Sterol Carrier Protein 2, a Critical Host Factor for Dengue Virus Infection

Home , Sterol

Sterol Carrier Protein 2, a Critical Host Factor for Dengue Virus Infection, Alters the Cholesterol Distribution in Mosquito Aag2 Cells Author(s): Qiang Fu, Bahar Inankur, John Yin, Rob Striker, and Que Lan Source: Journal of Medical Entomology, 52(5):1124-1134. Published By: Entomological Society of America URL: http://www.bioone.org/doi/full/10.1093/jme/tjv101

BioOne (www.bioone.org) is a nonprofit, online aggregation of core research in the biological, ecological, and environmental sciences. BioOne provides a sustainable online platform for over 170 journals and books published by nonprofit societies, associations, museums, institutions, and presses. Your use of this PDF, the BioOne Web site, and all posted and associated content indicates your acceptance of BioOne’s Terms of Use, available at www.bioone.org/page/ terms_of_use. Usage of BioOne content is strictly limited to personal, educational, and non-commercial use. Commercial inquiries or rights and permissions requests should be directed to the individual publisher as copyright holder.

BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofit publishers, academic institutions, research libraries, and research funders in the common goal of maximizing access to critical research. VECTOR/PATHOGEN/HOST INTERACTION,TRANSMISSION Sterol Carrier Protein 2, a Critical Host Factor for Dengue Virus Infection, Alters the Cholesterol Distribution in Mosquito Aag2 Cells

1,2 3 3 4 1,5 QIANG FU, BAHAR INANKUR, JOHN YIN, ROB STRIKER, AND QUE LAN

J. Med. Entomol. 52(5): 1124–1134 (2015); DOI: 10.1093/jme/tjv101 ABSTRACT Host factors that enable dengue virus (DENV) to propagate in the mosquito host cells are unclear. It is known that cellular cholesterol plays an important role in the life cycle of DENV in human host cells but unknown if the lipid requirements differ for mosquito versus mammalian. In mosquito Aedes aegypti, sterol carrier protein 2 (SCP-2) is critical for cellular cholesterol homeostasis. In this study, we identified SCP-2 as a critical host factor for DENV production in mosquito Aag2 cells. Treatment with a small molecule commonly referred to as SCPI-1, (N-(4-{[4-(3,4-dichlorophenyl)-1,3- thiazol-2-yl]amino}phenyl)acetamide hydrobromide, a known inhibitor of SCP-2, or knockdown of SCP- 2 dramatically repressed the virus production in mosquito but not mammalian cells. We showed that the intracellular cholesterol distribution in mosquito cells was altered by SCP-2 inhibitor treatment, suggesting that SCP-2-mediated cholesterol trafficking pathway is important for DENV viral production. A comparison of the effect of SCP-2 on mosquito and human cells suggests that SCPI-1 treatment decreases cholesterol in both cell lines, but this decrease in cholesterol only leads to a decline in viral titer in mosquito host cells, perhaps, owing to a more drastic effect on perinuclear cholesterol storages in mosquito cells that was absent in human cells. SCP-2 had no inhibitory effect on another enveloped RNA virus grown in mosquito cells, suggesting that SCP-2 does not have a generalized anti-cellular or antiviral effect. Our cell culture results imply that SCP-2 may play a limiting role in mosquito–dengue vector competence.

KEY WORDS Aedes aegypti, Dengue, SCP-2, Cholesterol, host factor

Dengue fever and dengue hemorrhagic fever has re- the spread of DENV may require a more complete un- surged during the past decades. Tens of million cases derstanding of mosquito vector competence and target- of dengue fever and 500,000 cases of dengue hemor- ing of mosquito factors that allow robust viral rhagic fever occur every year. It is estimated that 2.5 propagation in Aedes mosquitoes. billion people, almost half of the world population, are In human cells, the life cycle of Flavivirus is tightly at risk of infection (Kyle and Harris 2008). Currently related to cellular membranes and cholesterol (Fernan- there is no effective vaccine or specific treatment for dez-Garcia et al. 2009). The enveloped DENV (Kuhn dengue disease. Dengue virus (DENV), a positive- et al. 2002) enters cells via receptor-mediated endocy- stranded RNA virus in Flaviviridae family, is transmit- tosis (van der Schaar et al. 2008, Acosta et al. 2008, ted mainly by the mosquito species Aedes aegypti and Chu and Ng 2004), followed by fusion of the viral enve- Aedes albopictus. Human host cell factors that are re- lope proteins with late endosomal membrane to release quired for DENV propagation and pathogenesis have the virus RNA in the cytoplasm. However, in insect been extensively studied (reviewed in Pastorino et al. cells, entry occurs at the plasma membrane (Hase et al. 2010, Lindenbach and Rice 2003, Rodenhuis-Zybert 1987, 1989; Vancini et al. 2013). Flaviviral infections in- 2010). Although DENV infection requires an extensive duce dramatic intracellular membrane rearrangement number of host factors, only a limited number of them and proliferation (Pena and Harris 2012, Perera et al. have been identified in mosquito (Sessions et al. 2009, 2012, Welsch et al. 2009). Virus RNA replication and Tchankouo-Nguetcheu 2010). New strategies to control assembly of the virus takes place in the membrane- bound complexes derived from the endoplasmic reticulum (ER) (Welsch et al. 2009; Gillespie et al. 2010). Virus particles are assembled and the mature viruses 1 Department of Entomology, University of Wisconsin-Madison, are released via cell secretory machinery (Mackenzie WI. and Westaway 2001). In mammalian cell studies, cho- 2 Correspondence author, e-mail: [email protected]. 3 lesterol-rich membranes have been shown to mediate Chemical and Biological Engineering Department, University of Flavivirus viral entry (Valle et al. 2005, Lee et al. 2008, Wisconsin-Madison, WI. 4 Department of Medicine, University of Wisconsin-Madison, WI. Medigeshi et al. 2008) as well as viral replication 5 Deceased. (Aizaki et al. 2007, Mackenzie et al. 2007).

VC The Authors 2015. Published by Oxford University Press on behalf of Entomological Society of America. All rights reserved. For Permissions, please email: [email protected] September 2015 FUETAL.: SCPI-1 INHIBITS DENGUE VIRUS PRODUCTION IN MOSQUITO CELL 1125

