Enzootic Nasal Tumor Virus Envelope Requires a Very Acidic pH for Fusion Activation and Infection

Abstract

Enzootic nasal tumor virus (ENTV) is a close relative of jaagsiekte sheep retrovirus (JSRV), and the two viruses use the same receptor, hyaluronidase 2 (Hyal2), for cell entry. We report here that, unlike the JSRV envelope (Env) protein, the ENTV Env protein does not induce cell fusion at pHs of 5.0 and above but requires a much lower pH (4.0 to 4.5) for fusion to occur. The entry of ENTV Env pseudovirions was substantially inhibited by bafilomycin A1 (BafA1) but was surprisingly enhanced by lysosomotropic agents and lysosomal protease inhibitors following a 4- to 6-h treatment period; of note, prolonged treatment with BafA1 or ammonium chloride completely blocked ENTV entry. Unlike typical pH-dependent viruses, ENTV Env pseudovirions were virtually resistant to inactivation at a low pH (4.5 or 5.0). Using chimeras formed from ENTV and JSRV Env proteins, we demonstrated that the transmembrane (TM) subunit of ENTV Env is primarily responsible for its unusually low pH requirement for fusion but found that the surface (SU) subunit of ENTV Env also critically influences its relatively low and pH-dependent fusion activity. Furthermore, the poor infectivity of ENTV pseudovirions in human cells was significantly improved by either replacing the SU subunit of ENTV Env with that of JSRV Env or overexpressing the functional Hyal2 receptor in target cells, suggesting that ENTV SU-Hyal2 interaction is likely to be the limiting step for viral infectivity. Collectively, our data reveal that the fusogenicity of ENTV Env is intrinsically lower than that of JSRV Env and that ENTV requires a more acidic pH for fusion, which may occur in an intracellular compartment(s) distinct from that used by JSRV.

Enzootic nasal tumor virus (ENTV) and jaagsiekte sheep retrovirus (JSRV) are two related simple betaretroviruses that cause contagious respiratory tumors in sheep and goats. ENTV targets the upper-airway epithelial cells and induces nasal adenocarcinomas (12), whereas JSRV infects the lung airway epithelial cells, causing pulmonary adenocarcinomas (17). One unique feature of these sheep retroviruses is that their envelope (env) genes function as active oncogenes, inducing transformation in cultured cells (1, 2, 11, 14, 29, 31, 47) and causing tumors in animals (6, 56, 57). Prior efforts have been devoted largely to developing an understanding of the mechanisms of oncogenic transformation by the Env proteins, in particular that of JSRV, while fusion and cell entry mediated by these Env proteins are poorly understood. Recently, we showed that JSRV Env-mediated fusion requires a low pH and that the cytoplasmic tail (CT) of JSRV Env negatively regulates its fusion activity (3, 9). The mechanisms of ENTV Env-mediated fusion are currently not known.

As the Env proteins in all other retroviruses, Env in ENTV is a type I transmembrane protein composed of surface (SU) and transmembrane (TM) subunits (10). The SU subunit contains a receptor-binding domain that recognizes the entry receptor, hyaluronidase 2 (Hyal2). Hyal2 is a glycosylphosphatidylinositol-anchored cell surface molecule that also serves as the receptor for JSRV (14, 37, 47). The TM subunit of ENTV Env is directly involved in membrane fusion, and it contains an amino (N)-terminal fusion peptide, a central coiled-coil region, a membrane-spanning domain, and a carboxyl (C)-terminal CT (10). One interesting feature of ENTV Env is that, similar to JSRV Env, it has a relatively long CT of 47 amino acid residues (10). More importantly, the CT of ENTV Env shows only ∼50% sequence identity to that of JSRV at the amino acid level, in contrast to an overall sequence identity of 89% for the full-length Env proteins (10). The significance of these sequence differences between ENTV and JSRV Env proteins is not known but may be related to the oncogenic transformation potential, Env trafficking, fusion modulation, or virus assembly (30).

ENTV uses human, ovine, or bovine Hyal2 for binding and entry in a fashion similar to JSRV but appears to have a restricted host range (14). A Moloney murine leukemia virus (MoMLV) oncoretroviral vector bearing JSRV Env efficiently transduces cell lines derived from most species, including sheep, human, bovine, and canine cell lines, with the notable exception of rodent cell lines (46, 47). In contrast, an equivalent MoMLV vector bearing ENTV Env does not transduce or only weakly transduces these cell lines, including those from humans and sheep (1, 14, 51). Interestingly, the overexpression of human Hyal2 renders most of these cell lines highly transducible for ENTV pseudotypes, indicating that Hyal2 overexpression can overcome the inefficient binding of ENTV Env to Hyal2 (51). By using soluble SUs of ENTV and JSRV Env proteins that were respectively fused to a human immunoglobulin G Fc fragment, it was shown previously that the binding affinity of ENTV Env for human Hyal2 is much lower than that of JSRV Env, with dissociation constants (K_ds) of ∼175 and ∼10 nM for ENTV and JSRV, respectively (26, 51). A subsequent surface plasmon resonance study revealed an even tighter association between JSRV Env and human Hyal2, with a K_d of 32 pM (52), although a similar study has not been performed for ENTV Env. Whether or not Hyal2 plays a direct role in the activation of ENTV and JSRV Env-mediated fusion is currently not known.

Using syncytium induction and cell-cell fusion assays that we recently developed for JSRV Env (9), herein we have examined the fusion properties of ENTV Env. We found, unexpectedly, that the fusogenicity of ENTV Env is much lower than that of JSRV Env and that ENTV Env requires a very acidic pH for fusion activation and cell entry. We determined that the TM subunit of ENTV Env dictates the extremely low pH requirement for fusion, yet we discovered that the SU subunit of ENTV Env also critically influences its relatively low and pH-dependent fusion activity, suggesting that Hyal2 may play a direct role in fusion activation. The remarkable differences in fusogenicity and low-pH requirements between ENTV and JSRV Env proteins, as well as their responses to treatments with lysosomal protease inhibitors and lysosomotropic agents, strongly argue that ENTV may use an intracellular compartment(s) that is distinct from that used by JSRV for its pH-dependent fusion and cell entry.

MATERIALS AND METHODS

Reagents and antibodies.

The fluorescent dye 5 (and 6)-(((4-chloromethyl)benzoyl)amino)tetramethylrhodamine (CMTMR) and Lipofectamine 2000 were purchased from Invitrogen (Carlsbad, CA). Ammonium chloride (NH₄Cl) and chloroquine were purchased from Sigma (St. Louis, MO). Bafilomycin A1 (BafA1), leupeptin hemisulfate, and cathepsin inhibitor III were purchased from Calbiochem (Darmstadt, Germany). The anti-FLAG monoclonal antibody beads (EZviewTM Red), anti-FLAG antibody, anti-β-actin monoclonal antibody, and secondary anti-mouse immunoglobulin G coupled to fluorescein isothiocyanate or phycoerythrin were purchased from Sigma (St. Louis, MO). The rabbit anti-murine leukemia virus (MLV) Gag antiserum was a kind gift of Alan Rein (National Cancer Institute, Frederick, MD).

