Vol. 41, No. 3 Printed in U.S.A.

BAcTzmuoLoGIcAL Ruvizws, Sept. 1977, p. 543-567 Copyright 0 1977 American Society for Microbiology

Antiviral Activity of Interferons ROBERT M, FRIEDMAN

Laboratory ofExperimental Pathology, National Institute ofArthritis, Metabolism, and Digestive Diseases, Bethesda, Maryland 20014

INTRODUCTION.............................................................. 543 .................... 544 ESTABLISHMENT OF THE ANTIVIRAL STATE ........... ............................. 544 Interferon Binding ................... 546 Development of Antiviral Activity LOCUS OF THE INTERFERON-INDUCED INHIBITION OF VIRUS GROWTH. . 547 ........... 547 Evidence that Interferon Treatment Inhibits Virus Uncoating ....... Evidence that Interferon Treatment Inhibits Transcription of the Viral Genome. 547 ...... 550 Evidence that Interferon Treatment Inhibits Viral Protein Synthesis ..... Observations in virus-infected cells ........................... 551 Observations in cell-free systems ....... .................... 552 Evidence that Interferon Treatment Inhibits Terminal Events in the Replication ................. 557 Cycle of Murine Leukemia Viruses ........ INTERFERON TREATMENT IS INEFFECTIVE IN SYSTEMS IN WHICH THE SV40 GENOME IS INTEGRATED INTO AN INTERFERON-RESISTANT VIRUS OR A HOST GENOME ........................... 559 DISCUSSION ............................. 560 562 LITERATURE CITED ........................... .................................

Principles should not be unnecessarily multiplied. - William of Ockham, about 1320 I beseech you, in the bowels of Christ, think it possible you may be mistaken. -Oliver Cromwell, 1650

INTRODUCTION Interferons are proteins, the production of which can be induced in animal cells by a variety of stimulating substances; interferons inhibit a wide range of viruses by inducing an intracellular antiviral state, yet most interferons are animal species specific in their range of antiviral activity. Interferon was discovered in 1957 (85), but to the present time there has been no entirely statisfactory explanation of its potent antiviral effect. Part of the reason for this lies in the impressive biological activity of interferons. A recent estimate of the specific antiviral activity ofhuman interferon was at least 2 x 108 reference units per mg of protein, and it may be 10 to 100 times more potent (71). With a molecular weight of about 25,000, this would mean that 1 active unit per ml is present in (at most) a 10-12 M interferon solution. Cholera and diphtheria toxins are marginally active at 10-9 M; therefore, on a molar basis interferon exceeds the specific activity of some of the most potent biologically active substances. As a consequence of this, interferon preparations containing very large amounts of antiviral activity turn out to have extremely small amounts of interferon. It is, therefore, very difficult to purify interferons completely, and only recently

has there been any great optimism that this can be accomplished (71). Interferon assays have added to this difficulty because they have been exclusively biological and are based on the ability of a preparation to inhibit the production of a virus or of a viral product in infected cells. They are timeconsuming and relatively inaccurate. Cells must be treated with interferon for several hours before virus infection until antiviral activity develops. Because of the inherent inaccuracies of the biological assays, a two- to threefold inhibition is considered just significant; therefore, a level of uncertainty is present that is usually considered intolerable in a biochemical or biophysical system. One other problem has been the multiple substances that bear the name interferon in any one species. There are several proteins of differing molecular weights that have an antiviral effect and are produced in differing cell systems. Human buffy coat interferon contains three antiviral proteins, two with molecular weights of about 20,000 and one of molecular weight 15,000. Interferon produced in human diploid fibroblasts antigenically resembles one of the proteins with a molecular weight of 20,000 but is distinct from the other buffy coat interferons. Finally, human T lymphocytes produce an "immune" interferon in response to antigens and mitogens; this is a distinct interferon species. All of these act to produce an intracellular antiviral state, but it is unclear whether they have the same mechanism of action (107).

543

544

FRIEDMAN

Finally, results of studies on the antiviral activity of interferons have often been confusing and contradictory. In some cases it is uncertain whether the reported effect was due to an interferon, because the concentration of impurities in interferon preparations far exceeded the concentration of interferon. Also, with distressing frequency, studies on the mechanism of action of interferon which appeared to herald a real breakthrough in our understanding have seemed to lack that most fundamental requirement of science, reproducibility. This review will concern itself with the antiviral activity of interferons. There is very convincing evidence that interferon also has other effects on cells, such as inhibition of cell and tumor growth (72, 77, 106, 132), alteration in levels of cellular enzymes (100), and regulation of the immune system (63). Although publications on some of these observations are appearing with increasing frequency, there are still too few reports to warrant a detailed review of these at this time.

ESTABLISHMENT OF THE ANTIVIRAL STATE When cells are treated with an appropriate interferon, they develop an antiviral state, and a wide variety of viruses grow poorly in such cells. In the absence of virus infection, however, it is usually difficult to show whether or not a cell has been treated with interferon. One can arbitrarily look at two aspects of the antiviral effect of interferon cells: first, how interferon treatment alters the uninfected cell to produce an antiviral state; and second, the fate of viruses in an interferon-treated cell. In this section I will discuss what is known about the induction of the antiviral state. Until recently, most studies on this phenomenon have paradoxically had to involve virus infection at one point or another in order to test the degree of antiviral activity developed. Interferon Binding In spite of the very low concentrations of interferon that are required to induce an antiviral state, there is strong evidence that only a small portion of the interferon in a solution interacts directly with cells to induce antiviral activity. No detectable interferon was removed from a preparation that was repeatedly used to induce antiviral activity (21). There is, however, some evidence that interferon is bound to the surface of cells in which an antiviral state is later induced. When cells were treated with interferon at 40C, no antiviral activity was present if the cells were washed and then im-

BACTERIOL. REV.

mediately infected at 370C with a rapidly replicating virus. If, however, the cells were allowed to incubate at 370C for even 1 or 2 h, antiviral activity was evident. There is, therefore, an alteration in the cells after incubation at 40C with interferon, but a period of active metabolism at 370C was necessary to develop an antiviral state (38). The nature of the interaction at 40C between interferon and cells is partially understood. After incubation with interferon at 40C, treatment of cells with trypsin in the cold inhibited the development of antiviral activity. It appeared that interferon was bound at 40C to cells in superficial sites that were accessible to trypsin (8, 38). After a few minutes at 370C, however, a complex series of changes took place. Interferon that had been bound to the cell became trypsin resistant, yet could still be recovered in an active form by extraction of the cells. With further incubation, cell-associated interferon tended to elute into the culture fluid (8, 126). The role of the portion of the interferon that did become cell associated is not entirely clear at present (76). There is no direct evidence that this interferon was bound to a specific cellular interferon receptor; however, interferon treatment has been shown to alter the binding of cholera toxin or thyrotropin (TSH) to plasma membranes (75) and to decrease the sensitivity of cells to diphtheria toxin (97). All of these biologically active substances bind to fairly well-characterized, specific receptors. The inhibition of their binding by interferon means either that interferon directly competes with these substances for their binding sites or that interferon treatment brings about a general alteration in membranes; one manifestation of the latter would be a decreased binding of other biologically active substances. If, however, the former is the case, it would appear that distinct binding sites must exist for interferon. The location, chemical nature, and specificity of such putative binding sites have been investigated. One study indicated that if such sites exist, they must be on the external surface of the plasma membrane. Interferon has one unusual biological property. The cell producing the humoral factor (interferon) may also be an effector cell. In nature this is not likely to be important, because the interferon-producing cell is already virus-infected and thus probably beyond salvation by interferon treatment. In one tissue culture study, however, human fibroblasts were stimulated to produce interferon by the double-stranded ribonucleic acid (RNA) inducer polyriboinosinic acid * polyribocytidylic acid [poly(I * C)]. Inclusion of anti-human inter-

VOL. 41, 1977

ANTIVIRAL ACTIVITY OF INTERFERONS

feron antibody in the medium prevented development of an antiviral state, even when the antibody was added after interferon synthesis had already been initiated (134). The results seemed to indicate that interferon must be externalized and interact with the outer surface of the plasma membrane in order to be effective. Similar results have been obtained in cells induced to make interferon and then treated with ouabain (79). The receptors for interferon would appear to be gangliosides or ganglioside-like structures with an oligosaccharide moiety at the functional binding site. Treatment of mouse cells with Phaseolus vulgaris phytohemagglutinin (PHA), a plant lectin that binds to carbohydrates, blocked the development of antiviral activity after subsequent interferon treatment (9). Also, Sepharose-bound interferon lost its antiviral activity after preincubation with gangliosides, especially GM2 and Gr, (10), and soluble interferon was bound to Sepharose-ganglioside beads (10). Sialyl-lactose reversed the ability of gangliosides to inhibit the antiviral action of interferon, and PHA inhibited interferon binding to gangliosides (12). Furthermore, ganglioside-deficient mouse cells were insensitive to induction of antiviral activity by interferon; treatment of these cells with gangliosides both increased cell membrane ganglioside content and, in two of three cell lines tested, significantly increased cell sensitivity to interferon treatment. Ifindeed the receptors for interferons are gangliosides, these receptors would resemble those of cholera toxin and several glycopeptide hormones (98, 99). It is difficult to comment as yet on the species specificity of the putative binding sites for interferon. Many substances that have no known intracellular function bind to cell surfaces. One cannot imagine that all such binding involves specific receptor sites. There is significant binding of some interferons to cells in which the interferons tested had no known biological activity (26, 75). The meaning of this is unclear; however, mouse interferon altered the binding of cholera toxin or TSH to plasma membranes of KB cells, a human cell line that is completely unresponsive to both mouse and human interferons (75). Similar conclusions have been reached from studies in rat embryo fibroblasts treated with rat and human interferons. Human amnion interferon has antiviral activity in these rat cells, but human leukocyte interferon does not; however, both human interferons block the antiviral activity of rat interferon. Thus, human leukocyte interferon must have interacted with the rat cells even though it did not induce antiviral activity (26). In the case of

545

human interferons, neutralization experiments appear to indicate that multiple reactive sites on the interferon molecule can interact with cells of different species to induce antiviral activity (107). These results taken together suggested that the species specificity of interferons does not reside solely at the point oftheir initial interaction with their binding sites and that binding of an interferon is a necessary, but not a sufficient, condition for biological activity; it is an overture that may not be followed by the opera. A number of substances that bind to and alter the cell membrane inhibit interferon action. Cytochalasin B (109, 142) and colchicine (26) both are active in this respect, although the effect of the former is probably due to extrusion of the cell nucleus. Mercaptopyridethylbenzimidazole, which inhibits membrane transport of nucleosides, blocks the development of antiviral activity in interferon-treated cells (46). As might be expected, chorionic gonadotropin, TSH, and cholera toxin, which bind to ganglioside receptors, also inhibit interferon action (11, 75). Antibody to a cell surface component coded for by human chromosome 21 inhibited the antiviral activity of interferon (112), although antibody directed nonspecifically against cell surface antigens of human fibroblasts had no such inhibitory effect (134). Since chromosome 21 in human cells appeared to be necessary for the development of antiviral activity after treatment with human interferon (32, 128, 129), these findings suggested that human chromosome 21 may determine an antigen important in the interferon receptor. Other studies, however, suggested that human chromosome 21 did not determine the interaction receptor but that some .other chromosome did (28, 32). Also, there was a suggestion that chromosome 16 carried a function that regulated the interferon-induced antiviral state (28). Interferon treatment appears to bring about a number of changes in plasma membranes of uninfected cells. Plasma membrane preparations of L-cells exposed to interferon developed a complex alteration in their ability to bind cholera toxin. At low interferon concentrations, binding of toxin was significantly increased; when the interferon concentration was increased, however, binding was inhibited (75). Similar complex effects on binding have been reported with other systems and are thought to be due to binding of a substance to both an active (at low concentrations) and inhibitory (at high concentrations) site (98). Interferon treatment also alters the surface charge on Lcells (73), and in AKR mouse cells treated with interferon there is a significant change in the

546

FRIEDMAN

density of purified plasma membrane preparations. In untreated cells, 80% of the plasma membrane banded in sucrose at a density of 1.20, and 20% banded at 1.23; in interferontreated cells, 25% banded at 1.20, and 75% banded at 1.23. This change may be related to alterations in plasma membrane ganglioside content after interferon treatment. Also, there is a marked increased in the number of intramembranous granular particles that can be observed by electron microscopy after freeze etching the membranes of interferon-treated AKR cells as compared with untreated controls (E. Grollman, F. T. Jay, and E. H. Chang, manuscript in preparation). Although the significance of these alterations in plasma membranes after interferon treatment is unclear, they are consistent with a profound change that appears to be induced by exposure to interferon and may well be related to the development of the antiviral state. Development of Antiviral Activity Very soon after Isaacs and Lindenmann discovered interferon they recognized the necessity for metabolic activity for interferon action (85). This finding has been repeatedly confirmed and is generally thought to mean that interferon itself is not antiviral but that the development of an antiviral state requires the intracellular production of an antiviral substance. There is one study which does suggest that interferon may itself be antiviral and that the necessary metabolic activity was simply to transport interferon into the cell (122). Although this notion is difficult to rule out, the previously discussed experiments, which indicated that interferon must be externalized to be active (134), make it rather unlikely, unless there is some portion of the interferon molecule that must be reintroduced. In addition, interferon bound to an insoluble, Sepharose matrix was active in inducing antiviral activity (4); this study did not, however, rule out the possibility that interferon that might have come off the Sepharose matrix was actually responsible for the induction. Interferon resembles, in some respects, glycopeptide hormones and, therefore, what has been learned about the mechanism of polypeptide hormone action may be useful in uncovering clues to the mechanism of interferon action. As an example, in transmitting its message to the cell, TSH is believed to require a sequence of events whose main contributors are in turn a specific receptor (probably a ganglioside), a significant alteration in the state of the membrane, an effector-responsive adenylate cyclase, and cell metabolism susceptible to reg-

BACTERIOL. REV.

