Venezuelan Equine Encephalitis virus infection of mosquito cells requires acidification as well as mosquito homologs of the endocytic proteins Rab5 and Rab7

Abstract

Venezuelan equine encephalitis virus (VEEV) is a New World alphavirus that can cause fatal encephalitis in humans. It remains a naturally emerging disease as well as a highly developed biological weapon. VEEV is transmitted to humans in nature by mosquito vectors. Little is known about VEEV entry, especially in mosquito cells. Here, a novel luciferase-based virus entry assay is used to show that the entry of VEEV into mosquito cells requires acidification. Furthermore, mosquito homologs of key human proteins (Rab5 and Rab7) involved in endocytosis were isolated and characterized. Rab5 is found on early endosomes and Rab7 on late endosomes and both are important for VEEV entry in mammalian cells. Each was shown to have analogous function in mosquito cells to that seen in mammalian cells. The wild-type, dominant negative and constitutively active mutants were then used to demonstrate that VEEV requires passage through early and late endosomes before infection can take place. This work indicates that the infection mechanism in mosquitoes and mammals is through a common and ancient evolutionarily conserved pathway.

Introduction

Alphaviruses are classified as either New World or Old World viruses, based on geographic origin. Only the New World alphaviruses have the ability to cause fatal encephalitis in humans. Venezuelan equine encephalitis virus (VEEV) is the most important human pathogen among the New World alphaviruses and it remains a naturally emerging disease as well as a highly developed biological weapon (Weaver et al., 2004). Hundreds of thousands of people have been involved in sporadic outbreaks of febrile and neurological disease caused by VEEV since 1938 (Weaver et al., 2004). There is currently no publicly available vaccine for VEEV and the experimental military vaccine has poor efficacy (Russell, 1999). In addition, no specific antiviral regimens are available for treatment of VEEV disease and the broad spectrum antiviral Ribavirin is ineffective (Canonico et al., 1984). To develop successful treatments for VEEV infection, a better understanding of the pathogenesis of this virus is necessary.

VEEV particles, like those of other alphaviruses, have icosahedral symmetry and are surrounded by a lipid bilayer. Embedded in this bilayer are two viral glycoproteins, E1 and E2. These envelope proteins (envs) are responsible for viral attachment to a susceptible host cell receptor and fusion with host cell membranes. After membrane fusion, the viral RNA core can be released into the cellular cytoplasm (Weaver et al., 2004). In mammalian cells, alphavirus membrane fusion is generally thought to be triggered by acidification during endocytic uptake of the virus (Marsh, Kielian, and Helenius, 1984). The mechanism of entry that a virus uses to gain access to the cell is an important aspect of pathogenesis but little is known about the VEEV entry process.

It has been well established that the Old World alphavirus Semliki Forest Virus (SFV) enters mammalian cells by receptor-mediated, clathrin-dependent endocytosis (Marsh, Kielian, and Helenius, 1984; Sieczkarski and Whittaker, 2003). Upon binding a cellular receptor, SFV is internalized in clathrin-coated pits then routed through endosomes along the endocytic trafficking pathway. Inhibition of endosomal acidification by chemical inhibitors blocks SFV nucleocapsid release into the cellular cytoplasm (Helenius, Marsh, and White, 1982). It has been shown that SFV requires a pH of 6.2 for fusion with a cellular membrane, which likely occurs in the early endosome (Sieczkarski and Whittaker, 2003). In contrast, work with another Old World alphavirus, Sindbis virus (SIN), indicated productive infection without the need for a pH-dependent endocytic pathway (Hernandez, Luo, and Brown, 2001; Paredes et al., 2004). This disparity in entry pathways seen within Old World alphaviruses indicates that multiple entry pathways may be used and the entry pathway of New World alphaviruses, such as VEEV, cannot be inferred based on experiments done with Old World alphaviruses.

VEEV is maintained in nature in a rodent reservoir and is transmitted to humans by mosquito vectors. Mosquito salivary glands need to be infected with virus before transmission can take place (Weaver et al., 2004). As for mammalian cells, little is known about the pathogenesis of this virus in the mosquito. Again, there is controversy as to the entry mechanism and pathway used to infect mosquito cells. It was recently reported that exposure to low pH was not a requirement for the RNA of SIN to penetrate into the cytoplasm of mosquito cells (Hernandez, Luo, and Brown, 2001). In this case, it remains unclear if an endocytic route was followed or not. The entry route followed by VEEV to infect mosquitoes is also unknown. A thorough understanding of the virus entry process could lead to new targets that may help break the human/vector infection cycle.

Recent insights into endosomal trafficking mechanisms have provided new tools for dissecting entry pathways taken by natural ligands and pathogens such as viruses. To drive endosomal movement and maturation, specific proteins are added and removed on the cytoplasmic face of the vesicle throughout the trafficking process. The GTPases known as Rab proteins are known to be crucial for regulating all intracellular membrane transport activities as well as generating and maintaining vesicles along the endocytic pathway (Bucci et al., 1992; Zerial and Stenmark, 1993). Two well-characterized endosomal proteins that have been used to follow endocytic trafficking as well as disrupt the endocytic pathway are the small GTPases Rab5 and Rab7. Rab5 plays a key role in the trafficking of ligands from the cell surface into the cell via early endosomes. Rab7 regulates transport from the early endosome to the late endosome and is involved in the biogenesis of lysosomes (Bucci et al., 1992; Bucci et al., 2000). Consequently, Rab5 is located on early endosomes and Rab7 can be found on late endosomes.

Dominant-negative mutants of both proteins, as well as a constitutively active form of Rab5, have been widely used to inhibit various steps along the endocytic trafficking pathway in mammalian cells (Galperin and Sorkin, 2003; Li et al., 1994). These proteins have also been used to inhibit viral infection for viruses that utilize an endocytic pathway to gain entry into cells. Rauma et al showed that adenovirus uptake is increased when wild-type Rab5 is overexpressed and decreased upon expression of the dominant negative form of Rab5 (Rauma et al., 1999). Many other viruses, including influenza, have been shown to require both Rab5 and Rab7 for infection using the wild-type and dominant negative forms of the proteins in mammalian cells (Sieczkarski and Whittaker, 2002b). In our previous work it was shown that in mammalian cells, VEEV requires both Rab5 and Rab7 for productive entry and infection. This suggests that the virus needs to travel through both early and late endosomes before fusion can occur and the viral genome can get released into the cytoplasm (Kolokoltsov, Fleming, and Davey, 2006).