Env constructs and cell lines.

The parental JSRV Env gene derived from the JS7 strain (13) and the parental ENTV Env gene derived from the ENTV1 strain (10) were cloned into the pCI-Neo vector, with the introduction of a FLAG sequence at both the N and C termini of the products, as described previously (9, 28). The chimeric Env proteins constructed from the ENTV and JSRV Env proteins were generated by swapping the SU and TM subunits of each Env, resulting in EJ, which contains the ENTV SU and the JSRV TM, and JE, which contains the JSRV SU and the ENTV TM. The 10A1 amphotropic MLV Env-expressing vector and the vesicular stomatitis virus G protein (VSV-G)-expressing plasmid have been described previously (9, 40). The Ebola virus glycoprotein (GP) expression vector was a kind gift of Gary Kobinger (National Microbiology Laboratory, Winnipeg, Canada).

293T/GFP, 293/LH2SN, NIH 3T3/LH2SN, HTX/LH2SN, and 293/GP-LAPSN cells have been described previously (9, 38, 47). All cell lines were cultured in Dulbecco's modified Eagle medium (Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum (HyClone, Logan, UT) and were maintained at 37°C in a 10% CO₂-air atmosphere at 100% relative humidity.

Syncytium induction and cell-cell fusion assays.

The syncytium induction assay was performed as described previously (9). Briefly, 1.7 × 10⁶ 293/LH2SN cells plated into 6-well plates were transfected with 2 μg of plasmid DNA encoding Env or VSV-G plus 0.5 μg of a pCMV plasmid encoding green fluorescent protein (GFP; a kind gift from François-Loïc Cosset, Lyon, France) by the calcium phosphate method. At 24 h posttransfection, cells were treated with phosphate-buffered saline (PBS)-10 mM HEPES-10 mM morpholineethanesulfonic acid (MES) solution (Sigma) at pHs ranging from 3.0 to 7.0 for 1 or 5 min at 37°C. Cells were then incubated at 37°C in complete growth medium, and syncytium induction from 10 min to 24 h after treatment was examined. Pictures were taken 1 h after treatment by using a fluorescence microscope (Carl Zeiss, Goettingen, Germany) with a green filter for GFP expression.

The cell-cell fusion assay was also performed as described previously (9). Briefly, 1.7 × 10⁶ 293T/GFP cells were transfected with 2 μg of DNA encoding Env or VSV-G by the Lipofectamine 2000 method. At 24 h posttransfection, cells were detached using PBS plus 5 mM EDTA and cocultured at a 2:1 ratio with effector HTX/LH2SN cells that were prelabeled with CMTMR as described previously (9). After 1 h of coculture at 37°C, cells were treated for 1 min with PBS-10 mM HEPES-10 mM MES, pH 4.0, 5.0, or 7.0, and allowed to recover in normal growth medium for 1 h. The cells were trypsinized and analyzed by flow cytometry using a FACSCalibur (BD Biosciences, Mississauga, Ontario, Canada). The level of surface expression of Env was determined by anti-FLAG staining and analyzed by flow cytometry as described previously (9).

Pseudotyping, infection, and pH inactivation of pseudovirions.

GFP-encoding MoMLV pseudovirions were produced by cotransfecting 2.5 × 10⁶ 293T cells with a plasmid encoding individual Env proteins, VSV-G, or Ebola virus GP and a packaging plasmid encoding the MoMLV Gag-Pol proteins (pCMV-gag-pol-MLV) and a transfer vector encoding GFP (pCMV-GFP-MLV) (both pCMV plasmids were kind gifts from François-Loïc Cosset, Lyon, France). Alternatively, the packaging cell line 293/GP-LAPSN was transfected with a plasmid encoding each Env to produce alkaline phosphatase (AP)-expressing pseudovirions. Supernatants were harvested 48 to 72 h posttransfection, and cell debris was removed by centrifugation at 3,200 × g. Viruses were used immediately for infection or were stored at −80°C.

Target cells were infected with appropriate amounts of viral supernatant in the presence of 5 μg of Polybrene (Sigma)/ml and were assessed for GFP expression by flow cytometry 48 h postinfection (9) or for AP-positive foci by cell staining 72 h postinfection as described previously (27). For experiments involving inhibitors, cells were pretreated with appropriate concentrations of the agents for 1 to 2 h and subsequently infected with viruses in the presence of the drugs for 4 to 6 h or 16 to 48 h, unless otherwise stated. Comparable multiplicities of infection were used for all viral pseudotypes in each experiment, and typically, a multiplicity of infection of 0.05 to 0.2 was used for all infections. Noninternalized viruses were inactivated with citrate buffer (pH 3.15) after the infection period, and viral infectivity was determined 48 h postinfection. To assess the stabilities of ENTV pseudoparticles at different pHs, GFP-encoding viruses were incubated in buffers at pHs ranging from 3.0 to 7.0 at 37°C for 30 or 5 min and then neutralized with culture medium. Appropriate amounts of viruses were used to infect HTX/LH2SN cells, and titers were determined by flow cytometry 48 h postinfection.

Overcoming BafA1-mediated entry block by low-pH pulse.

HTX/LH2SN cells were pretreated with 30 nM BafA1 for 1 h and incubated with ENTV Env pseudovirions at 4°C for 2 h. Following three washes with cold PBS, the virion-cell complexes were either directly exposed to a low-pH solution ranging from pH 3.5 to 7.0 for 5 min or were preincubated at 37°C for 1 h in the presence of 30 nM BafA1 and then incubated in a low-pH solution for 5 min. In both cases, the total period of infection in the presence of 30 nM BafA1 was 4 h. Noninternalized virus was inactivated using citrate buffer (pH 3.15) after the infection period, and viral infectivity was determined by flow cytometry 48 h after the initiation of infection.

Cell cycle analysis.

Cell cycle analysis was performed as described previously (8) in order to determine potential side effects of BafA1 and lysosomotropic agents on cell proliferation. Briefly, HTX/LH2SN cells were treated for 4 to 6 h with the highest concentrations of BafA1 or lysosomotropic agents, i.e., 30 nM BafA1, 30 μM chloroquine, or 30 mM NH₄Cl, that were used in the infection assays. Sixteen hours later, cells were trypsinized and fixed in ice-cold ethanol overnight. Cells were washed and treated with 5 μg of propidium iodide (Sigma, St. Louis, MO)/ml and 100 μg of RNase A (Qiagen, Chatsworth, CA)/ml for 30 min at 37°C and were analyzed by flow cytometry using FACSCalibur (BD Biosciences) and software from FlowJo, LLC (Ashland, OR).

Env protein incorporation into a MoMLV vector and immunoblotting.

