ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, Aug. 1985, p. 252-258

Vol. 28, No. 2

0066-4804/85/080252-07$02.00/0 Copyright C 1985, American Society for Microbiology

Antiviral Activity of 2'-Deoxy-2'-Fluoro-3-D-Arabinofuranosyl-5lodocytosine Against Human Cytomegalovirus in Human Skin Fibroblasts JOSEPH M. COLACINOt* AND CARLOS LOPEZ Memorial Sloan-Kettering Institute for Cancer Research, New York, New York 10021 Received 20 June 1984/Accepted 1 May 1985

2'-Deoxy-2'-fluoro-3-D-arabinofuranosyl-5-iodocytosine (FIAC) was shown to be a selective anti-human cytomegalovirus agent in vitro with a 50% antiviral effective dose of 0.6 ,uM (J. M. Colacino and C. Lopez, Antimicrob. Agents Chemother. 26:505-508, 1983) and a 50% cell growth inhibitory dose of 8 ,M. Antiviral activity was more readily reversed with 10-fold excess thymidine, whereby the 50% effective dose was increased to 11.3 ,uM. FIAC-induced cytotoxicity was more readily reversed with 10-fold excess of deoxycytidine, whereby the 50% inhibitory dose was increased to greater than 100 ,M. Thymidine was unable to reverse completely the antiviral activity of FIAC. Although, the extent of phosphorylation of thymidine, deoxycytidine, and deoxyuridine was 6-, 4-, and 4-fold greater, respectively, in human cytomegalovirus-infected cell lysates than in uninfected cell lysates, the extent of phosphorylation of FIAC was only 1.3-fold greater in human cytomegalovirus-infected cell lysates than in uninfected cell lysates. By comparison, the extent of FIAC phosphorylation was 500 times greater in herpes simplex virus type 1-infected cells than in uninfected cell lysates. Methotrexate was 400 times more effective against human cytomegalovirus replication than it was against herpes simplex virus type 1 replication, indicating that thymidylate synthetase may be important for human cytomegalovirus replication. However, 10 ,uM FIAC did not inhibit thymidylate synthetase activity in uninfected or virus-infected cells as determined by their metabolism of [6-3H]deoxyuridine in the presence or absence of drug. FIAC at 1 ,uM suppresses and FIAC at 10 ,uM completely inhibits human cytomegalovirus DNA replication as indicated by Southern blot analysis. This inhibition was reversible. FIAC incorporation into the DNA of human cytomegalovirus strain AD169-infected cells was stimulated relative to that in nondividing, uninfected cells.

Most effective antiviral agents useful in the treatment of herpesvirus infections depend, at least in part, on their selective phosphorylation by a thymidine kinase (TK) that is virus specific and has an altered substrate specificity. Herpes simplex virus type 1 (HSV-1) TK phosphorylates antiviral compounds such as bromovinyldeoxyuridine (3), acycloguanosine (10), and 2'-deoxy-2'-fluoro-p-D-arabinofuranosyl-5-iodocytosine (FIAC) (13) much more efficiently than does cellular TK. Studies with HSV-1 mutants lacking TK (TK-), showed that these strains were much less susceptible to the antiviral activity of these compounds than were the wild-type HSV-1 viruses (14, 18). In contrast to most other herpesviruses, human cytomegalovirus (HCMV) does not code for a virus-specific TK. The elevated pyrimidine nucleoside kinase activities found in HCMV-infected cells were cell specific (9, 23). We have reported, however, that FIAC, an anti-HSV-1 agent which derives its selectivity, in part, from its phosphorylation by HSV-1 TK, was also effective and selective in inhibiting HCMV replication in cell culture (6). We have also shown the lack of selective drug phosphorylation in extracts derived from HCMV-infected human fibroblast cells (6). To investigate further the mechanism of action of this drug against HCMV, we have conducted additional studies; the results are presented in this report.

MATERIALS AND METHODS

Deoxycytidine (dCyd), thymidine (dThd), NaF, lithium chloride, creatine phosphate, creatine phosphokinase, Trishydrochloride, ATP, tetrahydrouridine, and bovine serum albumin were purchased from Sigma Chemical Co., St. Louis, Mo. Dithiothreitol was purchased from Eastman Kodak Co., Rochester, N.Y. MgCl2 was purchased from J. T. Baker Chemical Co., Phillipsburg, N.J. Radiolabeled nucleosides were purchased from New England Nuclear Corp., Boston, Mass. Unlabeled and 2-'4C-labeled FIAC (203.8 ,ug/ml, 2.038 ,uCi/ml) were prepared with appropriate structural verification and assayed for radiochemical purity by J. J. Fox and K. A. Watanabe of this institute (17, 21). Methotrexate was provided by F. M. Sirotnak of this institute. Plaque reduction assays. In determining the extent of reversal of the antiviral activity of FIAC by dCyd or dThd, a plaque reduction assay was performed as reported previously (6). FIAC was incorporated into the overlay at various concentrations with no added nucleoside or an equimolar or 10-fold excess concentration of dCyd or dThd. When distinct viral plaques were observed in the drug-free controls, cells were washed with phosphate-buffered saline, fixed with methanol, and stained with Giemsa. Viral plaques were counted, and the 50% effective drug dose was calculated as reported previously (6). Tests were carried out in duplicate, and each test was performed three times. The reversal of FIAC-induced cytotoxicity by dCyd and