It was previously unknown if mosquito cells contain functional Rab5 or Rab7 counterparts to those in mammalian cells. Furthermore, as indicated above, it is unclear if alphaviruses use an endocytic pathway to enter and infect mosquito cells or which proteins may be involved. Recently, the genomes for Aedes aegypti and Anopheles gambiae mosquitoes were sequenced and made available (Holt et al., 2002; Kaufman, Severson, and Robinson, 2002; Loftus, 2005). Here, we have identified and isolated mosquito homologs of human Rab5 and Rab7. Both genes share greater than 80% amino acid identity with those found in mammalian cells. We characterized each protein and show that they function analogously to human Rab5 and Rab7. Furthermore, their dominant negative (DN) and constitutively active (CA) mutants give phenotypes similar to the mammalian proteins. Using these, as well as the human genes, the proteins were expressed in mosquito cells and their effects on VEEV entry and infection were studied. We find that VEEV enters mosquito cells through a pH-dependent endocytic pathway requiring both Rab5 and Rab7 and that acidification of the endosome is required for productive infection to take place.

Results

VEEV enters and infects cells through a pH-dependent mechanism

It was previously shown that SFV and VEEV enter mammalian cells through a pH-dependent endocytic pathway (Helenius, Marsh, and White, 1982). However, SIN, another alphavirus, was shown to infect both mammalian and mosquito cells at neutral pH (Paredes et al., 2004). It remains unclear why such a difference exists but could be explained by experimental design differences or differences in entry mechanism of each alphavirus. This led us to first determine if VEEV, a New World alphavirus, enters mosquito cells via a pH-dependent endocytic pathway. Monensin, chloroquine and ammonium chloride were used to treat mosquito cells and then the cells were challenged with virus. Monensin is a cationic ionophore which binds potassium ions and uncouples the sodium/potassium gradient across endosomal membranes. This ion gradient supplies the energy required to pump protons into the endosomes. Ammonium chloride and chloroquine are primary amines that cross the endosomal membrane and buffer against acidification.

All experiments were done with C710 Aedes albopictus mosquito cells since VEEV can infect Aedes mosquitoes (Brault, Powers, and Weaver, 2002). SIN83-GFP virus was used for initial infection studies. This virus is composed of the VEEV structural proteins but contains a recombinant SIN genome encoding GFP and can be used at a BSL-2 level. It infects cells identically to VEEV, and has been characterized elsewhere (Paessler et al., 2003). At 8 h post-infection, more than 60% of untreated cells became infected. In contrast, none of the drug treated cells were infected (Fig. 1). Even though a range of drug concentrations was tested, the block to infection was similar with 1, 5, 10 or 50 μM monensin, 10, 20 or 30 mM ammonium chloride and 5, 10 or 20 mM chloroquine and is therefore not shown. Similar results were obtained when the same set of compounds was used on HEK293 cells. The mammalian vacuolar ATPase inhibitor, Bafilomycin A1, was also tested but did not inhibit infection on mosquito cells unless used at levels that were obviously toxic to the cells. This drug was effective at inhibiting infection in HEK293 cells, so we must conclude that the mosquito vacuolar ATPase is resistant to its effects (not shown). These observations indicated that VEEV requires acidification of endosomes in order to infect mosquito cells.

Figure 1. Infection of C710 cells and HEK293 cells with SIN83 is inhibited by lysomotropic agents.

A. Cells were incubated in medium alone (DMEM) or medium with ammonium chloride (DMEM+NH₄Cl) and then SIN83 virus encoding a GFP reporter gene was added to measure infection. C710 cells (left) or HEK293 cells (right) were used. Images are a composite of phase contrast and epifluorescence images. B. Virus titer for cells treated with each indicated compound, expressed as a percentage of the titer observed for untreated cells. Each was used at the following concentrations: 20 mM ammonium chloride, 20 mM chloroquine, 10 μM monensin. C710 cells (open bar) or HEK293 cells (solid bar) were used. Virus was added at an MOI of 0.5. The average +/- standard deviation for a triplicate sample is shown.

While VEEV infection was effectively blocked by each of the agents used, it was unclear which step in infection was affected. It has been reported that chloroquine, monensin and ammonium chloride block SFV uncoating and nucleocapsid release into the cytoplasm in mammalian cells (Helenius, Marsh, and White, 1982). Studies done with SIN, another Old World alphavirus, indicated that chloroquine and ammonium chloride instead inhibited infection of mammalian cells by blocking the synthesis of viral RNA (Cassell, Edwards, and Brown, 1984). In another study using mosquito cells, ammonium chloride also inhibited SIN RNA synthesis while chloroquine increased infection (Hernandez, Luo, and Brown, 2001). However, low pH induces fusion of both SFV and SIN viruses to membranes and ammonium chloride inhibits this fusion in mammalian cells (Glomb-Reinmund and Kielian, 1998). To determine if the block in VEEV infection was at the point of entry rather than at a downstream step, the inhibitors were tested using an entry assay.

Recently, a new entry assay was developed to monitor penetration of viruses into the cellular cytoplasm in real-time (Saeed, Kolokoltsov, and Davey, 2006). The entry assay permits blocks at the level of virus entry to be quantitatively identified with high sensitivity. The assay utilizes viral pseudotypes that contain a murine leukemia retroviral (MLV) core and have viral envs on the surface. A protein expressed from the HIV nef gene is used to create a nef-luciferase fusion protein that is encapsulated into the virus particle after transfection and budding of the pseudotype. This assay measures virus entry by the principle of contents mixing, directly detecting the release of luciferase that will occur after fusion of cellular and viral membranes. It overcomes limitations imposed by infection assays, allowing the functions of the virus envs for entry to be assessed in live cells without interference from downstream steps such as replication (Fig. 2).

Figure 2. Production of pseudotype viruses for use in luciferase virus entry assay.

A. 293FT cells were transfected with plasmids encoding the indicated recombinant genes. The first plasmid codes for viral envelope proteins (VEEV, SFV, VSV or 10A1 MLV were used), the second codes for infection GFP marker (Ψ-GFP), the third encodes the MLV core structural proteins and polymerase (gag-pol) and the fourth for the nef-luciferase fusion protein. After 48 h, virus containing encapsulated luciferase was collected, purified and used for virus entry assays. B. Schematic representation of the luciferase-based virus entry assay. Virus entry occurs after fusion of virus and cell membranes. Luciferase encapsulated inside the virus particle (star) is then released into the cell cytoplasm through contents mixing. The release is measured by perfusing the cell with luciferin, the substrate for luciferase which enters through high activity cell membrane permeases (Craig et al., 1991). Emitted light is then measured.