Virion-producing 293/GP-LAPSN cells were lysed in a lysis buffer as described previously (28), and the lysate was precleared by centrifugation for 10 min at 13, 000 × g and 4°C, followed by boiling for 5 min in sodium dodecyl sulfate (SDS) loading buffer (0.4% SDS and 0.01% beta-mercaptoethanol). Supernatant containing viral particles from the transfected 293/GP-LAPSN cells was first precleared by centrifugation for 5 min at 3,200 × g and 4°C. Viruses in the supernatant were purified by centrifugation on a 20% sucrose cushion for 3 h at 185, 000 × g and 4°C, the pellet was resuspended in SDS loading buffer (1% SDS and 0.02% beta-mercaptoethanol), and the suspension was boiled for 10 min. Cell lysates and viral particles were subjected to SDS-10% polyacrylamide gel electrophoresis, followed by transfer onto polyvinylidene difluoride membranes and immunoblotting using an anti-FLAG antibody as described previously (28).

Statistical analysis.

Statistical analysis was performed using two-tailed Student's t test, and results were adjusted by Bonferroni corrections unless otherwise noted. Data plotted in all histograms or shown in the tables are means ± standard deviations, unless only two data points were included, for which only the means are given. Typically, three to eight independent experiments were used for statistical analysis.

RESULTS

ENTV Env induces syncytium formation and cytoplasmic content transfer at pH 4.5 or below but not at pH 5.0 or above.

We recently showed that the JSRV Env protein induces cell-cell fusion and syncytium formation following a 1-min treatment at pH 5.0 (9). Given the high level of sequence similarity between ENTV and JSRV Env proteins (10, 13, 14) and the use of the same receptor for entry by the two viruses (14, 47), we asked whether ENTV Env also induces fusion under similar conditions. Unexpectedly, while JSRV Env and VSV-G induced extensive syncytia after a treatment at pH 5.0 for 1 min, as shown previously (9), no syncytia induced by ENTV Env were observed (Fig. 1A, upper panels). In contrast, the treatment of ENTV Env-expressing cells at pH 4.0 for 1 min induced substantial syncytium formation (Fig. 1A, upper panels), although it took longer for the ENTV Env-induced syncytia to appear (∼20 min) than it did for the JSRV Env-induced syncytia (∼10 min), even at pH 5.0 (9). Occasionally, a few syncytia were observed in cells expressing ENTV Env at pH 4.5, but the syncytia were extremely small (Fig. 1A, upper panels). Noticeably, prolonged treatment of cells expressing ENTV Env, for 5 min, resulted in extensive syncytia at pH 4.5 and occasionally at pH 5.0, yet the syncytia were very small and were often indistinguishable from normal cells (Fig. 1A, lower panels). Taken together, these results indicated that the fusogenicity of ENTV Env is much lower than that of JSRV Env and that a more acidic pH (∼4.5) is required for the ENTV Env-mediated cell fusion to occur.

FIG. 1.

ENTV Env requires a very low pH to induce cell fusion. (A) Syncytium induction. 293/LH2SN cells cotransfected with plasmids encoding ENTV Env and GFP were treated with a solution ranging from pH 4.0 to 7.0 for 1 min (upper panels) or 5 min (lower panels) and monitored for syncytium formation over time. Pictures were typically taken 1 h after the pH treatment by using a fluorescence microscope. Pictures representative of results from at least four independent experiments are shown. (B) Cell-cell fusion. 293T/GFP cells transfected with plasmid DNA encoding ENTV Env, JSRV Env, or VSV-G were cocultured with HTX/LH2SN cells that were prelabeled with CMTMR. Cells were treated with a solution ranging from pH 3.0 to 7.5 for 1 min and analyzed for fusion by flow cytometry following a 1-h recovery period. For each fusion protein, the highest level of fusion activity at an optimal pH was set to 100%, and the fusion activities at other pHs were calculated relative to the highest values and were plotted against the pH values used. The results shown are the averages ± standard deviations of results from at least three independent experiments. (C) Env surface expression. Cells were stained with an anti-FLAG mouse antibody and then incubated with a fluorescein isothiocyanate (293/LH2SN)- or phycoerythrin (293T/GFP)-labeled anti-mouse antibody and analyzed by flow cytometry. Solid lines represent the fluorescence shifts of ENTV Env, broken lines represent the fluorescence shifts of JSRV Env, and filled areas represent cells incubated with secondary antibody alone. Results representative of those from three independent experiments are shown.

To confirm the above-described results and quantify the fusion activity of ENTV Env, we next used a cell-cell fusion assay that we recently developed for JSRV Env (9). The fusion indexes for Env or VSV-G, as determined by the percentages of fusion at certain pHs relative to that at the optimal pH, were plotted against the pH values used (Fig. 1B). VSV-G exhibited close-to-maximal fusion activity over a broad pH range, with a sharp drop at pH ∼6.0, and fusion was undetectable at pH 6.5 (Fig. 1B). This pattern was similar to that reported previously (19, 60). JSRV Env also exhibited fusion activity over a broad pH range yet, interestingly, with maximal fusion at pH 3.5; its fusion activity was detected up to a pH of 6.5, after which background fusion was observed, as we have shown previously (9) (Fig. 1B). In sharp contrast, ENTV Env-induced fusion occurred within a very narrow pH range; it started to appear at pH 4.0, and the protein exhibited maximal fusion activity at pH ∼3.0, beyond which cells were essentially detached by lower-pH treatments (Fig. 1B and data not shown). Notably, the fusion activity of ENTV Env was also lower than that of JSRV Env, approximately 15% of that of JSRV Env at pH 4.0 (Table 1; also see Fig. 6B). The pH values corresponding to the half-maximal fusion activities of ENTV Env, JSRV Env, and VSV-G were estimated to be 3.7, 5.4, and 6.2, respectively. It should be emphasized, however, that the cell-cell fusion assays described here and throughout this study were performed with 1 min of treatment of cocultured cells and that, because of technical limitations, we were unable to treat the cocultured cells for 5 min (cells became detached), as we did for the syncytium induction assay (Fig. 1A), in order to determine if ENTV Env can also cause cell-cell fusion at pH 4.5. Nonetheless, these results confirm that ENTV Env indeed has low fusogenicity compared with that of JSRV Env and requires an unusually low pH for fusion.

TABLE 1.

Fusion activities of ENTV Env, JSRV Env, and their chimeras^a

Env protein	% Fused cellsb at:
	pH 4.0	pH 5.0	pH 7.0
ENTV	2.22	1.78	1.34
EJ	11.12	7.25	1.31
JSRV	15.07	10.32	1.14
JE	5.73	1.96	1.45
None	0.42	0.42	0.61

^aEffector 293T/GFP cells were transfected with plasmids encoding individual Env proteins; 24 h posttransfection, cells were split and cocultured for 1 h with HTX/LH2SN target cells that were prelabeled with CMTMR. Cocultured cells were treated with buffers at different pHs for 1 min, and the percentages of fused cells were determined by flow cytometry after a 1-h recovery period. Experiments were performed in duplicate, and the results from one representative experiment are shown. Values are averages of results for two duplicate samples. See Fig. 6B for a comprehensive summary of results from multiple experiments.

^b% Fused cells, number of fused cells in a given sample as a percentage of the total number of cells in that sample.

FIG. 6.