ulation by increased production of cyclic 3',5'adenosine monophosphate (cAMP) (99). There is some evidence that interferon action involves the first two steps listed (75, 133). In addition, there is a report that cAMP analogs such as dibutyl-cAMP increased the level of antiviral activity induced by a preparation of interferon (45). I am not aware of any reported studies on the effect of interferon treatment on adenylate cyclase activity. I believe that Chany's model of an interferon-binding site and an associated but distinct activation or amplification site that is actually responsible for the development of antiviral activity is probably correct (20, 26, 27) and resembles to a great extent the glycopeptide hormone models. Several cell functions are known to be required for the induction of antiviral activity by interferon. Studies (27, 28, 32, 128, 129) have shown the relationship of specific chromosomes to the induction of antiviral activity in human and mouse cells. Human cells trisomic for chromosome 21 are more sensitive to human interferon than cells that are diploid. The latter in turn are significantly more sensitive than cells monosomic for chromosome 21 (28, 128). The human interferon activity gene would appear to lie on the distal portion of the long arm of chromosome 21, since the translocation of this locus to another chromosome transfers sensitivity to interferon (35). This site is closely linked to the locus of the enzyme indophenol oxidase (129). The nature of the gene function required for antiviral activity is to some extent known through studies with antimetabolites. In 1964, Taylor showed that interferon action was inhibited in cells treated with actinomycin D (130); however, established antiviral activity was not decreased by treatment with actinomycin D (130). Several other inhibitors of RNA synthesis and ribonucleoside analogs also inhibited interferon action and, taken together, these studies indicated that cellular RNA synthesis is probably necessary for interferon action (124). These findings have recently been confirmed by studies with enucleated cells. In cells in which nuclei were removed by exposure to cytochalasin B, interferon treatment failed to induce antiviral activity. Once established, however, antiviral activity was not reversed by removing the nucleus (109, 142). The results of several studies with inhibitors of protein synthesis or amino acid analogs also suggested that interferon action required protein synthesis (124). These experiments were not as convincing as those with actinomycin D, which suggested a need for RNA synthesis, probably because the activity of inhibitors of

VOL. 41, 1977

ANTIVIRAL ACTIVITY OF INTERFERONS

protein synthesis must be reversed in order to test (by virus growth) for the presence of antiviral activity (48). These results suggest a by now familiar model: interferon treatment induces the production of a cellular messenger RNA (mRNA), which is translated to an antiviral protein. It is the latter that actually inhibits the growth of viruses. This theory was first suggested to me by J. A. Sonnabend in 1963, and to date no report has refuted it. By the same token, there has not yet been a direct demonstration of an antiviral protein or its mRNA. There are two studies purporting to define a significant difference in a specific RNA constituent between interferon-treated and control cells. Levy et al. (82, 112a) have investigated a small increase in the size of mRNA's and transfer RNAs (tRNA's) from interferon-treated mouse cells by polyacrylamide gel electrophoresis. Although these differences were small, they were reported to be consistent. However, almost all of the RNA species examined from interferon-treated cells seemed to trail behind the analogous species from extracts of untreated cell by about the same distance. Since, under the conditions employed, the distance migrated by an RNA species is thought to vary directly with the logarithm of its molecular weight, the finding would imply that interferon induces very large differences in the molecular weight of large RNA species, but quite small differences in small species. There would seem to be no easy explanation for these findings at present, but such studies as these are important since they may uncover significant differences that may exist between the RNAs of interferon-treated and control cells.

LOCUS OF THE INTERFERON-INDUCED INHIBITION OF VIRUS GROWTH Interferon inhibits the replication of a surprisingly wide variety of viruses, and incidently of some nonviral infectious agents (62). Many of these viruses and infectious agents would appear to have very little in common. A number of virologists have, therefore, investigated this problem and employed several viruscell systems in attempts to find a specific virus function which is blocked in interferon-treated cells. At least four possible sites of action have been seriously suggested by a number of laboratories. It is unfortunate that most of these suggested sites of action involve phases of virus replication about which we known quite little. Perhaps this is more a comment on the general state of the science (art) of virology than on interferon research; on the other hand, it may also be symptomatic of the perversity of the

547

latter. At any rate, I shall discuss each of the theories about interferon action currently in vogue in the order that they generally appear in the virus replication cycle. Evidence that Interferon Treatment Inhibits Virus Uncoating Until recently, there seemed little reason for seriously considering this as a site of action for interferon. Several studies with infectious RNA from poliovirus or western equine encephalitis virus indicated that interferon inhibited the replication of these RNA forms (33, 55, 59). Barring hydrolysis of the RNAs by ribonuclease (RNase) left from the interferon preparations used to treat the cells, the data clearly indicated that interferon must act at a site beyond the uncoating step. Uncoating is also not inhibited in interferon-treated cells infected with reovirus (135). Studies with simian virus 40 (SV40) seemed, however, to contradict the findings with infectious RNA (140). In agreement with the findings of Oxman et al. (96, 102, 103), the production of SV40 early mRNA and T antigen was markedly inhibited in interferon-treated monkey cells, when these cells were infected with intact SV40. If, however, infectious SV40 deoxyribonucleic acid (DNA) was employed, the results were directly opposite those obtained with infectious RNA. The synthesis of early SV40 mRNA and T antigen were only slightly inhibited in interferon-treated cells infected with the SV40 DNA (140). This would appear to indicate that in interferon-treated cells, SV40 uncoating, or a step soon after it, was inhibited, and that this is the explanation for the apparent inhibition of SV40-directed transcription under these conditions. A similar block in uncoating or an event soon thereafter has been described at nonpermissive conditions for the SV40 mutant, tslOl (113). The results in interferon-treated cells infected with SV40 DNA are clear-cut, but have not as yet been confirmed. The most likely explanation for the finding was discussed above, but there is one other possible explanation. In the SV40 systems employed in this study, there is a multiplicity-dependent reversibility of the interferon-induced inhibition of virus replication (140). This has been reported in another system (60). To carry out the experiments with SV40 DNA, 0.075 or 0.75 jig of SV40 DNA was employed. This amount of infectious DNA could conceivably be equivalent to a high virus-cell multiplicity, if any great part of the DNA actually induced infection. The explanation for these extraordinary findings may lie, therefore, in the SV40 multiplicity reversal of

548

FRIEDMAN

interferon action also reported in the same publication (140). Still, the possibility that interferon action involves, at least under some conditions, an inhibition of viral uncoating is important, and until such time as this work is directly contradicted, any theory that attempts to explain interferon action will have to make provision for these findings. Evidence that Interferon Treatment Inhibits Transcription of the Viral Genome The work in this area of interferon research is quite controversial, and, like scripture, can be quoted to support any side of an argument. In this case the disagreement revolves around whether interferon treatment causes an inhibition in the primary transcription of the genetic information of viruses. All of the possible answers (yes, no, or sometimes) can be found in various publications. Part of the confusion undoubtedly is due to the close relationship between viral RNA synthesis and viral protein synthesis (124). If a virus requires a polymerase not present in cells and has no endogenous RNA polymerase as a structural element, inhibition of viral protein synthesis will result in inhibition of all virusdirected RNA synthesis, because the viral polymerase is among the proteins whose synthesis would be inhibited. Arboviruses and picornaviruses, among the RNA viruses commonly used in interferon research, are lacking structural polymerases. In the case of those agents containing structural polymerases, analysis of the site of inhibition can be quite complicated. Since a polymerase is already present, inhibition of protein synthesis does not cause inhibition of early RNA synthesis. Primary transcription takes place in these cells, but secondary transcription, which depends on the elaboration of new polymerases, will be inhibited. Results obtained employing such systems require careful analysis, since significant inhibition of RNA synthesis does not necessarily indicate that primary RNA synthesis is the site of action. Furthermore, the use of inhibitors of protein synthesis in these experiments introduces some additional problems. First, the endogenous viral polymerases may be subject to negative control by viral proteins produced early in the infection process. When an inhibitor of protein synthesis is employed, the level of primary RNA synthesis in controls may be elevated (13). This could lead to difficulties in interpretation of data suggesting inhibition of primary transcription by interferon. In addition, related to the general problem of using inhibitors of protein synthesis is a problem relating specifi-

BACTERIOL. REV.

cally to the use of cycloheximide, which has been the most commonly employed inhibitor of protein synthesis; under these conditions, cyheximide inhibits the elongation of peptides, it protects and stabilizes mRNA (3, 123). Interferon treatment might alter this stabilization, if it tended to restrict the initiation of viral protein syntehsis; under these conditions, cycloheximide might not have protective effects on mRNA in interferon-treated cells. This could well lead to large differences in the concentrations of intracellular mRNA in cells treated with cycloheximide alone as compared with cells treated with both cycloheximide and interferon. These differences would not of course be indicative of an interferon effect on primary transcription. Among RNA viruses, there have been detailed reports on the effect of interferon on the primary transcription of vesicular stomatitis virus (VSV), influenza virus, and reovirus. In the case of DNA viruses, vaccinia virus and SV40 have been well studied. Inhibition of primary transcripton of VSV has been reported in human (88), chick (89), and monkey cells (91). All three studies employed cycloheximide to block secondary transcription and actinomycin D to inhibit host RNA synthesis, and then measured viral RNA synthesis in the presence or absence of interferon. In the experiments performed in human cells, there was a quantitative decrease in viral mRNA but no qualitative change (88). Similar conclusions were drawn from experiments carried out in chick cells infected with influenza Ao/WSN, but an entirely different method was employed. 32P-labeled virus infection was established, and at different times after infection, RNase-resistant RNA was determined before and after annealing. In control cells treated with cycloheximide, up to 20% of the infecting RNA became RNase insensitive after annealing. In actinomycin D-treated or interferon-treated cells, almost none of the viral RNA became RNase resistant (7). These results certainly suggested that at least one effect of interferon was to inhibit primary transcription of RNA viruses with structural polymerases. Contradictory results have, however, been obtained in a study of interferon inhibition of VSV replication in interferontreated monkey or human cells. With concentrations of interferon that significantly inhibited virus replication in these systems, there was only a 30% decrease in the specific activity of viral RNA synthesis in the presence of cycloheximide and interferon in monkey cells as compared with those treated with cycloheximide alone. In the human cells, the comparable

VOL. 41, 1977

ANTIVIRAL ACTIVITY OF INTERFERONS

figure was 53%. Those viral RNA forms that were most sensitive to inhibition by interferon were forms that required the elaboration of a new polymerase to produce, i.e., were due to secondary transcription. In addition, a concentration of interferon that inhibited viral RNA synthesis by only 10% inhibited viral protein synthesis by 60% (6). The increment of inhibition seen with both cycloheximide and interferon over that with cycloheximide alone was felt to be due to the above-described protective effect of cycloheximide on mRNA and its possible absence in interferon-treated cells (3, 123). A study of the effect of interferon on early production of complementary RNA in VSV or influenza virus-infected chick or mouse cells also reached the conclusion that inhibition of primary transcription was probably not the basis of interferon action. The results of this study are, however, ambiguous to me because the authors seemed to ignore what might be significant differences in hybridization data between cycloheximide- or interferon-treated cell extracts. In fact, in view of a possible protective effect of cycloheximide on mRNA, a cycloheximide control for an interferon-treated cell culture may give very misleading results, and it was not surprising, therefore, that some differences were indeed noted in this study; these differences were ignored, presumably because the authors doubted their significance (111). In an important study on the mechanism of inhibition of reovirus replication by interferon, the authors employed a different tack, which obviated the use of cycloheximide. Among the temperature-sensitive mutants of reovirus type 3, ts447 is blocked in its formation of progeny RNA at nonpermissive temperatures (38.50C); therefore, only primary transcription takes place at this temperature. A study of the growth of ts447 in interferon-treated cells incubated at 38.50C indicated that a concentration of interferon that inhibited virus yields by 80%, inhibited primary transcription by only 12% (135). Under approximately the same conditions, however, virus-directed translation was significantly inhibited (see below). One additional finding in interferon-treated cells infected with an RNA virus will be discussed in this section although it does not, strictly speaking, deal with an inhibition of transcription. Marcus et al. (92) have also reported that treatment with chick interferon caused an elevation of the enzymatic activity of a membrane-associated RNase, with optimal activity at an alkaline pH (assays were run at pH 8). If this is a general finding, it could explain some of the divergent observations that have been made on interferon action in various

549

laboratories. It would, for instance, account for why viral transcription seems to be inhibited in some interferon-treated cells, whereas in others, viral translation seems to be the site of interferon action. By this theory of Marcus et al. (92), what results one obtains would depend on what was being measured. Intracellular viral mRNA concentrations would be decreased and virus-directed translation would be inhibited. In cell-free systems measuring viral translation, the increased nuclease in extracts of interferon-treated cells might hydrolyze the added viral mRNA, and this in turn would cause an apparent primary inhibition in translation. In order for these notions to be considered as a general mechanism of interferon action, however, it would have to be shown that increased alkaline RNase was induced by homologous interferon in cells of species other than chicks and only chick cells were used in the study of Marcus et al. (92); indeed, Maenner and Brandner have reported no increase in nuclease activity in poly(I 0)-treated monkey, hamster, quail, or duck cell extracts (86a). In addition, no specificity was shown by this system, in that the RNase activity was tested only with VSV mRNA. Since interferon action appears to show specificity for virus functions, an RNase that is thought to be responsible for interferon action should show a restricted spectrum of activity. Vaccinia virus was the first of the DNA viruses to be studied intensively with respect to interferon (50). In many ways vaccinia virus infection lends itself well to this sort of study, because the virus contains an RNA polymerase that is activated with removal of the outer coat of the virus. The polymerase, which is located in the viral core, then elaborates viral mRNA, which is translated to yield several active proteins, including one that inhibits activity of the virus structural RNA polymerase and another that is responsible for final uncoating of the viral DNA (87). The effects of interferon on this well-studied system are fairly clear-cut in both mouse (64, 65, 94) and chick cells (36, 65, 93). Early vaccinia mRNA synthesis was increased (64), but final uncoating of the viral DNA was inhibited (87). Although the production of viral mRNA depends on a virus-associated polymerase, the inhibition of the activity of the polymerase and the uncoating of the viral DNA depend on synthesis of new proteins. It would appear in this system, then, that the site of interferon-induced antiviral action must lie between virusdirected transcription and translation. In one study, however, inhibition of early vaccinia virus mRNA synthesis was reported in

