Production and Characterization of a Soluble, Active Form of Tva, the Subgroup A Avian Sarcoma and Leukosis Virus Receptor

Abstract

The receptor for the subgroup A avian sarcoma and leukosis viruses [ASLV(A)] is the cellular glycoprotein Tva. A soluble form of Tva, sTva, was produced and purified with a baculovirus expression system. Using this system, 7 to 10 mg of purified sTva per liter of cultured Sf9 cells was obtained. Characterization of the carbohydrate modification of sTva revealed that the three N glycosylation sites in sTva were differentially utilized; however, the O glycosylation common to Tva produced in mammalian and avian cells was not observed. Purified sTva demonstrates significant biological activity, specifically blocking infection of avian cells by ASLV(A) with a 90% inhibitory concentration of ∼25 pM. A quantitative enzyme-linked immunosorbent assay, developed to assess the binding of sTva to ASLV envelope glycoprotein, demonstrates that sTva has a high affinity for EnvA, with an apparent dissociation constant of approximately 0.3 nM. Once they are bound, a very stable complex is formed between EnvA and sTva, with an estimated complex half-life of 6 h. The soluble receptor protein described here represents a valuable tool for analysis of the receptor-envelope glycoprotein interaction and for structural analysis of Tva.

The entry of enveloped viruses is determined by specific interactions between receptors on the cell surface and envelope glycoproteins on the virion. For most retroviruses, this process is pH independent, suggesting direct fusion between the viral and cell surface membranes (43). It has been hypothesized that the cellular receptor not only provides a binding site for the virus but also induces conformational changes in the viral envelope proteins, ultimately leading to membrane fusion and virus entry. The mechanism by which this occurs, however, is poorly understood.

The avian sarcoma and leukosis virus (ASLV) family is an excellent model system for analyzing the mechanism of pH-independent virus entry. There are five major subgroups (A to E) of ASLV based on host range, infection interference, and immunological reactivity of the viral envelope glycoproteins. The five ASLV subgroups utilize cellular receptors encoded by three dominant autosomal loci, TVA, TVB, and TVC (16–18, 50), that allow for comparison of common, required features for virus entry. TVA and TVC confer susceptibility to ASLV(A) and ASLV(C), respectively. The TVB locus is more complex, with various alleles determining infection by ASLV(B), ASLV(D), and ASLV(E). Several alleles of TVB have recently been cloned (1, 10).

Tva was the first ASLV receptor identified (7, 72) and has been mapped to the TVA locus (6). While it has been shown to function as a viral receptor, the cellular function of Tva remains unknown. Due to alternative splicing, two isoforms of Tva (Tva₉₅₀ and Tva₈₀₀) that differ in their membrane anchoring domains are expressed in avian cells. Both isoforms are functional viral receptors (7), indicating that the determinants for receptor function are contained in the conserved, 83-amino-acid extracellular domain. Within this domain is a 40-amino-acid, cysteine-rich motif homologous to repeats (LDL-A modules) first identified in the human low-density lipoprotein receptor (hLDLR). The LDL-A module in Tva, when appended to a heterologous membrane anchor, is sufficient to mediate ASLV(A) entry (51). Mutations within the LDL-A module severely affect subgroup A envelope (EnvA) binding and virus entry, which suggests that the envelope binding site is composed of discontinuous regions within the LDL-A module (8, 54, 74, 75). Recent work demonstrates that three amino acid substitutions in the hLDLR module A4 convert this module into an ASLV(A) receptor (53). In addition, since Tva contains a single LDL-A module, EnvA-Tva interactions may serve as a model system for studying LDLR-ligand binding in addition to dissecting determinants of virus entry (52).

Soluble receptors have been an effective tool for dissecting early entry events for several viruses. Preincubation of soluble CD4 (sCD4) with human immunodeficiency virus type 1 (HIV-1) envelope results in dissociation of gp120 from gp41 (47), acquisition of protease sensitivity by HIV-1 Env (56), and alterations in monoclonal antibody binding to HIV-1 Env (31, 41, 56) and promotes the ability of Env to interact with coreceptors (4, 33, 65, 71). A soluble form of the poliovirus receptor irreversibly alters poliovirus to an intermediate similar to that identified during poliovirus infection (38). sTva produced with stably transfected quail cells was shown to inhibit ASLV(A) infection (15). However, the amount of receptor produced was not adequate for rigorous biochemical analysis of the Tva-EnvA interaction or for structural determination of Tva by crystallography. We therefore chose to use a baculovirus expression system (BES) to produce sTva. The BES produces high levels of secreted protein that retains biological activity as well as many of the posttranslational modifications.

In this report, we describe the production and purification of large quantities of sTva with a BES. The purified sTva is biologically active; it specifically and potently blocks ASLV(A) infection, with a 90% inhibitory concentration (IC₉₀) of 26 pM. Additionally, we quantitatively evaluated sTva-EnvA interactions by enzyme-linked immunosorbent assay (ELISA). sTva has a high affinity for EnvA, with an apparent dissociation constant (K_d) of 0.3 nM, and once bound, forms a very stable complex with an estimated half-life of 6 h.

MATERIALS AND METHODS

Cells, viruses, and reagents.