Fusogenicities of ENTV Env chimeras in comparison with those of parental ENTV and JSRV Env proteins. EJ consists of the ENTV SU and the JSRV TM, while JE consists of the JSRV SU and the ENTV TM. (A) A syncytium induction assay was performed as described in the legend to Fig. 1A, except that different Env-encoding plasmids were used to transfect 293/LH2SN cells and cells were treated with solutions at the indicated pHs for 1 min. The results shown are representative of results from three to five independent experiments. (B) Cell-cell fusion was performed as described in the legend to Fig. 1B, except that different Env-encoding plasmids were tested under the three indicated pH conditions. The relative fusion activities of each Env at different pHs were calculated relative to the fusion activity of JSRV Env at pH 5.0 (set to 100% [dashed line]). Due to complex comparisons, the P values were not labeled in this graph but were instead provided in the text. Also note that the P values were not corrected by Bonferroni adjustment because in this particular case only two Env proteins of interest were compared to each other. No-Env, parental 293T/GFP cells transfected with no plasmid. (C) Env surface expression in 293T/GFP cells was assessed by flow cytometry as described in the legend to Fig. 1C, with geometric fluorescence means being acquired for each Env. All values were normalized to that for the parental ENTV Env, whose geometric mean was set to 100%. Averages ± standard deviations of results from four to six independent experiments are shown in panels B and C.

The surface expression of the ENTV Env SU, along with that of JSRV, on both 293/LH2SN and 293T/GFP cells was examined by flow cytometry. We consistently observed higher levels of SU expression for ENTV Env than for JSRV Env in both cell lines (based on geometric means) (Fig. 1C), indicating that the relatively low fusion activity of ENTV Env and its unusually low pH requirement are not due to the levels of ENTV SU expression on the cell surface.

The entry of ENTV Env pseudovirions is pH dependent, but the pattern appears to be different from that of JSRV.

The low-pH-dependent fusion activity of ENTV Env predicted that ENTV entry should be pH dependent. To test this, we generated MoMLV pseudovirions bearing the Env of ENTV, JSRV, or 10A1 or VSV-G and tested their abilities to enter cells in the presence of BafA1, a specific, nonreversible endosomal proton pump inhibitor (4), as well as two commonly used lysosomotropic agents (36), ammonium chloride (NH₄Cl) and chloroquine. Due to the limited host range of ENTV (14), we used cell lines that overexpress human Hyal2 in this and all other experiments throughout this study, unless otherwise specified. BafA1 inhibited ENTV Env pseudovirion infection of HTX/LH2SN cells by ∼25% (P < 0.05 at 30 nM; n = 3), in contrast to JSRV Env and VSV-G pseudovirion infection, which was inhibited by ∼50 to 70% (P < 0.05; n = 4) and ∼60 to 85% (P < 0.05 or 0.01; n = 3), respectively (Fig. 2A). As expected, entry by 10A1 Env pseudovirions was not affected by the BafA1 treatment (Fig. 2A), consistent with the fact that 10A1 is a pH-independent retrovirus (35, 41). No apparent cytotoxity was observed in the BafA1-treated cells, a finding which was also confirmed by independent cell cycle analyses showing cell proliferation profiles (G₀/G₁, M, and G₂/M) similar to that of untreated cells (data not shown). Similar results were also obtained with human 293 and mouse NIH 3T3 cells overexpressing human Hyal2 (data not shown).

FIG. 2.

Effects of BafA1, chloroquine, and ammonium chloride (NH₄Cl) on ENTV entry. (A to C) HTX/LH2SN cells were pretreated with BafA1 for 2 h (A) or chloroquine (B) or NH₄Cl (C) for 1 h at the indicated concentrations and were infected with GFP-encoding pseudovirions bearing ENTV Env, JSRV Env, VSV-G, or amphotropic 10A1 MLV Env in the presence of the respective agents for 6 h (A) or 4 h (B and C). (D to F) HTX/LH2SN cells were pretreated with BafA1, chloroquine, or NH₄Cl at the indicated concentrations for 1 h and infected with each pseudotype in the presence of the agents for 16 h (D and E) or 48 h (F). In all cases, viral infectivity was determined 48 h postinfection by flow cytometry. Percentages of infection were calculated as the number of infected cells relative to the number of infected untreated cells, and values are averages ± standard deviations of results from two to eight independent experiments. Two-tailed Student's t tests were used for statistical analysis; * indicates a P value of <0.05, and ** indicates a P value of <0.01.

The treatment of cells with chloroquine or NH₄Cl yielded somewhat surprising results. While chloroquine strongly inhibited the entry of VSV-G pseudovirions, by ∼60 to 90% (P < 0.05; n = 4), its inhibitory effect on JSRV Env pseudovirions was much weaker, approximately 20 to 40% (P < 0.05; n = 4) following a similar 4-h treatment (Fig. 2B), as we observed previously (3). Surprisingly, the entry of ENTV Env pseudovirions was not inhibited but was rather enhanced by chloroquine treatment, although the difference was not statistically significant (P = ∼0.34; n = 4) (Fig. 2B). Even more strikingly, a 4-h NH₄Cl treatment performed in parallel significantly increased, in a concentration-dependent manner, the entry of ENTV Env pseudovirions (P < 0.05 and P < 0.01 for 15 and 30 mM, respectively; n = 3) (Fig. 2C) while substantially blocking the entry of JSRV Env (∼50% [P = 0.06; n = 3]) and VSV-G (∼80%) pseudovirions (Fig. 2C), similar to previous results (3). Cell cycle analyses of HTX/LH2SN cells that were treated with 30 mM NH₄Cl or 30 μM chloroquine for 4 to 6 h were also performed, and no difference in cell cycle profiles was found between agent-treated and untreated cells (data not shown). The strong enhancement of ENTV entry by lysosomotropic agents may be due to the inhibitory effects of these agents on viral particle degradation in the lysosomes (see below), as has been shown previously for human immunodeficiency virus type 1 (18, 53), and to other factors described below.

The relatively mild inhibitory effect of BafA1 on ENTV entry and the apparent enhancement of ENTV entry by lysosomotropic agents also raised the possibility that ENTV Env pseudovirions that accumulated in the intracellular vesicles during the agent treatment period may have remained transduction competent and that, once these agents were withdrawn from the cells, viral transductions could quickly resume as efficiently as those in the absence of agent treatment. To test this possibility, we treated HTX/LH2SN cells with BafA1, chloroquine, or NH₄Cl in the presence of viruses for a prolonged period of time, i.e., for 16 or 48 h, depending on the cell tolerances to these treatments, and determined the effects of the treatments on viral infectivity 48 h postinfection. We found that, indeed, ENTV entry was completely blocked by 5 or 10 nM BafA1 (P < 0.01; n = 4) and by 15 mM NH₄Cl (P < 0.01; n = 4) after the prolonged treatments (Fig. 2D and F). Similarly, a 16-h treatment with 10 or 30 μM chloroquine also significantly inhibited (P < 0.01; n = 8), but interestingly failed to completely block, the ENTV entry (Fig. 2E). Further increasing the chloroquine concentration or prolonging the treatment duration was too toxic to the cells, and the results were not informative (data not shown). It is of note that chloroquine has indeed been shown previously to have no or little inhibitory effect on the entry of a number of pH-dependent viruses, including retroviruses such as equine infectious anemia virus (22) and foamy virus (44). Importantly, the entry of JSRV Env and VSV-G pseudovirions was also completely blocked, or significantly inhibited, by these prolonged treatments, whereas the entry of 10A1 was not affected, again demonstrating the specificities of these prolonged treatments for pH-dependent viral entry (Fig. 2D, E, and F). Altogether, these results strongly suggest that viral particles that accumulated in the intracellular vesicles during the 4- to 6-h agent treatment period remained transduction competent and that a prolonged treatment is required for the chemicals to completely block the entry of ENTV, and perhaps that of JSRV as well.