550

FRIEDMAN

interferon-treated chick cells (13). This result may have been due to the use of cycloheximide in this particular study, but similar experiments in both mouse and chick cells with or without cycloheximide have repeatedly shown that transcription of vaccinia virus mRNA is increased rather than decreased in interferontreated cells (93). A study performed in frog polyhedral cytoplasmic deoxyribovirus (FPCD)-infected fathead minnow (FHM) cells indicated that the early mRNA synthesis by FPCD was inhibited by FHM interferon. The results also showed that FPCD production was inhibited by FHM interferon and that FPCD virus-associated polymerase activity was not inhibited by an early virus protein; however, there was only a 57% inhibition of early viral mRNA synthesis by 10 U of FHM interferon, a concentration that reduced virus replication by about 90%o, and a 75% mRNA inhibition by 100 U under conditions in which 20 U completely abolished virus replication. In addition, cycloheximide was employed in some of these studies (53). These results would appear to be open to the same criticisms that can be made of some of the studies with RNA virus-use of cycloheximide and a relatively small effect on primary RNA synthesis as compared with large effects on viral replication. SV40 differs from the other DNA viruses discussed in that it lacks a structural RNA polymerase and so at least early in infection must depend on a cellular polymerase. Interferon treatment inhibited SV40 T antigen production in acutely infected, but not in transformed, cells (101, 102). Transformation of mouse 3T3 cells by SV40 was also inhibited by interferon treatment (131). These results were consistent with a primary effect on either virus uncoating (140) or virus-directed transcription (103) or translation (52). Oxman and Levin next investigated the effect of interferon treatment on early SV40-specific RNA synthesis in monkey cells (103). The results indicated a marked inhibition of very early RNA synthesis under conditions where no late RNA could be formed, that is, in cells treated with enough cytosine arabinoside (20 4g/ml) to inhibit viral DNA synthesis by more than 99%. A subsequent study showed that the inhibition of early SV40 RNA synthesis in interferon-treated cells was probably not the result of inhibition of virus adsorption, penetration, or uncoating; or of increased degradation of either the viral DNA template or viral RNA; or, finally, to an inhibition of translation of a very early mRNA that might have a secondary

BACTERIOL. REV.

inhibitory effect on further early viral RNA synthesis (96). These results with SV40 were somewhat unexpected, because the virus must use a cell polymerase early in infection. Although these studies are of great interest, the previously mentioned findings that suggested an altered uncoating of SV40 virus in interferon-treated cells could account for the inhibition of early SV40 RNA synthesis (140). It will be important to see whether these experiments with infectious DNA of SV40 can be repeated. One other study with the SV40 system is of present interest. SV40 RNA prepared in vitro with SV40 DNA and a bacterial polymerase was microinjected into SV40-permissive monkey cells that had been treated with interferon. Under these conditions, T-antigen production was blocked, although it was evident in controls microinjected but not treated with interferon (52). This result indicated that, in interferon-treated and SV40-infected cells, translation of virus genetic information may be inhibited; however, this result does not necessarily mean that inhibition of translation is an important mechanism of interferon action in tbe SV40 system in vivo. Although many of the papers reviewed in this section suggested that interferon treatment caused an inhibition in early virus-directed transcription, in the case of most virus groups studied there is at least one publication that indicated no direct effect of interferon treatment on primary transcription. With RNA viruses having virus-associated polymerases, the most meaningful studies of interferon action at this point would appear to be those that employ viral mutants blocked just after primary transcription has been completed. Only one study has so far been reported with such a system (135), and the general conclusion reached was that, although there was some inhibitory effect of interferon treatment on early reoviruses RNA synthesis, it was not enough to account for the profound inhibition of virus replication. The authors felt that in interferon-treated cells, transport of subviral particles from phagocytic vacuoles and lysosomes, where they were formed, to locations in the cytoplasm where they transcribe RNA might be slowed by interferon; however, they advanced no direct evidence for this hypothesis (135). As for the DNA viruses, the weight of the evidence in the case of vaccinia virus infection is that interferon treatment has no inhibitory effect on primary viral RNA synthesis. The jury is still out on SV40, however. I, for one,

VOL. 41, 1977

ANTIVIRAL ACTIVITY OF INTERFERONS

look forward with great anticipation to further publications on the effect of interferon treatment on the replication of SV40 DNA. Evidence that Interferon Treatment Inhibits Viral Protein Synthesis It should be noted again, at the onset of the discussion of this material, that viral protein and RNA synthesis are so interdependent that it is often difficult to distinguish which of the two is the primary site of action of an antiviral substance such as interferon, since progeny RNA molecules are usually responsible for most of the total virus-directed protein and RNA synthesis that is carried on during infection. Therefore, whether interferon treatment acts directly to inhibit RNA or protein synthesis early in infection, later both transcription and translation are profoundly inhibited (124). In order to show which ofthe two is the primary site of interferon action, it is necessary to study its effect under conditions in which viral RNA and protein synthesis are clearly dissociated. This has been effected in two ways: either by studying viral protein synthesis in cell-free systems; or, among RNA viruses, by studying in infected cells the messenger function of parental (input) RNA of viruses the genome of which is an mRNA (positive-stranded RNA viruses) or of viral mRNA that is made by using the input RNA as a template in association with a viral polymerase (negative-stranded RNA viruses). There have also been important studies on one DNA virus. Experiments on viral mRNA function in cell-free systems derived from interferon-treated cells will also be discussed below. Semliki Forest virus (SFV, an arbovirus) and mengovirus (a picornavirus) have been used to determine whether positive-stranded RNA virus messenger function is a primary site of interferon action. Of the negative-stranded RNA viruses, VSV and reovirus protein synthesis have been employed. Among the DNA viruses, there has been extensive study of the effect of interferon treatment only on the early protein synthesis of vaccinia virus. Observations in virus-infected cells. In infection with SFV, inhibition of viral protein synthesis in interferon-treated cells does not appear to be a result of inhibition of viral RNA synthesis. Under conditions in which viral RNA synthesis was almost completely inhibited, early SFV protein synthesis was unaffected. If cells were treated with interferon, however, no virus-specific proteins could be identified (39). In this system, therefore, parental RNA does not seem to be translated; the RNA of parental SFV also does not become

551

RNase resistant (42, 49). The latter observation would indicate that the infecting virus fails to enter the double-stranded form and so cannot become part of a replication complex as such. In SFV infection the replication complex is a membranous structure associated with the viral polymerase and single- and doublestranded viral RNA (43, 54). One report indicated that in interferon-treated cells, parental RNA of SFV entered into a membranous structure that appeared to be a replication complex. The RNA of this structure remained in the single-stranded form, presumably because the viral polymerase was not produced (51). Some insight into the mechanism of how interferon might inhibit viral protein synthesis was provided by studies on mengovirus-infected L-cells. Less radioactivity from input radioactive mengovirus was precipitated in a post-mitochondrial fraction from interferon-treated Lcells than from untreated control cells. This precipitated fraction was found to contain 50S and 240S components, both of which had decreased activity in the extract from interferontreated cells. The author felt that the 240S component represented viral polysomes and the 50S component represented a complex between the viral RNA and the 40S ribosomal subunit (81). Although interferon treatment did appear to inhibit the formation of these structures, their nature was not clearly established, and the levels ofradioactivity were very low. Therefore, the significance of these findings must remain uncertain. Only one study has been carried out on the effect of interferon treatment on protein synthesis by the negative-stranded RNA virus VSV (141). The results indicated that 3 h after infection in interferon-treated rabbit kidney cells, the formation of all structural and nonstructural VSV proteins was inhibited. Viral RNA synthesis was not studied, but the inhibition of viral protein synthesis was so profound that, if decreased RNA synthesis were the primary factor, only a very marked inhibition could possibly account for these results. The findings in interferon-treated, VSV-infected cells (see above) indicated that, even if it does occur, inhibition of virus-directed primary transcription might not be as great as would be necessary to completely account for the observed inhibition of virus-directed protein synthesis. In the study of the effect of interferon on the growth of the reovirus mutant ts446, the results of which were discussed above with respect to primary viral transcription, inhibition of viral protein synthesis was also studied (135). Under

552

FRIEDMAN

conditions in which only primary transcription took place, there was a 20% inhibition in viral RNA synthesis but a 40 to 72% inhibition of the synthesis of various individual virus-specific polypeptides. The apparent inhibition of viral RNA synthesis did not seem great enough to the authors to account for all of the observed inhibition of virus translation. In interferon-treated cells infected with vaccinia virus, as indicated above, all (36, 64, 87, 93, 94) but one (13) of several studies indicated that early viral mRNA synthesis was not inhibited. The vaccinia virus mRNA formed in interferon-treated cells has a normal content of polyadenylic acid (70), but does not associate with ribosomes to form polyribosomes as readily as does mRNA ordinarily formed early in vaccinia virus infections (19, 64, 94, 95). This failure to form polyribosomes readily is probably the result of an inhibition of the initiation steps in viral protein synthesis, although some inhibition of chain elongation was also reported (95). There was a suggestion that the site of inhibition of initiation of peptide synthesis might be in the formation of a complex between a vaccinia virus mRNA-protein (ribonucleoprotein) and the 40S ribosomal subunit (95); however, the findings did not conclusively show this. SV40 is the only other DNA virus in which virus-directed protein synthesis in interferontreated cells has been studied at all. In interferon-treated cells there is marked inhibition of the synthesis of viral T antigen in lytic infections (102). This could have been due, however, either to the previously discussed inhibition of virus uncoating (140) or to inhibition of virusdirected transcription (103). As previously noted, there was inhibition of the translation of SV40 mRNA microinjected into interferontreated cells (52). If this as yet unconfirmed finding is correct, it suggests that inhibition of SV40 translation might take place in interferon-treated cells; however, significant inhibition of virus uncoating (140) or transcription (103) would preclude there being very much SV40 mRNA produced in interferon-treated cells. Observations in cell-free systems. All of the observations that have been made on the effects of interferon in cell-free systems have been carried out since 1966. Although late in starting, this has been a rapidly developing field, and many interesting and potentially important observations have been made; however, the results here have also been somewhat disappointing to me. Although it apparently has been possible in cell-free systems to imitate the interferon-induced inhibition of virus-directed

BACTERIOL. REV.

translation, this has not as yet led to firm conclusions on the basic mechanism of interferon's action, and, as will be seen from the discussion of the results obtained so far in these studies, there is even some disagreement as to which system best represents what is going on in the interferon-treated, virus-infected intact cell. Most studies in cell-free systems have been carried out by investigators convinced at the onset that the interferon-induced antiviral state was aimed at virus-directed translation. So far, it has been impossible to mimic the purported inhibitory action of interferon on transcription, although in the absence of publications on the subject, I am not sure how hard such studies have been pursued. The activity of viral polymerases is markedly decreased in extracts from the cytoplasm of interferon-treated cells, but this could have been due to a decrease in the production of the enzyme rather than to an inhibtion of its action (125). Neither interferon, nor extracts from interferon-treated cells, virus-infected or uninfected, inhibited viral polymerase activities (J. A. Sonnabend, personal communication). This is not to say that such inhibitory activity will not be demonstrated, but so far, there are no reports ofconditions where such inhibitory activity could be repeatedly found. In the study of interferon's effect on virusdirected protein synthesis, two sorts of systems have been employed. In the earlier group of experiments, viral RNA was incubated with extracts from control or interferon-treated cells, and the interactions were analyzed with respect to the binding ofthe RNA to the cellular components. Later, when cell-free systems employing components from animal cells became available for the production of specific polypeptides, several laboratories began to study the ability of viral RNA to stimulate amino acid incorporation by cytoplasmic fractions from interferontreated and control cells. Marcus and Salb (90), employing viral RNAbinding studies, postulated that interferon induced the production of a new translational inhibitory protein that could combine with ribosomes to inhibit their ability to translate viral (but not cellular) mRNA. When they mixed tritiated Sindbis virus RNA with cytoplasmic extracts containing 74S monomeric ribosomes, a 250S polysomal structure was formed at 00C. On incubation at 370C, this 250S polysome was broken down, and it was thought that this represented the translation of the mRNA because no breakdown occurred in the presence of cycloheximide. When ribosomes from interferontreated cells were used, however, they bound viral RNA to a decreased extent, but more im-

VOL. 41, 1977

ANTIVIRAL ACTIVITY OF INTERFERONS

portantly, the breakdown of the 250S polysome at 370C was decreased. This was taken to mean that these ribosomes were unable to translate viral mRNA. Employing similar methods in the mengovirus-L-cell system, Carter and Levy (22) showed that less viral RNA was bound at 00C to ribosomes from interferon-treated cells to form polysomes than to ribosomes from control cells. A later publication using a similar system suggested, however, that more, not less, viral RNA was bound under these conditions (30). Results from other laboratories were directly contradictory to the findings of Marcus and Salb. Kerr et al. (69) reported that no discrete polysomes were formed when viral RNA was incubated with ribosomes at 4°C, and there was no difference in the attachment of viral RNA to ribosomes from interferon-treated or control cells. When the mixtures were incubated at 370C there was breakdown of the complexes that had been formed at 40C, but this breakdown was not correlated with protein synthesis. Indeed, RNase activity was probably responsible for the breakdown of the viral RNA-ribosome complexes observed at 37°C, since there was RNase activity in cell sap and ribosome preparations. Similar observations were made by R. Lockart, Jr. (personal communication). These results, taken together, were quite depressing. There is no strong evidence that the observations of Marcus and Salb (90) or Carter and Levy (22) were related to virus-directed protein synthesis. The observations on the effect of interferon on the system could not be repeated by Kerr et al. (69). No further confirmatory publications have appeared to strengthen the binding theories advanced by these authors. It may yet turn out that these notions are the correct explanations for the antiviral activity of interferons, but it will probably not be the data in these publications that will prove the binding theories to be correct. The first publication that attempted to employ viral RNA as a messenger in a cell-free system was also by Carter and Levy (23). They showed that, although both polyuridylic acid [poly(U)]- and tobacco mosaic virus (TMV) RNA-directed protein synthesis were not inhibited in cell-free amino acid-incorporating systems derived from interferon-treated cells, mengovirus RNA-directed synthesis was. They felt that the lack of activity of the ribosomes was due to their alteration as a consequence of interferon treatment ofthe cell from which they had been derived. The nature of the products stimulated by mengovirus RNA was not investigated. Moreover, the work was difficult to repeat, partially because the salt concentra-