The entry of VEEV, SFV and VSV env pseudotyped viruses into C710 mosquito cells was compared using the luciferase release assay. Plasmids used in the production of these pseudotypes are shown in Figure 2A. VEEV and SFV were used to compare the entry behaviour of New to Old World alphaviruses, respectively. The VSV pseudotype was included as a positive control since it is well established that VSV enters cells through a pH-dependent, clathrin-dependent endosomal route and utilizes an endocytic pathway in both mammalian and mosquito cells (Marsh and Helenius, 1989; Matlin et al., 1982; Superti et al., 1987). Virus titer and luciferase activity for a set of virus preparations is shown in Table 1. The luciferase activities of each of the virus pseudotypes varied over a six-fold range. Virus titer did not correlate well to particle-associated luciferase activity and is likely due to the observation that similar numbers of MLV particles bud from cells irrespective of the presence of an env on their surface (Sharma et al., 1997). However, luciferase activity on cells showed a closer correlation with virus titer and indicated that luciferase activity reflected env-dependent entry of virus into cells.

Table 1. Comparison of virus pseudotype titer to entry assay luciferase activity.

Viral env pseudotype	aPseudotype titer (cfu/ml)	bLuciferase activity of detergent lysed virus Counts (×103)/sec/μl	cLuciferase activity on cells Counts (×103)/sec//106 cells
VSV	2.3 × 108	6490 +/- 258	3,241 +/- 318
VEE	1.5 × 105	925 +/- 84	9.22 +/- 0.8
SFV	2.5 × 106	4144 +/- 129	17.39 +/- 1.2
10A1 MLV	8.0 × 106	1021 +/- 126	52.21 +/- 4.8

^aA GFP-encoding MLV pseudotyped virus was used. Virus titer was determined by limiting dilution and counting colonies of GFP-expressing cells 48 h after addition of virus to HEK293 cells and expressed as colony forming units (cfu) per ml.

^bLuciferase activity was determined by lysing purified particles in cell lysis buffer supplied as part of the Steadyglo luciferase assay kit (Promega). Luciferase assay buffer was added and activity was then measured in a Turner 20/20 luminometer. The average of three measurements +/- standard deviation is shown for each viral pseudotype.

^cOne milliliter of viral supernatant was added to 10⁶ HEK293 cells and the relative luciferase activity (counts/sec) was measured after 1 h incubation. The average of three measurements +/- standard deviation is shown for each viral pseudotype.

Since virus entry kinetics in mosquito cells was poorly understood, a time course of entry was first established. The pseudotyped viruses were incubated with either HEK293 or C710 cells and samples were tested for luciferase activity at ten time points, up to 135 min. The HEK293 cells were incubated at 37°C while the C710 cells were incubated at 27°C since these are the optimal growth temperatures for mammalian and insect cells, respectively. At 75 min the entry signal for the VEEV pseudotype peaked in both C710 and HEK293 cells. The signal for the VSV pseudotype entry peaked later at 105 min in both cell types. In HEK293 cells, the SFV pseudotype entry signal peaked at 105 min while in C710 cells the signal peaked earlier, at approximately 90 min (Fig. 3). Based on these findings, an incubation time of 60 min was used for the remaining entry studies. The decrease in signal following the plateau was likely the product of luciferase degradation in the cells (T_1/2 of luciferase is 30 min in mammalian cells).

Figure 3. Kinetics of pseudotyped virus entry.

HEK293 cells (top panel) or C710 cells (bottom panel) were incubated with viral pseudotypes bearing the envs of VEEV (circles), VSV (squares) or SFV (triangles). Excess virus was washed free of cells. At the times indicated, entry was measured and was expressed as a percentage of the maximum signal obtained for each viral pseudotype. Entry was measured at ten time points, from time zero to 135 minutes. Assay was done in duplicate and the average +/- standard deviation was plotted. The effective MOI for this experiment was determined to be 0.1 by measuring virus infection by GFP infection marker expression as described in the methods section.

To determine the effects of inhibiting endosomal acidification on virus entry, C710 cells were preincubated with monensin, chloroquine and ammonium chloride at the same concentrations used for infection studies for 1 h at 27°C (1, 5, 10 and 50 μM monensin, 10, 20 and 30 mM ammonium chloride and 5, 10 and 20 mM chloroquine). The luciferase-containing virus particles were then added to the cells for an hour and entry signals were determined via measurement of luciferase activity. Each of the chemical inhibitors reduced the entry signal to a similar low level for the viral pseudotypes used and the effect of a middle dose is shown (Fig. 4). VSV and VEEV signals were completely abolished. This is similar to the pattern seen with VSV and VEEV in previous work using mammalian cells (Kolokoltsov, Fleming, and Davey, 2006). Low levels were still detected in the presence of the inhibitors for the SFV pseudotype (Fig. 4, lower panel), with a small portion of the virus fusion from untreated cells remaining in the presence of each inhibitor. This suggested that SFV entry pathway or mechanism may differ from VEEV but will need a more detailed analysis to be confirmed. A pseudotype was also made bearing the envs of the 10A1 strain of MLV, which is known to enter cells at the cell surface through a pH-independent mechanism (Blanchard et al., 2006), and was used in the entry assay as a control for both cell viability and assay function. The 10A1 MLV pseudotype efficiently entered the C710 cells in the presence of all three inhibitors (Fig. 4) and indicated that cells were competent for virus infection and that the assay was unaffected by each chemical treatment. These findings indicated that the inhibitors block VEEV infection by acting at the level of virus entry and that acidification, likely in an endocytic compartment is required to allow VEEV entry to occur in a mosquito cell.

Figure 4. Lysomotropic agents prevent alphavirus entry into mosquito cells.

C710 cells were incubated for 1 h at 27°C with each of the indicated agents. Each inhibitor was used at the following concentrations: 20 mM ammonium chloride, 20 mM chloroquine and 10 μM monensin. Entry was then measured using the luciferase entry assay with pseudotypes bearing the envs of VEEV, SFV, VSV and 10A1 MLV as indicated. Entry was measured at 1 h after addition of virus to cells. Average luciferase activity (log percent of control) is shown for assays performed in triplicate +/- standard deviation.