Sf9 cells were routinely maintained in SF900II medium (Gibco BRL, Gaithersburg, Md.). Turkey embryo fibroblasts (TEF) were maintained in TEF medium (1× M199, 10% tryptose phosphate broth, 5% fetal calf serum, 1% chick serum, 2 mM l-glutamine, penicillin [100 U/ml], and streptomycin [100 μg/ml]). Stocks of the recombinant ASLV(A) virus RCAS(A)AP (34), which expresses the histochemical marker alkaline phosphatase, were generated as described previously (72). The Baculogold baculovirus transfection kit was from Pharmingen (San Diego, Calif.).

Plasmid pSP73 was from Promega (Madison, Wis.). pVL1393 was from Pharmingen. pCB6 (7) and pVTBac (62) were described previously.

Wheat germ agglutinin (WGA) agarose and concanavalin A (ConA) agarose were from Vector Laboratories, Inc. (Burlingame, Calif.). Peptide N-glycosidase F (PNGase F) was from New England Biolabs, Inc. (Beverly, Mass.). Neuraminidase, O-glycosidase, and tunicamycin were from Boehringer Mannheim Biochemicals (Indianapolis, Ind.). ImmunoPure NHS-LC-biotin and 2,2′-azinobis(3-ethylbenzothiazoline)-6 sulfonic acid diammonium salt were from Pierce (Rockford, Ill.). Rabbit polyclonal Tva antiserum 40 was previously described (7, 72). Acrylamide was from National Diagnostics (Atlanta, Ga.). Ni²⁺-nitriloacetic acid (NTA) agarose was from Qiagen (Chatsworth, Calif.).

Construction of transfer vectors.

The 968-bp fragment of pg950 (7) containing the complete open reading frame of quail Tva₉₅₀ was cloned into pSP73, which was digested with Asp718 and EcoRI to make pSP73-0.95. A six-histidine tail was appended to the carboxy terminus of the Tva ectodomain by digesting pSP73-0.95 with EagI and inserting a linker (OS9, 5′-GGCCACCATCATCACCATCACTAGGAATTC-3′, and OS10, 5′-GCCTGAATTCCTAGTGATGGATGATGATGGT-3′), encoding six histidine residues and containing a stop codon and an EcoRI site, to make pSP73-0.95H₆RI. This insertion deleted the codon for the ultimate carboxy-terminal residue of the ectodomain, Arg83.

To generate a eukaryotic expression vector that expresses sTva, with a six-histidine carboxy terminus, pSP73-0.95H₆RI was digested with KpnI and ClaI (found in the polylinker of pSP73) and the resulting 830-bp fragment was cloned into pCB6 to make pCB6WT-Tva_HIS6.

Two baculovirus transfer vectors were used to generate recombinant baculovirus. The first construct preserves the Tva signal sequence. pSP73-0.95H₆RI was digested with BamHI (found in the polylinker of pSP73) and EcoRI to yield a 790-bp fragment that was cloned into the baculovirus transfer vector pVL1393 to generate pVL1393Tva. The second baculovirus transfer vector contains the melittin leader sequence. pSP73-0.95H₆RI was digested with AgeI and EcoRI, and the 720-bp fragment was cloned into pVTBac digested with BamHI and EcoRI to make pVTTva by using a short oligonucleotide adaptor (OS13, 5′-GATCCGC-3′, and OS14, 5′-CCGGGCG-3′) to connect the vector BamHI terminus to the insert AgeI terminus. Expression of sTva from this construct was predicted to yield a protein with four amino acids (AspProProGly) appended to the amino terminus by the linker (Fig. 1B).

FIG. 1.

Comparison of sTva and naturally occurring forms of quail Tva. (A) Diagram of sTva and quail Tva₉₅₀ and Tva₈₀₀. The arrow indicates the site of Tva leader peptide cleavage; the sTva melittin leader peptide is cross hatched, the LDL-A motif (residues 11 to 50) is hatched, transmembrane domain is solid, the glycosylphosphotidylinositiol linkage signal is vertically hatched, the cytoplasmic tail is shaded, and the six-histidine tail is stippled. The numbers on the right indicate the sizes (in amino acids) of the mature proteins. (B) Comparison of WT Tva and sTva amino and carboxy termini. The arrows indicate the signal peptide cleavage sites. The four additional N-terminal amino acids in sTva are marked with a line above. Potential N-linked glycosylation sites are underlined. MSD, start of the membrane-spanning domain.

Generation of recombinant baculovirus.

Generation of recombinant baculovirus (Autographa californica nuclear polyhedrosis virus) (Baculogold; Pharmingen) was performed as described by the manufacturer. The recombinant baculovirus produced from pVL1393Tva is designated bVL1393Tva. Likewise, pVTTva produced the recombinant virus bVTTva.

Infections for protein expression.

A suspension culture of 1.2 × 10⁷ Sf9 cells/ml in an Erlenmeyer flask was infected with the recombinant baculovirus at a multiplicity of infection of 2 and incubated in a rotary shaker at 90 rpm and 28°C for 1 h. The cells were diluted 1:2, and infection was allowed to proceed for between 48 and 96 h postinfection (p.i.), depending on the experiment. To harvest the recombinant protein, the medium was clarified by two-step centrifugation at 430 × g and then at 2,300 ×g. The clarified medium was used directly for biochemical analysis of sTva or for protein purification. The cell culture viability was determined by trypan blue exclusion. The percentage of viable cells was equal to the number of unstained cells divided by the total number of cells.

