In vivo tumorigenesis by Jaagsiekte sheep retrovirus (JSRV) requires Y590 in Env TM, but not full length orfX open reading frame

Abstract

Jaagsiekte retrovirus (JSRV) causes ovine pulmonary adenocarcinoma (OPA), a transmissible lung cancer of sheep. The envelope (Env) glycoprotein protein of JSRV functions as a dominant oncoprotein in vitro and in vivo. An SH2 binding domain (YXXM) in the cytoplasmic tail of the JSRV Env is one of the main determinants of viral transformation at least in vitro. In these studies, we report the first in vivo tests of site-specific mutants of JSRV in their natural host, the sheep. We show that, in vivo, JSRV₂₁ with the cytoplasmic tail YXXM mutated to DXXM did not cause disease nor detectable infection, indicating that this motif is absolutely required for virus replication and possibly transformation in vivo. In contrast, mutation of the JSRV open reading frame orfX, for which no function has yet been attributed, did not alter the disease induced by JSRV₂₁.

Introduction

Jaagsiekte sheep retrovirus (JSRV) is the causative agent of ovine pulmonary adenocarcinoma (OPA), an infectious lung tumor of sheep. The disease is of interest both as a real veterinary problem, and as a potential model for some human lung adenocarcinomas (Perk and Hod 1982; Palmarini and Fan, 2001; Mornex et al., 2003). In recent years we and others have investigated the mechanism of oncogenesis by JSRV using in vitro systems. JSRV itself does not replicate efficiently in cell culture, but expression of the JSRV envelope glycoprotein (Env) is able to morphologically transform various cell lines including rodent fibroblast and epithelial cells, and human lung epithelial cells (Allen et al., 2002; Chow et al., 2003; Danilkovitch-Miagkova et al., 2003; Liu and Miller, 2005; Maeda et al., 2001; Palmarini et al., 2001; Rai et al., 2001). Thus, JSRV Env appears to function as an oncogene, a unique situation for mammalian oncogenic retroviruses.

We found that the in vitro transforming activity of JSRV Env is dependent on the short cytoplasmic tail of the C-terminal transmembrane (TM) domain. In particular, a single tyrosine at position 590 is critical since mutation of this residue (Y590F, Y590D or Y590A) abolishes transformation in rodent cells (Hull & Fan, 2006). The Y590 is in the sequence motif YXXM, which if phosphorylated could bind the SH2 domain of the p85 regulatory subunit of phosphatidyl inositol 3-kinase (PI3K). Indeed, constitutive phosphorylation of Akt, a downstream substrate of PI3K, was detected in JSRV-transformed rodent fibroblasts; the phosphorylation was blocked by PI3K inhibitors or mutation of Y590 (Albert et al., 2002; Palmarini et al., 2001). However, Y590 has not been found phosphorylated in Env-transformed cells and there is no evidence of p85 binding the JSRV Env. In additionin some cell lines such as the chicken fibroblast line DF-1, transformation by JSRV Env does not absolutely require Y590 (Allen et al., 2001) although transformation is less efficient (Zavala et al., 2003). Akt phosphorylation can be detected in vivo in tumors induced by the related enzootic nasal tumor virus (ENTV), but only in some OPA tumors (Zavala et al., 2003; Suau et al., 2006). Thus the importance of the Y590 residue for JSRV oncogenesis in vivo remains to be determined. JSRV carries an alternate reading frame within the pol gene, designated OrfX (York et al., 1992). This reading frame is conserved in all exogenous JSRV isolates sequenced to date and also in some of the JSRV-related sheep endogenous retroviruses present in normal sheep (Palmarini et al., 2000; Rosati et al., 2000). Caprine ENTV (ENTV-2) also has an open OrfX reading frame, but there are two stop codons in the OrfX reading frame of ovine ENTV (ENTV-1) (Cousens et al., 1999). OrfX has an unusual codon usage, which suggests that the putative protein may be expressed at low levels. Weak amino acid similarity of the putative OrfX protein to the adenosine A3 receptor has been noted (Bai et al, 1999). The OrfX peptide has proven difficult to express in vitro, so it has not been possible to raise antisera that might detect expression of the protein in infected cells or tumors. We detected a 3.2 Kb spliced JSRV RNA in OPA tumors and transiently transfected 293T cells containing the entire OrfX reading frame which could be the mRNA for this protein (Palmarini et al., 2002).

