Journal of General Virology (2014), 95, 2060–2070

DOI 10.1099/vir.0.066852-0

Prototype endogenous avian retroviruses of the genus Gallus Melanie Ann Sacco1 and Venugopal K. Nair2 Correspondence Melanie Ann Sacco [email protected]

Received 13 April 2014 Accepted 2 June 2014

1

Center for Applied Biotechnology Studies, Department of Biological Science, California State University Fullerton, Fullerton, CA 92834-6850, USA

2

Pirbright Institute, Compton Laboratory, Newbury, Berkshire RG20 7NN, UK

Ancient endogenous retroviruses (ERVs), designated endogenous avian retrovirus (EAVs), are present in all Gallus spp. including the chicken, and resemble the modern avian sarcoma and leukosis viruses (ASLVs). The EAVs comprise several distinct retroviruses, including EAV-0, EAV-E51 and EAV-HP, as well as a putative member previously named the avian retrotransposon of chickens (ART-CH). Thus far, only the EAV-HP elements have been well characterized. Here, we determined sequences of representative EAV-0 and EAV-E51 proviruses by cloning and data mining of the 2011 assembly of the Gallus gallus genome. Although the EAV-0 elements are primarily deleted in the env region, we identified two complete EAV-0 env genes within the G. gallus genome and prototype elements sharing identity with an EAV-E51-related clone previously designated EAV-E33. Prototype EAV-0, EAV-E51 and EAV-E33 gag, pol and env gene sequences used for phylogenetic analysis of deduced proteins showed that the EAVs formed three distinct clades, with EAV-0 sharing the last common ancestor with the ASLVs. The EAV-E51 clade showed the greatest level of divergence compared with other EAVs or ASLVs, suggesting that these ERVs represented exogenous retroviruses that evolved and integrated into the germline over a long period of time. Moreover, the degree of divergence between the chicken and red jungle fowl EAV-E51 sequences suggested that they were more ancient than the other EAVs and may have diverged through mutations that accumulated post-integration. Finally, we showed that the ART-CH elements were chimeric defective ERVs comprising portions of EAV-E51 and EAV-HP rather than authentic retrotransposons.

INTRODUCTION Endogenous retroviruses (ERVs) that exist as stable genetic elements have been shown to be distributed widely in all vertebrate species examined, including the class Aves (Be´nit et al., 1999; Gifford et al., 2005; Herniou et al., 1998; Martin et al., 1999). The genome structures of the avian ERVs determined to date primarily resemble avian exogenous Ctype retroviruses of the avian sarcoma and leukosis viruses (ASLVs) in the genus Alpharetrovirus (Boyce-Jacino et al., 1992; Coffin et al., 1983; Dunwiddie et al., 1986; Hughes et al., 1981; Ruis et al., 1999; Sacco et al., 2000). The first chicken ERVs identified were the subgroup E avian leukosis viruses (ALV-E), previously termed ev loci, which exist at various segregating loci in the chicken (Gallus gallus subsp. domesticus) genome as defective elements or intact infectious proviruses (reviewed by Crittenden, 1991). These The GenBank/EMBL/DDBJ accession numbers for the sequences pEAV5, pEAV3, pEAV21, pEAV27 and pEAV-E51 are AM418554, AM418555, AM418557, AM4185560 and AM418553, respectively. Five supplementary figures are available with the online version of this paper.

2060

appear to be the most recently integrated of the ERVs based on their restricted distribution in the species G. gallus (Frisby et al., 1979), which includes the chicken and its undomesticated progenitor, the red jungle fowl (G. gallus subsp. gallus), for which the genome sequence has been obtained (Hillier et al., 2004). Other chicken ERVs, designated endogenous avian retroviruses (EAVs), were described subsequently (Boyce-Jacino et al., 1989; Dunwiddie & Faras, 1985; Dunwiddie et al., 1986). EAVs appear to be more ancient than the ALV-E elements, as suggested by their distribution in all species of the genus Gallus (Boyce-Jacino et al., 1989; Resnick et al., 1990; Sacco et al., 2000) and the presence of shared loci between different Gallus spp., because germline integrations are rare and identical proviral insertions likely indicate integration occurred before divergence of the two species (Resnick et al., 1990; Sacco et al., 2001). The EAVs comprise several distinct retrovirus elements designated EAV-0, EAV-E51 and EAV-HP. EAV-0 was first identified following hybridization with Rous sarcoma virus (RSV) probes to DNA from line 0 white leghorn chickens bred to be free of ALV-E loci (Dunwiddie & Faras, 1985). 066852 G 2014 The Authors

Printed in Great Britain

Endogenous avian retroviruses

The EAV-E51 sequences were identified subsequently in line 0 genomic DNA (Boyce-Jacino et al., 1992). Only partial provirus elements were sequenced in initial published reports; however, both EAVs appeared to be defective, primarily as a consequence of deletions in their genomes of various sizes (Boyce-Jacino et al., 1992). The EAV-HP proviruses (also designated ev/J) have been examined in the most detail due to their similarity with and likely contribution of env sequences to the exogenous ALV subgroup J (ALV-J) by recombination (Bai et al., 1995; Ruis et al., 1999). Whilst most of the EAV-HP provirus sequences have large deletions, including the entire pol region, intact and expressed proviral sequences have been identified that segregate within the chicken population (Sacco et al., 2004). A final member ascribed to the EAV family is an endogenous retroelement named the avian retrotransposon of chickens (ART-CH) due to the absence of env sequences (Gudkov et al., 1992). ART-CH was found subsequently to share a high degree of identity with EAVHP and EAV-E51, suggesting it is a defective EAV chimeric retrovirus with the env sequences lost, rather than a bona fide retrotransposon; however, a portion of the ART-CH genome did not match any known ERV sequences (Sacco et al., 2000). Further interest in avian ERVs also developed from the concerns for human health arising from detection of reverse transcriptase (RT) activity in ALV-free chicken embryos or embryo fibroblast cells used as substrates for human vaccine production (Robertson et al., 1997; Weissmahr et al., 1997). Whilst the EAV-HP elements in line 0 chickens appear to be completely devoid of pol gene sequences (Sacco et al., 2004), the EAV-0 elements appear to have a complete pol region and the structures of EAV-E51 elements are unknown, suggesting that these elements may be the source of RT activity (Boyce-Jacino et al., 1992). The study presented here was aimed at obtaining full-length representative sequences to define the structures of the EAV-0 and EAV-E51 elements. We also aimed to determine the relationships among the different EAV family members and recombinant proviruses, such as ART-CH elements, and to ascertain the ability of these elements to produce infectious virus.