Corresponding author. t Present address: Department of Microbiology and Immunology, Louisiana State University Medical Center, Shreveport, LA 71130. *

252

VOL. 28, 1985

dThd was determined with a trypan blue dye exclusion assay as reported previously (6). FIAC at 100, 10, 1, 0.1, 0.01, or 0.001 ,uM with no added nucleoside or an equimolar or 10-fold excess concentration of dThd or dCyd was incorporated into the growth medium. Preparation of cell cytosols. Uninfected and HCMVinfected cell cytosols were prepared for the evaluation of pyrimidine kinase activity by the procedure of Kit et al. (12) as used previously (6). Briefly, uninfected cells were approximately 80 to 90% confluent at 72 h after mock infection, when they were harvested for cytosol preparation. HCMV strain AD169-infected cells (multiplicity of infection, 1) were harvested 72 h after infection to obtain optimal levels of induced pyrimidine kinase activity (4, 9, 23). HSV-1 strain 2931-infected cell cytosols w.ere obtained in the same manner with virus-infected cells (multiplicity of infection, 1) harvested 12 h after virus inoculation. Determinations of the protein content of cell cytosols were carried out by the method of Bradford (2) with bovine serum albumin as the standard. Deoxypyrimidine nucleoside kinase assays. All phosphorylation reactions contained 50 mM Tris-hydrochloride (pH 8.0), 10 mM ATP, 25 mM NaF, 0.3 mM creatine phosphate, 10 U of creatine phosphokinase per ml, 16 mM dithiothreitol, 16 mM MgCl2, and 0.05 mg of bovine albumin per ml. Reaction mixtures contained 6.7 ,uCi of [methyl-3H]dThd (20.0 Ci/mmol), [6-3H]deoxyuridine (dUrd) (20.8 Ci/mmol), or [5-3H]dCyd (27.7 Ci/mmol) plus cold substrate to yield a substrate concentration of 1.5 to 2 ,uM. The extent of FIAC phosphorylation was evaluated by using [2-14C]FIAC (13 mCi/mmol), adjusted to 1.6 ,uM, and 2 mM tetrahydrouridine to prevent the deamination of FIAC. Phosphates made from radiolabeled substrates were measured by DE-81 anionexchange chromatography. The linearity of all phosphorylation reactions was established over 60 min by monitoring the reaction every 15 min. Role of thymidylate synthetase in mediating antiviral activity of FIAC. To study the role of thymidylate synthetase in HCMV replication, plaque reduction assays were performed with methotrexate. The effects of FIAC on thymidylate synthetase were studied by examining the in vitro metabolism of [6-3H]dUrd in uninfected and virus-infected cells by the method of Cheng et al. (5). Flasks (150 cm2) were seeded with human fibroblasts and were mock infected or infected with either HCMV (AD169) or HSV-1 (2931) at a multiplicity of infection of 1. Mock-infected cells were approximately 60 to 70% confluent at the start of the experiment. HCMV-infected cells were used 60 h after infection, and HSV-1-infected cells were used 12 h after infection. At the start of the experiment, monolayers were rinsed free of growth medium and replenished with 4.5 ml of growth medium containing 0 or 10 puM FIAC and incubated on a rocker at 37°C for 0.5, 1.5, 2.5, 3.5, or 4.5 h. At these time points, 0.5 ml of medium containing 0 or 10 ,uM FIAC plus 5 ,uCi of [6-3H]dUrd (1 Ci/ml, 20 Ci/mmol) was added to the monolayers, and the cells were incubated for an additional 15 min. The cells were then rinsed twice with phosphate-buffered saline and trypsinized from the flask. Cell pellets were washed twice with phosphate-buffered saline and suspended in 0.5 N perchloric acid (PCA) and kept on ice for 10 min. After two cycles of freezing and thawing, PCA-insoluble material was separated from soluble material by microcentrifugation at 4°C in an Eppendorf microcentrifuge. The PCA-insoluble material was suspended in 1 ml of 0.5 N PCA and assayed for radioactivity by scintillation spectometry with Biofluor (New England Nu-