We previously generated a mutant molecular clone of JSRV DNA containing two stop codons in the OrfX reading frame without altering the amino acids in the overlapping pol frame (Maeda et al., 2001). In in vitro transformation assays, this mutant gave the same number of transformed foci in NIH-3T3 cells as the parental wild-type JSRV DNA clone (Maeda et al., 2001). Thus OrfX protein does not appear to be involved in the oncogenic properties of JSRV, but this did not rule out effects on virus replication.

Given the lack of an efficient cell culture system for the propagation of JSRV we cannot test in vitro viral mutants. In these studies, we report the first in vivo tests of site-specific mutants of JSRV in their natural host, the sheep. Our previous isolation of an infectious and oncogenic molecular clone of JSRV, and development of methods to prepare infectious virus from it, made this possible (Palmarini et al., 1999a). In this report, mutants in env (Y590D) and in orfX were studied.

Results

Generation of infectious JSRV mutants in OrfX or env

To test the potential importance of the orfX and the Y590 residue of the JSRV Env in oncogenesis in vivo, it was first necessary to generate infectious JSRVs containing mutations in these genes. We previously isolated a complete integrated JSRV provirus (clone 21) from an OPA tumor and showed that placing the viral sequences under control of the highly active human cytomegalovirus (CMV) immediate early promoter allowed production of infectious virus by transient DNA transfection of human 293T cells (Palmarini et al., 1999a). The availability of this virus allowed us to establish an in vitro infection system for JSRV in sheep choroid plexus (CP) cells and other ovine cell lines (Palmarini et al., 1999b), although infection was inefficient due to the fact that the JSRV is transcriptionally specific for lung epithelial cell lines, and no ovine lung epithelial lines that retain their differentiation properties exist. In vitro, infectivity could be assessed by infection of CP cells, followed by serial passage of the infected cultures and PCR analyses for JSRV DNA at different passages.

We previously described a mutant form of the CMV-driven JSRV proviral plasmid in which two stop codons were created in the orfX reading frame by single base mutations that did not affect the amino acids in the corresponding parts of the integrase protein (pCMV2JS21ΔorfX, Maeda et al., 2001). We showed that pCMV2JS21ΔorfX is capable of transforming NIH-3T3 cells in culture, indicating that the putative OrfX protein is not necessary for transformation (Maeda et al., 2001). We tested if JSRV virus with the orfX mutation is infectious in vitro. JSRV ΔorfX was prepared by transient transfection of 293T cells, and then tested for infection of CP cells (Fig 1A). Compared to CP cells infected with wild-type JSRV, similar amounts of JSRV DNA could be detected in cells infected with JSRV ΔorfX, even after prolonged passage (23 passages). In contrast, infection of CP cells with heat-inactivated viruses resulted in no PCR amplification products, indicating that the positive PCR signals represented genuine viral infection. These results indicated that OrfX protein is not necessary for JSRV replication in vitro, and they provided a means for generation of orfX mutant virus.

Figure 1. Replication capacity of JSRVΔorfX.

Sheep choroid plexus cells (CP-1) were infected with i) heat-inactivated JSRV21, ii) JSRV21, iii) heat-inactivated ΔorfXJSRV or iv) ΔorfXJSRV, and serially passaged. At different passages DNA was extracted from the cells and tested by PCR for the presence of proviral JSRV DNA. Representative PCR results from an intermediate passage are shown (ca. passage 6); PCR signals for both wild-type and ΔorfX JSRV were detectable at equivalent levels through passage 23. No PCR signals were detected in the cultures infected with heat-inactivated viruses. +: PCR positive control, pCMV2JS21 plasmid DNA. -: PCR negative control, water.