RESULTS Characterization of prototype EAV-0 elements and retrieval of novel env sequences We initiated a study to determine prototype sequences of the EAV-0- and EAV-E51-type ERVs from the chicken initially using molecular biology approaches, and continued this endeavour by capitalizing on the release of the G. gallus genome sequence (Hillier et al., 2004). PCR amplification of EAV-0 proviral DNA using long terminal repeat (LTR)specific primers produced two products of ~2.8 and 5.8 kbp (data not shown). The 5.8 kbp PCR product was the expected size based on previously described EAV-0 element transcripts amplified by PCR (Weissmahr et al., http://vir.sgmjournals.org

1997) or mapped by restriction enzyme digestion and hybridization (Resnick et al., 1990). A representative clone (pEAV5) was found to be 5811 bp in length, including the virus-derived sequences in the oligonucleotide primers with NotI restriction sites, and showed 98 % identity to the corresponding regions in previously reported partial sequences. The EAV-0 provirus had a large deletion of the env gene, as previously described for these ERVs, and was grossly intact for the gag and pol genes, although clone pEAV5 had a single-base-pair deletion resulting in a –1 frameshift within the protease gene that would render it defective in those genes. A BLAST search using the deduced amino acid sequence of the EAV-0 pol gene demonstrated 66 % identity and 80 % similarity to the translated ASLV pol genes, and 66 % identity and 79 % similarity to the translated EAV-HP pol gene. The less-conserved gag genes showed 45–47 % identity and 55–57 % similarity between the deduced amino acid sequences of the EAV-0 prototype and the ASLVs; the same region demonstrated only 37 % identity and 48 % similarity to EAV-HP. Using the full-length EAV-0 sequence, a search of the red jungle fowl genome database produced two hits with additional env coding sequences (Table 1), both from contigs that had not yet been mapped and for which only partial provirus sequences were available at the time of submission of this manuscript. These sequences were designated EAV-0Un1 and EAV-0Un2 (EAV-0 unmapped sequences 1 and 2, respectively). The sequences of EAV0Un1 and EAV-0Un2 spanned from the integrase-coding region of the pol gene corresponding to nt 5061 and 4652 of the pEAV5 clone, respectively. Alignment of the EAV-0 clone pEAV5 with EAV-0Un1 allowed the size of the EAV-0 env gene deletion in pEAV5 to be determined as 1189 bp. The deduced translated env sequences demonstrated an intact env gene in EAV-0Un1 and a premature stop codon in the transmembrane (TM) protein sequence of the partial EAV-0Un2. Comparison of this EAV-0 deduced amino acid sequence with other avian retroviral sequences demonstrated divergence within the envelope protein peptide leader sequence (Fig. S1, available in the online Supplementary Material). When the more-conserved envelope surface (SU) and TM glycoprotein-coding regions were examined, EAV-0 shared 43–44 % identity and 57 % similarity with the ALV-J and EAV-HP sequences, and had the closest similarity (46 % identity and 73 % similarity) with the previously described EAV-E51 endogenous element. Intact EAV-E51 elements exist at few segregating genomic loci PCR amplification using LTR-specific primers for EAVE51 was performed to obtain a full-length proviral clone as described for the EAV-0 prototype provirus above; however, repeated attempts amplified products of only ~3 kbp in size (data not shown). A clone of the PCR products (pEAV27) was sequenced and the small size (2959 bp) of the provirus was found to be mainly due to a large deletion of the pol 2061

M. A. Sacco and V. K. Nair

2062

Table 1. EAV endogenous retrovirus loci from the red jungle fowl genome used in this study EAV species and locus name EAV-0 EAV-0Chr1 EAV-0Chr2 EAV-0Chr2A EAV-0Chr2B EAV-0Un1 EAV-0Un2 EAV-E51 EAV-E51LChr2 EAV-E51LChr28 EAV-E51LChrZ EAV-E51LChrZB EAV-E51LUn EAV-E33 EAV-E33LChr1 EAV-E33LChr4 EAV-E33LChr6

NCBI reference sequence

Contig location

LTR identity (%)

Provirus sequence size (bp)

Retrovirus structure spanned*

Flanking direct repeat

NW_003763650.1 NW_001471639.2 NW_001471639.2 NW_003763661.1 NW_003779871.1 NW_003778217.1

21 841 561–21 846 150 8 095 311–8 101 488 21 006 029–21 012 227 6 587 549–6 593 731 1–2161D 1–1790d

95 100 100 100

4590 6178 6200 6183 2161 1790

59-LTR–Dgag–pol–Denv–LTR-39 59-LTR–gag–pol–Denv–LTR-39 59-LTR–gag–pol–Denv–LTR-39 59-LTR–gag–pol–Denv–LTR-39 59-env–LTR-39 59-env–LTR-39

GCTGAC none CCTTT TAAAC

NW_001471637.2 NW_003764302.1 NW_003764319.1 NC_006127.3 NW_003773815.1

3 060 739–3 068 550 160 868–168 457 4 190 360–4 183 955 49 054 924–49 061 458 2747–9716I

97 96 99 98

7812 7590 6406 6534 6969

59-LTR–gag–pol–env–LTR-39 59-LTR–gag–pol–env–LTR-39 59-LTR–gag–pol–env–LTR-39 59-LTR–gag–pol–env–LTR-39 59-LTR–gag–pol–env–DLTR-39

C(C/T)(A/G)GA TGTAC none GTTGT

NW_003763647.1 NW_003763739.1 NW_003763812.1

859 363–867 090 3 009 329–3 013 895 10 867 263–10 861 262

7728 4567 6002

59-LTR–gag–pol–env–LTR-39 59-pol–env–LTR-39 59-Dgag–pol–env–LTR-39

CCTGTGTA

ND ND

ND

98 ND ND

Journal of General Virology 95

ND, LTR and flanking sequence analysis not determined due to lack of sequence for one side. *Retroviral gene sequence present in locus with a partial deletion is indicated by D. DBeginning of contig in env gene. dContig location is the full contig length for the accession. IRight LTR incomplete due to poor-quality sequence before position 2747 of contig. Sequence size was determined after subtracting a region of duplicated and uncalled sequence in the contig.

