Inhibition of rotavirus replication by downregulation of fatty acid synthesis

Recently the recruitment of lipid droplets (LDs) to sites of rotavirus (RV) replication was reported. LDs are polymorphic organelles that store triacylglycerols, cholesterol and cholesterol esters. The neutral fats are derived from palmitoyl-CoA, synthesized via the fatty acid biosynthetic pathway. RV-infected cells were treated with chemical inhibitors of the fatty acid biosynthetic pathway, and the effects on viral replication kinetics were assessed. Treatment with compound C75, an inhibitor of the fatty acid synthase enzyme complex (FASN), reduced RV infectivity 3.2-fold (P 5 0.07) and modestly reduced viral RNA synthesis (1.2-fold). Acting earlier in the fatty acid synthesis pathway, TOFA [5-(Tetradecyloxy)-2-furoic acid] inhibits the enzyme acetyl-CoA carboxylase 1 (ACC1). TOFA reduced the infectivity of progeny RV 31-fold and viral RNA production 6-fold. The effect of TOFA on RV infectivity and RNA replication was dose-dependent, and infectivity was reduced by administering TOFA up to 4 h post-infection. Co-treatment of RV-infected cells with C75 and TOFA synergistically reduced viral infectivity. Knockdown by siRNA of FASN and ACC1 produced findings similar to those observed by inhibiting these proteins with the chemical compounds. Inhibition of fatty acid synthesis using a range of approaches uniformly had a more marked impact on viral infectivity than on viral RNA yield, inferring a role for LDs in virus assembly and/or egress. Specific inhibitors of fatty acid metabolism may help pinpoint the critical structural and biochemical features of LDs that are essential for RV replication, and facilitate the development of antiviral therapies.


INTRODUCTION
The double-stranded (ds) RNA rotaviruses (RVs) comprise a genus of the Reoviridae family and consist of at least seven phylogenetically defined species (Matthijnssens et al., 2012).Species A RVs (Estes & Kapikian, 2007) cause an estimated 440 000 child deaths per year (Parashar et al., 2003;Tate et al., 2012).Two live attenuated RV vaccines have been licensed in various countries since 2006, and their overall effectiveness is being assessed (Ruiz-Palacios et al., 2006;Vesikari et al., 2006;Yen et al., 2011).Currently there are no specific antiviral drugs approved to treat RV disease.
RVs replicate in mature enterocytes of the small intestine, in which they form cytoplasmic inclusion bodies known as 'viroplasms' (Petrie et al., 1982), where early viral morphogenesis and RNA replication take place.Recently it was demonstrated that lipid droplet (LD) organelles co-localize with viroplasms during RV infection in vitro (Cheung et al., 2010).LDs sequester lipids including sterol esters, diacyland triacylglycerols; they exhibit marked morphological and compositional plasticity (Martin & Parton, 2006).The LD proteome comprises over 100 proteins including perilipin (PLIN), which is localized exclusively to the LD phospholipid monolayer (Clifford et al., 2000).
Initially, biogenesis of the phospholipid, triacylglycerol and fatty acid components of LDs follows a common pathway beginning in the cytoplasm with the conversion of acetyl-CoA to malonyl-CoA.This rate-limiting step, catalysed by acetyl-CoA carboxylase 1 (ACC1), predetermines entry into the fatty acid synthesis cycle.One molecule each of acetyl-CoA and malonyl-CoA combine to form a four-carbon compound that undergoes a series of biochemical modifications catalysed by the fatty acid synthase (FASN) enzyme complex.A new malonyl-CoA then enters the cycle and contributes two carbons for elongation of the substrate, and this cycle continues until a sixteen-carbon compound, palmitoyl, is produced and subsequently activated as palmitoyl-CoA.Palmitoyl is the precursor molecule for the synthesis of phospholipids, triacylglycerols and fatty acids (Fig. S1, available in JGV Online).Chemical inhibition of triacylglycerol synthesis by triacsin C decreases the production of infectious RV in vitro (Cheung et al., 2010).We have investigated compounds targeting various stages of palmitoyl-CoA synthesis for their ability to inhibit RV replication.Compound C75 (tetrahydro-4-methylene-2Roctyl-5-oxo-3S-furancarboxylic acid) inhibits the enzyme complex FASN, and in doing so significantly reduces LD accumulation (Schmid et al., 2005).TOFA [5-(tetradecyloxy)-2-furoic acid] acts at the earlier rate-limiting step of fatty acid biosynthesis by inhibiting the enzyme ACC1 (Parker et al., 1977).Here we report that the production of infectious viral progeny and of rotaviral RNA in vitro is decreased in the presence of either C75 or TOFA and that the two compounds act synergistically.

