Inhibitors of Animal Virus Replication

CHAPTER 8

Inhibitors of Animal Virus Replication

Yasushi Watanabe

I. Introduction 237 II. Compounds Affecting the Early Steps of Virus Replication 238

A. Amantadine 238 B. Guanidine and 2-(a-Hydroxybenzyl)benzimidazole 241

III. Compounds Affecting the Late Steps of Virus Replication 244 A. Isatin-j8-thiosemicarbazone and Related Compounds 244

B. Rifampicin 246 IV. Inhibitors of Virus-Specific RNA Synthesis 250

A. Actinomycin D 250 B. Inhibitors of RNA-Dependent R N A Polymerase 257

V. Inhibitors of Viral Protein Synthesis: Cycloheximide, Streptovitacin A,

and Puromycin 258 A. Inhibition of Protein Synthesis 258

B. Use of Protein Synthesis Inhibitors in Virus Research 259

VI. Inhibitors of Viral D N A Synthesis 262

A. 5-Halogen-2/-deoxyuridine 262

B. 5-Trifluoromethyl-2'-deoxyuridine 266 C. ljft-D-Arabinofuranosylcytosine 266

References 267

I. INTRODUCTION

There have been two substantial rewards derived from the past decade's intensive search for antiviral compounds having therapeutic potential. One has been the successful chemotherapy of viral keratitis with halogenated pyrimidine derivatives (see Section VI,A,1) and the other the confirmation of the usefulness of methisason in the prophylaxis and treatment of smallpox infection (see Section ΙΙΙ,Α,Ι). But apart from these successes there have been a variety of other potentially chemotherapeutic agents that were investigated and subsequently dis­

carded because of their ineffectiveness in man.

237

Among these, amantadine (see Section II,A,3) and certain biguanide compounds were studied extensively in hope of providing a drug against influenza virus infection, as vaccination against it had been most ineffec- tive. For a similar reason, rhinovirus has been regarded as another im- portant target virus in recent research (see Section III,A,3). Unfortu- nately, no compound has yet proved useful in the chemotherapy of these virus infections in man.

However, the knowledge accumulated in the hunt for virus inhibitors helped to expand our fundamental understanding of the viral infection process in animals as well as in cultured cells. Moreover, the use of these inhibitors has greatly facilitated studies of the molecular biology of virus replication, as a result of which the existence of many virus- specific events distinguishable from cellular metabolism has been eluci- dated. It has become possible to utilize such molecular biological infor- mation in designing new screening systems for selective virus inhibitors.

Some of the recent work along this line concerns a search for an inhibitor of the RNA-dependent R N A polymerase induced by the R N A viruses

(see Section IV,B) and for an inhibitor of the RNA-dependent D N A polymerase of R N A tumor viruses (see Section III,B,4).

The present chapter is not intended to give a complete bibliography or a chemotherapeutic evaluation of the reported potential antiviral agents. Rather, an attempt will be made to discuss representative com- pounds in which the antiviral mechanism of action has been extensively studied in tissue culture. Some of these compounds are regarded as "spe- cific inhibitors^,, of virus replication. The others act nonspecifically on both virus replication and cellular metabolism but are nevertheless useful in the studies of the molecular biology of virus replication.

Emphasis will be placed on the mechanism of action of the inhibitors and on their utilization in the elucidation of the virus infection process in tissue culture. It is hoped that the references selected will provide an introduction to the field.

II. COMPOUNDS AFFECTING THE EARLY STEPS OF VIRUS REPLICATION

A. Amantadine

Amantadine (I) (1-adamantanamine hydrochloride), a symmetric pri- mary amine with an unusual structure, has been under investigation

for the treatment of influenza in man (1).

8. INHIBITORS OF ANIMAL VIRUS REPLICATION 239

Amantadine · HC1 (I)

1. ANTIVIRAL ACTIVITY

Amantadine inhibits the multiplication of influenza A, A^a, and A² in tissue culture and in ovo [2-6). Several other viruses, e.g., Sendai (3), pseudorabies (3), rubella (5), and avian sarcoma viruses (7), are also inhibited by amantadine, but the inhibition is not as consistent as with influenza A group viruses. The compound does not inhibit the replication of influenza Β viruses, Newcastle disease virus, or mumps virus (2-6).

Antiviral activity against the infection with influenza A viruses has also been demonstrated in mice (2, 4, 6, 8, 9).

2. MECHANISM OF ANTIVIRAL ACTION

The primary site of amantadine action is not known. The compound does not inactivate the virus per se (2, 10, 11), nor does it inhibit the adsorption of virus to cells (2, 10, 11). Strong inhibition of virus produc­

tion was observed only when amantadine was administered prior to viral infection. If added later than 10 minutes after infection, the inhibitory effect was virtually absent (2, 10, 11). These data suggest that the inhibitory action of amantadine takes place during the very early stage of viral infection, presumably during the penetration or uncoating of the virus.

a. Inhibition of Penetration. Two reports have indicated that the com­

pound may block or slow down the penetration of influenza virus into cells (2, 10). This conclusion was derived from the following observa­

tions, (a) Virus particles adsorbed to cells in the presence of amantadine remained susceptible to inactivation by antibody for several hours with­

out being taken up into the cells, and (b) virus particles on the cell surface could also be detected by hemadsorption even after several hours of incubation at 37°C. B y contrast, the virus adsorbed in the absence of amantadine rapidly became undetectable.

Amantadine has also been shown to inhibit the general phagocytic

activity of macrophages of rabbit or mouse (12). The mode of inhibition of phagocytosis by amantadine shares many characteristics with the inhibition of virus uptake by the drug (12). It may be that amantadine primarily affects the pinocytotic activity of host cells, thereby blocking the uptake of adsorbed virus.

b. Inhibition of Uncoating. Kato and Eggers, using a fowl-plague- virus-chick-embryo system, reported that amantadine has only a slight effect on virus penetration but that it markedly inhibits virus uncoating (11). In their experiments, the extent of virus penetration was deter- mined by the decrease in susceptibility to neutralizing antibody, and uncoating was measured by the loss of photosensitivity in the virus labeled with neutral red. Since their method of determining penetration was essentially the same as that employed by other groups (2, 10), the reason for the discrepancy is unknown.

The antiviral effect of amantadine appears, however, to be mediated through reaction with the host cells. Possibly this reaction acts to inhibit the "enzyme(s)" in cell membrane involved in the initial steps of virus infection (10, 12). Such "enzyme (s)" might act on a cell membrane function involved in both penetration and uncoating of virus. Thus, if the uncoating process in the pinosomes is inhibited by amantadine (11), this effect could be mediated through an initial reaction of amantadine with the "enzyme (s)" required for virus penetration. Alternatively, the compound may inhibit simultaneously two different enzyme systems, one involved in penetration and one in uncoating.