553

tions employed in the incubation mixtures were omitted from the published form of the manuscript and partially because, until relatively recently, it has been difficult in other hands to significantly stimulate animal cell systems with TMV RNA to incorporate amino acid into specific polypeptides (74). Since 1972, the pace of activity in this area of interferon research has quickened. This has been due, in great part, to the development of animal cell-free protein-synthesizing systems that could translate both viral and cell mRNA's with fidelity. Some very puzzling results have been published using such systems in interferon studies, and the contrasting findings of several laboratories have only partially been reconciled. One of the basic disagreements in this field is whether the antiviral activity of interferon is manifest in extracts of interferon-treated cells in the absence of some activating step. Working with L-cells, a group at Mill Hill, London (44), observed that only a small decrease in virusdirected amino acid incorporation and in the size of viral peptides formed was present in extracts derived from cells treated with interferon. When the interferon-treated cells were also infected with vaccinia or encephalomyocarditis (EMC) virus there was a marked decrease in the ability oftheir extracts to translate EMC RNA, but the stimulatory action of poly(U) on phenylalanine incorporation was not inhibited. The initiation and the elongation of virus-specific peptides were affected in the extracts from interferon-treated and virus-infected cells, but initiation seemed inhibited to a greater extent (68). One additional, and somewhat curious, finding in this system was that EMC peptide formation initiated with formylated methionyl initiator tRNA escaped the major interferon inhibition at the time of peptide chain initiation (67). This latter finding remains unexplained but tended to give rise to the notion that the interferon effect studied was an inhibition of an early step in the initiation of virus protein synthesis. It might be important to note here that, in initiation with formylated methionyl initiator tRNA, there was a loss of the requirement for one of the (normally required) initiation factors (121a). Current efforts of Kerr's group have been directed toward investigation of the apparent necessity for virus infection of cells to activate the inhibitory effect ofinterferon in the cell-free system derived from them. Addition of poly(I-C) or double-stranded RNAs from reovirus or P. chrysogenum phage to extracts from interferon-treated cells that had not been infected with viruses resulted in marked inhibition of

554

FRIEDMAN

viral RNA-directed translation (66). Indeed, addition of small volumes of a post-ribosomal supernatant fraction from interferon-treated cells to a mixture of extracts of control cells and double-stranded RNA also caused an inhibition in the translation of viral mRNA, whereas cell sap from interferon-treated cells or doublestranded RNA alone did not cause such inhibition when added to extracts from untreated cells. The observed inhibition in the presence of interferon cell sap and double-stranded RNA also required adenosine 5'-triphosphate (ATP), and preincubation of interferon cell sap with both ATP and double-stranded RNA greatly increased its inhibitory capacity. Both the double-stranded RNA and the ATP could be removed after this preincubation step without impairing the inhibitory activity of the treated cell sap (114). The results suggested that interferon treatment caused the formation of an inactive inhibitor, the activation of which was due to double-stranded RNA and probably is related to a phosphorylation step. The activated inhibitor then interacted with the proteinsynthesizing system to inhibit virus-directed translation. I presume that the addition of double-stranded RNAs to the extracts from interferon-treated cells served the same function as did infection of intact, interferontreated cells with EMC or vaccinia virus, but this remains to be proven. These results are of added interest because the regulation of protein synthesis in reticulocyte lysates required phosphorylation of an initiator methionyl-tRNAbinding factor by a protein kinase that is present in a translational inhibitor from hemedeficient systems (80). After addition of double-stranded RNA to extracts from interferon-treated mouse cells, fractionation indicated that the inhibitory effect was associated with both the microsomal and cell sap preparations. Addition of microliter quantities of cell sap from interferon-treated cells to cell-free systems from control cells made them sensitive to inhibition by double-stranded RNA forms, but the inhibition induced by double-stranded RNA in interferon cell sap was not, in their hands, reversed by the addition of tRNA. The abnormal distribution of EMC polypeptides synthesized in systems inhibited by interferon cell sap and double-stranded RNA resembles that obtained on translation of EMC RNA in cell-free systems from interferontreated, virus-infected cells. Finally, although there was an enhanced breakdown of EMC RNA in inhibited systems, it was not clear whether this was due to an induced nuclease activity or to the degradation of mRNA not engaged in translation in inhibited systems

BACTERIOL. REV.

(66a). This uncertainty is somewhat reminiscent of the above discussion of the purported inhibition of viral transcription in the presence of cycloheximide. A group from Yale has suggested that treatment with double-stranded RNA of extracts from interferon-treated Ehrlich ascites tumor cells resulted in induction of an endonuclease activity that degraded viral RNA (20a). In these extracts there was an ATP-dependent phosphorylation of at least two proteins, the apparently critical one having a molecular weight of 67,000 (78). A group from Israel (143) has presented similar findings, and in both studies (78, 143) the phosphoprotein was ribosome-associated but could be removed by a salt wash. After addition of double-stranded RNA the more rapid degradation of viral mRNA in extracts from interferon-treated cells required ATP. The degradation was biphasic: doublestranded RNA and ATP were required for the first phase (activation), whereas in the second phase (endonuclease action), neither had to be present (119a). In many respects these findings were similar to those of the group at Mill Hill, with the difference that the latter group considered the degradation possibly to be secondary to the inhibition of protein synthesis and not its cause (66a). In addition, Kerr's group reported (114a) finding at least three major phosphorylated substances in their extracts after addition of ATP and double-stranded RNA, one substance with a molecular weight of about 60,000, another of about 30,000, and the third of low molecular weight. The activated inhibitor itself was also of low molecular weight and was stable to 900C for 5 to 15 min. The heated but active inhibitor had no protein kinase activity and did not function as a kinase activator in cell-free systems. It is possible that this inhibitor was identical to the low-molecular-weight phosphorylated component. Lest the reader feel, however, that this aspect of the mechanism of action of interferon is close to being solved, it should be pointed out immediately that there is also a well-nourished body of literature which insists that extracts of uninfected, interferon-treated cells show antiviral activity with no particular manipulation or activation. In mixed cell-free protein-synthesizing systems containing ribosomes from Krebs ascites cells and cell sap fractions from interferon-treated chick fibroblasts, EMC RNA was translated somewhat less efficiently than in mixed systems containing control cell sap (65a). Also, the same Israel- and Yale-based groups that carried out the just-discussed work on a ribosome-associated phosphoprotein acti-

VOL. 41, 1977

ANTIVIRAL ACTIVITY OF INTERFERONS

vated by interferon treatment have published several important studies on inhibition of viral mRNA translation in extracts from uninfected cells. In the Israeli studies, extracts from uninfected L-cells treated with interferon translated the RNA of mengovirus less efficiently than extracts from control cells (37). The concentration range of interferon used was higher than was necessary in the previously discussed studies with cells that were both interferon-treated and later infected with virus (44). The virus inhibitory effect was associated with a ribosomal factor that could be washed off the ribosome by a buffer containing 0.5 M KCl. Somewhat similar results were obtained by the group at Yale University with a system employing extracts from Ehrlich ascites tumor cells and EMC RNA or reovirus mRNA (57). The inhibitory activity studied by the Yale group was not dialyzable and seemed to be ribosome associated. During incubation of the extracts at 30TC for 90 min, however, at least some of the translational inhibitory activity was released from ribosomes into the supernatant and therefore did not sediment at 200,000 x g for 150 min. The specificity of this inhibitory activity was also investigated. Although the antiviral activity of mouse interferon in intact L-cells was associated with a specific inhibition of viral, but not of cellular, protein synthesis, the translation of both viral and cellular mRNA's was inhibited in incubation mixtures employing extracts from interferontreated L-cells (56); therefore, the cell-free system did not seem to reflect very well what was going on in the intact cells. The Israeli group also found the same lack of discrimination by their interferon-treated, L-cell-derived system (37). The group from Mill Hill had reported a similar finding, but they had used extracts from virus-infected cells in which cellular protein synthesis was markedly inhibited; therefore, it was not surprising that translation of cell mRNA's would be inhibited in their systems (44). In general, the inability of extracts from interferon-treated cells to distinguish between host and viral mRNA's in cell-free systems has been disappointing because there is excellent discrimination in intact cells (note, however, the single exception below). A further development in the studies on inhibition of translation in extracts from uninfected, interferon-treated cells indicated that the interferon-induced inhibition could be corrected by the addition of tRNA from animal, but not bacterial, sources (30, 31, 58). Only some species of animal tRNA's seemed very

555

active in this respect. In extracts from interferon-treated cells, kinetic studies showed that translation of viral mRNA proceeded normally in L-cells but stopped after 20 to 30 min. Addition of tRNA allowed translation to resume normally. The restoration of translation of globin mRNA by tRNA seemed to require different tRNA's than did restoration for viral mRNA (31). Most of the inhibitory activity was directed against elongation of peptide chains, and inhibition of chain initiation seemed secondary to the effect on elongation. In contrast to the findings of the Mill Hill group there was a decreased binding of formylated methionyl initiator tRNA to the initiation complex in extracts from interferon-treated cells (31). In effect, therefore, the results suggested that nonfunctional polysomes were formed in extracts of interferon-treated cells and that this was due to interferon action. The formation of such inactive polysomes was, in turn, a result of a defect in some species of tRNA (31, 58). The nature of this tRNA was further elucidated by chromatographic purification of minor species of leucyltRNA (143). Addition of some of these tRNA forms allowed proper translation of mengovirus RNA in extracts from interferon-treated cells, whereas others permitted globin mRNA translation. In spite of these findings, the authors felt that the main difference between extracts from control cells and extracts from interferontreated cells was the presence in the latter of a ribosome-associated inhibitor of translation that blocked peptide chain initiation and elongation when some tRNA's were present in limited concentrations (143). Again, the group at Yale had similar findings in extracts from Ehrlich ascites cells, except that viral mRNA translation proceeded at a lower rate for 30 min in extracts from interferon-treated cells before halting; but, large, inactive polysomes were formed. This effect was also reversed with animal tRNA's. The basic defect here was traced to an impairment of amino acid acceptor activity of lysyl-, seryl-, and, especially, of leucyl-tRNA's. This was due to a faster rate of inactivation of these tRNA forms in extracts from interferon-treated cells than from control cells (119). There were, however, some disturbing findings related to this neat explanation for interferon action (29, 119). There was no difference in the amounts of tRNA from interferontreated or control cells required to correct the block in translation in extracts from interferontreated cells. In addition, extracts from interferon-treated cells inactivated leucine-specific tRNA's from interferon-treated or control cells at the same rate. Finally, and most impor-

556 FRIEDMAN tantly, filtration of an extract from interferontreated cells through Sephadex G-25 or dialysis was necessary for the extract to develop the capacity to rapidly inactivate leucyl-tRNA. In extracts that had not been so treated, there was only very marginal inhibition of the translation of viral mRNA (119). These results suggested that interferon treatment alone does tend to alter the cytoplasmic

BACTERIOL. REV.

KC1 wash fraction also inhibited the translation of viral mRNA's in extracts from untreated Krebs II ascites cells. In further work with this system, Samuel (115a) found that formation of reovirus methionyl-X initiation dipeptides was only slightly inhibited in extracts of interferon-treated cells. Also, mouse ascites cell tRNA partially reversed the inhibition of reovirus mRNA transextracts in the following manner: after passage lation, and reovirus 32P-labeled mRNA was stathrough Sephadex or dialysis, an activity is ble after a 15-min incubation with extracts from produced that tends to destabilize the charging interferon-treated cells. These results sugcapacity of some species of tRNA. This appears gested that the interferon-induced inhibition of to be an effect of interferon treatment, but the viral mRNA translation took place at a step relationship of this to the antiviral activity subsequent to formation of the first peptide developed in interferon-treated cells is very far bond and involved participation of tRNA. These findings are of great interest, since it is from being clear. It may be that these findings are related to those previously discussed in the only work I am aware of that claims to find that, in the case of extracts from uninfected the same specificity of interferon action in cellcells, either dialysis or filtration through free systems derived from uninfected cells as is Sephadex acts as an activation step, just as found in intact cells. It also suggested that the does the addition of double-stranded RNA (66); specific antiviral activity of interferon treathowever, the effects of the activation by dialy- ment was mediated by a ribosome-associated sis or Sephadex filtration (119) seem different polypeptide that inhibited virus-directed profrom those seen with extracts treated with tein synthesis but had no effect on cellular prodouble-stranded RNA (66) or with extracts tein synthesis. Unfortunately, this work has from interferon-treated and virus-infected cells not yet been confirmed. Attempts in my labora(44). It remains to be determined whether all tory to repeat the findings of Samuel and Joklik of these can be somewhat pulled together as were unsuccessful (L. Pinkus, unpublished obbeing basically a single effect of interferon servations). Therefore, although the conclusions reached by this work would go a long way treatment viewed under different conditions. Three additional publications in this area of in establishing the biochemical basis for the research merit discussion at this point. All are antiviral action of interferon, final judgment of potential interest but are so far unconfirmed. must await its confirmation by other groups. Another rather exciting finding by the Yale Samuel and Joklik (116) investigated the translation of reovirus and vaccinia virus mRNA's in group must also await confirmation (120, 120a). extracts of uninfected Krebs II ascites cells. This was the report of an inhibition of reovirus Normally, such extracts translated endogenous mRNA methylation in extracts of interferonmRNA, and added poly(U), Krebs cell or L-cell treated Ehrlich ascites cells. Most, but not all, mRNA's, or viral mRNA's quite well. When viral (and cellular) mRNA's contain blocked the cells were treated with interferon in the and methylated 5' ends, and this terminal ascites state, their extracts translated viral methylation was shown to be necessary for the mRNA's poorly, but they did translate the translation of reovirus mRNA in extracts from other mRNA's mentioned as well as extracts animal cells. The Yale group demonstrated from control cells. The unusual feature of this that the methylation of unmethylated reovirus system was that such treatment with inter- mRNA was impaired in extracts from interferon actually had to be carried out in intact feron-treated cells. This impairment was not mice by intraperitoneal injections of high-titer due to cleavage, irreversible inactivation of unmethylated reovirus mRNA, depletion of mouse interferon preparations. The results of Samuel and Joklik (116) also methyl donors, or accumulation of methylation suggested that the inhibition of virus mRNA inhibitors in the reaction mixture. The methyltranslation was due to a ribosome-associated ation inhibitor was a heat-labile macromolefactor that could be recovered by treatment of cule that decreased methylation of 5'-terminal the ribosomes with 1.0 M KC1. In addition, a 0.3 guanylate residues. In a study of the in vitro to 0.6 M KCI wash fraction of the ribosomes rate of reovirus mRNA methylation, methylafrom interferon-treated cells contained a pep- tion by core-associated enzymes was inhibited tide with a molecular weight of 48,000, which by extracts from interferon-treated cells (120a). was not found in ribosomes or washes from Again, to my knowledge, there has been no ribosomes of untreated cells. The 0.3 to 0.6 M confirmation of this work in this or any other

