The title of this paper merits some examination. It was suggested to me, and it sounded so imposing that I could not bring myself to change it. Dr. Ellem and I have had ample time in which to regret that hasty decision. It seemed to us that the title suggested a discussion of possible mechanisms of viral tumorigenesis. We concluded that we were faced with two alternatives. We could have attempted to distill into a brief discussion the many excellent, but largely speculative, papers on this subject. We might then have given our paper the subtitle, "The Best of Dulbecco, Luria, Stanley, Latarjet, and Lwoff et ai."

The second alternative-and the one we have chosen-is something of a compromise. We shall permit ourselves the luxury of some speculation, but, in addition, present some of our observations concerning the interaction of a mammalian cell with a viral ribonucleic acid (RNA), and discuss some of the factors that influence the response of the cell to the nucleic acid. The ribonucleic acid in question is not of tumor-virus origin, a fact which may seem to reflect a departure from the spirit of this symposium. However, we do not feel that this is true. Oncogenic viruses may in some cell types and under certain conditions act as cytocidal agents. Continuous replication of a virus without destruction of the host cell is a feature apparently not confined to tumor viruses. For the moment we would suggest that the basic mechanisms in the response of a cell to a tumor virus may be substantially the same as those involved in the response to a non-tumor virus. Thus a non-tumor virus system might be considered as a model that could provide a baseline for investigations of the special characteristics of tumor-virus systems.

The now classical experiments of Hershey and Chase (1) provided the first demonstration of the unique importance of bacteriophage deoxyribonucleic acid (DNA) in the reproduction and genetic continuity of these viruses. The corresponding central role played by the RNA of the tobacco mosaic virus (TMV) was conclusively and directly demonstrated by Gierer and Schramm (2). That the genetic information contained in the TMV particle for the deteiraination of the progeny virus lay in the nucleic acid core, and not in the protein coat, was also deduced by Fraenkel-Conrat and coworkers (S, 4). By mixing the nucleic acid of one strain of TMV with a preparation of protein from another strain, they effected reconstitution of viral activity. Progeny arising from an active infection initiated by this artificial virus always exhibited the characteristics of the virus from which the nucleic acid had been derived.

The results of Gierer and Schramm were quickly extended to a number of animal viruses. Colter et al. (5) demonstrated the infectivity of RNA solutions isolated from tissues infected with poliomyelitis, West Nile, and Mengo encephalomyelitis viruses. Alexander et al. (6) reported the successful isolation of infectious RNA from type I poliomyelitis virus. Since then, the fist of viruses from which biologically active RNA has been isolated has grown rapidly. I shall not undertake to reproduce it here, nor will I recount the various criteria by which RNA has been distinguished from the corresponding virus particles. It may be worth noting, however, that the viruses which make up the list are all small and are viruses with either neurotropic or myotropic potentialities.

A number of laboratories are involved in studies designed to demonstrate the biological activity of nucleic acids from tumor viruses. In addition, there is renewed interest in attempts to isolate viral agents from human tumors. As a part of this latter program, considerable attention is being given to nucleic acids isolated from tumor specimens-an attention which has no doubt received considerable impetus from the observations of Holland et al. (7) that viral RNA has in some cases a wider spectrum of activity than has the native virus. It would seem then to be of some value to consider certain factors which influence the interaction of RNA with the host cell.

In our laboratory, an examination of the optimal conditions for the interaction of Mengo RNA with strain L mouse fibroblast cells has been made (8). In brief, we have used the following system: Strain L cells are incubated, as a suspension of single cells, with Mengo RNA. After incubation, the cell-RNA mixture is diluted tenfold with normal growth medium and applied to preformed monolayers of strain L cells. After 1 hour the medium is removed and the monolayer, with the infected cells affixed thereto, is overlaid with agar. Plaques are read after 2 and 3 days.

It was found that the number of infectious centers formed in mixtures of RNA and strain L cells is a function of the osmotic pressure of the medium in which cells and nucleic acid interact. Maximum efficiency of infection was obtained in solutions of sodium chloride or in solutions of sucrose in physiological saline, when the osmotic pressure of the diluent was 4 times that of physiological saline. The number of infectious centers formed in the optimal sodium chloride or sucrose solutions was 20 to 30 times the number formed in physiological saline.

The infectivity of intact Mengo virus in a similar suspended cell system was adversely affected by increasing the osmotic pressure of the extracellular medium. The number of infectious centers formed from virus was reduced to less than 0.2 percent by an increase in the concentration of sodium chloride to 0.4 m and to 3 percent by the addition of sucrose to 0.7 m.

Studies are in progress which we hope will elucidate the mechanism whereby solutions of high osmolarity stimulate RNA infectivity. The data suggest that such solutions act not by affecting the RNA itself, or by stimulating the uptake of the nucleic acid, but rather by creating a hypertonic intracellular environment more suitable to the stabilization of the absorbed infectious RNA unit. It has been suggested that the increased RNA infectivity in solutions of high ionic strength may be due to the inhibition of intracellular ribonucleases (9); ribonuclease is known to be inhibited by solutions of high ionic strength (10). Another possible mechanism is suggested by our preliminary observations that the infectivity of Mengo RNA is reduced by ribonuclease-free proteins when nucleic acid and protein are mixed in solutions of physiological ionic strength. At higher ionic strengths inhibition does not occur. A hypertonic intracellular environment could then act by preventing nonspecific immobilization of the viral RNA by cytoplasmic proteins.

The period for which RNA and cells are incubated is quite critical. In the Mengo RNA-L cell system, at 37° C, and in the optimal sodium chloride or sucrose media, maximum production of infectious centers is obtained with a 15-minute incubation period. The number of infectious centers formed drops precipitously when incubation is continued for longer than 15 minutes. This probably reflects a loss of viability of the strain L cells in solutions of high osmotic pressure, and is very likely a function of the particular cell being used.