ANTI-HCMV ACTIVITY OF FIAC

253

clear). A 0.5-ml sample of the soluble material was also assayed for radioactivity. Samples of the PCA-soluble fraction collected at 2.5 h were neutralized with 3 M KOH and applied to a polyethyleneimine cellulose thin-layer chromatography plate (Macherey-Nagle, Duren, Germany) to determine the amount of radioactivity associated with mono-, di-, and triphosphates. The chromatogram was developed in 5 M lithium chloride-88% formic acid-water (1:1:8, vol/vol/vol) for 18 cm. The plate was then cut into 1-cm2 sections, and each section was assayed for radioactivity by scintillation counting with Omnifluor (New England Nuclear) in toluene (4 g/liter). The Rfs for dUMP, dUDP, and TTP markers were 0.82, 0.39, and 0.09, respectively. Southern blot analysis. The effects of FIAC on HCMV DNA replication were studied by Southern blot analysis. Flasks (25 cm2) of confluent human fibroblasts were either mock infected or infected with HCMV AD169 at a multiplicity of infection of 1. Virus was allowed to adsorb to the monolayer for 2 h, and then the monolayers were rinsed with phosphate-buffered saline and refed with fresh Dulbecco modified Eagle medium plus 10% fetal calf serum containing 1 or 10 ,uM FIAC. At the indicated times, the monolayers were rinsed twice with phosphate-buffered saline and treated with 1 ml of lysis solution consisting of lysis buffer (0.1 M NaCl, 0.01 M Trizma-base [pH 7.3], 0.01 M EDTA disodium salt, 0.5% sodium dodecyl sulfate), 1.0 mg of proteinase K (EM Biochemicals), and 0.3 ml of 20% sodium dodecyl sulfate. The cells were then allowed to lyse overnight. The cell lysate was extracted twice with 1 ml of TE (10 mM Tris-hydrochloride, 1 mM EDTA, pH 8.0)-saturated phenol and 1 ml of chloroform-isoamyl alcohol (24:1). The extraction was mixed by inversion and centrifuged at 2,500 rpm for 5 min to separate the phases. The upper aqueous layer was carefully removed and extracted twice with 1 volume of anhydrous ether. The ether phase was separated from the bottom aqueous phase by centrifugation at 2,500 rpm for 5 min and discarded. Residual ether was blown off by passing a gentle current of air over the aqueous phase. DNA in the remaining aqueous phase was precipitated by the addition of 0.1 volume of 3 M sodium acetate and 2 volumes of ice-cold 95% ethanol. The precipitated DNA was then centrifuged at 5,000 rpm at 4°C. The entire supernatant was removed, and the remaining pellet was dried in a vacuum for 10 min. The DNA was suspended in 0.3 ml of TE and allowed to stand at room temperature for 4 h to ensure the uniform solubilization of the entire pellet. DNA prepared as above was studied by Southern blot analysis. Samples (10 to 15 p.l) of extracted DNA from uninfected or HCMV AD169-infected cells were restricted for 2 h with BamHI (Bethesda Research Laboratories, Rockville, Md.) at 37°C according to the manufacturer's directions. The restriction reaction mix contained 0.1 p.l of restriction enzyme, 2.0 p.1 of 1Ox restriction buffer (1 mM dithiothreitol, 10 mM MgCl2, 50 mM Tris-hydrochloride [pH 7.5]), and 100 mM NaCl and sterile glass-distilled water to bring the volume to 20 RI. The restricted DNA was electrophoresed in 0.5% agarose in lx TAE (0.04 M Tris-acetate, 0.002 M EDTA). The gel was allowed to run approximately 600 V-h and was then visualized by staining with ethidium bromide and observing under UV illumination. The gel was then submerged in denaturing solution (1.5 M NaCl, 0.5 M NaOH) and allowed to stand in neutralizing buffer (1 M Trizma-base [pH 7.3], 1.5 M NaCl) for 45 min. The electrophoresed DNA was transferred to a nitrocel-

254

COLACINO AND LOPEZ

ANTIMICROB. AGENTS CHEMOTHER.

lulose filter (Schleicher & Schuell Co., Keene, N.H.) by the method of Southern (19). After blotting, the filter was washed with 2x SSC (lx SSC is 0.15 M NaCl plus 0.015 M sodium citrate) and baked for 3 h at 80°C. The filter was prehybridized for 1 h at 68°C in 6x SSC-lx Denhardt solution-100 ,ug of salmon sperm DNA per ml. The filter was then hybridized overnight at 68°C with 106 cpm (as determined by Cerenkov counting) of XbaI-digested HCMV DNA (a gift from Adam Geballe, Stanford University), which was nick translated with [32P]dCTP (New England Nuclear). Hybridization was carried out in 6x SSC-lx Denhardt solution-100 ,ug of salmon sperm DNA per ml. The nitrocellulose filter was then oven -dried at 80°C for 10 min and autoradiographed for 10 days at room temperature with Kodax X-AR film. Incorporation of [3H]dThd and [2-14C]FIAC into total cellular DNA. In some experiments, 25-cm2 flasks of uninfected or HCMV AD169-infected cells were radiolabeled for 4-h pulses at the indicated times. The radiolabel consisted of either [methyl-3H]dThd (853,00 cpm per 25-cm2 flask) or [2-14C]FIAC containing 2 mM tetrahydrouridine (53,800 cpm per 25-cm2 flask). The radiolabel was added in 3 ml of growth medium per flask. DNA was extracted and prepared exactly as for Southern analysis.

