Reticuloendotheliosis Virus Sequences within the Genomes of Field Strains of Fowlpox Virus Display Variability

Abstract

Nine field strains of fowlpox virus (FPV) isolated during a 24-year span from geographically diverse outbreaks of fowlpox in the United States were screened for the presence of reticuloendotheliosis virus (REV) sequences in their genomes by PCR. Each isolate appeared to be heterogeneous in that either a nearly intact provirus or just a 248- or 508-nucleotide fusion of portions of the integrated REV 5′ and 3′ long terminal repeats (LTRs) was exclusively present at the same genomic site. In contrast, four fowlpox vaccines of FPV origin and three originating from pigeonpox virus were genetically homogeneous in having retained only the 248-bp LTR fusion, whereas two other FPV-based vaccines had only the larger one. These remnants of integrated REV presumably arose during homologous recombination at one of the two regions common to both LTRs or during retroviral excision from the FPV genome. Loss of the provirus appeared to be a natural event because the tripartite population could be detected in a field sample (tracheal lesion). Moreover, the provirus was also readily deleted during propagation of FPV in cultured cells, as evidenced by the detection of truncated LTRs after one passage of a plaque-purified FPV recombinant having a “genetically marked” provirus. However, the deletion mutants did not appear to have a substantial replicative advantage in vitro because even after 55 serial passages the original recombinant FPV was still prevalent. As to the in vivo environment, retention of the REV provirus may confer some benefit to FPV for infection of poultry previously vaccinated against fowlpox.

The continual exposure of poultry to a variety of advantageous pathogens can provide an environment for distinct entities to encounter each other. In this regard, chickens have been diagnosed as being concurrently infected with both fowlpox virus (FPV) and the avian herpesvirus infectious laryngotracheitis virus (8, 25). Although no genetic exchange between these viruses has been documented, it should be noted that homologs of several FPV genes are present in the genome of another avian herpesvirus, Marek's disease virus (4). Moreover, short nucleotide stretches exhibiting >70% homology with the R and U3 regions of the long terminal repeat (LTR) of the avian retrovirus reticuloendotheliosis virus (REV) have been found in the DNA of two serotype I oncogenic strains of Marek's disease virus (10). Likewise, remnants of the REV LTR have been retained in the genomes of all FPV vaccine strains so far examined (9, 15, 22).

Of more concern is the demonstration that nearly intact REV provirus has remained as an integral part of the DNAs of both Australian and United States field strains of FPV (6, 9, 13, 22) and also in the now discontinued Australian FPV standard (S) vaccine strain (9). Interestingly, the 3′ LTR of the provirus in the genome of the S vaccine strain, and presumably in those of the field strains, has undergone rearrangement and lacks part of the U3 and U5 regions and all of the R region (9). Despite these alterations, the integrated retroviruses have maintained infectivity, as evidenced by the generation of anti-REV antibodies (9, 22) and the detection of REV (6, 9) in chickens infected with FPV field strains and the Australian S vaccine strain.

Although the genesis of REV insertion into the FPV genome is unknown, this event occurred at least once more than 50 years ago (13). Whether there was only one progenitor or others resulting from independent insertions is also a point of conjecture. However, it should be noted that this association with a retrovirus may be novel in avipoxviruses, since no REV sequences have been found in the DNAs of either canarypox virus, quailpox virus, or a field isolate of pigeonpox virus (13, 15). Of equal interest is the origin of the attenuated strains currently used as vaccines. Presumably, these have arisen from FPV field strains that have lost the provirus, as their genomes contain only a partial and rearranged LTR (9) or individual LTR (15). Although such deletion mutants have also been detected in field virus populations (9, 13, 22), the continuous presence of an intact, integrated retrovirus implies that REV expression may be advantageous to their survival and ability to successfully infect poultry. In fact, to date only one field virus isolated in 1956 has been found to contain only an LTR remnant, not the entire provirus, in its genome (3). Moreover, even immunized flocks in geographically diverse areas of the United States have not been entirely resistant to subsequent infection by REV-containing FPV (22).

Previous characterizations of FPV field strains endemic to the United States have relied primarily on the use of PCR to consistently demonstrate the presence of REV provirus in the genome of each virus with the exception of one isolated in 1956 (13, 22). A similar screening of vaccines used in this country demonstrated their genetic homogeneity in having only one truncated or complete LTR (15). The former is identical in nucleotide sequence to the LTR remnant found in the genome of the mild Australian vaccine virus (9), while the latter represents a previously undiscovered form. Based on a comparison of the sizes of the amplicons obtained with poxvirus DNA as the template, at least a portion of each United States FPV field isolate appeared to have retained the incomplete LTR (13, 22). Although a PCR product presumably corresponding to the single, intact LTR was generated from the genome of a vaccine strain (22), its existence was not detected in any of the field viruses examined. Thus, currently, a potential, natural source of FPV containing an individual LTR has not been identified.

