolysate (E-Lac); medium 199; 1 M MgCl2; 0.067 M phosphate buffers at pH 7.0, 7.5, 8.0, and 8.5, and 0.067 M phosphate buffers which, when mixed with equal volumes of those just listed, would give a pH of 7.2 (Humason,
1962); and a solution called saline Y, containing 7.14 g of NaCl, 0.204 g of KCI, 0.136 g of KH2PO4, 0.174 g of K2HPO4, 0.038 g of NaHCO3, and 0.020 g of phenol red, per 100 ml of final solution in deionized water. To these were added, where indicated, newborn bovine serum (Bo), chicken serum (Ch), or a solution in 100 ml of Hanks' balanced salt solution of 2 g of gelatin (G). The additives and their concentrations are indicated by the above abbreviations and a subscript denoting the percentage of additive (v/v) employed. Thus EBSS Bo2 refers to Earle's balanced salt solution with an added 2% of calf serum. Where possible, the virus titers (in PFU) of the filtrates have been compared directly with those of the same virus suspensions before filtration. The procedural complexities of certain experiments have made it necessary to compare the observations with a "null count." The null count is an estimate, projected from the controls in the same experiment, of the number of plaques which would have been observed if membrane filtration had been totally without effect. RESULTS
Preliminary experiments. Po-1 (2 ml) was diluted in 8 ml of medium 199, and was filtered serially through Gelman membranes of 200, 100, and 50 m, in CH on the centrifuge (1,500 X g). The 50-m,A membrane was inverted and backwashed with 9 ml of medium 199 in the same manner. Each of these four filtrates was sampled and titrated by the plaque method. The titers, in PFU/ml, were found to be 2.3 X 107, 8 X 105,