Identification of Rab5 and Rab7 in mosquito cells

Having established that VEEV uses a pH-dependent route to enter mosquito cells, we then wanted to identify if an endocytic pathway was involved. Recently it was shown that VEEV needs both functional Rab5 and Rab7 endosomal GTPases for entry in mammalian cells (Kolokoltsov, Fleming, and Davey, 2006). Rab5 is found on early endosomes and Rab7 is found on late endosomes. Both are involved in endosomal fusion, trafficking and maturation (Bucci et al., 1995; Bucci et al., 1992; Bucci et al., 2000; Feng et al., 2001; Feng, Press, and Wandinger-Ness, 1995). If VEEV follows a similar pathway in mosquito cells as in mammalian cells, then entry and infection should have a similar dependence on mosquito homologs of these proteins. The genome of Anopheles gambiae was recently sequenced and published and the sequencing of the Aedes aegypti genome is near completion (Holt et al., 2002; Kaufman, Severson, and Robinson, 2002; Loftus, 2005). By sequence comparison, mosquito protein homologs were identified that showed 82% and 88% amino acid identity to human Rab5 and Rab7 genes, respectively. The aligned mosquito sequences had E values of < 1e^-83 for human Rab7 and 2e^-69 for human Rab5. Since the E value represents the chance of obtaining a similar alignment strictly by chance these low E values indicated that the genes identified were highly likely to be related genes and possibly functional homologs. Interestingly, the human genome contains three Rab5 isoforms (Bucci et al., 1995) but only one mosquito equivalent was identified within each sequence database. Both humans and mosquitoes appear to only have one form of the Rab7 protein.

cDNA encoding related Rab5 and Rab7 mosquito genes were isolated from Anopheles gambiae and Aedes albopictus cDNA libraries respectively. The sequences of the isolated DNA matched those identified in the databases except for several silent mutations and have been submitted to Genbank (NCBI) with accession numbers: Aedes albopictus Rab7 – 858647 and Anopheles gambiae Rab5 - 863568. Alignments to human and Drosophila homolog amino acid sequences are shown in Figure 5. Each was inserted into an insect expression plasmid, pAc5.1/V5-HisA (Invitrogen, CA), as a fusion to GFP, (GFP tag was placed at the N-terminus of each protein). The human Rab5 and Rab7 genes, each similarly tagged with GFP, were also cloned into the same mosquito expression construct for comparison with the mosquito genes. Each construct was transfected into C710 mosquito cells and expression was observed after 24 h. Typically, than 2-5 % of the cells expressed detectable fluorescent protein.

Figure 5. Amino acid alignment of human and mosquito Rab5 and Rab7 proteins.

Amino acid alignment of the three isoforms of the human (H. sap.) Rab5 GTPase, the Drosophila melanogaster (D. mel.) Rab5 and an Aedes aegypti mosquito homolog (A. aeg.) isolated from a cDNA library. Sequences were aligned using ClustalW (Thompson, Higgins, and Gibson, 1994) and regions of homology shaded using Boxshade software. The GTP-binding domains are indicated by horizontal bars (GTP-BD) and the amino acid substitutions used to create the DN (S34N) and CA (Q79L) forms of Rab5 and DN Rab7 (T22N) are indicated by vertical arrows. The conserved prenylation domain is also indicated.

To confirm that mosquito Rab5 and Rab7 genes were functionally homologous to their mammalian counterparts in endosome formation and maturation, amino acid substitutions at conserved residues were made in both proteins that give well characterized dominant negative (DN) and constitutively active (CA) phenotypes in the corresponding mammalian proteins (Bucci et al., 2000; Feng, Press, and Wandinger-Ness, 1995; Stenmark et al., 1994). The location of these mutations is shown within the alignments in Figure 5. To characterize the function of each protein bearing these substitutions, plasmids containing the mosquito genes were transfected into mosquito cells and those coding for human proteins were transfected into human cells. The cells were examined by confocal microscopy to compare expression and localization within the cell.

The Rab5 wild type (WT) protein from both humans and mosquitoes shared a similar expression pattern in human (Hela) and mosquito (C710) cells, respectively. Each protein was visible as punctate, vesicular staining within the cellular cytoplasm (Fig. 6A and B). Similarly, both the human and mosquito WT Rab7 proteins expressed equally well in human and mosquito cells and gave similar staining, though less punctate and more diffuse than the Rab5 expression. The mutant forms of both human and mosquito Rab5 and Rab7 also expressed well in the mosquito cells and gave similar phenotypes as in the mammalian cells. In each case the pattern of expression of the human protein matched that previously described and the mosquito proteins gave similar expression patterns. This indicated that each protein likely served similar roles in endosome biogenesis and function in both human and mosquito cells.

Figure 6. Characterization of mosquito homologs of human Rab5 and Rab7.

All cells were fixed and observed by confocal microscopy. Individual representative cells are shown. Pictures are of middle cell slice clearly showing cytoplasm vs. membrane. All experiments were repeated three times with similar outcomes. Cells were transfected with expression plasmids encoding either GFP alone or Rab5 WT, Rab5 Q79L (constitutively active) or Rab5 S34N (dominant negative) genes fused to GFP. (A) Putative mosquito homologs of Rab5 genes and those carrying amino acid substitutions making DN or CA forms were transfected into C710 mosquito cells and (B) human homologs were transfected into Hela cells. Cells were then incubated with AlexaFlour 594-labelled transferrin for 30 min at 27°C (C710 cells) or 37°C (Hela cells) and cells were then imaged. For Rab5Q79L in C710 cells, DAPI counterstain was used to outline cell nucleus. (C) C710 cells were transfected with insect expression plasmids encoding the mosquito homolog of Rab7WT fused to GFP. Cells were incubated with LysoTracker Red for 1 h at 27°C and then cells were imaged by microscopy.

Roles of Rab5 and Rab7 in mosquito cells and effects of mutant proteins on function

To further examine if the mosquito Rab5 and Rab7 genes made protein with similar function to their human counterparts, transfected cells were then incubated with labelled human transferrin. Transferrin uptake in mammalian cells occurs primarily through clathrin-mediated endocytosis and disruption of this pathway results in accumulation of transferrin prior to the block (Damke et al., 2001; van Dam and Stoorvogel, 2002). Human transferrin has been successfully used in drosophila cells for study of endocytosis (Blitzer and Nusse, 2006) but it was unknown if it would function in mosquito cells. Human transferrin was incubated with the mosquito cells and then the cells were fixed and stained with antibody against cytoplasmic tail of the human transferrin receptor (this domain is highly conserved across many species including mosquitoes). Colocalization of transferrin and antibody staining confirmed that the mosquito transferrin receptor was involved in the endocytic uptake of the human transferrin protein (Fig. 7).