3 . PROPHYLAXIS TRIALS IN M A N

In 1 9 6 3 , it was reported that amantadine was effective in man in preventing experimental infection with an attenuated strain of influenza A² virus, as judged by seroconversion and incidence and severity of clinical illness (13). Similar results in experimental and naturally occur- ring influenza A² infection have since been reported by several investiga- tors. Others were unable to detect any protective effect with amantadine

(see review 14). Recent trials carried out during epidemics of the Hong Kong and other influenza A² strains indicate that, in most cases, amantadine has some protective effects (15,16).

4 . RELATED COMPOUNDS

Ammonium ions and various amines have been shown to be similar to amantadine in antiviral activity (17, 18). They inhibit the growth

8. INHIBITORS OF ANIMAL VIRUS REPLICATION 2 4 1

of influenza A virus in tissue culture and in ovo. They do not inactivate virus and do not inhibit neuraminidase or adsorption of virus to cells.

As with amantadine, it has been suggested that penetration is the sus­

ceptible step in antiviral activity (18). However, these compounds show no antiviral activity in mice (18). Possibly, unlike amantadine, they cannot reach the site of infection in the body without undergoing changes Rimantadine (α-methyl-1-adamantanmethylamine hydrochloride) has been shown to have a virustatic effect against influenza A virus in tissue culture (19, 20), in experimental animals (19, 20), and in therapeutic trials against influenza A2 infection in man (21). The compound has been reported to have higher antiviral activity in tissue culture and broader chemotherapeutic index in animals than amantadine (19).

Strains of influenza A² virus with increased resistance to the compound have been isolated from tissue culture or from mice during passage of the virus in the presence of the compound (19, 22). These resistant strains should be valuable tools for studying the mode of action of amantadine.

B. Guanidine and 2-(a-Hydroxybenzyl) benzimidazole

1. ANTIVIRAL SPECTRA

Guanidine (II) inhibits the replication of many members of the picornavirus group in tissue culture without affecting cellular growth (23). Included in this category are poliovirus types 1, 2 , and 3 , many (18).

HN=C; NH2

NH.

Guanidine (Π)

Η HBB

(ΠΙ)

types of echovirus, Coxsackie A and Β viruses, H G P strain of rhino- virus (23, 24), and several strains of foot-and-mouth disease virus (25).

The antiviral spectrum of 2-(a-hydroxybenzyl)benzimidazole (HBB) ( I I I ) is somewhat different from that of guanidine; it does inhibit vari­

ous strains of rhinovirus (26) and lymphocytic choriomeningitis virus (27).

Guanidine (23) and H B B (28, 29) do not inhibit the replication of R N A viruses belonging to the myxovirus, reovirus, or arbovirus groups [with the exception of Semliki forest virus, which is inhibited by guani­

dine (30, 31)], nor do they inhibit the replication of D N A viruses.

2. MECHANISM OF INHIBITION OF POLIOVIRUS REPLICATION BY GUANIDINE

Although the antiviral action of guanidine appears to be similar to that of H B B , the primary site(s) of action of these drugs are not identical as judged by the slight difference in their antiviral spectra.

Further support of this view was provided by studies of mutants of poliovirus which are resistant to or dependent upon guanidine or H B B (32-36). Mutants resistant to one of the drugs were found to be not resistant to the other (32).

The mode of action of guanidine on poliovirus infection has been ex­

tensively studied, but the primary site of action is still obscure. Several possibilities have been proposed, three of which will be considered here.

a. Blockage of Formation of RNA Polymerase. Early studies using poliovirus showed that guanidine did not inactivate the virus nor block its adsorption and penetration (37, 38) but that it did inhibit the forma­

tion of viral protein (37), viral R N A (37, 39, 40), and RNA-dependent R N A polymerase (41)- The simplest interpretation of these data was that guanidine first inhibited the formation of viral R N A polymerase, and the inhibition, in turn, made synthesis of viral R N A and protein impossible. The fact that guanidine did not interfere with the cell-free viral R N A polymerase reaction (41) was an important basis of this interpretation. Lwoff (36, 42) suggested that guanidine, acting as "an allosteric effector," might interfere with assembly of active polymerase molecules from protein subunits.

b. Direct Intereference with Viral RNA Synthesis. Guanidine, when added to cells, rapidly inhibited ongoing synthesis of viral R N A (40).

8. INHIBITORS OF ANIMAL VIRUS REPLICATION 243 This inhibition of viral R N A synthesis was manifested before protein synthesis in the infected cells was substantially affected (43, 44)- These data suggested that guanidine might affect viral R N A synthesis directly rather than acting indirectly through blockage of the viral R N A poly- merase formation.

Caliguiri and Tamm (45), on the basis of their data, proposed that guanidine inhibits the initiation of new viral R N A chain synthesis but does not inhibit elongation and completion of chains whose synthesis has already started. The data presented by Noble and Levintow (46), however, do not support this hypothesis. Baltimore (47) suggested that guanidine inhibition of poliovirus R N A synthesis may be related to the fact that a newly completed viral R N A chain cannot leave the site of synthesis in the presence of the drug. At the moment, it is difficult to draw any conclusions on the mechanism of guanidine inhibition of viral R N A synthesis.

c. Involvement of Viral Structural Proteins. From genetic studies, Cooper's group concluded that the guanidine sensitivity of poliovirus is determined by the structural proteins rather than the viral R N A poly- merase (48). This conclusion was derived from the following observa- tions, (a) The locus determining, guanidine sensitivity coincided with the structural protein region of the genetic map (49), and (b) although there were two exceptions, ten independent changes in temperature sensi- tivity in the structural protein genes were accompanied by changes in guanidine sensitivity, whereas seven independent changes in nonstruc- tural protein genes were not.

3. COMPOUNDS T H A T ANTAGONIZE GUANIDINE

Many compounds have been reported which counteract guanidine in- hibition of virus replication. These are choline (50), methionine, valine, alanine, leucine (36), trimethylamine, tetramethylammonium iodide, methyl formate, methyl acetate, methanol, triethylamine, tetraethylam- monium iodide, hemicholinium HC-3, ethionine (51), dimethylethanol- amine, dimethylpropanolamine, and ethanolamine (52).

The mechanism of antagonism with any of these compounds has not been delineated. Possible competition between these compounds and guanidine appears unlikely (36, 52), as does the involvement of methyla- tion or C-l compound transfer (35, 51, 52).

III. COMPOUNDS AFFECTING THE LATE STEPS OF VIRUS REPLICATION

A. Isatin-^-thiosemicarbazone and Related Compounds

1. PRACTICAL U S E IN THE TREATMENT OF POXVIRUS INFECTION

The antiviral activity of isatin-/?-thiosemicarbazone (ITSC) and iV-methyl-ITSC (methisason or Marboran) (IV) was recognized as early

s

R

R = H , ITSC R = C H3, Methisason

(IV)

as 1 9 5 1 - 1 9 5 3 , when Thompson's group first reported on the therapeutic activity of ITSC against vaccinia virus infection in mice (53). Bauer

(54, 55) extended this work and found that N-ethyl- and N-methyl- ITSC were more active than free ITSC [for the activity-structure rela- tionship see also Sadler (56)]. Following extensive testing in animals, he pioneered in the practical use of methisason against smallpox infection in man and gave ample statistical proof supporting the prophylactic usefulness of the drug (57, 58). A similar observation was reported by Ribeiro D o Valle (59). Methisason has also proved useful in the treat- ment of the progression of vaccinia gangrenosa (60) and of eczema vaccinatum (see 58).

