Naturally Occurring Frameshift Mutations in the tvb Receptor Gene Are Responsible for Decreased Susceptibility of Chicken to Infection with Avian Leukosis Virus Subgroups B, D, and E

ABSTRACT

The group of highly related avian leukosis viruses (ALVs) in chickens are thought to have evolved from a common retroviral ancestor into six subgroups, A to E and J. These ALV subgroups use diverse cellular proteins encoded by four genetic loci in chickens as receptors to gain entry into host cells. Hosts exposed to ALVs might be under selective pressure to develop resistance to ALV infection. Indeed, resistance alleles have previously been identified in all four receptor loci in chickens. The tvb gene encodes a receptor, which determines the susceptibility of host cells to ALV subgroup B (ALV-B), ALV-D, and ALV-E. Here we describe the identification of two novel alleles of the tvb receptor gene, which possess independent insertions each within exon 4. The insertions resulted in frameshift mutations that reveal a premature stop codon that causes nonsense-mediated decay of the mutant mRNA and the production of truncated Tvb protein. As a result, we observed that the frameshift mutations in the tvb gene significantly lower the binding affinity of the truncated Tvb receptors for the ALV-B, ALV-D, and ALV-E envelope glycoproteins and significantly reduce susceptibility to infection by ALV-B, ALV-D and ALV-E in vitro and in vivo. Taken together, these findings suggest that frameshift mutation can be a molecular mechanism of reducing susceptibility to ALV and enhance our understanding of virus-host coevolution.

IMPORTANCE Avian leukosis virus (ALV) once caused devastating economic loss to the U.S. poultry industry prior the current eradication schemes in place, and it continues to cause severe calamity to the poultry industry in China and Southeast Asia, where deployment of a complete eradication scheme remains a challenge. The tvb gene encodes the cellular receptor necessary for subgroup B, D, and E ALV infection. Two tvb allelic variants that resulted from frameshift mutations have been identified in this study, which have been shown to have significantly reduced functionality in mediating subgroup B, D, and E ALV infection. Unlike the control of herpesvirus-induced diseases by vaccination, the control of avian leukosis in chickens has relied totally on virus eradication measures and host genetic resistance. This finding enriches the allelic pool of the tvb gene and expands the potential for genetic improvement of ALV resistance in varied chicken populations by selection.

INTRODUCTION

For retroviruses to infect host cells, they must fuse themselves to a cellular membrane of the target host with their viral membrane. This fusion process, followed by entry into the host cells, is initiated by an interaction between the viral envelope glycoprotein and a cellular surface protein that acts as a receptor (1, 2). When this interaction leads to a perfect fusion between the viral envelope proteins and host receptors, virus entry succeeds (3). In spite of the complexity and specificity of this interaction, highly related family members of retroviruses have evolved by altering their envelope glycoprotein sequences, which demand different cellular proteins on the host side as receptors to mediate successful entry (4). The family of avian leukosis virus (ALV) in chickens contains six closely related subgroups, A to E and J, which are thought to have evolved from a common viral ancestor. The susceptibility of host cells to these viruses is determined by four genetic loci in chickens, namely, tva, tvb, tvc, and tvj (5), which encode different cell surface proteins as receptors. The tva locus encodes the Tva protein, the receptor for ALV subgroup A (ALV-A), which is a member of the low-density lipoprotein receptor (LDLR) family (6, 7). The tvb locus encodes the Tvb protein, the receptor for ALV-B, ALV-D, and ALV-E, which is related to the tumor necrosis factor receptor (TNFR) family (8–10). The tvc locus encodes the Tvc protein, the receptor for ALV-C, which is related to a member of the immunoglobulin protein Ig family (11). The membrane-spanning cell surface protein Na⁺/H⁺ exchanger type 1 (NHE1), encoded by the tvj locus, was identified as the receptor for ALV-J (12). Additionally, chicken annexin A2 (chANXA2) was also considered a receptor for ALV-J (13). Thus, the fact that different subgroups of ALV in chicken successfully acquired perfectly matching receptors illustrates the capability and evolutionary successes of avian leukosis viruses to sustain their survival and replication through modifying their highly conserved envelope glycoproteins (4, 5, 14).

Selection forces imposed by novel ALV subgroups, on the other hand, drive the positive selection of receptors with variants that decrease or even abrogate binding affinity to viral glycoproteins. The complete resistance or decreased susceptibility of host cells to a particular ALV subgroup can be derived from genetic mutations of a specific receptor locus, which may either eliminate the receptor protein synthesis through interruption of the gene expression or alter the receptor protein structure to lower the binding affinity between the viral glycoproteins and the receptor. Consequently, the susceptibility of host cells to ALV strictly depends on the expression and proper display of host receptor proteins. The genetic defects that confer host cell resistance to ALV subgroups A, B/D/E, and C, the tva^r, tvb^r, and tvc^r alleles, respectively, have been identified and reported in several inbred lines of White Leghorns as well as in some Chinese broiler lines (10, 11, 15–19). The resistance can be caused by premature termination codons (PTCs) or frameshift mutation in the receptor-encoding genes (11, 15, 16), disrupted mRNA splicing due to intronic deletions (17, 18), and even single cysteine residue substitutions in the receptor peptide sequence (10, 19). In addition, deletions of the critical tryptophan 38 in the first extracellular loop (ECL1) of NHE1 explain the resistance to ALV-J in galliform species (20, 21). From the point of view of coevolution between ALVs and host, it is attractive to speculate that slight structural changes affecting the virus binding sites yet retaining the normal cellular functions of the receptor proteins could be the starting point for coevolution of the specific subgroups of ALV and the host cells.

The Tvb protein contains three extracellular cysteine-rich domains (CRD1 to -3) and a functional cytoplasmic death domain that is postulated to activate apoptosis in cells (22, 23). Multiple alleles of the tvb gene and corresponding Tvb receptors have been identified. The tvb^s1 allele encodes a Tvb^S1 receptor that confers susceptibility to ALV-B, ALV-D, and ALV-E, while the tvb^s3 allele encodes a Tvb^S3 receptor that confers susceptibility to only ALV-B and ALV-D (10). Tvb^S3 differs from Tvb^S1 by a C62S substitution in CRD2 of Tvb^S1, which presumably alters the structure of CRD2, resulting in the loss of ALV-E binding affinity and therefore causing resistance to ALV-E. The turkey homolog of this gene, the tvb^t allele, encodes an ALV-E-specific receptor (9). The major determinants of the Tvb receptor critical for the binding and entry of ALV-B, ALV-D, and ALV-E were previously reported (24, 25). Amino acid residues 32 to 46 in CRD1 of Tvb^S1 are sufficient for it to act as a receptor for ALV-B and ALV-D. Specifically, Leu-36, Gln-37, and Tyr-42 appear to be the essential amino acid residues to support the fundamental functionality of the receptor (24). In contrast, amino acid residues Tyr-67, Asn-72, and Asp-73 in CRD2 of Tvb^S1 are necessary for efficient ALV-E binding and entry (25). The tvb^r allele contains an in-frame stop codon at residue 57, which confers complete resistance to ALV-B, ALV-D, and ALV-E infection (15). Furthermore, the tvb^r2 allele encodes a Tvb^R2 receptor with a C125S substitution in CRD3, which significantly reduces the susceptibility to ALV-B and ALV-D infection and virtually eliminates ALV-E infection in vitro and in vivo (19).

