TEXT

Top Abstract Text References

The common brushtail possum (Trichosurus vulpecula) is an Australian marsupial from the family Phalangeridae, which also includes the cuscuses (Ailurops ursinus, Phalanger spp., Spilocuscus spp., Strigocuscus spp.), the scaly-tailed possum (Wyulda squamicaudata), and the other members of the brushtail (Trichosurus) genus (northern brushtail [T. arnhemensis], mountain brushtail [T. caninus], and coppery brushtail [T. johnstonii] possums) (8). Extinct possums belonging to the brushtail (Trichosurus) genus are present in sediments deposited 20 million years ago, but gaps in the fossil record preclude accurate determination of the time of speciation of the modern brushtail species (8). However, T. vulpecula, now one of the most widely distributed possums, appears to have proliferated only within the last 1 million years (8). Common brushtail possums were introduced into New Zealand in the late 1800s and early 1900s to establish a fur trade and have since flourished to such an extent that they are now considered a major environmental pest and economic threat (18).

The type D retroviruses had, until recently, been observed only in primates. The genomes of all Cercopithecine Old World monkeys contain endogenous type D retroviruses (simian endogenous retroviruses, SERVs) (31). Exogenous type D retroviruses, namely simian retrovirus type 1 (SRV-1), simian retrovirus type 2 (SRV-2), and Mason-Pfizer monkey virus (MPMV, also known as SRV-3), which appear to have evolved from the endogenous SERVs (31), have been detected in captive macaques worldwide (9). An endogenous type D retrovirus, squirrel monkey retrovirus (SMRV), is present in the genome of the New World squirrel monkey (7, 10). Recently, type D endogenous retroviruses, named MusD, from mice have been reported (14, 26).

We have been investigating the presence of ERVs in the genome of the common brushtail possum, and here we report the isolation and characterization of a type D ERV from possums.

We began this project looking for exogenous retroviruses that infect the common brushtail possum. In preliminary experiments (results not shown here), reverse transcriptase (RT)-PCR using degenerate primers derived from retroviral pol genes (11) was performed on RNA isolated from filtered possum blood serum. A ~130-bp product was generated and was cloned and sequenced and confirmed to be of retroviral origin. A 3' rapid amplification of cDNA ends (RACE) approach was used to amplify the 3' end of the retroviral RNA. Subsequently, the provirus from which the RNA was derived was detected, by PCR, in the genomic DNA of all possums tested and was assumed to be an endogenous possum retrovirus. Using a variety of PCR and RT-PCR approaches a contiguous sequence, named TvERV(D) contig (for Trichosurus vulpecula endogenous retrovirus type D), of the entire possum endogenous retrovirus was assembled.

Because the TvERV(D) contig sequence was assembled from a number of PCR products which were potentially derived from related yet distinct TvERV(D) proviruses, we deemed it necessary to amplify and sequence a single copy of this retrovirus. Liver samples were taken from possums euthanatized by intraperitoneal injection of sodium pentobarbital (Pentobarb 300; Chemstoch Animal Health Ltd., Auckland, New Zealand), and genomic DNA was isolated by the proteinase-K/sodium dodecyl sulfate and phenol-chloroform extraction method of Sambrook et al. (27). Genomic DNA was PCR amplified using the U5' for (5'-GTCTCCTTCCTCTCCGTGAT-3') and 3'UTRrev (5'-GCAACTTGGGTCTGATAATGAG-3') primers derived from the TvERV(D) contig and the Expand Long Template PCR system (Boehringer Mannheim). Products of a range of sizes, most between ~7.0 and ~9.3 kb, were generated (data not shown). An ~9.3-kb product, corresponding to the size expected from the sequence of the TvERV(D) contig, was gel purified and cloned into the pGEM-T vector (Promega). One clone, named pTvERV(D), was sequenced by primer-walking using an ABI PRISM 377 automated DNA sequencer and BigDye terminator chemistry (PE Applied Biosystems), and its sequence is shown in Fig. 1.

View larger version (194K): [in this window] [in a new window]   FIG. 1.   Nucleotide sequence and deduced amino acid sequences of pTvERV(D). Locations of regulatory regions within the nucleotide sequence, including the internal repeat (IR) at the 3' end of U5, the PBS, and SD, are overlined. The deduced amino acid sequences of the gag, pro, and pol ORFs, as well as a short region with TM homology, are shown below the nucleotide sequence. The shaded regions within each protein correspond to conserved domains described in the text. A shill (/) indicates a frameshift.

