The Bipartite Sequence Motif in the N and C Termini of gp85 of Subgroup J Avian Leukosis Virus Plays a Crucial Role in Receptor Binding and Viral Entry

Infection of a cell by retroviruses requires the attachment and fusion of the host and viral membranes. The specific adsorption of envelope (Env) surface proteins to cell receptors is a key step in triggering infections and has been the target of antiviral drug screening. ALV-J is an economically important avian pathogen that belongs to the genus Alpharetrovirus and has a wider host range than other ALV subgroups. Our results showed that the amino acids 38 to 131 of the N terminus and 159 to 283 of the C terminus of ALV-J gp85 controlled the efficiency of gp85 binding to chNHE1 and were critical for viral infection. In addition, the glycosylation sites (N6 and N11) and cysteines (C3 and C9) of gp85 played a crucial role in the receptor binding and viral entry. These findings might help elucidate the mechanism of the entry of ALV-J into host cells and provide antiviral targets for the control of ALV-J.

ABSTRACT

Subgroup J avian leukemia virus (ALV-J), belonging to the genus Alpharetrovirus, enters cells through its envelope surface unit (gp85) via specifically recognizing the cellular receptor chicken Na⁺/H⁺ exchanger type I (chNHE1), the 28 to 39 N-terminal residues of which were characterized as the minimal receptor functional domain in our previous studies. In this study, to further clarify the precise organization and properties of the interaction between ALV-J gp85 and chNHE1, we identified the chNHE1-binding domain of ALV-J gp85 using a series of gp85 mutants with segment substitutions and evaluating their effects on chNHE1 binding in protein-cell binding assays. Our results showed that hemagglutinin (HA) substitutions of amino acids (aa) 38 to 131 (N terminus of gp85) and aa 159 to 283 (C terminus of gp85) significantly inhibited the interaction between gp85 and chNHE1/chNHE1 loop 1. In addition, these HA-substituted chimeric gp85 proteins could not effectively block the entry of ALV-J into chNHE1-expressing cells. Furthermore, analysis of various N-linked glycosylation sites and cysteine mutants in gp85 revealed that glycosylation sites (N6 and N11) and cysteines (C3 and C9) were directly involved in receptor-gp85 binding and important for the entry of ALV-J into cells. Taken together, our findings indicated that the bipartite sequence motif, spanning aa 38 to 131 and aa 159 to 283, of ALV-J gp85 was essential for binding to chNHE1, with its two N-linked glycosylation sites and two cysteines being important for its receptor-binding function and subsequent viral infection steps.

IMPORTANCE Infection of a cell by retroviruses requires the attachment and fusion of the host and viral membranes. The specific adsorption of envelope (Env) surface proteins to cell receptors is a key step in triggering infections and has been the target of antiviral drug screening. ALV-J is an economically important avian pathogen that belongs to the genus Alpharetrovirus and has a wider host range than other ALV subgroups. Our results showed that the amino acids 38 to 131 of the N terminus and 159 to 283 of the C terminus of ALV-J gp85 controlled the efficiency of gp85 binding to chNHE1 and were critical for viral infection. In addition, the glycosylation sites (N6 and N11) and cysteines (C3 and C9) of gp85 played a crucial role in the receptor binding and viral entry. These findings might help elucidate the mechanism of the entry of ALV-J into host cells and provide antiviral targets for the control of ALV-J.

INTRODUCTION

Avian leukosis virus (ALV) belongs to the Alpharetrovirus genus of the Retroviridae family. Based on their host range, receptor specificity, interference patterns, and cross-reactivity with neutralizing antibodies, ALVs are divided into 10 different subgroups (designated A through J) (1). Recently, a new subgroup named K subgroup has been reported (2). The interaction of the glycoproteins of the viral envelope (Env) with specific cellular receptor proteins is known to be the primary and key stage of the infection by retroviruses. The Env glycoprotein of ALVs is a typical type I transmembrane protein, composed of a surface (SU) and a transmembrane (TM) unit. The env gene of ALV encodes a precursor polyprotein, which is then assembled into a trimer at the endoplasmic reticulum (3). After their glycosylation in Golgi, polyproteins are cleaved into the SU (gp85) and TM (gp37) units. Like other retroviruses, the specific interaction between the gp85 of ALV and its cellular receptor triggers a conformational change in the trimeric envelope glycoprotein, exposing the fusion peptide in gp37 to mediate the fusion of the viral and cellular membranes, which is followed by the delivery of the viral genome into the cytoplasm (4–7). In contrast to the lentiviruses of human immunodeficiency virus (HIV), feline immunodeficiency virus (FIV), and simian immunodeficiency virus (SIV) that employ a dual coreceptor to invade cells (8–10), ALVs, which are simple retroviruses similar to murine leukemia viruses (MLV) (11), require only a single functional receptor to infect target cells. The interactions between the ALV Env proteins and their receptors have been widely used as a model system for studying retroviral entry because different ALV subgroups have been reported to use different cellular receptors for entry into cells (5, 6, 12, 13).

Several cellular receptors have been identified for the ALV subgroups. A receptor for ALV-A and ALV-K has been shown to be related to the low-density lipoprotein receptor called tva (14, 15). Alleles of tvb, which are known to belong to the tumor necrosis factor receptor, have been shown to control the susceptibility to ALV-B, ALV-D, and ALV-E (tvbS1 for subgroups B, D, and E, as well as tvbS3 for subgroups B and D) (16, 17). The tvc gene encoding a 488-amino-acid protein related to mammalian butyrophilins, which are members of the immunoglobulin protein family, has been reported as the receptor of ALV-C (18). The chicken Na⁺/H⁺ exchanger type 1 (chNHE1) has been identified as the functional cellular receptor of ALV-J (19). The subgroup-specific receptor usage determinants of ALVs are known to be mainly located within the central region of the gp85, where contained sequences are known to vary among the ALV subgroups, thus taking charge of recognizing different cell receptors (20).