A pH of 4.5 or lower, but not a pH of 5.0, overcomes a BafA1-mediated block of ENTV entry.

The requirement for an extremely low pH for fusion by ENTV Env prompted us to further examine if such a low pH is also required for overcoming a BafA1-imposed block on ENTV entry. To test this, we pretreated HTX/LH2SN cells with 30 nM BafA1 for 1 h and incubated the cells with ENTV Env pseudovirions at 4°C for 2 h. Following extensive washes with cold PBS, the virion-cell complexes were incubated directly with a low-pH solution for 5 min or were preincubated at 37°C for 1 h and subsequently incubated with a low-pH solution for 5 min. None of the low-pH buffers rescued the ENTV Env pseudovirion infectivity when the cells were treated immediately after the virus binding (Fig. 3). Notably, the preincubation of the virion-cell complex for 1 h at 37°C followed by treatment with a pH 3.5, 4.0, or 4.5 solution substantially reduced the BafA1-mediated block of ENTV Env pseudovirion entry (Fig. 3) (P < 0.05 for pH 3.5 [n = 3], P < 0.01 for pH 4.0 [n = 5], and P = 0.054 for pH 4.5 [n = 5]). In contrast, treatment with a pH 5.0 solution in parallel had no effect (P = 0.35; n = 5), similar to treatment with a pH 7.0 solution (Fig. 3). Taken together, these results demonstrate that pHs of 4.5 and below, but not a pH of 5.0, overcome the BafA1 block of ENTV entry, further supporting the notion that the unusually low pH is necessary for ENTV Env-mediated fusion and cell entry. The greater inhibitory effect of BafA1 on ENTV entry observed here than in the experiments whose results are shown in Fig. 2A was likely due to the removal of unbound viruses in this experiment, suggesting that the constant presence of virus is another important factor that had contributed to the less inhibitory effect of BafA1 shown in Fig. 2A. Importantly, our results also indicate that a 37°C incubation period is critical for the low-pH rescue of ENTV entry to occur, presumably because the higher temperature is required for the formation of new vesicles, which in the presence of BafA1 can be acidified only by an extracellular low pH.

FIG. 3.

A pH of 4.5 or below overcomes a BafA1-imposed block on ENTV entry. HTX/LH2SN cells pretreated with BafA1 were incubated with GFP-encoding pseudovirions bearing ENTV Env for 2 h at 4°C. Unbound pseudovirions were removed by washing cells with cold PBS, and virus-bound cells either were directly treated with buffers at different pHs for 5 min (no incubation) or were preincubated at 37°C for 1 h in the presence of BafA1 and then treated with buffers at different pHs for 5 min (1-h incubation). Note that in both cases, the total period of infection at 37°C in the presence of BafA1 was 4 h. Percentages of infection were calculated as the numbers of infected cells relative to the number of infected untreated cells or of pH 7.0 buffer-treated cells, and values are averages ± standard deviations of results from three to six independent experiments. * indicates a P value of <0.05, and ** indicates a P value of <0.01.

Lysosomal protease inhibitors enhance the infectivity of ENTV Env pseudovirions.

For some pH-dependent viruses, low-pH-dependent lysosomal/endosomal protease activation, rather than a low pH per se, is critical for membrane fusion (5, 7, 21, 23, 24, 42, 43, 48, 49). Given the extremely low pH requirement for the fusion of ENTV Env and the potential role of the lysosomes in this process, we next asked if lysosomal protease activities are important for ENTV Env-mediated cell entry (although the ability of ENTV Env to directly induce syncytium formation and cell-cell fusion at a low pH argues against this possibility). While the infectivity of VSV-G pseudovirions was not affected by leupeptin, a broad-spectrum lysosomal protease inhibitor, the infectivity of Ebola virus GP pseudovirions was significantly reduced (P < 0.01; n = 3) (Fig. 4A), which agreed with the findings in previous reports (7, 48). Interestingly, ENTV Env pseudovirion infection was not inhibited but rather was significantly enhanced by leupeptin treatment in a concentration-dependent manner (P < 0.05; n = 4) (Fig. 4A). JSRV Env pseudovirion infection was also enhanced, but to a lesser extent (Fig. 4A). Similarly, cathepsin inhibitor III, a broad-spectrum cathepsin inhibitor, significantly inhibited Ebola virus GP pseudovirion infectivity (P < 0.01; n = 4) but had no significant effect on the entry of ENTV Env, JSRV Env, and VSV-G pseudovirions (Fig. 4B). Due to the severe cytotoxicity of this drug, we were unable to treat cells with higher concentrations or for prolonged periods of time to determine if additional effects could be achieved. Nevertheless, these results show that lysosomal proteases are unlikely to be required for ENTV or JSRV Env-mediated entry at a low pH, as opposed to the entry of other noncanonical pH-dependent viruses such as Ebola virus. The observed significant enhancement of ENTV pseudovirion infectivity by leupeptin was probably due to decreased degradation of pseudoviral particles in the lysosomes, where ENTV likely fuses.

FIG. 4.

Effects of lysosomal protease inhibitors on ENTV entry. HTX/LH2SN cells were pretreated with leupeptin (A) or cathepsin inhibitor III (B) at the indicated concentrations for 1 h and infected with GFP-encoding pseudovirions bearing ENTV Env, JSRV Env, VSV-G, or Ebola virus GP in the presence of each drug for 3 h (A) or for 4 h (B). Percentages of infection were calculated as the numbers of infected cells relative to the number of infected untreated cells, and values are averages ± standard deviations of results from two to four independent experiments; * indicates a P value of <0.05, and ** indicates a P value of <0.01.

ENTV Env pseudovirions are resistant to inactivation at low pHs (4.5 and 5.0).