DENV replication in human host cells depends on compared SCP-2’s function in DENV infection be- both de novo cholesterol biosynthesis and intracellular tween insect and mammalian cells. Our findings sug- cholesterol transport. The key cholesterol synthesis en- gest that there are differences in the requirements for zymes, such as mevalonate diphospho decarboxylase cholesterol trafficking between dengue replication in and 3-hydroxy-methylglutaryl-CoA (HMGCoA) reduc- human versus mosquito host cells. tase (HMGCR), are necessary for the life cycle of Fla- viviruses (Mackenzie et al. 2007, Rothwell et al. 2009). Materials and Methods The cholesterol biosynthesis pathway has been suggested as a target for antiviral drugs. For example Chemicals. Chemicals and reagents were pur- HMGCR inhibitors (statins, such as lovastatin) lower chased from Sigma (St. Louis, MO), Fisher Scientific cholesterol levels (Endo 1992) and reduce DENV titers (Pittsburgh, PA), and Life Technologies (Carlsbad, in epithelial Vero cells and endothelial HMEC-1 cells CA), except when their origins are mentioned in the (Martinez-Gutierrez et al. 2011) as does Hymeglusin, text. N-(4-{[4-(3,4-dichlorophenyl)-1,3-thiazol-2-yl]ami- an HMGCoA synthesis inhibitor (Rothwell et al. 2009). no}phenyl)acetamide hydrobromide (SCPI-1) was pur- Inhibitors of squalene synthase, the enzyme that cata- chased from Chembridge Chemical (San Diego, CA). lyzes the committed step of sterol synthesis, signifi- Virus and Cell Culture. Ae. aegypti Aag2 cells cantly reduces DENV production in human K562 cells were maintained in Schneider’s Drosophila Medium (Rothwell et al. 2009). In human cells, the Niemann– and Ae. albopictus C6/36 cells were maintained in Lei- Pick disease Type C (NPC)-mediated intracellular cho- bovitz L-15 medium at 28 C(withoutCO2). The mos- lesterol trafficking pathway is likely involved in DENV quito cell culture medium was supplemented with 10% replication (Poh et al. 2012). DENV entry and replica- fetal bovine serum (FBS), 50 unit/ml penicillin, and tion in mammalian host cells are suppressed by 50 mg/ml streptomycin. Monkey kidney Vero cells and U18666A (Poh et al. 2012), which is a chemical inhibi- Human Huh7.5 cells were maintained in Dulbecco’s tor of intracellular transport pathways of low-density li- Modified Eagle Medium with 10% FBS, 1% non- poprotein-derived cholesterol from lysosomes to essential amino acids, 50 unit/ml penicillin, and 50 mg/ plasma membranes (Liscum et al. 1989) and from ml streptomycin at 37 C with 5% CO2. Passages of plasma membranes to intracellular membranes (Ha¨r- cells were conducted every 7–10 d with a 1:10 dilution ma¨la¨ et al. 1994). of cells. DENV has both invertebrate (mosquito) and verte- To prepare the virus stock, the DENV serotype 2 brate (human) hosts, but mosquitoes are cholesterol New Guinea C was used to infect Ae. albopictus C6/36 auxotrophs and cannot synthesize cholesterol de novo cell in a T75 flask at multiplicity of infection (MOI) of (Clayton 1964, Rawson 2003). The importance of cellu- 0.01 and incubated at 28C. The medium with virus lar cholesterol trafficking and distribution in mosquito– was harvested 7–10 d post-infection and was concen- virus interactions has not been addressed. DENV in- trated using 100-kDa filters (Millipore, Billerica, MA). fection in mosquito C6/C36 cells has been previously For VSV stock preparation, the VSV-DsRed-Express characterized as cholesterol-independent (Umashankar was used to infect BHK-21 cells in a T75 flask at an et al. 2008), similar to another enveloped RNA virus, MOI of 0.001. The infection was carried out for 15-18 vesicular stomatitis virus (VSV), which can also infect hours at 37 C and 5% CO2. The supernatant contain- mosquito cells (Phalen and Kielian 1991). Sterol carrier ing virus was then collected and filtered using 0.22-mm protein 2 (SCP-2) is a cytosolic protein involved in cho- filters (Millipore, Billerica, MA). lesterol binding and transport (Krebs and Lan 2003, Cytotoxicity Assays. Cytotoxicity of sterol carrier Lan and Massey 2004), and is especially important for protein inhibitor 1 (SCPI-1) was determined using the mosquito to acquire dietary cholesterol (Blitzer et al. CellTiter 96 AQueous Cell Proliferation Assay Kit 2005, Dyer et al. 2008). Knocking down of SCP-2 ex- (Promega, Madison, WI) following the manufacturer’s pression leads to reduced growth rate in larvae and direction. Briefly, cells (1 104/well) were seeded in lowered fertility in adults (Blitzer et al. 2005; Peng 96-well plates overnight and treated with serial dilu- et al. 2011, 2012). When SCP-2 has been mutated such tions of SCPI-1 for 72 h. Reaction solution (20 ml) pro- that the protein lacks cholesterol-binding capacity, cells vided in the kit was then added to medium and expressing these SCP-2 mutants no longer facilitate incubated at 37C for 1 hour, and the absorbance was cholesterol uptake in Ae. aegypti embryonic Aag2 cells subsequently read at 490 nm. The absorbance was nor- (Radek et al. 2010). In addition, SCP-2 protein inhibi- malized to control and plotted against the log transfor- tors (SCPIs) have been identified and act as insecti- mation of the concentration of SCPI-1. The 50% cides to mosquito larvae and adults (Kim et al. 2005, cytostatic concentration (CC50) value was determined Larson et al. 2008). We demonstrated that over-expres- using GraphPad Prism v4.0 software (San Diego, CA) sion of SCP-2 protein in mosquito Aag2 cells facilitated using nonlinear regression. uptake of [3H] cholesterol (Lan and Massey 2004, DENV Infection and Titration. Aag2 or Huh7.5 Radek et al. 2010). Given the role of SCP-2 in choles- cells (5 105/well) were seeded in 24-well plates over- terol uptake and metabolism in mosquito, we hypothe- night and treated with 0.1 mM, 1 mM, or 10 mM SCPI-1 sized that SCP-2 is a potential host factor that for 24 h before infection. The control group was treated facilitates DENV infection in mosquito host cells. In with the corresponding amount of DMSO to dissolve this study, the role of SCP-2 protein in virus–host inter- SCPI-1. DENV was then added (MOI ¼ 0.5–1) in action was investigated in Aag2 cells. In addition, we 0.5 ml serum-free medium and incubated at 37Cfor 1126 JOURNAL OF MEDICAL ENTOMOLOGY Vol. 52, no. 5