Since outbreaks of fowlpox in the United States have continued to occur despite routine vaccination against this disease, the issue of REV integration in the genomes of FPV field isolates is being revisited. In contrast to a previous study (22) involving just four viruses, the number of isolates from previously immunized flocks was doubled to more accurately reflect their possible diversity. Moreover, a field isolate obtained approximately 20 years earlier than the rest as well as nine viruses used to vaccinate chickens against fowlpox were also analyzed. In addition to the intact REV provirus and LTR remnant previously detected in the DNAs of field viruses, a novel, larger LTR remnant was also found at the same genetic location in all nine field isolates. Both incomplete LTRs appeared to have resulted from homologous recombination between the 5′ and 3′ LTRs of the provirus, and one or the other was exclusively present in each vaccine virus's genome.

MATERIALS AND METHODS

Viruses and cells.

FPV field strains associated with outbreaks of fowlpox occurring between 1992 and 1999 in previously vaccinated chicken flocks located in the United States and in an unvaccinated flock at the University of Illinois in 1974 (24) were examined in the present study. These viruses were designated based on the geographical site and year of the disease occurrence as follows: IL-74 (Illinois, 1974); MN-92 (Minnesota, 1992); NE-92 (Nebraska, 1992); PA-1996 (Pennsylvania, 1996); MN-97 (Minnesota, 1997); CA-98 (California, 1998); NY-98 (New York, 1998); OK-98 (Oklahoma, 1998); and GA-99 (Georgia, 1999). For comparison, the fowlpox challenge virus utilized by the Animal and Plant Health Inspection Service (APHIS) at the Center for Veterinary Biologics (Ames, Iowa) and nine commercially available fowlpox vaccines of FPV or pigeonpox virus origin were also included. The latter were obtained from the indicated manufacturers and randomly assigned letters as follows: VAC-1 (TCP-Blen; CEVA Laboratory, Overland Park, Kans.); VAC-2 (Chick-n-Pox; Fort Dodge Animal Health, Fort Dodge, Iowa); VAC-3 (FP-VAC; Intervet Inc., Millsboro, Del.); VAC-4 (fowlpox; Mallinckrodt Veterinary Inc., Millsboro, Del.); VAC-5 (Chick VI Pox; Vineland Laboratories, Vineland, N.J.); VAC-6 (fowlpox; Intercontinental Biologics, Millsboro, Del.); VAC-7 (pigeonpox; Intervet Inc., Millsboro, Del.); VAC-8 (Schering-Plough Animal Health, Omaha, Neb.); and VAC-9 (Select Laboratories, Gainesville, Ga.).

Virus present in cutaneous or tracheal lesions of afflicted chickens was initially propagated in the chorioallantoic membranes (CAMs) of embryonating chicken eggs as described by Singh et al. (22). After confirmation of the suspected etiology by histopathological examination of the resultant pocks on the CAMs (26), attempts were made to grow the viruses in the QT-35 quail cell line as previously described (17). Of the nine FPV field strains, only NY-98 was not successfully propagated in these cells.

DNA isolation.

DNA was isolated from lesions on the comb or trachea of FPV-infected chickens and from pocks on virus-infected CAMs as previously described (22). Briefly, after maceration of the tissue with either a mortar and pestle or a glass homogenizer in the presence of TE buffer (10 mM Tris [pH 8.0], 1 mM EDTA), the ground material was clarified at 1,500 × g and 4°C for 10 min. Supernatants were incubated in an equal volume of extraction buffer (10 mM Tris-HCl [pH 7.4], 100 mM NaCl, 10 mM EDTA, 0.5% sodium dodecyl sulfate, 2% β-mercaptoethanol, 0.125% proteinase K) for 2 h at 55°C. DNA was extracted from the digested samples in the presence of a phenol-chloroform (48:2) mixture, ethanol precipitated, and then resuspended in TE buffer. This procedure was also used to obtain DNA from lyophilized FPV and pigeonpox virus vaccine preparations after their rehydration in TE buffer (1 ml per vial).

DNA was also isolated from virus nucleocapsids that had been obtained from infected QT-35 cell monolayers as described by Schnitzlein et al. (17). In this case, extraction of the DNA involved the use of DNAzol (Invitrogen Life Technologies, Carlsbad, Calif.) as per the manufacturer's instructions.

PCR amplification.

Dual amplifications of a 1,222-bp region of FPV open reading frame (ORF) 108 encoding a major antigenic envelope protein (5) and a 282-bp region of the 5′ long terminal repeat (LTR) of the REV retrovirus integrated into the FPV genome were conducted with primers FPV.G1 and FPV.G2 and primers REV.F and REV.R (Fig. 1A; Table 1), respectively, as described previously (22) except that the primer annealing temperature was increased from 56°C to 58°C. Likewise, this modification was applied to single amplifications (22) involving primers TR-1 and TR-2 (Fig. 1; Table 1), which correspond to regions in FPV ORFs 201 and 203, respectively, flanking the REV integration site in the FPV genome, as well as to the protocol (22) used to generate a 642-bp portion of the integrated REV envelope (gp73) gene in conjunction with primers Renv-1 and Renv-2 (Fig. 1A; Table 1). This modified procedure was also used for amplification of the variable-sized FPV ORF 168 (3, 21, 23) in the presence of primers 39K.1 and 39K.2 (21) (Fig. 1; Table 1).