2 . MECHANISM OF INHIBITION OF POXVIRUS REPLICATION

Earlier work showed that ITSC inhibits multiplication of poxviruses in tissue culture at a concentration that is nontoxic for cells (61-64).

The drug does not interfere with the synthesis of vaccinia virus D N A or the formation of many of the viral antigens, nor does it inhibit the appearance of virus-induced cytopathic changes. On the basis of these observations, Easterbrook (63) and Bach and Magee (64) suggested that ITSC affects only the late stages of the vaccinia virus replication

8. INHIBITORS OF ANIMAL VIRUS REPLICATION 245 cycle. Electron microscopy revealed that only immature virions are formed in the presence of the drug (56, 63, 65).

a. Inhibition of Late Protein Formation. Unequivocal evidence that the block is related to the inhibition of "late" protein synthesis has been presented by two groups of workers. Magee and Bach (66) found that "early" viral antigens were formed in the presence of ITSC but that "late" antigens, produced after 4 hours in the normal growth cycle, failed to be induced. The "early" proteins induced by virus were func- tionally normal in that the increase in thymidine kinase and the forma- tion of viral D N A occurred normally. Furthermore, the viral D N A syn- thesized in the presence of the drug also appeared to be functionally normal; i.e., it was incorporated, after the removal of ITSC, into virions that were infectious. A similar conclusion was reported by Woodson

and Joklik (67), who showed that in the presence of ITSC viral protein synthesis proceeded at the same rate as in the absence of drug for about 3 hours but then abruptly declined (at this time viral D N A begins to be synthesized at the maximum rate). The viral D N A synthesized was not coated, presumably because of lack of necessary proteins. Indeed, it has been recently shown that one of the polypeptide components of viral core is synthesized late (68).

b. No Inhibition of mRNA Synthesis. According to Woodson and Joklik (67), the observed failure in the synthesis of "late" viral protein

is not due to a failure to synthesize "late" viral mRNA from progeny virus D N A . The nascent mRNA synthesized after 3 hours of infection in the presence of ITSC was normal with respect to both its rate of synthesis and its sedimentation coefficient and was incorporated into polyribosomes. However, its translation was interrupted because the polyribosome containing the viral m R N A was rapidly degraded in the presence of ITSC and, at the same time, the sedimentation coefficient of the m R N A decreased from 16 S to 8 S. It is unknown whether the breakage of m R N A results from an intrinsic defect in the m R N A syn- thesized in the presence of ITSC or whether the drug specifically affects the stability of polyribosomes containing viral mRNA.

Apart from the primary site of action, of interest is the fact that the manifestation of the inhibitory effect of ITSC in the infected cells was prevented by actinomycin D at a concentration that did not interfere with virus production (69). This observation could mean either that the synthesis of an R N A or protein factor that is a true inhibitor is induced by ITSC, or that ITSC must be metabolized to a true active

form by an enzyme that is induced in the cells upon the addition of the drug.

3. EFFECTS ON OTHER ANIMAL VIRUS INFECTIONS

It was once thought that the spectrum of antiviral activity of ITSC and its derivatives was limited to the poxvirus group. However, methisa- son turned out to be highly active against certain types of adenovirus in cell culture (types 3, 7, 9, 11, 16, 17, 21, 28 adenoviruses and a simian adenovirus, SV15) (70), although the treatment of adenovirus 3 infection in man with methisason was unsuccessful (71). Furthermore, ITSC, methisason, and related heterocyclic compounds have been reported to be effective against certain strains of rhinovirus (72). The dialkyl-sub- stituted derivatives of ITSC such as iV-methylisatin-/?-4', 4'-dibutylthio- semicarbazone (busatin) are active primarily against R N A viruses (56, 73, 74). The manner in which busatin inhibits poliovirus production in cell culture was studied by Pearson and Zimmerman (75), who pro- vided evidence that the drug interferes with virus-specific R N A synthesis in infected cells as well as in the cell-free poliovirus R N A polymerase system. Thus, the mode of action of these compounds against R N A viruses seems to be distinct from that reported for vaccinia virus. At the moment, busatin may be classified as an inhibitor of RNA-dependent R N A polymerase.

B. Rifampicin

Rifampicin (V) [3- (4-methylpiperazinyliminomethyl)rifamycin SV]

(76) is a semisynthetic derivative of rifamycin SV (77), which was

Rifampicin (V)

8. INHIBITORS OF ANIMAL VIRUS REPLICATION 247 isolated from the culture broth of Streptomyces mediterranei n. sp.

Rifampicin has been reported to inhibit replication of gram-positive bacteria, mycobacteria (78-80), trachoma agent (81, 82), and bacterio­

phages (83, 84). Recently, the compound has been found to prevent replication of the DNA-containing animal viruses.

1. INHIBITION OF R N A POLYMERASE IN BACTERIA

Rifampicin and rifamycin SV inhibit R N A synthesis in bacteria by binding to DNA-dependent R N A polymerase (85-89). The precise mech­

anism of the inhibition of E. coli R N A polymerase was reported by di Mauro et al. (90), who found that rifampicin binds to core enzyme of R N A polymerase (without affecting σ factor) and thereby competi­

tively inhibits the binding of the first ribonucleotide to R N A polymerase.

Mutants of E. coli resistant to rifampicin have been found to possess an altered R N A polymerase (89, 91). This observation provided strong evidence that the target of rifampicin action in bacteria is exclusively the DNA-dependent R N A polymerase.

2. INHIBITION OF ANIMAL D N A VIRUS REPLICATION

In contrast to its inhibitory action in bacterial systems, rifampicin has little effect on the activity of mammalian DNA-dependent R N A polymerase (86, 88). This suggested an interesting idea. If the R N A polymerase coded for by animal D N A viruses resembled the bacterial enzyme and differed from that of animal cells, rifampicin might selec­

tively prevent viral replication. In view of this, the first report by an Israeli group, Heller et al. (92), that rifampicin does selectively inhibit vaccinia virus (WR strain) in cell culture greatly encouraged animal virologists engaged in the chemotherapy of virus diseases.

Subak-Sharpe et al. (93) independently reported the inhibition by rifampicin of vaccinia, cowpox, and adenovirus replication. They also isolated a mutant of vaccinia virus resistant to rifampicin. The mutation frequency they determined was of the order of 10~⁷, suggesting that the drug may act on a single protein coded for by the virus genome.

