<<

Broad-spectrum antiviral that interferes with de novo pyrimidine biosynthesis

Hans-Heinrich Hoffmanna, Andrea Kunza,b,1, Viviana A. Simona,b,c, Peter Palesea,b,2, and Megan L. Shawa,2

aDepartment of , bDepartment of Medicine, and cGlobal Health and Emerging Pathogens Institute, Mount Sinai School of Medicine, New York, NY 10029

Edited by Thomas E. Shenk, Princeton University, Princeton, NJ, and approved February 25, 2011 (received for review February 4, 2011)

Compound A3 was identified in a high-throughput screen for Favipiravir (T-705), which is currently in Phase III trials in inhibitors of influenza replication. It displays broad-spectrum Japan, also displays broad-spectrum antiviral activity, but only antiviral activity, and at noncytotoxic concentrations it is shown to among RNA . In vivo efficacy has been demonstrated for inhibit the replication of negative-sense RNA viruses (influenza influenza viruses (A, B, and C), flaviviruses (West Nile virus, viruses A and B, Newcastle disease virus, and vesicular stomatitis Yellow Fever virus), bunyaviruses (Punta Toro virus, Rift Valley virus), positive-sense RNA viruses (Sindbis virus, hepatitis C virus, fever), arenaviruses (Pichinde virus), and (foot- West Nile virus, and dengue virus), DNA viruses (vaccinia virus and and-mouth disease virus) (11, 12). T-705 is converted to the human adenovirus), and (HIV). In contrast to mamma- ribofuranosyl triphosphate form by host , and it was lian cells, inhibition of viral replication by A3 is absent in chicken shown that the antiviral activity of T-705 can be reversed by an cells, which suggests species-specific activity of A3. Correspond- excess of purines (13). Nevertheless, the mechanism of T-705 is ingly, the antiviral activity of A3 can be linked to a cellular , not fully understood. It does not affect the synthesis of cellular dihydroorotate dehydrogenase (DHODH), which is an in DNA or RNA (13, 14), and may therefore target viral RNA- the de novo pyrimidine biosynthesis pathway. Viral replication of dependent RNA directly or may be preferentially both RNA and DNA viruses can be restored in the presence of excess incorporated into viral RNA, thereby causing hypermutations. uracil, which promotes pyrimidine salvage, or excess orotic acid, Brequinar was described as a broad-spectrum antiviral agent ca- which is the product of DHODH in the de novo pyrimidine bio- pable of inhibiting both negative- and positive-strand RNA viruses fl synthesis pathway. Based on these findings, it is proposed that A3 (15), and le unomide and a derivative, FK778, have been repor- MICROBIOLOGY acts by depleting pyrimidine pools, which are crucial for efficient ted to inhibit human cytomegalovirus (HCMV) and herpes sim- virus replication. plex virus 1 (HSV-1) (16–19). This group of compounds all target the cellular enzyme dihydroorotate dehydrogenase (DHODH) (20–22), although leflunomide shows the weakest activity against mall-molecular-weight compounds with antiviral activity can this enzyme (23). DHODH is a key player in the pyrimidine de act by inhibiting viral or host cell proteins that are S novo biosynthesis pathway and converts dihydroorotate (DHO) required for virus replication. Although drugs directed at viral into orotate (24). Both the pyrimidine de novo biosynthesis proteins are more virus specific, they can easily lead to the se- pathway and the uracil salvage pathway channel into the pro- lection of resistant mutants. For example, the use of adamant- duction of uridine monophosphate (UMP) (25), which is the anes which target the M2 ion channel of influenza A viruses is precursor for all pyrimidine nucleotides needed for RNA (UTP, now precluded due to wide-spread resistance (1, 2). By targeting CTP) and DNA (dTTP, dCTP) synthesis. Leflunomide (Arava) is host cell proteins, resistance is less likely to occur, and if that a Food and Drug Administration (FDA)-approved therapy for protein is necessary for replication of a variety of different vi- rheumatoid arthritis, and its immunosuppressive properties are ruses, broad-spectrum antiviral activity can be achieved. Such related to inhibition of T-cell proliferation, which is heavily reliant compounds are presumed to be more prone to toxicity, and it is on pyrimidine pools (26). Virus replication is also dependent on important to identify targets that are not critical for cell growth large nucleotide pools, and therefore the antiviral activity of these but that are absolutely essential for the virus. compounds is likely due to pyrimidine depletion. Recently, two small molecules were reported, LJ-001 (3) and Here we report that the small-molecular-weight compound dUY11 (4), that demonstrate broad antiviral activity against all A3, which was identified in a previously reported influenza virus enveloped viruses. These so-called rigid amphiphiles resemble high-throughput screen (27), possesses broad-spectrum activity phospholipids and are incorporated into viral membranes, where against RNA-, DNA-, and retroviruses and acts by targeting they modulate the membrane curvature needed for the mem- pyrimidine metabolism. brane fusion event and they therefore inhibit . The fi virus speci city of these compounds takes advantage of struc- Results tural differences in virion membranes versus cellular membranes Antiviral Activity of Compound A3. Approximately 61,600 com- and also the lack of repair mechanisms for viral membranes. To mercial, small-molecular-weight compounds were screened in date, the only approved broad-spectrum antiviral that is effective duplicate in a high-throughput screen (HTS) assay described against both RNA and DNA viruses is ribavirin (5), which is currently used in combination with IFN for hepatitis C therapy (6). Ribavirin is a ribosyl purine analog, the carboxamide group of which can resemble adenosine or guanosine, depending on its Author contributions: H.-H.H., P.P., and M.L.S. designed research; H.-H.H. and A.K. per- formed research; A.K. and V.A.S. contributed new reagents/analytic tools; H.-H.H., A.K., rotation. As a prodrug, it is sequentially phosphorylated by cel- V.A.S., P.P., and M.L.S. analyzed data; and H.-H.H., V.A.S., and M.L.S. wrote the paper. lular kinases, and all its intermediates such as ribavirin mono- Conflict of interest statement: A provisional patent application has been filed by Mount (RMP), di-(RDP) and triphosphate (RTP) are inhibitors of certain Sinai School of Medicine covering the A3 compound. viral RNA-dependent RNA polymerases (7). RTP is incorporated This article is a PNAS Direct Submission. into RNA and pairs equally well with uracil and cytosine, inducing 1Present address: Institute of Tropical Medicine and International Health, Charité, Univer- lethal hypermutations (8), and RMP has been shown to target sity Medicine Berlin, Berlin 14050, Germany. the cellular inosine monophosphate dehydrogenase (IMPDH) and 2To whom correspondence may be addressed. E-mail: [email protected] or peter. thereby depletes intracellular pools of GTP (9, 10). This lack of [email protected]. GTP may explain the inhibition of DNA viruses, as well as the This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. general cytotoxicity of ribavirin. 1073/pnas.1101143108/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1101143108 PNAS Early Edition | 1of6 Downloaded by guest on September 28, 2021 A B 1E+8 120 100 1E+7 Fig. 1. Compound A3 and its antiviral activity against influenza A/WSN/33 virus. (A) Chemical 80 structure of compound A3 and its molecular 1E+6 IC50=0.178μM CC50=268μM weight (MW). (B) A549 cells were infected with 60 influenza A/WSN/33 virus (MOI = 0.01) in the 1E+5 ter [pfu/ml] ter presence of increasing concentrations of com- 40