RESULTS Effects of natural nucleosides on the biological activities of FIAC. dCyd was more effective in reversing the cytotoxic effects of FIAC than was dThd, but dThd was relatively more effective in suppressing antiviral activity. The concentration of FIAC required to inhibit cell replication by 50% in the presence of lOx dCyd or lOx dThd was greater than 100 or 46 ,uM, respectively, compared with 7.6 ,uM in the absence of any added nucleoside. The concentration of FIAC required to inhibit HCMV plaque formation by 50% in the presence of lOx dCyd or lOx dThd was 1.5 or 11.3 ,uM, respectively, compared with 0.6 p.M in the absence of any added nucleoside (Table 1). In light of the studies of De Clercq et al. (8), who showed that the antiviral activity of some antiviral analogs against HSV-1 can be reversed completely with dThd, the reversal of anti-HCMV activity by dThd was not impressive. Phosphorylation studies. We were able to show, as were others (4, 9, 23), that HCMV infection of human fibroblasts resulted in the stimulation of deoxypyrimidine nucleoside kinase activity relative to that in uninfected cells. The extent of phosphorylation of dThd, dCyd, and dUrd was six-, four-, TABLE 1. Effect of natural nucleosides on the biological activity of FIAC Overlaya

FIAC FIAC plus FIAC plus FIAC plus FIAC plus

1x dCyd

1Ox dCyd 1x dThd

lOx dThd

Cytotoxicityb

(I±M)

Antiviral effectc (jj.M)

8 60 >100 23 46

0.6 0.9 1.5 2.1 11.3

1X indicates an equimolar concentration of dCyd or dThd relative to concentration of FIAC at each point in the dose-response curve. 1Ox indicates a 10-fold greater concentration of dCyd or dThd relative to concentration of FIAC at each point in the dose-response curve. b Concentration of FIAC required to inhibit cell growth by 50%. c Concentration of FIAC required to inhibit HCMV plaque formation by 50%. a

TABLE 2. Phosphorylation of dThd, dCyd, dUrd, and FIAC by cell lysates of uninfected, HCMV-infected, and HSV-1-infected human fibroblasts Phosphorylation (nmollmin per mg of protein) Lysate

Uninfected HCMV infected HSV-1 infected

dThd

dCyd

dUrd

FIAC

0.010 0.060 0.49

0.009 0.04 0.05

0.008 0.03 0.31

0.0017 0.0022 0.89

and fourfold greater, respectively, in HCMV-infected cell lysates than in uninfected cell lysates (Table 2). The increased extent of dUrd phosphorylation in HCMV-infected cell lysates was of interest. In lieu of a virus-specific TK, the increased phosphorylation of dUrd by induced cellular enzymes may be important for HCMV replication by supplying substrate for thymidylate synthetase and thereby providing a de novo source of TTP for incorporation into DNA. The extent of phosphorylation of these nucleosides by HSV-1infected cell lysates was included for comparison (Table 2). The extent of phosphorylation of FIAC was only slightly greater in HCMV-infected cell lysates (0.005 nmol, 45 min) than in uninfected cell lysates (0.004 nmol, 45 min). This represented about a 1.3-fold increase. By comparison, the extent of phosphorylation of FIAC in HSV-1-infected cell lysates was 2.0 nmol in 45 min, a 500-fold increase over the extent of FIAC phosphorylation in uninfected cells. This indicates that the anti-HCMV effect of FIAC was not determined at the level of drug phosphorylation. The antiviral and cytotoxic effects of FIAC could also be a result of the inhibition of TK activity. Therefore we examined the effect of FIAC on TK activity in uninfected, HCMV-infected, and, for comparison, HSV-1-infected cell lysates. Figure 1 reveals that only HSV-1 TK activity was significantly inhibited (40 to 50%) by 100 ,uM FIAC. No significant inhibition of TK activity was observed in either uninfected or HCMV-infected cell lysates. Thymidylate synthetase studies. To study the role of thymidylate synthetase in the replication of HCMV, plaque reduction assays were carried out with methotrexate, a folate antagonist able to block specifically folate-dependent enzymes such as thymidylate synthetase (16). The 50% effective dose of methotrexate against HCMV AD169 was 0.0018 ,uM, whereas the 50% effective dose of this compound against HSV-1 was 0.72 ,uM. Methotrexate was 400 times more effective against the TK- HCMV than it was against the TK+ HSV-1, indicating that thymidylate synthetase may be important for HCMV replication. De Clercq et al. (7) have proposed that thymidylate synthetase may be a target enzyme for some antiviral nucleoside analogs, especially the 5-substituted 2'-dUrd analogs. Therefore, studies were carried out to determine whether FIAC has a specific effect on this enzyme. The uptake of dUrd by uninfected and virus-infected cells and the conversion of this nucleoside to the nucleotide (dUMP) was not inhibited significantly in any of the cells treated with 10 p.M FIAC. PCA-soluble fractions from cells treated with 10 ,uM FIAC for 2.5 h were examined for radiolabeled nucleotide content by thin-layer chromatography. In eucaryotic cells, dUTP is rapidly hydrolyzed to dUMP by dUTP nucleotidylhydrolase. The activity of dUTP nucleotidylhydrolase minimizes the incorporation of dUTP into DNA and also provides the cell with dUMP, which serves as the substrate for thymidylate synthetase. In addition, HSV-1 infection of permnissive cells results in the