1.5 h. The virus-containing medium was removed after transfected with 0.2 ml transfection medium. After 5 h, infection. The cells were washed with serum-free the transfection medium was replaced with regular medium and cultured in the regular growth medium growth medium. The transient transfected cells were with SCPI-1. For the time-of-addition experiment, used for SCPI-1 treatment or virus infection 16 h post there were three different treatments: cells were only transfection. treated for 48 h with 10 mMSCPI-1beforeinfection siRNA Transfection. Small interference RNA (pre-incubation), only treated during the 1.5 h infection (siRNA) for mosquito SCP-2 was synthesized by IDT period (co-infection), or only treated for 48 h after (Coralville, IA), while siRNA for human SCP-2 was infection (post-infection), respectively. DENV was har- predesigned and purchased from Qiagen vested from the medium 48 h post-infection. (Cat#SI04950757). The sequences of siRNA are listed The virus was tittered using TCID50 assay described in Supp Table 2 [online only]. SiRNA (20 pmol) and before (Che et al. 2009). Briefly, Vero cells were grown 1 ml of Lipofectamine2000 (Life Technologies) were to 50–60% confluent monolayers in 96-well plates, dilutedin50mlofserum-freemediumfor5min, infected with 10-fold serial dilutions (each dilution in respectively, before combined. The mixture was then triplicates) of virus, and incubated at 37C with 5% incubated at room temperature for an additional CO2. Cytopathic effects were monitored after 5–7 d, 20 min and added to the host cells in a well of a 24- and the 50% tissue culture infective dose (TCID50)was well plate containing 0.5 ml growth medium. The calculated using the Reed and Muench method (1938). treated cells were used for virus infection 48 h post Virus titers were expressed as TCID50 units per ml. transfection. VSV Infection and Titration. The infections were Western Blot. Western blotting analysis was per- carried out using a fluorescent VSV strain incorporating formed as described, using SDS 15% PAGE. The pro- a DsRed-Express-DR gene (VSV-DsRed-Express); tein blots were incubated with 1:1000 dilution of rabbit VSV-DsRed-Express has similar infectivity as the wild- polyclonal anti-AeSCP-2 antibody (Lan and Massey type VSV (Swick et al. 2014). Aag2 cells were seeded in 2004). The goat anti-rabbit horseradish peroxidase con- 12-well plates of 2.5 105 cells/well and cultured for jugated secondary antibody (Jackson ImmunoResearch 24 h until the cells formed 70–90% confluent mono- Laboratory, West Grove, PA) was used at 1:2000 dilu- layers. The cells were treated 24 h before infection with tion. The ECL Western blot substrate (Pierce) was 1 lMSCPI-1,10mM SCPI-1, or DMSO only (mock used to visualize the bound antibodies. The blots were control). Cells were infected at a multiplicity of 10 in placed in a cassette and pressed against the film for 0.5 ml serum-free medium and incubated at 37Cfor 30 min. 1.5 h to allow for adsorption. The inoculum was then Cellular Cholesterol Assay. Aag2 or Huh7.5 cells removed, cells were washed twice with serum-free were seeded in 24-well plates overnight and treated medium, and 0.5 ml/well of medium with 10% FBS with DMSO or SCPI-1 for 48 h as described above. and appropriate concentration of SCPI-1 was added to The cells were washed three times with 1 ml PBS in cells. Infection was allowed to progress in a humidified the well, then were dislodged and transferred to a 1.5- incubator at 28C. At 24 hours after the initial virus ml centrifuge tube and briefly spun at 800g to decant addition, supernatants were removed from cells and the extra PBS. Cellular cholesterol was extracted with stored at 80C. The experiment was conducted with 0.5 ml hexane/isopropanol (hexane:isopropanol ¼ 3:2). full biological duplicates for every sample. The lipid extract and protein pellets were separated by Virus titers from each sample were quantified using centrifugation at 10,000g. The liquid phase was trans- standard plaque assays on BHK monolayers. Virus was ferred to a new tube, air-dried overnight in the fume serially diluted and then adsorbed to cells in six-well hood, and re-dissolved in 0.5 ml reaction buffer (0.1 M plates for 1 h at 37 C and in 5% CO2. After removing potassium phosphate, 0.05 sodium chloride pH ¼ 7.4, inoculum and washing with DPBS, cells were overlaid 5 mM cholic acid, 0.1% Triton X-100). The total choles- with semi-solid MEM media with 2% FBS containing terol concentration was determined using Amplex red 0.6% agar noble. Plates were incubated for 18–20 h cholesterol assay kit (Life technologies) following the and scanned with a GE Typhoon FLA 9000 Biomolec- manufacturer’s instruction. A standard curve was calcu- ular Imager (555/580 nm) under BSL2 conditions. Pla- lated each time based on the cholesterol standards on ques were counted from images of two replicate wells the same 96-well plate with samples. The protein pel- for each sample, and PFU/ml was calculated. lets were dissolved in 0.2 M KOH at 65C for 2 h and Plasmid and Transfection. The coding region of the soluble protein concentration was determined using AeSCP-2 gene was cloned into pIE1hr expression vec- BCA assay (Pierce, Rockford, IL). tors (Lan and Massey 2004). The plasmid was purified Staining for Cellular free Cholesterol and using the Plasmid Maxi Kit (Qiagen, Germantown, Lipid Droplets. Cells were grown on sterile cover- MD). The DNA transfection procedure in Aag2 cells slips in six-well plates overnight, followed by 10 mM was described (Lan and Massey 2004). Briefly, Aag2 SCPI-1 treatment and/or virus infection (MOI ¼ 10). cells were seed at 0.5 ml of 5 105 cells/ml in a 24-well To detect the cellular free cholesterol, the coverslips plate overnight. To prepare the transfection medium, were washed three times with 2 ml PBS and fixed in 16 mg of plasmid DNA and 32 ml Lipofectin were incu- 4% paraformaldehyde for 30 min. The coverslips were bated together in 0.2 ml serum-free medium for washed again with PBS and incubated in 1.5 mg/ml gly- 30 min, and then 1.8 ml of additional serum-free cine solution for 10 min, and then stained with 50 mg/ medium was added. Then, each well of cells was ml Filipin complex in PBS for 1 h. To visualize the lipid September 2015 FUETAL.: SCPI-1 INHIBITS DENGUE VIRUS PRODUCTION IN MOSQUITO CELL 1127 droplets, the cells with the same treatments were fixed in 10% formalin in PBS for 30 min. The coverslips were washed briefly in distilled water and 60% isopropanol, and then stained with filtered 1.8 mg/ml Oil Red O (ORO) in 60% isopropanol for 30 min. The excess ORO was removed by washing several times with distilled water. The nuclei were stained using 300 nM DAPI dihydrochloride (Sigma) in PBS for 10 min. After staining, the coverslips were washed with distilled water and mounted to a glass slide using fluoromount (Sigma). Photos were taken using Leica (Solms, Germany) DMRB fluorescent microscope with N2.1 and I3 filter. Statistical Analyses. GraphPad Prism v4.0 software was used for statistical analyses. P values were obtained from a two-tailed unpaired Student t-test.