We have extensively studied the JSRV Env protein and its role in in vitro transformation of rodent cells (Alberti et al., 2002; Chow et al., 2003; Hofacre & Fan, 2004; Hull & Fan, 2006; Maeda et al., 2002, 2003, 2005; Palmarini et al., 2001; Zavala et al., 2003). For the transformation studies, we used a CMV-driven Env expression plasmid derived from pCMV2JS21 that only encodes Env protein, pCMV3ΔGP (Maeda et al., 2001). Previous site-directed mutagenesis experiments indicated that a YXXM motif (putative binding site for the p85 subunit of phosphatidyl inositol 3-kinase [PI3K]) is essential for efficient transformation in mouse and rat cells (Hofacre & Fan, 2004; Maeda et al., 2005; Palmarini et al., 2001). In particular, mutation of the YXXM tyrosine residue (Y590F or Y590D) abolished or greatly reduced transformation (Palmarini et al., 2001). To test if these Env mutations allowed production of infectious virus, they were transferred from pCMV3ΔGP to pCMV2JS21, to give the plasmids pCMV2JS21- TMY590F and pCMV2JS21-TMY590D. Mutant viruses were then obtained by transient transfection of the proviral plasmids into 293T cells, and they were then tested for replication in vitro in CP cells as described above. Both mutant viruses were capable of replication in CP cells, as evidenced by positive PCR signals in the infected cultures. However, the PCR signal for JSRV TMY590F was lost after 3-4 cell transfers, while the PCR signal for JSRV TMY590D persisted for the same number of transfers as for cultures infected with wild-type JSRV (>6 transfers). These results indicated that the Y590F mutation in TM reduced infectivity of the virus, while Y590D mutation did not appear to show this defect within the limitations of the assay. Therefore, JSRV TMY590D was chosen for in vivo studies.

Generation and quantification of JSRV stocks

Viral stocks of wild-type and mutant JSRVs were prepared by large-scale transient transfection of 293T cells as described in the Materials and Methods. Two preparations of wild-type JSRV were generated, one in parallel with JSRV ΔorfX for one experiment (JS21-04) and one in parallel with JSRV TMY590D for the other experiment (JS21-03). During the transient transfections, the relative amounts of the wild-type and mutant viruses produced on the different days of harvest were monitored by SDS-PAGE and western blotting for JSRV CA protein (26 kDa) (Fig. 2). The amount of JSRV ΔorfX produced by the transfected cells was similar to that for wild-type JSRV (top panels). On the other hand, the amount of JSRV TMY590D was consistently less than that produced by the wild-type JSRV-producing cultures (ca. ten-fold less, lower panels). Therefore, more transfected cultures and higher degrees of concentration were used to generate the JSRV TMY590D stocks; they were adjusted to the concentration of the wild-type JSRV stock based on western blotting for JSRV CA.

Figure 2. Comparison of JSRV dose in the inocula.

a. Western blot comparison of JSRV. 5μl of inoculum run on a 10% SDS-PAGE gel, electro blotted onto nitrocellulose and reacted with 1:200 antiserum JS382 (raised against recombinant JSRV CA protein (Salvatori et al., 2004)). b. qRT-PCR estimation of JSRV copy number. Serially diluted in vitro transcribed RNA was used as standard from which to estimate the RNA copy number of the JSRV in 5ml of inoculum. Error bars shown are the standard error of the mean from four replicates.

Subsequent to the experimental inoculations, we developed a quantitative RT-PCR assay for JSRV RNA (Cousens et al., in preparation), and it was applied to the viral inocula as an independent measure of the amounts of virus inoculated. The results indicated that in the first experiment comparing wild-type and TMY590D JSRV, the amounts of viral RNA in the inocula were equivalent; in the second experiment comparing wild-type and ΔorfX JSRV, the amount of JSRV ΔorfX RNA was approximately five-fold higher than wild-type JSRV RNA (Figure 2). Thus fairly well-matched inocula were used in the experiments. All inocula were checked by RT-PCR amplification and sequencing of the regions encompassing the mutations. For the mutant viruses, the expected mutations were detected, and there was no evidence for the presence of wild-type JSRV in the stocks.