VOL. 41, 1977

ANTIVIRAL ACTIVITY OF INTERFERONS

system, so that its significance must remain in doubt. Moreover, interferon treatment inhibits replication of EMC virus, which appears not to have 5'-terminal methyl groups (L. Pinkus, unpublished data). However, demonstration of a specific inhibitory effect on methylation of viral mRNA's would also go far in establishing yet another biochemical basis of interferon action. A last contribution ofpotential interest is the finding that extracts from cells treated with partially purified mouse interferon caused the deacylation of aminoacylated TMV RNA or of EMC RNA. There was no alteration of aminoacylated tRNA, and a similar hydrolyzing activity was present in crude interferon preparations (118). The chief problem with this work is that there is as yet no known function in animal cells for amino acids linked to viral RNAs and, therefore, hydrolysis of such structures is of unknown significance. Also, the experiments in this study did not, to my mind, entirely rule out the possibility that the activity observed was due to a contaminating ribonucleolytic activity. Evidence that Interferon Inhibits Terminal Events in the Replication Cycle of Murine Leukemia Viruses The last major group of findings to be discussed is concerned with an apparent interferon-induced inhibition of the assembly of RNA tumor viruses. Studies on this observation have been going on for less time than those on transcription or translation; perhaps that is why contradictions and paradoxes are somewhat less common here. It has been known for some time that RNA tumor viruses are sensitive to interferon (5) and that in interferon-treated cells newly infected with murine leukemia viruses (MLV), virus yields are significantly inhibited (117). Surprisingly, there was also inhibition of virus yields in interferon-treated rat and mouse cells chronically infected with a murine RNA tumor virus (61, 132a). This was unexpected because interferon was thought to be effective only in situations where cells were treated before virus infection (124). Replication of endogenous type C RNA tumor viruses was sensitive to interferon treatment (17, 47), whereas high concentrations of interferon did not seem to prevent induction of intracisternal type A particles after bromodeoxyuridine or dimethyl sulfoxide treatment (17); however, quantitation of Aparticles is quite difficult, and their nature is in dispute. Another group found that in their system, mouse mammary tumor virus (Bparticle) production was also sensitive to interferon treatment (J. Strauchen, N. Young, and R. Friedman, manuscript in preparation).

557

The work on RNA tumor viruses and interferon took an unexpected twist when evidence was uncovered simultaneously in two laboratories that inhibition of RNA tumor virus production was not correlated with inhibition of some intracellular steps in virus replication (14, 15, 40, 47). In interferon-treated AKR cells in which there was marked inhibition of production of both endogenous MLV particles and infectious MLV, the intracellular concentration of viral p30, group-specific (gs) antigen was unaffected or even increased (40, 47, 108); furthermore, transmission electron micrographs indicated that the number of cell-associated virus particles was not depressed in interferon-treated MO-P cells, which were chronic producers of Kirsten murine sarcoma virus MSV (MLV)

(14).

More detailed studies followed these initial observations. S. Z. Shapiro, M. Strand, and A. Billiau (121) found that synthesis of the proteins p30, gp 69/71, and p15 was not inhibited in interferon-treated mouse 3T3 fibroblast cells. Synthesis and cleavage of the precursors of these proteins were also unaffected. In interferon-treated AKR cells, scanning electron micrograph (SEM) studies confirmed and extended the finding that there was no change in the number of cell-associated virus particles. SEM (24) and transmission micrographs (16) in fact clearly indicated that in mouse systems the number of surface particles was increased. This may have been the reason why cell-associated p30 (gs) antigen was increased in these systems after interferon treatment. Since the viral RNA-dependent DNA polymerase also proved to be insensitive to interferon treatment in AKR cells (40), intracellular concentrations of all of the groups of known MLV structural proteins seemed unaffected by interferon treatment. Therefore, interferon inhibition of MLV production probably does not involve inhibition of virus protein synthesis. One interesting fact to turn up in these studies was that as long as interferon was present in the medium of AKR cells that were infected with an endogenous MLV, virus production was inhibited. After removal of the interferon, full virus production again resumed and, within 24 h, it was equal to that of untreated cells (40). The reason for this probably was that, although MLV production was greatly inhibited by the interferon treatment, the provirus remained integrated into the cellular genome and upon removal of interferon, there was no reason why normal production of virus could not resume. This explanation was similar to that given for why interferon treatment did not inhibit production of T antigen by an inte-

558 FRIEDMAN grated SV40 genome (101, 105) (see below). These findings differed from those in infections with a lytic virus where repeated treatment of cells with interferon may result in complete suppression of virus replication and apparently in disappearance of all traces of the viral genome from cultures (41, 91). A number of cell cultures taken from different mouse fibroblast strains chronically infected with MLVs were tested for sensitivity to the virus-inhibitory effect of interferon (using a lytic virus infection) and for the ability of interferon to inhibit chronic MLV production (40). In all cell lines sensitive to interferon, MLV production was inhibited. In interferon-resistant lines there was no effect on MLV production. Pitha et al. (108) also tested the effect of interferon on several types of MLV infection in AKR cells. In the case of exogenous infection, virus production was delayed but not suppressed, since virus production began when interferon was removed. During inhibition by interferon of exogenous or induced infection, the findings were similar to those previously discussed in that the interferon block appeared to occur before virus assembly so that, although there was no inhibition of viral p30 (gs) antigen, there was a decrease in released virus. They also reported, however, a decrease in cell-associated viral particles, and in this respect their paper differed from previous reports on chronically infected cells. Their findings in chronically infected cells also differed in one very interesting respect from those previously discussed in that they found that the interferon-induced inhibition of assembly and release seemed to result in a relatively small decrease in virus particle release but a much greater inhibition in the production of infectious MLV (108). This observation has since been confirmed in both chronic and acute Moloney MLV infection in TB cells (25, 138). It would seem, therefore, that although in some systems interferon treatment resulted in marked inhibition of virus release, in others, particle production was almost normal, but the virus released was quite deficient in infectivity. In view of this conclusion, it would be interesting to check a report (which is an apparent exception to the above-discussed finding) that, in an interferon-sensitive cell line, interferon treatment seemed not to inhibit MLV production. Allen et al. (2) found that in JLS-V9R cells, interferon inhibited VSV production but had little effect on MLV particle production. A study of the infectivity of the viral particles produced would have been interesting in this system, for here, as in TB cells (25, 138) or in the AKR system of Pitha et al. (108), the interferon-treated cells may have re-

BACTERIOL. REV.

leased almost as many particles as control cells; however, the released particles may have been deficient in infectivity. One of the most important recent observations in RNA tumor virology was the discovery that latent RNA tumor virus infections could be activated by treatment of cells with halogenated thymidine analogs or by inhibitors of protein synthesis (1, 86). It was natural that investigators would attempt to study the effect of interferon treatment on this induction phenomenon, and it was not difficult to demonstrate that interferon pretreatment (i.e., before induction) inhibited the yield of MLV obtained in Kirsten sarcoma virus-transformed BALB/3T3 cells (KBALB/3T3) exposed to iododeoxyuridine (IUdR) (18). Interpretation of these results was difficult. Only a small percentage of the cells in a culture exposed to IUdR go on to produce virus, and the virus induced could have reins fected KBALB/3T3 cells; therefore, it was unclear whether the interferon treatment inhibited virus induction or only the subsequent infection of susceptible cells in the induced culture. As noted above, similar findings were obtained in AKR-2B cells induced by IUdR, except that marked inhibition of virus yields was obtained when interferon was present during and after IUdR treatment. The interferon treatment inhibited the number of cells producing MLV but had no effect on p30 (gs) antigen production (108). Again, however, these results were complicated by the fact that the effects of the interferon treatment rapidly wore off when the interferon was removed, and the virus produced could then reinfect susceptible cells in the culture. One study did, however, suggest that interferon might inhibit an event in the induction process itself (110). Cycloheximide activated KBALB/3T3 cells to produce only a xenotropic MLV that could infect rat cells but not mouse cells (1). In a system using cycloheximide induction of KBALB/3T3 mouse cells with rat embryo kidney as detector cells, interferon treatment also decreased the virus yield, so it appeared that some event in the induction process was inhibited by interferon treatment. The findings of Wu et al. (139) differed in two important respects from those just described. They found that interferon treatment had no effect on the induction by IUdR of an N-tropic virus from KBALB cells, whereas induction of a xenotropic virus was inhibited in the same system. Also, the intracellular concentration of viral gs antigen was decreased when virus production was inhibited by interferon treatment. In other studies, the tropism of virus seemed to make no difference as far as interferon treat-

VOL. 41, 1977

ANTIVIRAL ACTIVITY OF INTERFERONS

ment was concerned, since N, B, NB, or xenotropic virus production and the replication of the NZB agent were inhibited (40). Also, in several other studies, the intracellular concentrations of p30 (gs) antigen were not decreased by interferon treatment, and indeed were in some cases actually increased (47, 108). The reasons for the differences between the results of Wu et al. (139) and those of other investigators are not clear. In the case of the p30 (gs) antigen, it may be that the differences are due to the methods employed for detecting the antigen, since Wu et al. (139) used a complement fixation technique, whereas others employed a radioimmunoprecipitation inhibition assay (47, 108). As the results obtained with interferon treatment of MLV-infected cells seem at least superficially to be inconsistent with the findings in other systems, it would be useful to pin down the site of this action, if possible, to a specific event in the virus replication cycle. In order to try to do this, advantage has been taken of the properties of a temperature-sensitive (ts) mutant of Moloney MLV, ts3 (136, 137). At nonpermissive temperatures (390C), this mutant fails to bud normally from the surface of TB cells. When the temperature was lowered, however, ts3 virus particles collected on the cell surface were released within 30 min, even in the presence of cycloheximide. The SEM of TB cells infected with ts3 at 390C, therefore, strikingly resembled those from interferon-treated AKR cells infected with MLV. It was of interest to establish the temporal relationship in the growth cycle between the site of the ts3 mutation and the interferon-sensitive step. TB cells chronically (25) or acutely (138) infected with the ts3 mutant were employed for these experiments. In all cases in this sort of study, there was a very rapid release of virus particles in interferon-treated cells after a temperature downshift (25, 138). As previously found by Pitha et al. (108) there was a much greater inhibition of viral infectivity than of virus particle production; however, the early release of viral antigen and transcriptase-containing viral particles after the temperature downshift was about as rapid in interferontreated as in control cells. The results suggested that, in interferon-treated TB cells infected with ts3 and held at a nonpermissive temperature, there was a population of virus particles that was released from the cell surface immediately upon the temperature downshift. These particles must, therefore, have been insensitive to the effect of interferon, and the results taken together would suggest that the block in ts3 replication at 390C must occur just

559

after the interferon-induced block. The results indicated that the morphogenesis and release of MLV is a complex event and that interferon treatment inhibits a comparatively early stage in the process. Morphologically, many more virus particles were in a very early budding phase in the interferon-treated cells than in controls (138), a result that also tended to confirm the notion that interferon inhibited an early event in the morphogenesis of MLVs. The decrease in infectivity of the particles that are actually released from the interferon-treated cells in some systems (25, 108, 138) is probably also related to an interferon effect on this early step in morphogenesis. INTERFERON TREATMENT IS INEFFECTIVE IN SYSTEMS IN WHICH THE SV40 GENOME IS INTEGRATED INTO AN INTERFERON-RESISTANT VIRUS OR HOST GENOME A special case where interferon is inactive in a situation where normal inhibitory activity might be expected requires additional consideration. Interferon treatment inhibited lytic infection with SV40 virus (101). Not only was infectious virion formation inhibited but also production of T-antigen, an early gene product, was decreased (102). In mouse cells, SV40 does not undergo a complete replication cycle; nevertheless, treatment with interferon inhibited both viral T-antigen production and viralinduced cell transformation (131). The stimulation of cellular DNA synthesis in BHK-21 hamster cells after infection with polyoma virus (which is closely related to SV40) was also inhibited by treatment with hamster interferon (34). All of these findings clearly indicated that SV40 and polyoma virus functions are sensitive to interferon treatment. Evidence has already been discussed indicating that the site of interferon action in this case may be at any one of several stages in the virus replication cycle including the uncoating step, or virus-directed transcription or translation. It was, therefore, surprising that in SV40 virus-transformed mouse cells the production of SV40 T-antigen was insensitive to interferon treatment in spite of the fact that such treatment made them resistant to VSV infection. That is, the same interferon system that recognized and inhibited new lytic infections with SV40 and other viruses apparently failed to inhibit viral fimction when it was exercised by an integrated tumor virus genome (101). This interpretation was strengthened by findings in cells infected with an adeno-SV40 hybrid virus that is infectious and presents an interesting genetic combination of an interferon-sensitive

560

FRIEDMAN

BACTERIOL. REV.