Characterization of sTva. (i) Western immunoblotting.

Detection of sTva was performed by Western blotting with polyclonal Tva antiserum 40 as described previously (51).

(ii) Lectin binding.

WGA agarose and ConA agarose were washed with Triton lysis buffer (50 mM Tris [pH 8], 5 mM EDTA [pH 8], 150 mM NaCl, 1% Triton X-100) containing 10 mM MnCl₂. sTva from infected cell lysates was incubated with 1/20 volume of lectins, rocked 1 to 2 h at 4°C, and then washed three times with Triton lysis buffer containing 10 mM MnCl₂. The lectin-bound protein was eluted with Laemmli buffer and analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (12.5% acrylamide) and Western blotting. sTva precipitation from media of infected cells with WGA and ConA was performed as described above, except the washes were performed with phosphate-buffered saline (PBS) containing 10 mM MnCl₂.

(iii) Enzymatic analysis of carbohydrates.

sTva was precipitated from medium with 10% trichloroacetic acid and incubated at 0°C for 15 min. The precipitate was washed once with ice-cold acetone and resuspended in a minimal volume of 20 mM NaH₂PO₄ (pH 7.2)–1% SDS. NaOH (2 N) was used to adjust the sample to a pH of ∼9, and the sample was heated to 95°C for 5 min to completely denature sTva. The sample was then diluted 10-fold in 20 mM NaH₂PO₄ (pH 7.2) plus 0.5% Triton X-100 to yield a final SDS concentration of 0.1%. The pH was adjusted to ∼7.5 with 6 N HCl. The denatured sTva was digested overnight at 37°C with the endoglycosidases PNGase F (0.4 U/Rx), O-glycosidase (2.5 mU/Rx), and neuraminidase (0.5 U/Rx), individually or in combination. The digested products were analyzed by SDS-PAGE (12.5% acrylamide) and Western blotting as previously described (51).

To produce deglycosylated sTva for the ASLV inhibition assay (see below), purified sTva in 20 mM NaH₂PO₄ (pH 7.2) was incubated with PNGase F (10 μg of sTva/0.09 U of PNGase F) overnight at 37°C. The deglycosylated protein was purified from PNGase F on a Ni²⁺-NTA agarose column.

(iv) Tunicamycin treatment of infected cells.

Sf9 cells were infected with bVTTva as described above with the addition of 4 μg of tunicamycin/ml. The infections were allowed to proceed in the presence of tunicamycin for 48 h, and then the insect cell medium was clarified and analyzed by SDS-PAGE and Western blotting as described above.

Purification of sTva.

An Ni²⁺-NTA agarose column was washed with one bed volume of wash buffer 1 (100 mM NaH₂PO₄, 10 mM Tris, pH 7.0). sTva-containing medium that was dialyzed with 3,000-molecular-weight-cutoff dialysis tubing overnight in PBS was passed through the column twice at ∼0.5 ml/min. The column was washed with 10 bed volumes of wash buffer 1, 5 bed volumes of wash buffer 2 (initial experiments used 100 mM NaH₂PO₄–10 mM Tris–60 mM imidazole [pH 7.0]; subsequent experiments used 20 mM imidazole), and eluted with 2 bed volumes of elution buffer (100 mM NaH₂PO₄, 10 mM Tris, 500 mM imidazole, pH 7.9). Removal of imidazole, buffer exchange, and concentration of purified sTva was performed with a 3K Macrosep concentrator (Filtron, Northborough, Mass.) as described by the manufacturer, initially with 20 mM NaH₂PO₄ (pH 7.2) and subsequently with 10 mM HEPES (pH 7.0). The purity of sTva was evaluated by SDS-PAGE (12.5% acrylamide) followed by Coomassie blue staining. The sTva concentration was determined by spectrophotometry (1A₂₈₀ unit = 0.70 mg/ml, as calculated from the molar extinction coefficient determined by the computer program Protean [DNA✫, Madison, Wis.]). The concentration was verified by SDS-PAGE (12.5% acrylamide) and Coomassie blue staining of sTva compared with those of bovine serum albumin standards.

Inhibition of ASLV infection.

Serial dilutions of purified sTva (or the deglycosylated form) were added to 2.1 × 10³ infectious units of RCAS(A)AP in 1 ml (final volume) of complete TEF medium and incubated for 30 min at 22°C. Following incubation, the sTva-virus mixture was added to 1.5 × 10⁵ TEF/well of a six-well plate and placed in an incubator for 4 h. To limit virus diffusion and allow foci of infected cells to form, the cells were washed three times with PBS and overlaid with 3 ml of complete TEF medium containing 0.5% agarose. After 72 h at 37°C, the overlays were removed and the cells were fixed and stained for alkaline phosphatase activity as described previously (51). The stained foci were enumerated, and these values were plotted against the concentration of sTva. Linear regression analysis with the Cricket graph program (Computer Associates International, Inc., Islandia, N.Y.) was used to determine the IC₅₀ and IC₉₀ for sTva. Experiments were done in triplicate with two independent preparations of sTva.