The gp85-encoding sequences of the A to E ALV subgroups are highly conserved, showing 80% to 85% identity to each other, except for five variable regions in the central region of gp85, including vr1, vr2, hr1, hr2, and vr3 (11, 21). Several studies concentrating on the receptor-binding affinity of the envelope glycoproteins of ALV-A to ALV-E have suggested that the main binding interactions between the envelope glycoproteins and the host receptor occur in the two host range-determining regions, hr1 and hr2 (22–25). The vr1 and vr2 regions have been reported not to be essential for receptor specificity, whereas vr3 might contribute to receptor recognition but does not appear to participate in binding specificity (23). The sequence of ALV-J, however, has been demonstrated to exhibit large divergence from those of other ALV subgroups. The sequence encoding the gp85 domain of the ALV-J prototype strain HPRS-103 was shown to be 93 to 126 bases shorter than those of viruses from the A to E subgroups, with only 40% overall average identity to the corresponding sequences of other subgroups (26). Therefore, we speculated that the variable regions and receptor usage determinants in gp85 of ALV-J might totally differ from those of other ALV subgroups. Our previous study demonstrated that the minimal functional domain responsible for chNHE1 binding of ALV-J gp85 and efficient mediation of ALV-J cell entry were located in the membrane-proximal region of the N terminus of the extracellular loops 1 of chNHE1 (27). This geometric location would not allow the chNHE1 to reach a binding pocket at the side of gp85 of ALV-J, suggesting that the receptor-binding domain on ALV-J gp85 might likely be located on its apex. To determine the precise organization and properties of the interaction between gp85 of ALV-J and chNHE1 and better understand the mechanism of the invasion of the host cell by ALV-J, we expressed, in the present study, soluble forms of wild-type and chimeric ALV-J gp85 and tested their chNHE1-binding abilities using a protein-cell binding assay. We also tested their ability to block the receptor of the ALV-J entry. Our results revealed, for the first time, that the chNHE1-binding domain was located on the bipartite sequence motif spanning amino acids (aa) 38 to 131 in the N terminus and aa 159 to 283 in the C terminus of the gp85 region of ALV-J, and identified two glycosylation sites (N6 and N11) and two cysteines (C3 and C9) in gp85, which served as critical determinants for its receptor binding and viral entry.

RESULTS

Design and expression of the ALV-J chimeric gp85 for binding domain mapping studies.

To map the chNHE1-binding domain of the ALV-J gp85, we designed a panel of chimeric gp85 proteins containing a hemagglutinin (HA; a 9-aa linear epitope peptide) substitution in different sites (Fig. 1A). A total of 33 chimeric gp85 (HA1 to HA33) were successfully expressed in 293T cells (Fig. 1B). Results of Western blotting showed that all 33 chimeric gp85 proteins migrated with the same electrophoretic mobility as that of wild-type gp85 and could react with an anti-HA monoclonal antibody (Fig. 1C).

FIG 1.

Design and expression of ALV-J chimeric gp85. (A) Schematic representation of the strategy for constructing a chimeric gp85 protein. The amino acid sequence of wild-type gp85 was divided into 33 segments, which were sequentially substituted with an HA tag to construct the chimeric gp85 plasmids. (B) SDS-PAGE analysis of the wild-type and 33 chimeric gp85 proteins. Plasmids containing wild-type and chimeric gp85 with HA tag substitutions were transfected into 293T cells. At 48 h posttransfection, the harvested supernatant of cells was centrifuged and purified using protein A resin. Subsequently, the purification of the gp85 proteins was verified by SDS-PAGE analysis. (C) Western blotting of the wild-type and 33 chimeric gp85 proteins. The wild-type or chimeric purified gp85 proteins were separated on 12.5% SDS-PAGE gels and analyzed by Western blotting with anti-HA monoclonal antibody. After incubation with IRDye 680RD secondary antibody, proteins were visualized using the Odyssey system. (D) Calculated secondary structures of the wild-type and 33 chimeric gp85 proteins based on CD data. From bottom, bar segments represent α-helix (blue), β-strand (firebrick), turns (yellow), and other conformations (green). CD data are shown as the means of three independent experiments.

To evaluate whether HA tag substitutions affected the stability of chimeric gp85 proteins, we assessed the similarity to the secondary structure of wild-type gp85 using circular dichroism (CD). We analyzed the chimeric and wild-type gp85 proteins for differences in their secondary structure at 25°C. Our results indicated that the CD spectra of the wild-type and 33 chimeric gp85 proteins exhibited a similar percentage of α-helix, β-strand, and turns, which suggested that all chimeric gp85 proteins maintained a similar structure and stability to that of the wild-type gp85 protein (Fig. 1D).

Mapping the chNHE1-binding domains of ALV-J gp85.

To map the chNHE1-binding domain of ALV-J gp85 that mediates the interaction with the chNHE1 receptor, we measured the binding capacities of different chimeric gp85 proteins to chNHE1-expressing 293T cells using the protein-cell binding assay (27). The results illustrated in Fig. 2A demonstrated that the HA1 to HA4 (aa 1 to 37) tag substitution on the N terminus of gp85, the HA15 to HA17 (aa 132 to 158) on the middle, and HA32 to HA33 (aa 284 to 304) on the C terminus of gp85 were able to bind chNHE1 with greater than 75% efficiency compared with the wild-type gp85. In contrast, the HA5 to HA14 (aa 38 to 131), HA18, and HA20 to HA31 (aa 159 to 283) chimeric gp85 were shown to almost abrogate the ability for chNHE1 binding.

FIG 2.

Mapping the chNHE1-binding domain of ALV-J gp85. (A) Binding capacity of the wild-type and 33 chimeric gp85 proteins to chNHE1 expressed on the surface of transfected 293T cells. The binding capacity of wild-type gp85 (gp85 wt) was set as 100%, with the values of chimeric gp85 proteins calculated as its proportion. The values of transfected 293T cells without protein incubation were set as the negative control (NC). (B) SDS-PAGE analysis of the soluble chECL1. Soluble chECL1 plasmids were transfected into 293T cells. At 48 h posttransfection, the harvested culture supernatant was centrifuged and purified using protein A resin. Subsequently, the purification of the soluble chECL1 was verified by SDS-PAGE analysis. (C) The binding affinity between soluble chECL1 and wild-type and 33 chimeric gp85 proteins was analyzed by solid-phase binding ELISA described in Materials and Methods. The OD value of wild-type gp85 (gp85 wt) was set as 1, with the OD values of chimeric gp85 proteins calculated as its proportion. The values of protein-coated wells without protein incubation were set as the negative control (NC). Three independent experiments were performed, and data are shown as means ± SDs of triplicates from a representative experiment.

We previously demonstrated that the extracellular loop 1 of chNHE1 (chECL1) was the critical functional ECL interacting directly with the ALV-J gp85 to mediate viral entry (27). To further determine the binding affinity of different chimeric gp85 proteins to their receptors, we expressed the soluble chECL1 in 293T cells (Fig. 2B). We then tested the binding affinity of different chimeric gp85 proteins to the soluble chECL1 using solid-phase binding enzyme-linked immunosorbent assay (ELISA) (28). Results of ELISA absorbance showed that the binding affinities of the HA1 to HA4, HA15 to HA17, and HA31 to HA32 chimeric gp85 to the soluble chECL1 were similar to those of the wild-type gp85, with optical density (OD) values of more than 0.7 (Fig. 2C), whereas the binding affinities of the HA5 to HA14 (aa 38 to 131) and HA18 to HA31 (aa 159 to 283) chimeric gp85 proteins to soluble chECL1 were demonstrated to be significantly reduced compared with the wild-type gp85, with OD values less than 0.4 (Fig. 2C). The results of ELISA were almost consistent with the results from the above binding assay. In addition to HA19 chimeric gp85, the protein-cell binding assay showed that HA19 could bind chNHE1 with an efficiency of about 60%, although ELISA results indicated that it could not bind soluble chECL1. Taken together, these results suggested that the chNHE1-binding domain of gp85 was located on the bipartite sequence motif spanning aa 38 to 131 and aa 159 to 283.