Figure 7. Human transferrin colocalizes with mosquito transferrin receptor in C710 cells.

To verify that fluorescently labelled human transferrin binds to the mosquito transferrin receptor, live cells were incubated at 27°C with Alexa₅₉₄-labeled human transferrin (red) for 30 min and then fixed in paraformaldehyde. Cells were then stained with antiserum raised against a conserved cytoplasmic-domain peptide of the transferrin receptor (green). The image is a single optical slice taken through the center of the cell by confocal microscopy. Arrowheads indicate transferrin staining that is colocalized (orange) or adjacent to transferrin receptor staining in one representative cell.

Since Rab5 controls early endosome formation, disrupting Rab5 function in mammalian cells results in inhibition of rapid transferrin uptake and sequestration of transferrin close to the cell surface (Bucci et al., 1995; Stenmark et al., 1994; Trischler, Stoorvogel, and Ullrich, 1999). Similarly, in mosquito cells transferrin uptake was rapid and evident at 15 min after the addition of labelled transferrin, consistent with the rapid endocytic uptake seen in mammalian cells. Confocal microscopy revealed similar patterns of staining for transferrin in the human Rab5 WT -expressing Hela cells and in the mosquito Rab5 WT -expressing C710 cells with transferrin often associated with Rab5-containing vesicles (Fig. 6A and B). The activity of the Rab5 DN mutant (S34N) was also determined by transferrin uptake. The expression of both human and mosquito DN Rab5 gave diffuse staining in the cell cytoplasm with punctate, vesicular staining along the cell membrane. This expression blocked transferrin internalization into both cell types, with the transferrin remaining at the cell surface in small vesicles and possibly coated pits. This is consistent with the function of Rab5 in early endosome formation and trafficking (Fig. 6A and B). The human and mosquito CA Rab5 (Q79L) was present on large fluorescent vesicles that contained transferrin (Fig. 6A and B). This is consistent with the CA form of Rab5 causing accumulation of early endosomes by preventing fusion of early with late endosomes and inhibiting the maturation of endosomes beyond the early endosome. In each case, the mosquito gene functioned equivalently to the human gene indicating that the mosquito Rab5 homolog is similar to that found in human and other mammalian cells. The pattern of transferrin uptake and the phenotypes of each mutant appeared consistent with a conserved role in insect cell endosome formation and maturation.

The function of the Rab7 mosquito homolog was then tested. Transferrin is recycled back to the cell surface from the early endosome and therefore its pathway should be unaffected by the DN Rab7 mutant, since it is located on and influences late endosomes. Indeed, expression of both WT Rab7 and DN Rab7 (T22N) allowed transferrin uptake into the mosquito cells. Rab7 has been reported as important in late endosome formation and lysosome biogenesis (Bucci et al., 2000; Feng, Press, and Wandinger-Ness, 1995). To assess if Rab7 associates with vesicles in mosquito cells, markers for late endosomes and lysosomes must be identified in mosquitoes. Mammalian antibodies recognizing late endosome and lysosome markers have yet to be made for insect cells. Instead, the cells were transfected with WT Rab7 and then incubated with Lyso Tracker (Invitrogen, CA). This weak base accumulates and reaches peak fluorescence in highly acidic compartments, typically late endosomes and lysosomes and has previously been used to characterize Rab7 function in mammalian cells (Bucci et al., 2000; Gutierrez et al., 2004). It was observed that cells expressing WT Rab7 accumulated Lyso Tracker in perinuclear vesicles and that the Rab7 colocalized with the Lyso Tracker (Fig. 6C). Together with the lack of association between labelled Rab7 with transferrin, the co-localization with Lyso Tracker indicated that the Rab7 homolog was present in late endocytic vesicles and/or lysosomes as expected. In summary, this portion of the work indicates that mosquito Rab5 is necessary for early endosome formation and Rab7 likely has a role in endocytic trafficking towards lysosomes. These observations indicate that Rab5 and Rab7 have conserved roles for endosome function and maturation in insect cells as in mammalian cells and will be useful for study of virus entry pathways.

Functional Rab5 and Rab7 are necessary for VEEV infection of mosquito cells

Given the important roles of Rab5 and Rab7 in endosome trafficking, it was expected that expression of the DN and CA Rab proteins in mosquito cells should disrupt VEEV infection if endocytosis was required for infection. Rab5 function will be crucial if VEEV enters the cell and utilizes early endosomes for trafficking from the membrane. Rab7 function will be important if the late endosome must be accessed by the virus before fusion can occur and the genome can be released. Flow cytometry was used to get a quantitative and accurate measurement of how expression of each protein affected subsequent infection of the mosquito cells by VEEV. After transfection of the plasmids encoding for the Rab proteins, cells were infected with SIN83. Cells were analyzed by flow cytometry to detect expression of GFP-Rab5/Rab7 fusion proteins and subsequent infection with the SIN83 virus. Infection of cells expressing the indicated protein was compared to infection of cells expressing the transfected gene fused to GFP. Controls were virus alone, expression of GFP alone and virus infection of cells expressing GFP alone. Flow cytometry allowed a detailed analysis of the affects of expressing each of the recombinant Rab genes on virus infection and gave internal controls for variation in cell number, MOI and transfection efficiency. The raw flow cytometry data is shown in Figure 8.

Figure 8. FACS analysis of infection in mosquito cells expressing wild type and DN or CA forms of mosquito Rab5 and Rab7 proteins.

C710 cells were transfected with each indicated gene fused to GFP. After 1 d cells were infected with SIN83 virus. Cells were then stained for VEEV env expression. Cells were analyzed by FACS and divided into 4 quadrants (gates a,b,c,d, shown top left panel) where a=cells infected but not expressing the GFP-fusion protein, b= cells not infected and not expressing the GFP fusion protein, c= cells infected and expressing the GFP fusion protein at moderate to high levels and d=cells not infected but expressing the GFP fusion protein. The impact of gene expression on infection was calculated as shown at lower right where the proportion of infected cells expressing the GFP fusion protein is expressed as a percentage of uninfected cells expressing the GFP-tagged protein. This analysis internally controls for experimental variation in: virus infectivity, expression plasmid transfection efficiency and total cell number. The top three panels serve as controls showing staining with virus alone, GFP alone (no added virus) and virus infection in the presence of GFP. The lighter shades indicate higher numbers of cells.