In vivo infection with JSRV TMY590D

In the first experiment, twenty newborn (1-7 day old) lambs from an OPA-free flock were inoculated intra-tracheally with wild-type JSRV (stock JS21-03), and the same number of animals were inoculated with an equivalent amount of JSRV TMY590D (stock Y590D-03). Sixteen of the wild-type JSRV-inoculated animals appeared to show early signs of OPA between 6 and 12 weeks p.i. and were sacrificed and necropsied. OPA was confirmed by histopathology in 14 out of the sixteen animals. The remaining four animals were sacrificed and analyzed at 9 months p.i., and three of them were confirmed to have OPA. Thus in total at least 17/20 of animals inoculated with the cloned JSRV₂₁ stock developed OPA by 9 months. None of the animals inoculated with JSRV TMY590D showed symptoms of OPA over the course of the experiment (9 months). One animal was sickly and was euthanized at 3 weeks of age. Eight animals were sacrificed and necropsied at 6 months p.i., and the remainder at 9 months. None of the animals showed evidence of OPA upon histological examination. This would be consistent with the TMY590D mutation abolishing in vivo oncogenicity for the virus. To test if JSRV TMY590D could replicate in vivo, we tested blood at various times post-inoculation and mediastinal lymph nodes at post-mortem for the presence of viral DNA by PCR (Salvatori et al., 1999). We have previously found that low levels of infected cells in the blood can be detected by PCR in animals experimentally inoculated with JSRV (Holland et al., 1999; Palmarini et al., 1996; Salvatori et al., 1999) and this can be taken as an indicator of productive in vivo infection. Likewise, some blood samples from animals naturally-infected with JSRV test positive by PCR, but this is highly variable when longitudinal samples from the same animal are tested (D. Salvatori PhD thesis; Caporale et al., 2005; De las Heras, et al., 2005; Gonzalez, et al., 2001). Even though sequencing confirmed the presence of only the Y590D mutation in the input inoculum, when PCR amplification products from the env gene from the blood of the infected animals were sequenced, in all cases they showed sequences corresponding to wild-type JSRV env and not the Y590D mutant. One possible explanation could be contamination during the PCR amplification, even though results were only considered when appropriate preparation and PCR control samples were negative. Another explanation could be that the Y590D mutant virus was defective for replication in vivo and there was in vivo selection for reversion of the Y590D mutation to wild-type Env. However, PCR from Y590D-infected lambs yielded wild-type Env even in blood taken 1 day post-infection. This seems too short a time for the appearance of wild-type revertants. Thus contamination during the PCR amplification must be seriously considered although we cannot rule out infection by wild type JSRV of some of our housed sheep although no OPA cases have been recorded in the flock since it was purchased 7 years ago.

To try to find conclusive evidence of infection with JSRV TMY590D, samples of mediastinal lymph node (MLN) collected at post mortem examination were tested by PCR for the presence of JSRV. None of 20 TMY590D-inoculated animals tested contained detectable JSRV in MLN samples. Taken together with the PCR amplifications on infected animal blood, there was no evidence for infection in vivo by JSRV TMY590D.

In vivo infection with JSRV ΔorfX

In the second experiment, ten newborn lambs were inoculated intratracheally with wild-type JSRV (stock JS21-04) and twelve animals were inoculated with JSRV ΔorfX (stock ΔorfX-04). The clinical pictures of both groups were very similar. Early clinical signs were first noted in both groups at around 7 weeks of age. All the animals were culled over the next 8 weeks as clinical signs became more noticeable. The last animals were culled at 15 weeks p.i. Histopathology confirmed OPA in all 10 JSRV₂₁-inoculated animals and in 11 of the 12 animals inoculated with JSRV ΔorfX. There was no discernable difference in the pathology caused by the two viruses (Fig. 3). Lung fluid was collected at post mortem from 7 of the 12 animals inoculated with JSRV ΔorfX, and 7 of the 10 inoculated with JSRV₂₁. The volumes collected varied from 2 to 40 ml. By comparison, lung fluid collection from naturally occurring field cases of OPA may be from 0 to 500ml of lung fluid per day. Large volumes of lung fluid are generally associated with more extensive tumor lesions, while the experimentally inoculated animals typically had multiple small lesions.

Figure 3. Histology of experimentally-induced OPA.

Haematoxylin & eosin stained sections of OPA induced by a) JSRV21 and b) JSRVΔorfX show indistinguishable tumours composed of papillary proliferation of cuboidal epithelial cells in the lung parenchyma irrespective of whether they were induced by the wild-type or Δorfx virus.

To test if the inoculated JSRV ΔorfX mutant virus itself had induced disease and was being produced by the tumour, RNAs from lung fluids from two animals inoculated with this virus were extracted and amplified by RT-PCR. Sequencing of the amplicons indicated the presence of both mutations in the JSRV ΔorfX virus (Table 2). As expected, equivalent analysis of lung fluids from two animals inoculated with wild-type JSRV₂₁ showed the predicted wild-type sequences. These results therefore indicate that mutation of the orfX reading frame does not interfere with either in vivo replication or oncogenesis by JSRV.

Table 2.

Sequence of JSRV around orfX point mutations.