FIG. 1.

Schematic representation of REV provirus integrated into the genomes of FPV field and vaccine strain viruses. The relative positions of PCR primers used for amplification of the entire or portions of the intact provirus (A) as well as of the long (B) and short (C) REV LTR remnants present in the FPV genome are shown.

TABLE 1.

PCR primers designed for amplification of portions of the FPV genome and/or the REV provirus

Primer	Specificitya	Nucleotide sequence
FPV.G1	FPV	CAT ACA TTA CTC TTA ATT CGT TTC
FPV.G2	FPV	TTG TAA CTG TTC TAT TAG TGC C
39K.1	FPV	CAG GAA TTC GCT GAG AAC TTC CAC A
39K.2	FPV	TTC CAG CTC GAG TTA AGG AAT AAT AGC
TR-1	FPV	AAC AAT GAT ACG TCT CTT CC
TR-2	FPV	CAC ACG AAT ATA CCA ATA AGG
REV.F	REV	CAT ACT GGA GCC AAT GGT T
REV.R	REV	AAT GTT GTA GCG AAG TAC T
REV-2	REV	TCA GAT TGG CAG TGA GAG
Renv-1	REV	TCG ATT GCG GTA GCT CCA C
Renv-2	REV	CCA TCG AGA GTG ACA TTG C
Renv.F	REV	GGA ATT CCT GAC AAC CAA GAA GAA TGG
Renv.R	REV	CCT CGA GGC TTG ACC TAG GGT ATC CAT CTC

^aSite of primer binding, FPV genome or REV provirus.

For both single and dual long-range PCR amplifications, the Elongase amplification system (Invitrogen) was used. The 50-μl PCR mixtures consisted of 60 mM Tris-SO₄ (pH 9.1), 18 mM (NH₄)₂SO₄, 1.5 mM MgSO₄, a 0.2 mM concentration of each of the four deoxynucleoside triphosphates, 400 nM each primer, 2 μl of DNA, and 1 μl of Elongase enzyme mix. After initial denaturation for 30 s at 94°C, the DNA was amplified for 35 cycles of 30 s of denaturation at 94°C, 30 s of annealing at 56°C, and 9 min of extension at 68°C, with a final extension step of 10 min at 68°C. Included in some reactions together with primer TR-2 was a novel forward primer, REV-2 (Table 1), which encompasses the R and U5 juncture of the 5′ LTR of the integrated REV (9) and was designed via the Prime program of the Wisconsin package (version 9.1; Genetics Computer Group, Madison, Wis.). An additional pair of primers, Renv.F and Renv.R (Fig. 1A; Table 1), was used to amplify the entire envelope gene of the integrated provirus. Originally developed for cloning the REV envelope gene based on the sequence of the provirus present in the genome of FPV UI-74, these primers have EcoRI and XhoI restriction sites at their respective 5′ termini. The forward primer is located within 18 nucleotides upstream of the initiation codon of the envelope gene, while the reverse primer is positioned immediately upstream of the termination codon of this gene.

Determination of nucleotide sequence of PCR products.

Amplicons in completed PCRs were precipitated in the presence of linear polyacrylamide (Sigma, St. Louis, Mo.), reconstituted in TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 8.0), and then electrophoresed in 1.0% low-melting-point agarose gels. Products were excised from the gels, purified with a Zymoclean gel DNA recovery kit (Zymo Research, Orange, Calif.), and then sequenced with a Big Dye Terminator version 2.0 cycle sequencing ready reaction kit (Applied Biosystems, Foster City, Calif.).

Gel electrophoresis and Southern blotting.

All amplicons generated by PCR amplification were analyzed in agarose gels. Likewise for Southern blotting, the genomes of various FPV isolates were electrophoresed in a 0.8% agarose gel after digestion with BglII. The separated DNA fragments were transferred under alkaline conditions (13) to a Zeta probe nylon membrane (Bio-Rad, Hercules, Calif.) and then fixed during exposure to UV light in a Bio-Rad cross linker. The membrane-bound DNA was then hybridized in Perfect-Hyb solution (Sigma, St. Louis, Mo.) with a radioactively labeled portion of the 5′ LTR of the integrated provirus. Annealed probe was detected by autoradiography with X-ray film.

Generation of the 390-bp SstI-HindIII REV LTR probe consisted of its excision from a plasmid member of a HindIII genomic library of the FPV IL-74 strain, labeling with [³²P]dCTP with a random primer labeling kit (Invitrogen), and then purification by passage through a Centri^.spin CS-20 column (Princeton Separations, Adelphia, N.J.).

