Conservation and variation of the parapoxvirus GM-CSF-inhibitory factor (GIF) proteins

Page created by Gabriel Peterson
 
CONTINUE READING
Journal of General Virology (2009), 90, 970–977                                                                DOI 10.1099/vir.0.006692-0

                                      Conservation and variation of the parapoxvirus
                                      GM-CSF-inhibitory factor (GIF) proteins
                                      D. Deane,1 N. Ueda,2 L. M. Wise,2 A. R. Wood,1 A. Percival,1 C. Jepson,1
                                      N. F. Inglis,1 S. B. Fleming,2 A. A. Mercer2 and C. J. McInnes1
  Correspondence                      1
                                       Moredun Research Institute, Pentlands Science Park, Bush Loan, Penicuik EH26 0PZ,
  C. J. McInnes                        Scotland, UK
  mcinc@mri.sari.ac.uk                2
                                       Virus Research Unit, University of Otago, Dunedin, New Zealand

                                      The GIF protein of orf virus (ORFV) binds and inhibits the ovine cytokines granulocyte–
                                      macrophage colony-stimulating factor (GM-CSF) and interleukin-2 (IL-2). An equivalent protein
                                      has so far not been found in any of the other poxvirus genera and we therefore investigated
                                      whether it was conserved in the parapoxviruses. The corresponding genes from both the
                                      bovine-specific pseudocowpox virus (PCPV) and bovine papular stomatitis virus (BPSV) were
                                      cloned and sequenced. The predicted amino acid sequences of the PCPV and BPSV proteins
                                      shared 88 and 37 % identity, respectively, with the ORFV protein. Both retained the six cysteine
                                      residues and the WSXWS-like motif that are required for biological activity of the ORFV protein.
                                      However, an analysis of the biological activity of the two recombinant proteins revealed that,
                                      whilst the PCPV GIF protein bound to both ovine and bovine GM-CSF and IL-2 with very similar
  Received 19 August 2008             binding affinities to the ORFV GIF protein, no GM-CSF- or IL-2-binding activity was found for the
  Accepted 10 December 2008           BPSV protein.

INTRODUCTION                                                            orthopoxvirus protein (VAC C21L), which inhibits the
                                                                        activation of the complement cascade (Isaacs et al., 1992).
The parapoxviruses (PPVs), members of the genus
                                                                        Some, such as the vFLIP protein (MC159) of molluscum
Parapoxvirus of the family Poxviridae, cause a contagious
                                                                        contagiosum virus, which inhibits apoptosis, may be
pustular dermatitis in ruminants. Orf virus (ORFV) is the
                                                                        species-specific (Shisler & Moss, 2001; Thurau et al.,
type species of the genus and causes disease in sheep and
                                                                        2006). ORFV has been shown to encode a protein
goats, whereas bovine papular stomatitis virus (BPSV) and
                                                                        (GIF) that is capable of binding and inhibiting both of
pseudocowpox virus (PCPV) affect cattle. All three viruses
                                                                        the ovine cytokines interleukin-2 (IL-2) and granulocyte–
can infect humans. As is common amongst other
                                                                        macrophage colony-stimulating factor (GM-CSF) (Deane
poxviruses, ORFV has been shown to encode factors that
                                                                        et al., 2000; McInnes et al., 2005). Such a protein has
either mimic host immunoregulatory proteins (Fleming
                                                                        not been found so far in any other poxvirus; we there-
et al., 1997; Haig & Fleming, 1999; Seet et al., 2003) or have
                                                                        fore investigated whether it was conserved in other PPVs.
the potential for interacting with components of the host
immune system (Haig et al., 1998; McInnes et al., 1998,
2005; Deane et al., 2000). In doing so, it is thought that
these virus-encoded factors are capable of subverting the               METHODS
host immune response to infection, creating an envir-                   Viruses. ORFV strain NZ2 (Mercer et al., 2006), PCPV strains BO74
onment suitable for efficient virus replication. Some of the            (kindly obtained from Dr Mathias Buettner, Bavarian Health and
poxvirus factors, such as the double-stranded RNA-                      Food Safety Authority, Oberschleissheim, Germany) and VR634
                                                                        (Gassmann et al., 1985) and BPSV strains V660 (Gassmann et al.,
binding protein (VAC E3L), which inhibits the antiviral
                                                                        1985) and A599 (isolated de novo in the UK) were used throughout
effects of interferon (Haig et al., 1998; McInnes et al.,               this study and were maintained by passage in primary bovine testis or
1998), are conserved across two or more poxvirus genera,                fetal lamb muscle cells.
whilst others appear to be genus-specific, such as the
                                                                        Expression of the PPV GIFs. The entire open reading frames of the
                                                                        three PPV GIF genes were amplified by PCR and cloned individually
The GenBank/EMBL/DDBJ accession numbers for the sequences of            into the pEE14 expression vector (Celltech) (Cockett et al., 1990). The
the PCPV and BPSV GIF genes are EU999744 and EU999745,                  integrity of the genes was verified by sequencing prior to transfection
respectively.                                                           into COS-7 fibroblasts or CHO cells. Stable cell lines expressing each
Supplementary figures showing ELISA and Northern blotting results are   of the GIF proteins were established as described previously (Deane
available with the online version of this paper.                        et al., 2000). Cell lines expressing the recombinant ORFV and PCPV