Chemical inhibition of FASN
To determine whether inhibition of the fatty acid synthase enzyme complex would disrupt RV replication, infected MA104 cells were treated with compound C75 at 15 mM (Fig. 1) [The distribution of NSP5 (viroplasms) and perilipin inside the infected cell is shown at higher magnification in Fig. S2 and can also be inspected in Cheung et al., 2010].The dose of 15 mM was not cytotoxic to MA104 cells, but close to the minimal cytotoxic dose (20 mM; Fig. S3).RV propagated in treated cells showed a 3.2-fold reduction in infectivity (P50.07) and a 1.2-fold decrease in viral RNA production (Table 1).Significantly fewer cells were infected when treated with C75 (P50.008;Tables 2 and S2).However, the number of viroplasms in C75-treated cells was not reduced (P50.252;Table 2).
As a comparison, treatment was undertaken with 3isobutyl-1-methylxanthine (IBMX) and isoproterenol (IP), previously shown to inhibit RV replication (Cheung et al., 2010), alongside experiments in which TOFA was either added to the virus inoculum 30 min pre-infection, or to RV infected cells at 30 min post-infection (p.i.).Under all three conditions, infectivity was significantly reduced and viral RNA synthesis was decreased as measured by a decrease in VP6 RNA transcription (Table S4).No difference in infectivity or RNA detection was seen when virus was incubated with TOFA prior to infection compared with TOFA added 30 min p.i. (Table S4).
TOFA treatment over a range of 400 to 1200 nM revealed a dose-dependent reduction in virus infectivity and viral RNA production (Fig. 2).The reduction in infectivity was more marked than the reduction in RNA production; this was more obvious with increasing TOFA concentrations.When TOFA was added at later times in relation to the time of infection, the reduction in virus infectivity was smaller.Nevertheless, addition of the drug as late as at 4 h p.i. was associated with a reduced cytopathicity, a decrease in viral RNA synthesis and a reduction in infectious viral progeny (Fig. 3, Table S5).Pre-treatment of cells with TOFA did not inhibit production of infectious virus progeny or RNA synthesis any further than TOFA treatment at the time of infection (Fig. 3, Table S5).

Combined drug treatment
Combined treatment with C75 (15 mM) and TOFA (1.2 mM) was no more cytotoxic than either of the drugs alone (results not shown).When TOFA and C75 were used in combination at these concentrations at the time of infection (Fig. 1) infectivity titres were reduced a further 10-fold compared with TOFA treatment alone, and RNA production was also synergistically reduced (Table 1).The number of infected cells was significantly reduced in cells treated with both drugs in combination compared with untreated cells (P,0.0001;Table 2) and C75-only treated cells (P,0.0001) but not when compared with TOFA-only treated cells (P.0.999).The number of viroplasms in cells treated with both drugs was significantly reduced compared with untreated cells (P,0.0001;Table 2) and cells treated with either C75 or TOFA alone (P.0.0001 for both).