Envelope binding assay.

A 96-well ELISA plate was coated overnight at 4°C with 100 μl of saturating concentrations (1:1,000 ascites fluid) of α-gD monoclonal antibody 1D3 (provided by Gary Cohen and Roselyn Eisenberg, University of Pennsylvania)/well in 20 mM Tris HCl, pH 8.5, and 0.1 M NaCl. The following morning, the plate was washed three times with 200 μl of ELISA wash buffer (1× PBS, 0.05% Tween 20)/well with agitation for 15 s between washes. To reduce nonspecific binding, the plate was subsequently blocked for 45 min at room temperature with 200 μl of ELISA block solution (1× PBS, 0.5% gelatin, 0.05% Tween 20, 0.05% NaN₃)/well. 100 μl of gD-EnvA (52) cell lysates/well in ELISA blocking buffer was added, and the solution was incubated for 1 h at 4°C. The volume of lysates used was predetermined to be sufficient to saturate the 1D3 antibody on the plate. The plate was then washed three times with ELISA wash buffer. Various concentrations of purified, biotinylated sTva (see below) in a total of 100 μl of ELISA blocking buffer/well were added and incubated at either 4, 22, or 37°C. After 1 h of incubation, the wells were washed with ELISA wash buffer three times. The bound biotinylated sTva was detected with 100 μl of avidin-horseradish peroxidase (1:2,500 dilution in 1× PBS–0.5% gelatin–0.05% Tween 20)/well for 30 min at 4°C, followed by three washes with 2× ELISA wash buffer. 2,2′-Azinobis(3-ethylbenzothiazoline)-6 sulfonic acid diammonium salt (100 μl/well) dissolved in 10 ml of ELISA substrate solution (0.1 M NaOAc, 0.1% Tween-20 [pH 4.2], 5 μl of 3% H₂O₂) was added to visualize the horseradish peroxidase activity. The plates were read with a Vmax ELISA reader (Molecular Dynamics, Sunnyvale, Calif.) at 405 nm. All experiments were performed in triplicate and repeated several times with comparable results. The data was plotted as maximum optical density as a function of sTva concentration and analyzed by nonlinear least-squares fitting with Cricket graph. The coefficients of correlation for the fitted lines were 0.97, 0.96, and 1.0 for 37, 22, and 4°C, respectively. The apparent K_d was estimated from half-maximal binding as described previously (45).

EnvA-sTva complex stability assay.

To analyze the stability of the EnvA-sTva complex, the ELISA was modified as follows: 5 nmol of biotinylated sTva in a total of 100 μl of ELISA blocking buffer was added to each well and incubated for 1 h at 4°C. After being washed three times with ELISA wash buffer, 5 nmol of unlabeled sTva in a total of 100 μl of ELISA blocking buffer was added to each well and incubated for 0, 15, 30, 60, 120, and 240 min at either 4 or 22°C. The wells were washed three times, and then bound, labeled sTva was detected as described above. The data was plotted as the percent of labeled sTva bound (maximum optical density) as a function of time. All experiments were performed in triplicate and repeated several times with comparable results.

Biotinylation of purified sTva.

Twenty micrograms of sTva was added to a 1 mg/ml stock of Sulfo-NHS-LC-biotin in PBS to yield a 400-μl reaction solution. This was incubated at 4°C for 45 min. The reaction was quenched with 40 μl of 1 M glycine for 5 min at 4°C. The biotin-labeled sTva was dialyzed against PBS for 24 h with several buffer changes. The concentration of labeled sTva was subsequently determined by spectroscopy.

RESULTS

Vector construction.

To express the subgroup A ASLV receptor protein (Tva) in a form and quantity sufficient for biochemical and structural analysis, we used a BES to produce secreted sTva. The region encoding the Tva signal sequence and extracellular domain was PCR amplified and cloned into the A. californica nuclear polyhedrosis transfer vector pVL1393. To aid in protein purification, six histidine residues were appended to the carboxyl-terminal end of the Tva extracellular domain to generate pVL1393Tva. This His₆ tag replaced one residue at the boundary of the membrane-spanning domain (Fig. 1B). A second construct, pVTTva, was generated by using the A. californica nuclear polyhedrosis transfer vector pVT-Bac (62). This vector substitutes the honeybee mellitin signal peptide, which substantially increases the secretion of proteins from insect cells, for the Tva signal peptide (58, 62). The amino terminus of the mature form of the sTva protein produced from pVTTva was predicted to contain four additional amino acids (AspProProGly) from the mellitin signal sequence and an oligonucleotide adaptor (Fig. 1B). A comparison of sTva with the two alternatively spliced forms of Tva is shown in Fig. 1A. The pVTTva and pVL1393Tva transfer vectors were used to generate recombinant baculoviruses, which were cloned by three rounds of plaque purification. Finally, to allow expression of secreted receptor protein in mammalian and avian cells, the tva sequence from the pVL1393 vector was cloned into an expression vector under control of the cytomegalovirus immediate-early promoter.

Expression of Tva in insect cells.