970                                                                                        006692 G 2009 SGM Printed in Great Britain
Parapoxvirus GM-CSF-inhibitory factor proteins

GIF (rGIF) proteins were selected by assaying for GIF activity using a      Healthcare Life Sciences), according to the manufacturer’s instruc-
modification of the ovine (ov) GM-CSF-specific ELISA (Entrican et           tions, and exposure to Hyperfilm ECL for 0.5–5.0 min before
al., 1996; Deane et al., 2000). The pEE14 vector contains a glutamine       development.
synthetase (GS) gene; therefore, cell lines transfected stably with the
plasmid can be selected on the basis of their survival in methionine        Ligand blotting. GIF–cytokine binding was detected by ligand
sulphoximine (MSX; Sigma). The cell line expressing the BPSV GIF            blotting as described previously (Deane et al., 2000). Briefly, 250–
was selected based on its survival in MSX and upon reactivity with a        400 ng each cytokine was separated by denaturing SDS-PAGE (15 %
rabbit anti-ORFV GIF serum. Verification that the anti-ORFV GIF             gel) and then transferred to nitrocellulose membranes. The mem-
serum recognizes the BPSV GIF was obtained by direct ELISA using            branes were blocked in blocking buffer, washed in PBS containing
an anti-ORFV GIF mAb (see Supplementary Fig. S1, available in JGV           0.05 % Tween 20 (wash buffer) and then incubated with 125I–GIF
Online). The recombinant proteins used in these studies were                (15–25 pM) for 2 h at room temperature. After washing three times
prepared by culturing the cells for 4 days after reaching confluence        with wash buffer, bound GIF was detected by autoradiography using
in Glasgow’s modified Eagle’s medium without serum or MSX. The              Hyperfilm MP X-ray film (GE Healthcare Life Sciences).
serum-free supernate was harvested, clarified by centrifugation at
1000 g, dispensed in aliquots and stored at 220 uC. Recombinant             Scatchard analysis of GIF activity. Purified cytokines were radio-
ORFV and PCPV GIF were purified from CHO cells by affinity                  iodinated by the chloramine-T method and a soluble ligand-binding
chromatography using purified rovGM-CSF bound to CNBr–                      assay was performed for each cytokine as described previously (Deane
Sepharose (GE Healthcare Life Sciences). Cell-free supernates from          et al., 2000). Briefly, a range of 125I–GM-CSF and 125I–IL-2
the rGIF-transfected cells were applied to the GM-CSF–Sepharose             concentrations (2–20 nmol) was incubated with GIF (100 ng)
affinity column. Eluted fractions were tested for GIF activity by using     containing 5 % fetal bovine serum for 2 h at room temperature.
the ovine GM-CSF-specific ELISA. Recombinant BPSV GIF was                   Bound proteins were precipitated by the addition of 20 %
partially purified, from CHO cell-free supernates, by Mono-Q anion-         polyethylene glycol (PEG 6000; Sigma) in PBS and incubation on
exchange chromatography followed by gel filtration on Superdex 200          ice for 30 min. The precipitated material was collected under vacuum
(GE Healthcare Life Sciences). For ligand blotting, the purified GIF        onto GF/C filter discs (Whatman) and washed four times with ice-
proteins were radio-iodinated by the chloramine-T method                    cold 10 % PEG 6000 in PBS. The radiolabelled complexes were
(McConahey & Dixon, 1980) and further purified on Superdex 200              detected and quantified in a gamma scintillation counter. The level of
columns in PBS with 0.04 % (v/v) CHAPS buffer, pH 7.2.                      non-specific binding of 125I-labelled cytokines to the filter discs in the
                                                                            absence of GIF was measured and the data were adjusted accordingly.
Assay of GIF activity. The presence of biologically active ORFV GIF         Scatchard analysis was performed on best-fit plots [bound counts
can be measured indirectly by its ability to interfere with the detection   min21/free counts min21 (y-axis) versus bound counts min21 (x-
of ovine/bovine GM-CSF by specific ELISA. Cell-free supernates and          axis)], generated by using the Origin software package (OriginLab
cell lysates from GIF cDNA-transfected and virus-infected cells were        Corporation).