siRNA inhibition of FASN
To further investigate the role of FASN in the RV replication cycle, MA104 cells were transfected with siRNA directed against FASN.Knockdown was verified by Western blotting (Fig. 4).Despite a considerable reduction in the amount of FASN detected (48.7 % compared with non-targeting siRNA control), no significant reduction in virus infectivity was observed on siRNA treatment, and the VP6 gene-based transcriptional levels did not differ between siRNA-treated and mock-treated RV-infected cells (Table 3).No difference was seen between the number of infected cells in FASN siRNA-treated cells compared with those treated with the non-targeting siRNA (P.0.999;Table S6).

siRNA inhibition of ACC1
MA104 cells were transfected with siRNA directed against ACC1, and knockdown was verified by Western blotting (95.2 % compared with non-targeting siRNA control; Fig. 4).Infectivity of RV in ACC1 knockdown cells was reduced by 8.5-fold (significant, P,0.01), whereas the VP6 genebased transcriptional levels were unchanged by siRNA  3).The numbers of infected cells treated with ACC1 siRNA compared with those treated with the non-targeting siRNA were not significantly different (P50.116;Table S6).

DISCUSSION
LDs are increasingly appreciated as dynamic and motile cellular organelles that increase in size in response to elevated fatty acid levels (Martin & Parton, 2006).Roles for LDs in the replication cycle of several viruses, including RVs (Cheung et al., 2010), hepatitis C virus (HCV) (Barba et al., 1997;Miyanari et al., 2007), dengue virus (Samsa et al., 2009), and human parechovirus 1 (Krogerus et al., 2007) have been identified.Lipids are involved in diverse events during virus replication including endocytosis, formation of replication complexes and assembly and budding (Chukkapalli et al., 2012;Heaton & Randall, 2011).
We have used compounds that inhibit different stages of fatty acid biosynthesis and determined their effects on RV replication, viral RNA production and viroplasm formation in vitro.At the doses used, we found that TOFA, an inhibitor of ACC1, had a more potent antiviral effect than Table 1.Effects of host cell treatment with compounds C75 (15 mM) and TOFA (1.2 mM) on RV infectivity and RV RNA production (for further details see Tables S2 and S3)  et al., 2000) and is being considered as an anti-cancer drug in humans (Guseva et al., 2011).Critically, we found that TOFA reduced production of infectious RV progeny even when cells were treated several hours after infection.Inhibition of the enzymic drug targets by the alternative approach of siRNA knockdown verified the advantage of targeting ACC1, the enzyme catalysing the rate-limiting step of the palmitoyl-CoA synthetic pathway, over FASN.Conceivably, ACC1 inhibition is more successful as this enzyme is pivotal to the efficiency by which this pathway proceeds.Nevertheless, it is worth considering whether ACC1 might play some other role in RV replication.For example, ACC1 but not FASN is thought to be found on the surface of LDs (Goodman, 2009), and so might play a crucial role via a direct or indirect interaction with one or more viral components.
Human cytomegalovirus (HCMV) upregulates ACC1 during infection, and TOFA treatment of HCMV-infected cells attenuates the replication of this virus (Spencer et al., 2011).Addition of 5 or 10 mg ml 21 TOFA (15.4 mM or 30.8 mM, respectively) at the time of infection reduced HCMV replication by 10-to 100-fold, demonstrating a dose-dependent effect comparable to that seen for RVs.
Similarly, treatment with TOFA up to 48 h p.i. was able to reduce HCMV replication, as shown for RVs (for TOFA addition at up to 4 h p.i.).HCV recruits LDs to its replication complexes (Barba et al., 1997;Miyanari et al., 2007).It has been suggested that this recruitment either provides a structural framework for HCV assembly (Miyanari et al., 2007) or that LDs are exploited as vehicles to relocate core protein from sites of RNA translation and replication to sites of virion assembly (Bartenschlager et al., 2011).TOFA (5 mg ml 21 , 15.4 mM) reduced HCV replication around 3-fold (Kapadia & Chisari, 2005).The difference in sensitivity of HCV to TOFA compared with RV may relate to a number of factors including the involvement of LDs at different stages of the viral replication cycles.