Source	Sequence at 5′ site	Sequence at 3′ site
	♦	♦
JSRV21	CCTCAAACCCTTAAGACAGA	TTTGGACGTCTTAAATATGT
ΔorfX	CCTCAAACCCTAAAGACAGA	TTTGGACGTCTAAAATATGT
enJSRVa	CCTCAAACCCTTAAGACAGA	TTTGGACGTCTTAAATATGT
A357b-LF	CCTCAAACCCTAAAGACAGA	TTTGGACGTCTAAAATATGT
A385 b-LF	CCTCAAACCCTAAAGACAGA	TTTGGACGTCTAAAATATGT
A361c-LF	CCTCAAACCCTTAAGACAGA	TTTGGACGTCTTAAATATGT
A362c-LF	CCTCAAACCCTTAAGACAGA	TTTGGACGTCTTAAATATGT

^aJSRV-like endogenous sheep retrovirus.

^bLung fluid (LF) produced by animals inoculated with JSRV-ΔorfX.

^cLung fluid (LF) produced by animals inoculated with JSRV₂₁.

Comparative quantitation of the inocula used for the in vivo experiments is shown in Figure 2 and a summary of the outcome of these inoculations is shown in Table 1.

Table 1.

Summary of results of in vivo inoculation with JSRV.

Inoculum	No. animals inoculated	No. animals with OPAa in ≤9mo.	No. animals producing LF	Clinical signs first noted in some animals
JSRV21-03	20	17	5	6 wk
TMY590D-03	20	0	0	n/a
JSRV21-04	10	10	6	7 wk
ΔorfX-04	12	11	8	7 wk

^aAnimals noted with OPA lesions at histopathological examination of post mortem lung samples.

Discussion

In the studies reported here, we tested for the first time the biological activities of JSRV mutants in vivo in their native host, the sheep. These experiments were facilitated by the isolation of an infectious molecular clone of JSRV (JSRV₂₁), and development of the transient transfection approach for recovering infectious virus (Palmarini et al., 1999a). Given the logistic and cost challenges of performing in vivo experiments on large animals, the use of an in vitro infectivity assay as a preliminary screen for replicative ability of viral mutants was also important. These experiments describe in vivo tests of JSRV mutants in the YXXM motif in the cytoplasmic tail of the envelope TM protein (important for efficient transformation or rodent cells in vitro), and in the orfX reading frame that is conserved among all JSRV isolates but for which a function has not been assigned.

In the first experiment, we tested if mutation of the TM cytoplasmic tail YXXM motif led to loss of tumorigenesis for JSRV, as might be predicted from the in vitro transformation experiments. Indeed, lambs inoculated with JSRV TMY590D did not develop OPA, which would be consistent with this hypothesis. The Y590D mutant was selected because in the in vitro infection assay, it showed replication equivalent to wild-type JSRV while the Y590F mutant showed a replication defect. If animals infected with JSRV TMY590D showed evidence for infection but no tumors, this would have strongly supported the importance of the YXXM motif in the TM cytoplasmic tail in lung tumorigenesis. The input inoculum was confirmed to contain the Y590D mutation but the only virus that could be detected in these animals was of wild-type sequence and it is therefore questionable whether this represents genuine infection. Reversion would be relatively feasible, since the Y590D mutation was due to a single base change, but wild type sequence was detected as early as day 1 p.i., so sample contamination during the PCR must be strongly considered. A possible explanation for our inability to detect virus with the Y590D mutation in these animals is that it did not efficiently replicate in vivo. As shown in Figure 2, the amount of Y590D virus produced from transiently transfected 293T cells was substantially less than from cells producing wild-type virus. This would be consistent with a replication defect in vivo, even though within the limits of the assay, the in vitro infectivity assay did not indicate a defect. In the related betaretrovirus Mason-Pfizer monkey virus (MPMV), the cytoplasmic tail of TM contains a tyrosine residue in the analogous location to the Y590 for JSRV TM, and mutation of the MPMV tyrosine results in defects in transport/release of cytoplasmic viral core particles from the cells (Blot et al., 2006) At the very least, the experiments with JSRV TMY590D indicate that the Y590 residue is essential for tumorigenesis and/or replication of JSRV. It may be that cell transformation or enhanced cell division is necessary for efficient production and spread of JSRV. In the future it might be interesting to test other non-transforming TM cytoplasmic tail mutants to see if they are not tumorigenic. We have carried out alanine scanning mutagenesis of the cytoplasmic tail, and identified quite a number that do not transform cells in vitro (Hull and Fan, 2006). We have generated pCMV2JS21 plasmids containing each of these mutations and found that the majority of them (unlike JSRV TMY590D) produce the same amounts of virus during transient transfections of 293T cells. It will be interesting to test if any of the JSRVs with non-transforming mutations in the TM cytoplasmic tail show replication equivalent to wild-type JSRV in the in vitro infection of CP cells. If a mutant with a non-transforming env gene shows normal levels of virus production and replication in vitro, it is possible that such a mutant would show in vivo replication but no tumorigenesis. We also identified alanine substitutions in the TM cytoplasmic tail that showed enhanced transformation in vitro (Hull and Fan, 2006). It would be interesting to investigate if such mutants show enhanced tumorigenicity in animals.