assayed for GM-CSF-binding activity as described previously (Deane
et al., 2000). Binding of GIF to ovine/bovine IL-2 was similarly            Cytometric bead array (CBA) assay. Cell-free supernates were
measured by ELISA. Briefly, recombinant ovine/bovine IL-2 was               collected from virus-infected cells and concentrated approximately 10
isolated from CHO cell-free supernates by Mono-Q anion-exchange             times by using Centricon YM-10 concentrators (Millipore). These
chromatography followed by gel filtration on a Superdex-200 column          and purified rGIF proteins were assayed for chemokine-binding
(GE Healthcare Life Sciences). ELISA plates were coated overnight           activity with the human chemokines IL-8, MCP-1, MIP-1a, MIP-1b
with 1 mg IL-2 ml21 in 0.1 M NaHCO3 (pH 9.5) before blocking with           and RANTES in a modified CBA assay (BD), carried out as outlined
4 % non-fat milk powder in PBS. Cell-free supernates and cell lysates       in the manufacturer’s instructions. Briefly, 50 ml aliquots of infected
from GIF cDNA-transfected and virus-infected cells were assayed for         cell supernates, rGIFs (10–20 mg ml21) or control samples were pre-
IL-2-binding activity. Bound GIF was detected with 2 mg ml21 of an          incubated with 50 ml individual chemokine standards from
affinity-purified IgG fraction from a rabbit anti-GIF serum, followed       Chemokine kit II (1250 pg ml21) for 1 h at 4 uC. The mixture was
by a 1 : 1000 dilution of goat anti-rabbit IgG conjugated with              then added to a 50 ml aliquot of mixed capture beads and PE
horseradish peroxidase (HRP; DAKO) in wash buffer [PBS containing           detection reagent and incubated for 3 h at room temperature in the
0.02 % Tween 20 (Sigma)]. For colour development, TMB peroxidase            dark. After washing, the beads were analysed for bound PE in a
substrate (SureBlue; Kirkegaard & Perry Laboratories) was added; the        FACScalibur flow cytometer (BD). Levels of each chemokine present
reaction was stopped after 20–30 min by the addition of 0.1 M HCl           in the samples were calculated from the data by using CBA analysis
and A450 was read. The specificity of the interaction between GIF and       software (BD).
IL-2 was verified by pre-incubating the IL-2-coated ELISA plate with
an anti-ovine IL-2 mAb, 1E10 (a gift from Dr P. Wood, CSIRO
Animal Health, Parkville, Australia; Pedersen et al., 2002), before
incubating with GIF.                                                        RESULTS
                                                                            The ORFV GIF gene (ORFV117) was found approximately
Western blot analysis. In order to verify protein expression from
                                                                            20 kbp from the right terminus of the genome (Deane et al.,
the various GIF constructs, the cell-free supernates and/or cell lysates
from the COS-7/CHO cells were electrophoresed in a denaturing               2000; Mercer et al., 2006) and we therefore explored the
SDS–polyacrylamide gel (12 %) and transferred to Hybond ECL                 corresponding regions of the PCPV, strain VR634, and
nitrocellulose membranes (GE Healthcare Life Sciences) for 3 h at           BPSV, strain V660, genomes. Sequencing of the right-hand
2 mA (cm gel)22. The membranes were washed in PBS containing                end of the PCPV EcoRI D fragment (Gassmann et al., 1985)
4 % non-fat milk powder for 1 h at room temperature before being            revealed a homologue of the ORFV GIF gene. Sequencing
probed with a rabbit anti-GIF IgG fraction (1 mg ml21) in blot wash         the ends of the 20.7 kb BPSV BamHI A fragment and
buffer [PBS containing 0.35 M NaCl and 0.5 % (v/v) Tween 80
(Sigma)]. Binding of antibody to immobilized proteins was visualized
                                                                            comparing the sequence with that of ORFV suggested that
by a further 1 h incubation with a 1 : 1000 dilution of goat anti-rabbit    the BPSV GIF gene was likely to be located in this
IgG conjugated with HRP (DAKO) in wash buffer, followed by                  fragment. We found partial sequences of the BPSV GIF
treatment with the enhanced chemiluminescence reagent ECL (GE               gene in SalI subfragments of the BPSV BamHI A fragment,

