Functional Pseudotyping of Human Immunodeficiency Virus Type 1 Vectors by Western Equine Encephalitis Virus Envelope Glycoprotein

Abstract

We investigated the ability of western equine encephalitis virus envelope glycoproteins (WEEV GP) to pseudotype lentiviral vectors. The titers of WEEV GP-pseudotyped human immunodeficiency virus type 1 (HIV) ranged as high as 8.0 × 10⁴ IU/ml on permissive cells. Sera from WEEV-infected mice specifically neutralized these pseudotypes; cell transduction was also sensitive to changes in pH. The host range of the pseudotyped particles in vitro was somewhat limited, which is atypical for most alphaviruses. HIV vectors pseudotyped by WEEV GP may be a useful tool for characterizing WEEV cell binding and entry and screening for small-molecule inhibitors.

Human immunodeficiency virus type 1 (HIV)-based vectors are capable of being pseudotyped with a range of viral glycoproteins. Recent examples include glycoproteins from Venezuelan equine encephalitis (VEEV) (10), Semliki Forest (7), Ross River (7, 8), and Sindbis (1, 14) viruses. In the last case, specificity has been enhanced by inserting various ligand recognition domains within the viral E2 gene (1, 14, 15). While these pseudotyped particles may broaden the host range of the HIV vector, the titer varies. VEEV pseudotypes have a titer of 10⁶ IU/ml and a broad host range. Semliki Forest virus pseudotypes usually have lower titers (10³ to 10⁴ IU/ml), whereas those of Ross River virus pseudotypes are at least 10- to 100-fold higher. In the latter case, transduction efficiencies can be variable and dependent upon culture conditions and cell type (8, 19).

Western equine encephalitis virus (WEEV) is a natural recombinant alphavirus that causes sporadic outbreaks of human and equine disease in the United States. It is also a potential biological weapon (6), as recognized by its NIH-NIAID priority B pathogen status (www.3.niaid.nih.gov/topics/BiodefenseRelated/Biodefense/research/CatA.htm). WEEV chiefly infects passerine birds as reservoir hosts and propagates through a silent enzootic cycle. An arbovirus, WEEV is transmitted by mosquitoes (Culex tarsalis). Larger mammals and humans infected by WEEV are usually dead-end hosts. Since 1964, there have been fewer than 700 confirmed cases of WEEV infection in humans (http://www.cdc.gov/ncidod/dvbid/arbor/weefact.htm). Symptoms begin 5 to 10 days following infection and range from mild flu-like symptoms to encephalitis causing coma and death. WEEV is typically propagated in vitro using BHK21 (3) or Vero (5, 16) cells, and due to its similarity to Sindbis virus, E1 and E2 glycoproteins are thought to be responsible for cell binding and entry (4, 11, 17, 21).

We show here that HIV vectors can be pseudotyped by WEEV envelope glycoproteins (WEEV GP). Entry of WEEV GP-pseudotyped virus is mediated by a pH-dependent pathway, and cell transduction is markedly inhibited by specific antisera. The host range of the WEEV GP-pseudotyped vector appeared narrower in vitro than that of other alphaviruses. These studies expand the known range of HIV vector pseudotypes and may provide useful tools for future characterization of WEEV cell binding and entry.