The resistance alleles not only confer host resistance to specific subgroups of ALV but also provide host targets for antiviral strategies. The natural ALV infection of chickens causes tremendous economic losses for the poultry industry as a consequence of tumor-induced mortality, growth retardation, serious immunosuppression, and costs incurred from management of prevention and eradication of the viruses (26, 27). Therefore, it is urgently needed to identify naturally occurring mutations that confer host resistance to ALV infection. ALVs are prevalent on commercial poultry farms worldwide, particularly in China and Southeast Asia (27–29). Thus, populations of chickens exposed to ALVs may be subjected to selection pressure, which mediates development of host complete or partial genetic resistance to ALV infection. In order to identify additional resistance alleles of ALV receptor loci, we systematically screened specific receptor genes in a panel of Chinese commercial broiler lines for naturally preexisting polymorphisms (30). We have previously identified two tva alleles causing reduced susceptibility to ALV-A infection that resulted from intronic deletions within the tva receptor gene (18). In this study, we have identified novel molecular defects that account for decreased susceptibility to ALV-B, ALV-D, and ALV-E infection in some Chinese commercial broiler lines. The molecular defects include two separate insertions (c.291_292insAG, named the tvb^r4 allele, and c.359_360insA, named the tvb^r5 allele) within exon 4 of the tvb gene, which caused a frameshift and resulted in a premature stop codon that leads to the generation of truncated Tvb proteins unable to function as efficient receptors for ALV-B, ALV-D, and ALV-E.

RESULTS

Identification of novel tvb^r4 and tvb^r5 alleles in Chinese broiler lines.

In our previous study, we identified intronic deletions of the tva receptor gene that result in reduced susceptibility to subgroup A ALVs (18). This finding is suggestive for the presence of similar genetic mutations within other ALV receptor genes, and Chinese broiler lines that have undergone long-term virus eradication measures may manifest such recessive alleles (30). Two tvb allelic variants were identified in Chinese commercial broiler lines by generating a series of amplicons for the genomic region of the tvb receptor gene followed by sequencing, which were designated the tvb^r4 and tvb^r5 alleles. tvb^r4 has a 2-nucleotide insertion, and tvb^r5 has a single nucleotide insertion. Both of the insertions are located within exon 4 of the tvb gene but at separate spots (Fig. 1A and B). The distributions of the tvb^r4 and tvb^r5 alleles in Chinese commercial broiler lines were obviously different; the tvb^r4 allele was prevalently distributed in line CB03, while the tvb^r5 allele was prominently distributed in line CB08 (data not shown).

FIG 1.

The tvb^r4 and tvb^r5 alleles contain a 2-nucleotide insertion and a 1-nucleotide insertion, respectively. (A) Schematic representation of the structure of the tvb receptor gene. (B) Sequence traces obtained from genomic DNA of homozygous wild-type and heterozygous and homozygous tvb^r4 birds, showing the insertion of the AG dinucleotide in the mutant (left), and of homozygous wild-type and heterozygous and homozygous tvb^r5 birds, showing the insertion of the A mononucleotide in the mutant (right). The position of the inserted nucleotides is underlined in bold. (C) Partial nucleotide and deduced amino acid sequences of exon 4 surrounding the AG dinucleotide insertion that distinguishes tvb^r4 from tvb^s1 and the A mononucleotide insertion that distinguishes tvb^r5 from tvb^s1 (numbered from the start methionine codon). The 2-nucleotide insertion and altered amino acid sequence of tvb^r4 allele, as well as the 1-nucleotide insertion and altered amino acid sequence of tvb^r5 allele, are underlined.

To confirm the insertions in the tvb^r4 and tvb^r5 transcripts, we amplified the entire tvb coding sequence from the cDNAs of predetermined tvb^s1/s1 (wild-type [wt]), tvb^r4/r4, and tvb^r5/r5 chicken embryo fibroblasts (CEFs). Analyses of the complete transcripts assembled from the reverse transcription-PCR (RT-PCR) product sequences of the tvb^s1/s1, tvb^r4/r4, and tvb^r5/r5 CEFs showed that the tvb^r4 transcript contained the 2-nucleotide insertion (AG) located in exon 4 at nucleotide positions 291 and 292 of the tvb cDNA (c.291_292insAG), while the tvb^r5 transcript contained a single nucleotide insertion (A) in exon 4 at nucleotides 359 and 360 of the tvb cDNA (c.359_360insA). We then determined the deduced amino acid sequences of the tvb cDNA products and compared the peptide sequence of tvb^s1 with the peptide sequences of tvb^r4 and tvb^r5. These two insertions both resulted in alteration of the amino acid sequence of the Tvb receptor protein (Fig. 1C).

Mutant tvb mRNAs are targeted by the NMD pathway.

Both c.291_292insAG and c.359_360insA cause a frameshift mutation, which results in a premature stop codon in the fifth and seventh exons of the 10 exons of the tvb gene, respectively. The mutant mRNAs were therefore predicted to undergo nonsense-mediated mRNA decay (NMD) (31). To test this prediction, we first compared the tvb transcript levels of the heterozygotes between the wild type and each of the mutant types (c.291_292insAG and c.359_360insA) in CEFs by direct sequencing of RT-PCR products encompassing the insertions. As shown in Fig. 2A and B, the mutant transcripts were scarcely detectable. We then compared the levels of tvb mRNA in CEFs of the tvb^s1/s1, tvb^s1/r4, and tvb^r4/r4 genotypes and in CEFs of the tvb^s1/s1, tvb^s1/r5, and tvb^r5/r5 genotypes using reverse transcription-quantitative PCR (RT-qPCR) performed with primer sets targeting the 5′ and 3′ ends of the tvb mRNA separately. Highly significant reductions in tvb mRNA levels were observed in tvb^s1/r4 and tvb^s1/r5 heterozygotes compared to the homozygous wild-type individuals (37% ± 5% and 76% ± 8% of control values, respectively, for the 5′ system and 10% ± 3% and 35% ± 4% of control values, respectively, for the 3′ systems), while tvb mRNA levels in tvb^r4/r4 CEFs were less than 5% and those in tvb^r5/r5 CEFs were less than 10% of those for the homozygous wild types, respectively (Fig. 2C and D). Taking our findings together, we demonstrated that both the allelic imbalance and RT-qPCR experiments indicated degradation of the mutant transcripts by NMD.

FIG 2.

Nonsense-mediated RNA decay of c.291_292insAG and c.360_361insA mutant tvb transcripts. (A and B) Direct sequencing of tvb amplicons spanning the c.291_292insAG and c.360_361insA mutations obtained from genomic DNA and pulmonary cDNA of the heterozygous tvb^s1/r4 (A) and tvb^s1/r5 (B) CEFs, showing the virtually exclusive detection of the wild-type allele among transcripts. The position of the inserted nucleotide is underlined in bold, and the sequencing direction is represented by a triangle. (C and D) Comparison of tvb mRNA levels in the CEFs of wild-type, tvb^s1/r4, and tvb^r4/r4 birds (C) and in the CEFs of wild-type, tvb^s1/r5, and tvb^r5/r5 birds (D). Data are shown for two amplicons at the 5′ and 3′ ends of the tvb mRNA. Error bars correspond to standard errors of the means over three replicates per sample.