Moreover, their radioautographic experiment showed that the rifampicin inhibited the incorporation of [3H ] uridine into cells infected with vac­

cinia virus. These results were taken to indicate that the rifampicin prevented vaccinia virus replication by inhibiting the virus-specific DNA-dependent R N A polymerase. However, there existed a considerable difference in the drug action between the bacterial and animal virus sys­

tems, (a) Rifamycin SV, which is as active as rifampicin in bacteria,

had no inhibitory effect on the replication of poxviruses; (b) rifampicin inhibited the replication of these viruses only when used at concentra- tions ( > 100 /xg/ml) close to the toxic level for animal cells; and (c) the drug was inhibitory even when added at late stages of the virus replication cycle.

3. MECHANISM OF INHIBITION OF VACCINIA VIRUS REPLICATION

a. No Effect on Viral mRNA Synthesis. Subsequent work by Moss, McAuslan, Becker, and their colleagues have clearly shown that virus- specific R N A synthesis is not a primary target of rifampicin action.

These groups agreed that rifampicin inhibits neither the intracellular synthesis of "early" mRNA, which is transcribed from viral D N A in parental virus core (94, 95), nor the "late" mRNA, which is transcribed after the synthesis of progeny D N A (94-96). The mRNA synthesized in vivo in the presence of 100 /xg/ml of rifampicin was normal with respect to rate of synthesis (94, 95), sedimentation coefficient (94, 96), and association with polyribosomes (96). Furthermore, rifampicin did not inhibit in vitro the DNA-dependent R N A polymerase associated with the virus core (94).

b. Little Effect on Viral DNA Replication. Viral D N A synthesis was also not greatly affected (93, 94, 97), and since the viral D N A synthe- sized in the presence of 100 /xg/ml of the drug was incorporated into mature infectious virions (94, 97) upon the drug's removal, it appeared to be normal.

c. Inhibition of Late Protein Synthesis. In contrast to viral RNA and D N A synthesis, vaccinia-virus-specific protein synthesis was de- pressed by rifampicin at a late stage of the virus growth cycle (7 hours after infection) (96-98). Some differences of opinion exist concerning the effect of rifampicin on the synthesis of late viral proteins. Ben-Ishai et al. (96) reported that rifampicin interferes with late viral protein formation in general and thus prevents virus maturation. Tan and McAuslan (98 ) showed that a virus-specific particulate R N A polymerase did not increase in the presence of rifampicin and concluded that the failure in the polymerase induction was probably not due to a general reduction in the rate of late protein synthesis but rather to failure in the formation of necessary viral structure required for the polymerase activity. This conclusion was in agreement with the recent finding by

8. INHIBITORS OF ANIMAL VIRUS REPLICATION 2 4 9

Nagayama et al. (99) that four kinds of enzyme activities associated with the viral structure were coordinately suppressed by rifampicin.

d. Block of Viral Envelope Formation. Moss et al. (97) demonstrated that all the viral proteins necessary for virus maturation were made in the presence of rifampicin even at late stage of the viral replication cycle. They found by electron microscopy that, instead, rifampicin blocks morphological maturation of vaccinia virion at a stage in the formation of viral envelope, thus causing the accumulation of immature viral mem- brane (WO, 101). A more complicated finding by Katz and Moss (102) was that the formation of a core peptide (MW 7 6 0 0 ) from a precursor

(MW 1 2 5 , 0 0 0 ) was prevented by rifampicin. This observation may be reconciled with the other data if it is assumed that the cleavage of this precursor occurs during a stage in formation of viral structure and that interruption of envelope development by rifampicin prevents subse- quent events necessary to reach this stage.

Experiments on inhibition of poxviruses with various derivatives sug- gest that the hydrazone side chain of rifampicin is essential for its anti- viral activity (108).

4. EFFECTS ON OTHER ANIMAL VIRUS INFECTIONS

Animal viruses whose replication is inhibited by rifampicin are vac- cinia virus (WR), cowpox virus, and adenovirus, while herpes, pseudo- rabies viruses, and R N A viruses such as poliovirus, echovirus 1, Coxsackie virus B 6 , influenza virus AO (NWS), Sendai virus, encephalo- myocarditis virus, vesicular stomatitis virus, and reovirus are all rifampicin resistant (93).

The effect of rifampicin on R N A tumor viruses should not be over- looked. Diggelman and Weissmann (104) reported that transformation (as manifested by focus formation) of chick fibroblasts infected with Rous sarcoma virus (RSV) was suppressed at least 1 0 times more strongly by rifampicin than was virus replication itself (the inhibition of virus replication was only 3 0 - 4 0 % at 6 0 /Ag/ml). They determined the inhibition of focus formation after exposure of the chick cell mono- layers infected with RSV to rifampicin at different time periods and concluded that there exists a rifampicin-sensitive period (between 3 6 and 6 0 hours after infection) during which a process necessary for the eventual focus formation is completed. It may be interesting to correlate this rifampicin inhibitor of focus formation to a postulated D N A tran- scription from viral R N A in the early stage of virus infection (see Sec-

tion IV,A,2). However, an RNA-dependent D N A polymerase of the R N A tumor virus was not strongly inhibited in vitro by rifampicin. Marked inhibition was observed only with a iV-demethylated derivative of rifampicin (103, 105). Further studies of the precise mechanism by which the rifampicin inhibits transformation of the cells infected with R N A tumor virus might provide a clue to the molecular events leading to transformation of animal cells.

IV. INHIBITORS OF VIRUS-SPECIFIC RNA SYNTHESIS

A. Actinomycin D

Actinomycin D (VI) binds to deoxyguanosine residues in the D N A helix, thereby inhibiting DNA-directed R N A synthesis both in intact

Actinomycin D (VI)

cells and in cell-free systems (106). Synthesis of D N A is inhibited only at higher concentrations of the drug (106, 107). Because of its interfer- ence with the D N A in its role as a template, actinomycin D inhibits expression of various functions of D N A viruses as well as their reproduc- tion (107-113).

8. INHIBITORS OF ANIMAL VIRUS REPLICATION 2 5 1

1. USEFULNESS OF ACTINOMYCIN D IN THE STUDY OF R N A VIRUS

Reich et al. (107) first reported that actinomycin D does not inhibit replication of mengovirus, an R N A virus, in L cells at 2 /xg/ml, a concen- tration that blocks cellular R N A synthesis and vaccinia virus replication.

Their finding implied that the replication of viral R N A is a process different from ordinary R N A synthesis directed by cellular or viral D N A . This contention was soon confirmed by the discovery of an RNA-de- pendent R N A polymerase (114) and of a replicative intermediate (RI) or double-stranded R N A (RF) (115).

The use of actinomycin D in R N A virus systems, therefore, allowed observation of virus-specific R N A synthesis without interference from cellular R N A synthesis in the infected cells. A successful application of this new approach was first reported by Shatkin (116) and Zimmerman et al. (117), who demonstrated that, in the presence of actinomycin D , [1 4C]uridine was preferentially incorporated into poliovirus R N A in infected HeLa cells. This technique has since been widely utilized in the studies of virus-mediated R N A synthesis in cells infected with picornaviruses, paramyxoviruses, arboviruses, and reovirus, all of which are little affected by the antibiotic [reovirus was initially reported to be sensitive (118) but turned out to be rather insensitive (119, 120)].