Receptor blockade of entry of ALV-J by chimeric gp85 proteins.

To verify the mapping results of the chNHE1-binding domain of ALV-J gp85, we further tested the chimeric gp85 protein for its ability to block the entry of ALV-J into cells. First, the replication-competent avian leucosis sarcoma (RCAS)(J)-enhanced green fluorescent protein (EGFP) recombinant virus was successfully rescued in DF-1 cells and reached a titer of 10^4.36 50% tissue culture infective dose (TCID₅₀)/ml. The subgroup specificity of the RCAS(J)-EGFP recombinant virus was confirmed by indirect immunofluorescent assay (IFA) with 4A3 anti-ALV-J gp85 monoclonal antibody (MAb) (29) (Fig. 3A). Then, we determined the optimum concentrations of gp85 to block the ALV-J receptor to be used for preincubation with DF-1 cells. As shown in Fig. 3B, 200 μg/ml of wild-type gp85 protein could almost block the entry of the RCA(J)-EGFP recombinant virus into DF-1 cells. Therefore, we used the concentration of 200 μg/ml in the subsequent block assay of different chimeric gp85 proteins. We measured the blocking capacities of different chimeric gp85 proteins to viral entry using fluorescence microscopy, flow cytometry, and real-time PCR. Our results showed that preincubation of DF-1 cells with the HA1 to HA4, HA15 to HA17, and HA32 to HA33 chimeric gp85 significantly reduced the number of green fluorescence-positive cells (less than 40%) infected with the RCAS(J)-EGFP recombinant virus, similar to the blocking effect of the wild-type gp85 (Fig. 3C and D). Preincubation of cells with HA19 reduced green fluorescence-positive cells by about 55%. However, preincubation of DF-1 cells with the HA5 to HA14, HA18, and HA20 to HA31 chimeric gp85 had almost no influence on the positive rate (over 90%) of green fluorescence of RCAS(J)-EGFP recombinant virus-infected cells, exhibiting a lack of a blocking effect on viral entry (Fig. 3C and D).

FIG 3.

Receptor blockade of gp85 proteins. (A) Immunofluorescent detection of RCAS(J)-EGFP recombinant virus. DF-1 cells were infected with different RCAS(J)-EGFP recombinant virus for 72 h and then immunostained with an 4A3 monoclonal antibody and analyzed using the fluorescence microscope. Normal DF-1 cells were used as negative control (Mock). (B) Wild-type gp85 protein blocks the entry of the RCAS(J)-EGFP recombinant virus into DF-1 cells. DF-1 cells were incubated with different concentrations of wild-type gp85 or Fc protein at 4°C for 1 h. The RCAS(J)-EGFP recombinant virus was incubated with DF-1 cells at 37°C for 1 h. To detect viral entry, cells infected with the RCAS(J)-EGFP recombinant virus were identified by flow cytometry at 48 h postinfection. The Fc protein was used as negative control. (C) Immunofluorescent detection of the wild-type or 33 chimeric gp85 proteins blocking the entry of the RCAS(J)-EGFP recombinant virus into DF-1 cells. DF-1 cells were incubated with different gp85 proteins at 4°C for 1 h. The RCAS(J)-EGFP recombinant virus was incubated with DF-1 cells at 37°C for 1 h. To detect viral entry, cells infected with the RCAS(J)-EGFP recombinant virus were visualized under a fluorescence microscope at 48 h postinfection. Virally infected normal DF-1 cells (without protein treatment) were used as the positive control (PC). Normal DF-1 cells (without protein treatment) were used as the negative control (NC). (D) Flow cytometry detection of the wild-type or 33 chimeric gp85 proteins blocking the entry of the RCAS(J)-EGFP recombinant virus into DF-1 cells. DF-1 cells were incubated with different gp85 proteins at 4°C for 1 h and infected with the RCAS(J)-EGFP recombinant virus. Infected DF-1 cells were detached using a solution of PBS and 5 mM EDTA at 48 h postinfection, and their fluorescence was determined by flow cytometry. The positive rate of cells of the positive control was set as 1, with the values of chimeric gp85 proteins calculated as its proportion. (E) The RCAS(J)-EGFP recombinant virus was detected by quantitative PCR (qPCR) analysis. DF-1 cells were incubated with wild-type or 33 chimeric gp85 proteins at 4°C for 1 h and infected with the RCAS(J)-GFP recombinant virus. To detect viral copies, cell cultures were harvested and subjected to qPCR analysis at 48 h postinfection. Virally infected normal DF-1 cells (without protein treatment) were used as positive control (PC). Normal DF-1 cells (without protein treatment) were used as the negative control (NC). For panels D and E, three independent experiments were performed, and data are shown as means ± SDs of triplicates from a representative experiment.

Further, the results of real-time RCR indicated that the copies of the RCAS(J)-EGFP recombinant virus in the DF-1 cells preincubated with the HA1 to HA4, HA15 to HA17, and HA32 to HA33 chimeric gp85 exhibited 500- to 10,000-fold decreases at 72 h postinfection compared with those of cells preincubated with the HA5 to HA14, HA18, and HA20 to HA31 chimeric gp85 (Fig. 3E), suggesting that HA5 to HA14 (aa 38 to 131) and HA18 and HA20 to HA31 (aa 159 to 283) replacement of gp85 did disrupt the ability of the virus to bind to chNHE1 and also lost the ability of blocking viral entry. These findings, together with the above mapping results, strongly suggested that the chNHE1-binding domain of ALV-J gp85 was located within the sequence between amino acids 38 to 131 and 159 to 283 of the gp85 protein.

N-linked glycosylation sites 6 and 11 were essential for receptor binding.