First, the mosquito cells were transfected with human Rab5 and Rab7 genes and the effect expression on virus infection was measured in mosquito cells. The wild-type human Rab5 had little effect on infection but both the DN (S34N) and the CA (Q79L) forms of human Rab5 significantly decreased the SIN83 infection (Fig. 9, upper panel) by 45 and 30%, respectively (P<0.01). Next, the mosquito genes were tested. Interestingly, as the amount of mosquito Rab5 WT homolog expression increased, the infection by SIN83 also increased, rising 2.5-fold over the infection seen for cells transfected with only GFP. This suggested that Rab5 activity in mosquito cells is normally limiting for infection. Similar to the human Rab forms, the expression of the mosquito Rab 5 DN (S34N) and CA (Q79L) proteins also decreased infection but to a greater extent, by an average of 53 and 57%, respectively (P<0.01) (Fig. 9, middle panel). Neither the human nor the mosquito Rab5 S34N protein expression and subsequent infection with SIN83 differed significantly from the Rab5 Q79L protein expression and infection (P>0.05). The results of the flow cytometry assay clearly illustrated that Rab5 function and therefore early endosomes are necessary for mosquito cell infection with the SIN83 virus. The experiment was repeated with WT and DN (T22N) mosquito Rab7 proteins. The transfection of the WT Rab7 did not have a significant impact on SIN83 infection, instead increasing infection by a small amount (20%). In contrast, expression of the DN Rab7 mutant (T22N) significantly lowered the infection of the cells (Fig. 9, lower panel), by 45% (P<0.05). Together, these results suggest that functional early as well as late endosomes are important for VEEV infection of mosquito cells. We conclude that VEEV has a distinct requirement for endosomal acidification and endosomal trafficking in mosquito cells.

Figure 9. Rab5 and Rab7 are necessary for infection of mosquito cells with SIN83.

C710 cells were transfected with an insect expression plasmid containing either GFP alone or genes for either the human or mosquito proteins Rab5 WT, DN Rab5 (S34N), CA Rab5 (Q79L), Rab7 WT, DN Rab7 (T22N) fused to GFP. After 24 h, cells were challenged with SIN83 virus. Cells were fixed 12 h post-infection and stained with an anti-VEEV env antibody, as a marker of infection. FACS analysis was used to determine the impact of gene expression on infection and infection efficiency was calculated as described in Fig. 8. The effects of both human Rab5 expression (upper), mosquito Rab5 homolog expression (middle) and mosquito Rab7 homolog expression (bottom) are shown. The average of three separate experiments is shown +/- standard deviation. * P<0.05, ** P<0.01

Discussion

Many families of viruses enter host cells by an endocytic pathway rather than at the cell surface. Endocytosis is an important step in virus entry, especially for pH-dependent enveloped viruses which require low pH-induced fusion to occur between the viral and cellular membranes in order to release the viral genome (Marsh and Helenius, 2006; Sieczkarski and Whittaker, 2002a). West Nile, Vesicular Stomatitis and Influenza A viruses require a functioning endocytic pathway in order to enter and infect cells and disabling steps along this pathway causes a marked reduction in virus entry (Chu and Ng, 2004; Sieczkarski and Whittaker, 2002a). Most of the information on alphavirus attachment and entry into host cells has come from work with Old World alphaviruses such as SFV and SIN in mammalian cell types. Little is known about the infection pathway taken by New World alphaviruses in mammalian cells and especially into the cells of their vector, mosquitoes.

By using a real-time contents mixing assay to measure VEEV entry, we first established that for contents mixing to take place, acidification was needed. This assay can be used with any pseudotyped virus and can precisely detect entry by measuring the fusion of the viral membrane with the cellular membrane and release of virion contents. We show for the first time that this assay can be effectively used in mosquito cells to gain insight into virus entry in a non-mammalian host. Both the New World VEEV and Old world SFV had similar requirements with entry being impaired by all lysosomotropic agents tested except Bafilomycin. Bafilomycin was not effective at concentrations that block entry in mammalian cells. At higher doses significant toxicity was observed in mosquito cells and the control virus used, 10A1 MLV, which enters cells through a pH-independent pathway (Blanchard et al., 2006) was also affected. We suspect that the mosquito vacuolar ATPase is resistant to the actions of this drug. However, overall, we conclude that acidification is a requirement of virus entry into Aedes mosquito cells.

In mammalian cells, endocytic vesicle trafficking and fusion is partly controlled by small GTPases known as Rab proteins. Rab5 is responsible for early endosome function and regulation, while Rab7 is responsible for the maturation and regulation of late endosomes (Sieczkarski and Whittaker, 2003). There has been a great deal of research on these proteins in mammalian cells and this has led to a greater understanding of the endocytic mechanisms of the cell. Several Rab proteins have been identified in the fruit fly Drosophila, and have been shown to be important in endocytosis (Guha et al., 2003; Satoh et al., 1997; Satoh, Tokunaga, and Ozaki, 1997; Wucherpfennig, Wilsch-Brauninger, and Gonzalez-Gaitan, 2003) but prior to the current work nothing was known about such proteins in the mosquito. Our results demonstrate for the first time that mosquitoes have counterparts to the mammalian proteins that are essential for the functioning of the endocytic pathway.

By expressing the wild type, DN or CA forms of Rab5 and Rab7 mosquito proteins in mosquito cells, we demonstrated that the mosquito cell also utilizes these proteins for endocytosis much like the mammalian cell does. Each appears to be a true functional homolog of its mammalian counterpart. Protein distribution and ligand uptake are similar. The affects of making amino acid substitutions in the conserved GTP-binding domains also gives similar outcomes. The DN form of mosquito Rab5 prevented the transferrin uptake while the DN form of Rab7 affected lysosome biogenesis as in mammalian cells (Li et al., 1994). The CA Rab5 protein formed the classical large vesicles that accumulated transferrin in the mosquito cell as seen for the mammalian counterpart. Our results indicate that Rab5 and Rab 7 have an evolutionary conserved role in endosome biogenesis that is shared between mosquitoes and humans.