The region of TvERV(D) not covered by pTvERV(D), namely, the 3' UTR and most of the long terminal repeat (LTR), was also amplified and cloned. The envfor primer (5'-CAGCAGGAAGAGCGACTACAAT-3'), corresponding to nucleotides (nt) 8795 to 8816 of pTvERV(D), and the U5rev primer (5'-CAGGTCACACAATCGTGGGT-3'), the reverse complement of nt 99 to 118 of pTvERV(D), were used to PCR amplify the 3' end of the env gene, the 3'UTR, and almost the entire 3' LTR of a TvERV(D) element. The ~890-bp product was cloned and sequenced, and it was named p3'UTR/LTR.

The TvERV(D) sequences contained several regulatory regions and open reading frames (ORFs) characteristic of retroviruses. The TvERV(D) LTR contains characteristic TG and CA inverted repeat sequences, which are located at the termini of all retroviral LTRs (4), at its 5' and 3' ends, respectively. Potential CAAT box, TATA box, and poly(A) signal sequences were also apparent (data not shown). Otherwise, no similarity to other retroviral LTRs was detected. The sequence from nt 133 to 151 of pTvERV(D) (5'-TGGCGCCCAAGCGTGGGGC-3') was, when the G residue at nt 143 was removed, perfectly complementary to the 18 nt at the 3' end of mammalian tRNA_1,2^Lys (24) and probably represents the primer binding site (PBS) of TvERV(D). Other retroviruses with PBSs complementary to tRNA_1,2^Lys include the exogenous type D retroviruses of Old World monkeys (MPMV, SRV-1, SRV-2), an endogenous type D retrovirus of the New World squirrel monkey (SMRV), the exogenous and endogenous type B/type D retroviruses in sheep and goats (JSRV, ENTV, ESRV), human foamy virus, and Visna Maedi virus (22). A potential polypurine tract (PPT) was located immediately 5' to the LTR sequence of p3'UTR/LTR (results not shown). A potential splice donor (SD) was located at nt 211 to 218 of pTvERV(D); the sequence at this position (5'-AGGTAAGT-3') was identical to the consensus eukaryotic SD sequence (19). Although no splice acceptors (SAs) were identified in the sequence of pTvERV(D), a potential SA was present in the sequence of pEnv(D) (see below).

pTvERV(D) possessed three large ORFs (Fig. 1). Database searches using the amino acid sequences of these ORFs confirmed that they encoded uninterrupted retroviral Gag, Pro, and Pol polyproteins. The gag, pro, and pol ORFs were all in separate reading frames, an arrangement which resembles those of the type B, type D, and human T-cell leukemia virus/bovine leukemia virus groups of retroviruses (22). Although a complete ORF encoding an Env polyprotein was not present, a short region of nucleotide sequence encoded an amino acid sequence with homology to the transmembrane (TM) proteins of several retroviruses (Fig. 1).

The pTvERV(D) Gag polyprotein, encoded by nt 238 to 2361 of pTvERV(D) (Fig. 1), was 708 amino acids (aa) long. One region (encoded by nt 1705 to 1764) (Fig. 2) bore significant similarity to the major homology region, a highly conserved region within the capsid proteins of all retroviruses (Fig. 1) (1, 4). Two CCHC motifs (nt 2038 to 2079 and nt 2137 to 2178) (Fig. 1) were located near the C terminus of the Gag protein and presumably lie within the nucleocapsid (2, 33).

View larger version (64K): [in this window] [in a new window]   FIG. 2.   Nucleotide sequence and deduced amino acid sequences of pTvERV(D)-env. The putative SA is overlined. The deduced amino acid sequences of the 3' end of the pol gene and the entire env gene are shown below the nucleotide sequence. The shaded regions within each protein correspond to the conserved domains described in the text. Sites where the Env polyprotein is cleaved are indicated by arrows (). Asparagine (N) residues which are within A-X-S/T motifs (where X is any amino acid except proline) and are therefore potentially glycosylated are underlined.

The pTvERV(D) Pro ORF (nt 2193 to 3131) (Fig. 1) was 313 codons long. The amino-terminal half of this ORF contained five motifs conserved in all dUTPs (6, 17). The type B and type D retroviruses encode dUTPs in the 5' halves of their pro ORFs (6). The carboxy-terminal half of the pTvERV(D) Pro protein contained three motifs present in all retroviral proteases and cellular aspartyl proteases (25, 32), including the active site triplet DSG (Fig. 1).