Results SCPI-1 Suppresses DENV Viral Production in Mosquito Aag2 Cells. SCPI-1 was one of the Aedes SCP-2 (AeSCP-2) protein-specific inhibitors identified previously, and the affinity of SCPI-1 to AeSCP-2 protein is comparable with that of cholesterol (Kim et al. 2005). The Aag2 cells were pretreated by SCPI-1 and infected by DENV. The result showed that SCPI-1 at 1 mM was able to cause a 10-fold decrease (P ¼ 0.005) in infectious viral particles produced 48 h post-infection (Fig. 1A). More importantly, in the presence of 10 mM of SCPI-1, the infectious viral particle in the medium was below the detection limit of TCID50 assay (Fig. 1A), dropping by more than 6 logs. The suppressed virus production in SCPI-1-treated cells is not due to cytotoxicity of SCPI-1, as the viability of Aag2 cells was not affected by SCPI-1 at 10 mM(Fig.1B). The 50% cytostatic concentration (CC50) was 22.2 mM, Fig. 1. SCPI-1 suppressed DENV production in Aag2 which is similar to its LC50 to mosquito fourth-instar cells. (A) Aag2 cells were pre-treated with SCPI-1 for 24 h larvae (Larson et al. 2008). Order-of-addition experi- before infection. The cells were infected with DENV ments with SCPI-1 demonstrated that exposure of (MOI ¼ 0.5) at 37C for 90 min in serum-free S2 medium. Aag2 cells to SCPI-1 only prior to viral infection was Infection medium was then removed and replaced with not sufficient to inhibit viral replication and production, regular growth medium with SCPI-1. The virus-containing but suggests that SCPI-1 does act on an early step in medium was harvested 48 h post-infection and the infectious the viral life cycle (Supp Fig. 1 [online only]). virus titer was measured. ND: None detected. (B) Cell viability assay. Aag2 cells were seed in 96-well plates and AeSCP-2 Protein is Critical for DENV treated with SCPI-1 for 72 h before assay for viability. The Production in Aag2 Cells. Although SCPI-1 was data represent the average and standard deviations of the able to suppress viral production (Fig. 1A), it was still results of three independent experiments. CC50: 50% uncertain whether its antiviral effect was owing to cytostatic concentration. inhibiting AeSCP-2 protein or other non-target effects. Aag2 cells express SCP-2 endogenously (Lan and Mas- sey 2004), but DENV infection did not up-regulate additional heterogeneous AeSCP-2 did not enhance SCP-2 mRNA level in Aag2 cells (data not shown). To DENV propagation. In the Aag2 cells transfected with validate the role of AeSCP-2 protein in viral infection, siRNA against AeSCP-2 gene, there was a significant we transiently over-expressed or knocked down the decrease (P < 0.001) of viral particle production by AeSCP-2 expression in Aag2 cells, as described (Lan approximately 100-fold (Fig. 2B, siGFP vs siSCP2). and Massay 2004). The over-expression/knock down Indeed, the virus production was affected (Fig. 2B) was verified by examination at the protein level 48 h when the SCP-2 protein level was lowered (Fig. 2A). post transfection (Fig. 2A). Then we examined its Therefore, the results confirmed that AeSCP-2 is an impact on infectious viral particle production. There important host factor in the DENV production in Aag2 was no significant increase (P ¼ 0.62) in the viral par- cells, although over-expression of AeSCP-2 itself did ticles produced in AeSCP-2 over-expressing Aag2 cells not enhance the viral propagation. compared with the siGFP control (Fig. 2B), indicating SCPI-1 has no Effect on VSV Production in that the amount of endogenous AeSCP-2 was sufficient Aag2 Cells. To further investigate whether functional to support DENV propagation in Aag2 cells and SCP-2 is required for production in mosquito host cells 1128 JOURNAL OF MEDICAL ENTOMOLOGY Vol. 52, no. 5

Fig. 3. Effects of SCPI-1 on VSV-DsRed-Express production in Aag2 cells. Aag2 cells were pre-treated with SCPI-1 for 24 h, infected with VSV-DsRed-Express (MOI ¼ 10) at 37C for 90 min in serum-free S2 medium lacking SCPI-1. Infection medium was then removed and replaced with regular growth medium containing SCPI-1. The virus-containing medium was harvested 24 h post- infection and the infectious virus titer was measured.

Human SCP-2 is not Critical for DENV Infection in Human Huh7.5 Cells. SCP-2 is a ubiq- uitous protein found in nearly all mammalian cells Fig. 2. Effects of SCP-2 dysfunction on dengue examined (Gallegos et al. 2001). Mammalian SCP-2 production in Aag2 cells. Aag2 cells were transfected with plays a major role in cholesterol homeostasis through either 40 nM siRNA against SCP-2 or pIE1-SCP-2 plasmid uptake and transhepatocyte transfer (Moncecchi et al. from 24 h before infection (MOI ¼ 1) throughout the 1996, Atshaves et al. 2009). As Human SCP-2 protein experiment. The siRNA against GFP was used as control. (A) (HsSCP-2) shared the similar function as an intracellu- Western blot analysis using polyclonal antibodies to detect lar cholesterol transporter (Krista et al. 2010) as that of SCP-2 protein level after transfection. (B) The virus was mosquito SCP-2 (Lan and Massey 2004), it is of inter- collected 48 h post-infection for titration. The data represent est to investigate if it is also an important human host the average and standard deviations of the results of three factor in DENV production. The HsSCP-2 was independent experiments. siGFP: siRNA against GFP; siSCP-2: siRNA against SCP-2, SCP-OE: SCP-2 over- expressed endogenously in the hepatocytes’ Huh7.5 expression. cells, and predesigned siRNA was able to effectively knock down the HsSCP-2 expression (Fig. 4A). In vitro studies showed that SCPI-1 was an effective inhibitor by other arboviruses, VSV, an enveloped virus in the of HsSCP-2 protein (Kriska et al. 2010). However, nei- Rhabdoviridae family, was used to infect SCPI-1- ther siRNA nor 10 mM SCPI-1 was able to suppress treated Aag2 cells. VSV replicates in mosquito cells DENV production in Huh7.5 cells (Fig. 4B). The CC50 (Phalen and Kielian 1991, Marquardt et al. 1993). VSV- of SCPI-1 was 14.6 mM in Huh7.5 cells (lower than DsRed-Express that has similar infectivity as the wild- insect cells), and the cell viability was not affected at type VSV (Swick et al. 2014) readily replicates in Aag2 10 mM(Fig. 4C). It seemed that HsSCP-2 did not play cells, producing high levels of infectious virus particles the same role in DENV production in Huh7.5 cells as 24 hours post-infection (Fig. 3, DMSO control). Fur- that of the AeSCP-2 in mosquito Aag2 cells (Fig. 2B), ther, SCPI-1 treatments had minimal effects on VSV despite having cellular toxicity at a lower drug production in Aag2 cells, with merely a 3% reduction concentration. in virus titer at 10 lM SCPI-1 concentration (Fig. 3). SCPI-1 Treatment Reduced Cellular These results suggest that VSV production in Aag2 Cholesterol Levels in Both Human and Mosquito cells does not require functional AeSCP-2, which is in Cells. Under normal growth conditions, cells can sharp contrast to DENV-2 (Fig. 1A) and provides acquire cholesterol in the cholesterol-rich medium further evidence that 10 lMSCPI-1isnottoxictomos- (300 mg/ml in FBS according to the manufacturer). quito cells generally but rather specifically blocks SCP-2 protein facilitates the cholesterol uptake (Mon- DENV-2 production. These results are also consistent cecchi et al. 1996, Lan and Massey 2004). SCPI-1 with previous studies that show that manipulating cellu- treatment reduced [3H] cholesterol incorporation in lar cholesterol had little effect on VSV production in Aag2 cells cultured in sterol-free serum medium (Kim mosquito cells (Phalen and Kielian, 1991, Marquardt et al. 2005). Here we performed the experiment under et al. 1993). normal growth conditions and measured the total September 2015 FUETAL.: SCPI-1 INHIBITS DENGUE VIRUS PRODUCTION IN MOSQUITO CELL 1129

Fig. 5. The effect of SCPI-1 treatment on levels of total cellular cholesterol. Aag2 or Huhu7.5 cells were seeded in 24-well plates overnight, and then were treated with DMSO (solvent for SCPI-1) or SCPI-1 for 48 h in normal growth medium. The total cellular cholesterol contents were determined and were normalized to soluble cellular protein quantity (mg/mg of protein). The data represent the average and standard deviations of six individual wells.