ND ND

ND

ND ND

As a first step to obtain more complete EAV-E51 sequence data, an E51 probe obtained by PCR amplification of broiler chicken line 21 DNA was used to screen by hybridization a lambda genomic DNA library of the same chicken line. A positive clone (pEAV-E51) was isolated and subcloned for sequencing, but was found to be only a partial clone, comprising 4625 bp of the 39-end sequence. The pEAV-E51 sequence was used to search the G. gallus genome database to obtain the complete provirus sequence for this locus, which was found on the Z chromosome and designated EAV-E51LZ (EAV-E51-like provirus from red jungle fowl genome chromosome Z; Table 1). Comparison of the original EAV-E51 sequence, our pEAV-E51 clone from line 21 and the red jungle fowl EAV-E51LZ sequence indicated that they were 98–99 % identical to each other with identical 39 integration junctions. Hybridizations of chicken genomic DNA with probes for the EAV-E51 env gene have shown previously that these ERVs exist as moderately repeating elements within the chicken genome (Boyce-Jacino et al., 1992). The detection of only the truncated 3 kbp EAV-E51 proviruses by LTRprimed PCR suggested that these deleted structures may be the predominant form of these elements within the chicken genome, although preferential amplification of smaller amplicons could prevent amplification of more intact proviral templates. In order to examine the distribution of EAV-E51 proviruses with intact pol genes, a 275 bp RT region, which was deleted from the 3 kbp structures, was hybridized to EcoRI-digested Gallus spp. genomic DNA using high-stringency conditions. Only two hybridizing fragments from line 0 and brown leghorn chickens were detected, whilst the broiler chicken line 21 and two jungle fowl species showed additional hybridizing bands (Fig. 1), indicating that at least some of these EAV-E51 loci segregate within the chicken population. The red jungle fowl fragments were poorly resolved; however, a search of the red jungle fowl genome with the 275 bp probe sequence identified 11 potentially hybridizing loci in this subspecies, with sequence identities with the probe clone ranging from 88 to 98 %. In our initial work to obtain a prototype sequence for the EAV-E51 59 end, line 0 chicken DNA was PCR amplified with the EAV-E51 forward LTR primer and a reverse INregion primer (L4E9) designed for the sequence obtained for the 39-end sequence. A 3.9 kbp PCR product that was amplified and cloned was designated pEAV21 (EAV-E51like clone from line 21). The pEAV21 sequence represented an EAV element that was related to EAV-E51, but somewhat diverged, with only 82 % nucleotide identity to E51LZ in the non-coding leader sequence including the retrovirus packaging signal (Y) and 90–93 % identity over the length of the coding region. The pEAV21 sequence http://vir.sgmjournals.org

SJF

R JF

BR L

L 21

gene, producing a gag–env fusion. Additional smaller deletions were present in the remaining provirus sequence, including several single-base-pair deletions disrupting the reading frame of the grossly intact gag gene that rendered it non-functional.

L0

Endogenous avian retroviruses

kbp 12.2 7.1 5.1 4.1 3.1 2.0 1.6 1.0

0.5

Fig. 1. EAV-E51 elements segregate in G. gallus. The distribution of EAV-E51 RT sequences in the Gallus genome were observed by Southern blot hybridization performed on EcoRI-digested genomic DNA from three chicken lines [line 0 (L0), line 21 (L21) and brown leghorn (BRL)] and two jungle fowl species [red jungle fowl (RJF) and Gallus sonneratii (Sonnerat’s jungle fowl; SJF)] using a 275 bp EAV-E51-specific 32P-labelled RT probe under high-stringency conditions. Molecular mass size markers are indicated. Segregating proviruses are apparent within the chicken genomes. Although red jungle fowl bands were poorly resolved, the G. gallus genome database was used to count 11 potentially hybridizing proviruses from this subspecies.

was also highly defective, with multiple insertions/deletions throughout the gag coding region causing frameshifts in the deduced translated protein. A search of the G. gallus database also localized the pEAV21 ERV sequence to the Z chromosome, with a 100 % identical match that was designated EAV-E51LChrZB (Table 1). Phylogenetic analysis of EAV family member relationships In order to examine the relationships among the EAVs and ASLVs, sequences of the novel EAV-0- and EAV-E51-like clones were used to obtain additional G. gallus proviral sequences with complete pol and/or env genes for alignment of their deduced translated sequences; some of these more complete proviruses also contained gag genes that could be included for analysis (Table 1). An overall intact provirus on chromosome 1 most closely resembled the previously described clone designated EAV-E33 (Boyce-Jacino et al., 1992) and represented the full-length prototype sequence for EAV-E33, designated EAV-E33L1. Comparison of the deduced amino acid sequences of the EAV-E51 gag genes demonstrated a region of divergence 2063

M. A. Sacco and V. K. Nair (a)

(b) 0.1

EAV-HPJF

100

EAV-HP

100

0.1

EAV-HPJF

EAV-HP 100

EAV-E33LChr6

76

EAV-E33LChr1

EAV-E33LChr4

EAV-E51LChr28 EAV-E51LChrZ

98

EAV-E51LUn

EAV-E51LChrZB

87

EAV-E51LChr2

91 91

86

86

99 EAV-0Chr2B 88 EAV-0Chr2 100

100

100 RCAS-A

EAV-0Chr2 EAV-0Chr1 85 ALV-E

ALV-J 100

ALV-E

RSV-SRB RCAS-A LPDV

MMTV

EAV-HP

100

0.1

EAV-E51LChrZ

EAV-0

98 RSV-RB

(c)

EAV-E51LChr28

EAV-0Chr2B

EAV-0Chr3

100

EAV-E51LChr2

EAV-0Chr2A

EAV-0 78

EAV-E33LChr1

100

100

ALV-J 100 86 93

EAV-E51LChr2 81

EAV-E51LChr28 EAV-E33LChr1

90

100

EAV-E51LChrZ EAV-E51 EAV-E51Un

EAV-E33LChr6 91

100

EAV-E33LChr4 EAV-0Un1 EAV-0Un2 ALV-E 100 RSV-B 99 RAV-2 85

100

Subgroup E Subgroup B

MAV-2 RSV-D

Subgroup D

98 MAV-1/2 ALV-RSA 78

RCAS-A RAV-1

Subgroup A

MAV-1 RSV-C

2064

Subgroup C

Journal of General Virology 95

Endogenous avian retroviruses

Fig. 2. Phylogenetic analysis of the EAV sequences from the G. gallus genome. The CLUSTAL alignments of (a) CA, (b) RT and (c) SU proteins (Figs S3–S5) were used to construct phylogenetic trees in SeaView using the neighbour-joining method. MMTV CA and LPDV RT were included as phylogenetically distant outgroups. Bar, 0.1 expected amino acid replacements per site. Numbers at the branching points represent bootstrap confidence levels determined from 1000 data replicates that exceeded 75 %.

encompassing the C-terminal portion of the matrix or membrane-associated (MA) protein (p19) to the start of the capsid (CA) protein, which was shown previously to be dispensable for ASLV budding and infectivity in vitro (Nelle & Wills, 1996). Overall there was little similarity in deduced sequences in these regions and the size of peptides between MA and CA varied, making it difficult to predict borders of protease cleavage products to determine if the polyprotein structure was the same as delineated for the ASLV Gag. However, several important elements required for retroviral budding were apparent, including the first 87 aa of the MA (the conserved N-terminal M domain), and the presence of the virus late (L) domain motif PPPY found in the ASLV p2 and common to most retroviruses except the lentiviruses (Bowles et al., 1994; Pepinsky et al., 1996) (Fig. S2). Similarly, EAV sequences also demonstrated the presence of other polyprotein signature motifs, including the conserved three-residue TDW sequence seen in the ASLV p10, two identically spaced zinc-finger motifs (CX2CX4HX4C) in the nucleocapsid (NC; p12) and the protease (PR; p15) DSG active-site motif (Fig. S2) (Arad et al., 1995; Me´ric & Spahr, 1986).