Monolayer cultures of Sf9 cells were infected with either the recombinant virus bVTTva or bVL1393Tva or with the parental baculovirus. To address the possibility of clonal variation in expression of Tva, three recombinant virus clones for both bVTTva and bVL1393Tva were examined. Lysates and media from infected cells were analyzed for Tva expression with a rabbit antiserum to the receptor protein (7). SDS-PAGE and Western blotting of lysates and media with the polyclonal Tva antiserum revealed that cells infected with any of the bVTTva recombinant clones expressed very high levels of Tva. Greatly reduced levels of a protein of similar size were detected in the bVL1393Tva-infected cells (data not shown). No Tva protein was detected in lysates or media from cells infected with the parent virus (data not shown). This data supports previous work suggesting that the mellitin leader aids in expression of foreign glycoproteins in insect cells (58, 62). Based on these results, one of the bVTTva recombinant virus clones was selected for large-scale production of sTva.

Optimization of sTva production.

In order to determine the optimal time for harvesting the sTva protein from the supernatant of infected Sf9 cells, a time course analysis was performed (data not shown). Sf9 cells in suspension were infected with bVTTva at a multiplicity of infection of 2, and medium was removed at various time points for analysis by Western blotting. sTva could be readily detected in the culture supernatant as early as 24 h p.i. (data not shown). Maximum levels of sTva in the medium were observed at 48 h p.i. At 72 h p.i., significant amounts of sTva were observed in the medium; however, degradation products were also apparent. By 96 h p.i., little sTva was detected. The level of sTva in the supernatant correlated with the viability of the infected cell culture. As determined by trypan blue exclusion, the infected cell culture remained >95% viable up to 50 h p.i. (data not shown). By 72 h p.i., the viability dropped to nearly 40%, suggesting that the sTva proteolytic products seen in the culture supernatant are due to cleavage by cellular or viral proteases released from dead cells. This expression pattern was highly reproducible; therefore, the 48-h p.i. time point was used for sTva production.

Characterization of sTva.

To begin characterizing the soluble receptor protein, sTva produced in insect cells was compared with Tva₉₅₀ and sTva (wild-type [WT] Tva_His6) expressed in human 293T cells. sTva produced in insect cells infected with bVTTva migrates significantly faster on SDS-PAGE than Tva₉₅₀ expressed in human cells (Fig. 2A). sTva appears as a group of four bands approximately 18 to 26 kDa in mass, while Tva₉₅₀ from human cells produces a characteristic heterogeneous population of bands migrating with apparent molecular masses between 29 and 43 kDa. The discrepancy in size between the insect cell-produced sTva and Tva₉₅₀ expressed in human cells suggested that sTva was modified differently than membrane-anchored Tva₉₅₀. Consistent with this hypothesis, expression of sTva (WT Tva_His6) in murine (data not shown) or human (Fig. 2A) cells results in a series of bands with apparent molecular masses ranging from 18 to 29 kDa (additional protein bands due to nonspecific antibody cross-reactivity [32 to 43 kDa] and proteolytic degradation [14 to 18 kDa] are also observed).

FIG. 2.

Comparison of Tvas from mammalian and insect cells and analysis of purified sTva. (A) Cell lysates from human 293T cells transiently expressing either Tva₉₅₀ or sTva and lysates from Sf9 cells infected with bVTTva were subjected to SDS-PAGE (12.5% acrylamide), and Western blotting was performed with a polyclonal α-Tva serum. (B) Analysis of purified sTva. To assess the purity of sTva, 0.5 (left) and 10 (right) μg of the same preparation were analyzed by SDS-PAGE (12.5% acrylamide) and Coomassie blue staining. Molecular mass standards (in kilodaltons) are indicated.

sTva contains three potential N-linked glycosylation sites as well as a region rich in serine and threonine residues that could be modified by O-linked carbohydrates. To determine if the four bands observed when sTva is produced in insect cells represent differential carbohydrate modifications, sTva was precipitated from Sf9 cell lysates with lectin-agarose beads. Both WGA- (data not shown) and ConA-agarose precipitated the three largest species of sTva from infected Sf9 cell lysates but not the fastest-migrating species (Fig. 3A). WGA binds to N-acetylglucosamine residues in both N- and O-linked oligosaccharides, whereas ConA binds to α-linked mannose residues in the N-linked core oligosaccharide. Thus, these results suggest that the upper three slower-migrating bands of sTva represent forms of the soluble receptor with one, two, or three N-linked oligosaccharides, and because sTva was precipitated by WGA, it may contain O-linked glycosylation.

FIG. 3.

sTva is modified by asparagine-linked glycosylation. (A) sTva binds to ConA. Lysates of Sf9 cells infected with bVTTva were incubated with ConA sepharose. After extensive washing, ConA-bound proteins were eluted and analyzed by Western blotting as described in the legend to Fig. 2. (B) Analysis of sTva glycosylation by glycosidase digestion. sTva was acid precipitated from the medium of bVTTva-infected Sf9 cells. Pelleted sTva was resuspended in the appropriate glycosidase buffer and digested with PNGase F (F; lanes 2 and 5), neuraminidase (N; lanes 3, 4, and 5), and O-glycosidase (lanes 4 and 5). Following glycosidase digestion the samples were analyzed by SDS-PAGE and Western blotting as described in Fig. 2. Molecular mass standards (in kilodaltons) are indicated on the left.