The unusually low-pH-dependent fusion of ENTV Env and the possible role of lysosomes in this process also raised the question of whether or not ENTV pseudovirions are stable in an acidic-pH environment. To address this issue, we incubated the ENTV Env pseudovirions encoding GFP at 37°C in buffers at different pHs for 30 and 5 min and measured viral infectivity in HTX/LH2SN cells. As shown in Fig. 5, treatment with a pH 3.0 solution for 30 min almost completely inactivated ENTV pseudovirion infectivity (P < 0.01; n = 6), similar to the results observed for JSRV (3). Treatment at pH 3.5 and 4.0 for 30 min also substantially inactivated ENTV Env pseudovirion infectivity, by ∼75 and ∼40% (P < 0.05; n = 6), respectively (Fig. 5). In sharp contrast, treatment at pH 4.5 or 5.0 for 30 min had no significant effect (Fig. 5) on ENTV Env pseudovirion infectivity, and this result was slightly different from that for JSRV Env pseudovirions, whose infectivity was reduced by ∼50% following treatment at pH 4.5 (3). Noticeably, the treatment of ENTV Env pseudovirions for 5 min at 37°C substantially decreased the efficiencies of inactivation of ENTV infectivity at low pHs, in particular, pH 3.0, 3.5, and 4.0 (Fig. 5), suggesting that longer incubation is required for these low pHs to completely inactivate the ENTV infectivity. Taken together, our data show that ENTV Env pseudovirions are virtually resistant to low-pH inactivation and appear to be more stable than JSRV Env pseudovirions in acidic environments, which supports the extremely low pH requirement of ENTV Env for fusion. However, we cannot exclude the possibility that other factors besides a low pH are involved in or required for the low-pH inactivation (see Discussion).

FIG. 5.

ENTV Env pseudovirions are resistant to low-pH inactivation. GFP-encoding MoMLV pseudovirions bearing ENTV Env were treated with different solutions ranging from pH 3.0 to 7.0 at 37°C for 30 or 5 min. Following neutralization with culture medium, virions were used to infect HTX/LH2SN cells, and the percentages of infected cells were measured 48 h postinfection by flow cytometry. The percentage of viral infectivity at each pH was calculated relative to that of the pH 7.0 solution-treated viral stocks (set to 100%); the results shown are averages ± standard deviations of results from four or six independent experiments. * indicates a P value of <0.05, and ** indicates a P value of <0.01.

The TM subunit of ENTV Env is responsible for the ability of the protein to fuse at an unusually low pH, and the SU subunit contributes to its low and pH-dependent fusion activity.

We generated two chimeras from ENTV and JSRV Envs in order to dissect the respective roles of the ENTV Env SU and TM in the process of low-pH-dependent fusion activation. The first chimera, referred to as EJ, consisted of the ENTV SU and the JSRV TM. The second chimera, JE, consisted of the JSRV SU and the TM of ENTV. The synthesis and processing of each of these Env chimeras were similar to those of the other chimera and their parental Env molecules, according to metabolic labeling assays with 293T cells (data not shown) and an immunoblotting analysis of virus producer cell lysates (see below and Fig. 7A).

FIG. 7.

Env expression by ENTV, JSRV, and their chimeras in virus producer cells (A) and in pseudovirion particles (B). 293/GP-LAPSN cells were transfected with a plasmid encoding each Env, and supernatants containing pseudovirions were harvested 48 to 72 h posttransfection. (A) Cell lysates were analyzed by immunoblotting using anti-FLAG antibody (upper panel); the same membrane was then stripped and reblotted using anti-β-actin (lower panel). (B) Pseudovirions were purified on a sucrose cushion, and Env incorporation was analyzed by immunoblotting using an anti-FLAG antibody (upper panel). The membrane was stripped and reblotted with anti-MLV Gag (lower panel). The densitometric quantification of TM (upper panels) or Gag (lower panels) bands was performed using the Quantity One software (Bio-Rad), and the relative intensities were obtained by setting the signals of JSRV Env to 1.00. No Env, parental 293/GP-LAPSN cells transfected with no plasmid. Results representative of those from three to four independent experiments are shown.

Syncytium induction was performed to examine the fusogenicities of these chimeras relative to those of their parental Envs. As shown in Fig. 6A, the replacement of the ENTV TM with that of JSRV, as in EJ, resulted in almost 100% syncytium induction at pH 4.5 and ∼80% at pH 5.0, and small syncytia were still detectable at pH 6.0, a pattern similar to that induced by JSRV Env. Of note, the percentages and the sizes of syncytia induced by EJ generally were smaller than those of syncytia induced by JSRV Env, particularly at pH 4.5 and above (Fig. 6A), suggesting that the SU subunit of ENTV Env influences the protein's fusogenicity. Conversely, the replacement of the JSRV TM subunit with that of ENTV, as in JE, essentially blocked the fusion activity of JSRV Env at pH 5.0 but permitted fusion at pH 4.0 or 4.5, a pattern which was similar to that of ENTV Env (Fig. 6A). Again, the fusion activity of JE was higher than that of ENTV Env under these low-pH conditions (pH 4.0 and 4.5) (Fig. 6A).

To confirm the above-described results and better quantify the fusion activities of these chimeric Envs relative to those of the parental Envs, cell-cell fusion assays were carried out at pH 4.0, 5.0, or 7.0. No cell-cell fusion was detected for ENTV Env at pH 5.0 following a 1-min treatment (Table 1 and Fig. 6B), consistent with the results of the syncytium induction assay shown in Fig. 1 and Fig. 6A. Interestingly, we detected only a very low level of cell-cell fusion activity at pH 4.0 (Table 1 and Fig. 6B), compared to the substantial syncytium formation at pH 4.0 shown in Fig. 1A and Fig. 6A. In contrast, chimeric EJ, containing the JSRV TM, showed significantly higher fusogenicity than did the parental ENTV Env at both pH 5.0 (P = 0.003; n = 4) and pH 4.0 (P = 0.028; n = 4) (Fig. 6B). Yet the fusion activity of EJ was still lower than that of JSRV Env (P = 0.01 and 0.04 for pH 4.0 and 5.0, respectively; n = 4) (Fig. 6B). In agreement with the results of the syncytium induction assay (Fig. 6A), chimeric JE, containing the ENTV TM, failed to induce cell-cell fusion at pH 5.0 and also exhibited much less fusion activity at pH 4.0 than did the parental JSRV Env (P = 0.01; n = 4) (Fig. 6B). We noticed that the fusion activity of JE was still higher than that of ENTV Env at pH 4.0 (P = 0.02; n = 4) (Fig. 6B). Altogether, our data suggest that the TM subunit of ENTV Env is likely to be responsible for the unusually low pH requirement for fusion but that the SU subunit may modulate the low and pH-dependent fusion activity of ENTV Env.

Flow cytometric analysis of a portion of 293T/GFP effector cells revealed that the level of surface expression of the SU of EJ was in fact higher than that of JSRV Env but was comparable to that of ENTV Env (Fig. 6C). By contrast, the level of surface expression of the SU of JE was lower than that of ENTV Env yet was similar to that of JSRV Env (Fig. 6C). These results indicated that the lower fusogenicities of EJ and ENTV Env than those of JSRV Env and JE, respectively, were not due to Env expression on the cell surface. Collectively, these results revealed that the TM subunit of ENTV Env is responsible for the protein's unusually low-pH-dependent fusion activation and that the SU subunit of ENTV Env also critically influences its low and pH-dependent fusogenicity.

Incorporation of ENTV Env and its chimeras into a MoMLV retroviral vector and efficiencies of transduction of cells expressing different levels of Hyal2.