Generation of recombinant FPV field strain expressing GFP.

A recombinant FPV having an expressible green fluorescent protein (GFP) gene inserted into the envelope (gp73) gene of the integrated provirus was produced as described below. First, the vaccinia virus P11 promoter was excised as an XbaI-BamHI fragment from pVBX5 (18) and inserted into a BamHI-linearized pEGFP vector (Clontech, Palo Alto, Calif.) which had been partially digested with XbaI. The resultant plasmid has the GFP gene positioned in frame with the retained, upstream portion of the poxvirus gene. The entire transcriptional unit was then released by XbaI digestion of the plasmid, and its termini were blunt ended with Klenow polymerase. The modified DNA was then ligated into a DraI site located within a cloned copy of the envelope gene of the integrated REV. The resultant transfer vector, having approximately 600 and 1,000 bp of provirus DNA flanking the GFP gene, was transfected into MN-97 FPV-infected QT-35 cells in the presence of Lipofectamine (Invitrogen). Progeny-infected cells that contained GFP were separated from nonfluorescing cells with a Cytomation MoFlo sorter (Cytomation Inc., Fort Collins, Colo.). Afterwards, GFP-expressing virus (MN-97-Renv-GFP) was identified and purified seven times by selecting plaques exhibiting fluorescence when viewed through a 470- to 490-nm excitor filter in a fluorescent microscope.

Identification of loss of REV provirus from recombinant FPV expressing GFP.

A 100-fold dilution of the progeny from the original infection with plaque-purified MN-97-Renv-GFP and from each successive passage was propagated in monolayers of approximately 2 × 10⁶ QT-35 cells. After every fifth round, the resultant progeny were further diluted prior to being screened for the presence of potentially provirus-deficient FPV through examination of infected QT-35 cell monolayers for individual nonfluorescent plaques. Viruses predicted to have lost the provirus were then visually evaluated during two consecutive plaque purifications. Those still exhibiting the nonfluorescent phenotype were genotypically verified based on the ability of their genomes to serve as template for amplification of only the FPV, not the REV, envelope gene.

RESULTS

Variability in REV proviral nucleotide sequences present in genomes of FPV field and vaccine strains.

To screen for an association between REV and FPV, PCRs designed to amplify either a portion of an REV proviral LTR (2) or envelope (gp73) gene (22) were used. In either situation, a positive result was only observed with DNA isolated from FPV NY-98-infected cells or nucleocapsids of any of the remaining eight field strains as the template (Fig. 2A and 2B; Table 2). Failure to detect REV provirus nucleotide sequences in DNA extracted from any of the vaccine formulations or the FPV challenge strain (APHIS) was not due to inadequate purification, because a 1.2-kb product corresponding to a portion of FPV ORF 108 was observed in specific PCRs involving any of these preparations as well as those obtained from the field viruses (Fig. 2A; Table 2). Moreover, amplicons that varied from approximately 0.8 to 1.2 kb depending on the number of 36-nucleotide repeat sequences present within the amplified portion of the FPV ORF 168 template (3, 21, 23) were also seen with any of the virus genomes used as the PCR template (Table 2). Rather, as reported previously for vaccine strains of FPV (other than the discontinued Australian S strain) and pigeonpox virus origin (9, 15, 22), the genomes of the vaccine viruses examined in this study did not appear to contain intact REV provirus. As for the APHIS FPV challenge strain, the unsuccessful amplifications were anticipated because a cursory inspection of the nucleotide sequence of its DNA (1) revealed the presence of only a partial REV LTR.

FIG. 2.

PCR amplification of REV proviral and FPV genomic DNAs. Templates consisted of the genomes of reconstituted FPV vaccine strains VAC-3 (lane 1) and VAC-1 (lane 2) and FPV field strains NE-92 (lane 3), MN-97 (lane 4), PA-96 (lane 5), and OK-98 (lane 6) propagated in QT-35 cells. Amplicons obtained by PCR in the presence of primers FPV.G1, FPV.G2, REV.F, and REV.R (A), Renv-1 and Renv-2 (B), or TR-1 and TR-2 (C) were electrophoresed in agarose gels. Amplicon sizes are indicated on the right (in base pairs).

TABLE 2.

PCR amplification of portions of the FPV genome and REV provirus

Virus	Amplificationa
	REV 5′ LTR	REV gp73 gene	REV 248-bp LTR remnant	REV 508-bp LTR remnant	FPV ORF 168 (kbp)
IL-74	+	+	+	+	0.8
MN-92	+	+	+	+	1.0
NE-92	+	+	+	+	1.0
PA-96	+	+	+	+	1.0
MN-97	+	+	+	+	0.8
CA-98	+	+	+	+	0.8
NY-98	+	+	+	+	0.8
OK-97	+	+	+	+	0.8
GA-98	+	+	+	+	0.8
APHISb	−	−	+	−	0.8
VAC-1	−	−	−	+	1.2
VAC-2	−	−	+	−	0.8
VAC-3	−	−	+	−	1.2
VAC-4	−	−	+	−	1.2
VAC-5	−	−	−	+	1.2
VAC-6	−	−	+	−	1.2
VAC-7	−	−	+	−	1.2
VAC-8	−	−	+	−	1.2
VAC-9	−	−	+	−	1.2