phylogenetic tree of these viruses (Fig. 2b). Two additional EAV-0 proviruses in the red jungle fowl genome with deletions in the gag gene as well as two additional EAVE51-like elements that showed more sequence similarity to EAV-E33 (E33L4 and E33L6 for EAV-E33-like chromosome 4 and 6, respectively) were included in the analysis, and the lymphoproliferative disease virus (LPDV) of turkeys, a phylogenetically distant alpharetrovirus, was used as the outgroup. The neighbour-joining tree based on the RT sequence showed congruence with the CA tree. The inclusion of additional EAV-E51-related RT sequences demonstrated that the EAV-E51-related proviruses formed two distinct subclades of EAV-E51- or EAV-E33-like sequences. The branch lengths of the EAVE51 clade indicated a greater degree of divergence than within the EAV-0, EAV-HP or ALV clades. This degree of divergence may reflect the antiquity of these sequences, which is further supported by the divergence between the chicken and red jungle fowl Z chromosome EAV-E51 locus, the decreased LTR sequence identity (Table 1) and the accumulation of one or multiple stop codons in the pol gene within this clade.

As a result of the variability in length of the Gag polyprotein sequences, the high degree of divergence between the membrane-binding domain and CA, and the presence of point deletions/insertions causing frameshifts in many of the proviruses, the deduced CA sequence alone was used to generate an alignment for phylogenetic analysis of the EAV and ASLV Gag proteins (Fig. S3), particularly as CA has strongly conserved boundaries for recognition by the viral protease during processing of the polyprotein and contains the major homology region that is conserved in nearly all avian and mammalian retroviruses (Mammano et al., 1994). The CA of a betaretrovirus, mouse mammary tumour virus (MMTV), was used as the phylogenetically distant outgroup. The three different EAV-type elements formed distinct clades with strong bootstrap support, with EAV-E33 and EAV-E51 placed as sister taxa (Fig. 2a). Similar to the ALV clade, branch lengths were very short within the EAVHP and EAV-0 clades, indicating that these clade members were very closely related. The EAV-0 sequences were also observed to share the last common ancestor with the modern ALVs. When considered together with the high degree of LTR sequence identity within individual EAV-0 proviruses, the clustering of the EAV-0 and ALV clades might hint at the EAV-0 proviruses being the most recently integrated EAVs.

A final neighbour-joining tree was constructed using an alignment of the complete deduced sequence of the SU proteins (Fig. S5). The relationships among the ALV and EAV clades based on this SU/TM phylogenetic tree (Fig. 2c) were again congruent with the data obtained using the gagor pol-encoded sequences. The EAV-HP clade included the recently emerged ALV-J envelope sequence, which likely arose in meat-type chicken flocks by recombination between an unknown ALV and EAV-HP genomic RNA transcribed from a locus identified previously in line 15I white leghorn chickens (Sacco et al., 2004). Members of the EAV-0 and EAV-HP clades again showed short branch lengths, indicating their close relationships, whilst the EAV-E51 clade had branch lengths similar to or longer than those in the ALV clade. The divergent sequences in the ASLV env were focused primarily in the variable and hypervariable regions defined in the SU protein that determined receptor binding and consequently the virus subgroup (Fig. S1). Conversely, the EAV-E51 residue substitutions appeared to be random in SU and TM, most notably for the E33L4 sequence that appeared to be the most divergent of the clade; the sequences included one or two nonsense mutations in several of the proviruses, as well as different frameshifts in both the chicken and red jungle fowl EAV-E51 proviruses and in the EAV-E33 provirus. Taken together, these data suggested the EAVE51 proviruses may be more ancient than the EAV-0 and EAV-HP proviruses.

Alignment of the RT sequences of the EAVs and representative ALVs was also performed (Fig. S4) to generate a http://vir.sgmjournals.org

2065

M. A. Sacco and V. K. Nair

Chimeric EAVs, including ART-CH, have resulted from recombination with a deleted EAV-E51 Previous reports on EAVs have indicated that recombination events have resulted in multiple ERV elements that share regions of identity (Sacco et al., 2000). PCR amplification of EAV-0 sequences produced an ~2.8 kbp product (pEAV3), which was found upon pair-wise sequence comparison with the EAV-0 prototype and the 3 kbp EAV-E51 clone (pEAV27) to represent a chimeric provirus: the 550 bp 59 end derived from EAV-0 and the remainder of the sequence derived from the EAV-E51 provirus. The full-length published sequence of the ERV designated ART-CH is ~3 kbp long and appears to be a chimeric provirus (Gudkov et al., 1992; Sacco et al., 2000). Previous sequence comparisons indicated that the ART-CH 59 region was derived from EAV-HP, including the U5 up to nt 793 in the gag gene, whilst the 39 sequences corresponding to the partial env gene to the U3 region were derived from EAV-E51; the remainder of the gag region had no close nucleotide identity with any published sequence (Sacco et al., 2000). When we compared the defective EAV-E51 provirus of clone pEAV27 and ART-CH, the sequences were found to have 90 % sequence identity in the region of the provirus 39 to the putative recombination point shown in Fig. 3(a). Moreover, the ART-CH provirus structure was found to have the same pol deletions as this 3 kbp EAV-E51 clone, suggesting ART-CH arose by recombination between an already defective EAV-E51 and an EAV-HP. Interestingly,

the recombination points for the EAV-0/E51 chimera and the EAV-HP/E51 chimera differed, but were in close proximity within a highly conserved nucleotide block (Fig. 3a). Phylogenetic analysis further illustrated the relationship of the gag gene sequences deriving from EAV-HP 59 to the recombination site (Fig. 3b) and deriving from EAV-E51 on the 39 side of the recombination point (Fig. 3c). Using the full-length ART-CH sequence to search the G. gallus genome database, nine complete matches were obtained (Table 2), although many more elements with further deletions or that matched the 3 kbp EAV-E51 sequence were also observed. These nine G. gallus EAVHP/E51 proviruses shared 98–99 % sequence identity to each other and at least 97 % overall identity to the sequence of the chicken ART-CH clone. A comparison of the left and right LTR sequences of the individual proviruses revealed that seven loci had 100 % identical LTRs, suggesting recent integration activity of these elements.

DISCUSSION The presence of ERVs in the genome has been postulated to have some importance for host evolution, accounting for the retention of the loci within the genome (Sverdlov, 2000). This is exemplified by studies showing the involvement of human ERV (HERV) loci in genomic rearrangements (Hughes, 2001) and the adaptation of a HERV Env glycoprotein to placental development (Mi et al., 2000).