The pTvERV(D) pol ORF (nt 3098 to 5710) (Fig. 1) encoded 871 aa. The amino-terminal two-thirds of the pol ORF encoded the RT protein, including eight motifs conserved in the polymerase domains of all retroviral RTs (12) and six motifs conserved in retroviral RNase H (12, 16). The carboxy-terminal one-third of the pol ORF encoded the integrase and contained an amino-terminal HHCC motif and a central D,D(35)E motif (3, 13, 23).

pTvERV(D) did not possess an env ORF. However, a short region of the sequence (nt 8518 to 8920) (Fig. 2) encoded an amino acid sequence with homology to the TM domains of Env proteins of the simian type D retroviruses and related viruses. Of the remaining sequence 3' to the pTvERV(D) pol gene, the reverse complement of nt 6608 to 6817 encoded a short sequence of amino acids with limited identity (22 of 73 amino acids) to the envelope protein of porcine reproductive and respiratory syndrome virus (5). Otherwise, the 3' region of pTvERV(D) did not encode any detectable proteins.

To determine whether TvERV(D) elements with intact env genes resided within the possum genome, genomic DNA was PCR amplified with the polfor primer (5'-CCAAGCACGTTATCAATCACA-3') from within the pol gene and the 3'UTRrev primer described above. PCR products ranging in size from ~4 kb [the size expected from pTvERV(D)] to ~1.4 kb were generated, with prominent bands at ~2.8 and ~2.0 kb (data not shown). The ~2.8- and ~2.0-kb bands were gel purified, cloned, and sequenced. The nucleotide sequence of one of the clones of the ~2.8-kb fragment, named pTvERV(D)-env, is shown in Fig. 2. The 5' and 3' ends of pTvERV(D)-env were highly similar, but not identical, to regions of pTvERV(D). The 534 nt at the 5' end of pTvERV(D)-env were 99% identical to the corresponding region of pTvERV(D), and the 749 nt at the 3' end of pTvERV(D)- env were 93% identical to the corresponding region of pTvERV(D).

pTvERV(D)-env encoded an uninterrupted Env polyprotein, as determined by database searches and comparison with the Env proteins of other retroviruses. A candidate SA was located just downstream of the pol ORF of pTvERV(D)-env (nt 540 to 543) (Fig. 2). This SA is probably used, in conjunction with the SD located just upstream of the gag gene, to generate subgenomic mRNAs for translation of the env gene. The Env polyprotein was 620 aa in length and possessed several motifs that are conserved in other retroviral Env proteins (Fig. 2). By comparison with the Env polyproteins of type D retroviruses (35, 39, 46, 47) and using software available on the internet (20), the Env polyprotein encoded by pTvERV(D)-env was predicted to comprise a signal peptide of 17 aa, a surface (SU) protein of 372 aa, and a TM protein of 231 aa (Fig. 2). The SU protein contained five potential sites for glycosylation, and the TM protein contained one. Two hydrophobic regions within the TM protein were identified: the fusion peptide, at the amino terminus, and the membrane-spanning domain near the carboxyl terminus of the TM protein (Fig. 2). The amino acid sequence encoded by nt 1943 to 2047 of pTvERV(D)-env is identical to the immunosuppressive peptide sequences of the simian type D retroviruses (29).

The deduced amino acid sequences encoded by the pTvERV(D) pol ORF and the pTvERV(D)-env env ORF were aligned with corresponding sequences of other retroviruses using the CLUSTAL X application (30), with default gap opening and extension parameters and the BLOSUM series protein weight matrix. Neighbor-joining trees were generated from the alignments using CLUSTAL X (gaps excluded; 1,000 bootstrap replicates) (Fig. 3). The tree constructed using pTvERV(D) Pol grouped pTvERV(D) with the exogenous (MPMV, SRV-1, SRV-2) and endogenous (SERV 23.1, SERV 25.2) type D retroviruses of Old World monkeys, an endogenous type D element in mice (MusD), and an endogenous type D retrovirus in the New World monkey (SMRV-H) (Fig. 3A). Within this group, pTvERV(D) and SMRV-H formed a well-supported clade. However, the pTvERV(D) and SMRV-H Pol sequences are highly divergent from each other and from those of the other type D retroviruses, as evidenced by the long branch lengths shown in Fig. 3A.