N-linked glycosylation has been shown to be important for folding, protein stability, immune evasion, and receptor usage (30, 31). There are 15 potential N-linked glycosylation sites (N1 to N15) on the ALV-J gp85 sequence in total. All these glycosylation sites were compromised by the HA substitution in the chimeric gp85 proteins. The N2 to N13 sites are known to be located in the chNHE1-binding domain and distributed in 11 chimeric gp85 proteins (HA7, HA9, HA11, HA13, HA14, HA18, HA20, HA23, HA24, HA25, and HA28) (Fig. 4A). To evaluate the roles of these glycosylation sites in the binding of gp85 to chNHE1, we constructed 11 glycosylation sites with revertant mutated gp85 expression plasmids, in which each glycosylation site was reverted to the wild-type sequence (NXT/S) at the corresponding position (Fig. 4A), using the 11 chimeric gp85 constructs with the compromised glycosylation sites as backbones. These revertant mutated gp85 proteins were successfully expressed and confirmed by SDS-PAGE (Fig. 4B). The protein-cell binding assay results indicated that most of the glycosylation site revertant mutants still exhibited under 20% binding capacity to the chNHE1 expressed in 293T cells except for the N6 and N11 revertant mutants, which showed more than 73.1% binding capacity of gp85 to chNHE1 (Fig. 4C). Using ELISA, the binding affinities of the glycosylation site revertant mutants to the soluble chECL1 revealed that the N6 and N11 revertant mutants could effectively bind to the soluble chECL1 (Fig. 4D), almost consistent with the above binding assay results. These results suggested that the N6 (Gly116) and N11 (Gly219) (positions at aa 116 and 219 of gp85, respectively) N-glycosylation sites of gp85 were important for its receptor binding.

FIG 4.

Analysis of the receptor binding capacity of the 12 N-glycosylation site revertant mutants of gp85. (A) Schematic diagram of the gp85 sequences and annotation positions of the individual N-linked glycosylation sites. The number in the red circle represents the number of glycosylation sites. The yellow shading represents the chNHE1-binding domain of gp85. The numbers at both ends represent the amino acid position. The N2 to N13 sites were revertant mutated based on the 11 chimeric gp85 plasmids (HA7, HA9, HA11, HA13, HA14, HA18, HA20, HA23, HA24, HA25, and HA28), respectively. (B) SDS-PAGE analysis of the 11 glycosylation site revertant mutated gp85 proteins. Plasmids containing 11 glycosylation site revertant mutated gp85 were transfected into 293T cells. At 48 h posttransfection, the harvested supernatant of cells was centrifuged and purified using protein A resin. Subsequently, the purification of the gp85 proteins was verified by SDS-PAGE analysis. (C) Binding capacity of the 11 glycosylation site revertant mutated gp85 proteins to chNHE1 expressed on the surface of transfected 293T cells. The binding capacity of the wild-type gp85 was set as 100%, with the values of revertant mutated gp85 proteins calculated as its proportion. The values of transfected 293T cells without protein incubation were set as the negative control (NC). (D) The binding affinity between soluble chECL1 and the 11 glycosylation site revertant mutated gp85 proteins was analyzed by solid-phase binding ELISA described in Materials and Methods. The OD value of the wild-type gp85 (gp85 wt) was set as 1, with the OD values of revertant mutated gp85 proteins calculated as its proportion. The values of protein-coated wells without protein incubation were set as the negative control (NC). In panels C and D, three independent experiments were performed, and data are shown as means ± SDs of triplicates from a representative experiment.

Cysteines 3 and 9 are essential for receptor binding.

Cysteines are often related to the formation of disulfide bonds that have much impact on protein structure (5, 32). The ALV-J gp85 unit has 14 cysteines in total, with 12 of them being found in the chNHE1-binding domain (Fig. 5A). Similarly, all these cysteines were compromised in the chimeric gp85 proteins with the HA tag substitution. To test if the disability for binding of the HA-substituted gp85 to chNHE1 was related to the compromise of cysteines, we constructed 10 revertant mutants, which contained the 12 cysteine sequences at the corresponding positions, respectively, based on the chimeric gp85 constructs with the compromised cysteines, and successfully expressed them in cells (Fig. 5B). The results of protein-cell binding assays showed that the third (C3) and ninth (C9) cysteine revertant mutants could increase the binding capacity of gp85 to chNHE1 to more than 62.4% (Fig. 5C). Solid-phase binding ELISA results indicated that C3 and C9 revertant mutants could significantly increase the binding affinity of gp85 to soluble chECL1 by 3 to 4 times compared with other revertant mutants (Fig. 5D). We inferred from these results that C3 (Cys60) and C9 (Cys197) (position at aa 60 and 197 of gp85, respectively) of gp85 were involved in receptor-gp85 binding.

FIG 5.

Analysis of the receptor-binding capacity of the 12 cysteine revertant mutants of gp85. (A) Schematic diagram of the gp85 sequences and annotation positions of the individual cysteine residues. The number in the red circle represents the number of cysteines. The yellow shading represents the chNHE1-binding domain of gp85. The numbers at both ends represent the amino acid position. The C2 to C7 and C9 to C14 sites were revertant mutated based on the 10 chimeric gp85 plasmids (HA5, HA7, HA8, HA12, HA13, HA22, HA24, HA25, HA29, and HA31), respectively. (B) SDS-PAGE analysis of the 10 cysteine revertant mutated gp85 proteins. Plasmids containing the 11 cysteine revertant mutated gp85 were transfected into 293T cells. At 48 h posttransfection, the harvested supernatant of cells was centrifuged and purified using protein A resin. Subsequently, the purification of the gp85 proteins was verified by SDS-PAGE analysis. (C) Binding capacity of the 10 cysteine revertant mutated gp85 proteins to chNHE1 expressed on the surface of transfected 293T cells. The binding capacity of the wild-type gp85 was set as 100%, with the values of revertant mutated gp85 proteins calculated as its proportion. The values of transfected 293T cells without incubating protein were set as negative control (NC). (D) The binding affinity between soluble chECL1 and the 10 cysteine revertant mutated gp85 proteins was analyzed by solid-phase binding ELISA described in Materials and Methods. The values of protein-coated wells without protein incubation were set as the negative control (NC). The OD value of the wild-type gp85 (gp85 wt) was set as 1, with the OD values of revertant mutated gp85 proteins calculated as its proportion. For panels C and D, data are means ± SDs from three independent experiments.

Two N-linked glycosylation sites and two cysteines are essential for viral entry.

To further test the function of these two N-linked glycosylation sites and that of the two cysteines in viral entry, we constructed RCAS(J)-luciferase recombinant viruses with wild-type gp85 and mutated glycosylation sites (N6 and N11) and cysteines (C3 and C9), as well as their revertant mutants of the chimeric gp85 (Fig. 6A), to compare their efficiency for rescue and ability for viral entry. The results of IFA (Fig. 6B) showed that four recombinant viruses (rN6, rN11, rC3, and rC9) with revertant mutants of the glycosylation sites (N6 and N11) and cysteines (C3 and C9) were successfully rescued. The virus titers of the rN6, rN11, rC3, and rC9 recombinant viruses were demonstrated to range from 10^4.0 to 10^4.8 TCID₅₀/ml, similar to the titers of the wild-type RCAS(J)-luciferase recombinant virus (rALV-J/Luc) (Fig. 6C). However, the four recombinant viruses with mutated glycosylation sites and cysteines (ΔN6, ΔN11, ΔC3, and ΔC9) exhibited almost no fluorescence, and their viral titers could not be detected (Fig. 6C). These results clearly indicated that the RCAS(J)-luciferase recombinant virus constructs with glycosylation site and cysteine revertant mutations were successfully rescued, whereas these recombinant viruses with compromised glycosylation sites or cysteines could not be.