cell viability [%] pound A3. Viral titers were determined at 24 h MW: 356 g/mol viral 1E+4 20 postinfection and the IC50 calculated (left-hand scale, blue curve). Mean of three replicates ± SD

1E+3 0 are shown. Cell viability (CC50) was determined 0.001 0.01 0.1 1 10 100 independently for a 24-h incubation period (right- hand scale, red curve). Mean of five replicates ± compound concentraon [μM] SD are shown.

previously (27). Briefly, a compound was defined to be a strong the , leading to the hypothesis that replication and inhibitor of influenza virus replication if the luminescence signal may be targeted. of an influenza virus inducible firefly luciferase reporter was We addressed this question by performing an influenza virus decreased by at least 90% compared with that in untreated cells. minigenome assay to determine whether influenza virus poly- Compound A3 (Fig. 1A) from the Asinex1 library displayed a merase activity is affected by A3. A549 cells were transfected very strong effect on viral replication, such that there was no with expression plasmids for the influenza virus luminescence detectable. The compound was further evaluated proteins (PB1, PB2, and PA), the (NP), and the for cytotoxity, and its CC50 in A549 cells was determined to be previously described influenza virus-specific firefly luciferase re- 268 μM over a 24-h incubation period (Fig. 1B). To confirm the porter (27). To normalize for transfection efficiency, a Renilla results of the primary screen, A3 was tested at noncytotoxic luciferase plasmid was cotransfected. Compounds were added at concentrations in viral replication assays performed at an MOI 4 h before transfection and were present for the duration of the of 0.01. A reduction in viral titers of 4 logs was detected at a assay. A3 strongly inhibits influenza virus polymerase function by concentration of 2 μM, and the IC50 over a 24-h replication 98% compared with the DMSO control, without affecting cel- period was determined as 0.178 μM (Fig. 1B). This resulted in lular gene expression as monitored by Renilla luciferase activity a selective index (SI = CC50/IC50) of 1,505, which indicates (Fig. 3A). Diphyllin was shown to inhibit influenza virus entry compound A3 to be a very strong inhibitor with very little tox- (28) and was included as a negative control, whereas ribavirin, icity. The inhibitory effect of A3 was even more pronounced in a known polymerase inhibitor of RNA viruses, was included as primary human tracheal–bronchial epithelial (HTBE) cells (Fig. a positive control. A dose–response assay indicated that influ- S1A). Viral titers were reduced by 4 logs after 24 h, and the IC50 enza virus polymerase activity is inhibited to the same degree by was determined to be 0.04 μM, which resulted in an SI of >2,380 2 μM A3 as with 100 μM ribavirin (Fig. S2), suggesting that A3 is (CC50 >100 μM). In contrast, the same experiment performed in 50 times more potent than ribavirin. To determine whether A3 is mouse embryo fibroblast (MEF) cells (Fig. S1B) reduced viral affecting viral RNA synthesis, primer extension assays were titers by less than 2 logs at a similar concentration of A3. The performed on the NA segment to observe synthesis of v-, c-, and IC50 was determined as 1 μM, which resulted in an SI of >100 mRNA. To control for cellular replication, the levels of 5S rRNA (CC50 >100 μM). were also monitored. A3 was shown to fully inhibit production of all three viral RNA species at a concentration of 2 μM (Fig. 3B) A3 Inhibits Influenza Virus Polymerase Activity. Next, we tested A3 without decreasing levels of cellular 5S rRNA. Similar results at different time points during the viral life cycle to understand were obtained for 100 μM ribavirin. whether it acts early or late in infection. First we compared its effect on replication when added pre- or postinfection. A549 A3 Displays Broad-Spectrum Antiviral Activity. A3 was shown to cells were infected at a high MOI with influenza A/WSN/33 virus inhibit a number of influenza virus strains of both H1N1 and and compound treatment started 2 h preinfection or at several H3N2 subtypes (Fig. S3); therefore, to address the question of time points postinfection. Inhibition was observed when A3 was whether A3 displays broad-spectrum antiviral activity, we ex- present preinfection or when added up to 2 h postinfection but amined the replication of viruses representing several different not later (Fig. 2A). The inhibition was further demonstrated by families (Table 1). All viruses were tested in A549 cells (unless reduced levels of viral proteins in the presence of A3 (Fig. 2B). otherwise indicated) at a concentration of 10 μM A3 or less. At These data indicated that A3 was acting at early to mid stages in this concentration, A3 does not appear to be either cytostatic or