Frameshift mutations lead to truncated Tvb receptors.

c.291_292insAG of tvb^r4 causes a frameshift mutation in the tvb mRNA at the 98th codon in exon 4 and introduces a premature TGA stop codon at the 131st codon in exon 5, which is presumably to append a frameshifted 33-residue peptide to a truncated Tvb receptor missing a portion of residues in CRD3, the membrane-spanning domain (MSD), and the cytoplasmic death domain (Fig. 3A and B). Similarly, c.359_360insA of tvb^r5 causes a frameshift in the open reading frame (ORF) of the tvb mRNA at the 121st codon in exon 4 and also introduces a premature TGA stop codon at the 190th codon in exon 7, which is presumably to append a frameshifted 69-residue peptide to a truncated Tvb receptor missing the cytoplasmic death domain (Fig. 3A).

FIG 3.

The tvb^r4 and tvb^r5 alleles both generate a truncated Tvb protein product. (A) Schematic drawings of the Tvb^S1, Tvb^R4, and Tvb^R5 proteins. The numbering of the Tvb^S1 protein includes the signal peptide (SP), cysteine-rich domains (CRD1 to -3), membrane-spanning domain (MSD), and cytoplasmic death domain. The positions of the frameshift mutations in Tvb^R4 and Tvb^R5 are highlighted with a triangle, while the premature TGA stop codons in Tvb^R4 and Tvb^R5 are indicated by an asterisk. (B) Comparison of the deduced amino acid sequences of the CRDs of the known Tvb receptors. The putative disulfide bonds in the CRDs of Tvb^S1 are indicated based on comparisons with the closely related TNFR proteins and the X-ray structure of the DR5 TRAIL receptor protein. The C62S substitution in Tvb^S3, the C125S substitution in Tvb^R2, and the putative cysteine substitutions in Tvb^R4 and Tvb^R5 are circled. The residues in the frameshifted peptides of Tvb^R4 and Tvb^R5 are indicated in bold. (C) Expression of the wt and mutant forms of Tvb. 293FT cells transfected with expression plasmids encoding wt or mutant forms of Tvb receptor constructs, as well as the pEGFPC1 or empty vectors (mock), were subjected to SDS-PAGE and Western blot analysis using an anti-GFP rabbit polyclonal antibody, followed by a horseradish peroxidase (HRP)-conjugated goat anti-rabbit IgG (H+L) antibody, and the bound protein complexes were visualized by chemiluminescence. The control lanes 1 and 2 are the same as lanes 3 and 4 in Fig. 3B in our recent study (32). Molecular masses in kilodaltons are shown on the right.

To verify these predictions, we constructed Tvb expression vectors encompassing the wild-type tvb^s1 (pEGFPC1-tvb^s1) and the mutant types tvb^r4 and tvb^r5 (pEGFPC1-tvb^r4 and pEGFPC1-tvb^r5). The pEGFPC1 vector itself was used as a control. The pEGFPC1 vector and the Tvb expression constructs were used to transfect 293FT cells, followed by Western blotting analyses of the culture supernatants. The green fluorescent protein (GFP) and Tvb/GFP fusion proteins migrated as broad bands ranging from 27 kDa to 70 kDa due to glycosylation (Fig. 3C). A band corresponding to Tvb^S1-GFP fusion proteins with a molecular mass of ca. 70 kDa was observed (the sizes of the Tvb^S1 and GFP proteins are 43 kDa and 27 kDa, respectively). In contrast, the band sizes of the Tvb^R4-GFP and Tvb^R5-GFP fusion proteins were 41 kDa and 48 kDa, respectively (the sizes of Tvb^R4 and Tvb^R5 are 14 kDa and 21 kDa, respectively). Endogenous Tvb expression was barely visible as shown by the signal in mock-transfected 293FT cells (Fig. 3C). Consequently, the Western blot data were consistent with the predictions for the transcripts of the novel tvb^r4 and tvb^r5 alleles, which differed from the transcript of the tvb^s1 allele due to the introduction of the premature stop codons.

The cytoplasmic death domain of the Tvb receptor can activate cellular apoptosis signaling pathways to trigger cell death (8, 23). To further evaluate whether the Tvb^R4 or Tvb^R5 receptor protein lost the cytoplasmic death domain, we evaluated the cellular apoptosis activity of the Tvb^R4 and Tvb^R5 receptors. As anticipated, when Tvb^S1 was overexpressed in 293FT cells after 24 h, more than 24.2% of the transfected cells underwent morphological changes characteristic of apoptosis (Fig. 4A and B). However, the morphological structure and percentages of apoptosis of cells overexpressing Tvb^R4 or Tvb^R5 were not significantly different from those of the mock vector-transfected cells (Fig. 4A and C). Collectively, these data provided the key evidence that both the tvb^r4 and tvb^r5 alleles yield truncated Tvb receptors and that the deletion of the death domains of Tvb^R4 and Tvb^R5 did occur.

FIG 4.

Ectopic expression of Tvb^R4 or Tvb^R5 inhibits cellular apoptosis activity. 293FT cells were transiently transfected with expression plasmids encoding wt or mutant forms of Tvb receptor constructs or with the pEGFPC1 vector. At 24 h after transfection, cells were stained with an annexin V-FITC/PI apoptosis detection kit and examined. (A) Morphological changes of 293FT cells overexpressing Tvb^S1, Tvb^R4, Tvb^R5, or vector alone. Representative fields of view were captured at 24 h posttransfection with an inverted fluorescence microscope. (B and C) The apoptotic cells were quantitated by flow cytometry, and the percentage of apoptotic cells is indicated as a mean from three parallel dishes. Error bars show the standard error of mean wherever they are large enough to be visible. To determine the statistical significance of differences between the amounts of apoptotic cells triggered by wt and mutant Tvb receptors, unpaired two-tailed t tests were performed (**, P < 0.01).

Frameshift mutations significantly reduce the binding affinity between the Tvb receptors and the ALV-B, -D, and -E envelope glycoproteins.

The conserved cysteines in the CRDs are critical for the proper folding of the Tvb protein. Compared with the deduced amino acid sequences of the CRDs of the Tvb^S1 wild-type receptor, six conserved cysteines in CRD2 and CRD3 of the Tvb^R4 receptor and four conserved cysteines in the CRD3 of the Tvb^R5 receptor were replaced due to the frameshift mutations (Fig. 3B). Presumably, the replacements of these cysteine residues altered the structure of the Tvb receptor protein by abolishing the formation of the correct intrachain disulfide bonding patterns. Despite these defects, the Tvb^R4 and Tvb^R5 proteins can still function somehow as receptors for ALV-B, ALV-D, and ALV-E, albeit at a much lower efficiency. We hypothesized that the frameshift mutations reduced the binding affinity of the Tvb^R4 and Tvb^R5 receptors for the ALV-B, ALV-Ds and ALV-E glycoproteins.