http://vir.sgmjournals.org                                                                                                                       971
D. Deane and others

with the complete sequence being determined by the             also used for comparative purposes. No GM-CSF- or IL-2-
primer-walking method. Subsequent publication of the           binding activity was found with this strain. To verify that
complete BPSV genome sequence (Delhon et al., 2004)            the GIF gene was being expressed by the BPSV strains, a
confirmed this to be the equivalent of the ORFV GIF gene.      Northern blot of RNA isolated from infected cells 24 h
The sequences of the PCPV and BPSV GIF genes have been         post-infection was probed with a 32P-labelled BPSV GIF
deposited in GenBank under accession numbers EU999744          cDNA (see Supplementary Fig. S2, available in JGV
and EU999745, respectively. An alignment of the predicted      Online). As had been reported previously for ORFV GIF
GIF protein sequences from the three viruses is shown in       (Deane et al., 2000), no discrete mRNA was detected, but
Fig. 1.                                                        rather a smear representative of poxvirus late mRNA
                                                               transcripts was present after 24 h. This is also consistent
All three GIF proteins are predicted to be 265 aa in length,
                                                               with the T-rich late promoter-like sequences found
with both ORFV and PCPV GIF possessing a 19 aa signal
                                                               upstream of the start codons of the GIF genes of all three
peptide, whilst that of BPSV GIF is predicted to be 20 aa in
                                                               viruses (data not shown).
length. The ORFV and PCPV predicted proteins share 88 %
identity, whereas both share only 37 % identity with the       Supernates from CHO cells transfected with the PPV GIF
predicted BPSV protein. The ORFV and PCPV GIFs share           cDNAs were analysed for GIF activity as above. As with the
four conserved potential Asn-linked glycosylation sites,       native GIF, the recombinant ORFV and PCPV GIFs bound
only two of which are conserved in the BPSV protein,           ovine and bovine GM-CSF and IL-2, but the rGIF from
which has a further three potential sites. The mature ORFV     BPSV bound neither (Figs 2 and 3). Again, a Northern blot of
GIF protein is predicted to contain eight Cys residues,        the RNA from the BPSV GIF cDNA-transfected cells
whilst both PCPV and BPSV GIFs are predicted to contain        indicated that the gene was being expressed (see
seven. Six of these, however, are positionally conserved in    Supplementary Fig. S3, available in JGV Online). In addition,
all three proteins. The WDPWV motif, important for the         the cell-free supernate from the BPSV GIF cDNA-transfected
biological activity of ORFV GIF (McInnes et al., 2005), is     cells was analysed by liquid chromatography electrospray
perfectly conserved in the PCPV GIF sequence, but              ionization tandem mass spectrometry to verify that the BPSV
although present in the BPSV GIF sequence, it shows two        GIF protein was being produced (data not shown).
amino acid changes, WSPWT.
                                                               Western and ligand-blot analysis
Native and recombinant GIF activity
                                                               In order to verify that each of the three PPV GIF species
Native GIF activity was assessed in the supernates from        was being expressed, partially purified proteins from the
fetal lamb muscle cell cultures 72 h post-infection with the   transfected CHO cells were blotted onto nitrocellulose and
three species of PPV at an m.o.i. of 10. Whilst both ORFV      probed with an IgG fraction of a rabbit antiserum raised
and PCPV native GIFs were found to bind both ovine and         against ORFV GIF. Under reducing conditions, protein
bovine GM-CSF and IL-2, GIF from BPSV did not bind             bands were detected at approximately 43 and 55 kDa
either cytokine (Figs 2 and 3). As the BPSV strain used        (Fig. 4). The 43 kDa protein represents the native and
(V660) is a well-used laboratory strain that has been          highly glycosylated form of the predicted 28 kDa
passaged in culture numerous times, a second BPSV strain       monomer, whereas the 55 kDa form is thought to be a
(A599) that has not been repeatedly passaged in vitro was      GIF dimer (McInnes et al., 2005).

                                                                                 Fig. 1. Alignment of the predicted mature
                                                                                 amino acid sequences of the three PPV GIF
                                                                                 proteins. The predicted ORFV GIF sequence
                                                                                 is given as the consensus and identical amino
                                                                                 acids in the other proteins are indicated by an
                                                                                 asterisk (*). Missing amino acids are indicated
                                                                                 by a dash (–). The predicted signal peptides
                                                                                 are not shown. The six Cys residues and
                                                                                 potential glycosylation sites are shaded and
                                                                                 the WDPWV motif known to be important for
                                                                                 biological activity of ORFV GIF is boxed.