(a) EAV-HP ART-CH EAV-E51 EAV-3 EAV-0

(b)

0.1

EAV-0

(c)

0.1

EAV-0

EAV-3

EAV-HP

EAV-E51 EAV-HP

100 100

ART-CH

ART-CH 100

EAV-E51

78 EAV-3

Fig. 3. Multiple chimaeric EAV proviruses have likely arisen through recombination in a conserved region. (a) Sequence alignment of representative sequences demonstrating the recombination point between EAV-HP and EAV-E51 giving rise to the recombinant ART-CH provirus, and the recombination point between EAV-0 and EAV-E51 giving rise to the recombinant provirus represented by clone pEAV3. Sequences identical between the parent genomes where strand transfers likely occurred during replication are indicated by a line above or below the alignment. The deduced amino acid sequence is shown below the alignment for the block of nucleotide sequence that is highly conserved among all EAVs that includes both recombination points. (b, c) Maximum-likelihood analysis of EAV sequences 59 (b) or 39 (c) to the putative recombination point of chimeric proviruses. Bar, 0.1 expected nucleotide replacements per site. Bootstrap values are shown as per cent occurrence in 500 random samplings. 2066

Journal of General Virology 95

Endogenous avian retroviruses

Table 2. EAV-HP/E51 (ART-CH) from the red jungle fowl genome identified in this study Chromosome 1 1 1 2 2 3 3 3 5

NCBI reference sequence

Contig location

LTR identity (%)

Provirus sequence size (bp)

Flanking direct repeat

NW_001471551.2 NW_003763650.1 NW_003763646.1 NW_003763690.1 NW_003763661.1 NW_003763720.1 NW_001471673.2 NW_001471668.2 NW_003763748.1

3 924 891–3 928 244 2 803 238–2 806 585 503 544–506 922 348 872–352 219 20 918 248–20 921 603 28 501 302–28 504 648 2 099 517–2 102 903 42 668 903–42 672 289 4 216 429–4 219 813

99 100 100 100 99 100 100 100 100

3354 3348 3379* 3350 3356 3347 3387* 3387* 3385*

CTCTT CACTCT GGCTG GGCCAG CAGAC GTTTC AGGGAC CTGA

ND

ND,

Sequence unavailable adjacent to left LTR. *Sequence includes a 30 bp duplication in the env gene portion of the provirus.

Recently, an EAV-HP provirus was shown to determine blue eggshell colour phenotype through overexpression of the SLCO1B3 gene activated by the proviral LTR (Wang et al., 2013; Wragg et al., 2013). However, endogenous ALV (ev loci) activity has been shown to correlate with a reduction in poultry productivity traits, including fecundity and feed conversion rates (Bacon et al., 1988, 2004), and may be detrimental by causing autoimmunity through their spontaneous reactivation, as demonstrated by activation of ALV-E production in the Smyth line chicken that serves as a model for autoimmune human vitiligo (Sreekumar et al., 2000). Similarly, HERVs have been implicated in human disease (reviewed by Ryan, 2004). The maintenance of ERVs, including the EAVs, in most cases likely demonstrates a lack of selection pressure to lose elements that have become non-infectious as a result of coding defects or transcriptional silencing by host methylation (Groudine et al., 1981). Previous examination of the G. gallus genome suggested that purifying selection has operated to eliminate ERVs that disrupt gene function as both ASLVs and HIV favour integration into active transcription units (Barr et al., 2005). In our analysis of the G. gallus genome, no EAV proviruses that represent complete and intact retroviruses capable of producing infectious virus were identified. This suggests that a similar purifying selection has likely operated on the EAVs to eliminate viruses capable of producing infectious particles that would reduce bird fitness. In addition to the numerous frameshift and stop codon mutations of the ancient EAVs, most EAV-E51-related viruses had large deletions within the pol gene, whilst only two EAV-0 proviral env sequences could be retrieved. Phylogenetic analysis of the deduced EAV Env proteins demonstrated that the EAV-0 and EAV-HP sequences form clades with short branch lengths to a common ancestor compared with the ASLV Envs with different host range determinants (subgroups A, B, C and E), suggesting each clade comprises members with specificity of a single subgroup. However, the E51-related proviruses appear to http://vir.sgmjournals.org

radiate from a common ancestor to a similar or greater degree than the modern ASLV clade. These proviruses may represent different ancient EAV-E51 subgroups with different host range determinants as seen for the ASLVs. Alternatively, the EAV-E51 proviruses may have resulted from repeated germline integration of an evolving quasispecies with further divergence accumulating over time by random mutation since integration. This latter possibility is more likely for at least some of the members of this clade as most of the divergent residues appear to be randomly distributed through SU and TM sequences rather than concentrated in regions that would correspond to hypervariable or variable regions defined for SU that are under host selection pressure (Bai et al., 1995; Bova et al., 1988; Dorner et al., 1985; Rainey et al., 2003). The phylogenetic analysis may also provide a clue to the chronology of EAV integration into the Gallus genome, as the clustering of the EAV-0 and ASLV clades suggests that the EAV-0 proviruses integrated most recently of the EAVs as they are more closely related to the modern alpharetroviruses. The spread of defective ERVs has been postulated to occur during exogenous retrovirus infections, whereby exogenous virus proteins mediate packaging of the defective ERV genomes, and their subsequent reverse transcription and integration at an another genomic site. The apparent emergence of the ALV-J subgroup by recombination between an EAV-HP and an exogenous ALV provided support for this idea, showing recent activity of an EAV and its mobilization by exogenous virus-mediated packaging (Sacco et al., 2004). The data presented in this paper provide support for mobilization of defective EAVs by exogenous retrovirus infection. As retroviral replication generates identical LTRs that diverge in ERVs over the course of time through the accumulation of random mutations, the degree of divergence correlates with the time since integration (Cantrell et al., 2005). EAV-0 elements examined on chromosome 2 possessed identical LTRs, suggesting recent integration of these EAVs; however, the chromosome 1 locus (EAV-0Chr1) LTR sequences showed divergence of 2067