1E+8 A B compound added [hours p.i.] Fig. 2. Inhibition of influenza A/WSN/33 DMSO -2 +1 +2 +4 mock virus replication by A3 added at different times during the viral life cycle. (A) A549 fl 1E+7 NP cells were infected with in uenza virus mL] 55 kDa A/WSN/33 (MOI = 1). Compound A3 was

[PFU/ 26 kDa M1 present in the culture medium 2 h before infection or added to the medium at in- 1E+6 dicated time points postinfection at its

viral ter ter viral CC . Viral titers were determined 24 h 17 kDa M2 10 postinfection by plaque assay. Assay was β-acn performed in triplicate; results are pre- 1E+5 sented as mean ± SD. (B) levels (NP, M1, and M2) from infections DMSO 2+1+2+4 shown in A were determined by Western compound added [hours p.i.] blot analysis using specific .

2of6 | www.pnas.org/cgi/doi/10.1073/pnas.1101143108 Hoffmann et al. Downloaded by guest on September 28, 2021 Fig. 3. A3 inhibits influenza viral polymerase ac- A FF-Luciferase Ren-Luciferase B DMSO Ribavirin [μM] A3 [μM] tivity. (A) A549 cells were transfected with protein 160 10 100 2 10 expression plasmids for influenza A/WSN/33 virus 140 polymerase subunits PB1, PB2, PA, and nucleopro- tein NP. An influenza virus-specific firefly luciferase 120 reporter and a Renilla luciferase expression plas- 100 mRNA mid were cotransfected. Transfections were per- 150 nt cRNA formed in the presence of DMSO or A3 at its CC10. 80 Ribavirin (replication inhibitor) is included as posi- 60 tive control and diphyllin (entry inhibitor) is in- [%] acvity reporter cluded as negative control. Cells were harvested 40 vRNA 24 h posttransfection, and activation of the lucif- 5S rRNA erase reporter by the viral polymerase was mea- 20 sured. DMSO control is set to 100%. Assay was 0 performed in triplicate; results are presented as DMSO Ribavirin A3 Diphyllin mean ± SD. (B) A549 cells were infected with in- fluenza A/PR/8/34 virus at an MOI of 7. Infections were performed in the presence of DMSO, ribavirin, or A3. Viral RNA was extracted 9 h postinfection and subjected to primer extension analysis to determine the levels of v-, c-, and mRNA. Host cell-derived 5S rRNA was used as an internal standard.

cytotoxic over a 48-h period (Fig. S4). For influenza A/WSN/33 Antiviral Activity of A3 Is Linked to Pyrimidine Metabolism. Anti- virus (representing the ), viral titers for infec- virals with broad-spectrum activity often affect nucleotide syn- tions performed under multicycle replication conditions (MOI = thesis or incorporation. To address the question of whether A3 0.01) are decreased by 5 logs compared with a 2-log reduction for acts in such a way, we tested its activity in the presence of dif- infections performed at an MOI of 1 (Fig. S5A). Sendai virus ferent purines and pyrimidines in plaque reduction assays. Uracil (SV52), a paramyxovirus, was inhibited by 3 logs (Fig. S5B), and was found to be the only base that is able to reverse the in- hibitory effect of A3 and to restore viral replication. Surprisingly, an even stronger effect of A3 on viral replication was found for at these concentrations, the effect was seen only in Madin-Darby vesicular stomatitis virus (), the titer of which was canine kidney (MDCK) cells and not in A549 cells (Fig. 4 A–C).