Both Rab5 and Rab7 proteins have been shown to be important for entry and infection for several viruses. Entry of influenza virus requires functional Rab5 as well as Rab7, while SFV and VSV only need functional Rab5 present for productive infection (Sieczkarski and Whittaker, 2003). SFV was shown to require Rab7 function in mammalian cells in order to reach the late endosome but it is unclear if access to this compartment leads to productive infection (Sieczkarski and Whittaker, 2003; Vonderheit and Helenius, 2005). In mammalian cells, VEEV infection was shown to require both Rab5 and Rab7, indicating that the virus must access both early and late endosomes for productive infection (Kolokoltsov, Fleming, and Davey, 2006). In addition to regulating endocytosis, Rab5 and Rab7 may act as signalling molecules to create the appropriate environment in the mosquito cell for virus entry and infection. Here, disruption of the endocytic pathway by expression of DN and CA forms of Rab5 and Rab7 demonstrated the importance of this pathway for VEEV infection of mosquito cells. The expression of the wild type form of the mosquito Rab5 homolog increased infection with VEEV (SIN83) virus showing that early endosome formation may be limiting for infection. In contrast, the DN and CA forms greatly and significantly reduced mosquito cell infection. Similarly, DN Rab7 expression decreased infection with VEEV. We conclude that functional Rab5 and Rab7 are required for VEEV entry into mosquito cells indicating a need to access a late endosome after the early endosome in order to establish an infection.

There are many proteins involved in the mammalian endocytic pathway that most likely have homologs within the mosquito genome. The recent sequencing of both the Anopheles gambiae and Aedes aegypti mosquito genomes provides a novel opportunity to find these genes and determine how they function in the mosquito endocytic pathway as well as what their roles may be in virus entry and infection. Many viruses that are highly infectious and sometimes fatal to humans are transmitted by mosquitoes or other insect vectors. Knowledge of the molecular mechanisms of the endocytic pathway in mosquito and other insect cells could greatly improve our abilities to break the human-vector infection studies in highly pathogenic arboviruses. This research represents the first identification and characterization of mosquito endocytic proteins and, importantly, the involvement of these proteins in virus entry. Mosquito Rab5 and Rab7 are most likely involved in the entry of other alphaviruses and they may prove useful in investigating arbovirus entry as well as developing strategies to prevent arbovirus infection of mosquitoes.

Materials and Methods

Chemicals

Chloroquine and monensin were from Calbiochem (San Diego, CA). Ammonium chloride was from Sigma (St. Louis, MO).

Cell lines and cultivation

C710 Aedes cells (kindly provided by Dr. Ilya Frolov, UTMB, TX) were used for entry assays as well as transfection/infection studies. The cells were grown in DMEM supplemented with 10% heat-inactivated fetal bovine serum (Gemini Bioproducts, CA), 1% penicillin-streptomycin and 1% tryptose phosphate broth (Sigma, MO). 293FT cells (Invitrogen, CA) were used to generate envelope pseudotyped virions and were grown in the same medium with 0.5 mg/ml of Geneticin (Invitrogen, CA).

Expression plasmid constructs

All plasmids were prepared using Qiagen kits (Valencia, CA). A pcDNA3 (Invitrogen, CA) expression plasmid containing the VEEV envs from E3 to E1 from subtype IC strain 3908 was used to make VEEV env pseudotypes (Kolokoltsov, Weaver, and Davey, 2005). Similarly, a pcDNA3 expression plasmid containing the SFV envs was used for SFV pseudotype production and was provided by Dr. D. Sanders (Indiana University Medical School, IN) and is described elsewhere (Kahl et al., 2004). For production of the MLV pseudotype, p10A1 (Clonetech, CA) was used in the transfection. A plasmid encoding VSV-G (pVSV-G, BD Biosciences, CA) was used to make vesicular stomatitis virus (VSV) pseudotypes as described previously (Kolokoltsov, Weaver, and Davey, 2005). pGAG-POL which encodes the murine leukemia virus (MLV) gag and polymerase was a gift of Dr. J. Cunningham (Harvard Medical School, MA). The marker gene encoding plasmid was based on pFB (Stratagene, CA) and encodes enhanced green fluorescent protein (GFP) under control of the MLV LTR and virus packaging sequence (pFB-GFP).

Production of env pseudotyped MLV

MLV particles bearing the envs of VEEV, VSV and SFV were made according to previous work (Kolokoltsov, Weaver, and Davey, 2005). Briefly, 293FT cells (Invitrogen) were transiently transfected using calcium phosphate (Chen and Okayama, 1987) with pGAG-POL, pFB-GFP and plasmids encoding appropriate envs. The VSV-G-encoding plasmid was used at 1 μg per transfection and other plasmids were used at 5 μg each. After overnight incubation, the medium was replaced. When virus production peaked after a total of 48 h, the supernatants were collected and filtered through a 0.45 μm cellulose acetate filter. The filtrate was used directly, or else virus was purified by pelleting through 20% (w/v) sucrose in 20 mM Tris-HCl, pH 7.4. Titer of the MLV pseudotypes was determined by limiting dilution on HEK293 cells and by counting the number of GFP positive colonies 2 d post-infection.

Luciferase-based entry assay

The rapid entry assay requires that luciferase enzyme is enclosed within an intact viral envelope. This was achieved by fusing the luciferase protein to the HIV nef protein. The nef-luciferase encoding plasmid (pcDNA3-Nef-luc) used in making virus for entry assays was made by amplifying the nef gene and cloning it into the pcDNA3 vector that contained the firefly luciferase gene. This plasmid is described elsewhere (Saeed, Kolokoltsov, and Davey, 2006). After transient transfection of plasmids encoding viral structural proteins and the nef-luciferase fusion construct (1 μg), virions budding from the cell surface incorporate the nef protein inside the viral envelope. Previously this method was used exclusively for mammalian cells but here we show it can be applied to mosquito cells as well. Entry assays were performed on C710 cells and involved incubating virus-containing supernatant with 10⁵ cells for 1 h at room temperature at an MOI of <1. Cells were then pelleted, washed with PBS and lysed by freeze/thaw to ensure that luciferin could permeate subcellular compartments. This treatment does not disrupt virus particles yet ensures that luciferase inside the cell can be accessed by the luciferin in the buffer. The thawed cells were then resuspended in Steady-glo luciferase assay buffer (Promega, #E2510) and light emitted by luciferase was detected after 12 sec using a luminometer (Turner Biosystems model 20/20).

Use of inhibitors of endosomal acidification

Ammonium chloride and chloroquine were dissolved directly in DMEM. Monensin was made as a 50 mM stock in ethanol. All were diluted to the concentrations indicated in DMEM immediately before use. For entry assays, cells were pre-incubated with drug for 1 h at room temperature and then virus was added as described above.

