The apparently uncontrolled production of viral components may be viewed in terms of the genetic control of the host cell. The replication and expression of a normally occurring nucleic acid is undoubtedly the result of a fine balance betweeen activating and suppressor controls. The nucleic acid of a virus carries new information which is sufficiently intelligible to the cell to misinform it, but which is sufficiently foreign so that the cell has no control or suppressor systems for it. Consequently the viral nucleic acid and other viral constituents are synthesized at the cost of disturbing the homeostasis of the cell. Integration of a virus may mean greater recognition of the viral information, and this may be reflected in partial control and a degree of compatibility between the viral and host genomes. Nonintegrative relationships perhaps imply little or no recognition and consequently uncontrolled and deleterious viral multiplication.

The genetic information of the virus may be stored in the converted cell in at least two ways: (1) The viral nucleic acid may persist as such, replicating at least as fast as the cell, and exerting an influence on the cell phenotype more or less directly via the abnormal proteins it determines.

For the RNA viruses, models exist in some of the cytoplasmic factors in sexual and cellular heredity (32) and perhaps in the microsomes of normal cells, while, for DNA viruses, a parallel is seen in the Kappa particles of the Paramecium. However, the limited amount of genetic information which, on current hypotheses, it seems reasonable to expect in the small and medium-sized viruses (33) may incline us to view a detached seat of power, relatively independent of the host genetic apparatus, with some intuitive feelings of misgiving.

(2) A fashionable alternative is to consider the possibility of integration of the viral nucleic acid with the host genome. In this regard it is of interest to recall the radiological data of Rubin and Temin (34). They found that the ability of chick fibroblasts to produce Rous sarcoma virus and their ability to divide were equally sensitive to X-ray inactivation. The capacity for another virus of a cytocidal nature in this cell was many times more resistant to irradiation. They felt that this evidence suggested that a necessary early step in Rous sarcoma virus replication involves integration with the host genome, and that a common radiosensitive target controls the ability of the cells to divide and their ability to produce Rous sarcoma virus.

As a means of incorporating information more or less permanently into the host cell from viruses containing DNA, the model of lysogeny exists.

In the RNA viruses it is conceivable that transfer of genetic information could take place from viral RNA to host DNA by a mechanism akin to that postulated by Stent (35) for the replication of phage DNA. On the basis of P32 suicide and utraviolet inactivation data, and of a copy choice mechanism for DNA synthesis, he postulated that a ribonucleo-protein (RNP) strand was synthesized in the deep groove of the parental DNA helix, a ribonucleoprotein which carried the imprint of the DNA duplex, since the base pairs of the latter would select its constituent ribotides. The newly synthesized RNP molecule would then unwind itself from the parent DNA duplex and remain independent until meeting and "mating" with an identical macromolecule with which it would entwine to form an identical duplex. This structure could then serve as a template for the synthesis of DNA, a complementary pair of deoxyribotides being attracted by the single RNA base.

The studies of Doudney and Haas (86) on the production of stable bacterial mutants by ultraviolet irradiation have provided data which, by suggesting an important role for RNA in the replication of DNA, are at least compatible with Stent's hypothesis. These investigators demonstrated a close correlation between postirradiation RNA synthesis and mutation fixation. The latter was interrupted if RNA synthesis was disturbed, and was complete before any DNA synthesis occurred. They suggested that abnormal pyrimidines were formed in the metabolic pools by the ultraviolet irradiation, and that these were incorporated into ribonucleoprotein, which in turn imprinted faulty information in the base sequence of DNA synthesized on it.

Stent's postulates could explain how genetic information might be transmitted from DNA to RNA and thence back to DNA. For a similar mechanism to operate for an RNA virus, at least three conditions would have to be met: (1) The active viral nucleoprotein or RNA would have to be in proximity to the nuclear apparatus. All available evidence suggests that the nucleoprotein of the small and medium-sized RNA viruses is synthesized in the nucleus. (2) At least two viral genomes would have to be present in the cell to form an identical RNP duplex. Since one virus particle is sufficient to initiate an infection (80), some viral replication would have to proceed before this RNP duplex could be formed. It follows from this that DNA synthesis cannot be a prerequisite to viral replication, but merely an incidental phenomenon. (3) There would have to be considerable homology between the base ratios of the hypothetical DNA and the host DNA which the viral nucleic acid would determine. It may be pertinent that, in bacterial systems, successful genetic recombination has been observed only between organisms containing DNA's of the same base ratio group, and the same sort of "code similarity" is probably required for the successful integration of a phage or prophage with a bacterial chromosome (88).

A corollary arising from this hypothesis is that it might be possible to isolate from some virus-induced tumors both DNA and RNA capable of transmitting the biological properties of the intact virus.

The new nucleic acid added to the host genome might influence other genes in a fashion similar to that seen in the alteration in the activity of loci near the site of attachment of certain genetic elements to the chromosome, such as the fertility factor in E. coli (87) or the controlling elements in maize (88). It may also perhaps induce host modifications as an intrinsic property of its own genetic constitution. As examples, one may cite the toxin production in Corynebacterium diphtheriae, which is inextricably related to a phage infection (89, 40) or in the suppression by X prophage of the multiplication of certain mutants of the T-even phages (41). This would in effect be somatic mutation, as has been pointed out by Luria. Successive somatic mutations have been suggested as the basis for the evolution of the cancer cell (42, 48) and may be the final common pathway for the activity of many carcinogens.

The foregoing considerations suggest a mechanism by which a hereditable defect might be built into a mammalian cell-the introduction, by a viral agent, of an abnormal ribonucleoprotein, or the intracellular synthesis of such an abnormal molecule after irradiation. The consequent alteration in DNA could lead to deletions, resulting, for example, in a loss of systems responsible for negative feedback on the specific enzyme-forming systems required for cell division (44), or the gain of abnormal potentialities in differentiation (45) and other phenotypic expressions of malignancy.