decreased by 5 logs (Fig. S5C). The positive-sense RNA viruses, Thus, A3 is able to inhibit viral replication efficiently in both cell MICROBIOLOGY Sindbis virus (togaviridae) and hepatitis C virus (flaviviridae), lines, but its effect can be more easily compensated for by uracil were both inhibited by 2 logs in the presence of A3. Surprisingly, in MDCK cells than in A549 cells. This result suggested a tissue DNA viruses were also affected by A3, and titers of the human or species specificity of A3, and therefore viral replication in the adenovirus 5 (hAd5) () were reduced by 6 logs in the presence of A3 was tested in a panel of cell lines of different presence of A3 (Fig. S5D). It was noted that viruses that depend species. A strong inhibition of viral replication was seen among on DNA synthesis, such as vaccinia virus (), adenovi- human cells (A549, 293T) and primate cells (Vero, CV1), in- rus, and HIV-1 (retroviridae), were also sensitive to A3. The dependent of the tissue origin, and only high concentrations of IC of A3 against HIV-1 (NL4-3) is in the mid nanomolar range uracil (100-fold) were able to partially restore viral replication 50 fi (205 nM, Fig. S6A), which is remarkably similar to the IC for (Fig. S7A). Vero cells, unlike CV1, are de cient in type I IFN, 50 but both cell lines are derived from the same species (Cercopi- influenza virus (178 nM; Fig. 1). Time-of-addition experiments thecus aethiops) and tissue (kidney). This suggested that the in- with A3 in conjunction with FDA-approved HIV-1 antiretroviral volvement of IFN could be excluded in the antiviral actions of drugs show that A3 remains active even when added as late as A3. This was further confirmed via quantitative RT-PCR, which 12 h after infection, whereas inhibitors of or showed no induction of IFN-β and other antiviral genes (ISG56, integration lose activity when added 4–8or10–12 h postinfection IRF7, RIG-I, and TNF-α) upon treatment with A3 alone or in (Fig. S6B). These findings suggest that the mechanism of action combination with influenza virus infection in A549 cells. In of A3 is at the step of transcription after integration but before nonprimate mammalian cells, complete inhibition of viral repli- maturation (Fig. S6B). cation by A3 was seen in both MDCK and bronchial–tracheal

Table 1. Viruses tested for their susceptibility to A3 Viruses Group Family Inhibition of Degree of inhibition

Influenza A virus (A/WSN/33) (−) ssRNA Orthomyxoviridae Virus ∼5 logs (MOI = 0.01) Influenza B virus (B/Yamagata/88) (−) ssRNA Orthomyxoviridae Virus ∼1.5 logs (MOI = 1) Newcastle disease virus (La Sota) (−) ssRNA Virus ∼2 logs (MOI = 1) Sendai virus (SV52) (−) ssRNA Paramyxoviridae Virus ∼3 logs (MOI = 1) Vesicular stomatitis virus (−) ssRNA Rhabdoviridae Virus ∼5 logs (MOI = 0.01) Sindbis virus (+) ssRNA Togaviridae Virus ∼2 logs (MOI = 1) Hepatitis C virus* (+) ssRNA Virus ∼2 logs (n.a) West Nile virus (+) ssRNA Flaviviridae Replicon cell line ∼1 logs (NA) Dengue I virus (+) ssRNA Flaviviridae Replicon cell line ∼1 logs (NA) † Vaccinia virus (NYVAC) dsDNA Poxviridae Virus ∼3 logs (MOI = 1) Adenovirus (hAd5) dsDNA Adenoviridae Virus ∼6 logs (MOI = 5) ‡ HIV-1 ,§ ssRNA RT Retroviridae Virus ∼2 logs (NA)

NA, not applicable. *Tested in Huh 7.5 cells. † Tested at 1 μMA3 ‡ Tested in TZM bl cells. §Tested at 2 μM A3.

Hoffmann et al. PNAS Early Edition | 3of6 Downloaded by guest on September 28, 2021 A 120 100 A549 80 MDCK in %

60

40 replicaon 20

0

A3

B C DMSO no Ura 10xUra 20xUra 50xUra 100xUra

1E+9 1E+9 1E+8 1E+8 ] ]

mL 1E+7 1E+7 mL 1E+6 1E+6 [PFU/ 1E+5 1E+5 1E+4 1E+4 1E+3 1E+3 viral ter ter viral viral ter [PFU/ ter viral 1E+2 1E+2 1E+1 1E+1 0hpi 12hpi 24hpi 36hpi 0hpi 12hpi 24hpi 36hpi

D 1E+9 E 1E+9 24hpi 1E+8 12hpi 1E+8

36hpi ] 1E+7 24hpi

mL] 1E+7 1E+6 1E+6 1E+5 [PFU/mL [PFU/ 1E+4 1E+5 1E+3 1E+4 viral ter ter viral viral ter ter viral 1E+2 1E+3 1E+1 A3 + A3 + A3 + A3 + A3 + A3 + A3 + A3 + A3 + A3 + A3 + A3 + A3 + A3 + A3 + A3 + 10x 33x 100x 333x 10x 33x 100x 333x 10x 33x 100x 333x 10x 33x 100x 333x

DMSO A3 DHO Oro DMSO A3 DHO Oro

Fig. 4. Inhibition of influenza virus replication by A3 is reversed by excess uracil and orotate. (A) Plaque-reduction assay with influenza A/WSN/33 virus was performed in A549 (human) and MDCK (canine) cells in the presence of DMSO, 10 μM A3 alone, or A3 and indicated purine and pyrimidine bases (10-fold excess relative to A3). Percent replication was calculated and set relative to DMSO control. (B and D) A549 (human) cells and (C and E) MDCK (canine) cells were infected with influenza A/WSN/33 virus (MOI = 0.001). (B and C) medium was supplemented with A3 (5 μM) and in addition with increasing concentrations of uracil (Ura), as indicated by fold excess relative to A3. (D and E) Cell culture medium was supplemented with A3 (10 μM) and in addition with increasing concentrations (fold excess relative to A3) of dihydroorotate (DHO) or orotate (Oro). Viral titers were determined by standard plaque assay at indicated time points.