972                                                                                              Journal of General Virology 90
Parapoxvirus GM-CSF-inhibitory factor proteins

                                                                               Fig. 2. Cell-free supernates from virus-
                                                                               infected cells or from CHO cells transfected
                                                                               with expression plasmid constructs of the
                                                                               three PPV species of GIF were incubated with
                                                                               recombinant ovine GM-CSF (filled bars) or
                                                                               bovine GM-CSF (empty bars) for 1 h at 37 6C.
                                                                               The mixtures were then subjected to the GM-
                                                                               CSF-specific ELISA. A high ELISA reading
                                                                               indicates no interference in the detection of
                                                                               GM-CSF (”ve GIF activity), whilst a low ELISA
                                                                               reading indicates interference with detection
                                                                               (+ve GIF activity), as indicated by the bar on
                                                                               the right. Values are means±SEM of quad-
                                                                               ruplicate samples. *P,0.001 compared with
                                                                               the medium control.

The partially purified proteins were then radio-iodinated     above that of the non-specific control samples was
and tested for GM-CSF- and IL-2-binding activity by           detected.
ligand blotting (Fig. 5). Under normal blotting conditions,
125
    I-labelled PCPV GIF bound both sets of cytokines
                                                              Investigation of chemokine binding by PPV GIFs
strongly. In contrast, 125I-labelled ORFV GIF bound all
four cytokines only under less stringent washing conditions   It had previously been suggested that the ORFV GIF
and, in addition, a much longer exposure time to X-ray        protein showed significant similarity to a family of
film was required to detect binding. No binding was           poxvirus chemokine-binding proteins; more specifically,
observed with 125I-labelled BPSV GIF, even with mild          those found in members of the genera Orthopoxvirus and
washing conditions and extended exposure times.               Leporipoxvirus that bound a variety of CC chemokines
                                                              (Seet et al., 2003). In addition, significant similarity was
                                                              also found between it and another ORFV protein, encoded
Determination of PPV GIF binding constants
                                                              by a gene found upstream of the GIF gene in the ORFV
The affinity of the binding of PCPV and ORFV GIFs to          genome. This protein also displayed chemokine-binding
both ovine and bovine GM-CSF and IL-2 was determined          activity, but in addition to binding CC chemokines, was
by Scatchard analysis (Table 1). Although no binding          also shown to bind the C chemokine lymphotactin. As a
activity had been detected with the BPSV GIF in either the    consequence of the similarity between the sequences of
ELISA or ligand-blot assays, it was still included in this    these proteins, it had been suggested that the GIF protein
analysis. However, again no binding activity significantly    may also exhibit chemokine-binding activity. As no GM-

                                                                               Fig. 3. Cell-free supernates from virus-
                                                                               infected cells or from CHO cells transfected
                                                                               with expression plasmid constructs of the
                                                                               three PPV species of GIF were incubated with
                                                                               ovine IL-2 (filled bars) or bovine IL-2 (empty
                                                                               bars) pre-bound to 96-well plates. After
                                                                               washing to remove non-specifically bound
                                                                               protein, the PPV GIFs were detected by using
                                                                               an affinity-purified IgG fraction from a rabbit
                                                                               anti-GIF serum, followed by a 1 : 1000 dilution
                                                                               of goat anti-rabbit IgG conjugated with HRP.
                                                                               Values are means±SEM of quadruplicate
                                                                               samples. *P,0.01 compared with the anti-IL-
                                                                               2+GIF control.

http://vir.sgmjournals.org                                                                                               973
D. Deane and others

Fig. 4. Western blot of PPV GIF proteins with a rabbit anti-rGIF-
ORFV GIF polyclonal antiserum. Partially purified recombinant PPV
GIF proteins were separated on SDS-PAGE gels (12 %) run under
reducing conditions. Proteins were electrotransferred to nitrocel-
lulose sheets and probed with an IgG fraction of a rabbit antiserum
raised against rGIF-ORFV. Bound antibody was visualized by
treating the blots with an HRP-conjugated goat anti-rabbit IgG
fraction and ECL reagent. The positions of molecular mass markers
are shown on the left. The 43 kDa protein isolated from rGIF-
ORFV represents the native and highly glycosylated form (of the
predicted 28 kDa monomer). The 55 kDa form is thought to be a
GIF dimer.