FIG 6.

Analysis of the influence of N-glycosylation sites and cysteines of gp85 on viral entry. (A) Schematic diagrams showing the construction of RCAS(J)-luciferase recombinant viruses with wild-type gp85, mutated glycosylation sites (N6 and N11) and cysteines (C3 and C9), and their revertant mutants. (B) Immunofluorescent detection of RCAS(J)-luciferase recombinant viruses. DF-1 cells were infected with different RCAS(J)-luciferase recombinant viruses for 72 h and then immunostained with an 4A3 monoclonal antibody and analyzed using a fluorescence microscope. Cell nuclei were counterstained with DAPI (blue). Normal DF-1 cells were used as the negative control. (C) Viral titers of RCAS(J)-luciferase recombinant viruses. At 7 days after infection with recombinant viruses, the infectious viral titers in DF-1 cells were determined by TCID₅₀ analysis using 96-well plates. (D) Luciferase detection of the entry of the RCAS(J)-luciferase recombinant virus into DF-1 cells. Levels of viral entry were determined using a luciferase reporter assay, as described in Materials and Methods. The entry level of rALV-J/Luc into DF-1 cells was set to 1, with the values of the revertant mutated recombinant viruses calculated as its proportion. In panels C and D, data are means ± SD from three independent experiments.

Next, we tested the ability of the revertant mutated recombinant viruses (rN6, rN11, rC3, and rC9) for entry into host cells. These recombinant viruses were inoculated into DF-1 cells, and their efficiency for viral entry was tested 72 h after inoculation. As shown in Fig. 6D, the rN6, rN11, rC3, and rC9 recombinant viruses could effectively enter into DF-1 cells, with their luciferase values being similar to those of rALV-J/Luc. Taken together, these data further demonstrated that the glycosylation sites (N6 and N11) and cysteines (C3 and C9) of ALV-J gp85 were important for both receptor binding and viral entry.

DISCUSSION

The entry of retroviruses into cells is known to be mediated by specific interactions between glycoproteins on the viral surface and receptors on cell membranes, which trigger a series of conformational changes in the envelope that lead to membrane fusion (30). On that note, ALVs are useful for studying early events in retroviral infection because they use distinct cellular receptors to gain entry into cells. ALV-J has been reported to be derived by recombination of an exogenous ALV with ancient endogenous avian viral sequences in the chicken genome and thus has a broader host range than that of the A to E subgroups (26, 33, 34). Several studies concentrating on the binding of the interactions between cell receptors and envelope glycoproteins of ALV-A, ALV-B, ALV-C, and ALV-E have suggested that the main binding occurs in the hr1 and hr2 regions of gp85 (22–25). However, the receptor-binding determinants of the gp85 unit of ALV-J are still unclear. In this study, our results, summarized schematically in Fig. 7, showed that the 38 to 131 amino acids of the N terminus and 159 to 283 aa of the C terminus of ALV-J gp85 control the efficiency of gp85 binding to chNHE1, and thus, these sequences are critical for viral infection of receptor-expressing cells. More importantly, our results demonstrated that the N-glycosylation sites (N6 and N11) and cysteines (C3 and C9) of gp85 were important for its receptor binding and ALV-J entry into host cells.

FIG 7.

Schematic diagram of the localization of chNHE1-binding domain of ALV-J gp85 and N-glycosylation sites and cysteines of different ALV gp85 (reference numbers are provided in parentheses). Alignment of gp85 sequences of ALV subtypes A to E and J. The numbers on the left represent the amino acid position. The letters on the right represent different ALV subgroups. A, ALV-A (Schmidt-Ruppin A; GenBank accession no. S83510) (11). B, ALV-B (Schmidt-Ruppin B; GenBank accession no. AF052428) (11). C, ALV-C (Prague C; GenBank accession no. J02342) (18). D, ALV-D (Schmidt-Ruppin D; GenBank accession no. D10652) (17). E, ALV-E (RAV-0; GenBank accession no. M12172) (16). J, ALV-J (HRPS-103; GenBank accession no. Z46390) (20). The yellow shading represents the chNHE1-binding domain of ALV-gp85. The variable regions identified in the A to E ALV subgroups gp85 genes (vr1, vr2, vr3, hr1, and hr2) (11, 21) are shown above the sequences. The red stars represent the N-linked glycosylation sites of ALV-A to E. The blue stars represent the N-linked glycosylation sites of ALV-J. The red circles represent the cysteines of ALV-A to E. The blue circles represent the cysteines of ALV-J.

The gp85 sequences of the A to E ALV subgroups have been shown to be highly conserved. Studies on the host range and receptor binding of these ALV subgroups have suggested that the principal receptor-binding regions of gp85 were located in the hr1 variable region, the last third of the N terminus, and hr2, and the central region of the C terminus of gp85 (23). In particular, the hr2 region has been reported to be closely related to receptor binding (23). For example, two basic residues (R213 and K227) in the hr2 region of ALV-A gp85 were demonstrated as important determinants of receptor binding, while four residues (R210, R213, R223, and R224) were critical for efficient ALV-A infection (23). However, ALV-J is known to be a multiple recombinant of exogenous ALVs and 1 EAV (endogenous avian retroviral) sequence from the chicken genome. The gp85 of ALV-J has been shown to be closely related to that of the defective EAV-E51 but divergent from those of other ALV subgroups (A to E) (26). Unlike the receptor-binding regions of SU of other subgroups, our mapping results clearly indicated for the first time that the chNHE1-binding domain of ALV-J g85 was present in the central region (aa 38 to 131) of the N terminus and the central region (aa 159 to 283) of the C terminus of gp85 (Fig. 7). The vr1, vr2, and the first half of hr1 of the gp85 gene of ALV-A to ALV-E were shown to be included within the binding domain of the N terminus of ALV-J, whereas the hr2 and vr3 of ALV-A to ALV-E were included within the binding domain of the C terminus of ALV-J; all these regions are important in the binding to chNHE1 and thus important for viral infection. In contrast, vr1 and vr2 of ALV-A to ALV-E did not seem to be essential for receptor specificity, while the vr3 region might recognize the receptor but did not appear to be involved in the binding to the receptor (23). Therefore, five variation regions (vr1, vr2, vr3, hr1, and hr2) identified in other subgroups of the gp85 gene (ALV-A to ALV-E) were shown to not be suitable for ALV-J. The structural organizations of the gp85 receptor-binding region were different between ALV-J and other subgroups.