ferret cells, but this could be easily reversed by the addition of note, attempts to generate A3-resistant influenza viruses have uracil. Replication was only partially inhibited by A3 in baby been unsuccessful, which is another indication that a host factor hamster kidney (BHK) cells and pig kidney (PK-15) cells, and may be the target of A3. the addition of uracil had no obvious effect. In contrast, A3 had The finding that uracil can reverse the inhibition of viral no effect on viral replication in chicken fibroblast (DF-1) cells replication by A3, suggests that the pyrimidine pathway may be (Fig. S7B). This further underlines the possible species specificity affected. Uracil is converted by the pyrimidine salvage pathway of A3 and suggests that A3 may target a host cell protein. Of to uridine monophosphate (UMP), the base from which all other

4of6 | www.pnas.org/cgi/doi/10.1073/pnas.1101143108 Hoffmann et al. Downloaded by guest on September 28, 2021 pyrimidines are manufactured in the cell. Besides the salvage of acceptor. After the release of orotate, ubiquinone binds to a sec- uracil, UMP is also generated via de novo biosynthesis. To dis- ond site, where it reoxidizes the cofactor. Leflunomide interferes tinguish between de novo biosynthesis and the salvage pathway, with the ubiquinone binding site, and brequinar was shown to be we tested all intermediates of the pyrimidine biosynthesis cycle a noncompetitive inhibitor versus DHO but competitive versus for their ability to restore viral replication. Only orotic acid, ubiquinone (20). Whether A3 interacts with the DHO or the which itself is converted in a two-step reaction by UMP synthase ubiquinone binding sites or whether it competes with the cofactor into UMP, was able to reverse the inhibition by A3 in both A549 binding site needs to be determined by cocrystallization. The and MDCK cells. Viral replication was tested in both cell lines in cofactor binding site seems to be a plausible target, as different the presence of A3 alone or in combination with increasing species use distinct cofactors (20). This could explain the lack of concentrations of dihydroorotate (DHO) and orotate (Fig. 4 D activity in avian cells and perhaps reduced activity in nonhuman/ and E). DHO is converted by DHODH into orotate, which can nonprimate cells, making it easier for excess uracil to reverse A3- rescue viral replication in a dose-dependent manner in contrast mediated inhibition and to restore UMP levels. Our attempts to to DHO. Although we cannot guarantee that DHO and orotate generate A3-resistant influenza viruses were most likely un- are equally efficient at entering cells or are equally stable, the successful due to the fact that a lack of nucleosides is hard to dose–response seen with orotate provides a strong indication overcome. Nevertheless, Qing et al. (15) recently reported a bre- that DHODH activity may be targeted by A3. Leflunomide is a quinar-resistant dengue virus in which a in the NS5 gene known weak inhibitor of DHODH and was tested in A549 and (RNA-dependent RNA polymerase) conferred resistance through MDCK cells at a concentration of 10 μM for its effect on in- enhancement of viral RNA synthesis. fluenza virus replication. No inhibition was observed at this In summary, the broad-spectrum antiviral activity of A3 is concentration, perhaps indicating that A3 is a superior inhibitor associated with de novo pyrimidine synthesis, and this feature of DHODH compared with leflunomide. A sequence alignment contrasts with the activity of ribavirin and favipiravir, both of of known DHODHs reveals ∼90% sequence identity between which target the purine pathway. Further studies with A3 and its the human and canine DHODH proteins but only ∼73% be- derivatives are required to investigate the in vivo efficacy of this tween the human and chicken DHODH, which supports the idea unique structural group of antivirals. of species-specific activity of A3. Material and Methods Discussion Cell Lines, Viruses, and Plasmids. A549, HEK 293T, TZM bl (HeLa), Vero, CV1, We describe the properties of a small-molecular-weight com- MDCK, BHK, PK-15, chicken fibroblast (DF1) cells, and primary human tra- pound named A3 that inhibits a broad spectrum of viruses. At cheal–bronchial epithelial (HTBE) cells were obtained from the American MICROBIOLOGY least eight different families including RNA (orthomyxoviridae, Type Culture Collection (ATCC). MEF cells were kindly provided by Benjamin paramyxoviridae, rhabdoviridae, flaviviridae, togaviridae), DNA tenOever (Mount Sinai School of Medicine, New York, NY). Bronchial tra- (poxviridae, adenoviridae), and retroviruses (retroviridae) are po- cheal ferret cells were kindly provided by Randy Albrecht (Mount Sinai School of Medicine, New York, NY). Human liver hepatoma (Huh 7.5) cells tently inhibited (Table 1). A time-of-addition experiment per- were kindly provided by Charles Rice (Rockefeller University, New York, NY). formed with influenza A virus suggested that the viral fi Vero cells expressing the replicon system of West Nile virus (WNV) and replication step is affected by A3. This was con rmed in mini- dengue I virus (DENV-I) were kindly provided by Pei Yong Shi (Novartis In- genome assays, and the primer extension assay demonstrated stitute for Tropical Diseases, Singapore). A549, HEK 293T, Huh 7.5, TZM bl, that A3 treatment resulted in complete inhibition of viral RNA Vero, CV1, DF-1, BHK, PK-15, MEF, and ferret cells were cultured in DMEM synthesis. Because a variety of different virus families were (Invitrogen) supplemented with 10% FBS (HyClone). MDCK cells were cultured inhibited by A3, it appeared likely that a cellular factor that is in Eagle’s minimum essential medium (MEM) (Invitrogen) supplemented with essential for virus replication was the target. Several compounds 10% FBS. HTBE cells were cultured in Bronchial Epithelial Cell Growth Medium described to have broad-spectrum antiviral activity act on cel- (BEGM) supplemented with the BEGM SingleQuot kit (Lonza). lular nucleotides, and here we demonstrate that the antiviral Influenza viruses A/WSN/33 (H1N1) and B/Yamagata/88 were grown and effect of A3 can be reversed specifically by uracil. Uracil is a key titered in MDCK cells. Sindbis virus (SINV) and vesicular stomatitis virus player in pyrimidine metabolism. Homeostasis of cellular UMP expressing the green fluorescence protein (VSV-GFP; kindly provided by John levels is ensured by salvage of uracil and also by the pyrimidine Hiscott, McGill University, Montreal, QC, Canada) were grown and titered in Vero cells. Newcastle disease virus (NDV-LaSota) was grown in 10-d-old de novo biosynthesis pathway. It is known that different cell ’ fl types rely to varying degrees on the de novo biosynthesis versus embryonated hens eggs and titered in DF1 cells. The in uenza viruses A/ Hong Kong/68 (H3N2), A/Victoria/3/75 (H3N2), A/Sw/Texas/98 (H1N1), A/ salvage pathways (29), and this could be one explanation for the Moscow/10/99 (H3N2), A/NY/2008 (H1N1), A/PR/8/34 (H1N1), A/Udorn/72 observation that excess uracil can more easily restore virus rep- (H3N2), and Sendai Virus (SV52) were grown in 8-d-old embryonated hens’ lication in MDCK than in A549 cells. Alternatively, this could eggs and titered in MDCK or Vero cells. Human adenovirus (hAd5) was indicate species-specific activity of A3, particularly as the phe- grown and titered in A549 cells. Vaccinia virus (NYVAC) was grown and notype in A549 cells was also seen in other human and primate titered in CV1 cells. HIV-1 (NL4-3) viral stocks were produced by transfection cells independent of their tissue origin. This hypothesis is sup- of HEK 293T and titered as described previously (30). Hepatitis C virus ported by the complete absence of A3 antiviral activity in (Jc1FLAG2[p7-nsGluc2a]) was kindly provided by Charles Rice (Rockefeller avian cells. University, New York, NY) and grown as described previously (31). The restorative activity of uracil could indicate that A3 is af- The influenza minigenome reporter (pPolI-Luc) was previously described (27). fecting either the uracil salvage pathway or the pyrimidine de novo biosynthesis pathway. However, the finding that orotic acid Compounds. A3 was purchased from ASINEX. Diphyllin was purchased from also has the ability to reverse A3-mediated inhibition of viral ChemDiv.Ribavirin,leflunomide,purines(adenine,inosine,2′-deoxyadenosine, replication led to the conclusion that A3 targets the pyrimidine 2′-deoxyguanosine, adenosine, guanosine, hypoxanthine), pyrimidines (thy- de novo biosynthesis pathway. The de novo synthesis pathway mine, cytosine, uracil), and components of the pyrimidine de novo biosynthe- consists of six synthetic steps; orotate is the product of the fourth sis (aspartic acid, carbamoyl phosphate, carbamoyl aspartate, dihydroorotic acid, orotic acid, orotidylic acid) were purchased from Sigma-Aldrich. Lam- step, which is catalyzed by DHODH. This strongly suggests a role ivudine, nevirapine, , and amprenavir were obtained through the for this enzyme in the inhibitory mechanism of A3 but an in vitro NIH AIDS Research and Reference reagent program, Division of AIDS, National enzymatic assay is needed to verify this hypothesis. Interestingly, Institute of Allergy and Infectious Diseases, National Institutes of Health. the chemical structure of A3 is distinct from those of leflunomide and brequinar, which are known inhibitors of DHODH, so Cell Viability Assay. The CellTiterGlo Cell Viability Assay (Promega) was used to perhaps the mode of DHODH inhibition may differ. DHODH detect cell viability. A549 cells were seeded into 96-well plates at 1.25 × 103 has two binding sites. The substrate dihydroorotic acid (DHO) cells per well and after 24 h incubation, the medium was replaced with binds to the first site and is oxidized via a cosubstrate electron 100 μL fresh DMEM containing the compounds. After a further 24-or 48-h