To produce pseudotyped virus, the cytomegalovirus immediate-early promoter of pcDNA3 (Invitrogen) was used to regulate expression of WEEV GP of strain CO92-1356 (Fig. 1A). The expression plasmid included viral capsid, E3, E2, 6k, and E1 genes and was sequence verified. 293T producer cells were cotransfected with HIV vectors encoding enhanced yellow fluorescent protein (eYFP) or luciferase (LUC) or genes conferring resistance to puromycin (puro) or hygromycin (hygro), along with either WEEV GP or vesicular stomatitis virus (VSV) G expression plasmids. Vector titers were determined by transducing HOS TK⁻ cells and analyzing cells 48 to 72 h later for either eYFP epifluorescence, LUC expression, or resistance to hygromycin or puromycin (read 7 to 9 days later). The titer of WEEV GP-pseudotyped HIV-eYFP was 4.8 × 10⁴ IU/ml, approximately 300-fold less than that of VSV G-pseudotyped HIV-eYFP (1.5 × 10⁷ IU/ml). In other experiments, WEEV GP-pseudotyped-vector titers of 8 × 10⁴ to 10 × 10⁴ IU/ml were obtained, which were still only 0.3% relative to VSV G titers (approaching 3 × 10⁷ IU/ml). The titers of WEEV GP-pseudotyped HIV-LUC (190 relative LUC units/ml), HIV-puro (3.0 × 10³ IU/ml), and HIV-hygro (1.25 × 10³ IU/ml) were between 200- and 400-fold lower than those of the corresponding VSV G pseudotypes (Fig. 1B). Figure 1C shows the flow cytometric analysis of HOS TK⁻ cells transduced by 3.0 ml of WEEV GP-pseudotyped HIV-eYFP. Using only 0.01 ml of VSV G-pseudotyped HIV-eYFP resulted in a comparable percentage of positively transduced cells, with a similar mean fluorescence intensity (MFI). Frameshifting the WEEV E2 gene at a unique ClaI site resulted in no pseudotyping (titer of <1 IU/ml; data not shown). To ensure that the WEEV GP plasmid was functional, transiently transfected 293T cells were metabolically labeled with [³⁵S]cysteine and [³⁵S]methionine. Figure 1D shows immunoprecipitation of labeled and processed viral proteins E1 and E2 (∼48 kDa) by use of ascites fluid from mice immunized by the Fleming WEEV strain (CDC VSO12). WEEV GP expression on the surfaces of 293T cells was analyzed by flow cytometry. The same ascites fluid was used as the primary (1°) antibody at a 1:100 dilution, and anti-mouse antibody conjugated to phycoerythrin was used as the secondary (2°) antibody at a 1:200 dilution. WEEV GP-transfected 293T producer cells incubated with both 1° and 2° antibodies showed positive events (∼80 out of 7,200 events) compared to cells that were mock transfected and treated with both 1° and 2° antibodies (2 out of 6,000 events) or WEEV GP-transfected cells treated with 1° antibody alone (no events) (data not shown). The MFI of the positive events was ∼3-fold higher than that of the negative cells. Neutralization assays were conducted using different sera obtained from infected mice (against the Fleming or BFS1703 WEEV strain). For the BFS1703 strain, Swiss-Webster mice were immunized subcutaneously and sera obtained 3 weeks later. WEEV GP- or VSV G-pseudotyped HIV-eYFP was incubated with sera at 37°C for 1 h. The mixture was then added to target HOS TK⁻ cells, and 2 days later, the titers were determined. In all cases, the titers of WEEV GP-pseudotyped HIV-eYFP were significantly reduced in a dose-dependent fashion; inhibition using 10 μl (1:100) of antisera reduced the titers by several orders of magnitude (Fig. 2A). None of the WEEV-reactive sera or ascites affected the titers of VSV G-pseudotyped HIV-eYFP.

FIG. 1.

WEEV GP can pseudotype HIV vectors. (A) Schematic diagram of cytomegalovirus (CMV) promoter-regulated expression of WEEV GP (pCDNA3-WEEV GP). (B) 293T producer cells were transiently cotransfected with the indicated envelope glycoprotein plasmid and HIV vector, and their titers on HOS TK⁻ cells were determined. The titer of WEEV GP-pseudotyped virus (black bars) is shown as a percentage relative to the VSV G titer (white bars) for each vector. The numbers above the bars reflect the percentages relative to the VSV G titers. eYFP expression was analyzed by epifluorescence microscopy. (C) Flow cytometric analysis of HOS TK⁻ cells transduced by VSV G- or WEEV GP-pseudotyped HIV-eYFP. Percentages of positive events and MFIs of positive cells are shown. (D) Immunoprecipitation of ^S35-labeled 293T cell lysates after transient transfection with the indicated envelope glycoproteins. The arrow indicates a prominent band at 48 kDa (WEEV GP lane). In all cases, the transfection efficiency was ∼80%, as judged by cotransfection of an eYFP marker.

FIG. 2.