Inhibition of fatty acid synthesis, and as a consequence LD biogenesis, had a much greater inhibitory effect on the infectivity of RV progeny than on RV RNA production.This is consistent with an effect on virus assembly or egress rather than an effect at the earlier stage of replication of viral RNA.For example, LDs may serve as a scaffold for building infectious virions.
This latest data set confirm and substantiate the physical and functional interaction of the cellular organelles' LDs with RV-induced viroplasms or viroplasm-like structures.
The finding that disturbance of LD homeostasis in various ways interferes with RV replication emphasizes the importance and significance of this interaction.
LDs have been shown to be intrinsically involved in the replication of members of several virus families.As fatty acid upregulation promotes LD synthesis, the inhibition of the fatty acid biosynthetic pathway provides a mechanism by which replication of several viruses can be decreased.
Here we have demonstrated the effectiveness of this approach in the context of RV replication, and there is the potential of targeting this pathway for therapeutic developments.Sigma; 20 mM as previously described (Cheung et al., 2010)] were checked for cytotoxic effects using trypan blue staining, as previously described (Tennant, 1964).When suitable working concentrations were established for the drugs, they were verified by measuring ATP activity as a direct correlate of the number of viable cells using the CellTiter Glo assay (Promega, G7570) in accordance with the manufacturer's instructions.Furthermore, as it has previously been reported that TOFA is associated with activation of the apoptotic pathway at the low concentrations established as suitable by these methods (Guseva et al., 2011), both drugs were analysed to determine whether they were associated with caspase activation using the Caspase-Glo 3/7 Assay (Promega, G8090) according to the manufacturer's instructions.Luminescence for the CellTiter Glo Caspase-Glo assays was measured using the Glomax Multi-Detection System (Promega) with program settings as provided by the manufacturer.
RNAi assay.Pooled siRNAs (Dharmacon) were used to target individual FASN (L-003954-00-0005) and ACC1 (L-004551-00-0005), simultaneously with non-targeting (irrelevant, irr) siRNA (D-001810-10-05) and transfection controls.As these siRNAs target human genes, their complementarity with the African green monkey homologues (from which the MA104 cell line used is derived) was confirmed by sequence comparison in the Simmonics v1.0 program (Simmonds, 2012) (representative genome sequences were downloaded from GenBank).Day 1: 80 nM siRNA was complexed with N-TER peptide (N-TER nanoparticle transfection system, Sigma), according to the manufacturer's recommended protocol for reverse transfection, using medium containing 5 % FCS.Then 250 ml aliquots of the Nanoparticle Formation Solution (NFS) produced were dispensed into 24-well plates; ~1610 5 MA104 cells in 250 ml medium (5 % FCS) were then added to each well and cultured overnight.Day 2: cell confluency of ¢90 % was confirmed.Medium was replaced by fresh NFS (80 nM siRNA) in the presence of serum (manufacturer's forward transfection protocol with a total volume of 500 ml/well) and incubated overnight.Day 3: medium was replaced, with or without FCS depending on the confluency of the cells.Day 4: cells were infected with RV at an m.o.i. of ~1 and incubated overnight (~16 h).
Day 5: cells were harvested for immunofluorescence testing, infectivity titrations, qRT-PCR and Western blotting.
Determination of viral infectivity by TCID 50 .Infected cells were harvested by scraping and freeze-thawing, and inoculated in 10-fold dilutions onto confluent MA104 cells.CPEs were recorded, and infectivity was expressed as TCID 50 ml 21 , calculated according to the method of Reed and Muench (Reed & Muench, 1938) as described previously (Arnoldi et al., 2007).