Aag2 cells (Fig. 2BvsFig. 4B). It was also noticed that 1 mM SCPI-1 was able to suppress DENV propagation in Aag2 cells (Fig.1A), whereas the total cellular cholesterol content was not affected at the same concentration of treatment (Fig. 5). The results suggest that lowering total cellular cholesterol alone may not be the underlying mechanism of SCP-2-dependent DENV propagation in Aag2 cells. SCPI-1 Treatment Altered the Intracellular Cholesterol Distribution Only in Aag2 Cells. To investigate the mechanism in which SCPI-1 affects DENV production in mosquito host cells, we closely examined the distribution of cellular cholesterol in Fig. 4. Effects of SCP-2 dysfunction on dengue Aag2 cells. Filipin stain has been widely used to localize production in human Huh7.5 cells. (A) Treatment of siRNA on HsSCP-2 expression (triplicates of independent samples). free cholesterol in various cells and tissues, as it does (B) SCP-2 dysfunction and DENV production. (C) Cell not recognize esterified cholesterol (Rudolf and Curcio viability of Huh7.5 cells treated with SCPI-1. Huh7.5 cells 2009, Gimpl 2010). The infected or non-infected Aag2 were treated with either 40 nM siRNA or 10 mM SCPI-1 from cells were treated with either DMSO or 10 mMSCPI-1 24 h before infection throughout the experiment (MOI ¼ 1). for 48 h before fixing and staining. In the DMSO- The mRNA level of HsSCP-2 was measured via RT-PCR 24 h treated control Aag2 cells, the plasma membrane and after siRNA transfection. The virus titration and viability numerous perinuclear free cholesterol-rich compart- assay were conducted as described (See Methods). The data ments were stained with Filipin (Fig.6A). When the represent the average and standard deviations of the results cells were infected with DENV, the perinuclear free of three independent experiments. CC50: 50% cytostatic concentration. cholesterol-rich compartments were still visualized with bright Filipin staining (Fig. 6C), similar to non-infected cells (Fig. 6A vs C). The result indicates that DENV cellular cholesterol levels. When the SCP-2 was inhib- infection in Aag2 cells did not significantly alter the ited by 10 mM SCPI-1 treatment, there is a 24% intercellular free cholesterol distribution. Interestingly, decrease (P < 0.001) of the total cholesterol content in when treated with 10 mM SCPI-1, the intensity of Fili- Aag2 cells and a 15% decrease (P < 0.001) in Huh7.5 pin staining in the perinuclear free cholesterol-rich cells (Fig. 5). Total cellular cholesterol levels were simi- compartments was significantly reduced, no matter lar in both cell lines (Fig. 5, DMSO). Although SCPI-1 whether the cells were infected with DENV or not was able to significantly deplete cellular cholesterol lev- (Fig. 6B and D). The result suggests that SCP-2 may els in both cell lines, it only affected viral production in be involved in cellular cholesterol homeostasis via its 1130 JOURNAL OF MEDICAL ENTOMOLOGY Vol. 52, no. 5

Fig. 6. The intracellular free cholesterol distribution in Aag2 cells. The Aag2 cells were infected with 10 MOI of DENV (C, D) and/or treated with DMSO (A, C, E) or 10 mM SCPI-1 (B, D, F) for 48 h. Filipin staining for free cholesterol (A-D) and Oil Red O (ORO) staining for neutral lipids (E, F) are shown. The arrows indicate the perinuclear compartments. N: nuclei. cholesterol transport function. As siRNA against SCP-2 previously (Samsa et al. 2009). To investigate whether treatment suppressed DENV production in Aag2 cells the perinuclear free cholesterol-rich compartments are (Fig. 2B), we speculate that SCP-2-mediated intracellu- lipid droplets, we stained the cells using ORO to local- lar cholesterol trafficking is critical for DENV propaga- ize the lipid droplets (Rudolf and Curcio 2009). The tion in mosquito host cells. results showed that lipid droplets in Aag2 cells were The perinuclear free cholesterol-rich compartments perinuclear but were not as numerous as the perinu- have been reported in Ae. albopictus C6/C36 cells clear free cholesterol-rich compartments (Fig. 6Evs (Rolls et al. 1997). The nature of those free cholesterol- A). On the other hand, SCPI-1 treatment did not sig- rich compartments is unknown, although they nificantly affect the intensity of ORO staining of the resembled the lipid droplet structures reported lipid droplets (Fig. 6E vs F). The result is consistent September 2015 FUETAL.: SCPI-1 INHIBITS DENGUE VIRUS PRODUCTION IN MOSQUITO CELL 1131