WEEV GP pseudotyping is neutralized by antisera and inhibited by AZT. (A) HOS TK⁻ cells transduced with WEEV GP (black bars)- or VSV G (open bars)-pseudotyped HIV-eYFP. “1” indicates Fleming ascites fluid, and “2” to “4” indicate antisera from individual mice immunized subcutaneously with the BFS103 strain. Nonimmune serum (NIS) was used as a negative control. Dilutions (1:100 or 1:1,000) of the antisera were utilized (total volume of 1.0 ml). eYFP expression was analyzed by epifluorescence microscopy. * indicates P values of <0.05 for comparison to nonneutralized WEEV GP-pseudotyped vector (Student's t test). None of the treatments affected the titers of VSV G-pseudotyped HIV-eYFP (not shown). The experiment was repeated five or more times with similar results. Ab, antibody. (B) Increasing concentrations of AZT inhibited transduction of HOS TK⁻ cells by both WEEV GP (right)- and VSV G (left)-pseudotyped HIV-eYFP vectors, in triplicate, as measured by epifluorescence microscopy (means ± standard deviations). Numbers above bars indicate percentages relative to the control level (set at 100% for each envelope). * indicates P values of <0.05 for comparison to the no-AZT control (Student's t test).

After transduction of HOS TK⁻ cells by WEEV GP-pseudotyped HIV-eYFP, the percentage of eYFP-positive cells remained unchanged for several weeks in culture. To further exclude pseudotransduction (cell transduction by nongenomic virus-like particles), the reverse transcriptase inhibitor azidothymidine (AZT) was employed. Treatment of cells with AZT significantly decreased the titers of both WEEV-GP- and VSV G-pseudotyped HIV-eYFP in a dose-dependent manner (Fig. 2B).

Alphaviruses gain cellular entry via an endocytic pathway sensitive to pH changes (2, 13). In alphavirus infections, a decrease in pH alone is enough to cause fusion of virus to the cell membrane, and most alphaviruses are sensitive to changes in pH (12, 22, 23). Although not an alphavirus, VSV also depends on acidification of endosomes to initiate fusion of virus to the endosomal membrane and release of viral genome into the cytoplasm. Compounds that inhibit endosomal acidification should therefore inhibit transduction by both VSV G- and WEEV GP-pseudotyped vectors. To test this, target HOS TK⁻ cells were pretreated with 10 mM ammonium chloride and 50 μM chloroquine for 1 hour. Vector supernatants were added to target cells for 3 hours, and cells were then washed extensively. The titers were determined 2 days later, with HIV-eYFP pseudotyped with a murine leukemia virus amphotropic envelope (MLV-ampho) serving as a negative control and VSV G-pseudotyped HIV-eYFP as a positive control. Chloroquine treatment resulted in markedly decreased titers for both WEEV GP (98%)- and VSV G-pseudotyped HIV-eYFP (>98%) (Fig. 3). Treatment with ammonium chloride resulted in similar titer decrements for both. As expected, neither treatment affected the titers of MLV-ampho-pseudotyped HIV-eYFP.

FIG. 3.

Inhibition of acidification of endosomes decreases transduction by WEEV GP-pseudotyped HIV-eYFP. The titers of WEEV GP (black), VSV G (open), and amphotropic (striped) envelope glycoproteins on HOS TK⁻ cells treated with ammonium chloride (10 mM) or chloroquine (50 μM) were determined. eYFP expression was analyzed by epifluorescence microscopy. * indicates P values of <0.05 for comparison to untreated cells (Student's t test). Data are representative of two separate experiments.

The host range of alphaviruses is broad and diverse, varying from arthropods to large mammals, both in vivo and in vitro (4). To examine cell host range for WEEV GP-pseudotyped HIV, different mammalian cell lines were exposed to WEEV GP- and VSV G-pseudotyped HIV-eYFP. Interestingly, only eight of the cell lines tested (HOS TK⁻, MCF10A, BHK21, Vero, A23, 3T3, U87, and A549) were reasonably susceptible (Table 1). Relatively resistant cell lines (WEEV GP titers of <0.01% relative to the corresponding VSV G titers) included 293T, primary smooth muscle cells, human foreskin fibroblasts, MCF7, CaCo2, T-84, HT29, HT1080, HepG2, C8166, PM1, SAOS2, HeLa, BSC-40, CHO, LMTK⁻, A9, 3T6, and COS7 cells (not shown). Of note, transduction events were observed in most of those cell lines, suggesting incomplete resistance.

TABLE 1.