The goal of the in vivo test of JSRV TMY590D was to determine if the JSRV Env protein is functioning as an oncogene in JSRV-induced OPA. After these experiments were initiated, two experiments have strongly confirmed this hypothesis. Wootton, et al., (2005) showed that infection of immunodeficient mice with an adeno-associated virus (AAV)-based vector expressing the JSRV Env protein developed lung adenocarcinomas with histopathology reminiscent of OPA. In addition, we generated a JSRV-based vector expressing only the Env protein, and showed that inoculation of the vector into newborn lambs resulted in development of OPA (Caporale et al., 2006).

The experiment with JSRV ΔorfX showed that this virus induced OPA with a similar time course and histopathology compared to wild-type JSRV (Figure 3). This confirmed the indications from the in vitro experiments that orfX is not required for in vivo replication or transformation by JSRV. However, it was still possible that a truncated OrfX protein could be contributing to these processes, since JSRV ΔorfX could in principle encode a protein of 85aa. The mutations were chosen as these were sites where stop codons could be introduced into orfX without altering the amino acid sequence of the overlapping pol gene. Interestingly, ovine enzootic nasal tumor virus (ENTV-1) has stop codons in exactly the same places as those engineered into JSRV ΔorfX (Cousens et al., 1999). These experiments provided the first large-scale evaluation of the in vivo tumorigenicity of molecularly cloned JSRV. Originally, induction of JSRV-induced OPA employed JSRV inocula derived from concentrated lung fluids from animals with OPA (Martin et al., 1976; Salvatori et al., 2004; Summers et al., 2002 & 2005). These previous studies showed induction of OPA within weeks-to-months, although in some cases induction of disease has been less efficient (J.M. Sharp, personal communication). In a previous infection with cloned JSRV₂₁, 2 out 4 animals developed OPA after 4 months (Palmarini et al., 1999a). This might have suggested that the cloned JSRV was less oncogenic or infectious than lung fluid-derived JSRV. Therefore in the first experiment described here, 20 animals were infected with wild-type JSRV₂₁ as the positive control. As described, 17/20 of the JSRV₂₁-inoculated lambs developed OPA. Consequently, in the second experiment, a reduced number of animals were infected with JSRV₂₁ or JSRV ΔorfX. Combining the results with JSRV₂₁ in the first and second experiments, it can be predicted that 92.5 ± 10.6 % of animals (at least of the age and breed utilised in this study) will develop OPA when inoculated with the concentration of JSRV₂₁ used here under these experimental conditions. Although quantification of viral RNA in the inocula did not necessarily measure infectious virus, it appeared to be a reasonable way of equalizing virus dose within and between experiments. The results from these large scale in vivo experiments will allow better predictions of outcomes in future studies, resulting in using smaller groups of animals while maintaining statistical power, in line with the aims of the 3Rs in animal experimentation (Flecknell, 2002; Russell and Burch, 1959). Post mortem examinations showed that OPA tumors with discrete lesions had developed in the wild-type and JSRV ΔorfX-inoculated animals within a few weeks of infection. On the other hand, two animals suspected of having early physical symptoms of OPA did not show evidence of tumors by histopathology. This may have been because we were looking for subtle early signs of OPA which may be similar to symptoms of other lung infections. A recent field study also found that it was not possible to identify early stage OPA from clinical signs, and even late stage OPA may be misdiagnosed prior to post-mortem examination (C. Cousens et al., In Press).

In summary, this study demonstrated the efficacy of the JSRV₂₁ experimental OPA system as a model to test JSRV pathogenicity in vivo. We showed that a complete OrfX is not required for pathogenesis of JSRV. We also showed that mutation of Y590, part of the YXXM motif of envTM which is predicted to be involved in transformation by JSRV, appears to abrogate the ability of the virus to replicate and/or cause tumours in vivo.