A previous study showed that the R213 and K227 basic residues in the hr2 region of ALV-A gp85 were important determinants of receptor binding (35). In this study, we were not able to identify a single residue as being essential for receptor binding. Our results were similar to that of HIV, amphotropic MLV, and equine infectious anemia virus (EIAV) receptor-binding domains, which are known to be composed of a complex of discontinuous segments of the SU protein (36–40). These findings from different retroviruses suggested that multiple residues of the SU protein might be involved in the contact with their receptor. Previous studies have shown that the binding interfaces between proteins depend on multiple intermolecular contacts involving 10 to 30 side chains per protein (41, 42). Our previous studies also indicated that at least four residues (A30, V33, W38, and E39) of chNHE1 were involved in the binding to gp85 (26), further suggesting that the binding of gp85 to chNHE1 was likely to be the result of the participation of multiple amino acids.

The SU units of retroviruses that bind to membrane proteins or receptors are known to be the important physiological factors that trigger infection. To date, the receptor-SU unit binding of gammaretroviruses (MLV, FIV), deltaretroviruses (human T-cell leukemia virus [HTLV]), and lentiviruses (HIV, EIAV, SIV, etc.) has been extensively analyzed and well characterized. These studies have shown that the receptor properties and binding domains of SU are different among different retroviruses. For gammaretrovirus, receptor recognition by envelope proteins has been attributed to the N-terminal half of SU. For example, all the receptor-binding domains of MLV and FLV SU have been shown to be located in two variable regions (VRA and VRB) in the N-terminal third of SU (43, 44). The receptor domain of deltaretroviruses was reported to be somewhat similar to that of the gammaretrovirus. The receptor-binding domain of the human T-cell leukemia virus is known to lie in the first 160 residues in the N terminus of SU, immediately upstream of a central proline-rich region (Env residues 180 to 205) (45), whereas, regarding lentiviruses, their receptor-binding domains have been shown to be mainly located in the C-terminal region of SU units. The CD4-binding sites of gp120 are known to be located in a depression formed at the interface of the outer and inner domain and the bridging sheet that connects these two domains, spanning a discontinuous segment in the C-terminal half region of gp120 (37, 38). The ELR1-binding domain of the equine infectious anemia virus has also been found to be present in a complex of discontinuous segments located in the C-terminal two-thirds of gp90 (39). In contrast, in the case of avian alpharetroviruses, there have been relatively few studies on the receptor-binding properties of gp85. Our results indicated that the chNHE1-binding domain of ALV-J lies in the middle of the N terminus and the middle of the C terminus of gp85. The receptor-binding regions of other subgroups have not been accurately identified. However, it is generally believed that the hr1 and hr2 regions of gp85 determine host range and receptor specificity (22, 23). Although the size of the receptor-binding region of ALV-J was found to be longer than that of hr1 and hr2 regions of the gp85 of other subgroups, their position was shown to be similar, with all of them located at the N and C termini of gp85, unlike other retroviruses in which their receptor-binding domains are found either at the N terminus or the C terminus of the SU units. This might be attributed to the receptor-binding properties of the SU units of avian alpharetroviruses.

It remains unclear whether the N and C termini of gp85 bind to chNHE1 simultaneously or sequentially in a 2-step manner. Our previous studies indicated that the gp85-binding domain of chNHE1 constituted of 12 amino acids in the membrane-proximal region of the N terminus of chNHE1, which was totally different from the membrane-distal binding regions in the receptors of SU units of other ALV subgroups and HIV (46–52). The geometric location of the gp85-binding sites would apparently not allow the simultaneous binding of the N and C termini of ALV-J gp85 to chNHE1 (27). We speculated that these two separate segments of gp85 might together form a single receptor-binding site. To address this hypothesis, the structure of the gp85 protein needs to be resolved by crystal or cryo-electron microscopy in the future.

The gp85 of ALV is a heavily glycosylated protein, with almost 50% of its molecular mass represented by carbohydrates. Analysis of the glycosylation sites in the gp85 indicated that the glycosylation patterns among ALV-A to ALV-E exhibit a high degree of conservation and are important for maintaining the conformation and structure of gp85 (53). Previous studies have shown that the glycosylation site 10 was essential for the receptor binding and ALV-A entry into target cells (54). The gp85 of ALV-J contains 15 glycosylation sites, with a glycosylation pattern different from that of the ALV-A to ALV-E subgroups (Fig. 7). For instance, the glycosylation site 10 (NAS close to the vr3 region) in the gp85 of ALV-A, was not found in the gp85 of ALV-J (Fig. 7). Our results indicated that the N6 (Gly116) and N11 (Gly219) glycosylation sites of ALV-J gp85, which are not found in the gp85 of other subgroups, were very important for its receptor binding and viral entry. Once again, these results showed that the ALV-J gp85 binds cellular receptors in a way or mechanism different from that of other subgroups (27).

In addition, cysteines are important components in the structure or function of envelope proteins (55, 62). However, unlike other ALV subgroups, HIV, and SIV (5, 32), the intrachain disulfide bond pattern of ALV-J gp85 has not yet been identified. Cysteines (such as C1, C2, C13, and C14) in the N and C termini are relatively conserved between ALV-J gp85 and other subgroups. Smith and Cunningham found that although C2 (Cys38) of gp85 was essential for infection by ALV-A, Cys38 was not required for SU binding to Tva, suggesting that the critical Cys38-dependent step(s) in the infection by ALV-A occur after receptor binding (5). Consistently, our results also indicated that the C2 of gp85 was not necessary for its receptor binding. In addition, the C1 (Cys25) of ALV-A gp85 has been demonstrated to bind to the last extracellular C19 cysteine (Cys438) in the TM of ALV-A (32). C1 is conserved among all ALV subgroups (Fig. 7), suggesting that C1 might play an important role in the formation of the gp85-TM dimer (32). Our results demonstrated that the C3 (Cys60) and C9 (Cys197) of ALV-J gp85 were able to restore its receptor-binding ability after revertant mutation, whereas the other 10 cysteines could not, indicating that C3 and C9 might be the essential components of the chNHE1-binding domain of ALV-J gp85.

In conclusion, our results suggested that the chNHE1-binding domain of ALV-J gp85 was formed by the bipartite sequence motif spanning aa 38 to 131 and aa 159 to 283, with the N-linked glycosylation sites (N6 and N11) and cysteines (C3 and C9) playing a crucial role in receptor binding and viral entry. These findings cannot only help us better understand the molecular basis of the interaction between ALV-J gp85 and its receptors but also help elucidate the molecular mechanism of the infection of cells by ALV-J.

MATERIALS AND METHODS

Cells and viruses.

293T cells and DF-1 cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM; Thermo Scientific, Rockford, IL, USA) supplemented with 10% heat-inactivated fetal bovine serum (FBS; Sigma-Aldrich, St. Louis, MO, USA) and 1% penicillin/streptomycin (Summus, Beijing, China) at 37°C under 5% CO₂. The RCAS(J)-luciferase recombinant virus was maintained at the Harbin Veterinary Research Institute, Chinese Academy of Agricultural Sciences (CAAS) (Harbin, China) (27).