CSF- or IL-2-binding activity had been detected for the
BPSV GIF, we explored the possibility that it could bind CC
chemokines. This was done by using a modified CBA assay.
Chemokine-binding activity was not detected with any of
the recombinant GIF proteins or supernates from the
BPSV-infected cells. However, chemokine binding was
detected in both ORFV- and PCPV-infected cells (Fig. 6).
This, however, differed between the two viruses, with
MCP-1, MIP-1a, MIP-1b and RANTES binding being
found with supernates from ORFV-infected cells, whilst
only RANTES binding was detected with the supernates                  Fig. 5. Partially purified ovine and bovine GM-CSF and IL-2 were
from PCPV-infected cells. This suggests that a factor(s),             separated by SDS-PAGE and transferred to nitrocellulose. These
                                                                      were then probed with radio-iodinated rGIF-ORFV (a), rGIF-PCPV
other than GIF, produced by ORFV and PCPV is
                                                                      (b) or rGIF-BPSV (c); the blots were washed with 0.02 % Tween
responsible for binding these chemokines.
                                                                      20 in PBS and then subjected to autoradiography. The gel in (b)
                                                                      was exposed for 48 h; those in (a) and (c) were exposed for
                                                                      7 days.
DISCUSSION
Orthologues of the ORFV GIF protein were found in the
two bovine-specific PPVs BPSV and PCPV. In addition to
                                                                      amino acid similarity with them. In addition, some of the
sharing 88 % overall amino acid identity with the ORFV
                                                                      features that are key to the biological activity of GIF, such
GIF protein, the PCPV protein was found to contain the
                                                                      as the six Cys residues, two of the putative Asn-linked
six Cys residues, the Asn-linked glycosylation sites and the
                                                                      glycosylation sites and the WSXWS-like motif, are
WDPWV (WSXWS-like) motif, all of which were shown to
                                                                      conserved in the BPSV protein. However, we did not
be necessary for the biological activity of ORFV GIF
                                                                      detect any interaction between it and any of the cytokines
(McInnes et al., 2005). As might have been predicted from
                                                                      tested. This may be due to a variety of factors. BPSV V660
the presence of these motifs, we were able to show that the
                                                                      is a laboratory strain of the virus that has undergone serial
PCPV protein was able to bind GM-CSF and IL-2.
                                                                      passage in culture and could therefore potentially have lost
Although it is only 37 % identical to the ORFV and PCPV               the ability either to express the GIF gene or, as a result of
proteins, the BPSV GIF protein shares approximately 60 %              cumulative mutations to the gene sequence, to produce

974                                                                                                      Journal of General Virology 90
Parapoxvirus GM-CSF-inhibitory factor proteins

Table 1. Binding affinities (in pM) of PPV GIF proteins with ovine and bovine GM-CSF and IL-2

 GIF                    Ovine GM-CSF          Ovine IL-2         Bovine GM-CSF             Bovine IL-2

 rGIF-ORFV                   405–500          1110–1390              400–468                496–556
 rGIF-PCPV                   248–278           536–714               237–263                250–338
 rGIF-BPSV                     NA                 NA                    NA                     NA

NA,   Not applicable.

biologically active GIF. We demonstrated, however, that                 the sequence reported here is representative of the true
the GIF mRNA from V660 and A599 was expressed                           BPSV GIF sequence and it is considered unlikely that the
similarly to that of PCPV GIF. We also tested cell-free                 few changes in amino acids noted would have resulted in
supernates from cells infected with A599, a different strain            an inactive protein. It is worth noting, however, that the
of BPSV that has been passaged fewer than 10 times since                WSXWS-like motif that is found in the BPSV GIF protein
its isolation. However, again no GM-CSF- or IL-2-binding                is different from those found in the ORFV and PCPV
activity was detected with this virus. The V660 sequence                proteins. This motif is found in a number of cytokine
reported here is 95.5 % identical to that of the published              receptors and is thought to be linked to the tertiary
BPSV sequence (Delhon et al., 2004) and 98 % identical to               structure of the receptor molecules (Bazan, 1990; Quelle
another isolate (B177; kindly provided by H.-J. Rziha,                  et al., 1992; Baumgartner et al., 1994; Ronco et al., 1995).
Friedrich-Loeffler-Institute, Federal Research Institute for            The sequence of the motif varies according to the cytokine
Animal Health, Tübingen, Germany), which we have                       being bound and, therefore, it may be that the amino acid
sequenced (results not shown). Therefore, it is likely that             changes within the BPSV GIF molecule could alter its