Materials and Methods

Vector constructs

Generation of the infectious molecular clone pCMV2JS21 has been described previously (accession number AF105220) (Palmarini et al., 1999). The TM mutant, JSRV₂₁TM-Y590D, has a single nucleotide substitution changing the tyrosine at position 590 to aspartic acid (Palmarini et al., 2001). The ΔorfX mutant of pCMV2JS21 has been described previously (Maeda et al., 2001); two T to A substitutions in the orfX region (nt 4565-5104) generate two stop codons in this reading frame (at 4821 TTA to TAA and at 4952 TTA to TAA) without altering the amino acid sequence encoded by the overlapping region of the pol gene.

In vitro virus production

Human 293T cells (3.2 ×10⁶ cells per 10cm plate) in 10 ml of D-MEM/10%FBS/Penicillin/Streptomycin were seeded approximately 24 hours before transfection. Cells were transfected with 28μg per plate pCMV2JS21 (Palmarini et al 1999a) using the CalPhos Mammalian Transfection Kit (Clontech) as per manufacturer's instructions. After 12 hours the transfection medium was replaced with 5 ml culture medium. Culture medium was harvested 24, 48 and 72 hours later. At each time point the supernatant was filtered (0.45 μM filters) and virus was concentrated by centrifugation (100,000g for 1 hour at 4°C) on 50% and 25% glycerol cushions as described previously (Palmarini et al 1999a). Pellets were resuspended in 0.1MNaCl, 0.01M Tris (pH 7.5), 1mM EDTA (one hundredth original volume) and stored at -70°C. Immediately prior to inoculation the concentrated stocks were pooled and adjusted to 66ml with PBS.

In vitro infection

Sheep choroid plexus cells (CP-1) were infected with wild-type or ΔorfX mutant JSRV produced from transfected 293T cells as described previously (Palmarini et al., 1999a). In parallel, virus that had been heat-inactivated (15 min at 65°C) was also infected into CP-1 cells. The infected cell cultures were serially passaged, and at different passages were DNA from cell extracts was tested for the presence of exogenous JSRV DNA by PCR employing primers P-I (TGGGAGCTCTTTGGCAAAAGCC) and P-III (CACCGGATTTTTACACAATCACCGG) (Palmarini et al., 1996)

Experimental infection

JSRV TMY590D

One to seven-day-old lambs from a known OPA-free flock were inoculated with 5ml of virus stock by the intratracheal route. Twenty lambs each received the equivalent of ten 10 cm plates per animal (all 3 harvests) of JSRV₂₁. Twenty lambs each received equivalent amounts of JSRV₂₁ TMY590D based on western blots for viral CA protein. Aliquots of the inocula were stored at -70°C for subsequent analysis. Animals were monitored every 2-3 days for early signs of OPA by observing respiration at rest and after exercise. Signs of abnormal breathing and/or slow recovery time were noted. Animals were euthanised when early OPA was suspected or at 6 or 10 months old. Blood samples were collected at 1 day, 1 week and 1 month post infection and at monthly intervals thereafter. Samples of blood, lung fluid (if present), lung tumour, normal and mediastinal lymph node were collected at necropsy for histopathological examination and/or for PCR analysis.

When available, lung fluid was collected by raising the animal's rear end and collecting the fluid dripping from its nostrils. The lung fluid was prepared as follows; mucus was removed by filtering through sterile gauze and cells and debris were removed by centrifuging 1ml aliquots in a microfuge for 5 minutes, discarding the pellet and repeating the microcentrifugation. 100μl aliquots of clarified lung fluid were stored at -70°C.

JSRVΔOrfX

Two to four-day-old lambs from a known OPA-free flock were inoculated with 5ml of in vitro produced virus by the intratracheal route. 10 lambs received the equivalent of seven and a half 10 cm plates (all 3 harvests) of JSRV21. Twelve lambs received JSRV21ΔorfX harvested from nine plates. Aliquots of the inocula were stored at -70°C for subsequent analysis. Animals were monitored and lung fluid was collected as above.