To confirm whether the Sf9-produced sTva was modified by N and/or O glycosylation, we subjected sTva to digestion with specific glycosidases. Neither neuraminidase nor O-glycosidase treatment affected the mobility of sTva (Fig. 3B, lanes 3 and 4), suggesting that the protein is not significantly modified by O-linked sugars or sialic acid. In contrast, treatment of sTva with PNGase F, which specifically removes N-linked carbohydrates, condensed the characteristic four bands of sTva to a single species that comigrated with the fastest-migrating form (18 kDa) seen in cell lysates (Fig. 3B, lanes 2 and 5). The PNGase F-treated sTva was incompetent to bind to either WGA or ConA (data not shown). Additionally, when bVTTva-infected cells are grown in the presence of tunicamycin, which inhibits N-linked glycosylation, only a single form of secreted sTva could be detected by Western blot analysis, migrating at 18 kDa (data not shown). Taken together, these results are consistent with N-linked modification but not O glycosylation of sTva, with the upper three forms of sTva representing proteins with one, two, or three N-linked sugars and the fastest-migrating species representing an unmodified form of the receptor.

Purification of sTva.

His₆-tagged sTva protein can be rapidly and quantitatively purified from culture supernatants of bVTTva-infected cells by using the protocol outlined in Materials and Methods. The main features of this protocol are an initial dialysis step followed by binding to metal-chelating resin (NTA), a low-concentration imidazole wash to remove contaminating proteins adsorbed to the column, and, finally, elution of sTva with a high concentration of imidazole. The dialysis step was added to the protocol for all cultures larger than 100 ml, since it appeared that compounds in the insect cell growth medium remove the nickel from the NTA column, interfering with sTva purification. The eluted soluble receptor protein is >95% pure as judged by staining with Coomassie blue following SDS-PAGE analysis of overloaded sTva samples (Fig. 2B, right). Furthermore, three glycosylated forms of sTva are present in the purified protein when the purified protein is analyzed by SDS-PAGE and Coomassie blue staining (Fig. 2B, left). The extinction coefficient predicted from the sTva primary amino acid sequence was used to estimate the yield of purified protein from various preparations at 7 to 10 mg/liter of infected Sf9 cells. This estimate was confirmed by SDS-PAGE by comparing purified sTva to bovine serum albumin concentration standards (data not shown).

Inhibition of ASLV infection by sTva.

To determine if the purified sTva retained biological activity, we assessed the ability of the purified protein to inhibit infection by ASLV(A). A subgroup A Rous sarcoma virus vector carrying an alkaline phosphatase reporter gene, RCAS(A)AP, was preincubated with various concentrations of purified sTva and then added to TEF. Following removal of the virus inoculum, the cells were overlaid with soft agar to allow foci to develop. Two days after infection, the overlay was removed, the cells were stained for alkaline phosphatase activity, and the alkaline phosphatase-positive foci were enumerated. sTva inhibited infection by RCAS(A)AP in a dose-dependent manner (data not shown). The IC₅₀ for sTva inhibition of RCAS(A)AP infection was found to vary between 6 and 25 pM in numerous experiments, with an average of 14 pM (Table 1). A PNGase F-treated form of sTva similarly inhibited RCAS(A)AP, indicating that saccharide moieties are not required for envelope-receptor interactions (data not shown). As expected, infection by subgroup C ASLV, which does not utilize Tva, was unaffected by sTva (data not shown). Thus, it appears that sTva retains specific biological activity effectively neutralizing ASLV(A).

TABLE 1.

Inhibition of ASLV(A) infection by sTva^a

IC (%)	Value (pM)
	Expt 1	Expt 2	Expt 3	Mean ± SDb
50	5.5	25	12	14 ± 10
90	8.6	44	25	26 ± 18

^aAssays were performed as described in Materials and Methods. 

^bSD, standard deviation. 

Analysis of sTva binding to ASLV envelope protein.

Neutralization of RCAS(A)AP by sTva suggested that the soluble, baculovirus-produced receptor was capable of binding to the subgroup A envelope glycoprotein, EnvA. To directly address this question and to quantitatively analyze the interaction between sTva and EnvA, an ELISA-based binding assay was developed. This assay is similar to an assay used to measure HIV-1 glycoprotein binding to a soluble form of the cellular receptor CD4 (45).

Epitope-tagged EnvA (gD-EnvA) was captured on the ELISA plate with a monoclonal antibody to the gD tag. sTva was labeled by biotinylation with a primary amine-reactive reagent. Since sTva contains no lysine residues, labeling with this reagent should attach a single biotin molecule only to the amino terminus of the protein. Labeled sTva was added at various concentrations to wells containing captured gD-EnvA and incubated for 30 min at the indicated temperature (Fig. 4A). After washing, bound sTva was detected with peroxidase-conjugated avidin. In numerous experiments, sTva bound to the captured gD-EnvA in a concentration-dependent manner, achieving saturable binding at concentrations in excess of 2 to 4 nM. Moreover, the data in Fig. 4A fits a simple bimolecular noncooperative-binding curve. From this data, the binding affinity of sTva was estimated to have an apparent K_d of 0.3 nM at 37°C. The apparent K_ds at 4 and 22°C were very similar, indicating that sTva binding to EnvA is temperature independent.