Fig. 7. The intracellular free cholesterol distribution in Huh7.5 cells. The Huh7.5 cells were treated with DMSO (A and C) or 10 mM SCPI-1 (B and D) for 48 h. The Filipin staining (A and B) and ORO staining (C and D) were performed as described (See Methods). N: nuclei. with the previous finding that the composition of lipid Results from our study suggested that intracellular cho- droplets are mainly neutral lipids including sterol esters lesterol distribution was not altered by SCPI-1 in (Walther and Farese 2012), which would not be stained Huh7.5 cells, although high concentration of SCPI-1 by Filipin (Rudolf and Curcio 2009, Gimpl 2010). Our did reduce cellular cholesterol levels in Huh7.5 cells results indicated that the perinuclear free cholesterol- (Fig. 5B). rich structures did not have the same composition as that of the lipid droplets. Moreover, inhibition of SCP- Discussion 2 function had a more pronounced effect on free cholesterol content in perinuclear free cholesterol-rich Production of enveloped viruses including DENV in compartments than that of the esterified lipids in the mammalian cells involves both de novo cholesterol syn- lipid droplet. Our study suggested that SCPI-1 treat- thesis (Mackenzie et al. 2007, Rothwell et al. 2009)and ment somehow disrupted the storage of free choles- intracellular transport of cholesterol (Poh et al. 2012). terol in the perinuclear compartments or the Previously though, DENV production in mosquito cells homeostasis of cellular cholesterol distribution. has been characterized as cholesterol-independent rela- We also examined the free cholesterol distribution tive to alphaviruses (Umashankar 2008). Here we show and lipid droplets in non-infected human Huh7.5 cells. that disruption of the cholesterol trafficking pathway Interestingly, perinuclear free cholesterol-rich compart- mediated by SCP-2 in mosquito, but not human, cells ments were not present in Huh7.5 cells (Fig.7AvsFig. lowers DENV production. Mosquito host cells are aux- 6A). SCPI-1 treatment did not significantly affect the otrophic for cholesterol, which provide us an optimal free cholesterol distribution (Fig.7BvsA).Therewere model to study the role of intracellular cholesterol traf- substantial amounts of lipid droplets in Huh7.5 cells, ficking for DENV production. This highlights an inter- but their lipid contents were not significantly affected esting and unanswered question of whether males and by SCPI-1 treatment, as the ORO staining intensities pre-bloodmeal females are less competent for DENV were similar under both conditions (Fig.7CvsD). production in part because cholesterol is limited, or 1132 JOURNAL OF MEDICAL ENTOMOLOGY Vol. 52, no. 5 perhaps a plant sterol can suffice (Martins et al. 2012, infection (Heaton and Randall 2010, Pena and Harris Buckneretal.2013). We focused on the function of 2012). Over-expression of SCP-2 in mouse intestine L mosquito SCP-2, given that the cholesterol transport cells alters the lipids and cholesterol composition of function of SCP-2 has been demonstrated in Aag2 cells lipid droplets (Atshaves et al. 2001). The subcellular (Lan and Massey 2004, Radek et al. 2010)andinAe. localization of human SCP-2 and mosquito SCP-2 is aegypti mosquito (Blitzer et al. 2005, Dyer et al. 2008). quite different, even though AeSCP-2 and HsSCP-2 Our study is the first report of a sterol transport pro- share 31% identity in amino acid sequences (Krebs and tein, AeSCP-2, as a mosquito host factor for DENV Lan 2003). In Aag2 cells, over 90% of the SCP-2 pro- propagation in mosquitoes. When the SCP-2 function tein was localized to the cytosol (Lan and Massey was inhibited, production of DENV viral particles was 2004), whereas the mammalian SCP-2 localizes largely impaired, while that of a different enveloped RNA (>50%) to the peroxisome, ER, and mitochondria (Gal- virus was not. legosetal.2001). Therefore, the HsSCP-2 trafficking HsSCP-2 protein does not appear to be an essential pathway might not be critical for the viral infection in host factor for DENV propagation in human cells. To human host cells. Rather, the release of free cholesterol understand the difference between Aag2 cells and from late endosomes mediated via the NPC transport Huh7.5, we examined the total cellular cholesterol level pathways may play a more important role in human after SCPI-1 treatment in both. The total cholesterol cells (Poh et al. 2012). was significantly reduced by 10 mMSCPI-1inboth Stored cellular cholesterol and/or cholesterol traffick- types of cells. SCPI-1 was less effective in reducing ing may be more important in DENV production in mammalian Huh7.5 cellular cholesterol level, possibly mosquito than human cells, as mosquitoes lack de novo owing to its ability to synthesize cholesterol de novo. cholesterol synthesis. Interestingly, about 80–90% of However, cellular cholesterol depletion as the result of cellular cholesterol in Aag2 cells is unesterified choles- SCPI-1 treatment could not fully explain the antiviral terol (data not shown) and there are not many lipid effect in mosquito host cells. First of all, 1 mM SCPI-1 droplets/cell (Fig.6E). The majority of cellular free cho- did not change the total cellular cholesterol level, but it lesterol was stored in the perinuclear compartment, as did affect DENV production. Second, DENV propaga- reported in mosquito C6/36 cells (Rolls et al. 1997). tion was not hampered in Huh7.5 cells, despite Possibly, the mosquito SCP-2 protein is involved in the total cholesterol being reduced by 15%. Third, pre- loading the cells with free cholesterol to the perinu- infection treatment with SCPI-1 did not affect DENV clear compartment, which is essential for DENV pro- production as much as that of post-infection treatment duction in mosquito host cells. Whether DENV in mosquito host cells, which is different from the cho- replication complex localizes to the perinuclear free lesterol depletion drug methyl-beta-cyclodextrin that is cholesterol-rich compartments needs further investiga- most effective in suppressing DENV infection by pre- tion. Although SCP-2 has moderate affinity toward fatty treating mammalian cells (Lee et al. 2008, Medigeshi acid (Gallegos et al. 2001, Vyazunova et al. 2007), et al. 2008). Therefore, the antiviral effect of SCPI-1 SCPI-1 specifically impacted cellular free cholesterol may not be solely due to cellular cholesterol depletion. distribution in Aag2 cells, because lipid contents of SCPI-1 treatment altered the cellular free cholesterol lipid droplets were not significantly affected by SCPI-1 distribution in either infected or non-infected Aag2 treatment. Therefore, our studies support other data cells but not in Huh7.5 cells. The results suggest that that DENV replication in mosquito and mammalian in mosquito host cells, the SCP-2-mediated cholesterol cells have different requirements. trafficking pathway is critical for DENV production. There are many differences in viral replication in mosquito cells versus human cells (Vasilakis et al. 2009), Supplementary Data but to our knowledge, ours is the first report showing that disrupting cholesterol pathways in mosquito cells Supplementary data are available at Journal of Medical limits DENV. While both siRNA and small molecule Entomology online. inhibitors can have cascading or secondary effects, here both approaches support that cholesterol pathways are important for DENV replication in mosquito cells. Acknowledgments We speculate that DENV utilizes different cholesterol trafficking pathways in two different host cells. In We thank Dr. Jorge Osorio for kindly providing DENV mammalian cells, inhibition of NPC-mediated intracel- New Guinea C virus and Dr. Israr H. Ansari for technical lular transport pathway leads to suppression of DENV assistance. We thank Dr. David Anstrom for critical com- propagation during initial infection as well as post- ments on the manuscripts. This study was funded by the graduate school of University of Wisconsin-Madison awards infection (Poh et al. 2012). Moreover, studies found 135-PRJ19QR and 135-PRJ61TK to QL. that the cholesterol containing lipid droplets were co- localized with DENV viral protein in BHK cells (Samsa et al. 2009). A more recent study using Huh7.5 cells References Cited revealed that the amount of lipid droplets remained Acosta, E. G., V. Castilla, and E. B. Damonte. 2008. Func- the same post-DENV infection, while the total lipid tional entry of dengue virus into Aedes albopictus mosquito droplet volume decreased (Heaton and Randall 2010), cells is dependent on clathrin-mediated endocytosis. J. Gen. indicating lipid droplets were consumed during viral Virol. 89: 474–484. September 2015 FUETAL.: SCPI-1 INHIBITS DENGUE VIRUS PRODUCTION IN MOSQUITO CELL 1133