PCR on lymph nodes

The QIAgen tissue kit was used (as manufacturer's instructions) to extract DNA from lymph nodes collected at post mortem. PCR tests were performed in triplicate as previously described (Salvatori et al., 2004). PCR conditions were as follows: 100ng of test DNA with 5μl of 10X PCR Buffer (Qiagen), 200μM dNTPs (Roche Diagnostics Ltd, East Sussex, UK), 200 μM of each primer P-I and P-III and 1.25U of Hot Start Taq polymerase (Qiagen) in a final volume of 50μl. Cycle conditions were 94°C for 16 minutes, followed by 35 cycles at 94°C for 30 seconds, 59°C for 30 seconds and 72°C for 30 seconds, with a final extension step of 5 minutes at 72°C. PCR products were visualized following electrophoresis in a 2% agarose gel containing 100μg-ml ethidium bromide. If any one of the triplicate reactions generated a product of the correct size the test was considered JSRV positive.

Quantitation of the inocula

The viral RNA copy numbers in the inocula were compared using quantitative real-time PCR (Cousens et al. in preparation). Briefly, virus was prepared using the QIAgen viral RNA kit according to the manufacturer's instructions and eluting with 100μl of elution buffer. Single step RT-PCR was conducted as follows using reagents from the Applied Biosystems real-time one step RT-PCR kit; (1× reaction buffer, 1 × RT, 10μM each dNTP), plus 200nM each primer P-I and P-III, 200nM fluorogenic probe JS-T (5′-FAM-AGCAAACATCCGARCCTTAAGAGCTTTCAAAA-TAMRA-3′), and 4μl of test or standard RNA template in a 50μl reaction volume. Amplification, data analysis and data acquisition were carried out on an ABI Prism SDS 7000 (Applied Biosystems). Cycling conditions were as follows: 48°C for 30 minutes, 94°C for 10 minutes followed by 40 cycles of 95°C for 15 seconds, 59°C for 30 seconds and 60°C for 30 seconds.

Sequence analysis of inocula and experimental samples

RT-PCR followed by sequencing was used to check the sequence of JSRV in the inocula and in virus produced in lung fluid of the animals with OPA. Fifty microlitres of clarified lung fluid or of inoculum was treated with DNase as follows: 0.1mM sodium acetate, 2.5mM MgCl₂ 5mM Tris pH8.5, 2μl DNase (RNase free) (Roche) in a total volume of 140μl, incubated at 30°C for 1 hour then 65°C for 10 minutes. RNA was then extracted using the QIAgen viral RNA kit according to the manufacturer's instructions and eluted in 50μl. Four microliters of RNA was used for RT-PCR as described below. The product was gel purified and recovered using the QIAgen gel purification kit according to the manufacturers instructions. Sequencing was done on a MegaBACE 500 using ET terminator chemistry. The data was processed using the Cimarron 3.12 basecaller.

JSRV TMY590F PCR

Hemi-nested PCR was conducted using primers TMfo (7004-7034 JSRV21) AAGATGAGAGTTGAAATGCTGCATATG and TMr (complement of 7225-7247) GTAGGAATACTTGTGTTTATTATAATTGTCG in the first round and TMfi (7086-7101 JSRV 21) GTAYTTGTCCTAGGCTTTGGCTT and TMr in the second round, to amplify the region of JSRV containing the Y590D substitution. Reaction conditions were 1× reaction buffer (QIAgen) plus 0.5mM MgCl₂, 0.2mM each dNTP, plus 200nM each primer in a 50μl reaction volume. Cycling conditions were as follows: 48°C for 30 minutes, 94°C for 10 minutes followed by 35 cycles of 95°C for 30 seconds, 57°C for 30 seconds and 72°C for 60 seconds, with a final extension of 5 minutes at 72°C. A number of the amplimers were sequenced. These were generated in separate RT-PCR runs from at least 2 independent DNA extractions for each sample and were sequenced individually. To check the sequence of the JSRV TMY590D inoculum RT-PCR was conducted using primers TMf1 and TMr with the one-step Superscript III RT-PCR kit (Invitrogen) according to the manufacturer's instructions, and 4 μl of RNA as target. Cycles were; 50°C 15 min, 95°C 2 min then 35 cycles of 95°C 15s, 57°C 30s.

OrfX PCR

DNA fragments spanning the orfX mutations were generated with primers orfx-f1 TCTGGCAAACAGATGTTACTCAC (4777-4797 of JSRV₂₁) and orfx-r2 GGCTAGTATACCCAGGTCC (C′ of 4966-4984) using the one-step Superscript III RT-PCR kit (Invitrogen) according to the manufacturer's instructions, and 4 μl of RNA as target. Cycles were as follows; 50°C 15min, 95°C 2 min then 35 cycles of 95°C 15s, 58°C 30s.