^aPositive amplification of the REV 5′ LTR corresponds to a 282-bp product obtained with primers REV.F and REV.R. Positive amplification of the REV gp73 gene corresponds to a 642-bp product obtained with primers Renv-1 and Renv-2. Positive amplification of the REV 248-bp LTR remnant corresponds to a 485-bp product obtained with primers TR-1 and TR-2. Positive amplification of the REV 508-bp LTR remnant corresponds to a 269-bp product obtained with primers REV-2 and TR-2 (FPV field strains and vaccine strains VAC-1 and VAC-5) and also a 745-bp product obtained with primers TR-1 and TR-2 (FPV VAC-1 and VAC-5 strains only). For FPV ORF 168, the size of the product amplified in the presence of the 39K.1 and 39K.2 primers is shown.

^bFowlpox challenge virus.

To ascertain whether REV integration had occurred at a previously described site in the FPV genome (9), the region flanking this location was amplified (Fig. 1). Initially, the PCR elongation step was of sufficient duration to ensure amplification of a truncated REV LTR similar in size to that present in both the mild and S Australian FPV vaccine strains (9). Although the expected, approximately 485-bp product was obtained in PCRs involving DNA from the FPV field isolates and challenge strain as well as the majority of the vaccine strains, the size of the amplicon increased to approximately 745 bp when the templates consisted of DNA from either vaccine strain VAC-1 or VAC-5 (Fig. 2C; Table 2).

A comparison of the nucleotide sequences of the two distinct products indicated that while their FPV genomic regions were identical to that of the Australian viruses (data not shown), the amounts of the proviral 5′ and 3′ LTRs present in each differed (Fig. 1 and 3). Whereas both lacked the 3′-terminal region of the 5′ LTR and at least the initial 60% of the 3′ LTR, a nearly complete copy of the 5′ LTR was only present in the larger amplicon. Moreover, although the smaller LTR remnant and the intact provirus 5′ LTR in the United States FPV genomes were 100% homologous with their counterparts in the Australian FPV DNAs (9), one nucleotide had been deleted and one base substituted in the 3′ LTR of the provirus associated with the Australian FPV (Fig. 3).

FIG. 3.

Alignment of 5′ and 3′ LTRs present in REV provirus integrated in the genomes of FPV field isolates with long and short LTR remnants present in the genomes of FPV and pigeonpox virus vaccine strains. Nucleotide sequences of REV provirus 5′ and 3′ LTRs were determined from cloned fragments of the FPV IL-74 genome. Nucleotide sequences of the two incomplete LTRs were determined from PCR products obtained with TR-1 and TR-2 primers and FPV vaccine strains VAC-1 and VAC-5 and field strains IL-74 and NE-92 (designated Vac 1) and FPV vaccine strain VAC-3 and pigeonpox virus vaccine strain VAC-7 (designated Vac 3) DNAs. Only one example from each group is shown because identical results were obtained for all members. Conserved nucleotides are indicated by a dot, and deletions are represented by dashes. Asterisks indicate regions of exact homology between the 5′ and 3′ proviral LTRs and each remnant LTR. These common sites are shown twice in the incomplete LTRs for ease of interpretation. However, only one copy of the respective regions is present in the remnant LTRs. The locations of the U3, R, and U5 regions of the provirus 5′ and 3′ LTRs and positions of primers REV.F, REV-2, and REV.R are also indicated. Differences between the nucleotide sequences of the 3′ LTR of the United States field isolate and the Australian S vaccine strain (GenBank accession no. AF006066) are designated by an arrow (additional G) or overlined (G-to-A transition).

When PCR conditions suitable for the amplification of 8 kb of DNA (approximately the entire length of the provirus) were employed, an amplicon of this size was detected only if FPV field isolate DNA was included in the reaction (Fig. 4). As previously observed, a 485-bp product was generated from these templates and also from the DNA of FPV vaccine strain VAC-3. Once again, the larger 745-bp amplicon was obtained in the presence of the FPV vaccine strain VAC-1 genome. When the forward primer (TR-1) was replaced by one (REV-2) homologous to a portion of the 5′ LTR retained in VAC-1's and not VAC-3's DNA (Fig. 1 and 3), the predicted 269-bp amplicon was present not only in reactions utilizing the former's genomes but also in those having DNA derived from the field strains FPV MN-97 and NE-92. To verify the origin of this PCR product, the nucleotide sequence of a 269-bp amplicon obtained with FPV PA-96 DNA as the template was determined. As expected, its nucleotide sequence was identical to the corresponding portion of the LTR fusion present in VAC-1's genome (comparison not shown).

FIG. 4.