Hoffmann et al. PNAS Early Edition | 5of6 Downloaded by guest on September 28, 2021 incubation, CellTiterGlo solution was added to each well, and luminescence Influenza Virus Minigenome Assay. A549 cells were seeded into 12-well plates 5 was measured using a Beckman Coulter DTX 880 plate reader (Beck- at 2 × 10 cells per well and incubated overnight at 37 °C, 5% CO2. The cells man Coulter). were transfected with pCAGGS constructs for influenza A/WSN/33 virus PB1, PB2, and PA (100 ng each) and NP (200 ng), the RNA polymerase II driven Viral Growth Assays in the Presence of Inhibitors. Cells were seeded into six- Renilla luciferase reporter pRLTK (Promega) (200 ng), and the influenza fi fi fl wellplatesat5× 105 cells per well or into 12-well plates at 2 × 105 cells per virus-speci c RNA polymerase I driven re y luciferase reporter (pPolI Luc) (150 ng). Four hours before transfection, the cells were cultured in DMEM well and incubated for 24 h at 37 °C, 5% CO2. Four hours before infection, the cells were washed with PBS (Invitrogen), and the medium was replaced supplemented with compounds at their CC10 or DMSO. The transfection was with DMEM containing the compound of interest at the indicated concen- performed with Lipofectamine 2000 (Invitrogen) in OptiMEM (Invitrogen), which was also supplemented with compounds or DMSO. OptiMEM was trations. Compounds were absent during the 1-h virus incubation but were replaced 4 h posttransfection with DMEM containing compounds or DMSO. present in the DMEM postinfection medium. For infections with influenza After a 20- to 24-h incubation period, cells were harvested, and fireflylu- viruses, NDV-LaSota and SV52 postinfection medium also contained 1 μg/mL ciferase and Renilla luciferase expression was determined using the Dual TPCK-treated trypsin (Sigma-Aldrich). When indicated, the medium was Luciferase Assay Kit (Promega). supplemented with uracil, dihydroorotic acid, and orotic acid. The infected cells were incubated at 37 °C with the exception of influenza B virus-infected Primer Extension Assay. A549 cells were seeded into 12-well plates at 2 × 105 cells, which were incubated at 33 °C. Viral titers were determined by stan- cells per well. After incubation for 24 h at 37 °C and 5% CO , the cells were dard plaque assay in MDCK cells. For influenza virus plaque reduction assays, 2 washed with PBS, and the medium was replaced with DMEM supplemented the overlay medium was supplemented with the indicated purines, pyr- with DMSO, A3 (2 and 10 μM) or ribavirin (10 and 100 μM). The cells were imidines, or components of the pyrimidine de novo biosynthesis pathway. incubated for 4 h before infection with influenza virus A/PR/8/34 (MOI = 7). Viral replication of WNV, DENV-I, and HCV was monitored by the production RNA was extracted 9 h postinfection using the QIAamp viral RNA kit (Qia- of Renilla luciferase, which was measured using the Renilla Luciferase Assay gen). Primers were synthesized for the NA segment of influenza A/PR/8/34 Kit (Promega) according to the specifications of the manufacturer. TZM-bl virus (c/m-RNA primer: 5′-tccagtatggttttgatttccg-3′ and v-RNA primer: 5′- reporter cells, which encode a Tat-responsive β-galactosidase indicator gene ggactagtgggagcatcatttc-3′) and the human 5S rRNA (5′-tcccaggcggtctcc- under the transcriptional control of the HIV-1 LTR, were used to assess HIV-1 catcc-3′). For the RT reaction, they were labeled with ATP-[γ-32P] using T4- infectivity in the presence of A3. β-Galactosidase activity was quantified 48 h kinase (Invitrogen) according to the specifications of the manufacturer. Viral after infection as described elsewhere (32). Infectivity values were normal- RNA (2 μg) was reverse transcribed with the SuperScript First-Strand Syn-