M. A. Sacco and V. K. Nair

~5 %. Using the autosomal mutation rate of 3.661029 substitutions per site per year as calculated by Axelsson et al. (2004), the interval between integration of the chromosome 1 and chromosome 2 elements is .4 million years. This interval is consistent with the observed distribution of EAVs in the genus Gallus, indicative of integration before separation of species from a common ancestor (Resnick et al., 1990), and the recent estimate that G. gallus and Gallus varius, the green jungle fowl, diverged from a common ancestor 3.6 million years ago (Sawai et al., 2010). The presence of nine chimeric EAV-HP/E51 proviruses in the red jungle fowl genome sharing 99 % overall sequence identity suggests that these elements may still be actively mobilized by infection with modern alpharetroviruses. We have shown previously that EAV-HP elements segregate within the chicken population (Sacco et al., 2001); the recent activity of the EAVs and the different EAV-E51 RT hybridization patterns observed for different chicken lines reported here support that other EAV elements segregate within the chicken population as well. Previously, low levels of RT activity have been found associated with virus-like particles (VLPs) in live attenuated virus vaccines derived from chicken cell substrates or from line 0 ALV-E-free chicken cells through the use of a sensitive PCR-based method for detection of the reversetranscribed product of a heterologous RNA substrate (Bo¨ni et al., 1996; Robertson & Minor, 1996). Detection of RT activity in chicken cells raised concerns for human safety; however, this RT activity did not appear to be associated with any replication-competent retrovirus (Khan et al., 1998). EAV-0 transcripts were found to be associated with VLPs obtained from live attenuated vaccines produced from chicken cell substrates (Weissmahr et al., 1997) and EAV-HP transcripts were detected in line 0 chicken VLPs (Sacco, 2001). The presence of EAV transcripts in VLPs suggests that ERVs capable of producing functional RT and Gag protein enable packaging of such transcripts; our findings in this paper suggest that only EAV-0 elements could provide the RT activity as EAV-E51-like elements appear to be highly defective and EAV-HP elements with intact pol genes were absent from line 0 (Sacco et al., 2001). A recent study of ERVs in the G. gallus genome determined that the EAVs present are outnumbered by retroviruses of the beta and gamma classes, and of an ancestral intermediate alphabeta-like class (Bolisetty et al., 2012), so ERVs that are yet to be studied could contribute to this RT activity. Our analysis of the distribution of EAV elements in the G. gallus genome has not been exhaustive as our purpose was to the define prototype sequences for complete EAV proviruses or, where that was not been possible, for complete genes of the various elements. Taken together with previous reports on EAV proviruses (Sacco et al., 2004), EAV proviruses that can produce functional Gag–Pro–Pol proteins are rare. The existence of an intact EAV-0 env gene devoid of mutations that would terminate Env translation prematurely may provide a reservoir for emergence of a new 2068

ALV subgroup in the future through recombination, in the same way that ALV-J is proposed to have emerged through recombination with an EAV-HP (Smith et al., 1999). It could be of interest to determine whether EAV-0 proviruses carrying the intact env sequences are present within breeder flocks in order to remove this potential threat.

METHODS Cloning of EAV-E51. To create an E51 hybridization probe, line 0

genomic DNA was amplified by PCR with primers E51For (59TCTGGGAGGTCCATGTTAC-39) and E51Rev (59-CCTACATAACCGTGCTGTC-39) designed from the published E51 clone env and host flanking sequences, respectively. PCR products were cloned into the pGEM-T cloning vector (Promega), and the E51-specific DNA probe was radiolabelled and hybridized to a line 21 chicken genomic DNA lambda library as described previously (Sacco, 2001). Filters were washed once with 26 SSC (30 mM, sodium citrate pH 7.0, 300 mM NaCl) and three times with 0.1 % (w/v) SDS for 15 min at 55 uC before exposure to Kodak BioMax MR film overnight at 270 uC with intensifying screens. Positive phage plaques were rescreened and an ~6 kb EcoRI fragment hybridizing to the E51 probe from a positive lambda library clone was subcloned into vector pGEM-3Z (Promega) and sequenced by MWG Biotech. PCR amplification of EAV sequences. For primary reactions, PCR was performed in 50 ml reaction mixtures on 100 ng genomic

DNA template using 2.5 U Herculase Hotstart DNA polymerase (Stratagene), 0.25 mM primers, 0.2 mM dNTPs and 4 % DMSO in 16 Herculase reaction buffer. Thermocycling was performed with 1 cycle of 95 uC for 2 min, and 35 cycles of 95 uC for 15 s, 50 uC for 30 s and 72 uC for 5 min, followed by one cycle of 72 uC for 5 min in a Mastercycler gradient thermocycler (Eppendorf). Nested PCR was performed similarly using 1 ml primary reaction as template. For EAV0, nested PCR was carried out with primers EAV0RForNot (59ATTGCGGCCGCCATTTTACCGCTCATC-39) and EAV0U3Rev1 (59GTTACATCCCGAGACCCC-39), followed by primers EAV0RForNot and EAV0U3Rev2Not (59-CTTGCGGCCGCACGCCTATTACAACATC-39). Nested amplification of the EAV-E51 3 kbp provirus was performed with primers E51U5For (59-GCCTAACACAGGTGGCTAAC-39) and E51U3Rev1 (59-GTCATGTCTCACCCTGACTG-39), followed by E51U5For and E51U3Rev2 (59-TTCACACCAGATAACATTGTGC-39). Amplification of the EAV-E51-related gag–pol fragments was carried out using a single PCR with primers E51U5For and L4E9 (59-CGTAAGGAGACAGCAAG-39). Agarose gel-purified PCR products were A-tailed and cloned into the pGEM-T Easy vector (Promega), and EAV clones (pEAV3, pEAV5 and pEAV21) were sequenced. Southern blot analysis of the E51 RT region. A 275 bp E51 RT

region was PCR amplified from line 21 chicken DNA with primers RT1 (59-CTTTTCGCTTGCTGCATGAC-39) and RT3 (59-TTGACAAATGGTGGGGGAG-39), and cloned into the pGEM-T Easy vector to produce clone pGEM-RT6. For preparation of the probe, the insert was amplified with SP6 and T7 primers, radiolabelled, and hybridized to membrane blots of EcoRI-digested genomic DNA as described previously (Sacco et al., 2000). High stringency washes were performed as detailed above for hybridization screening of the phage lambda library. Sequence and phylogenetic analysis. Sequence identity to

published retroviral sequences available in the National Center for Biotechnology Information database and Gallus_gallus-4.0 assembly (released November 2011) were performed using the BLAST search tools (www.ncbi.nlm.nih.gov/BLAST). For phylogenetic analysis of Journal of General Virology 95

Endogenous avian retroviruses deduced protein sequences, DNA sequences were translated to protein using the Translator in the JustBio suite (http://www. justbio.com/) and aligned using the CLUSTALW2 algorithm-based aligner program at the European Bioinformatics Institute (Goujon et al., 2010; Larkin et al., 2007). Phylogenetic trees were generated using SeaView version 4 (Gouy et al., 2010) using the Bio neighbourjoining (BioNJ) method with Poisson distances and confidence limits of branching points from 1000 bootstrapped datasets. Nucleotide sequence accession numbers. New nucleotide

sequences presented in this paper have been submitted to GenBank/ EMBL/DDBJ with accession numbers as follows: pEAV5 (AM418554), pEAV3 (AM418555), pEAV21 (AM418557), pEAV27 (AM4185560) and pEAV-E51 (AM418553). Sequences retrieved from the G. gallus genome database are indicated in Tables 1 and 2. Published sequences used in alignments for comparative analyses were as follows (accession numbers in parentheses): chicken transmembrane protein, 39 end, clone EAV-E51 (M95189), chicken transmembrane protein, clone EAV-E33 (M95190), G. gallus (clones 14, 15) complete retrotransposon ART-CH (L25262), layer-type chicken line 15I EAV-HP clone EAV-15I (AJ623289), grey jungle fowl EAV-HP clone EAV-JF1 (AJ292966), locus ev-1 ALV-E (AY013303), subgroup J ALV (ALV-J) prototype HPRS-103 (Z46390), RSV-B Schmidt-Ruppin B strain (AF052428), RSV-C Prague strain; (J02342), ALV subgroup A cloning vector RCAS-A (AF484679), MMTV p27 CA (NP 955569), Rousassociated virus RAV-1 (AAA67094), Myeloblastosis-associated virus MAV-1 (AAA46303) and LPDV of Meleagris gallopavo (turkey) Pol precursor (AAA62195).