Extended PCR amplification of REV proviral and FPV genomic DNAs. Templates consisted of the genomes of reconstituted FPV vaccine strains VAC-3 (lanes 1 and 2) and VAC-1 (lanes 3 and 4) and FPV field strains NE-92 (lanes 5 and 6), MN-97 (lanes 7 and 8), and OK-98 (lane 9 and 10) propagated in QT-35 cells. OK-98 from tracheal samples from a natural outbreak of fowlpox in chickens in Oklahoma in 1998 (lanes 11 and 12) and OK-98 from secondary lesions from chickens experimentally infected with this virus (lanes 13 and 14) was also tested. Products obtained by long-range PCR in the presence of primers TR-1 and TR-2 (lanes 1, 3, 5, 7, 9, 11, and 13) or REV-2 and TR-2 (lanes 2, 4, 6, 8, 10, 12, and 14) were electrophoresed in an agarose gel. Amplicon sizes are indicated on the right (in base pairs).

Heterogeneity in FPV field strains in retention of integrated REV provirus.

The ability to amplify either type of FPV vaccine strain-retained REV LTR remnant as well as the presumed intact provirus from the genomes of FPV field isolates was indicative of a heterogeneous population. However, since all of the viruses were propagated in secondary hosts such as CAMs or cultured avian cells prior to preparation of their DNA, it is possible that the loss of provirus was a consequence of replication under these conditions. To evaluate this possibility, DNA isolated from tracheal lesions arising during the FPV OK-98 outbreak in chickens as well as from subsequently tissue culture-adapted virus and secondary lesions present on experimentally FPV OK-98-infected chickens was subjected to long-range PCR. In each reaction, regardless of the pair of primers used, amplicons corresponding in size to those obtained from the genomes of other FPV field isolates were generated (Fig. 4). Therefore, loss of the REV provirus from the FPV genome appeared to be a natural event.

To approximate the relative frequencies of the two LTR remnants and the provirus retained within field virus populations, Southern hybridization of BglII-generated fragments of FPV MN-97 and PA-96 DNA preparations was performed. This restriction enzyme was selected because its recognition site was known to be present within the provirus and also in the FPV genomic regions flanking the REV insertion site but to be absent from the LTR remnants. When the probe consisted of approximately 76% of the provirus 5′ LTR, single bands of the predicted sizes of 1.27 and 1.53 kb were detected in the DNAs from FPV vaccine strains VAC-3 and VAC-1, respectively (Fig. 5). Whereas the larger LTR remnant could not be perceived in the genomes of either field virus, the smaller remnant was readily evident in FPV PA-96 DNA and could be barely discerned in FPV MN-97 DNA. In contrast, the 1.66- and 2.63-kb fragments, housing the 5′ and 3′ LTR, respectively, hybridized with the majority of the probe, and the resultant signals were of the intensities that would be expected for equimolar pieces. Their presence only in the genomes of the two field isolates and the overall lack of any complementary nucleotide sequences in fragments of novel size supports the contention that REV has integrated at only one location within FPV DNA.

FIG. 5.

Southern hybridization of FPV vaccine and field strains with REV LTR. Electrophoretically separated BglII fragments of the genomes of FPV vaccine strains VAC-3 (lane 1) and VAC-1 (lane 2) and field strains MN-97 (lane 3) and PA-96 (lane 4) were annealed with a radioactively labeled probe consisting of a portion of the integrated REV provirus 5′ LTR. Sizes of hybridized fragments are indicated on the right (in base pairs).

Loss of integrated REV provirus from a homogeneous FPV population.

To evaluate the ease with which the provirus is lost from the FPV genome, a transcriptional cassette consisting of the vaccinia virus P11 promoter coupled to the GFP gene was inserted within the integrated REV envelope gene. Whereas cells infected with the resultant recombinant virus (MN-97-Renv-GFP) would fluoresce when exposed to 470- to 490-nm light, loss of the GFP gene and presumably the provirus would render plaques produced by such altered viruses devoid of fluorescence. Based on the inability to observe any nonfluorescing plaques generated by the progeny from the prior as well as the final purification steps, a homogeneous population of MN-97-Renv-GFP appeared to have been obtained. Moreover, evidence of any lingering parental virus was not forthcoming when a portion of the REV envelope gene flanking the GFP gene insertion site was amplified (Fig. 6). In this case, only a 2.63-kb amplicon representative of inclusion of the P11 promoter-GFP gene was obtained from the preparation of DNA from purified MN-97-Renv-GFP. The approximately 1.78-kb product indicative of unaltered field virus was only seen when MN-97 genomes were used as a template.

FIG. 6.