2 . INHIBITION OF R N A TUMOR VIRUSES

Actinomycin D has been a useful probe for determining whether a given virus utilizes information transfer from D N A to R N A in its repli- cation cycle. In this regard, R N A tumor viruses are extremely interesting.

It has been observed that replication of murine and avian R N A tumor viruses (121-129) and of avian myeloblastosis virus (130) is suppressed by relatively low concentrations of actinomycin D ( 0 . 1 - 0 . 5 /xg/ml) added during infection or to virus-producing transformed cells. Furthermore, infection and transformation of cells by these viruses can be prevented by inhibitors of D N A synthesis when added at early stages of virus infection, indicating early requirement for D N A synthesis during virus infection (122, 123, 126). These rather puzzling findings led Temin to postulate that viral R N A is first transcribed into D N A , which is then integrated into host cell chromosome and serves as template for the syn- thesis of viral R N A (121, 126, 129).

Recently, the postulated inverted transcription from R N A to D N A has been proved real by Baltimore (131) and Temin and Mizutani (132), who discovered an RNA-dependent D N A polymerase as an integral com-

ponent of murine and avian R N A tumor viruses. Similar findings have since been made in several laboratories, and the overall process of D N A synthesis now appears to contain at least two steps (133-139): (1) synthesis of short segments of D N A from viral R N A template with the formation of R N A - D N A hybrids and (2) subsequent synthesis of double-stranded D N A by a second virion enzyme, a DNA-dependent D N A polymerase.

Step 2 was found to be inhibited by a rather high concentration of actinomycin D (137). However, this inhibition is probably unrelated to the actinomycin D inhibition of virus replication in the virus-produc- ing transformed cells, since it has been shown that D N A synthesis is not required for continuous shedding of R N A tumor virus once infection is established (129). At the moment, it appears that the step sensitive to actinomycin D during virus reproduction is most likely the replication of viral R N A from a D N A intermediate, as has been proposed by Temin.

3. INHIBITION OF INFLUENZA VIRUS AND FOWL PLAGUE VIRUS

Influenza virus and fowl plague virus have a phase sensitive to actino- mycin D early in the infection cycle (140-146). This sensitivity to the drug progressively decreases during the first 1.5-2 hours of infection;

if added later than these times, the inhibition is minimal (14?, 148).

These data, together with the fact that virus production was prevented by pretreatment of the cells with ultraviolet light (140), were interpreted to mean that the transcription of host genome is required for the comple- tion of the early steps of influenza virus infection (140, 141, 146)- Borland and Mahy found that the host cell DNA-dependent R N A poly- merase is enhanced in the cells infected with influenza virus and proposed that this enzyme may be involved in the virus replication (149).

An alternative interpretation proposed by Rott et al. (142, 144) is that the actinomycin D (150-153), like mitomycin C (154), induces breakdown of R N A in the nucleus and that through this process the viral R N A may be destroyed during the early stages of infection before having established an R N A synthesizing center in the infected cell nu- cleus. In fact, it has been reported that nucleocapsid antigen and viral R N A of influenza virus are formed in the nucleus (155-157) as con- trasted to cytoplasmic multiplication of Newcastle disease virus ( N D V ) , another myxovirus, whose reproduction is little affected by actinomycin D (158, 159).

Apart from this strong inhibition during early stages, Rott et al. (144) have shown that actinomycin D gradually inhibits influenza virus R N A

8. INHIBITORS OF ANIMAL VIRUS REPLICATION 253 synthesis even when added at late stages of infection. Pons [145) found that actinomycin D has an effect on the R N A duplex of the virus ( R F ) . When the antibiotic was added at 1.5 hours of infection, the R F already made disappeared, but when added at 3.5 hours the R F already present remained. According to Nayak and Baluda (148) and Scholtissek and Rott (160), however, the R N A duplex comprises only a small fraction of the total virus-specific R N A formed in the infected cells; most of the viral R N A (plus strand) and its complementary R N A (minus strand) exist in a single-stranded form in the infected cells. The latter authors (160) found that actinomycin D preferentially inhibits the syn- thesis of minus-strand RNA. Therefore, the inhibitory effect on R F for- mation might be a consequence of the inhibition of minus-strand R N A synthesis.

Since it is known that actinomycin D does not inhibit influenza-virus- induced R N A polymerase as tested in vitro (161-163), the mechanism by which the antibiotic inhibits the minus-strand formation remains unknown. Whatever the mechanism may be, if similar inhibitory action were operative at a very early stage of infection, then the synthesis of the first minus strand from input virus R N A would be depressed, resulting perhaps in the strong inhibition of all later events.

4. OTHER R N A VIRUSES

While a direct interaction of viral R N A with actinomycin D has not been found (106, 116, 117, 164), it has been reported that the antibiotic does affect the production of most R N A viruses under certain experi- mental conditions.

Poliovirus replication in HeLa cells is unaffected (108, 117), as men- tioned before, but is inhibited in H E P - 2 cells (165) and human amniotic cells (166). Infection of HeLa cells with an extracted viral R N A is slightly enhanced or unaffected by the presence of 0.1-2.0 /xg/ml of actinomycin D , whereas infection with a double-stranded R N A (RF)<

is inhibited tenfold (167). Replication of rabies virus and vesicular stomatitis virus (members of the rhabdovirus group) is inhibited by 90-95% at concentrations over 0.1 /xg/ml of actinomycin D in B H K cells, but not in VERO cells (168). Mengovirus replication is unaffected by the antibiotic in L cells (169), but both replication and virus-induced R N A polymerase formation are inhibited in Novikoff hepatoma cells

(170). The inhibition of R N A polymerase induction has also been re- ported for foot-and-mouth disease virus ( F M D V ) in B H K cells (171).

The production of N D V in chick embryo cells is unaffected or reduced

(140, 141, 172-175), but synthesis of virus-specific RNA, both plus and minus strands, is enhanced (176-178).

The manner in which actinomycin D reduces virus production has not been delineated in any of these cases. The following possibilities should be considered. First, the inhibition of virus production by actino- mycin D is seen mainly when the drug is administered prior to, or throughout, the virus infection, but such lengthy treatment may cause a general toxic effect in the infected cells as a result of blocking cellular R N A synthesis. Second, available evidence suggests that actinomycin D may affect cells in ways unrelated to its interference with DNA-de- pendent R N A synthesis (179-183). If this is so, actinomycin D may interact with cellular materials in a complex manner, thereby inducing various changes in the metabolic state of the cells. Data supporting this notion were provided by Cooper (166) when he showed that actino- mycin D inhibition of poliovirus production could be reversed by insulin contained in the serum used for cell culture.