To test this hypothesis, we evaluated the binding affinities of soluble forms of ALV SU glycoproteins, SU(B)-rIgG, SU(D)-rIgG, and SU(E)-rIgG, for Tvb receptors expressed on the surfaces of CEFs of defined origin. The chimeric immunoadhesins SU(B)-rIgG, SU(D)-rIgG, and SU(E)-rIgG, which are composed of the subgroup B-, D-, or E-specific SU subunit linked to the constant fragment (Fc) of rabbit IgG, were constructed and transfected into DF-1 cells. The integrity of the three soluble SU proteins was verified by Western immunoblot analysis (Fig. 5A). The concentration of each soluble SU protein stock was quantitated by enzyme-linked immunosorbent assay (ELISA) for rabbit IgG. Binding of SU(B)-rIgG, SU(D)-rIgG, or SU(E)-rIgG to Tvb receptors expressed on the surfaces of tvb^s1/s1, tvb^s1/r4, tvb^r4/r4, tvb^s1/r5, and tvb^r5/r5 CEFs was detected with a fluorescein-conjugated goat anti-rabbit antibody and measured by fluorescence-activated cell sorting (FACS). The percentage of fluorescein-positive cells reflects viral glycoprotein-receptor binding affinity. As expected, SU(B)-rIgG, SU(D)-rIgG, or SU(E)-rIgG bound the wild-type Tvb^S1 receptor expressed on the surfaces of tvb^s1/s1 CEFs efficiently, with 19.8%, 18.4%, and 18.3% of tvb^s1/s1 CEFs being bound at 2 h postinoculation, respectively, while all three soluble SU proteins bound Tvb receptors expressed on the surfaces of tvb^s1/r4 and tvb^s1/r5 CEFs with affinities almost similar to those for the Tvb^S1 receptor (Fig. 5B and C). Binding of all three soluble SU proteins to the Tvb^R4 or Tvb^R5 receptor expressed on the surfaces of tvb^r4/r4 or tvb^r5/r5 CEFs was detectable, but they bound with significantly lower affinities than to the Tvb^S1 receptor (P < 0.01) (Fig. 5B). The fluorescein-negative and fluorescein-positive cells are obviously disparate, as shown by the presence of two independent peaks in the FACS histograms (Fig. 5C). These findings clearly demonstrated that the frameshift mutations of the tvb gene caused a dramatic loss of binding affinity between the mutant Tvb^R4 and Tvb^R5 receptors and the ALV-B, ALV-D, and ALV-E glycoproteins.

FIG 5.

Binding affinities of Tvb receptors for ALV-B, ALV-D, and ALV-E envelope glycoproteins. (A) Western immunoblot analysis of the soluble forms of the SU glycoproteins SU(B)-rIgG (SUB), SU(D)-rIgG (SUD), and SU(E)-rIgG (SUE) immunoprecipitated with anti-rabbit IgG-agarose beads. The precipitated proteins were denatured, separated by SDS-PAGE, and transferred to nitrocellulose. The filters were probed with horseradish peroxidase-conjugated goat anti-rIgG, and the bound protein-antibody complexes were visualized by chemiluminescence using X-ray film. Molecular masses (in kilodaltons) are given on the left. (B) tvb^s1/s1, tvb^s1/r4, tvb^r4/r4, tvb^s1/r5, and tvb^r5/r5 CEFs were incubated with the same amount of each soluble form of SU-rIgG. The receptor-viral glycoprotein complexes were bound to goat anti-rabbit IgG linked to Alexa Fluor. The amount of Alexa Fluor bound to the cells was determined by FACS. Binding affinities of the receptor-viral glycoproteins are given as the percentage of Alexa Fluor-positive cells. The values shown are averages and standard errors of the means from three experiments. To determine the statistical significance of differences between binding of wt and the mutant Tvb receptors for all three soluble SU glycoprotein data sets, unpaired two-tailed t tests were performed (**, P < 0.01). (C) Examples of FACS histograms showing the percentages of Alexa Fluor-positive cells after incubation of tvb^s1/s1, tvb^r4/r4, and tvb^r5/r5 CEFs with the same amount of each soluble form of the SU-rIgG and empty control.

Frameshift mutations significantly reduce susceptibility to ALV-B, ALV-D, and ALV-E infection in vitro.

The RCASBP(B)-GFP, RCASBP(D)-GFP, and RCASBP(E)-GFP reporter viruses, transducing the green fluorescent protein (GFP) reporter gene, offered an ideal molecular tool to determine, in a highly sensitive manner, the effects of the frameshift mutations on the tvb receptor's susceptibility to ALVs (19). We thus constructed the GFP-transducing reporter vectors of subgroups B, D, and E based on the RCAS retrovirus vectors (33) and obtained the replication-competent viruses RCASBP(B)-GFP, RCASBP(D)-GFP, and RCASBP(E)-GFP, as previously described (18, 19). The tvb^s1/s1, tvb^s1/r4, tvb^r4/r4, tvb^s1/r5, and tvb^r5/r5 CEFs were infected with RCASBP(B)-GFP, RCASBP(D)-GFP, or RCASBP(E)-GFP, and virus spread was followed as the percentage of GFP-positive cells quantified by FACS subsequently for 7 days. The wild-type tvb^s1/s1 CEFs were used as a positive control since they are susceptible to subgroup B, D, and E ALV. As expected, RCASBP(B)-GFP, RCASBP(D)-GFP, and RCASBP(E)-GFP successfully infected the tvb^s1/s1 CEFs, with an initial infection rate of about 20% on day 1, and approached virtually complete infection of cells by day 7 (Fig. 6A). However, RCASBP(B)-GFP, RCASBP(D)-GFP, and RCASBP(E)-GFP infected the tvb^r4/r4 CEFs with much less efficiency, with about 1%, 5%, and 7% of cells infected on day 1, and spread through the cells very slowly, reaching 18%, 19%, and 20% of infected cells, respectively, on day 4 and 25%, 32%, and 34% on day 7 (Fig. 6A). Interestingly, the time course of infection in tvb^r5/r5 CEFs was very similar to that in tvb^r4/r4 CEFs with ALV reporter viruses. RCASBP(B)-GFP, RCASBP(D)-GFP, and RCASBP(E)-GFP infected only 2%, 6%, and 7% of tvb^r5/r5 CEFs on day 1 and then spread slowly, with infection remaining between 20% and 35% on days 4 to 7 (Fig. 6A). The GFP-negative and GFP-positive cells are obviously disparate, as shown by the presence of two independent peaks in the FACS histogram (Fig. 6B). In contrast to the tvb^r4/r4 and tvb^r5/r5 CEFs, the tvb^s1/r4 and tvb^s1/r5 CEFs were much more highly susceptible to infection by RCASBP(B)-GFP, RCASBP(D)-GFP, and RCASBP(E)-GFP and exhibited infection properties similar to those of the wild-type counterpart (Fig. 6A).

FIG 6.

Time course of infection of tvb^s1/s1, tvb^s1/r4, tvb^r4/r4, tvb^s1/r5, and tvb^r5/r5 CEFs with ALV reporter viruses. CEFs of defined origin were infected at an MOI of 10 with replication-competent ALVs encoding the GFP reporter proteins RCASBP(B)-GFP, RCASBP(D)-GFP, and RCASBP(E)-GFP. (A) The proportion of GFP-positive cells was determined by FACS on the indicated days postinfection, and the percentage of GFP-positive cells is indicated as a mean and standard error of mean from three parallel dishes. (B) Representative FACS histograms of CEFs of defined origin infected with RCASBP(B)-GFP, RCASBP(D)-GFP, and RCASBP(E)-GFP at 7 days postinfection. The relative GFP fluorescence is plotted against the cell count, and the percentage of GFP-positive cells is indicated.

To further confirm these results, we also infected the CEFs of defined origin with ALV-B and determined its growth kinetics by RT-PCR over a period of 5 days. Consistent with the infection by the reporter viruses, ALV-B strain SDAU09C2 hardly replicated and grew slowly in tvb^r4/r4 and tvb^r5/r5 CEFs, and it reached only low titers of 10^6.5 and 10^6.8 virus copies/ml, respectively, at day 5 postinfection (Fig. 7). In contrast, titers of 10⁸ and 10⁹ virus copies/ml were reached in tvb^s1/r4 and tvb^s1/r5 CEFs, respectively, at day 3 postinfection (Fig. 7). In tvb^s1/s1 CEFs, the ALV-B strain SDAU09C2 grew much rapidly and reached titers of about 10¹⁰ virus copies/ml by day 3 postinfection (Fig. 7).