(SV40 virus) with an interferon-insensitive (ad- cation. These generalities about interferon acenovirus) genome. In stimultaneous infection of tion resemble to some extent what is known cells with both complete viruses the sensitivity about the biological activities of glycopeptide of adenovirus or SV40 T-antigen production hormones and bacterial toxins. One peptide was characteristic of infection with either virus component of the hormone or toxin is responsiseparately; however, in infection with the ble for binding to specific glycolipid or glycoSV40-adenovirus hybrid, production of both T- lipid-like sites on the plasma membrane, and antigens was as resistant as adenovirus T- this binding effects alterations in the state of antigen production in infection with adenovirus the membrane. Another portion of the protein alone (105). acts through an adenyl cyclase-cAMP mechanThere is strong independent evidence that sim and is responsible for carrying out the spethe SV40 genome is covalently linked to the cific intracellular steps that characterize the adenovirus genome in the hybrid (104) or to action of the hormone or toxin (99). There are cellular DNA in SV40-transformed cells (115). thus some similarities to what is already The mRNA produced by the integrated SV40 known of the induction of interferon action. genome contains host sequences (84), and the Perhaps, the establishment of interferon-inmRNA of the hybrid contains both adenovirus duced antiviral activity will closely resemble and SV40 sequences (104). The resistance of these systems. SV40 T-antigen production to interferon treatThere has been great interest in the mechament in the case of integrated genomes may nism of action of the putative antiviral subindicate that the primary sequence of nucleo- stance that is thought to be responsible for tides in the genome does not determine sensi- interferon's activity in inhibiting virus replicativity to interferon. Other sites on the genome tion. The notion that interferon acts to inhibit such as those concerned with initiation or con- primary transcription of viral mRNA is based trol of genetic expression would then have to be for the most part on findings in cells infected the loci of interferon action, and the presence of with viruses that have a structural polymerase host or interferon-resistant virus RNA se- (7, 89) or with SV40 (103). In the case of the quences would render the virus mRNA resist- former, use of cycloheximide in many studies ant to interferon action. This inference does seems to me to have led to the erroneous conclunot, however, take into account the previously sion that primary transcription by viruses is discussed finding of Yamamoto et al. (140) that decreased in interferon-treated cells. This is interferon inhibited the uncoating of SV40 vi- possible because cycloheximide treatment in rus. If this finding is correct, the reason why the absence of interferon might tend to protect integrated SV40 genomes seem resistant to in- mRNA's by forming stable polysomes (3, 123). terferon treatment would be quite clear: the Recent findings suggest that after interferon virus uncoating is not a step necessary for ac- treatment, however, initiation of virus protein tivity of the viral genome. synthesis is inhibited and, if this is so, cycloheximide treatment would have less of a tenDISCUSSION dency to stabilize viral mRNA in interferonInterferon would appear to bind to glycolipids treated cells (37, 44). Furthermore, in a study of or glycolipid-like structures on the plasma a reovirus infection that employed a temperamembrane (10, 75, 133). This binding activity ture-sensitive mutant blocked at nonpermismay induce chemical, physical, and morpholog- sive conditions just after primary transcription, ical alterations in the plasma membranes, and interferon treatment had very little effect on these alterations may be directly related to the the production of viral mRNA (135). Somewhat induction of antiviral activity in cells in which similar results have been obtained with an such activity is produced. Binding alone is not analogous temperature-sensitive mutant of sufficient to bring about antiviral activity, VSV (P. I. Marcus, personal communication). though it induces some alterations in the mem- To me, the results of all of these studies, therebrane (75). Although nothing specific is known fore, do not conclusively indicate an interferon about what steps must follow binding in order effect on primary transcription. to give rise to antiviral activity, there is eviHowever, Oxman et al. (96, 103), on the basis dence that cAMP (45) may be involved in this of carefully conducted studies on SV40-infected process and that cellular RNA (130) and protein cells, have also reached the conclusion that insynthesis (48) are necessary for the develop- terferon treatment inhibits virus-directed tranment of antiviral activity; it has been thought scription early in infection. In spite of the care by many that an antiviral substance must be that Oxman et al. (96, 103) took to reach this induced by interferon treatment and that this is conclusion, the work of Yamamoto et al. (140) the active moiety in the inhibition ofviral repli- would seem to indicate that the effect of inter-

VOL. 41, 1977

ANTIVIRAL ACTIVITY OF INTERFERONS

may be on a step even earlier than pritranscription of SV40. In summary, I feel work published to the present time does not argue compellingly that interferon has a significant inhibitory effect on viral transcription. In addition, studies with temperature-sensitive mutants of VSV and reovirus (135) present data to suggest that there is no significant effect on viral transcription. Related to possible inhibition of translation in interferon-treated cells are reports of induced ribonucleolytic activity in cells after interferon treatment (92) or in extracts from such cells to which double-stranded RNA was added (119a). Although such a putative mechanism of action of interferon is of great possible interest, several observations suggest caution in accepting this as necessarily the explanation for interferon action. Repeated observations in vaccinia virus-infected interferon-treated cells indicated that viral mRNA production was increased over that found in appropriate controls (64, 94). This would be unlikely if interferon induced an RNase activity that was responsible for inhibiting virus replication. The RNase activity induced by addition of double-stranded RNA to cell-free systems from interferontreated cells may not be related to antiviral activity in these systems, because the kinase activity induced in such extracts resulted in the phosphorylation of several proteins clearly not related to the antiviral state and even of histones added to the system (114a). It is therefore possible that the RNase activity induced in such extracts was related to the kinase activation but was not functional in interferontreated cells and, indeed, could be detected only in extracts in which virus-directed translation had already been decreased as a consequence of an inhibition of viral mRNA-ribosome association induced by interferon treatment (66a). Studies in intact cells clearly point to a mechanism that involves inhibition of virus-directed translation under conditions where viral transcription did not take place (39). Several studies in cell-free systems involving translation of virus genetic information suggest that interferon treatment results in the production of an inactive precursor of the antiviral substance (78, 114, 143); some aspect of infection with virus (possibly production of a doublestranded RNA) may be the most important activator in vivo (66). The role of virus infection in these systems may be mimicked by incubating uninfected, interferon-treated cells in a buffered salt medium to reduce their level of protein synthesis (68). Cell-free systems from interferon-treated cells incubated in this way have a reduced capacity to translate viral

feron

mary

561

mRNA. Thus, virus-induced inhibition of host protein synthesis might be responsible for the activation of the interferon-induced antiviral factor. In cell-free systems, addition of doublestranded RNA or passage through Sephadex may also activate the system (120). This activation step may involve a phosphorylation process requiring ATP (78, 114, 143). Assuming all of this is correct, what virus function does the antiviral substance block? The best bet to my mind, at present, is a step in the initiation of protein synthesis. Results with formylated initiator tRNA point to a function involving an early event in the initiation of protein synthesis as the site directly affected in interferontreated cells (67). Also, the nature of the mRNA may be very important to interferon action, because SV40 mRNA with host sequences from interferon-insensitive adenovirus is translated normally in interferon-treated cells (101, 105). Inhibition of virus-directed translation at the initiation site might allow for the remarkable specificity shown by interferon.. There is, however, the unusual inhibitory effect of interferon treatment on the replication of RNA tumor viruses to consider. In these systems, virus protein and RNA synthesis do not seem to be decreased; however, virus assembly is disrupted, and this results in the production of very little extracellular virus (virus remains attached to the cell surface) or of virus with a greatly decreased infectivity (17, 47, 108). The lack of effect on these virus biosynthetic functions is probably related to the integration of the provirus of RNA tumor viruses into the cellular genome. As in the case of some SV40-transformed cells, the viral mRNA produced under these circumstances may include host sequences that act to prevent the interferon-induced mechanism from recognizing the genome as viral (101). The decrease in yields of infectious RNA tumor viruses could, on the other hand, be due to interferon-induced alterations in the plasma membrane, as could the reported inhibition of a very early step in SV40 replication which may involve virus uncoating (140). Such alterations may preclude normal assembly of murine leukemia viruses or uncoating and transcription of SV40. In the case of other viruses that assemble on the cell surface and bud from the plasma membrane (such as arboviruses or myxoviruses) the viral genome is not integrated, and the primary effect of interferon would be to inhibit the translation of virus genetic information. It might be, however, that, under some circumstances in lytic infections, virus with decreased infectivity could be produced in interferon-treated cells.

562

FRIEDMAN

If interferon has a single mechanism of action, it might be summarized as follows. Interferons bind to a specific receptor on the cell surface. In cells with the proper effector apparatus (sensitive cells) this binding causes changes in the cell membrane which, possibly through a cAMP mechanism, result in the production of an inactive precursor of an antiviral substance. After viral infection and formation of a double-stranded viral RNA, this precursor is activated in a step involving phosphorylation to produce an antiviral substance that selectively inhibits a step in the initiation oftranslation of viral mRNA. Although this scheme is speculative, the speculation is based on the results of many excellent studies. If, however, Ockham's razor does not apply here, interferon treatment may inhibit several or all of the steps in virus replication discussed in this review. An alteration induced in the plasma membrane by interferon treatment might help to account for the inhibition of MLV production and the uncoating of SV40, as well as for some of the putative non-antiviral actions of interferons. These include inhibition of cell replication (72, 77, 106, 132), enhancement of phagocytosis (62), regulation of the immune response (63), increased specific cytotoxicity of sensitized lymphocytes (83), and increased susceptibility to the toxicity of poly(I C) (127). There seem to be at least two "fail-safe" steps in the mechansim proposed. The correct interferon must be bound to the cell surface in order to produce the inactive precursor of the antiviral substance. Then, virus infection must follow to activate the precursor. One reasonable explanation for this intricate series of steps might be that interferon treatment affects the cell in ways other than simply to produce an antiviral activity. These would include the above-mentioned reports of interferon inhibiting cell replication and mediating the immune response. If these reports are correct, it is likely that the effects of interferon treatment must be carefully regulated by the cell, and such "failsafe" control mechanisms might be quite important. Lastly, one has to wonder why such a complicated series of biological processes as the interferon mechanism evolved to control virus infections. Interferon is an extracellular protein whose induction and production are carefully regulated. When interferon is bound to the proper cell, further actions are induced and these, also with careful regulation, result in an intracellular antiviral activity. This seems an extraordinarily complex mechanism for an apparently simple end, the inhibition of virus growth. Perhaps, what have been considered

BACTERIOL. REV.

here as "other" actions of interferon, the control ofcell growth and expression are, in reality, the main point of it all. On the other hand, if induction of antiviral activity is indeed the main biological role of interferon, nature would seem to have found it necessary to work in exceedingly complex ways. The only solace for a scientist working on interferon is that other biological mechanisms about which we know some details, such as the complement and blood clotting systems, or the inflammatory response seem, if anything, a great deal more complex than interferon action. LITERATURE CITED 1. Aaronson, S. H., and C. Y. Dunn. 1974. Endogenous C-type viruses of BALB/c cells: frequencies of spontaneous and chemical induction. J. Virol. 13:181-185. 2. Allen, P. T., H. Schellekens, L. J. L. D. Van Griensven, and A. Billiau. 1976. Differential sensitivity of Rauscher murine leukemia virus (MuLV-R) to interferons in two interferon-responsive cell lines. J. Gen. Virol. 31:429-435. 3. Allende, C. C., J. E. Allende, and R. A. Firtel. 1974. The degradation of ribonucleic acids injected into Xenopus laevis oocytes cell. Cell 2:189-196. 4. Ankel, H., C. Chany, B. Galliot, M. J. Chevalier, and M. Robert. 1973. Antiviral effect of interferon covalently bound to sepharose. Proc. Natl. Acad. Sci. U.S.A. 70:2360-2363. 5. Bader, J. P. 1962. Production of interferon by chick embryo cells exposed to Rous sarcoma virus. Virology 16:436-443. 6. Baxt, B., J. A. Sonnabend, and R. Bablanian. 1977. Effects of interferon on vesicular stomatitis virus transcription and translation. J. Gen. Virol. 35:325-334. 7. Bean, W. J., Jr., and R. W. Simpson. 1973. Primary transcription of the influenza virus genome in permissive cells. Virology 56:646651. 8. Berman, B., and J. Vilcek. 1974. Cellular binding characteristics of human interferon. Virology 57:378-386. 9. Besancon, F., and H. Ankel. 1974. Inhibition of interferon action by plant lectins. Nature (London) 250:784-786. 10. Besancon, F., and H. Ankel. 1974. Binding of interferon to gangliosides. Nature (London) 252:478-480. 11. Besancon, F., and H. Ankel. 1976. Inhibition de l'action de l'interf6ron par des hormones glycoproteiques. C. R. Acad. Sci. Paris 283:1807-1810. 12. Besancon, F., H. Ankel, and S. Basu. 1976. Specificity and reversibility of interferon ganglioside interaction. Nature (London) 259:576-578. 13. Bialy, H. S., and C. Colby. 1972. Inhibition of early vaccinia virus ribonucleic acid synthesis in interferon-treated chicken embryo fi-

VOL. 41, 1977

ANTIVIRAL ACTIVITY OF INTERFERONS

broblasts. J. Virol. 9:286-289. 14. Billiau, A., V. G. Edy, E. DeClercq, H. Hermans, and P. DeSomer. 1975. Influence of interferon on the synthesis of virus particles in oncornavirus carrier cell lines. III. Survey of effect on A-, B-, and C-type oncornaviruses. Int. J. Cancer 15:947-953. 15. Billiau, A., V. G. Edy, H. Sobis, and P. DeSomer. 1974. Influence of interferon on virus particle synthesis in oncornavirus carrier lines. II. Evidence for a direct effect on particle release. Int. J. Cancer 14:335-340. 16. Billiau, A., H. Heremans, P. T. Allen, J. De Maeyer-Guignard, and P. DeSomer. 1976. Trapping of oncornavirus particles at the surface of interferon-treated cells. Virology 73:537-542. 17. Billiau, A., H. Sobis, and P. DeSomer. 1973. Influence of interferon on virus particle formation in different oncornavirus carrier cell lines. Int. J. Cancer 12:646-653. 18. Blaineau, C., T. Kishida, M. Salle, and J. Peries. 1975. Inhibitory effect of interferon preparations on the endogenous mouse C-type viruses by iodeoxyuridine. IRCS Med. Sci. 3:258-261. 19. Bodo, G., W. Scheirer, W. Suh, B. Schultze, I. Horak, and C. Jungwirth. 1972. Protein synthesis in pox-infected cells treated with interferon. Virology 50:140-147. 20. Bourgeade, M. F. 1974. Interferon cell species specificity: role of cell membrane receptors. Proc. Soc. Exp. Biol. Med. 146:820-824. 20a.Brown, G. E., B. Lebleu, H. Kowakita, S. Shala, G. C. Sen, and P. Lengyel. 1976. Increased endonuclease activity in an extract from mouse Ehrlich ascites tumor cells which had been treated with a partially purified interferon preparation: dependence on double-stranded RNA. Biochem. Biophys. Res. Commun. 69:114-122. 21. Buckler, C. E., S. Baron, and H. Levy. 1966. Interferon: lack of detectable uptake by cells. Science 152:80-82. 22. Carter, W. A., and H. B. Levy. 1967. Ribosomes. Effect of interferon on their interactions with rapidly-labeled cellular and viral RNAs. Science 155:1254-1257. 23. Carter, W. A., and H. B. Levy. 1968. The recognition of viral RNA by mammalian ribosomes. An effect of interferon. Biochim. Biophys. Acta 155:437-443. 24. Chang, E. H., S. J. Mims, T. J. Triche, and R. M. Friedman. 1977. A morphologic study of interferon inhibition of murine leukemia virus release. Correlation with biochemical data. J. Gen. Virol. 34:363-367. 25. Chang, E. H., M. W. Myers, P. K. Y. Wong, and R. M. Friedman. 1977. The inhibitory effect of interferon on a temperature-sensitive mutant of Moloney murine leukemia virus. Virology 77:99-101. 26. Chany, C. 1976. Membrane-bound interferon specific cell receptor system: role in the establishment and amplification of the antiviral system. Biomedicine 25:148-157.