VOL. 28, 1985 z

ANTI-HCMV ACTIVITY OF FIAC

1.5

255

TABLE 3. Intracellular radiolabeled nucleotides in uninfected,

HCMV-infected, and HSV-1-infected human fibroblasts untreated or treated with 10 VtM FIAC and labeled with [6-3H]dUrd

0

%

ye(>M)

FIAC concn

Lysate

0~

Mono-

phosphate

1.0

Uninfected

0 10 0 10 0 10

HCMV infected

UR

HSV-1 infected

32.0 32.1 44.2 44.6 55.0 49.0

Radioactivity as: Diphos- Triphosphate

phate

16.0 17.1 18.5 19.0 13.3 15.9

15.0 20.8 17.0 21.5 10.8 8.9

IL

0~

mock infected or infected with HCMV AD169 were incubated for 1, 2, or 3 days in the presence or absence of 1 or 10 ,uM FIAC. HCMV-specific DNA was detected by hybridizing nitrocellulose-immobilized, infected cell DNA to denatured, 32P-labeled, XbaI-digested HCMV DNA. In the absence of FIAC, HCMV sequences increased in abundance from days 1 to 3 (Fig. 3, lanes 2, 5, and 8). In the presence of

0 z

15

30

(m/) kinase

45

TIME

a.

FIG. 1. Inhibition of thymidine activity byl100 pM FIAC. Kinase activity in cell lysates was evaluated with 100 p.M FIAC in the reaction mixture. Symbols: (0) 0 p.M FIAC, uninfected human skin fibroblast lysates; (0) 100 p.M FIAC, uninfected human skin fibroblast lysates; (U) 0 p.M FIAC, HCMV AD169-infected human skin fibroblast lysates; (LII) 100 p.M FIAC, HCMV AD169-infected human skin fibroblast lysates; (A) 0 p.M FIAC, IJS V-i 2931-infected human skin fibroblast lysates; (A) 100 p.M FIAC, HSV-1 2931infected human skin fibroblast lysates.

4 3 2 r-

b. induction of high levels of dUTP nucleotidyihydrolase, which is specific for virus-infected cells (22). Therefore, we assumed that in PCA-soluble fractions from cells labeled with [6_ H]dUrd, radioactivity associated with triphosphates represents TTP formed from dUMP through the action of thymidylate synthetase. A specific inhibition of thymidylate synthetase activity in uninfected or virus-infected cells treated with FIAC would have resulted in a significant decrease in the amount of radioactivity associated with triphosphates and a significant increase in the amount of dUMP and dUDP in the PCA-soluble fractions, as is the case when virus-infected cells are treated with 0.1 VLM 2'fiuorodeoxyuridine (5). FIAC did not alter the amounts of radiolabeled dUMP, dUDP, or dTTP contained in PCAsoluble fractions relative to those amounts in cells not treated with this drug (Table 3). Nucleoside analogs that specifically inhibit thymidylate synthetase inhibit the incorporation of dUrd into DNA (8). Our results (Fig. 2) demonstrate that 10 p.M FIAC, a concentration inhibitory to both cells and virus, did not prevent the uptake of dUrd into DNA as indicated by radioactivity in PCA-insoluble fractions. This was consistent with results obtained with PCA-soluble fractions. 2'Fluorodeoxyuridine, a compound that specifically interferes with thymidylate synthetase activity, has been shown to prevent completely the incorporation of dUrd into the DNA of HSV-1-infected cells at a concentration as low as 0.1 p.M (5). Inhibition of HCMV DNA replication by FIAC. To evaluate the effects of FIAC on HCMV DNA replication, Southern blot analysis was employed. Human fibroblasts that were

Un,ntected

1

a

3

4

3

4

3

4

HSV-1 Infected

b

4

X

3

0Q_ 2

U-

C. l 10

1

2 HCMV Infected

8 6 4 2 I

2

TIME (HR.) FIG. 2. Inhibition of [6-3H]dUrd incorporation into PCAinsoluble fractions of uninfected and virus-infected human skin fibroblasts. Cells were incubated in the presence or absence of FIAC and labeled with [6-3H]dUrd for 30 min at the indicated time points and digested in ice-cold 0.5 N PCA. Panels: (a) 50% confluent (growing) uninfected human skin fibroblasts; (b) HSV-1 2931infected human skin fibroblasts beginning 12 h after infection; (c) HCMV AD169-infected human skin fibroblasts beginning 48 h after infection. Symbols: (0) 0 ,uM FIAC; (0) 10 puM FIAC.