ized and IC50 values were computed by nonlinear regressions (GraphPad thesis System (Invitrogen) using labeled primers for v-, c/mRNA, and 5s rRNA. PRISM). For the time-of-addition experiments, TZM-bl reporter cells were Samples were separated on a 6% SDS/PAGE gel that contained 5 M urea, infected in duplicate with NL4-3 at an MOI of 0.1. FDA-approved HIV-1 transferred to membrane, and crosslinked. cDNA was visualized by exposure antiretroviral drugs were added at 0, 1, 2, 4, 6, 8, 10, and 12 h postinfection (24–72 h at −80 °C) to autoradiographic film (World Wide Medical Products at a multiple of the concentration required to inhibit 50% of the viral [WWMP]).

infectivity (IC50), as follows: lamivudine (a nucleoside analog reverse tran-

scriptase inhibitor, concentration 10 μM; 25-fold IC50), nevirapine (a non- ACKNOWLEDGMENTS. We thank Matthew Evans, Mila B. Ortigoza, and Jasmine Perez for technical advice and Adolfo García-Sastre for providing nucleoside reverse transcriptase inhibitor, concentration 1 μM; 11-fold IC50); reagents. This research was supported in part by National Institutes of raltegravir (an inhibitor, 1 μM; 75-fold IC ), and amprenavir (a 50 Health Grants U54 AI057158, U01 AI1074539, HHSN272200900032C, and μ protease inhibitor, 1 M; 50-fold IC50). A3 was used at a concentration of R21AI083673 (to M.L.S. and P.P.) and AI089246 and AI064001-06 (to μ fi β fi 1 M( vefold IC50). -Galactosidase activity was quanti ed 48 h after infec- V.A.S.). A.K. was supported by a fellowship from the Max Kade foundation. tion as previously described (32). V.A.S. is a Sinsheimer Scholar.