FIG. 4.

ELISA analysis of sTva-ASLV(A) envelope binding. (A) Titration of sTva. Increasing concentrations of purified sTva were incubated with captured gD-EnvA at the indicated temperature (○, 37°C; ◊, 25°C; □, 4°C) in an ELISA as described in Materials and Methods. The percent of sTva bound for each concentration is the average of three experiments. The error bars indicate standard deviations. The apparent K_ds for sTva-EnvA were estimated from the half-maximal binding and are 0.3, 0.2, and 0.5 nM at 37, 25, and 4°C, respectively. (B) Analysis of sTva-EnvA complex stability. The relative dissociation rate of the sTva-EnvA complex was determined by preincubating excess biotinylated sTva with gD-EnvA which was captured on an ELISA plate. After extensive washing, excess unlabeled sTva was incubated at 4°C for the indicated times. The percentage of labeled sTva bound is the average of three experiments. The error bars indicate standard deviations. The sTva-EnvA dissociation rates at 22 (○) and 4°C (●) are ∼5.5 and >6 h, respectively.

The ELISA was also used to analyze the stability of the sTva-EnvA interaction (Fig. 4B). A saturating amount (5 nM ≅18 × K_d) of biotinylated sTva was added to captured gD-EnvA and allowed to achieve equilibrium. After removal of the unbound labeled sTva and washing, excess unlabeled sTva (5 nM) was added and incubated for the indicated times at 4 or 22°C (Fig. 4B). Even after 6 h at 4°C, 50% of the labeled sTva remained bound to EnvA, suggesting a complex half-life of ∼6 h. Similar results were obtained at 22°C. This extremely slow off rate attests to the stability of the sTva-EnvA complex.

DISCUSSION

In this report, we describe the construction, expression, and characterization of a soluble form of Tva using a BES. sTva secreted into the medium of infected insect cells is differentially modified by N-linked glycosylation but does not appear to contain significant O link modifications. Large quantities of sTva can be purified from the medium of infected insect cells, and this purified protein is capable of efficiently and specifically neutralizing ASLV(A). In addition, using an ELISA-based binding assay, sTva-EnvA interactions were investigated and an apparent K_d of 0.3 nM and a complex half-life of 6 h were demonstrated.

Four different protein species of sTva from infected Sf9 cell lysates were identified, with masses of 18 to 26 kDa, suggesting that the three N-linked glycosylation sites in sTva are differently utilized. In contrast, expression of the same truncated form of Tva, as well as Tva₉₅₀, in mammalian cells resulted in additional protein bands, indicating differences in posttranslational processing of insect- and mammalian-derived proteins (Fig. 2A). Indeed, lectin binding and glycosidase digestion of sTva as well as tunicamycin treatment of infected insect cells confirmed that sTva contains N-linked glycans but does not appear to be O glycosylated (Fig. 3). This is in agreement with previous data demonstrating that proteins expressed in Spodoptera frugiperda cells are N link modified but are either not O glycosylated or are underglycosylated with either the monosaccharide N-acetylglucosamine (69) or N-acetylgalactosamine (63) or the disaccharide galactose(β1-3)N-acetylgalactosamine (29, 36, 63).

Using chimeric Tva-CD8 proteins, we have previously shown that there is no requirement for specific N or O glycosylation for ASLV(A) receptor function (51). The results with sTva confirm this observation. A PNGase F-treated form of sTva, in which there is no detectable remaining N-linked glycosylation, was fully competent to neutralize ASLV(A) infection of primary turkey cells, indicating that saccharide moieties have no apparent role in Tva viral receptor function.

Soluble forms of viral receptors for HIV (14, 22, 27, 35, 59, 64), Epstein-Barr virus (48), mouse hepatitis virus (55), poliovirus (38), and rhinovirus (42) have been shown to be biologically active by inhibition of virus infection, and therefore are most likely correctly folded. Previous experiments with media from mammalian cells expressing a secreted form of Tva suggested it also could block ASLV(A) infection (15). To determine whether purified sTva retained biological activity, we evaluated the ability of sTva to block ASLV(A) infection. Preincubation of sTva with virus potently inhibited virus infection of primary avian cells, with an IC₉₀ of 26 pM (Table 1), indicating that the protein is biologically active and therefore likely correctly folded. Moreover, sTva is a significantly more potent inhibitor of virus infection, with an IC₉₀ nearly 10³-fold lower than that of sCD4 for laboratory-adapted HIV-1 infection (14, 27, 35) and 10⁶-fold lower than that of soluble intercellular adhesion molecule 1 (sICAM-1) for rhinovirus infection (42). The differences among the antiviral activities of the three soluble receptors are likely due to mechanistic differences in the way the receptors neutralize their cognate viruses. sICAM-1 inhibits rhinovirus infection primarily by saturating the receptor binding sites on the virus (30). Similarly, sCD4 neutralization of HIV-1 appears to be mediated by receptor binding to virions; however, gp120-induced shedding from the virion may contribute to viral neutralization (23, 49).