Construction of various plasmids.

For mapping of receptor-binding domains, the gp85 gene was amplified from the CAG-sgp85-Fc expression vector (27), and the series of chimeric gp85 genes with HA tag (9-aa linear epitope peptide) substitutions (Fig. 1A) was constructed through an overlapping PCR with two hybrid primers designed to contain the HA tag sequences at the 5′ ends and ALV-J gp85-specific sequences at the 3′ ends of both primers. The overlapping products were then inserted into the CAG vector (Addgene, Cambridge, MA, USA) to generate 33 chimeric CAG-gp85-Fc plasmids named HA1 to HA33, respectively.

To construct the 11 N-linked glycosylation site revertant mutated gp85 expression plasmids (N2, N3, N4, N5, N6, N7, N8, N9, N10, N11, N12, and N13) (Fig. 4A), the sequences of the glycosylation sites at the corresponding positions of HA7, HA9, HA11, HA13 HA14, HA18, HA20, HA23, HA24, HA25, and HA28 expression plasmids were reverted to wild-type sequences (NXT/S) by site-directed mutagenic PCR, receptively.

To construct the 12 cysteine (C2, C3, C4, C5/C6, C7, C9, C10/C11, C12, C13, and C14) revertant mutants in the 10 chimeric gp85 expression plasmids (Fig. 5A), the cysteine sequences at the corresponding positions of HA5, HA7, HA8, HA12, HA13, HA22, HA24, HA25, HA29, and HA31 expression plasmids were reverted to wild type by site-directed mutagenic PCR, receptively.

All constructs were verified by DNA sequencing. The primer sequences of all oligonucleotides used in this study are available upon request.

Purification and quantification of gp85 mutants and soluble chECL1.

293T cells cultured in 10-cm-diameter culture dishes were transfected with the CAG-sgp85-Fc, all-chimeric gp85, or soluble chECL1 plasmids (27) using the X-tremeGene HP DNA transfection reagent (Roche, Indianapolis, IN, USA) according to the manufacturer’s instructions. Then, 48 h posttransfection, the culture supernatant was harvested and centrifuged at 8,000 × g at 4°C for 5 min to remove cell debris. After centrifugation, the supernatant was purified with protein A resin (GenScript, Jiangsu, China) that could specifically bind to Fc according to the manufacturer’s protocol. Different proteins were purified separately to avoid cross-contamination. To assess the expression level and molecular weight of expressed proteins, purified proteins were separated on 12.5% SDS-PAGE gels and analyzed by Coomassie blue staining. Gels were scanned using an Odyssey infrared imaging system (Li-Cor Biosciences). Purified proteins were quantitated using the Pierce BCA protein assay kit (Thermo Scientific, Rockford, IL, USA) according to the manufacturer’s protocol.

Western blotting.

The wild-type or chimeric purified gp85 proteins were separated on 12.5% SDS-PAGE gels and transferred onto nitrocellulose membranes (Hybond-C Super; GE Healthcare, Piscataway, NJ, USA). Membranes were blocked with 5% (wt/vol) skim milk at 4°C overnight. Then, membranes were incubated with a 1:3,000 dilution of an anti-HA monoclonal antibody produced in mouse (catalog no. H3663; Sigma, USA) at 25°C for 1 h and washed three times with PBS containing 0.05% Tween 20 (PBST, pH 7.2). Next, membranes were incubated with a 1:20,000 dilution of IRDye 680RD donkey anti-mouse IgG (H+L) antibodies (Li-Cor Biosciences) at 25°C for 1 h followed by washing thrice with PBST. Finally, membrane blots were scanned using an Odyssey infrared imaging system (Li-Cor Biosciences).

CD spectroscopy.

The CD spectroscopy was used to measure the secondary structures as previously described (56, 57). Briefly, wild-type and chimeric gp85 proteins were diluted to 0.6 mg/ml. CD spectra were then obtained on a Chirascan (Applied Photophysics Limited) using a 1-nm bandwidth with a 1-nm step resolution from 190 to 260 nm at 25°C with a scanning speed of 10 nm/min, respectively. Three scans were obtained for each sample. The α-helix, β-strand, turns, and other structures were calculated using the online BeStSel server.

Protein-cell binding assay.

The protein-cell binding assay was performed as previously described (27). Briefly, 293T cells transfected with pCAF-chNHE1 for 24 h were detached from dishes using a solution of PBS and 5 mM EDTA, pelleted (1,000 × g) for 10 min, and washed thrice with ice-cold PBS containing 5% (wt/vol) FBS. In parallel, 500 μl wild-type and chimeric gp85 proteins were incubated with cells for 1 h on ice, followed by staining with a 1:200 dilution of an anti-human IgG (Fc-specific)-fluorescein isothiocyanate (FITC) antibody produced in goat (catalog no. F9512; Sigma). Cells were then fixed in 4% (vol/vol) paraformaldehyde (Solarbio, Beijing, China) at 25°C for 15 min and incubated with a 1:200 dilution of monoclonal anti-FLAG M2 antibodies produced in mouse (catalog no. F1804; Sigma) for 1 h on ice. Then, cells were stained with a 1:200 dilution of anti-mouse IgG (whole-molecule)-tetramethyl rhodamine isocyanate (TRITC) antibodies (catalog no. T2402; Sigma). Samples were resuspended in 0.5 ml fluorescence-activated cell sorter (FACS) wash buffer and analyzed using a Becton, Dickinson FACSCalibur (BD Biosciences, San Jose, CA, USA). The stained cells were analyzed by FACS, using a FACS AriaII flow cytometer (BD Biosciences, Franklin Lakes, NJ). The binding level of the wild-type protein was set as 1, with the values of the binding of chimeric gp85 proteins calculated as its proportion. The values of transfected 293T cells without protein incubation were set as negative control.

Solid-phase binding ELISA.

First, 96-well polystyrene microtiter plates (Costar, New York, NY, USA) were coated with 2 ng/ml purified soluble chECL1 protein diluted in bicarbonate buffer at 4°C overnight, and then wells were blocked with 100 μl 5% (wt/vol) skim milk in PBS at 37°C for 4 h and washed five times with 300 μl PBST. Second, 100 ng/μl wild-type and chimeric gp85 proteins were, respectively, added to the protein-coated wells and inoculated at 37°C for 1 h. After, plates were washed five times using PBST, and a 1:3,000 dilution of the 4A3 (29) was added to the plates and incubated at 37°C for 1 h. Plates were washed again, and a 1:3,000 dilution of horseradish peroxidase (HRP)-labeled goat anti-mouse immunoglobulin G (IgG; Sigma, MO, USA) was added to the plates at 37°C for 1 h followed by washing five times with PBST. The reaction was detected using 100 μl 3,3,5,5,-tetramethylbenzidine substrate at 25°C in the dark for 15 min. Finally, the reaction was stopped with 100 μl of 2 M sulfuric acid solution (H₂SO₄). The optical density (OD) was determined at 450 nm using a model 680 microplate reader (Bio-Rad, Hercules, CA, USA). The OD value of the wild-type protein was set as 1, with the values of the chimeric gp85 proteins calculated as its proportion. The values of protein-coated wells without protein incubation were set as the negative control.