563

27. Chany, C., A. Grigoire, M. Vignal, J. Lemaitre-Moncuit, P. Brown, F. Besancon, H. Suarez, and R. Cassingena. 1973. Mechanism of interferon uptake in parental and somatic monkey-mouse hybrid cells. Proc. Natl. Acad. Sci. U.S.A. 70:557-561. 28. Chany, C., M. Vignal, P. Couillin, N. V. Cong, J. Boue, and A. Boue. 1975. Chromosomal localization of human genes governing the interferon-induced antiviral state. Proc. Natl. Acad. Sci. U.S.A. 72:3129-3133. 29. Colby, C., E. E. Penhoet, and C. E. Samuel. 1976. A comparison of transfer RNA from untreated and interferon-treated murine cells. Virology 74:262-264. 30. Content, J., B. Lebleu, U. Nudel, A. Zilberstein, H. Berissi, and M. Revel. 1975. Blocks in elongation and initiation of protein synthesis induced by interferon treatment in mouse L cells. Eur. J. Biochem. 54:1-10. 31. Content, J., B. Lebleu, A. Zilberstein, H. Berissi, and M. Revel. 1974. Mechanism of the interferon-induced block of mRNA translation in mouse L cells: reversal of the block by transfer RNA. FEBS Lett. 41:125-130. 32. Creagan, P. P., Y. H. Tan, S. Chen, and F. H. Ruddle. 1975. Somatic cell genetic analysis of the interferon system. Fed. Proc. 34:22222226. 32a.DeClercq, E., V. G. Edy, and J-J. Cassiman. 1976. Chromosome 21 does not code for an interferon receptor. Nature (London) 264:249-251. 33. DeSomer, P., A. Prinzie, P. Denys, Jr., and E. Schonne. 1962. Mechanism of action of interferon. I. Relationship with viral ribonucleic acid. Virology 16:63-70. 34. Dulbecco, R., and T. Johnson. 1970. Interferon-sensitivity of the enhanced incorporation of thymidine into cellular DNA induced by polyoma virus. Proc. Natl. Acad. Sci. U.S.A. 53:403-410. 35. Epstein, L. B., and C. J. Epstein. 1976. Localization of the gene AVG for the antiviral expression of immune and classical interferon to the distal portion of the long arm of chromosome 21. J. Infect. Dis. 133:A56-A62. 36. Esteban, M., and D. H. Metz. 1973. Inhibition of early vaccinia virus protein synthesis in interferon-treated chicken embryo fibroblasts. J. Gen. Virol. 20:111-115. 37. Falcoff, E., R. Falcoff, B. Lebleu, and M. Revel. 1973. Interferon treatment and the inhibition of in vitro mRNA translation in noninfected L cells. J. Virol. 12:421-430. 38. Friedman, R. M. 1967. Interferon binding: the first step in establishment of antiviral activity. Science 156:1760-1761. 39. Friedman, R. M. 1968. Interferon inhibition of arbovirus protein synthesis. J. Virol. 2:10811085. 40. Friedman, R. M., E. H. Chang, J. M. Ramseur, and M. W. Myers. 1975. Interferondirected inhibition of chronic murine leukemia virus production in cell cultures: lack of effect on intracellular viral markers. J. Vi-

564

FRIEDMANBBACTERIOL. REV.

rol. 16:569-574. 41. Friedman, R. M., and J. C. Costa. 1976. Fate of interferon-treated cells. Infect. Immun. 13:487-493. 42. Friedman, R. M., K. H. Fantes, H. B. Levy, and W. A. Carter. 1967. Interferon action on parental Semliki Forest virus ribonucleic acid. J. Virol. 1:1168-1173. 43. Friedman, R. M., J. G. Levin, P. M. Grimly, and I. K. Berezesky. 1972. Membrane-associated replication complex in arbovirus infection. J. Virol. 10: 504-515. 44. Friedman, R. M., D. H. Metz, R. M. Esteban, D. R. Tovell, L. A. Ball, and I. M. Kerr. 1972. Mechanism of interferon action: inhibition of viral messenger ribonucleic acid translation in L-cell extracts. J. Virol. 10:1184-1198. 45. Friedman, R. M., and I. Pastan. 1969. Interferon and cyclic-3'5'-AMP: potentiation of antiviral action. Biochem. Biophys. Res. Commun. 36:735-740. 46. Friedman, R. M., and I. Pastan. 1970. Mechanism of action of 2-mercapto-1 (f3-4-pyridethyl) benzimidazole: a reversible inhibitor of interferon activity. Proc. Natl. Acad. Sci. U.S.A. 65:104-109. 47. Friedman, R. M., and J. M. Ramseur. 1974. Inhibition of murine leukemia virus production in chronically infected AKR cells: a novel effect of interferon. Proc. Natl. Acad. Sci. U.S.A. 71:3542-3544. 48. Friedman, R. M., and J. A. Sonnabend. 1964. Effect of p-fluorophenylalanine on interferon action. Nature (London) 203:366-367. 49. Friedman, R. M., and J. A. Sonnabend. 1965. Inhibition by interferon of production of double-stranded Semliki Forest virus RNA. Nature (London) 206:532. 50. Friedman, R. M., J. A. Sonnabend, and H. McDevitt. 1965. Interferon inhibition of cytoplasmic DNA accumulation in vaccinia virus infection. A radioautographic study. Proc. Soc. Exp. Biol. Med. 119:551-553. 51. Friedman, R. M., and T. Sreevalsan. 1970. Membrane binding of input arbovirus ribonucleic acid: effect of interferon or cycloheximide. J. Virol. 6:169-175. 52. Graessmann, A., M. Graessmann, H. Hoffmann, J. Niebel, G. Brandner, and N. Mueller. 1974. Inhibition by interferon of SV40 tumor antigen formation in cells injected with SV40 cRNA transcribed in vivo. FEBS Lett. 39:249-251. 53. Gravell, M., and T. L. Cromeans. 1972. Inhibition of early viral ribonucleic acid synthesis in interferon-treated cells infected with frog polyhedral cytoplasmic deoxyribovirus. Virology 50:916-919. 54. Grimley, P. M., J. G. Levin, I. K. Berezesky, and R. M. Friedman. 1972. Specific membranous structures associated with the replication of group A arboviruses. J. Virol. 10:492503. 55. Grossberg, S. E., and J. J. Holland. 1962. Interferon and viral ribonucleic acid. Effect on

virus-susceptible and insusceptible cells. J. Immunol. 88:708-714. 56. Gupta, S. L., W. D. Graziadei III, H. Weideli, M. L. Sopori, and P. Lengyel. 1974. Selective inhibition of viral protein accumulation in interferon-treated cells; nondiscriminate inhibition of the translation of added viral and cellular messenger RNAs in their extracts. Virology 57:49-63. 57. Gupta, S. L., M. L. Sopori, and P. Lengyel. 1973. Inhibition of protein synthesis directed by added viral and cellular messenger RNAs in extracts of interferon-treated Ehrlich ascites tumor cells. Biochem. Biophys. Res. Commun. 54:777-783. 58. Gupta, S. L., M. L. Sopori, and P. Lengyel. 1974. Release of the inhibition of messenger RNA translation in extracts of interferontreated Ehrlich ascites tumor cells by added transfer RNA. Biochem. Biophys. Res. Commun. 57:763-770. 59. Ho, M. 1961. Inhibition of the infectivity of poliovirus RNA by interferon. Proc. Soc. Exp. Biol. Med. 107:639-644. 60. Ho, M. 1962. Kinetic considerations of the inhibitory action of an interferon produced in chick cultures infected by Sindbis virus. Virology 17:262-275. 61. Horvath, S. E., and R. M. Friedman. 1971. Nucleic acid and proteins isolated from a strain of murine sarcoma virus (MSV-O). Proc. Soc. Exp. Biol. Med. 137:1075-1081. 62. Huang, K-Y., R. M. Donahoe, F. B. Gordon, and H. R. Dressier. 1971. Enhancement of phagocytosis by interferon-containing preparations. Infect. Immun. 4:581-588. 63. Johnson, H. M., and S. Baron. 1976. Regulatory role of interferons in the immune response. IRCS Med. Sci. 4:50-52. 64. Joklik, W. K., and T. C. Merigan. 1966. Concerning the mechanism of action of interferon. Proc. Natl. Acad. Sci. U.S.A. 56:558565. 65. Jungwirth, C., I. Horak, G. Bodo, J. Lindner, and B. Schultze. 1972. The synthesis of poxvirus-specific RNA in interferon-treated cells. Virology 48:59-70. 65a. Kerr, I. 1971. Protein synthesis in cell-free systems: an effect of interferon. J. Virol. 7:448-459. 66. Kerr, I. M., R. E. Brown, and L. A. Ball. 1974. Increased sensitivity to double-stranded RNA of protein synthesis in cell-free systems from interferon-treated cells. Nature (London) 250:57-59. 66a. Kerr, I. M., R. E. Brown, M. J. Clemens, and C. S. Gilbert. 1976. Interferon-mediated inhibition of cell-free protein synthesis in response to double-stranded RNA. Eur. J. Biochem. 69:551-561. 67. Kerr, I. M., R. M. Friedman, R. E. Brown, L. A. Ball, and J. C. Brown. 1974. The inhibition of protein synthesis in cell-free systems from interferon-treated cells: further characterization and effect of formylmethionyltRNAf. J. Virol. 13:9-21.

VOL. 41, 1977

ANTIVIRAL ACTIVITY OF INTERFERONS

68. Kerr, I. M., R. M. Friedman, R. M. Esteban, R. E. Brown, L. A. Ball, D. A. Metz, D. Risby, D. R. Tovell, and J. A. Sonnabend. 1973. The control of protein synthesis in interferon-treated cells. Adv. Biosci. 11:109126. 69. Kerr, I. M., J. A. Sonnabend, and E. M. Martin. 1970. Protein-synthetic activity of ribosomes from interferon-treated cells. J. Virol. 5:132-144. 70. Klesel, N., W. Wolf, G. Heller, C. Jungwirth, and G. Bodo. 1974. Synthesis of poly(A) sequences in poxvirus-infected cells treated with interferon or poly(rI)-poly(rC). Virology 57:570-573. 71. Knight, E., Jr. 1976. Interferon: purification and initial characterization from human diploid cells. Proc. Natl. Acad. Sci. U.S.A. 73:520-523. 72. Knight, E., Jr. 1976. Antiviral and cell growth inhibitory activities reside in the same glycoprotein of human fibroblast interferon. Nature (London) 262:302-303. 73. Knight, E., Jr., and B. D. Korant. 1977. A surface alteration in mouse L-cells induced by interferon. Biochem. Biophys. Res. Commun. 74:707-713. 74. Knowland, J., T. Hunter, T. Hunt, and D. Zimmern. 1975. Translation of tobacco mosaic virus RNA and isolation of the messenger for TMV coat protein. Colloq. Inst. Nat. Sante Rech. Med. 47:211-216. 75. Kohn, L. D., R. M. Friedman, J. M. Holmes, and G. Lee. 1976. Use of thyrotropin and cholera toxin to probe the mechanism by which interferon initiates its antiviral activity. Proc. Natl. Acad. Sci. U.S.A. 73:36953699. 76. Kohno, S., C. E. Buckler, H. B. Levy, and S. Baron. 1975. Elution of cell-bound interferon as a cause of development of antiviral activity subsequent to washing of interferontreated cell culture. IRCS Med. Sci. 3:57. 77. Kuwata, T., A. Fuse, and N. Morinaga. 1976. Effects of interferon on cell and virus growth in transformed human cell lines. J. Gen. Virol. 33:7-15. 78. Lebleu, B., C. Sen, S. Shaila, B. Cabier, and P. Lengyel. 1976. Interferon, doublestranded RNA, and protein phosphorylation. Proc. Natl. Acad. Sci. U.S.A. 73:3107-3111. 79. Lebon, P., M.-C. Moreau, L. Cohen, and C. Chany. 1975. Different effects of ouabain on interferon production and action. Proc. Soc. Exp. Biol. Med. 149:108-112. 80. Levin, D. H., R. S. Ranu, V. Ernst, and I. M. London. 1976. Regulation of protein synthesis in reticulocyte lysates: phosphorylation on methionyl-tRNAf binding factor by protein kinase activity of translational inhibitor isolated for heme-deficient lysates. Proc. Natl. Acad. Sci. U.S.A. 73:3112-3116. 81. Levy, H. B., and W. A. Carter. 1968. Molecular basis of interferon action. J. Mol. Biol. 31:561-577. 82. Levy, H. B., and F. L. Riley. 1973. Effect of

565

interferon treatment on cellular messenger RNA. Proc. Natl. Acad. Sci. U.S.A. 70:38153819. 83. Lindahl, P., P. Leary, and I. Gresser. 1972. Enhancement by interferon of the specific cytotoxicity of sensitized lymphocytes. Proc. Natl. Acad. Sci. U.S.A. 69:721-725. 84. Lindberg, U., and J. E. Darnell. 1970. SV-40 specific RNA in the nucleus and polyribosomes of transformed cells. Proc. Natl. Acad. Sci. U.S.A. 65:1089-1096. 85. Lindenmann, J., D. C. Burke, and A. Isaacs. 1957. Studies on production, mode of action and properties of interferon. Br. J. Exp. Pathol. 38:551-562. 86. Lowy, D. R., W. P. Rowe, N. Tiech, and J. W. Hartley. 1971. Murine leukemia virus: high frequency activation in vitro by 5-iododeoxyuridine and 5-bromodeoxyuridine. Science 174:155-156. 86a.Maenner, H., and G. Brandner. 1976. Interferon action may not be effected by induction of a cellular ribonuclease. Nature (London) 250:637-638. 87. Magee, W. E., S. Levine, 0. V. Miller, and R. D. Hamilton. 1968. Inhibition by interferon of the uncoating of vaccinia virus. Virology 35:505-511. 88. Manders, E. K., J. G. Tilles, and A. S. Huang. 1972. Interferon-mediated inhibition of virion-directed transcription. Virology 49:573781. 89. Marcus, P. I., D. L. Englehardt, J. M. Hunt, and M. J. Sekellick. 1971. Interferon action: inhibition of vesicular stomatitis virus RNA synthesis induced by virion-bound polymerase. Science 174:593-598. 90. Marcus, P. I., and J. M. Salb. 1966. Molecular basis of interferon action: inhibition of viral RNA translation. Virology 30:502-516. 91. Marcus, P. I., and M. J. Sekellick. 1976. Cell killing by viruses. III. The interferon system and inhibition of cell killing by vesicular stomatitis virus. Virology 69:378-393. 92. Marcus, P. I., T. M. Terry, and S. Levine. .1975. Interferon action II. Membrane-bound alkaline ribonuclease activity in chick embryo cells manifesting interferon-mediated interference. Proc. Natl. Acad. Sci. U.S.A. 72:182-186. 93. Metz, D. H. 1975. The mechanism of action of interferon. Cell 6:429-439. 94. Metz, D. H., and M. Esteban. 1972. Interferon inhibits viral protein synthesis in L-cells infected with vaccinia virus. Nature (London) 238:385-388. 95. Metz, D. H., M. Esteban, and G. Danielescu. 1975. The effect of interferon on the formation of virus polyribosomes in L cells infected with vaccinia virus. J. Gen. Virol. 27:197209. 96. Metz, D. H., M. J. Levin, and M. N. Oxman. 1976. Mechanism of interferon action: further evidence for transcription as the primary site of action in simian virus 40 infection. J. Gen. Virol. 32:227-240.