1. Bright RA, Shay DK, Shu B, Cox NJ, Klimov AI (2006) Adamantane resistance among 17. Waldman WJ, et al. (1999) Inhibition of cytomegalovirus in vitro and in vivo by the influenza A viruses isolated early during the 2005-2006 influenza season in the United experimental immunosuppressive agent leflunomide. Intervirology 42:412–418. States. JAMA 295:891–894. 18. Waldman WJ, et al. (1999) Novel mechanism of inhibition of cytomegalovirus by the 2. Bright RA, et al. (2005) Incidence of adamantane resistance among influenza A experimental immunosuppressive agent leflunomide. Transplantation 68:814–825. (H3N2) viruses isolated worldwide from 1994 to 2005: A cause for concern. Lancet 366: 19. Evers DL, Wang X, Huong SM, Andreoni KA, Huang ES (2005) Inhibition of human 1175–1181. cytomegalovirus signaling and replication by the immunosuppressant FK778. Antiviral 3. Wolf MC, et al. (2010) A broad-spectrum antiviral targeting entry of enveloped Res 65:1–12. viruses. Proc Natl Acad Sci USA 107:3157–3162. 20. McLean JE, Neidhardt EA, Grossman TH, Hedstrom L (2001) Multiple inhibitor analysis fl 4. St Vincent MR, et al. (2010) Rigid amphipathic fusion inhibitors, small molecule of the brequinar and le unomide binding sites on human dihydroorotate dehy- – antiviral compounds against enveloped viruses. Proc Natl Acad Sci USA 107:17339– drogenase. Biochemistry 40:2194 2200. 17344. 21. Greene S, Watanabe K, Braatz-Trulson J, Lou L (1995) Inhibition of dihydroorotate fl 5. Sidwell RW, et al. (1972) Broad-spectrum antiviral activity of Virazole: 1-beta-D- dehydrogenase by the immunosuppressive agent le unomide. Biochem Pharmacol 50:861–867. ribofuranosyl-1,2,4-triazole-3-carboxamide. Science 177:705–706. 22. Liu S, Neidhardt EA, Grossman TH, Ocain T, Clardy J (2000) Structures of human 6. Reichard O, Schvarcz R, Weiland O (1997) Therapy of hepatitis C: Alpha and dihydroorotate dehydrogenase in complex with antiproliferative agents. Structure 8:25–33. ribavirin. Hepatology 26(3, Suppl 1):108S–111S. 23. Knecht W, Löffler M (1998) Species-related inhibition of human and rat 7. Toltzis P, O’Connell K, Patterson JL (1988) Effect of phosphorylated ribavirin on dihydroorotate dehydrogenase by immunosuppressive isoxazol and cinchoninic acid vesicular stomatitis virus transcription. Antimicrob Agents Chemother 32:492–497. derivatives. Biochem Pharmacol 56:1259–1264. 8. Crotty S, et al. (2000) The broad-spectrum antiviral ribonucleoside ribavirin is an RNA 24. Evans DR, Guy HI (2004) Mammalian pyrimidine biosynthesis: Fresh insights into an virus mutagen. Nat Med 6:1375–1379. ancient pathway. J Biol Chem 279:33035–33038. 9. Streeter DG, et al. (1973) Mechanism of action of 1-β-D-ribofuranosyl-1,2,4-triazole-3- 25. Li J, et al. (2007) Identification and characterization of human uracil phosphoribosyl- carboxamide (Virazole), a new broad-spectrum antiviral agent. Proc Natl Acad Sci USA transferase (UPRTase). J Hum Genet 52:415–422. 70:1174–1178. 26. Herrmann ML, Schleyerbach R, Kirschbaum BJ (2000) Leflunomide: An immuno- 10. Leyssen P, Balzarini J, De Clercq E, Neyts J (2005) The predominant mechanism by modulatory drug for the treatment of rheumatoid arthritis and other autoimmune fl which ribavirin exerts its antiviral activity in vitro against aviviruses and paramyxo- diseases. Immunopharmacology 47:273–289. – viruses is mediated by inhibition of IMP dehydrogenase. J Virol 79:1943 1947. 27. Hoffmann HH, Palese P, Shaw ML (2008) Modulation of influenza virus replication by 11. Gowen BB, et al. (2007) In vitro and in vivo activities of T-705 against arenavirus and alteration of sodium ion transport and protein kinase C activity. Antiviral Res 80:124–134. – bunyavirus infections. Antimicrob Agents Chemother 51:3168 3176. 28. König R, et al. (2010) Human host factors required for influenza virus replication. 12. Furuta Y, et al. (2009) T-705 (favipiravir) and related compounds: Novel broad- Nature 463:813–817. spectrum inhibitors of RNA viral infections. Antiviral Res 82:95–102. 29. Traut TW, Jones ME (1996) Uracil metabolism—UMP synthesis from orotic acid or 13. Furuta Y, et al. (2002) In vitro and in vivo activities of anti-influenza virus compound uridine and conversion of uracil to beta-alanine: Enzymes and cDNAs. Prog Nucleic T-705. Antimicrob Agents Chemother 46:977–981. Acid Res Mol Biol 53:1–78. 14. Kiso M, et al. (2010) T-705 (favipiravir) activity against lethal H5N1 influenza A viruses. 30. Mulder LC, Harari A, Simon V (2008) Cytidine deamination induced HIV-1 drug Proc Natl Acad Sci USA 107:882–887. resistance. Proc Natl Acad Sci USA 105:5501–5506. 15. Qing M, et al. (2010) Characterization of dengue virus resistance to brequinar in cell 31. Marukian S, et al. (2008) Cell culture-produced hepatitis C virus does not infect culture. Antimicrob Agents Chemother 54:3686–3695. peripheral blood mononuclear cells. Hepatology 48:1843–1850. 16. Knight DA, et al. (2001) Inhibition of herpes simplex virus type 1 by the experimental 32. Harari A, Ooms M, Mulder LC, Simon V (2009) Polymorphisms and splice variants immunosuppressive agent leflunomide. Transplantation 71:170–174. influence the antiretroviral activity of human APOBEC3H. J Virol 83:295–303.

6of6 | www.pnas.org/cgi/doi/10.1073/pnas.1101143108 Hoffmann et al. Downloaded by guest on September 28, 2021