Previous studies showed that ENTV has a limited host range compared with that of JSRV, based on the finding that a MoMLV vector bearing ENTV Env did not effectively transduce cell lines derived from most species, including humans and sheep (14). To determine if this limitation is due to a defect in the incorporation of ENTV Env into the MoMLV vector or to other factors, such as receptor binding or possibly fusion, here we generated AP-encoding MoMLV pseudotypes bearing ENTV Env, JSRV Env, or their chimeras and examined Env expression, incorporation, and vector transduction efficiencies in cells expressing different levels of Hyal2.

Env expression in virus producer cells and in viral particles was examined by immunoblotting using an anti-FLAG antibody. As shown in Fig. 7A, ENTV Env and its two chimeras were expressed efficiently and processed properly into the SU and TM in the virus producer cells, with efficiencies similar to that for JSRV Env. Of note, ENTV Env, in particular the SU subunit, migrated slightly slower than JSRV Env, likely due to their differences in molecular weight and posttranslational modification (e.g., there is one extra potential N-linked glycosylation site in the ENTV Env SU than in the JSRV Env SU) (Fig. 7A). Interestingly, the signal for the SU subunit of JSRV Env derived from cell lysates was consistently weaker than that for ENTV Env (Fig. 7A), which was in parallel with the relatively low level of JSRV Env SU surface expression on 293/LH2SN and 293T/GFP cells (Fig. 1C and 6C). The reason behind the low-level expression of the JSRV SU relative to that of the ENTV SU was not investigated further but may be associated with the different trafficking and endocytosis signals present in their respective Env CTs (30).

We found that ENTV Env was incorporated into the MoMLV vector as efficiently as JSRV Env (Fig. 7B), although this result did not take into account the relatively low levels of JSRV Env surface expression (∼50%) compared to those of ENTV Env surface expression (Fig. 1C and 6C). Indeed, ∼3-fold more EJ than ENTV Env was incorporated and ∼2-fold less JE than JSRV Env was incorporated into the vectors, based on quantitative analyses of Env and MoMLV Gag protein signals detected by immunoblotting (Fig. 7B). These data suggest that the TM subunit of ENTV Env is not as compatible as that of JSRV Env with incorporation into the MoMLV vector, as has been suspected previously (14).

The transduction efficiencies of MoMLV vectors bearing individual Envs for several cell lines was determined. In HTX cells expressing endogenous Hyal2, transduction by ENTV Env pseudovirions was extremely poor, with ∼26 AP⁺ foci per ml, which was over 1,000-fold lower than the result for JSRV Env (Table 2). The replacement of the ENTV Env TM with that of JSRV Env (as in EJ) increased the pseudovirion titer by ∼6-fold (Table 2), similar to the findings in a previous report (14, 51). Impressively, the replacement of the ENTV SU with that of JSRV (as in JE) significantly increased the ENTV titer to yield ∼10⁴ AP⁺ foci per ml, a level nearly comparable to that for JSRV Env (Table 2). In HTX or mouse NIH 3T3 cells that overexpress human Hyal2, the titers of all Env pseudovirions were significantly increased, particularly those of pseudovirions bearing ENTV Env and EJ (Table 2). Collectively, these results reveal that ENTV Env SU-Hyal2 interaction is likely to be the key determinant for the low infectivity of ENTV Env pseudovirions in HTX cells and most other cells expressing endogenous Hyal2.

TABLE 2.

Titers of MoMLV pseudovirions bearing ENTV Env, JSRV Env, or their chimeras in cells expressing different levels of Hyal2^a

Cell line	No. of AP+ foci/ml among cells expressing:
	ENTV Env	JSRV Env	EJ	JE
HTX	26 ± 2	(4.6 ± 2.8) × 104	1.6 × 102	(1.0 ± 0.5) × 104
HTX/LH2SN	(1.8 ± 0.7) × 105	(1.4 ± 0.3) × 106	(1.3 ± 0.7) × 106	(2.3 ± 1.2) × 105
NIH 3T3/LH2SN	1.5 × 104	(2.3 ± 1.4) × 105	(3.2 ± 1.2) × 105	(2.5 ± 1.4) × 104

^a293/GP-LAPSN cells expressing the MoMLV Gag-Pol proteins and AP were transfected with a plasmid encoding each Env, and viral supernatants were harvested 48 to 72 h posttransfection. Cells were infected with appropriate amounts of viral stocks in the presence of Polybrene, and the numbers of AP-positive foci were determined ∼72 h postinfection. The results shown are averages ± standard deviations of results from two to three independent experiments.

Noticeably, EJ pseudovirions exhibited a titer that was approximately 10- to 20-fold higher than that of ENTV Env pseudovirions and JE pseudovirions exhibited a titer that was ∼5- to 10-fold lower than that of JSRV Env pseudovirions in all three cell lines tested (Table 2). These titer differences cannot be explained simply by the two- to threefold decrease in ENTV Env incorporation indicated in Fig. 7B. One possible explanation is that Env pseudovirions containing the ENTV TM (i.e., JE and ENTV Env pseudovirions) may be more prone to lysosomal degradation due to their extremely low pH requirement for fusion than Env pseudovirions containing the JSRV TM (i.e., JSRV Env and EJ pseudovirions). Indeed, the treatment of HTX/LH2SN cells with 50 μM leupeptin substantially increased the titer of JE pseudovirions, by ∼33%, a degree that was similar to that of the increase in the titer of ENTV Env pseudovirions (Fig. 4A), while increasing the titer of EJ pseudovirions by only ∼10%, which was comparable to the increase in the titer of JSRV Env pseudovirions (Fig. 4A). These data therefore support the notion that the TM subunit of ENTV Env can modulate the vector transduction efficiency not only by slightly reducing the level of Env incorporation into the MoMLV vector but also by affecting the pH-dependent fusion process.

DISCUSSION

We show in this study that the fusogenicity of ENTV Env protein is much lower than that of JSRV Env and that ENTV Env requires a more acidic pH for fusion activation and cell entry (Fig. 1 and 6 and Table 1). The threshold pH value that activates ENTV Env-mediated fusion was estimated to be ∼4.5, which is much lower than that for JSRV Env (pH ∼6.0) (Fig. 1 and 6) and other typical pH-dependent viral glycoproteins (pH 4.8 to 6.5) (20), such as influenza virus hemagglutinin (pH 5.7) (50, 59), VSV-G (pH 6.3) (32, 34, 55), and Semliki Forest virus E1 (pH 6.2) (54, 55). While the cell-based fusion assays do not always mimic productive viral infection (15) and it is possible that ENTV may fuse with the cell membrane at pH 5.0 during infection in vivo, the differences between ENTV and JSRV Env proteins in fusogenicity, as well as in low-pH requirements for fusion activation in vitro, are intriguing. It would be ideal to confirm the results shown in this study with an infectious ENTV clone; however, a tissue culture system that can produce high titers of ENTV, which would allow us to perform such related experiments, is not currently available. The ENTV Env used in this study was derived from the first ENTV strain for which the full-length genome sequence was determined (also called ENTV1; GenBank accession numbers Y16627 and AF401741), which is considered to represent all other ENTV strains isolated thus far (10, 14, 51); however, we cannot rule out the possibility that other ENTV Env isolates may have somewhat different fusion activities and low-pH requirements.