Bolisetty, M., Blomberg, J., Benachenhou, F., Sperber, G. & Beemon, K. (2012). Unexpected diversity and expression of avian

endogenous retroviruses. MBio 3, e00344-12. Bo¨ni, J., Stalder, J., Reigel, F. & Schu¨pbach, J. (1996). Detection of

reverse transcriptase activity in live attenuated virus vaccines. Clin Diagn Virol 5, 43–53. Bova, C. A., Olsen, J. C. & Swanstrom, R. (1988). The avian retrovirus

env gene family: molecular analysis of host range and antigenic variants. J Virol 62, 75–83. Bowles, N., Bonnet, D., Mulhauser, F. & Spahr, P. F. (1994). Site-

directed mutagenesis of the P2 region of the Rous sarcoma virus gag gene: effects on Gag polyprotein processing. Virology 203, 20–28. Boyce-Jacino, M. T., Resnick, R. & Faras, A. J. (1989). Structural

and functional characterization of the unusually short long terminal repeats and their adjacent regions of a novel endogenous avian retrovirus. Virology 173, 157–166. Boyce-Jacino, M. T., O’Donoghue, K. & Faras, A. J. (1992). Multiple

complex families of endogenous retroviruses are highly conserved in the genus Gallus. J Virol 66, 4919–4929. Cantrell, M. A., Ederer, M. M., Erickson, I. K., Swier, V. J., Baker, R. J. & Wichman, H. A. (2005). MysTR: an endogenous retrovirus family in

mammals that is undergoing recent amplifications to unprecedented copy numbers. J Virol 79, 14698–14707. Coffin, J. M., Tsichlis, P. N., Conklin, K. F., Senior, A. & Robinson, H. L. (1983). Genomes of endogenous and exogenous avian retroviruses.

Virology 126, 51–72. Crittenden, L. B. (1991). Retroviral elements in the genome of

ACKNOWLEDGEMENTS

the chicken: implications for poultry genetics and breeding. Crit Rev Poultry Biol 3, 73–91.

This work was partly funded by the British Biotechnology and Biological Sciences Research Council (V. N. K.) and start-up funds from California State University, Fullerton (M. A. S.).

Dorner, A. J., Stoye, J. P. & Coffin, J. M. (1985). Molecular basis of

REFERENCES Arad, G., Chorev, M., Shtorch, A., Goldblum, A. & Kotler, M. (1995).

Point mutation in avian sarcoma leukaemia virus protease which increases its activity but impairs infectious virus production. J Gen Virol 76, 1917–1925. Axelsson, E., Smith, N. G., Sundstro¨m, H., Berlin, S. & Ellegren, H. (2004). Male-biased mutation rate and divergence in autosomal,

Z-linked and W-linked introns of chicken and Turkey. Mol Biol Evol 21, 1538–1547. Bacon, L. D., Smith, E., Crittenden, L. B. & Havenstein, G. B. (1988).

Association of the slow feathering (K) and an endogenous viral (ev21) gene on the Z chromosome of chickens. Poult Sci 67, 191–197. Bacon, L. D., Fulton, J. E. & Kulkarni, G. B. (2004). Methods

for evaluating and developing commercial chicken strains free of endogenous subgroup E avian leukosis virus. Avian Pathol 33, 233–243. Bai, J., Howes, K., Payne, L. N. & Skinner, M. A. (1995). Sequence of host-

host range variation in avian retroviruses. J Virol 53, 32–39. Dunwiddie, C. & Faras, A. J. (1985). Presence of retrovirus reverse

transcriptase-related gene sequences in avian cells lacking endogenous avian leukosis viruses. Proc Natl Acad Sci U S A 82, 5097–5101. Dunwiddie, C. T., Resnick, R., Boyce-Jacino, M., Alegre, J. N. & Faras, A. J. (1986). Molecular cloning and characterization of gag-, pol-, and

env-related gene sequences in the ev– chicken. J Virol 59, 669–675. Frisby, D. P., Weiss, R. A., Roussel, M. & Stehelin, D. (1979).

The distribution of endogenous chicken retrovirus sequences in the DNA of galliform birds does not coincide with avian phylogenetic relationships. Cell 17, 623–634. Gifford, R., Kabat, P., Martin, J., Lynch, C. & Tristem, M. (2005).

Evolution and distribution of class II-related endogenous retroviruses. J Virol 79, 6478–6486. Goujon, M., McWilliam, H., Li, W. Z., Valentin, F., Squizzato, S., Paern, J. & Lopez, R. (2010). A new bioinformatics analysis tools framework

at EMBL-EBI. Nucleic Acids Res 38 (Web Server issue), W695–W699. Gouy, M., Guindon, S. & Gascuel, O. (2010). SeaView version 4: a

multiplatform graphical user interface for sequence alignment and phylogenetic tree building. Mol Biol Evol 27, 221–224.

range determinants in the env gene of a full-length, infectious proviral clone of exogenous avian leukosis virus HPRS-103 confirms that it represents a new subgroup (designated J). J Gen Virol 76, 181–187.

Groudine, M., Eisenman, R. & Weintraub, H. (1981). Chromatin

Barr, S. D., Leipzig, J., Shinn, P., Ecker, J. R. & Bushman, F. D. (2005).

Gudkov, A. V., Komarova, E. A., Nikiforov, M. A. & Zaitsevskaya, T. E. (1992). ART-CH, a new chicken retroviruslike element. J Virol 66,

Integration targeting by avian sarcoma-leukosis virus and human immunodeficiency virus in the chicken genome. J Virol 79, 12035– 12044. Be´nit, L., Lallemand, J. B., Casella, J. F., Philippe, H. & Heidmann, T. (1999). ERV-L elements: a family of endogenous retrovirus-like

elements active throughout the evolution of mammals. J Virol 73, 3301–3308. http://vir.sgmjournals.org

structure of endogenous retroviral genes and activation by an inhibitor of DNA methylation. Nature 292, 311–317.