        Fig. 6. An adaptation of the CBA assay (BD) was used to assess chemokine-binding activity in cell-free supernates of PPV-
        infected cells or PPV GIF-transfected cells. Supernates were pre-mixed with a known amount of the standard chemokines
        MCP-1 (diagonally hatched bars), MIP-1a (filled bars), MIP-1b (horizontally hatched bars), RANTES (shaded bars) or IL-8
        (empty bars) before mixing with the capture beads and detection reagent. This was incubated for 3 h at room temperature
        before binding was measured using a FACScalibur flow cytometer (BD). A high reading indicates no binding activity in the
        supernates, whereas a low reading indicates interference by the cell-free supernates of binding between the chemokines and
        the capture beads. Data represent mean±SEM binding of quadruplicate samples. *P,0.001 compared with medium taken from
        untransfected CHO cells.

http://vir.sgmjournals.org                                                                                                           975
D. Deane and others

binding specificity in comparison to the ORFV and PCPV          (corresponding to VACV F13L) shares approximately 95 %
proteins.                                                       identity with the equivalent ORFV protein, but only 84 %
                                                                identity with that of BPSV (Inoshima et al., 2001).
The GIF proteins of the PPVs remain unique, with no
                                                                Similarly, there is approximately 68 % identity between
corresponding proteins yet found in any of the other
                                                                ORFV and PCPV VEGF proteins, but only 55 % between
genera of the subfamily Chordopoxvirinae. This is despite
                                                                the PCPV and BPSV VEGFs (Ueda et al., 2003). This
many computer-based structural-analysis programs sug-
                                                                difference is also reflected in the biological activities of the
gesting a close relationship with the family of type II         virus-encoded VEGF proteins. In addition to binding
chemokine-binding proteins (vCCI) encoded by members            VEGFR-2, the BPSV VEGF also has the distinct capability,
of the genera Orthopoxvirus and Leporipoxvirus (Carfi et al.,   amongst the poxviral VEGFs, of binding VEGFR-1 and
1999; Seet & McFadden, 2002). This relationship is              inducing monocyte migration (Inder et al., 2007). It had
primarily due to sequence similarity found between the          originally been thought that the two cattle viruses would be
GIF protein and a chemokine-binding protein also                more similar to each other than either would be to ORFV.
encoded by the PPVs (ORF 112). This protein has been            However, the data published here for the three GIF
shown to be related to the vaccinia virus (VACV)                proteins would appear to reinforce the suggestion that
chemokine-binding protein C23L and to possess both C-           PCPV is related more closely to ORFV than to BPSV. As it
and CC- chemokine-binding activity (Seet et al., 2003). As      would appear that PCPV does not cause disease in sheep or
we did not find any GM-CSF- or IL-2-binding activity with       goats and ORFV does not cause disease in cattle, a more
the BPSV GIF protein, we investigated whether the GIF           detailed comparison of the ORFV and PCPV genomes may
proteins exhibited CC chemokine-binding activity. No            reveal critical elements that help to dictate host range
such activity was detected with any of the rGIF proteins,       within the PPVs.
despite the fact that the supernates from both ORFV- and
PCPV-infected cells appeared to interact with human CC
chemokines. This activity was presumably due to the PPV         ACKNOWLEDGEMENTS
chemokine-binding protein characterized previously.
                                                                This work was funded by the Scottish Government and the Health
Interestingly, no activity was detected from BPSV-infected
                                                                Research Council of New Zealand, and made extensive use of reagents
cells. The putative BPSV chemokine-binding protein              developed under the BBSRC/RERAD Immunological Toolbox (grant
(Delhon et al., 2004) is predicted to be only 41 % identical    nos BBS/B/00255 and MRI/094/04).
to the ORFV protein (the PCPV protein is yet to be
characterized), with even lower similarity to the vCCI
orthopox proteins and, as a result, could have a different      REFERENCES
binding specificity. It has already been suggested that the
ORFV protein has shifted its binding specificity in             Baumgartner, J. W., Wells, C. A., Chen, C. M. & Waters, M. J. (1994).
                                                                The role of the WSXWS equivalent motif in growth-hormone
comparison to the orthopox and leporipox family of
                                                                receptor function. J Biol Chem 269, 29094–29101.
proteins, as it was shown to bind C chemokines as well as
                                                                Bazan, J. F. (1990). Structural design and molecular evolution of a
the CC chemokines.
                                                                cytokine receptor superfamily. Proc Natl Acad Sci U S A 87, 6934–
Our findings suggest that the ORFV and PCPV GIF proteins        6938.
behave similarly in terms of their ability to bind the          Carfi, A., Smith, C. A., Smolak, P. J., McGrew, J. & Wiley, D. C. (1999).
cytokines GM-CSF and IL-2. However, the same activity was       Structure of a soluble secreted chemokine inhibitor vCCI (p35) from
not found with the BPSV GIF protein. It is interesting that     cowpox virus. Proc Natl Acad Sci U S A 96, 12379–12383.
the pathological appearance of orf and pseudocowpox are         Cockett, M. I., Bebbington, C. R. & Yarranton, G. T. (1990). High level
also similar and that they both differ from bovine papular      expression of tissue inhibitor of metalloproteinases in Chinese
                                                                hamster ovary cells using glutamine synthetase gene amplification.
stomatitis, most notably in the proliferative nature of their
                                                                Biotechnology (N Y) 8, 662–667.
lesions. Whilst we are not suggesting that the GIF protein
                                                                Deane, D., McInnes, C. J., Percival, A., Wood, A., Thomson, J., Lear, A.,
has a role to play in the proliferation of ORFV and PCPV
                                                                Gilray, J., Fleming, S., Mercer, A. & Haig, D. (2000). Orf virus encodes a
lesions, we do think that it is possible that the BPSV GIF      novel secreted protein inhibitor of granulocyte–macrophage colony-
protein has evolved a quite separate spectrum of activity       stimulating factor and interleukin-2. J Virol 74, 1313–1320.
more suited to its biological niche.                            Delhon, G., Tulman, E. R., Afonso, C. L., Lu, Z., de la Concha-
The full genomic sequences of ORFV and BPSV have been           Bermelillo, A., Lehmkuhl, H. D., Piccone, M. E., Kutish, G. F. & Rock,
                                                                D. L. (2004). Genomes of the parapoxviruses ORF virus and bovine
published and show approximately 67–75 % nucleotide             papular stomatitis virus. J Virol 78, 168–177.
identity across the whole genome, with amino acid
                                                                Entrican, G., Deane, D., MacLean, M., Inglis, L., Thomson, J.,
identities between proteins ranging from 34 to 93 %, with       McInnes, C. & Haig, D. M. (1996). Development of a sandwich
an average of 71 % (Delhon et al., 2004; Mercer et al.,         ELISA for ovine granulocyte/macrophage colony-stimulating factor.
2006). Unfortunately, the full genomic sequence of PCPV         Vet Immunol Immunopathol 50, 105–115.
is not known, but the small amount of sequence that has         Fleming, S. B., McCaughan, C. A., Andrews, A. E., Nash, A. D. &
been published appears to be more similar to ORFV than it       Mercer, A. A. (1997). A homolog of interleukin-10 is encoded by the
is to BPSV. For example, the major envelope protein             poxvirus orf virus. J Virol 71, 4857–4861.