Aizaki, H., M. Fukasawa, H. Toni, Y. Matsuura, K. Hanada, Hase, T., P. L. Summers, and K. H. Eckels. 1989. M.Lai,T.Miyamura,T.Wakita,andT.Suzuki.2007. Flavivirus entry into cultured mosquito cells and human The cholesterol and sphingolipids of hepatitis C virus parti- peripheral blood monocytes. Arch. Virol. 104: 129–143. cles play critical roles in the viral infectivity. Hepatology 46: Heaton, N. S., and G. Randall. 2010. Dengue virus-induced 453A–454A. autophagy regulates lipid metabolism. Cell Host Microbe 8: Atshaves,B.P.,S.M.Storey,A.L.McIntosh,A.D.Pet- 422–432. rescu, O. I. Lyuksyutova, A. S. Greenberg, and F. Kim,M.,V.Wessely,andQ.Lan.2005.Identification of Schroeder. 2001. Sterol carrier protein-2 expression modu- mosquito sterol carrier protein-2 inhibitors. J. Lipid Res. 46: lates protein and lipid composition of lipid droplets. J. Biol. 650–657. Chem. 276: 25324–25335. Krebs, K., and Q. Lan. 2003. Isolation and expression of a ste- Atshaves,B.,A.McIntosh,G.Martin,D.Landrock,H. rol carrier protein-2 gene from the yellow fever mosquito, Payne, S. Bhuvanendran, K. Landrock, O. Lyuksyutova, Aedes aegypti. Insect Mol. Biol. 12: 51–60. J. Johnson, R. Macfarlane, et al. 2009. Overexpression Kriska, T., A. Pilat, J. C. Schmitt, A. W. Girotti. 2010. Sterol of sterol carrier protein-2 differentially alters hepatic choles- carrier protein-2 (SCP-2) involvement in cholesterol hydro- terol accumulation in cholesterol-fedmice.J.Lipid.Res.50: peroxide cytotoxicity as revealed by SCP-2 inhibitor effects. 1429–1447. J Lipid Res. 51: 3174–3184. Blitzer,E.,I.Vyazunova,andQ.Lan.2005.Functional analysis Kuhn, R. J., W. Zhang, M. G. Rossmann, S. V. Pletnev, J. of AeSCP-2 using gene expression knockdown in the yellow Corver,E.Lenches,C.T.Jones,S.Mukhopadhyay,P.R. fever mosquito, Aedes aegypti. Insect Mol. Biol. 14: 301–307. Chipman, E. G. Strauss, T. S. Baker, and J. H. Strauss. Buckner,E.A.,B.W.Alto,andL.P.Lounibos.2013.Verti- 2002. Structure of Dengue Virus: Implications for Flavivirus cal transmission of key west dengue-1 virus by Aedes aegypti Organization, Maturation, and Fusion. Cell. 108: 717–725. and Aedes albopictus (Diptera: Culicidae) mosquitoes from Kyle,J.,andE.Harris.2008.Global spread and persistence of Florida. J. Med. Entomol. 50: 1291–1297. dengue. Annu. Rev. Microbiol. 62: 71–92. Che, P., L. Wang, and Q. Li. 2009. The development, optimi- Lan, Q., and R. Massey. 2004. Subcellular localization of the zation and validation of an assay for high throughput antiviral mosquito sterol carrier protein-2 and sterol carrier protein-x. drug screening against Dengue virus. Int. J. Clin. Exp. Med. J. Lipid. Res. 45: 1468–1474. 2: 363–373. Larson,R.,V.Wessely,Z.Jiang,andQ.Lan.2008.Larvicidal Chu, J. J., and M. L. Ng. 2004. Infectious entry of West Nile activity of sterol carrier protein-2 inhibitor in four species of virus occurs through a clathrin-mediated endocytic pathway. mosquitoes. J. Med. Entomol. 45: 439–444. J. Virol. 78: 10543–10555. Lee,C.,H.Lin,C.Liao,andY.Lin.2008.Cholesterol effec- Chu, J., and P. Yang. 2007. c-Src protein kinase inhibitors tively blocks entry of flavivirus. J. Virol. 82: 6470–6480. block assembly and maturation of dengue virus. PNAS 104: Lindenbach,B.D.,andC.M.Rice.2003.Molecular biology 3520–3525. of flaviviruses. Adv. Virus Res. 59: 23–61. Clayton, R. B. 1964. The utilization of sterols by insects. J. Lipid Liscum,L.,andJ.R.Faust.1989.The intracellular transport Res. 5: 3–19. of low density lipoprotein-derived cholesterol is inhibited in Dyer,D.,V.Wessely,K.Forest,andQ.Lan.2008.Three- Chinese hamster ovary cells cultured with 3-beta-[2- dimensional structure/function analysis of SCP-2-like2 (diethylamino)ethoxy]androst-5-en-17-one. J. Biol. Chem. reveals differences among SCP-2 family members. J. Lipid. 264: 11796–11806. Res. 49: 644–653. Mackenzie, J. M., and E. G. Westaway. 2001. Assembly and Endo,A.1992.The discovery and development of HMG-CoA maturation of the flavivirus Kunjin virus appear to occur in reductase inhibitors. J. Lipid. Res. 33: 1569–1582. the rough endoplasmic reticulum and along the secretory Fallon,A.M.,andD.Sun.2001.Exploration of mosquito im- pathway, respectively. J. Virol. 10787–10799. munity using cells in culture. Insect Biochem. Mol. Biol. 31: Mackenzie,J.,A.Khromykh,andR.Parton.2007.Choles- 263–78. terol manipulation by West Nile virus perturbs the cellular Fernandez-Garcia, M. D., M. Mazzon, M. Jacobs, and A. immune response. Cell Host Microbe. 2: 229–239. Amara. 2009. Pathogenesis of flavivirus infections, using and Marquardt, M. T., T. Phalen, and M. Kielian. 1993. Choles- abusing the host cell. Cell Host Microbe. 5: 318–328. terol is required in the exit pathway of Semliki Forest virus. J. Gallegos,A.,B.Atshaves,S.Storey,O.Starodub,A.Pet- Cell. Biol. 123: 57–65. rescu, H. Huang, A. McIntosh, G. Martin, H. Chao, A. Martinez-Gutierrez, M., J. Castellanos, and J. Gallego- Kier, et al. 2001. Gene structure, intracellular localization, Gomez. 2011. Statins reduce dengue virus production via and functional roles of sterol carrier protein-2. Prog. Lipid. decreased virion assembly. Intervirology 54: 202–216. Res. 40: 498–563. Martins,V.E.P.,C.H.Alencar,M.T.Kamimura,F.M.de Gillespie, L. K., A. Hoenen, G. Morgan, and J. M. Macken- Carvalho Arau´jo,S.G.DeSimone,R.F.Dutra,andM. zie. 2010. The Endoplasmic Reticulum Provides the Mem- I. Guedes. 2012 Occurrence of Natural Vertical Transmis- brane Platform for Biogenesis of the Flavivirus Replication sion of Dengue-2 and Dengue-3 Viruses in Aedes aegypti and Complex. J Virol. 84: 10438–10447. Aedes albopictus in Fortaleza, Ceara´, Brazil. PLoS ONE 7: Gimpl, G. H. Jr. 2010. Cholesterol Protein Interaction: Meth- e41386 ods and Cholesterol Reporter Molecules. Subcell. Biochem. Medigeshi, G., A. Hirsch, D. Streblow, J. Nikolich-Zugich, 51: 1–45. and J. Nelson. 2008. West nile virus entry requires choles- Ha¨rma¨la¨,A.S.,M.I.Po¨rn, P. Mattjus, and J. P. Slotte. terol-rich membrane microdomains and is independent of al- 1994. Cholesterol transport from plasma membranes to pha v beta 3 integrin. J. Virol. 82: 5212–5219. intracellular membranes is inhibited by 3 beta- Moncecchi,D.,E.Murphy,D.Prows,andF.Schroeder. [2-(diethylamino)ethoxy]androst-5-en-17-one. Biochim. Bio- 1996. Sterol carrier protein-2 expression in mouse L-cell fi- phys. Acta 1211: 317–325. broblasts alters cholesterol uptake. Biochim. Biophy. Acta Hase,T.,P.L.Summers,K.H.Eckels,andW.B.Baze. Lipids Lipid Metabol. 1302: 110–116. 1987. An electron and immune- electron microscopic study Nishiura,H.,andS.Halstead.2007.Natural history of dengue of dengue-2 virus infection of cultured mosquito cells: Matu- virus (DENV)-1 and DENV-4 infections: Reanalysis of classic ration events. Arch. Virol. 92: 273–291. studies. J. Infect. Dis. 195: 1007–1013. 1134 JOURNAL OF MEDICAL ENTOMOLOGY Vol. 52, no. 5