Comparison of VP6 transcriptional levels by real-time qRT-PCR.Cells were harvested using RNA Protect Cell Reagent (Qiagen) according to the manufacturer's protocol.Nucleic acids were extracted using the Qiagen RNeasy Plus mini kit according to the manufacturer's instructions.Quantitative RT-PCR was then undertaken using the ABI 7500 Fast Real-Time PCR System with the Qiagen Quantifast Multiplex RT-PCR kit; 25 ml reaction mix contained 12.5 ml 2X Quantifast RT mix, 0.5 ml ROX reporter dye, 1 ml forward primer for viral protein VP6 detection (59-1089 AGGTATGAAYTG-GACTGATTTRATC) (10 mM stock, Sigma), 1 ml reverse primer for VP6 detection (39-1212 AAGTKGTTAGCTTGGTCCTCATTT) (10 mM stock, Sigma), 0.5 ml VP6_VIC conjugated probe (59-1153 GTATTTACAGTGGCTTCCATTAGAAGCATGCT) (10 mM stock, Applied Biosystems), 0.25 ml Quantifast RT mix and 5 ml RNA.Primers were designed for strain specific complementarity and optimized for real-time suitability in the Simmonics v1.0 and Oligocalc (Kibbe, 2007) programs.Cycling conditions were 50 uC for 20 min then 95 uC for 5 min, followed by 45 cycles of 95 uC for 15 s then 60 uC for 30 s.For chemical inhibition studies, RV-infected samples were processed in triplicate and mean C t values were calculated with outliers (.1 C t difference) being removed.For the siRNA experiments, infected cells treated with ACC1 and FASN siRNAs and infected cells treated with the non-targeting siRNA only were analysed in triplicate.
Confocal microscopy.MA104 cells grown on coverslips and then infected with RV were fixed by treatment with 2 % formaldehyde in PBS for 15 min.Cells were permeabilized with 1 % Triton X-100/ 0.1 % BSA for 30 min, and then washed and treated with primary antibodies (Table S1) in 1 % BSA for ¢1 h.Cells were then washed and treated with appropriate secondary antibodies (Table S1) with conjugated fluorophore in 1 % BSA and 5 ng ml 21 Hoechst nuclear stain 33342 (Invitrogen) for ¢1 h.Cells were washed, mounted onto slides using AntiFade Gold Mountant (Invitrogen) and inspected by confocal microscopy using the Leica DM Libre TCS SP instrument at 660 magnification.Counts of infected cells in the absence/presence of drugs were undertaken by random sampling of sixteen frames per Table 3.Effect of TOFA (post and pre infection) and of isoproterenol+IBMX on rotavirus infectivity and viral RNA synthesis Control infectivity titres were calculated by taking the average infectivity titre for the three controls (cells transfected with non-targeting siRNA, cells treated with the transfection reagent only and untreated cells).Change in infectivity was calculated as the difference between infectivity titre in RVinfected cells transfected with FASN/ACC1 siRNA compared with the mean infectivity of RV in the three controls.Changes in RNA levels were calculated by determining the difference between the C t value for VP6 RNA in cells transfected with targeting siRNA (FASN/ACC1) compared with the VP6 C t value in RV infected cells transfected with non-targeting siRNA (control) 3Joint senior and corresponding authors.Three supplementary figures and six supplementary tables are available with the online version of this paper.IP: 54.70.40.11On: Sat, 09 Mar 2019 02:51:37 IP: 54.70.40.11On: Sat, 09 Mar 2019 02:51:37 treatment (Table

Fig. 2 .
Fig. 2. Dose-dependent decrease in RV infectivity and RV RNA production in vitro after treatment of host cells with TOFA.The data indicate AM±SEM (n53).

Fig. 3 .
Fig. 3. Decrease of RV infectivity and viral RNA production when TOFA was added at different time points relative to the time of RV infection.The data indicate AM±SEM (n53).

FASNFig. 4 .
Fig. 4. Western blot of ACC1 and FASN after treatment of cells with specific siRNAs at the time of RV infection.Lane 1, siRNA; lane 2, non-targeting siRNA control; lane 3, N-TER transfection control.

Table 2 .
Counts of the number of infected cells and the number of viroplasms when infected cells were treated with compounds C75 and TOFA . Data are arithmetic mean±SEM.No. of experiments was 3 (infectivity) or 2 (RNA synthesis).