5. ENHANCEMENT OF VIRUS PRODUCTION BY A Low D O S E OF ACTINOMYCIN D

An occasional and varying enhancement of virus production was ob- served with poliovirus (166, 167), reovirus (184), Sendai virus (185), measles virus (186), Chikungunya virus (187, 188), and subacute scleros- ing panencephalitis (SSPE) virus (189). For measles and Chikungunya viruses, the increased virus production was observed after several cycles of virus growth in cell cultures inoculated wiih viruses at a low multi- plicity of input (186-188). In these cases, it was found that actinomycin D inhibited the synthesis of interferon in the infected cultures, thus preventing the decline in virus production that otherwise occurs following the first cycle of multiplication. Presumably SSPE virus production was

"stimulated" in a similar manner (189).

B y contrast, the enhancement of virus production with other viruses seems to be unrelated to interferon. Experiments were confined to a single cycle of infection, and in some cases a stimulation of virus produc- tion was apparent as soon as new virus became detectable at the end of eclipse period (185). The enhancement in these cases might be ac- counted for by an actinomycin-induced decay of host cell mRNA

(150-153) resulting in an increase in the number of unprogrammed ribo- somes (190) or in the pool of ribonucleotides. This interpretation may help to explain at least the observed increased synthesis of NDV-specific R N A in the presence of actinomycin D (177, 178) and also the results

8. INHIBITORS OF ANIMAL VIRUS REPLICATION 255 reported by Bukrinskaya and Zhdanov that the eclipse period of Sendai virus infection is shortened in the presence of actinomycin D (191).

6. INHIBITION OF D N A VIRUSES

Because the expression of viral genome is controlled largely at the level of transcription, actinomycin D has been useful for studying the temporal sequence of gene expression during the viral replication cycle.

a. SVIfl. Several groups have reported intriguing analyses of the SV40 infection process, especially in relation to the early events following viral infection. When African green monkey kidney (AGMK) cells are lyti- cally infected with SV40, the virus-specific tumor (T) antigen appears during the eclipse period of the infection cycle. The T-antigen formation precedes by about 10 hours the appearance of progeny D N A and virions and is the earliest detectable function of the SV40 genome. The data reported by Rapp et al. (192, 193) showed that the amount of T-antigen formed was markedly diminished when actinomycin D (1 μg/m\) was administered at the time of infection. This observation suggested that the synthesis of T-antigen requires transcription of R N A from viral D N A . On the other hand, subsequent reports by Defendi et al. (194) and Sabin (195) showed that actinomycin D at a concentration of 1 /xg/ml is insufficient to block the T-antigen formation, although this concentration is capable of blocking 85% of the R N A synthesis in in­

fected cells; a concentration of > 5 μ-g/ml was required for the inhibition of T-antigen formation. Defendi et al. (194) concluded that early transcription of viral D N A regulating the synthesis of T-antigen is con­

siderably more resistant to the effects of actinomycin D than is the transcription of viral D N A regulating the synthesis of other viral products.

Several workers have reported that the early transcription is distinct from the late with respect to the species of R N A transcribed and their templates. Rapp et al. (192) found that cytosine arabinoside, a specific inhibitor of D N A synthesis, did not affect T-antigen formation, suggest­

ing that the R N A required for the T-antigen synthesis is transcribed from parental virus genome. Aloni et al. (196) demonstrated the presence of at least two distinct classes of viral mRNA: an early R N A that is synthesized before the replication of viral D N A and a late R N A that is present after the onset of viral D N A replication. The difference between these two classes of R N A was recognized both by R N A - D N A hybridization and by base ratio analysis: 37% of the SV40 genome

is represented in the early mRNA, whereas at least 76% of the genome is represented in the late mRNA. Carp et al. (197) reported that R N A having the characteristics of early mRNA in the hybridization test was accumulated in large quantities in the SV40-infected AGMK cells in the presence of 0.5 /xg/ml of actinomycin D (a level capable of blocking the formation of progeny D N A ) . Taken together, these observations suggest that the restricted species of mRNA are transcribed from the input virus genome (the early mRNA) and that this process is rather resistant to the action of actinomycin D . At the present time there is no indication of the mechanism by which the transcription from parental genome resists the action of actinomycin D . It may be, however, that the insensitivity of the early transcription is due to an incomplete uncoating of the input viral genome, which acts to prevent the binding of actinomycin D to the viral D N A . That such partial uncoating is the case has been reported for vaccinia virus (198) and reovirus (199).

After infection, these viruses are only partially degraded, giving rise to the formation of a core particle. A small portion of the genome is then transcribed into mRNA by an R N A polymerase that resides in the core. Late mRNA formation takes place, as in SV40, only after the onset of the formation of progeny genome. For SV40, this partially un- coated structure has not been detected in infected cells, but the released genome was found to be associated with certain cellular proteins (200).

b. Poxvirus. The poxvirus genome has been known to contain the information for temporal regulation of the formation of early and late proteins which are involved in the synthesis of virus (201-208). Experi- ments in which R N A synthesis was inhibited by actinomycin D at vari- ous times during infection showed that the early induction of thymidine kinase and its subsequent "shutoff" are primarily regulated at the level of transcription (209, 210). Synthesis and translation of mRNA for the enzyme commenced within 2 hours of infection, and the translation was subsequently repressed at 6 hours after infection. For this "shutoff" to occur a new mRNA must be synthesized and translated into "repressor protein" between 4 and 6 hours after infection. Thus, the early synthesis of the mRNA for thymidine kinase, its translation into the enzyme, and the subsequent repression of the translation all appeared to be under the control of viral D N A . This regulatory mechanism is remarkably similar to that of early enzyme induction in certain phage-bacteria sys- tems (see review 211). Pseudorabies virus seems to display a similar shutoff mechanism for the synthesis of certain viral proteins (212, 213).

Perhaps a viral genome of large molecular weight is required for exerting

8. INHIBITORS OF ANIMAL VIRUS REPLICATION 257 such refined temporal control of virus reproduction, since no D N A virus with a smaller genome has been shown to contain such information.

B. Inhibitors of RNA-Dependent RNA Polymerase

It has been well established that an RNA-dependent R N A polymerase is induced in cells infected with R N A virus. Because this enzyme is specific to the R N A virus infection and is undetectable in uninfected cells, any agent that would specifically inhibit this R N A polymerase should be a good candidate as a possible antiviral agent. Recently, evi- dence that such specific inhibitors exist has been reported (see busatin in Section III,A,3).

Haruna et al. (214) employed a purified Qp R N A phage replicase system as a means of screening for an anti-RNA-virus agent. They found that 4-(2-propenyloxy)-/?-nitrostyrene and its related compounds strongly inhibited the phage R N A replicase but not the bacterial D N A - dependent R N A polymerase. Furthermore, the drug appeared to inhibit an RNA-dependent R N A polymerase isolated from the cells infected with polioviruses or R N A tumor viruses. Unfortunately, these nitro- styrene derivatives were ineffective in preventing the replication of vari- ous R N A viruses in cultured cells (215).