Pastorino,B.,A.Nougaire`de, N. Wurtz, E. Gould, and X. andA.V.Gamarnik.2009.Dengue virus capsid protein de Lamballerie. 2010. Role of host cell factors in flavivirus usurps lipid droplets for viral particle formation. PLoS infection: Implications for pathogenesis and development of Pathog. 5: e1000632. antiviral drugs. Antiviral Res. 87: 281–294. Sessions, O., N. Barrows, J. Souza-Neto, T. Robinson, C. Perera,R.,C.Riley,G.Isaac,A.S.Hopf-Jannasch,R.J. Hershey, M. Rodgers, J. Ramirez, G. Dimopoulos, P. Moore, K. W. Weitz, L. Pasa-Tolic, T. O. Metz, J. Ada- Yang,J.Pearson,etal.2009.Discovery of insect and hu- mec, and R. J. Kuhn. 2012. Dengue Virus Infection Per- man dengue virus host factors. Nature 458: 1047–1050. turbs Lipid Homeostasis in Infected Mosquito Cells. PLoS Swick,A.,A.Baltes,andJ.Yin.2014.Visualizing infection Pathog 8: e1002584. spread: Dual-color fluorescent reporting of virus–host inter- Pena,J.,andE.Harris.2012.Early dengue virus protein syn- actions. Biotech. Bioeng. 111: 1200–1209. thesis induces extensive rearrangement of the endoplasmic Tchankouo-Nguetcheu, S., H. Khun, L. Pincet, P. Roux, M. reticulum independent of the UPR and SREBP-2 Pathway. Bahut,M.Huerre,C.Guette,andV.Choumet.2010. PLoS ONE 7: e38202. Differential Protein Modulation in Midguts of Aedes aegypti Peng, R., V. Maklokova, J. Chandrashekhar, and Q. Lan. Infected with Chikungunya and Dengue 2 Viruses. PLoS 2011. In vivo functional genomic studies of sterol carrier pro- ONE 5: e13149. tein-2 gene in the yellow fever mosquito. PLoS ONE 6: Umashankar, M., C. Martin, M. Liao, B. Reilly, A. Guo, G. e18030. Taylor, and M. Kielian. 2008. Differential cholesterol Peng, R., Q. Fu, H. Hong, T. Schwaegler, and Q. Lan. 2012. binding by class II fusion proteins determines membrane fu- THAP and ATF-2 regulated sterol carrier protein-2 promoter sion properties. J. Virol. 82: 9245–9253. activities in the larval midgut of the yellow fever mosquito, Valle,J.,S.Chavez-Salinas,F.Medina,andR.delAngel. Aedes aegypti. PLoS ONE 7: e46948. 2005. Heat shock protein 90 and heat shock protein 70 are Phalen, T., and M. Kielian. 1991. Cholesterol is required components of dengue virus receptor complex in human for infection by Semliki Forest virus. J. Cell Biol. 112: cells. J. Virol. 79: 4557–4567. 615–623. Vancini, R., L. D. Kramer, M. Ribeiro, R. Hernandez, D. Poh, M. K., G. Shui, X. Xie, P. Y. Shi, M. R. Wenk, and F. Brown. 2013. Flavivirus infection from mosquitoes in vitro Gu. 2012. U18666A, an intra-cellular cholesterol transport reveals cell entry at the plasma membrane. Virology 435: inhibitor, inhibits dengue virus entry and replication. Antiviral 406–414. Res. 93: 191–198. van der Schaar, H. M., M. J. Rust, C. Chen, H. van der Radek,J.T.,D.H.Dyer,andQ.Lan.2010.Effects of muta- Ende-Metselaar,J.Wilschut,X.Zhuang,andJ.M.Smit. tions in Aedes aegypti sterol carrier protein-2 on the biological 2008. Dissecting the cell entry pathway of dengue virus by function of the protein. Biochem 49:7532–7541. single-particle tracking in living cells. PLoS Pathog. 4: Rawson, R. 2003. The SREBP pathway - Insights from insigs e1000244. and insects. Nat. Rev. Mol. Cell Biol. 4: 631–640. Vasilakis,N.,E.R.Deardorff,J.Kenney,S.L.Rossi,K.A. Reed,L.,andL.Muench.1938.A simple method of estimat- Hanley, and S. C. Weaver. 2009. Mosquitoes put the brake ing fifty percent endpoints. Am. J. Hyg. 27: 493–497. on evolution: Experimental evolution reveals slower mutation Rodenhuis-Zybert,I.A.,H.M.vanderSchaar,J.M.da accumulation in mosquito cells than vertebrate cells. PLoS SilvaVoorham,H.vanderEnde-Metselaar,H.-Y.Lei, Pathog. 5: e1000467 J. Wilschut, and J. M. Smit. 2010. Immature dengue virus: Vyazunova,I.,V.Wessley,M.Kim,andQ.Lan.2007. A veiled pathogen? PLoS Pathog. 6: e1000718. Identification of two sterol carrier protein-2 like genes in the Rolls, M. M., M. T. Marquardt, M. Kielian, and C. E. yellow fever mosquito, Aedes aegypti. Insect Mol. Biol. 16: Machamer. 1997. Cholesterol-independent targeting of 305–314. Golgi membrane proteins in insect cells. Mol. Biol. Cell 8: Walther, T. C., and R. V. Farese. 2012. Lipid droplets and cel- 2111–2118. lular lipid metabolism. Annu. Rev. Biochem. 81: 687–714. Rothwell,C.,A.LeBreton,C.Ng,J.Lim,W.Liu,S.Vasude- Walther,T.,R.Farese,andR.Kornberg.2012.Lipid drop- van, M. Labow, F. Gu, and L. Gaither. 2009. Cholesterol lets and cellular lipid metabolism. Annu. Rev. Biochem. 81: biosynthesis modulation regulates dengue viral replication. 687–714. Virology 389: 8–19. Welsch, S., S. Miller, I. Romero-Brey, A. Merz, C. K. Bleck, Rudolf,M.,andC.A.Curcio.2009.Esterified cholesterol is P. Walther, S. D. Fuller, C. Antony, J. Krijnse-Locker, highly localized to Bruch’s membrane, as revealed by lipid and R. Bartenschlager. 2009. Composition and three- histochemistry in wholemounts of human choroid. J. Histo- dimensional architecture of the dengue virus replication and chem. Cytochem. 57: 731–739. assembly sites. Cell Host Microbe. 5: 365–375. Samsa,M.M.,J.A.Mondotte,N.G.Iglesias,I.Assunc¸a˜o- Miranda, G. Barbosa-Lima, A. T. Da Poian, P. T. Bozza, Received 1 October 2014; accepted 24 June 2015.