Cell lines permissive to transduction by WEEV GP-pseudotyped HIV-eYFP

Cell line	Cell line origin	Titer (IU/ml)c		WEEV GP titer relative to VSV G titer (%)
		VSV G	WEEV GP
HOS TK−a	Human osteosarcoma	1.1 × 107	4.9 × 104	0.5
A549a	Human lung carcinoma	2.4 × 106	3.4 × 104	1.3
MCF10Aa	Human breast epithelium	1.0 × 107	5.5 × 104	0.5
U87-MGb	Human brain glioblastoma	4.1 × 106	9.2 × 103	0.2
BHK21a	Baby hamster kidney	1.3 × 106	1.5 × 104	1.2
Verob	African green monkey	1.0 × 106	1.5 × 103	0.1
3T3a	Murine animal	1.1 × 106	1.1 × 104	1.0
A23b	Hamster	1.8 × 106	4.2 × 103	0.2

^aTiters on these cells were determined by endpoint dilution and epifluorescence microscopy.

^bTiters on these cells were determined by flow cytometry (analysis of at least 10,000 events).

^cThe titers are from a representative experiment.

The WEEV GP-pseudotyped vector was concentrated by ultrafiltration (15-fold) and by ultracentrifugation (100-fold) without a substantial loss of infectious units, compared to the supernatant of the unconcentrated vector (Fig. 4). Unlike other viral glycoproteins, WEEV GP was unable to pseudotype vectors based upon MLV, simian immunodeficiency virus, equine infectious anemia virus, feline immunodeficiency virus, or bovine immunodeficiency virus (WEEV GP titers of <0.001% relative to the corresponding VSV G titers, with HOS TK⁻ cells used as targets; data not shown).

FIG. 4.

WEEV GP-pseudotyped HIV vector can be concentrated. WEEV GP- and VSV G-pseudotyped HIV-eYFP vectors were concentrated by ultrafiltration (UF) and ultracentrifugation (UC). The titers of the supernatants on HOS TK⁻ cells were determined by epifluorescence microscopy. The titers of unconcentrated (C) WEEV GP- and VSV G-pseudotyped vectors were 5.3 × 10³ IU/ml and 2.5 × 10⁶ IU/ml, respectively. Vectors concentrated by UF and UC had titers of 8.6 × 10⁴ IU/ml and 4.3 × 10⁵ IU/ml, respectively, for WEEV GP and 5.2 × 10⁷ and 1.0 × 10⁸ IU/ml, respectively, for VSV G. Numbers above the bars indicate the increases in titer. Data are representative of two separate experiments.

This is the first report to show that WEEV GP can pseudotype HIV vectors, with titers in the range of 5.0 × 10⁴ IU/ml. These appeared to be true transduction events since (i) they were specifically neutralizable; (ii) stable gene expression resulted, with MFIs comparable to those observed following VSV G transduction; and (iii) AZT inhibited transduction. Whereas most other alphavirus envelope glycoproteins are capable of pseudotyping a range of retroviral and lentiviral vectors (9, 10, 14, 18), WEEV GP cannot. The reason for these differences is unclear.

Unlike most other alphavirus envelope-pseudotyped particles, WEEV GP-pseudotyped HIV had a very limited host range. BHK21 and Vero cells are routinely used for WEEV virus production and infection assays. It is noteworthy that while Vero cells support WEEV replication and virus production, vector entry for WEEV GP-pseudotyped HIV was somewhat limited compared to that for susceptible human cells. Vero cells were positive when large volumes of vector supernatant were used for transductions; this requirement is likely due to the expression of African green monkey TRIM5α, which inhibits HIV at a postentry level (20).

Enveloped viruses typically enter cells by binding to receptors on the cell surface and then either directly fusing to the plasma membrane or fusing to an endosomal membrane following receptor-mediated endocytosis. WEEV GP-pseudotyped particles, like VSV and other alphaviruses, appear to enter cells via an endosomal pathway that is pH dependent. The actual cellular receptor used by WEEV is not presently known; pseudotyped particles may be a useful tool for identifying the receptors for both mammalian and arthropod cells.

WEEV infection is diagnosed by detecting antibodies specific to WEEV or by reverse transcription-PCR for WEEV in serum (4). Diagnostic methods such as isolation of virus in culture cells require biosafety level 3 containment. The lack of better diagnostic and treatment options for individuals infected with WEEV further makes the case for using WEEV GP-pseudotyped HIV as a tool for analytic testing and therapeutic screening. In the latter case, small molecules could tested be for their ability to specifically inhibit transduction by WEEV GP-pseudotyped vector in a high-throughput manner. Additionally, the restricted host range of these particles suggests possible uses as a targeted gene delivery vehicle.