Gliotoxin (VII), which was recognized as an antiviral agent by

Rightsel et al. (216), has been shown to inhibit intracellular replication of poliovirus in HeLa cells at the stage of viral R N A replication (217, 218). Chemotherapy trials were limited because of gliotoxin's high toxic- ity for animals (216). Ho and Walters (219) have found that this com- pound inhibits the influenza virus-induced R N A polymerase reaction in vitro without affecting cellular DNA-dependent R N A polymerase. Other known antivirals that belong to the same group of sulfur-containing

CH2OH Gliotoxin

(vn)

diketopiperazines, such as dehydrogliotoxin, chetomin, aranotin, and acetylaranotin, appear to display similar selective inhibition on viral R N A polymerase in vitro. In view of their specific inhibition of RNA- dependent R N A synthesis (as opposed to the action of actinomycin D ) , the usefulness of these compounds as well as Haruna's compounds in the study of R N A virus replication is promising. It is unknown, however, whether the inhibition results from the interaction of these compounds with the R N A template or with the enzyme.

V. INHIBITORS OF VIRAL PROTEIN SYNTHESIS:

CYCLOHEXIMIDE, STREPTOVITACIN A, AND PUROMYCIN

A. Inhibition of Protein Synthesis

Cycloheximide (VIII) (or Actidione, 3-[2-(3,5-dimethyl]-2-oxocyclo- hexyl)-2-hydroxyethyl]glutarimide), an antibiotic produced in Strepto­

myces griseus, is a highly potent inhibitor of protein synthesis in intact animal cells, in yeast, and in extracts derived from these cells, but it has no effect on bacterial systems (220-228). Protein synthesis in intact L cells is inhibited 50% by cycloheximide at concentrations of about 0.1 ju,g/ml and almost completely at concentrations of 10-50 /xg/ml (226, 228, 229).

Streptovitacin A (VIII), a 4-(OH) derivative of cycloheximide, is about 75% as active as cycloheximide in inhibiting protein synthesis in intact L cells but is ineffective in intact yeast cells, possibly due to impermeability (226). In extracts of rat liver, streptovitacin appeared to be five times more inhibitory than cycloheximide to cell-free amino

ο

Η R = Η, Cycloheximide R = OH, Streptovitacin A

(vm)

8. INHIBITORS OF ANIMAL VIRUS REPLICATION 259 acid incorporation (226). The action of these two drugs is reversible, and protein synthesis is readily restored soon after their removal from the cell culture medium (229). These drugs only partly depress D N A synthesis and have no effect on R N A synthesis (225, 227).

The mechanism by which cycloheximide inhibits protein synthesis is related to ribosome function (230). In cells treated with low concentra- tions of the antibiotic, the polyribosomes, perhaps because of a slowdown of ribosome movement along the mRNA, pile up (281-234), whereas with higher concentrations (sufficient to totally block protein synthesis) the polyribosomes break down (284). However, conclusive evidence on the exact mode of action is still lacking.

Puromycin (IX) (6-dimethylamino-9-[3-deoxy-3- (p-methoxyl-L-ph-

N(CH³)²

Puromycin (IX)

enylalanylamino)-/?-D-ribofuranosyl]purine), an antibiotic produced by Streptomyces alboniger (235), inhibits protein synthesis in various organisms, in mammalian cells, and in extracts derived from these (236).

Structurally puromycin resembles the aminoacyladenosine portion of aminoacyl transfer R N A (237, 238) and reacts with peptidyl transfer R N A on ribosomes giving rise to peptidyl puromycin and thereby inter- rupting peptide elongation (235, 239-241).

B. Use of the Protein Synthesis Inhibitors in Virus Research

Cycloheximide (242), streptovitacin A (243-245), and puromycin (246-259) prevent reproduction of a wide variety of D N A and R N A viruses. This antiviral activity is not unexpected in view of the absolute

requirement of protein synthesis for the reproduction of viruses, i.e., the proteins constituting virion and the enzymes involved in the synthesis of virus.

1. VACCINIA VIRUS UNCOATING E N Z Y M E

The use of these inhibitors in the study of molecular events during virus reproduction has led to several interesting discoveries. For example, the requirement for the synthesis of a new enzyme in the "uncoating"

of vaccinia virus was delineated using these inhibitors. The early vac- cinia virus infection involves three defined stages: (a) adsorption to the cell with subsequent uptake of virion into phagocytic vacuoles, (b) lysis of the outer viral membrane (thereby releasing the viral core), and (c) rupture of the core, allowing liberation of the viral D N A into the cytoplasm where foci of replication become established (see reviews 198, 250). For the third stage [termed "uncoating" by Joklik (251)]

to occur, there is an absolute requirement of protein synthesis. This conclusion was derived from the observation that treatment of infected cells with cycloheximide or streptovitacin A results in the accumulation of viral cores (252, 253) which, using a DNA-dependent R N A poly- merase contained in the viral core itself (254, 255), are capable of synthesizing early mRNA (256). These observations were interpreted as indicating that early mRNA codes for an enzyme which "uncoats"

the viral core and that inhibition of enzyme synthesis by these drugs results in accumulation of the core. Thus, the virus core appears to be capable of directing its own uncoating.

2 . S V 4 0

By employing cycloheximide, Kit et al. (257) have demonstrated that SV40 D N A replication requires concomitant synthesis of protein in lyti- cally infected cells. They suggested that a specific class of host proteins is required for both cellular and viral D N A replication; these proteins could be initiator proteins as postulated by Jacob et al. (258) and/or structural components of a D N A replication factory (257). In contrast, Hirai and Defendi found that new protein synthesis is not required for SV40 D N A to integrate into cellular D N A in nonpermissive cells, a process considered to be a primary event in the transformation of cells. They suggested that a preexisting enzyme is involved in the genome integration (259; see also Section VI,C). A similar finding has been reported for adenovirus (260).

8. INHIBITORS OF ANIMAL VIRUS REPLICATION 261

3 . REOVIRUS

The expression of R N A virus genome in infected cells is completely precluded if the protein synthesis inhibitors are given soon after infec- tion; inhibition is primarily attributed to the blocking of formation of RNA-dependent R N A polymerase (261). In reovirus-infected cells, however, a small amount of limited species of viral mRNA can be made in the presence of a high dose of cycloheximide added at the time of infection (199). On the basis of this finding, Watanabe et al. (199) predicted that some m R N A is copied from the input virus genome either by a preexisting cellular polymerase or, as has been shown with poxvirus, by a polymerase carried by the virion itself. Indeed, the existence of such an enzyme in the virion has been discovered by several groups

(262-265).

When the protein inhibitors are given to infected cells in which viral R N A synthesis is actively taking place, the viral R N A is continuously formed and accumulated but is not encapsidated (266, 267). For reovirus and influenza virus, however, the synthesis of certain species of virus- specific R N A is selectively inhibited by cycloheximide. In reovirus-in- fected L cells, both mRNA and double-stranded viral R N A begin to be synthesized between 6 and 7 hours after infection. When cyclo- heximide or puromycin is added at this time or later, the viral R N A formation is quickly interrupted, whereas the mRNA synthesis proceeds at an increased rate (229, 268). This result suggests that two distinct enzymes are involved in the synthesis of viral R N A and mRNA, one having a high turnover and responsible for the replication of viral double-stranded RNA, and the other having a low turnover and responsi- ble for the mRNA transcription (229).