FIG 7.

Growth curves of ALV-B wild-type virus in different CEFs. Growth curves for ALV-B strain SDAU09C2 were obtained by infection of tvb^s1/s1, tvb^s1/r4, tvb^r4/r4, tvb^s1/r5, and tvb^r5/r5 CEFs at an MOI of 1. Cell culture supernatant-associated viruses were harvested at 1, 2, 3, 4, and 5 days postinfection, and titration results were determined as virus copies using qPCR. Means ± SEM of data from three independent experiments performed in triplicates are shown.

Taken together, these results clearly demonstrated that both of the frameshift mutations in the tvb gene resulted in reduced susceptibility to and slow spread of subgroup B, D, and E ALVs in vitro.

Frameshift mutations significantly reduce susceptibility to ALV-B infection in vivo.

To explore whether the in vitro findings hold true in vivo, we conducted a challenge trial with ALV-B in randomly sampled commercial chicks from lines CB03 and CB08. One-day-old chicks were inoculated with 0.2 ml of ALV-B strain SDAU09C2 (ratio of the sample to the positive control for ALV p27 antigen [S/P value], 2.0) through the abdominal cavity and inoculated once again at 5 days of age. At 1 month postinfection, the ALV-B infection status of each chick was determined by RT-PCR of RNAs extracted from a whole blood sample from each of the birds. As positive controls, 12 specific-pathogen-free (SPF) chicks were inoculated with ALV-B strain SDAU09C2. All of the inoculated SPF chicks were positive for ALV-B based on the RT-PCR test (Table 1), which suggested that the challenge trial assay was successful. In a cohort of 27 chicks with the genotype tvb^s1/s1, all were positive for ALV-B (Table 1). However, the susceptibility of 19 tvb^r4/r4 and 25 tvb^r5/r5 chicks to ALV-B was decreased to 5.3% and 12%, respectively; the susceptibility of the tvb^s1/r4 and tvb^s1/r5 chicks was about 50% (Table 1). Due to lack of stocks of ALV-D and ALV-E in our laboratory, no corresponding in vivo trial was conducted. Since the tvb receptor gene determines the susceptibility to subgroup B, D, and E ALVs, we assumed that both the tvb^r4 and tvb^r5 alleles may confer a decrease in host susceptibility to ALV-D and ALV-E. These results demonstrated that the in vivo infection coincides with spread of the viruses in vitro and confirmed the reduced susceptibility of tvb^r4/r4 and tvb^r5/r5 birds to subgroup B, D, and E ALVs.

TABLE 1.

Incidence of infection by subgroup B ALV

Chicks	ALV	No. of positive samples/total no. of samples	Positive infection (%)
Line CB03	tvbs1/s1	19/19	100
	tvbs1/r4	6/12	50
	tvbr4/r4	1/19	5.3
Line CB08	tvbs1/s1	8/8	100
	tvbs1/r5	10/17	58.8
	tvbr5/r5	3/25	12
SPF	tvbs1/s1	12/12	100

DISCUSSION

We describe here the identification of two molecular defects in the tvb receptor gene of Chinese broiler lines, the c.291_292insAG and c.359_360insA insertion mutations in the tvb gene, which significantly reduce the susceptibility of these birds to infection by ALV-B, ALV-D, and ALV-E. Both of the insertions caused a frameshift resulting in a premature stop codon, which led to the production of truncated Tvb^R4 and Tvb^R5 proteins, respectively. In turn, Tvb^R4 and Tvb^R5 significantly decrease the binding affinity to ALV glycoproteins and account for the reduced infection efficiency of ALV-B, ALV-D, and ALV-E. To our knowledge, this study is the first to report that frameshift mutations in the receptor-encoding loci can decrease susceptibility to infection by specific ALV subgroups.

Nonsense-mediated mRNA decay (NMD) represents a molecular quality control process believed to degrade mRNAs that harbor a premature translation termination codon (PTC) in eukaryotes (34). NMD generally requires that PTCs be located upstream of at least one exon-exon junction, specifically, at least 50 to 55 bases upstream of the last junction (35, 36). In this way, the premature termination codons in the tvb^r4 and tvb^r5 transcripts positioned 71 nucleotides upstream of the exon 5-exon 6 junction and 64 nucleotides upstream of the exon 8-exon 9 junction, respectively, may be proper targets. Therefore, we speculate that mutant tvb mRNAs arising from frameshift mutations may also be targeted by NMD. Some mutations result in reduced, but not absent, levels of mutant RNA that are sufficient to be determined by either real-time or allele-specific PCR (37, 38). The quantification of gene expression results showed minimal mutant tvb expression compared to the normal expression (Fig. 2). This aspect is consistent with the NMD function as an mRNA surveillance mechanism to remove aberrant mRNAs. NMD can potentially prevent the production of C-terminally truncated proteins from PTC-containing mRNAs (31, 34). However, this is not always the case, as some alleles encoding PTCs produce only very low mRNA levels (39). Indeed, it is unlikely to apply in the case of the tvb^r4 and tvb^r5 alleles, since the truncated Tvb^R4 and Tvb^R5 protein products consisted of the 130 N-terminal amino acid residues and 189 N-terminal amino acid residues, respectively (Fig. 3A and C). This is likely due to the stabilization of the tvb mRNA or to proteasome activity. Many different human diseases are associated with mutations that generate a PTC-containing mRNA (40). Intriguingly, mutations causing PTCs in genes belonging to the TRAIL family, such as TRAILR1 and TRAILR2, were found to cause some inherited disorders and cancers (41, 42). It is attractive to presume that this is a common mechanism of mutagenesis in this gene family, since the Tvb receptor is most closely related to the human TRAIL receptors, TRAIL-R1 and TRAIL-R2 (8, 10, 23).

Several regions and specific residues in the Tvb receptor have previously been identified to be crucial determinants of the binding affinity with ALV-B, ALV-D, and ALV-E glycoproteins and to mediate efficient virus entry. A linear 15-amino-acid peptide representing amino acid residues 32 to 46 derived from CRD1 can serve as a minimal soluble Tvb receptor and mediate sufficiently for ALV-B and ALV-D binding and entry (24). Additionally, four residues (Leu-36, Gln-37, Leu-41, and Tyr-42) in the Tvb_32–46 domain are critical for peptide function. In sharp contrast to the case for ALV-B and ALV-D, all of the three CRDs of Tvb are required for efficient ALV-E binding and entry (19, 25, 43). Residues Tyr-67, Asn-72, and Asp-73 in CRD2 were identified as important ALV-E interaction determinants (25). Furthermore, the structural integrity of the protein, including the putative C-46–C-59 disulfide bond in CRD1, the C-62–C-77 disulfide bond in CRD2, and the C-125–C-143 disulfide bond in CRD3, is specifically required for ALV-E receptor function (10, 19, 43). A previous study has reported that there are two distinct types of Tvb^S1 receptor, which resulted from alternate patterns of protein folding and disulfide bonds located in CRD1 and CRD2 (43). The type 1 receptor is for ALV-B, ALV-D, and ALV-E, whereas the type 2 receptor is only for ALV-B and ALV-D. The tvb^s3 allele encodes the Tvb^S3 protein, which contains the C62S substitution that presumably alters the receptor conformation, leading to the loss of ALV-E binding and entry but having no detectable effect on binding and entry of ALV-B and ALV-D (10). CRD3 was originally thought to be dispensable for Tvb receptor function. Indeed, published studies have shown that a Tvb receptor with the deletion of CRD3 appear to function as efficiently as a wild-type Tvb^S1 receptor (24, 25, 43). However, a previous study has demonstrated a possible role of CRD3 in the efficient functioning of the Tvb receptor (19). The Tvb^R2 protein with the C125S substitution in CRD3 likely alters the folding and final structure of CRD3 and may further result in structural alterations in other CRD regions of the protein, such as CRD1 and CRD2, significantly reducing the binding to the ALV-B and ALV-D glycoproteins and virus infection efficiency and virtually eliminating ALV-E binding and infection (19). These pieces of evidence highlight the importance of the proper structure of the CRDs in the Tvb protein: single and combination cysteine mutations in CRD1 to -3 likely alter the folding and final structure of the CRDs, thereby changing the structure of the receptor protein in a way that blocks ALV-B, ALV-D, and ALV-E interaction and subsequent virus entry.