RCAS(J)-EGFP recombinant virus.

To rescue RCAS(J)-EGFP recombinant virus, DF-1 cells were transfected with RCAS(J)-EGFP plasmids (58) using the X-tremeGene HP DNA transfection reagent. Seven days posttransfection, DF-1 cells were harvested by centrifugation at 2,000 × g at 4°C for 10 min and stored at −80°C. The titer of the rescued recombinant viruses was determined on the basis of the TCID₅₀ value using the Reed and Muench method (59).

Indirect immunofluorescent assay (IFA).

To detect the rescue of recombinant viruses, the DF-1 cells cultured in 12-well plates were infected with 200 TCID₅₀ RCAS(J)-EGFP, recombinant mutant, or recombinant recovery mutant RCAS(J)-luciferase recombinant viruses at 37°C and 5% CO₂ for 2 h. After incubation, the cells were washed three times with room-temperature PBS and maintained in DMEM containing 5% (wt/vol) FBS at 37°C and 5% CO₂ for 72 h. Then, the cells were washed three times with cold PBS and fixed for 15 min with cold solute ethanol. After blocking in 5% (wt/vol) skim milk at 37°C for 1 h, the cells were washed three times with cold PBS and stained with 500 μl 1:200 dilution of the 4A3 (29) at 37°C for 1 h. After three washes with PBS, the DF-1 cells infected with RCAS(J)-EGFP recombinant viruses were inoculated with 500 μl of 1:200 dilution of the secondary antibody, anti-mouse IgG-TRITC (Sigma, USA), and the DF-1 cells infected with recombinant mutant or recombinant recovery mutant RCAS(J)-luciferase recombinant viruses were inoculated with 500 μl of 1:200 dilution of the anti-mouse IgG-FITC secondary antibody (Sigma). The DF-1 cells were incubated at 37°C for 1 h. Cells were further washed thrice with PBST and the nuclei stained with 4′,6-diamidino-2-phenylindole (DAPI) for 10 min at room temperature. After three washes with PBST, the DF-1 cells were then visualized under a fluorescence microscope (TU-80; Nikon, Tokyo, Japan). Normal DF-1 cells were used as negative controls.

Blocking assay of viral entry.

To build a sensitive method to evaluate the blocking effect, we used the RCAS(J)-EGFP recombinant virus in the blocking assay. First, 100 μg purified wild-type and chimeric gp85 proteins were incubated with DF-1 cells at 4°C for 1 h. DF-1 cells are known to be susceptible to exogenous ALVs (60). Second, cells inoculated with gp85 proteins were infected with 5,000 TCID₅₀ of the RCAS(J)-EGFP recombinant virus and incubated at 37°C for 1 h and then washed three times with PBS. Third, cells were maintained in DMEM containing 5% (wt/vol) FBS at 37°C and 5% CO₂ for 48 h and then were harvested. Cells were observed under fluorescence microscopy (TU-80; Nikon, Tokyo, Japan) and then detached with a solution of PBS and 5 mM EDTA and analyzed using a Becton, Dickinson FACSCalibur (BD Biosciences). At the same time, RCAS(J)-EGFP recombinant viral-infected cells were harvested through three freeze-thaw cycles for the detection of viral copies by real-time PCR, as described below. Normal DF-1 cells (without protein treatment) infected with the virus were used as positive control. Normal DF-1 cells (without protein treatment) were used as negative control. The positive rate of cells of the positive control was set as 1, with the values of chimeric gp85 proteins calculated as its proportion.

Real-time PCR.

Proviral DNA was isolated from RCAS(J)-EGFP recombinant viral-infected cells using the AxyPrepBody fluid viral DNA/RNA miniprep kit (Corning Life Sciences Co., Ltd., Suzhou, China) following the manufacturer's instructions. The real-time PCR was performed as previously described (61). The reaction conditions were as follows: predenaturation at 95°C for 5 min, followed by 40 cycles at 95°C for 10 s for denaturalization and 65°C for 40 s for elongation. Fluorescent signals were detected during the elongation step. Real-time PCRs were performed using a LightCycler 480 real-time thermocycler (Roche, Basel, Switzerland). Virally infected normal DF-1 cells (without protein treatment) were used as positive control. Normal DF-1 cells (without protein treatment) were used as negative control. The copy number of the positive control was set as 1, with the values of chimeric gp85 proteins calculated as its proportion.

Rescue of recombinant viruses.

To construct RCAS(J)-luciferase recombinant viruses with mutated N-linked glycosylation sites (N6 and N11) and their revertant mutants, the gp85 gene in the RCAS(J)-luciferase plasmid (27) was replaced with the gp85 gene of the HA13, HA24, N6, and N11 constructs (Fig. 6A), respectively. Similarly, to construct RCAS(J)-luciferase recombinant viruses with mutations of cysteines (C3 and C9) and their revertant mutants, the gp85 gene in the RCAS(J)-luciferase plasmid was replaced with the gp85 gene of the HA7, HA22, N6, and N11 constructs, respectively (Fig. 6A). To rescue these recombinant viruses, DF-1 cells were transfected with different RCAS(J)-luciferase plasmids using the X-tremeGene HP DNA transfection reagent. Seven days posttransfection, DF-1 cells were harvested by centrifugation at 2,000 × g at 4°C for 10 min and stored at −80°C. The RCAS(J)-luciferase recombinant viruses with mutated N-linked glycosylation sites (N6 and N11) and cysteines (C3 and C9) were named ΔN6, ΔN11, ΔC3, and ΔC9, while their revertant mutated recombinant viruses were named rN6, rN11, rC3, and rC9, respectively. The titer of the rescued recombinant viruses was determined on the basis of the TCID₅₀ value using the Reed and Muench method.

Viral entry assay.

DF-1 cells cultured in 12-well plates were infected with 5,000 TCID₅₀ RCAS(J)-luciferase and recombinant recovery mutant viruses (rN6, rN11, rC3, and rC9) at 37°C and 5% CO₂ for 2 h. After incubation, cells were washed three times at 25°C with PBS and maintained in DMEM containing 5% (wt/vol) FBS at 37°C and 5% CO₂ for 72 h. Levels of viral entry were determined by detecting the luciferase activity in cell lysates using the Steady-Glo luciferase assay system (Promega) following the manufacturer’s instructions. The entry level of the RCAS(J)-luciferase recombinant virus into DF-1 cells was set as 1, with the values of recombinant viruses calculated as its proportion.