256

COLACINO AND LOPEZ

1 ,uM FIAC at days 1, 2, and 3 (lanes 3, 6, and 9, respectively), HCMV sequences are detectable, but in much lower abundance relative to those in the drug-free control. At day 3 and 10 P,M FIAC, HCMV DNA is barely detectable, indicating an almost complete inhibition of viral DNA replication at this concentration of drug (lane 10). In addition, a dose-response effect was observed (compare lanes 8, 9, and 10). The inhibition of HCMV DNA replication appeared to be reversible. HCMV-infected cells were incubated in 10 ,uM FIAC for 1 day, rinsed free of drug, and incubated for an additional 2 days in drug-free medium. Figure 3, lane 11, demonstrates the hybridization of 32P-labeled HCMV DNA to DNA extracted from these cells. Clearly, HCMV DNA sequences are present, indicating the resumption of viral DNA synthesis after drug removal. Incorporation of [2-'4C]FIAC. The incorporation of [14C]FIAC into the DNA of nondividing uninfected and HCMV-infected cells was also studied. Human fibroblasts were synchronized by allowing them to grow to confluency in a 25-ml flask and incubating them in maintenance medium (Dulbecco modified Eagle medium plus 2% fetal calf serum) at 34°C for 48 h. The cells were then mock infected or infected with HCMV AD169 at a multiplicity of 1. At 24-h intervals after infection, the cells were pulse-labeled for 4 h with either [methyl-3HIdThd or [2-14C]FIAC with 2 mM tetrahydrouridine. DNA from these cells was then extracted and prepared exactly as for Southern blot analysis. Equal samples of DNA preparations were electrophoresed and visualized with ethidium bromide and UV illumination to

FIG. 3. Inhibition of HCMV DNA replication as indicated by Southern blot analysis of HCMV-infected cell DNA. Flasks (25 cm2) of confluent, mock-infected, or HCMV AD169-infected (multiplicity, 1) human skin fibroblasts were incubated in the presence or absence of FIAC. At days 1, 2, and 3, DNA was extracted, digested with BamHI, electrophoresed in 0.5% agarose, and transferred to nitrocellulose. The transferred DNA was then hybridized to 32p_ labeled, XbaI-digested HCMV DNA in 6x SSC. Lanes: 0, lambda phage DNA digested with EcoRI; 1, uninfected cell DNA; 2, 5, and 8, no FIAC, HCMV-infected cell DNA extracted on days 1, 2, and 3, respectively; 3, 6, and 9, 1 F±M FIAC, HCMV-infected cell DNA extracted from cells on days 1, 2, and 3, respectively; 4, 7, and 10, 10 ,uM FIAC, HCMV-infected cell DNA extracted on days 1, 2, and 3, respectively; 11, HCMV-infected cell DNA extracted from cells incubated in 10 ,uM FIAC for 1 day and in drug-free medium for an additional 2 days. Numbers to the right indicate the positions of fragments of EcoRI-digested lambda phage DNA (kilobase pairs).

ANTIMICROB. AGENTS CHEMOTHER.

30r

c

0

,L

0.0

20k

0

ox

E a: CL I0

iok

c

30

.2 0 o-

°- 20 x o._ E c

< U-

6

0

10

-I

1

2

--I

3

4

5

Days FIG. 4. Incorporation of [2-14C]FIAC into uninfected and HCMV AD169-infected human skin fibroblasts. Confluent and synchronized cells were mock infected or infected with HCMV AD169 (multiplicity, 1). Cells were then labeled for 4 h with [methyl3H]dThd (853,00 cpm per 25-cm2 flask) or [2-14C]FIAC containing 2 mM tetrahydrouridine (53,800 cpm per 25-cm2 flask). DNA was extracted once with one volume of phenol, once with chloroformisoamyl alcohol (24:1), and once with ether and precipitated with ice-cold 95% ethanol and sodium acetate. DNA pellets were dried in vacuo and reconstituted in 0.3 ml of TE, and 0.05 ml was assayed for radioactivity. The counts per minute were multiplied by 6 to obtain total radioactivity in the DNA pellet; 10,000 cpm of [3H]dThd represents approximately 1% of the total input radiolabel, whereas 300 cpm of [14C]FIAC represents approximately 0.6% of the total input radiolabel. Due to the low levels of [14C]FIAC incorporation, samples of DNA labeled with [14C]FIAC were counted for 100 min, and the background (25 cpm) was subtracted. Symbols: (0) uninfected human skin fibroblasts; (0) HCMV AD169-infected human skin fibroblasts.

estimate DNA content. Samples of DNA preparations were then monitored for radioactive content by liquid scintillation counting in 10 ml of Biofluor (New England Nuclear). [14C]FIAC-labeled DNA samples were counted for 100 min because of the low levels of incorporation in both mockinfected and virus-infected cells. The results of these experiments are given in Fig. 4. Incorporation of radioactivity from radiolabeled FIAC into the DNA of cells infected with HCMV appeared to be stimulated. At day 1, incorporation of dThd and radiolabel from FIAC was stimulated in virus-infected cells, perhaps due to the initial stimulation of cellular DNA synthesis induced by HCMV infection in stationary-phase cells (20). After day 1, incorporation of both dThd and FIAC decreased and remained relatively steady until day 5. In HCMVinfected cells, uptake of radiolabel began to increase at day 3 continuing to day 5; incorporation of dThd and radioactivity from FIAC was moderately stimulated in HCMVinfected cells relative to uninfected cells. DISCUSSION We have shown that FIAC-induced cytotoxicity can be reversed with 10-fold excess dCyd and that the antiviral activity of this drug is more readily reversed with dThd than