Although the underlying reason is not clear at present, several plausible explanations could account for the reduced ALV-B, ALV-D, and ALV-E susceptibility phenotype which arises from the frameshift mutations in the tvb receptor gene. Structural studies of the Tvb protein have shown that the conserved cysteine residues in the CRDs are critical to form intradomain disulfide bonds to stabilize the structure of the receptor protein (10, 24, 43). Therefore, the C101S substitution in CRD2 presumably alters the folding and final structure of CRD2 in the Tvb^R4 protein, since the C-101 residue in Tvb is predicted to exist as a disulfide bond with C-83 in CRD2 (Fig. 3B). Moreover, the C103L, C109S, C115P, C121S, and C125S substitutions likely hamper the folding and final structure of CRD3 in the Tvb^R4 protein by abolishing the formation of the correct intrachain disulfide bonding patterns. Furthermore, and unexpectedly, the C125S substitution in the Tvb^R2 protein was also in the Tvb^R4 receptor protein (Fig. 3B). In contrast to the phenotype of the Tvb^R2 receptor, the binding and entry of ALV-B, ALV-D, and ALV-E are affected in the Tvb^R4 receptor. However, we do not know the impact of replacements of other amino acid residues in CRD3, especially of several aromatic residues, which are critically required for the retrovirus-receptor interaction (43–46), on the structure and function of Tvb^R4. Therefore, if these cysteine substitutions alter the structures of CRD2 and CRD3, the structure of CRD3 in Tvb^R4 would likely be different from that in Tvb^R2. One possible explanation that reconciles our findings is that CRD2 and CRD3 of the Tvb^R4 proteins are misfolded, but a fraction of the molecules adopt a native conformation suitable for ALV-E binding and entry, albeit at a much lower efficiency. Another possible explanation is that the altered CRD2 and CRD3 structures may interfere with and/or alter the structure of CRD1 and might inhibit the interaction of ALV-B and ALV-D with CRD1, resulting in significantly reducing, but not eliminating, the binding and entry of ALV-B, ALV-D, and ALV-E. For Tvb^R5, the C121V, C125V, C135V, and C143V substitutions likely alter the folding and final structure of at least CRD3 in the Tvb^R5 protein, since the putative C-125–C-143 disulfide bond and C-121–C-135 disulfide bond were misfolded (Fig. 3B). However, the pivotal C-125 in CRD3 was substituted for V-125 in the Tvb^R5 protein. Moreover, given the existence of replacements of other amino acid residues in CRD3, it seems obvious to propose that the Tvb^R5 protein may exert the same effects as the Tvb^R4 protein. Therefore, in contrast with the phenotype of the Tvb^R2 receptor, the binding and entry of ALV-B, ALV-D, and ALV-E are affected in the Tvb^R5 receptor, albeit at a much lower efficiency.

Alternatively, despite these defects, the Tvb^R4 and Tvb^R5 proteins can still function as a receptor for ALV-B, ALV-D, and ALV-E, albeit at a much lower efficiency. Indeed, the tva^r allele contains Cys40Trp in the LDLR-related region, which is the functional ALV-A receptor interaction domain, resulting in drastically reduced ALV-A binding and entry. However, the ectopic expression of the Tva^R receptor at high levels in resistant cells increased the avidity of the interaction with the ALV-A glycoproteins and conferred susceptibility to infection by ALV-A at similar levels to those for wild-type Tva (16). Therefore, we speculate that when mutant Tvb proteins, such as Tvb^R4 and Tvb^R5, are expressed at wild-type levels, they may still function but at a lower efficiency.

In studies of naturally occurring mutations in the Tva, Tvb, and Tvc receptor alleles that confer resistance to ALV (subgroup A to E) infection, our initial findings were in favor of the notion that frameshift mutation which introduces a premature termination codon and leads to production of a truncated receptor protein may be a new molecular mechanism that reduces susceptibility to infection by specific ALV subgroups. The perfect analogy of tvb^r4 and tvb^r5 alleles, the tva^r2 allele contains 4-nucleotide insertion in exon 1 of the tva gene, which results in a change in reading frame and leads to eliminate the expression of the Tva receptor, explaining the complete resistance to ALV-A infection (16). Furthermore, in contrast to the tvb^r4 and tvb^r5 alleles, the tvb^r allele contains a premature stop codon at residue 57, explaining the complete resistance to infection by ALV-B, ALV-D, and ALV-E (15). Considering the features of tvb^r, it is predicted to result in the generation of a severely truncated protein. In fact, the premature stop codon leads to a lack of Tvb expression, and tvb^r is a null allele of the tvb gene (15). As a parallel to the tvb^r allele, the tvc^r allele also contains a premature stop codon at residue 55, which is predicted to produce a severely truncated Tvc receptor, including only the signal peptide plus 32 amino acid residues of the mature Tvc protein, and abolishes its use as an ALV-C receptor (11). However, none of the experimental evidence indicated the production of a truncated Tvc protein. Even though produced, the Tvc^R protein lacks Trp-48, Tyr-105, and a intact IgV domain of Tvc, which were identified as critical for efficient ALV-C infection (46), resulting in the loss of ALV-C binding and entry.

In summary, the present study identified two new tvb resistance alleles causing reduced susceptibility to infection by ALV-B, ALV-D, and ALV-E due to frameshift mutations in the tvb gene of Chinese broiler lines. Our findings should help in understanding the evolutionary mechanisms that drive the development of host resistance to retroviral infection and provide valuable insight into antiviral intervention.

MATERIALS AND METHODS

Chicken lines.

The tvb^r4 and tvb^r5 alleles exist separately in the commercial broiler line 03 and 08 (CB03 and CB08) based on our recent study (30). Lines CB03 and CB08 have been maintained at Guangdong Wen's Food Group Co., Ltd., Guangdong, People's Republic of China. Fertilized eggs were incubated at 37.8°C and 55% relative humidity in an incubator with a tilting motion through a 90° angle every 2 h. All experiments were approved by the Animal Ethics Committee at the South China Agricultural University, People's Republic of China (approval ID SCAU0011).

Preparation of CEFs and cell culture.