View larger version (17K): [in this window] [in a new window]   FIG. 3.   Neighbor-joining trees based on the derived amino acid sequences of the Pol proteins of type B and D retroviruses (A) and the Env proteins of the type D retroviruses and retroviruses with related pol genes (B). Amino acid sequences were aligned using CLUSTAL X, and the aligned sequences were used to construct neighbor-joining trees. Bootstrap values for 1,000 replicated trees are indicated. The Pol tree was rooted using MMTV as the outgroup, whereas the Env tree was rooted using Gibbon ape leukemia virus (GALV) and Moloney murine leukemia virus (MoMLV) as the outgroup. The accession numbers of retroviral sequences are as follows: BaEV, D10032; ENTV, Y16627; GALV, M26927; JSRV, M80216; MMTV, M15122; MPMV, AF033815; MoMLV, AF033811; RD114 env/Env, X87829; REV env/Env, 228842; SERV 23.1, U85505; SERV 25.2; U85506; SMRV-H, M23385; SNV Env, VCFVAS; SRV-1, M11841; and SRV-2, M16605.

pTvERV(D)-env and SMRV-H also clustered together on the basis of their Env sequences (Fig. 3B). In this case, however, they clustered with the Env proteins of reticuloendotheliosis virus and avian spleen necrosis virus (REV and SNV, respectively). The Old World monkey type D Env proteins formed a separate well-supported clade which also included the Env proteins of RD114 and BaEV (Fig. 3B). The MusD elements sequenced so far do not possess env genes (27) and therefore could not be included in the Env tree.

Southern hybridization analysis was used to estimate the number of copies of TvERV(D) in the possum genome. Genomic DNA was isolated from 12 possums from two geographically distinct regions of the North Island of New Zealand. Approximately 10 µg of genomic DNA from each possum was digested with PvuII [which cuts within the gag and pro genes of TvERV(D)], electrophoresed, and Southern blotted to Hybond N+ nylon membrane (Amersham Pharmacia Biotech) by the alkaline method of Sambrook et al. (27). A gag probe, which was generated by PCR from pTvERV(D) using primers from within the gag gene, was labeled with [-³²P]dCTP using a Rediprime II labeling kit (Amersham Pharmacia Biotech). The membrane was hybridized, washed to moderate stringency, and exposed to XAR film (Kodak) according to standard protocols (27). Approximately 30 copies of TvERV(D) were detected per possum genome (Fig. 4A). Although a few TvERV(D) junction fragments were conserved in all possums, the hybridization pattern was highly variable between possums from the same region and from different geographical regions (Fig. 4A).

View larger version (64K): [in this window] [in a new window]   FIG. 4.   Southern hybridization of restriction enzyme-digested genomic DNA from 12 possums with the TvERV(D) gag probe. (A) The gag probe detects junction fragments generated by PvuII cleavage 5' to the 5' LTR of TvERV(D) and at the 3' end of the TvERV(D) gag gene. Thus, each TvERV(D) provirus will generate a junction fragment of a distinct size, depending on the distance between the provirus and the nearest PvuII site. (B) The gag probe detects internal TvERV(D) fragments generated by NheI cleavage within the 5' LTR and env gene of TvERV(D) elements with intact env genes and within the 5' and 3' LTRs of TvERV(D) elements with disrupted env genes. Possums 1 through 6 were from the Wellington region and possums 7 through 12 were from the Waikato region of the North Island of New Zealand. Bands common to all 12 possums are indicated by arrows (right). DNA sizes are indicated at left.

Southern hybridization was also used to investigate the variability in TvERV(D) structure within the genomes of possums. NheI-digested genomic DNA from 12 possums was hybridized with the gag probe described above (Fig. 4B). The TvERV(D) LTRs contain a single NheI site within the U3 region. Neither the TvERV(D) contig nor pTvERV(D) contained any additional NheI sites, and hybridization of NheI-digested possum DNA to the gag probe would be expected to detect an ~9.5-kb band generated by restriction within only the LTRs of these proviruses. pTvERV(D)-env, on the other hand, possesses an NheI cut site at nt 1228 to 1233, and NheI digestion of TvERV(D) elements containing intact env genes would be expected to produce an ~6.6-kb band by digestion within the 5' LTR and env sites. As can be seen in Fig. 4B, NheI digestion of the DNA of 12 possums and hybridization with the gag probe detected the expected ~9.5- and ~6.6-kb bands. In addition, several other NheI fragments, ranging in size from ~7.0 to >10 kb, were present in the genomes of all possums. These fragments presumably represent TvERV(D) elements which are present in the genomes of all of the possums tested and which have gained or lost NheI sites or have undergone other rearrangements (insertions or deletions), relative to those which generated the ~9.5- and ~6.6-kb fragments. Numerous smaller and larger fragments were observed which were not conserved in all possums. These may represent TvERV(D) elements possessing alternative NheI sites and/or rearrangements which are not conserved in all possums.