VOL. 28, 1985

with dCyd. This suggests that FIAC may be metabolized differently in virus-infected and uninfected cells. However, 10-fold excess dThd was not able to reverse completely the antiviral effect of FIAC, indicating that the antiviral effect of this drug cannot be due solely to its ability to act as a dThd analog. These data expand those presented in our earlier report (6). Even though the extent of dThd, dCyd, and dUrd phosphorylation was four- to sixfold greater in HCMV-infected cell lysates relative to that in uninfected cell, lysates, the extent of FIAC phosphorylation was only increased marginally. By comparison, the extent of FIAC phosphorylation in HSV-1-infected cell lysates was 500-fold greater than that in uninfected cell lysates. Consequently, the selective activity of FIAC against HCMV is most likely not determined at the level of drug phosphorylation. Studies with FIAC in the phosphorylation reaction mixtures showed that only HSV-1 TK was inhibited to a significant degree by 100 FLM FIAC, whereas TK activity in uninfected and HCMV-infected cell lysates was not. These results indicate that the anti-HCMV activity of FIAC is probably not based on the ability of this compound to act as a selective inhibitor or alternate substrate at the deoxypyrimidine nucleoside kinase level. Studies with methotrexate demonstrated that thymidylate synthetase may be important in the replication of HCMV, since this compound was 400 times more effective in inhibiting HCMV plaque formation than in inhibiting HSV-1 plaque formation. However, examination of the effect of antiviral concentrations of FIAC on this enzyme by studying the metabolism of [6-3H]dUrd in virus-infected cells failed to demonstrate a significant inhibitory effect. This is consistent with our observation that, whereas dThd was better able to reverse antiviral activity than was dCyd, it did not do so readily. The antiviral activity of compounds that specifically inhibit thymidylate synthetase has been readily reversed by levels of dThd that are lower than that of the antiviral compound itself (8). Southern blot analysis revealed that HCMV DNA replication was inhibited by FIAC. Treatment of HCMV-infected cells with 1 ,uM FIAC resulted in a decrease in viral DNA replication. FIAC at 10 FiM resulted in the nearly complete inhibition of viral DNA replication. The inhibition of HCMV DNA replication appeared to be reversible, since viral DNA synthesis resumed after 10 FtM FIAC was removed and replaced with drug-free medium. This observation confirms those of Mar et al. (15), who used DNA-cRNA hybridization to show that the FIAC-induced inhibition of HCMV AD169 DNA replication was reversible. The reversible nature of FIAC inhibition of viral DNA synthesis has important implications for the therapeutic utility of this compound in patients. HCMV infections may be refractory to short-term, intermittent FIAC dosing schedules. Finally, the studies presented here indicate that in nondividing cells infected with HCMV, FIAC incorporation is stimulated relative to that in uninfected cells. Since we used quiescent cells, the stimulated incorporation of FIAC into total cellular DNA in HCMV-infected cells most likely was a result of alterations in cellular metabolism due to ongoing viral replication. This indicates that the increased incorporation of radioactivity from FIAC in virus-infected cells was due to the replication of viral DNA, but we have not determined whether FIAC or its metabolites was incorporated selectively into viral DNA. Since HCMV infection of human fibroblast cells results in the expression of a virus-specific DNA polymerase (11), the