The low fusogenicity of ENTV Env and its particular requirement for a very acidic pH for fusion are not due to abnormal Env trafficking, processing, or expression. In fact, the ENTV Env protein was synthesized and processed very efficiently (Fig. 7A), with levels of SU subunit surface expression (Fig. 1C and 6C) and basal levels of SU secretion into the culture medium that were consistently higher than those of JSRV Env (data not shown). Using chimeras constructed from JSRV and ENTV Envs, we demonstrated that the TM subunit of ENTV Env appears to be responsible for the extremely low pH requirement and largely accounts for the low fusion activity of the protein (Fig. 6A and B). This finding is consistent with the general notion that a low pH acts directly on the TM subunits of pH-dependent retroviral Env, triggering a series of conformational rearrangements that are required for membrane fusion (15, 16). It is also of note that the major difference between ENTV and JSRV Envs is located in the TM subunit, with ∼82% amino acid identity in this region compared to an overall ∼89% amino acid identity of the full-length Env proteins (1, 10, 14). Future efforts will be aimed at determining what regions or specific domains of the ENTV Env TM subunit are responsible for the extremely low pH requirement for fusion and at investigating the underlying mechanisms.

We demonstrated in this study that cell entry mediated by ENTV Env is also pH dependent but found, unexpectedly, that the patterns of ENTV entry are quite different from those of JSRV entry. First, a pH of 4.0 or 4.5, but not a pH of 5.0, substantially overcame a BafA1-mediated block of ENTV entry (Fig. 3), while the BafA1-mediated block of JSRV entry was readily eliminated by treatment at pH 5.0 (3). Second, although the infectivity of ENTV Env pseudovirions was inhibited by BafA1, the extent of this inhibition was consistently lower than that of JSRV pseudovirions over the 6-h incubation period (Fig. 2A). Moreover, the inhibition was not concentration dependent, which was also different from the inhibition of JSRV and VSV (Fig. 2A). Third, NH₄Cl and, to a lesser extent, chloroquine were not able to inhibit, but rather substantially enhanced, the entry of ENTV Env pseudovirions following a 4-h treatment period (Fig. 2B and C). Again, this result was in sharp contrast to that for JSRV pseudovirions, whose entry was substantially inhibited by similar treatments (Fig. 2B and C). Lastly, the treatment of cells with leupeptin, a broad-spectrum lysosomal protease inhibitor, appeared to enhance ENTV entry more significantly than JSRV entry (Fig. 4A). Taken together, our data strongly suggest that the endocytic compartments that mediate the fusion of ENTV Env are different from those required by JSRV Env. For example, ENTV Env-mediated fusion and cell entry may occur predominantly in more-downstream endocytic compartments, such as lysosomes (pH ∼4.0 to 5.0), while JSRV Env fusion may occur predominantly in late endosomes (pH ∼5.0 to 6.0). Thus, ENTV Env pseudovirions may be more prone to lysosomal degradation than JSRV Env pseudovirions, and this characteristic would explain the significant enhancement of ENTV infectivity by leupeptin relative to that of JSRV (Fig. 4A) and would also explain, at least partially, the differential effects of BafA1 and lysosomotropic agents on ENTV and JSRV infection during the 4- to 6-h treatment period (Fig. 2A to C). However, it should be emphasized that the less inhibitory effect of BafA1 and the apparent enhancement of ENTV infection by lysosomotropic agents following the short (4- to 6-h) incubation were also attributable to, and in conjunction with, other factors, such as the constant presence of viruses with the agent (Fig. 2A to C) and the ability of ENTV Env pseudovirions to resume transduction upon the withdrawal of these chemicals (Fig. 2D to F).

While virus fusion in cellular lysosomes is conceivably not advantageous for viral fitness, it is noticeable that a number of classical pH-dependent enveloped viruses, including influenza virus (33, 58), and some noncanonical pH-dependent viruses, such as Ebola virus, severe acute respiratory syndrome coronavirus, and MoMLV (7, 21, 24, 43, 48, 49), have been shown or suggested previously to use lysosomes and/or late endosomes to enter host cells. In this regard, it will be interesting to determine if ENTV Env-mediated fusion truly occurs in lysosomes or if other endocytic vesicles, such as late endosomes, are also involved in this process. Ultimately, the exact entry pathways of ENTV and JSRV will need to be elucidated.

For a typical pH-dependent virus, viral infectivity should be inactivated by a low pH because of premature conformational changes that can block further infection (15). However, we found in this study that ENTV Env pseudovirions are quite resistant to low-pH inactivation. For example, the infectivity of ENTV Env pseudovirions decreased only slightly after 30 min in a pH 4.5 or 5.0 buffer at 37°C (Fig. 5). This result was in contrast to that for JSRV, whose infectivity was reduced by almost 50% at pH 4.5 (3). While these results may be used to argue against the pH-dependent fusion of ENTV Env, we cannot rule out the possibility that there is still a pretriggering at pH 4.0 and below (Fig. 5) and that the conformational changes of ENTV Env induced by a low pH are reversible, as has been shown previously for VSV-G (45). Alternatively, a receptor-binding step may be required for the low-pH inactivation to occur, and if this is true, ENTV would resemble avian sarcoma and leukosis virus (ASLV), which uses a two-step fusion process (39). Although we favor the latter explanation, given some additional lines of evidence discussed below, we cannot rule out the first possibility.

The two-step fusion model described for ASLV states that a specific interaction between ASLV Env and its corresponding receptor induces the initial conformational changes of Env that are required to prime the subsequent low-pH trigger (39). We show here that, while the TM subunit of ENTV Env plays a major role in dictating the extremely low pH requirement for fusion and, therefore, largely accounts for the low level of fusion activity, the SU subunit of ENTV Env also critically modulates the protein's low-level and pH-dependent fusion activity (Fig. 6A and B; Table 1). This finding raises the possibility that ENTV Env SU-Hyal2 interaction may regulate the pH-dependent fusion process. Indeed, the binding affinity of the ENTV SU for human Hyal2 has been shown to be much lower than that of the JSRV SU for human Hyal2 (∼175 nM for ENTV versus ∼10 nM for JSRV) (26, 51), which may explain the relatively low fusogenicities of Env constructs harboring the ENTV Env SU (Fig. 6). However, we cannot rule out the possibility that other regions of ENTV SU not directly involved in receptor binding, such as the proline-rich motifs, may also regulate fusion, as has been shown previously for MLV (25). In addition, we have attempted to detect the oligomerization of ENTV Env in the viral particles at low pHs but have repeatedly failed to observe the formation of high-molecular-weight species by SDS-polyacrylamide gel electrophoresis (data not shown), again suggesting that other factors besides a low pH may be involved in the low-pH inactivation process. Taking our findings together, we propose that Hyal2 may play a very important role in the pH-dependent fusion of ENTV Env (and JSRV Env as well); however, additional experiments are required to critically test this hypothesis. Whether or not ENTV utilizes a two-step process for fusion and cell entry, like ASLV, will be explored in future studies.