PCR amplification of REV proviral and FPV parental and recombinant genomic DNAs. Templates consisted of the genomes of the parental FPV MN-97 (lane 1), the recombinant FPV MN-97-Renv-GFP (lanes 2 to 4), and the nonfluorescing plaque variant of FPV MN-97-Renv-GFP (lanes 5 to 7). Amplicons obtained by PCR in the presence of primers 39K.1, 39K.2, Renv.F, and Renv.R (lanes 1, 2, and 5), TR-1 and TR-2 (lanes 3 and 6), or REV-2 and TR-2 (lanes 4 and 7) were electrophoresed in agarose gels. Amplicon sizes are indicated on the right (in base pairs).

However, despite an apparently successful plaque purification, both types of LTR remnants could be detected in DNA isolated from the first, unscreened passage of MN-97-Renv-GFP in QT-35 cells (Fig. 6). In view of the relatively low frequency of these remnants in the parental virus population, as evidenced by limited detection of only the smaller one in a Southern blot of MN-97 DNA (Fig. 5), it is unlikely that the original ones were not eliminated during plaque purification of the recombinant virus. Rather, these remnants were probably the result of novel recombination events between the two provirus LTRs. Presumably, either the frequency of their creation was low or the decreased genomic size did not confer a replicative advantage over the larger provirus-containing FPV, since only after 55 serial passages of MN-97-Renv-GFP was an actual nonfluorescent plaque observed (Fig. 7). Genotypic characterization of the responsible virus indicated that it had lost at least the majority of the provirus envelope gene and presumably the remainder of the provirus except for the smaller LTR remnant (Fig. 6). In contrast, there was no apparent alteration in its ORF 168 (Table 2), since the PCR amplicon arising from this region was similar in size to the 0.8-kb product resulting from the parental virus genome (Fig. 6).

FIG. 7.

Plaques produced by recombinant FPV MN-97-Renv-GFP and its nonfluorescing plaque variant. Two plaques (indicated by arrows) that were produced by the progeny of the 55th unscreened passage of plaque-purified FPV MN-97-Renv-GFP were examined by fluorescent (A) and bright-field (B) microscopy. Bars, 10 μm.

DISCUSSION

REV can readily integrate into the genome of the avian herpesvirus Marek's disease virus during copassaging of the two viruses in cultured avian cells. This event, however, has usually resulted in only a solitary REV LTR or one having a 23-bp repeated sequence at its 5′ terminus that is retained during subsequent replication of the altered herpesviruses (10, 12, 14). In contrast, when the recipient was the herpesvirus of turkeys, a stable integrant having an intact provirus was found in addition to ones containing individual proviruses with internal deletions (11). In each instance, the last two nucleotides of the 3′ LTR were absent and a 5-bp host sequence adjacent to the point of insertion was duplicated.

Of these two physical characteristics of retroviral integration (19), only the former was apparent in Australian (9) and United States (unpublished observations) FPV DNAs. Moreover, unlike other integrated retroviruses, the primary structure of the 3′ LTR of the intact provirus associated with the Australian FPV S vaccine strain (9) and the United States FPV UI-74 strain differed from that of the corresponding 5′ LTR in that the terminal 5′ portion of its U3 region was nearly duplicated, part of its U3 and U5 regions and the entire intervening R region were missing, and nucleotide deletions, substitutions, and additions were evident in the remaining terminal portion of its U5 region. Although the origin of this aberrant LTR is unknown, at least some of the alterations may have arisen during intergenic exchange between the 3′ LTR and other regions of integrated proviruses as predicted for some proviruses present in the genomes of Marek's disease virus (12).

As to the fragmented LTRs remaining in the DNAs of FPV vaccine strains and field isolates, their genesis is most likely the result of intramolecular, homologous recombination between the 5′ and 3′ LTRs of the integrated provirus and not due to genetic rearrangement, as previously postulated (9). As illustrated in Fig. 3, the initial 476 and terminal 19 nucleotides of the larger REV remnant are identical to the corresponding regions of the 5′ and 3′ LTRs of the intact provirus and flank a 13-nucleotide stretch common to both LTRs. Conversely, the first 3 and last 52 nucleotides of the smaller REV fragment are homologous with the respective portions of the 5′ and 3′ LTRs of the provirus in FPV DNA and flank a 193-nucleotide expanse represented in both LTRs. Thus, the length of the LTR remnants would reflect where recombination occurred. Moreover, the relatively disproportionate sizes of the two LTR regions conducive to recombination should favor the generation of the 248-nucleotide fusion. Indeed, the smaller LTR remnant is more prevalent, because only this fusion was detected in the genomes of FPV field strains (Fig. 5) and the Australian S vaccine strain (9) by Southern hybridization. In fact, the relative scarcity of the larger LTR remnant is also indicated by the requirement for a novel primer (REV-2) that is homologous with a site unique to it and the intact REV 5′ LTR for PCR amplification of observable amounts of an amplicon representative of the larger LTR fusion.