ANTI-HCMV ACTIVITY OF FIAC

257

increased incorporation of FIAC into the DNA of HCMVinfected cells may be a result of a greater affinity of the viral DNA polymerase for FIAC triphosphate, as was shown for HSV-1-specific DNA polymerase (1). Although the extent of phosphorylation is not significantly greater in HCMVinfected cells, the viral DNA polymerase may use available FIAC triphosphate more efficiently as an alternate substrate for incorporation into DNA or may be more susceptible to analog inhibition than is the cellular enzyme. This would provide a basis for the selective action of FIAC against HCMV. ACKNOWLEDGMENTS We thank Edward Mocarski and Richard Spaete (Laboratory of Medical Microbiology, Stanford University) for providing their laboratory, supplies, and instruction in carrying out the Southern analysis. This work was supported in part by Public Health Service grants CA-08748 and CA-18856 from the National Cancer Institute. C.L. is the recipient of American Cancer Society faculty research award no. 193. LITERATURE CITED 1. Allaudeen, H. S., J. Descamps, R. K. Sehagal, and J. J. Fox. 1982. Selective inhibition of DNA replication in herpes simplex virus infected cells by 1-(2'-deoxy-2'fluoro-,3-D-arabinofuranosyl)-5-iodocytosine. J. Biol. Chem. 257:11879-11882. 2. Bradford, M. M. 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Bioch. 72:242-254. 3. Cheng, Y.-C., G. Dutschman, J. J. Fox, K. A. Watanabe, and H. Machida. 1981. Differential activity of potential antiviral nucleoside analogs on herpes simplex virus-induced and human cellular thymidine kinases. Antimicrob. Agents Chemother. 20:420-423. 4. Cheng, Y.-C., P. J. Hoffmann, M. Ostrander, S. Grill, S. Caradino, J. Tsou, M. R. Gallagher, and T. D. Flanagan. 1979. Properties of Herpesvirus specific thymidine kinase, DNA polymerase, and Dnase and their implication in the development of specific anti-herpes agents. Adv. Opthalmol. 38:173-186. 5. Cheng, Y.-C., K. Nakayama, and S. P. Grill. 1983. Roles of thymidylate synthetase activity in Herpes Simplex-infected HeLa cells. Biochem. Pharmacol. 32:1407-1410. 6. Colacino, J. M., andC. Lopez. 1983. Efficacy and selectivity of some nucleoside analogs as anti-Human cytomegalovirus agents. Antimicrob. Agents Chemother. 26:505-508. 7. De Clercq, E., J. Balzarini, P. J. Barr, A. S. Jones, G. Verhelst, and R. T. Walker. 1981. Thymidylate synthetase as a target enzyme for the inhibitory activity of 5-substituted 2'deoxyuridines on mouse leukemia L1210 cell growth. Mol. Pharmacol. 19:321-330. 8. De Clercq, E., J. Descamps, G.-F. Huang, and P. F. Torrence. 1978. 5 nitro-2'-deoxyridine and 5-nitro-2'deoxyuridine5'monophosphate: antiviral activity and inhibition of thymidylate synthetase in vivo. Mol. Pharmacol. 14:422-430. 9. Estes, J. E., and E.-S. Huang. 1977. Stimulation of cellular thymidine kinase by human cytomegalovirus. J. Virol. 24:13-21. 10. Furman, P. A., P. De Miranda, M. H. St. Clair, and G. B. Elion. 1981. Metabolism of acyclovir in virus-infected and uninfected cells. Antimicrob. Agents Chemother. 20:518-524. 11. Huang, E.-S. 1975. Human cytomegalovirus. III. Virus-induced DNA polymerase. J. Virol. 16:298-310. 12. Kit, S., W. C. Leung, D. Trkula, and G. Jorgensen. 1974. Gel electrophoresis and isoelectric focusing of mitochondrial and viral induced thymnidine kinases. Int. J. Cancer 13:203-218. 13. Kreis, W., L. Damin, J. Colacino, and C. Lopez. 1982. In vitro metabolism of 1-4-D-arabinofuranosylcytosine in normal and herpes simplex type 1 virus infected cells. Biochem. Pharmacol. 31:761-773. 14. Lopez, C., K. A. Watanabe, and J. J. Fox. 1980. 2'-Fluoro-5iodoaracytosine, a potent and selective anti-herpesvirus agent.

258

COLACINO AND LOPEZ

Antimicrob. Agents Chemother. 17:803-806. 15. Mar, E.-C., P. C. Patel, Y.-C. Cheng, J. J. Fox, K. A. Watanabe, and E.-S. Huang. 1984. Effects of certain nucleoside analogues on human cytomegalovirus replication in vitro. J. Gen. Virol. 65:47-53. 16. McBurney, M. W., and G. F. Whitmore. 1975. Mechanism of growth inhibition by methotrexate. Cancer Res. 35:586-590. 17. Reichman, U., K. A. Watanabe, and J. J. Fox. 1975. A practical synthesis of 2-deoxy-2fluoro-D-arabinofuranose derivatives. Carbohydr. Res. 42:233-244. 18. Smee, D. F., J. C. Marin, J. P. Verheyden, and R. R. Matthews. 1983. Anti-herpesvirus activity of the acyclic nucleoside 9-(1,3dihydroxy-2-propoxymethyl) guanine. Antimicrob. Agents Chemother. 23:676-682. 19. Southern, E. M. 1975. Detection of specific sequences among

ANTIMICROB. AGENTS CHEMOTHER.

20. 21.

22. 23.

DNA fragments separated by gel electrophoresis. J. Mol. Biol. 98:503-517. St. Jeor, S., T. B. Albrecht, F. D. Funk, and F. Rapp. 1974. Stimulation of cellular DNA synthesis by human cytomegalovirus. J. Virol. 13:353-362. Watanabe, K. A., U. Reichman, K. Hirota, C. Lopez, and J. J. Fox. 1979. Nucleosides 110. Synthesis and antiherpes activity of some 2'-fluoro-2'-deoxy-arabinofuranosyl-pyrimidine nucleosides. J. Med. Chem. 22:21-24. Wohlrab, F., and B. Francke. 1980. Deoxypyrimidine triphosphatase activity specific for cells infected with herpes simplex virus type 1. Proc. Natl. Acad. Sci. U.S.A. 77:1872-1876. Zavada, V., V. Erban, D. Rezacova, and V. Vonka. 1976. Thymidine kinase in cytomegalovirus infected cells. Arch. Virol. 52:333-339.