Use of human endocytic trafficking proteins to identify entry pathway of viruses

Rab5 is required for early endosome formation and function, and Rab7 is necessary for late endosome maturation. A dominant negative (DN) recombinant form of Rab5, S34N, blocks early endosome formation in mammalian cells. The constitutively active (CA) mutant, Q79L, of Rab5 prevents endosome fusion and results in accumulation of early endosomes. Both Rab5 S34N and Q79L encoding plasmids were provided by Dr. P. Stahl at Washington University Medical School. Rab7 is required for formation of late endosomes and expression of the mutant form, T22N (Dr. Wandinger-Ness, UNM) blocks this step and prevents lysosome biogenesis (Bucci et al., 2000; Feng et al., 2001). Each protein, when fused to the c-terminus of GFP, retains its function and permits detection by fluorescent microscopy and by flow cytometry. For expression in insect cells, human Rab5 and Rab7 genes similarly fused to GFP were cloned into an insect expression vector pAc5.1/V5-HisA (Invitrogen, CA). This vector contains the insect actin 5C promoter which allows for gene expression in mosquito as well as mammalian cells (Zhao and Eggleston, 1999).

Identification and use of mosquito homologs to human Rab5 and Rab7

Mosquito homologs of human Rab5 and Rab7 genes were identified using sequence alignment within the recently published Aedes aegypti and Anopheles gambiae genome sequences using the human genes as a template. The human amino acid sequences were aligned against all mosquito genome open reading frames (ORFs) using TBLASTN (Altschul et al., 1997) and putative mosquito gene sequences were returned. Exons were identified within these ORFs and deduced amino acid sequences were then aligned back to the human amino acid sequences. The mosquito amino acid sequences had over 80% similarity to the human genes. This level of similarity suggested the potential for analogous protein function between the two species. RNA was extracted from C710 mosquito cells using RNAqueous (Ambion, TX) and cDNA made by reverse transcription with a Superscript kit (Invitrogen, CA) according to the manufacturer's directions. For expression of insect genes, the pAc5.1/V5-HisA insect expression plasmid (Invitrogen, CA) was modified by insertion of the GFP coding sequence (pAc5.1-GFP.) Rab 5 and Rab 7 mosquito genes were amplified from the cDNA library by PCR and cloned into the pAc5.1-GFP vector. The GFP was placed in-frame and 5′ to the Rab encoding region. PCR oligos used to amplify the genes are as follows: mosquito Rab5 – 5′-CCTAAGCTTCATATGGCATGAGTCCGCGAG-3′ and 5′-AAGGATCCGCGGCCGCTCAAGCACAGCAGCCGCTG-3′; mosquito Rab7 –5′-CCTAAGCTTCATATGGCAACTAGGAAAAAGGTC-3′ and 5′- GAAGGATCCGCGGCCGCTTAGCACGAGCAGTTGTCTCC-3′; eGFP 5′- GACTCGAGCACGGTGAGCAAGGGCGAGGAG-3′ and 5′-CTGTCTAGAGCGGCCGCTTACTTGTACAGCTCGTCCATGCCG-3′. Restriction endonuclease sites used for cloning are underlined. The DN or CA mutant forms of the mosquito genes were made by making amino acid substitutions at conserved residues important for GDP/GTP exchange previously characterized in mammalian homologs, i.e. Rab5 S34N and Q79L and Rab7 T22N. The site directed mutagenesis was performed at the Recombinant DNA Core Facility at UTMB and changes confirmed by DNA sequencing.

Transfection of plasmids into mosquito cells

The pAc5.1-GFP plasmid containing either the human or mosquito genes were transfected into C710 cells using Effectene (Qiagen, CA) according to manufacturer's instructions. Briefly, for a 10 cm² plate, 4 μg of DNA was mixed with 500 μL buffer EC and 32 μL enhancer was added. This was allowed to incubate for 5 min on the benchtop. Then 30 μL Effectene reagent was added and the solution vortexed briefly. After 10 min incubation, the solution was added to the cells. Expression was observed after 24 h and peaked at 48 h.

Transferrin and antibody staining

Alexa₅₉₄-labeled human transferrin (Invitrogen, CA) was added to DMEM (without serum) at a 1/200 dilution and incubated with cells for 30 min at 27°C. The cells were then pelleted, washed with PBS and fixed in 2% paraformaldehyde before imaging by microscopy. For antibody staining, cells were fixed in paraformaldehyde and incubated with 1% BSA to block non-specific interaction with antibodies. The cells were then incubated with primary antibody diluted as indicated in 1% BSA in PBS for 1 h. Excess antibody was removed by 3 washes with PBS and the cells were incubated with secondary antibody, labelled with the indicated fluorescent dye. Excess was washed off and the cells were imaged by microscopy.

Infection with SIN83

When studying infection, SIN83, a virus comprising the structural proteins of VEEV but containing a SIN genome, was used as described elsewhere (Paessler et al., 2003). It is antigenically and structurally identical to VEEV but can be safely used at BSL-2. This virus has the receptor specificity and entry mechanism of VEEV dictated by the VEEV envs. It was used as a safe alternative to wild-type VEEV, which is a BSL-3 restricted virus and select agent. SIN83 was used in flow cytometry and microscopy experiments and was a gift of Dr. Ilya Frolov (UTMB, Galveston, TX). Another form of the virus, SIN83-GFP, had GFP as an infection marker under control of a second subgenomic 26S promoter and was used where indicated. Virus titer was determined by plaque formation and GFP expression, respectively.

Infection analysis by flow cytometry

Insect expression vectors encoding either the human or mosquito genes were transfected into C710 cells using Effectene according to manufacturer's instructions. Plasmids were used at 4 μg DNA per 10⁶ cells. After 1 d, cells were infected with SIN83 virus at an MOI of 1. At 12 h post-infection when viral proteins were expressed in cells, the cells were removed, pelleted at 350 × g, washed with PBS, pelleted again and fixed in 2% paraformaldehyde for 10 min at room temperature. The cells were washed and resuspended in PBS containing 5% BSA. Cells were then incubated with polyclonal serum raised against VEEV (ATCC #VR-1250AF) for 20 min at room temperature. Cells were again pelleted, washed, and incubated for 20 min with a secondary antibody labelled with Alexa Fluor 647 (Invitrogen, CA). Cells were analyzed using a Becton-Dickinson FACSCanto instrument in the Flow Cytometry Core Facility at UTMB.

Statistics

Statistics were performed using Graphpad software (GraphPad Prism version 4.00 for Windows, GraphPad Software, San Diego California USA, www.graphpad.com). Data were compared by one way ANOVA and analysis included the Tukey-Kramer post test.

Abbreviations

VEEV

Venezuelan equine encephalitis virus

SFV

Semliki forest virus

SIN

Sindbis virus

VSV

Vesicular stomatitis virus

MLV

Murine leukemia virus

WT

wild type

DN

dominant negative

CA

constitutively active

env

envelope protein