In contrast to the apparent invariance of the 248-nucleotide REV remnant in the FPV and pigeonpox virus genomes, the larger remnant in the DNAs of vaccine strains VAC-1 and VAC-5 examined in this study was only approximately 97% homologous with their counterpart in an FPV-vectored recombinant vaccine strain (15, 16). Since all 17 nucleotide differences were substitutions that occurred within the portion corresponding to the 5′ LTR of the integrated REV, the lack of alterations in the predicted DNA crossover region and the retained 3′ LTR terminus indicates that all three fusions originated in a similar manner. Whether these departures from the provirus 5′ LTR nucleotide sequence were present in the parental FPV field strain from which this vaccine virus originated or were acquired subsequently is unknown. Likewise, the existence of this type of remnant in other FPV vaccine strains has yet to be proven. Regardless, the extent of this deviation is striking in view of the nucleotide stability of the REV LTR fusions described in the current and a previous (9) study.

Although a REV LTR was previously detected in some FPV vaccine genomes in the absence of successful amplification of a portion of the REV envelope gene (15), this was not the case in the present study. Here, PCR verification of the presence of an REV LTR only occurred with the REV gp73 gene-containing DNAs of the FPV field isolates. This discrepancy was probably not due to the use of genetically distinct vaccines in the two studies, although a direct comparison cannot be made because of a lack of attribution in the former publication. Rather, the current use of a more stringent primer annealing temperature was responsible, because a reduction in this parameter enabled the formation of amplicons of the expected sizes (unpublished observations). Although both the provirus 5′ LTR and the 508-nucleotide LTR fusion had an identical region homologous to the forward PCR primer, differences in their nearly equidistant 3′ termini (reverse primer binding site), AGTATTTCGGTACAACATT versus AGTTCGGTACAACAGT (from flanking FPV DNA) would enable discrimination under the proper conditions. Thus, an LTR PCR test could be used confidently in the presence of positive (FPV field strain DNA) and negative (FPV vaccine strain VAC-1 or VAC-5 DNA) controls to confirm REV provirus contamination of FPV vaccine strains.

As shown previously (13, 22) and also in the current study, poxviruses responsible for outbreaks of fowlpox in the United States usually have a presumably intact REV provirus in their genomes. In fact, in only one instance occurring in 1956 could the perpetrator be considered to be a vaccine strain on the basis that its DNA had the small REV LTR fusion (13). That such FPV isolates can still be virulent was also shown by the ability of a virus exhibiting a similar REV genotype to produce lesions on the heads and conjunctiva of immunologically naïve ostriches (20). Quite possibly, FPV strains with such REV deletions in their DNAs or even unadulterated viruses still circulate in the environment, but continual immunization against fowlpox prevents their successful infection of poultry and subsequent detection. In this regard, whereas chickens immunized with either a commercial modified live fowlpox vaccine virus or one of five field isolates of FPV were resilient to subsequent challenge with the vaccine virus, vaccination with the commercial virus failed to confer complete protection against challenge with any of the field isolates (7). Thus, the requirement for FPV to retain an REV provirus may be more reflective of a selective pressure provided by routine vaccination than of a replicative advantage in alternative hosts. In that case, the inability to detect REV infiltration into the genomes of other avipoxviruses (13, 15) could reflect a possible symbiotic relationship between a poxvirus and a retrovirus that is not required for maintenance of avian poxviruses in wild bird populations.

Perhaps, in addition to the predicted biological selection, the limited homology between the LTRs flanking the rest of the REV provirus in the FPV genome compared to that shared by two unmodified LTRs may reduce the excision rate of the provirus and thus ensure its continuous presence within the natural FPV population. Clearly, such excisions appear to occur in vivo, based on PCR detection of the tripartite population of REV provirus-containing individuals within a primary poxvirus lesion in an infected chicken, although their frequency has not been established. Likewise, by PCR, similar deletion events were found to have happened in vitro as soon as during the initial propagation in cultured avian cells of a previously plaque-purified FPV field isolate that had been modified to express a GFP gene embedded within the REV provirus. Despite the rapid generation of such FPV deletion mutants, it was only after 55 unselected passages in the quail cell line that a non-green-fluorescing plaque arising from infection by a fowlpox virion which retained only the smaller REV LTR remnant was observed.

On the basis of the prolonged effort required to increase the frequency of FPV lacking an intact REV provirus to the level of microscopic detection, it would seem that partial elimination of the provirus from the FPV genome does not substantially enhance the rate of virus replication in cultured cells. Although a similar lack of replicative advantage has not been established during infection of poultry by the virus, it should be noted that three of four REV provirus-containing FPV field isolates were very virulent in chickens compared to two relatively attenuated FPV vaccine strains retaining only one of the two different-sized REV LTR remnants (22). Moreover, removal of the entire REV provirus from the FPV genome rendered the resultant virus less virulent in chickens than its parent, whereas reinsertion of the provirus enhanced the level of virulence associated with the restored virus (unpublished data). Since, in these studies, the extent of virulence was correlated with the development of secondary lesions and the persistence of the primary lesion at the site of inoculation, REV integration into the FPV genome may actually augment poxvirus replication in its natural host, as is the case for numerous genes deemed to be nonessential for virus growth in cultured cells.