The mechanism of sTva inhibition of ASLV(A) infection appears to be by irreversible inactivation of the ASLV virion. In sharp contrast to CD4, which requires a coreceptor for HIV-1 infectivity, Tva appears to be necessary and sufficient to mediate ASLV(A) entry (3, 7, 72). Binding of sTva to EnvA induces a conformational change in the SU subunit, exposing previously inaccessible protease sites (28), and leads to the formation of stable, higher-order oligomers of the TM subunit (3a) reminiscent of structures for the putative fusogenic forms of influenza HA2 (11, 13), Ebola virus GP2 subunits (67), and the murine leukemia virus (MLV) (26) and HIV-1 TM subunits (12, 60, 68). In addition, sTva binding to a soluble, oligomeric form of EnvA converts the envelope glycoprotein to a membrane-associated form consistent with activation of EnvA to a fusogenic state (21, 32). In this case, upon EnvA activation, sTva dissociates (21, 32), perhaps permitting it to reassociate with a new EnvA molecule. The extremely low IC₅₀ seen for sTva would suggest the soluble-receptor neutralization is mediated by a hit-and-run mechanism, where sTva induces irreversible conformational changes that inactivate EnvA and then dissociates to inactivate additional envelope molecules. Taken together, these results confirm the biological activity of sTva and demonstrate the utility of this reagent for dissecting the early events in ASLV(A) entry.

sTva inactivation of ASLV(A) is initiated by binding to EnvA, since it had no effect on ASLV(C) infection. To further evaluate the interaction between EnvA and sTva, we developed a sensitive ELISA. Half-maximal binding is a function of the equilibrium dissociation constant (K_d); however, because equilibrium was not maintained by the addition of sTva during wash and detection steps, the K_d value obtained is an approximation. With this caveat, the apparent K_d for sTva-EnvA is 0.3 nM at 37°C (Fig. 4A). The magnitude of the apparent K_d suggests that any dissociation is negligible. In addition, the apparent K_d for sTva-EnvA is temperature independent. While the K_ds for sCD4-gp120 at different temperatures are similar, the kinetics of binding vary with temperature (23). The kinetic parameters for EnvA-Tva will need to be defined to determine whether a similar situation exists. Table 2 summarizes the K_d values determined for a number of viruses. As Table 2 illustrates, the apparent K_d determined for the sTva-EnvA system is similar to that determined with a fluorescence-activated cell sorter-based assay (75). Moreover, it is within the range of many other virus-receptor K_ds, implying that strong affinity between viral glycoproteins and their cognate receptors is a common, conserved feature of a number of viruses, regardless of whether they are enveloped (e.g., HIV-1 and ASLV) or nonenveloped (e.g., poliovirus and reovirus).

TABLE 2.

Comparison of virus-receptor binding constants

Virusa	Kd (nM)	Reference(s)
HIV-1	0.83–11.6	23, 37, 39, 40, 44, 46, 59
HIV-2	48.5	45
ASLV(A)	1.5	75
ASLV(A)	0.3	This report
E-MLV	3.3	73
Poliovirus	0.21	9
HCMV	1.1	61
AcNPV	0.235	70
VSV	0.15	57
CPV	0.029	5
ASF	0.07	2
Reovirus type 1 and 3	0.5	24, 66

^aE-MLV, ecotropic MLV; HCMV, human cytomegalovirus; AcNPV, A. californica nuclear polyhedrosis virus; VSV, vesicular stomatitis virus; CPV, canine parvovirus; ASF, African swine fever virus. 

Consistent with the low apparent K_d, the sTva-EnvA complex is extremely stable. Complex half-life estimations indicate that 50% of the purified receptor is still bound to EnvA after up to 6 h and the binding is not significantly affected by temperature (Fig. 4B). In contrast, after 90 min, 50% of ecotropic MLV bound to the cell surface dissociates at 25°C (73). The difference in complex stability is even more striking considering the multivalent interactions between MLV and the cell surface versus the monovalent interaction between sTva and EnvA. Analysis of HIV-1 binding suggests this interaction is also relatively less stable, since 50% of the HIV-1 gp120-CD4 complex dissociates within ∼30 min at 37°C (44). A possible reason for the high stability of the Tva-EnvA interaction is that avian cells express levels of Tva on their surface that are too low to be detected by fluorescence-activated cell sorter, immunofluorescence, or Western blot analysis (7). Since cooperative binding appears to be required for EnvA activation (21), an extremely stable Env-receptor complex may be required to ensure enough envelope glycoproteins are bound for activation of membrane fusion. Interestingly, once Env is activated, receptor readily dissociates (21, 32), confirming the intimate relationship of Tva and EnvA in coordinating virus entry.

The quality and quantity of sTva produced from Sf9 cells has allowed us to begin structural analysis of the receptor. From crystal trials, we have identified conditions that produce crystals of sufficient size and quality to be competent to diffract X rays. Mapping of mutations in Tva that affect viral receptor function onto the structure will assist in understanding envelope-receptor interactions as well as LDLR-ligand interactions. Additionally, this will be the first structure of an LDL-A module in which biological activity can be demonstrated. It will be interesting to compare it to previously determined structures of hLDLR modules refolded from Escherichia coli-produced proteins (19, 20, 25).