The procedure for the preparation of chicken embryo fibroblasts (CEFs) from 10-day-old embryos from tvb^s1/s1, tvb^s1/r4, tvb^r4/r4, tvb^s1/r5, tvb^r5/r5, and SPF fertilized eggs has been described previously (47). CEFs, as well as the permanent chicken cell line DF-1 (48) and human 293FT cells, were grown in Dulbecco's modified Eagle's medium (DMEM) (Life Technologies) supplemented with 10% fetal bovine serum (FBS) (Life Technologies) and penicillin-streptomycin (Life Technologies). All cells were cultured at 37°C in a 5% CO₂ incubator.

DNA extraction, RNA isolation, and cDNA synthesis.

Genomic DNA was extracted from blood samples collected from CB03 and CB08 chickens, as well as chicks infected by ALV-B strain SDAU09C2 (kindly provided by Zhizhong Cui, Shandong Agricultural University, People's Republic of China), using a phenol-chloroform protocol. DNAs from the RCASBP(A)-EGFP, RCASBP(D), and RCASBP(E) plasmids were extracted using DNA extraction kits for plasmids (Bio-Rad Laboratories, Inc.). The proviral DNA of the isolate SDAU09C2 was extracted as described previously (29). Total RNA was isolated from tvb^s1/s1, tvb^s1/r4, tvb^r4/r4, tvb^s1/r5, and tvb^r5/r5 CEFs and CEFs infected with ALV-B strain SDAU09C2, as well as from blood samples of rabbits and chicks that were infected with ALV-B strain SDAU09C2, using TRIzol reagent (Life Technologies). cDNA was obtained by reverse transcription of 1 μg of DNA-free RNA using the PrimeScript RT reagent kit (TaKaRa) following the manufacturer's protocol. The prepared DNA and synthesized cDNA were stored at −20°C for further experiments.

Primers.

PCR primers that were specific for the whole tvb genomic region sequence clone and the gp85 (SU) coding region sequence clone of subgroup B, D, and E ALV were designed using Premier Primer 5.0 software. RT-PCR primers for the full-length tvb coding sequence clone, a portion of the tvb cDNA sequence clone encompassing the c.291_292insAG and c.359_360insA insertions, and the sequence clone of IgG heavy chain of rabbit were also designed using Premier Primer 5.0 software. RT-qPCR primers that were specific for the 5′ and 3′ ends of the tvb gene and chicken glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were also designed using the Premier Primer 5.0 software. All of the primers were synthesized by Sangon Biotech Co., Ltd. (Guangzhou, China), and are listed in Table S1 in the supplemental material.

Amplification and analysis of tvb alleles from lines CB03 and CB08.

The whole genomic region of the tvb gene was amplified from genomic DNAs of line CB03 and CB08 chickens using four specific primer pairs. Amplicons were directly sequenced by Sangon Biotech Co., Ltd. (Guangzhou, China). Multiple sequence traces of the tvb genes from line CB03 and CB08 birds were aligned and compared with sequence of the wild-type tvb^s1 allele using Lasergene version 7.1 (DNAStar, Inc., Madison, WI). The entire tvb coding sequence was amplified from the cDNAs of tvb^s1/s1, tvb^r4/r4, and tvb^r5/r5 CEFs using specific primers that can amplify the complete transcript. The purified RT-PCR products were cloned into the TA vector pMD19-T (TaKaRa) and then sequenced and analyzed as described above.

Allelic imbalance and RT-qPCR of nonsense-mediated mRNA decay (NMD).

A portion of tvb cDNA encompassing the c.291_292insAG and c.359_360insA insertions was amplified from cDNAs of tvb^s1/r4 and tvb^s1/r5 CEFs using standard procedures. The final RT-PCR products were ligated into the pMD19-T vector (TaKaRa) and sequenced as described above. The mRNA expression levels associated with the 5′ and 3′ ends of the tvb gene in the tvb^s1/s1, tvb^s1/r4, tvb^r4/r4, tvb^s1/r5, and tvb^r5/r5 CEFs were detected by RT-qPCR. The assays were performed using an Applied Biosystems 7500 Fast real-time PCR system (Applied Biosystems, Foster City, CA) with FastStart SYBR green Master Mix (Rox) (Roche) according to the manufacturer's instructions. The chicken GAPDH mRNA was used as an internal control. The specificity of the PCR products was verified by sequencing the PCR products and by melting curve analysis. All sample/gene combinations were analyzed in triplicate. The results were analyzed using the 2^−ΔΔCT method as previously described (49).

Construction of Tvb vectors and expression of Tvb receptors.

The full-length coding sequence of tvb was amplified from the cDNA of tvb^s1/s1, tvb^r4/r4, or tvb^r5/r5 CEFs as described above. The resulting PCR products were cloned into the pMD19-T vector (TaKaRa), and the entire tvb open reading frame was cut out using XhoI and HindIII (New England BioLabs) and ligated into XhoI-HindIII-linearized pEGFPC1 vector (Clontech) with the proper orientation. The resulting expression constructs were designated pEGFPC1-tvb^s1, pEGFPC1-tvb^r4, and pEGFPC1-tvb^r5, respectively. The Tvb expression vectors were verified by sequencing and restriction enzyme digestion analysis.

The expression constructs with wt and mutated tvb, as well as the pEGFPC1 vector, were transfected into 293FT cells. A total of 2 × 10⁵ cells were seeded onto a 6-well plate in the cultivation medium without antibiotics. Transfection was performed 24 h later with a mixture of 4 μl of Lipofectamine 3000 reagent (Life Technologies) and 2 μg of pEGFPC1-tvb^s1, pEGFPC1-tvb^r4, or pEGFPC1-tvb^r5 plasmid DNA, or pEGFPC1 and empty vectors, in Opti-MEM (Gibco BRL) per well. The rabbit anti-GFP polyclonal antibody (1:2,000; Sigma) was used for Western blot analysis of wt or mutated Tvb receptors and a GFP-positive control. Horseradish peroxidase (HRP)-conjugated goat anti-rabbit IgG (H+L) antibody (1:10,000; Bioss Inc.) was used as the secondary antibody. The results were analyzed with Odyssey Application software version 3.0 (Li-Cor Biosciences).

Apoptosis assay.

The pEGFPC1 vector or expression constructs with wt or mutated tvb were transfected into 293FT cells by using Lipofectamine 3000 reagent (Life Technologies) as described above. At 24 h after transfection, cells were stained with an annexin V-fluorescein isothiocyanate (FITC)/propidium iodide (PI) kit (Beyotime), and the percentage of apoptotic cells was quantitated using a Cytomics FC 500 analyzer (Beckman Coulter, USA). The mean values and standard errors of the means (SEM) were calculated from three independent experiments.

Construction and production of fusion proteins SU(B)-rIgG, SU(D)-rIgG, and SU(E)-rIgG.

The gp85 (SU)-coding regions, including the signal peptide, of ALV-B, ALV-D, and ALV-E were amplified by PCR from the isolate SDAU09C2 provirus DNA and RCASBP(D) and RCASBP(E) plasmid DNAs, respectively. The IgG heavy chain of rabbit (rIgGFc) was obtained by RT-PCR from the RNA of rabbit blood cells. The SU(B), SU(D), and SU(E) fragments and the rIgGFc were ligated with SacI/KpnI (New England BioLabs) and KpnI/BamHI (New England BioLabs) restriction enzyme sites, and the fusion fragment was cloned into the pUC18 vector (Promega) for sequencing. SU(B)-rIgG, SU(D)-rIgG, and SU(E)-rIgG then were recovered as SacI and BamHI fragments, subcloned into the Cla12Nco adapter plasmid, and then cloned into the RCASBP(A) replication-competent vector as a ClaI fragment. As expected, the recombinant retroviral vectors RCASBP(A)-SU(B)-rIgG, RCASBP(A)-SU(D)-rIgG, and RCASBP(A)-SU(E)-rIgG containing a 1.8-kb fragment which fused the receptor-binding parts of the ALV-B, ALV-D, and ALV-E SU glycoproteins and the Fc region of rabbit IgG were obtained and were transfected into DF-1 cells. The cell supernatants that expressed fusion protein SU(B)-rIgG, SU(D)-rIgG, or SU(E)-rIgG immunoadhesins were cleared by centrifugation at 2,000 × g for 10 min at 4°C and stored in aliquots at −80°C. The levels of the SU(B)-rIgG, SU(D)-rIgG, or SU(E)-rIgG fusion proteins in culture supernatants were quantitated by enzyme-linked immunosorbent assay (ELISA) for the rabbit IgG tag as previously described (50).