566

FRIEDMAN

BACTERIOL. REV.

97. Moehring, T. J., J. M. Moehring, and W. R. 110. Ramseur, J. M., and R. M. Friedman. 1976. Stinebring. 1971. Responses of interferonInterferon inhibits induction of endogenous treated cells to diphtheria toxin. Infect. Immurine leukemia virus production. Virology mun. 4:742-752. 73:553-556. 98. Mullin, B. R., S. M. Aloj, P. H. Fishman, G. 111. Repik, P., A. Flamand, and D. H. L. Bishop. Lee, L. D. Kohn, and R. 0. Brady. 1976. 1974. Effect of interferon upon the primary Cholera toxin interactions with thyrotropin and secondary transcription of vesicular storeceptors on thyroid plasma membranes. matitis and influenza viruses. J. Virol. Proc. Natl. Acad. Sci. U.S.A. 73:1679-1683. 14:1169-1178. 99. Mullin, B. R., P. H. Fishman, G. Lee, S. M. 112. Revel, M., D. Bash, and F. H. Ruddle. 1976. Aloj, F. D. Ledley, R. J. Winand, L. D. Antibodies to cell-surface component coded Kohn, and R. 0. Brady. 1976. Thyrotropinby human chromosome 21 inhibit action of interferon. Nature (London) 260:139-141. ganglioside interactions and their relationship to the structure and function of thyro- 112a.Riley, F. L., and H. B. Levy. 1977. Studies on tropin receptors. Proc. Natl. Acad. Sci. the mechanism of interferon action: characU.S.A. 73:842-846. terization of polysome-associated cellular 100. Nebert, D. W., and R. M. Friedman. 1973. RNAs modified by interferon. Virology Stimulation of aryl hydrocarbon hydroxylase 76:459-463. induction in cell cultures by interferon. J. 113. Robb, J. A., and R. G. Martin. 1972. Genetic Virol. 11:193-197. analysis of simian virus 40. III. Characteri101. Oxman, M. N., S. Baron, P. H. Black, K. K. zation of a temperature-sensitive mutant blocked at an early stage of productive infecTakemoto, K. Habel, and W. P. Rowe. 1967. tion in monkey cells. J. Virol. 9:956-968. The effect of interferon on SV40 T-antigen production in SV40-transformed cells. Virol- 114. Roberts, W. K., M. J. Clemens, and I. M. Kerr. 1976. Interferon-induced inhibition of proogy 32:122-127. tein synthesis in L-cell extract. An ATP de102. Oxman, M. N., and P. H. Black. 1966. Inhibition of SV40 T-antigen formation by interpendent step in the activation of an inhibitor feron. Proc. Natl. Acad. Sci. U.S.A. 55:1133by double-stranded RNA. Proc. Natl. Acad. Sci. U.S.A. 73:3136-3140. 1140. 103. Oxman, M. N., and M. J. Levin. 1971. Inter- 114a.Roberts, W. K., A. Hovanessian, R. E. Brown, M. J. Clemens, and I. M. Kerr. 1976. Interferon and transcription of early virus-specific feron-mediated protein kinase and low-moRNA in cells infected with simian virus 40. lecular-weight inhibition of protein syntheProc. Natl. Acad. Sci. U.S.A. 68:299-302. 104. Oxman, M. N., M. J. Levin, and A. M. Lewis, sis. Nature (London) 264:477-480. Jr. 1974. Control of simian virus 40 gene 115. Sambrook, J., H. Westphal, P. R. Srinivansan, and R. Dulbecco. 1968. The integrated state expression in adenovirus-simian virus 40 hyof viral DNA in SV40 transformed cells. brid viruses. Synthesis of hybrid adenovirus Proc. Natl. Acad. Sci. U.S.A. 60:1288-1295. 2-simian virus 40 RNA molecules in cells infected with nondefective adenovirus 2-sim- 115a.Samuel, C. E. 1976. Mechanism of interferon action. Studies on the mechanism of interian virus 40 hybrid virus. J. Virol. 13:322feron-mediated inhibition of reovirus mes330. senger RNA translation in cell-free protein 105. Oxman, M. N., W. P. Rowe, and P. H. Black. synthesis systems from mouse ascites tumor 1967. Studies of adenovirus-SV40 hybrid vicells. Virology 75:166-176. rus. VI. Differential effect of interferon on SV40 and adenovirus T antigen formation in 116. Samuel, C. E., and W. K. Joklik. 1974. A protein synthesizing system from interferoncells infected with SV40 virus, adenoviruses, treated cells that discriminates between celand adenovirus-SV40 hybrid viruses. Proc. lular and viral messenger RNAs. Virology Natl. Acad. Sci. U.S.A. 57:941-948. 58:476-491. 106. Paucker, K., K. Cantell, and W. Henle. 1962. Quantitative studies on viral interference in 117. Sarma, P. S., G. Shiu, S. Baron, and R. J. Huebner. 1969. Inhibitory effect of interferon suspended L cells. III. Effect of interfering in murine sarcoma and leukemia virus infecviruses and interferon on the growth rate of tion in vitro. Nature (London) 223:845-846. cells. Virology 17:324-334. 107. Paucker, K., B. J. Dalton, C. A. Ogburn, and 118. Sela, I., S. E. Grossberg, J. J. Sedmak, and A. H. Mehler. 1976. Discharge of aminoacylE. Torma. 1975. Multiple active sites on huviral RNA by a factor from interferonman interferons. Proc. Natl. Acad. Sci. treated cells. Science 194:527-529. U.S.A. 72:4587-4591. 108. Pitha, P. A., W. P. Rowe, and M. N. Oxman. 119. Sen, G. C., S. L. Gupta, G. E. Brown, B. Lebleu, M. A. Rebello, and P. Lengyel. 1976. 1976. Effect of interferon on exogenous, enInterferon treatment of Ehrlich ascites tudogenous and chronic murine leukemia virus mor cells: effect on exogenous mRNA transinfection. Virology 70: 324-338. lation and tRNA inactivation in the cell ex109. Radke, K. L., C. Colby, J. R. Kates, H. M. tract. J. Virol. 17:191-203. Krider, and D. M. Prescott. 1974. Establishment and maintenance of the interferon-in- 119a.Sen, G. E., B. Lebleu, G. E. Brown, M. Kawakita, E. Slattery, and P. Lengyel. 1976. duced antiviral state: studies in enucleated Interferon, double-stranded RNA and cells. J. Virol. 13:623-630.

VOL. 41, 1977

ANTIVIRAL ACTIVITY OF INTERFERONS

567

mation of mouse cell line 3T3. Proc. Natl. mRNA degradation. Nature (London) Acad. Sci. U.S.A. 54: 752-756. 264:370-373. 120. Sen, G. C., B. Lebleu, G. E. Brown, M. A. 132. Tovey, M., D. Brouty-Boy6, and I. Gresser. 1975. Early effect of interferon on mouse leuRebello, Y. Furuichi, M. Morgan, A. J. kemia cells cultivated in a chemostat. Proc. Shatkin, and P. Lengyel. 1975. Inhibition of reovirus messenger RNA methylation in exNatl. Acad. Sci. U.S.A. 72:2265-2269. tracts of interferon-treated Ehrlich ascites 132a.Van Griensven, L. J. D. L., M. F. Baudelaire, J. Peries, R. Emanoil-Ravicovitch, and M. tumor cells. Biochem. Biophys. Res. Commun. 65:427-434. Boiron. 1971. Studies on the biosynthesis of 120a.Sen, G. C., S. Shalia, B. Lebleu, G. E. Brown, murine leukemia virus RNA. I. Some properties of Rauscher leukemia virus isolated R. C. Derosiers, and P. Lengyel. 1977. Imfrom interferon treated JLSV5 cells, p. 145pairment of reovirus mRNA methylation in extracts of interferon-treated Ehrlich ascites 153. In Lepetit Colloquia on Biology and tumor cells: further characterization of the Medicine, Biology of Oncogenic Viruses, vol. phenomenon. J. Virol. 21:69-83. 2. North-Holland, Amsterdam. 121. Shapiro, S. Z., M. Strand, and A. Billiau. 1977. 133. Vengris, V. E., F. H. Reynolds, Jr., M. D. Synthesis and cleavage processing of oncorHollenberg, and P. M. Pitha. 1976. Internavirus proteins during interferon inhibition feron action: role of membrane gangliosides. of virus particle release. Infect. Immun. Virology 72:486-493. 16:742-747. 134. Vengris, V. E., B. D. Stollar, and P. M. Pitha. 121a.Shafritz, D. A., and W. F. Anderson. 1970. 1975. Interferon externalization by producFactor dependent binding of methionyling cell before induction of antiviral state. tRNAs to reticulocyte ribosomes. Nature Virology 65:410-417. (London) 227t918-920. 135. Wiebe, M. E., and W. K. Joklik. 1975. The 122. Sheaff, E. T., and R. B. Stewart. 1969. Interacmechanism of inhibition of reovirus replication of interferon with cells: induction of antion by interferon. Virology 66:229-240. tiviral activity. Can. J. Microbiol. 15:941- 136. Wong, P. K. Y., and R. Macleod. 1975. Studies 953. on the budding process of a temperature123. Singer, R. H., and Penman, S. 1972. Stability sensitive mutant of murine leukemia virus of Hela cell mRNA in actinomycin. Nature with a scanning electron microscope. J. Vi(London) 240:100-102. rol. 16:434-442. 124. Sonnabend, J. A., and R. M. Friedman. 1973. 137. Wong, P. K. Y., and J. A. McCarter. 1974. Mechanisms of interferon action, p. 201-239. Studies on two temperature-sensitive muIn N. Finter (ed.), Interferons and interferon tants of Moloney murine leukemia virus. Viinducers. North-Holland/American Elsevier, rology 58:396-408. New York. 138. Wong, P. K. Y., P. H. Yuen, R. MacLeod, E. 125. Sonnabend, J. A., E. M. Martin, E. Mecs, and H. Chang, M. W. Myers, and R. M. FriedK. H. Fantes. 1967. The effect of interferon man. 1977. The effect of interferon on de novo on the synthesis and activity of an RNA poinfection of Moloney leukemia virus. Cell lymerase isolated from chick cells infected 10:245-252. with Semliki forest virus. J. Gen. Virol. 139. Wu, A. M., A. Schultz, and R. C. Gallo. 1976. 1:41-48. Synthesis of type C virus particles from mu126. Stewart, W. E., II, E. DeClercq, and P. Derine-cultured cells induced by iododeoxyuriSomer. 1972. Recovery of cell-bound interdine. V. Effect of interferon and its interacferon. J. Virol. 10:707-712. tions with dexamethasone. J. Virol. 19:108127. Stewart W. E., II, E. DeClerq, A. Billiau, J. 117. Desmyter, and P. DeSomer. Increased sus- 140. Yamamoto, K., N. Yamaguchi, and K. Oda. ceptibility of cells treated with interferon to 1975. Mechanism of interferon-induced inhithe toxicity of polyriboinosinic -polyribocytibition of early simian virus 40 (SV40) funcdylic acid. Proc. Natl. Acad. Sci. U.S.A. tions. Virology 68:58-70. 69:1851-1854. 141. Yamazaki, S., and R. R. Wagner. 1970. Action 128. Tan, Y. H., E. L. Schneider, J. Tischfield, C. J. of interferon: kinetics and differential effects Epstein, and F. H. Ruddle. 1974. Human on viral functions. J. Virol. 6:421-429. chromosome 21 dosage: effect on the expres- 142. Young, C. S. H., C. R. Pringle, and E. A. C. sion of the interferon induced antiviral state. Follett. 1975. Action of interferon in enucleScience 186:61-63. ated cells. J. Virol. 15:428-430. 129. Tan, Y. H., J. Tischfield, and F. H. Ruddle. 143. Zilberstein, A. A., D. Dudok, H. Baressi, and 1973. The linkage of genes for the human M. Revel. 1976. Control of messenger RNA interferon-induced antiviral protein and indotranslation by minor species of leucyl-transphenol oxidase-B traits to chromosome G-21. fer RNA in extracts from interferon-treated J. Exp. Med. 137:317-330. L-cells. J. Mol. Biol. 108:43-54. 130. Taylor, J. 1964. Inhibition of interferon action 144. Ziberstein, A., P. Federman, L. Shulman, and by actinomycin. Biochem. Biophys. Res. M. Revel. 1976. Specific phosphorylation in Commun. 14:447-451. vitro of a protein associated with ribosomes 131. Todaro, G. J., and S. Baron. 1965. The role of of interferon-treated mouse L-cell. FEBS interferon in the inhibition of SV40 transforLett. 68:119-124.