In summary, we have amplified, cloned, and sequenced a near-full-length endogenous retrovirus, pTvERV(D), from the genome of a marsupial, the common brushtail possum (T. vulpecula). TvERV(D) is only the second full-length marsupial retrovirus sequence to be reported. pTvERV(D) possessed gag, pro, and pol genes which were uninterrupted by frameshift mutations or premature stop codons. Although no env gene was present, a short region encoding an amino acid sequence with homology to the TM domains of retroviral Env proteins was detected. In addition, a complete env gene was amplified from possum genomic DNA using primers derived from the pTvERV(D) sequence.

Phylogenetic analysis, using the deduced amino acid sequences of the pol ORF of pTvERV(D) and the env ORF of pEnv(D), placed TvERV(D) elements with the simian type D retroviruses, to the exclusion of the type B/type D retroviruses of sheep and goats (JSRV, ENTV) and the type B retroviruses of mice (mouse mammary tumor virus [MMTV]). Furthermore, TvERV(D) uses the same tRNA species (tRNA_1,2^Lys) as the majority of the simian type D retroviruses, encodes a dUTPase protein at the 5' end of its pro gene as do the type D retroviruses, and shares a genetic organization (requiring frameshifts at the gag-pro and pro-pol overlaps for translation of the gag-pro-pol polyprotein) with the type D retroviruses. Based on these findings, we suggest that TvERV(D) be classified as a type D retrovirus.

Southern hybridization and PCR (results not shown) revealed a range of TvERV(D) internal fragments common to all possums tested, suggesting that possums inherited TvERV(D) from their evolutionary ancestors. Southern hybridization detection of TvERV(D) junction fragments suggested that approximately 30 copies of TvERV(D) are present per possum genome. However, there was considerable variability in the patterns of junction fragments between possums, which suggests postspeciation (and perhaps ongoing) retrotranspositional activity. Alternative sources of the junction fragment variability could be loss or gain of restriction enzyme sites by point mutation and/or loss of the internal proviral sequences by homologous recombination between LTRs (15, 28). However, these processes would have had to have occurred at much higher rates than have previously been observed in order to have generated the variability seen in possums. That they possess intact gag, pro, and pol ORFs suggests that TvERV(D)s are at least capable of intracellular retrotransposition, and the presence of intact TvERV(D) env genes in the possum genome further suggests that these elements could produce infectious particles.

Thus, it seems likely that common brushtail possums, when they first evolved, already possessed several TvERV(D) elements, some with intact env genes and some without, in their genomes. Subsequent amplification within the germ line, either by intracellular retrotransposition or by extracellular reinfection of germ line cells, has given rise to the currently observed variability in distribution of TvERV(D) elements in the genomes of possums.

The recent report of a type D retroelement, named MusD, in mice (14, 26) suggests that rodents have played a role in the evolution of type D retroviruses. If type D retroviruses are more widespread in murids, this family of rodents might have been responsible for the dispersal of the type D retroviruses to Australia. Members of the Hydromyinae subfamily of the Muridae entered Australia at least 4 million years ago (35) and possibly as much as 8 million years ago (34), and members of the Rattus genus (subfamily, Murinae) reached Australia ~1 million years ago (35). However, it should be noted that Ristevski et al. (26) did not detect MusD elements in rat genomic DNA by Southern blot analysis. Perhaps sequencing of the rat genome (21) will allow the detection of MusD-related elements that could not be detected by Southern hybridization.

Clearly, much more work will be required to characterize the distribution of type D retroviruses, particularly in their murine or murid and possum or marsupial hosts, to test these hypotheses.

Nucleotide sequence accession numbers. The sequence of pTvERV(D) is located in the GenBank nucleotide sequence database under accession no. AF224725. The sequence of the TvERV(D) PPT and LTR are located together under accession no. AF286348. The sequence of pTvERV(D)-env has accession no. AF284693.