Western immunoblot analysis.

The SU(B)-rIgG, SU(D)-rIgG, and SU(E)-rIgG immunoadhesin proteins were immunoprecipitated separately from DF-1 cell culture supernatants with anti-rabbit agarose beads (Sigma) and analyzed by Western immunoblotting as previously described (51). Horseradish peroxidase-conjugated goat anti-rIgG (1:10,000; Bioss Inc.) was used as the secondary antibody. The protein complexes were detected with Lightning Chemiluminescence Reagent Plus (PerkinElmer Life Sciences Inc.).

Analysis of binding affinity by flow cytometry.

The tvb^s1/s1, tvb^s1/r4, tvb^r4/r4, tvb^s1/r5, and tvb^r5/r5 CEFs expressing the respective receptor proteins were trypsinized, washed with phosphate-buffered saline (PBS) supplemented with 2% calf serum (PBS-CS), centrifuged for 5 min at 800 × g, and resuspended in 100 μl of PBS-CS. The cells (ca. 1 × 10⁶) in PBS-CS were incubated with supernatant containing the same concentration (ca. 10 ng of ImmunoPure rIgG Fc fragment per ml) of SU(B)-rIgG, SU(D)-rIgG, or SU(E)-rIgG fusion protein at room temperature for 1 h. After three washes with PBS-CS, goat anti-rabbit IgG linked to Alexa Fluor 488 (Beyotime) was diluted 1:100 in PBS supplemented with 4% calf serum, and washed cells were incubated in 500 μl of diluted antibody at 37°C for 1 h. After incubation and three washes in PBS-CS, the cell-SU-rIgG-Alexa Fluor 488 complexes were resuspended in 500 μl PBS-CS, and the percentage of fluorescence-positive cells was quantitated by FACS using a Cytomics FC 500 analyzer (Beckman Coulter, USA).

Construction of subgroup B, D, and E ALV reporter vectors and virus propagation.

We previously have constructed an RCASBP(A)-EGFP subgroup A retrovirus vector transducing the enhanced green fluorescent protein (EGFP) reporter gene (18), which was stored at our laboratory. The construction of subgroup B, D, and E reporter vectors was the same as that of the subgroup A reporter vector. In brief, the EGFP gene was isolated from the Cla12Nco-EGFP adapter plasmid as a ClaI (New England BioLabs) fragment and cloned into the ClaI sites of the RCASBP(B), RCASBP(D), and RCASBP(E) retroviral vectors (33), which were kindly provided by S. H. Hughes (HIV Drug Resistance Program, National Cancer Institute, USA). The resulting constructs were named RCASBP(B)-EGFP, RCASBP(D)-EGFP, and RCASBP(E)-EGFP, respectively.

Infectious viruses encoding GFP were produced in DF-1 cells chronically transfected with RCASBP(B)-EGFP or RCASBP(D)-EGFP plasmid DNA and CEFs prepared from SPF chickens chronically transfected with RCASBP(E)-EGFP plasmid DNA. Infection and virus spread were observed as an increasing proportion of GFP-positive cells, and virus stocks were harvested on day 7 posttransfection. The cell supernatants were cleared of debris by centrifugation at 2,000 × g for 10 min at 4°C and stored in aliquots at −80°C. The virus titer, determined by terminal dilution and subsequent infection of fresh DF-1 cells, reached 10⁶ IU/ml.

The ALV-B strain SDAU09C2 was propagated in DF-1 cells. The value of the ratio of the sample to the positive control (S/P) for ALV p27 antigen of the ALV-B strain SDAU09C2 was determined by using the avian leukosis virus antigen test kit (Idexx) according to the manufacturer's instructions.

Susceptibility assay for CB03 and CB08 chicken embryo fibroblasts.

Susceptibility to ALV-B, ALV-D, or ALV-E were assessed by RCASBP(B)-GFP, RCASBP(D)-GFP, or RCASBP(E)-GFP virus spread as described in our previous study (18). Briefly, CEFs of the tvb^s1/s1, tvb^s1/r4, tvb^r4/r4, tvb^s1/r5, and tvb^r5/r5 genotypes were seeded in triplicate wells at a density of 5 × 10⁴ per well in a 24-well plate and were infected with either RCASBP(B)-GFP, RCASBP(D)-GFP, or RCASBP(E)-GFP virus at a multiplicity of infection (MOI) of 10 on the day after seeding. The percentage of GFP-positive cells was quantitated by fluorescence-activated cell sorting (FACS) using a Cytomics FC 500 analyzer (Beckman Coulter, USA) on days 1, 2, 4, and 7 postinfection. The cells of three wells were trypsinized and washed in phosphate-buffered saline (PBS) before the analysis.

We also assessed the susceptibility to ALV-B wild virus by virus growth kinetics and virus titration assays. The tvb^s1/s1, tvb^s1/r4, tvb^r4/r4, tvb^s1/r5, and tvb^r5/r5 CEFs were seeded in triplicate wells at a density of 2 × 10⁵ per well in a 6-well plate. At 24 h after seeding, CEFs of defined origin were infected with ALV-B strain SDAU09C2 (S/P value, 2.0) at an MOI of 1. After 2 h, the medium was replaced by fresh medium and incubated at 37°C in a 5% CO₂ atmosphere for 5 days. The infected cell cultures were harvested at days 1, 2, 3, 4, and 5 postinfection and were subjected to RNA extraction and cDNA synthesis. The titer of infectious progeny was determined as the virus copies per milliliter by RT-qPCR using an ALV-B-specific primer pair as previously described (52). The GAPDH gene was used as an internal control. The mean values and standard errors of the means were calculated from three independent experiments.

In vivo infections.

One-day-old chicks were randomly collected from lines CB03 and CB08 with 50 birds for each line and randomly divided into four groups, along with 25 commercial chicks and 3 SPF chicks that were used for controls in each group. Chicks were housed in four negative-pressure-filtered air isolators under quarantine conditions, and feed and water were provided ad libitum. Chicks were infected by abdominal cavity inoculation with ca. 0.2 ml of ALV-B strain SDAU09C2 (S/P value, 2.0) at 1 day old and infected once again at 5 days old. Genomic DNA from each chick at 7 days old was used to genotype the tvb^r4 and tvb^r5 alleles by direct sequencing. At 1 month postinfection, the status of infection by ALV-B strain SDAU09C2 in each chick was determined as previously described (53).

Statistical analysis.

Statistical comparisons were performed using GraphPad Prism (version 6.0) software (GraphPad Software Inc.). Results are presented as means ± SEM, and statistical significance was assessed at P values of <0.05 and <0.01.