1726–1736. Herniou, E., Martin, J., Miller, K., Cook, J., Wilkinson, M. & Tristem, M. (1998). Retroviral diversity and distribution in vertebrates. J Virol 72,

5955–5966. Hillier, L. D. W., Miller, W., Birney, E., Warren, W., Hardison, R. C., Ponting, C. P., Bork, P., Burt, D. W., Groenen, M. A. M. & other

2069

M. A. Sacco and V. K. Nair authors (2004). Sequence and comparative analysis of the chicken

genome provide unique perspectives on vertebrate evolution. Nature 432, 695–716. Hughes, D. C. (2001). Alternative splicing of the human VEGFGR-3/

FLT4 gene as a consequence of an integrated human endogenous retrovirus. J Mol Evol 53, 77–79. Hughes, S. H., Toyoshima, K., Bishop, J. M. & Varmus, H. E. (1981).

Organization of the endogenous proviruses of chickens: implications for origin and expression. Virology 108, 189–207. Khan, A. S., Maudru, T., Thompson, A., Muller, J., Sears, J. F. & Peden, K. W. (1998). The reverse transcriptase activity in cell-free

medium of chicken embryo fibroblast cultures is not associated with a replication-competent retrovirus. J Clin Virol 11, 7–18. Larkin, M. A., Blackshields, G., Brown, N. P., Chenna, R., McGettigan, P. A., McWilliam, H., Valentin, F., Wallace, I. M., Wilm, A. & other authors (2007). Clustal W and Clustal X version 2.0. Bioinformatics

23, 2947–2948. Mammano, F., Ohagen, A., Ho¨glund, S. & Go¨ttlinger, H. G. (1994).

Role of the major homology region of human immunodeficiency virus type 1 in virion morphogenesis. J Virol 68, 4927–4936. Martin, J., Herniou, E., Cook, J., O’Neill, R. W. & Tristem, M. (1999).

Interclass transmission and phyletic host tracking in murine leukemia virus-related retroviruses. J Virol 73, 2442–2449. Me´ric, C. & Spahr, P. F. (1986). Rous sarcoma virus nucleic acid-

binding protein p12 is necessary for viral 70S RNA dimer formation and packaging. J Virol 60, 450–459. Mi, S., Lee, X., Li, X., Veldman, G. M., Finnerty, H., Racie, L., LaVallie, E., Tang, X. Y., Edouard, P. & other authors (2000). Syncytin is

a captive retroviral envelope protein involved in human placental morphogenesis. Nature 403, 785–789. Nelle, T. D. & Wills, J. W. (1996). A large region within the

Rous sarcoma virus matrix protein is dispensable for budding and infectivity. J Virol 70, 2269–2276. Pepinsky, R. B., Papayannopoulos, I. A., Campbell, S. & Vogt, V. M. (1996). Analysis of Rous sarcoma virus Gag protein by mass spectrometry

indicates trimming by host exopeptidase. J Virol 70, 3313–3318. Rainey, G. J., Natonson, A., Maxfield, L. F. & Coffin, J. M. (2003).

Mechanisms of avian retroviral host range extension. J Virol 77, 6709– 6719. Resnick, R. M., Boyce-Jacino, M. T., Fu, Q. & Faras, A. J. (1990).

Phylogenetic distribution of the novel avian endogenous provirus family EAV-0. J Virol 64, 4640–4653. Robertson, J. S. & Minor, P. (1996). Reverse transcriptase activity in

significance of reverse transcriptase activity in chick cell-derived vaccines. Biologicals 25, 403–414. Ruis, B. L., Benson, S. J. & Conklin, K. F. (1999). Genome structure and expression of the ev/J family of avian endogenous viruses. J Virol 73, 5345–5355. Ryan, F. P. (2004). Human endogenous retroviruses in health and disease: a symbiotic perspective. J R Soc Med 97, 560–565. Sacco, M. A. (2001). Characterisation of the EAV-HP endogenous

retrovirus elements of the Gallus genus. PhD thesis, University of London, London, UK. Sacco, M. A., Flannery, D. M., Howes, K. & Venugopal, K. (2000).

Avian endogenous retrovirus EAV-HP shares regions of identity with avian leukosis virus subgroup J and the avian retrotransposon ART-CH. J Virol 74, 1296–1306. Sacco, M. A., Howes, K. & Venugopal, K. (2001). Intact EAV-HP

endogenous retrovirus in Sonnerat’s jungle fowl. J Virol 75, 2029– 2032. Sacco, M. A., Howes, K., Smith, L. P. & Nair, V. K. (2004). Assessing

the roles of endogenous retrovirus EAV-HP in avian leukosis virus subgroup J emergence and tolerance. J Virol 78, 10525–10535. Sawai, H., Kim, H. L., Kuno, K., Suzuki, S., Gotoh, H., Takada, M., Takahata, N., Satta, Y. & Akishinonomiya, F. (2010). The origin and

genetic variation of domestic chickens with special reference to junglefowls Gallus g. gallus and G. varius. PLoS ONE 5, e10639. Smith, L. M., Toye, A. A., Howes, K., Bumstead, N., Payne, L. N. & Venugopal, K. (1999). Novel endogenous retroviral sequences in the

chicken genome closely related to HPRS-103 (subgroup J) avian leukosis virus. J Gen Virol 80, 261–268. Sreekumar, G. P., Smyth, J. R., Jr, Ambady, S. & Ponce de Leon, F. A. (2000). Analysis of the effect of endogenous viral genes in the Smyth

line chicken model for autoimmune vitiligo. Am J Pathol 156, 1099– 1107. Sverdlov, E. D. (2000). Retroviruses and primate evolution. Bioessays

22, 161–171. Wang, Z. P., Qu, L. J., Yao, J. F., Yang, X. L., Li, G. Q., Zhang, Y. Y., Li, J. Y., Wang, X. T., Bai, J. R. & other authors (2013). An EAV-HP

insertion in 59 Flanking region of SLCO1B3 causes blue eggshell in the chicken. PLoS Genet 9, e1003183. Weissmahr, R. N., Schu¨pbach, J. & Bo¨ni, J. (1997). Reverse

transcriptase activity in chicken embryo fibroblast culture supernatants is associated with particles containing endogenous avian retrovirus EAV-0 RNA. J Virol 71, 3005–3012.

vaccines derived from chick cells. Biologicals 24, 289–290.

Wragg, D., Mwacharo, J. M., Alcalde, J. A., Wang, C., Han, J. L., Gongora, J., Gourichon, D., Tixier-Boichard, M. & Hanotte, O. (2013).

Robertson, J. S., Nicolson, C., Riley, A. M., Bentley, M., Dunn, G., Corcoran, T., Schild, G. C. & Minor, P. (1997). Assessing the

Endogenous retrovirus EAV-HP linked to blue egg phenotype in Mapuche fowl. PLoS ONE 8, e71393.

2070

Journal of General Virology 95