A transcriptase that was purified from infected cells was recently shown to be unrelated to the replication of viral R N A (265, 269, 270).

Another R N A polymerase, one that replicates viral double-stranded RNA, has also been discovered in extracts of infected cells and it appears to be distinct from the transcriptase (271, 272). It is yet unknown whether the replicase itself has a rapid rate of turnover or whether the replication requires a protein factor which is unstable.

4. INFLUENZA VIRUS

A similar selective inhibition of R N A synthesis has been reported for influenza virus, whose infection induces the synthesis of both viral R N A (plus strand) and its complementary R N A (minus strand) (see

Section IV,A,3). When cycloheximide was added to infected cells actively synthesizing these RNA's, the formation of plus strand was selectively depressed (160). Scholtissek and Rott (160) postulated, as one interpre­

tation of this phenomenon, the existence of two distinct enzymes, as may be the case in reovirus. They also offered an alternative explanation suggesting that a protein associated with a viral nucleocapsid might regulate the synthesis of plus and minus strands. If the synthesis of the protein part of the nucleocapsid is blocked by cycloheximide, the plus strand might remain functional as template, giving rise to the con­

tinued formation of minus-strand RNA.

VI. INHIBITORS OF VIRAL DNA SYNTHESIS

A. 5-Halogen-2'-deoxyuridine

1. ANTIVIRAL ACTION OF IUDR, B U D R , AND F U D R

5-Iodo-2'-deoxyuridine ( I U D R ) , 5-bromo-2'-deoxyuridine ( B U D R ) , and 5-fluoro-2'-deoxyuridine ( F U D R ) (X) are potent inhibitors of D N A

ο

HO Η R = CH3 , Thymidine R = Ι , IUDR R = Br, BUDR R = F, FUDR

(X)

virus replication in cell culture [vaccinia virus (273-277), herpes simplex virus (273, 278-282), adenovirus (283), pseudorabies virus (212, 284, 285), and SV40 (192, 193)], but they have no direct effect on R N A viruses (273, 276), with the exception of R N A tumor virus (see Section IV,A,2).

8. INHIBITORS OF ANIMAL VIRUS REPLICATION 263 The I U D R is one of several compounds being used to treat keratitis caused by herpes simplex or vaccinia viruses in rabbit and man (see review 286). This compound has also been clinically tested for herpes simplex encephalitis and congenital cytomegalovirus infection in man (287; also references cited in 288). Extensive reviews of these compounds giving their use in the treatment of viral infection and their mechanism of action have been published (286, 287, 289) and there is little that need be added here.

These pyrimidine analogs have been shown to poison host cells as well as inhibit virus replication (for example, 290) but are less toxic to cells not actively engaged in D N A synthesis. Probably, the low toxic- ity of I U D R for nondividing cells and the ease with which it reaches the site of viral infection are the basis for its successful use against viral keratitis. In view of the possibility of this kind of superficial "selec- tive effect," there has been a considerable amount of argument regarding the actual selectivity of the antiviral action of halogenated pyrimidine analogs. The present review attempts, therefore, to document the prob- lems related to the possible selectivity of the drug action. Prior to describing these problems it may be profitable to summarize the action of the analogs on D N A synthesis.

2. MECHANISM OF ACTION ON D N A SYNTHESIS AND CELL GROWTH

B y inhibiting the methylation of deoxyuridylic acid to thymidylic acid (a step in the normal pathway for the de novo synthesis of thymidylic acid), F U D R blocks D N A synthesis (291, 292). To be effec- tive F U D R must first be phosphorylated in the cell by thymidine kinase

(290); however, the phosphorylated product is not incorporated into D N A . When given to growing cells, F U D R arrests the cell replication cycle at a point before S phase (293) and seems to act lethally upon cells in S phase (294). These effects probably result from the deficiency in thymidylic acid in FUDR-treated cells, since the inhibition is reversed by exogenous thymidine (290, 293, 295). For this reason, cell lines con- taining a large pool of thymidine or its derivatives would not be effi- ciently inhibited by F U D R , nor would cell lines containing a high level of nucleoside phosphorylase, an enzyme that degrades F U D R to fluorouracil (295). The data on F U D R , up to 1965, have been extensively reviewed by Heidelberger (296).

In contrast, I U D R and B U D R apparently are substrates for many of the enzymes responsible for thymine metabolism. These analogs are

phosphorylated to mono-, di-, and triphosphates and are incorporated into D N A in place of thymine, thereby producing faulty D N A {297-299). The action of I U D R and B U D R is clearly contrasted to that of F U D R and F³T D R (XI) (described in the next section) when one adds them to a synchronized cell culture at the beginning of S phase and looks at their effect on cell division. The cells exposed to I U D R or B U D R synthesize D N A and undergo cell division, but they fail to go into a second cell cycle, whereas the cells exposed to the fluorinated derivatives do not synthesize D N A and do not undergo cell division (293). Like F U D R , I U D R and B U D R first must be phosphorylated by thymidine kinase to be active. Kit et al. demonstrated that a strain of mammalian cells lacking thymidine kinase is resistant to B U D R and I U D R (300). Furthermore, they isolated mutants of vaccinia and herpes viruses that are resistant to these analogs when grown in the cell line lacking thymidine kinase; these variant viruses appeared unable to induce their own thymidine kinase (301, 302).

In addition to faulty D N A production, the phosphorylated I U D R also exerts an allosteric feedback inhibition (mimicking the action of the corresponding natural D N A precursor, thymidine triphosphate) on deoxycytidylate deaminase (303, 304), on thymidine kinase (303), and on ribonucleoside diphosphate reductase (305), all of which are involved in the de novo synthesis of D N A precursors. The triphosphate deriva- tive of B U D R was also shown to inhibit the thymidine kinase of E.

coli in a similar manner (306).

3. POSSIBLE SELECTIVE ACTION ON D N A VIRUSES

The studies concerned with the effect of I U D R or B U D R on the multiplication of D N A viruses show that I U D R or B U D R replaces thymine residues of viral D N A to various extents [vaccinia (274, 275), pseudorabies (284)]. Thus, faulty viral D N A and viral antigen accumu- late in the infected cells and there is little assembly of normal virion

(192, 193, 213, 274, 275). Kaplan and Ben-Porat (212) have concluded, by using pseudorabies virus, that the failure in the assembly of normal virion is probably due to the faulty formation of proteins involved in virion maturation rather than a distortion of the D N A molecule per se.

In an attempt to test whether the action of I U D R is selective for viral D N A synthesis, Kaplan and Ben-Porat (285) determined the degree of incorporation of I U D R into newly synthesized D N A in infected and noninfected cells. They showed that in the viral D N A there was a considerable degree of substitution of I U D R for thymine, even when