976                                                                                                     Journal of General Virology 90
Parapoxvirus GM-CSF-inhibitory factor proteins

Gassmann, U., Wyler, R. & Wittek, R. (1985). Analysis of parapoxvirus      sequences of three isolates of orf virus reveals unexpected sequence
genomes. Arch Virol 83, 17–31.                                             variation. Virus Res 116, 146–158.
Haig, D. M. & Fleming, S. (1999). Immunomodulation by virulence            Pedersen, L. G., Castelruiz, Y., Jacobsen, S. & Aasted, B. (2002).
proteins of the parapoxvirus orf virus. Vet Immunol Immunopathol           Identification of monoclonal antibodies that cross-react with
72, 81–86.                                                                 cytokines from different animal species. Vet Immunol
Haig, D. M., McInnes, C. J., Thomson, J., Wood, A., Bunyan, K. &           Immunopathol 88, 111–122.
Mercer, A. (1998). The orf virus OV20.0L gene product is involved in       Quelle, D. E., Quelle, F. W. & Wojchowski, D. M. (1992). Mutations in
interferon resistance and inhibits an interferon-inducible, double-        the WSAWSE and cytosolic domains of the erythropoietin receptor
stranded RNA-dependent kinase. Immunology 93, 335–340.                     affect signal transduction and ligand-binding and internalization. Mol
Inder, M. K., Ueda, N., Mercer, A. A., Fleming, S. B. & Wise, L. M.        Cell Biol 12, 4553–4561.
(2007). Bovine papular stomatitis virus encodes a functionally distinct    Ronco, L. V., Doyle, S. E., Raines, M., Park, L. S. & Gasson, J. C.
VEGF that binds both VEGFR-1 and VEGFR-2. J Gen Virol 88, 781–             (1995). Conserved amino acids in the human granulocyte–macro-
791.                                                                       phage colony-stimulating factor receptor-binding subunit essential
Inoshima, Y., Murakami, K., Yokoyama, T. & Sentsui, H. (2001).             for tyrosine phosphorylation and proliferation. J Immunol 154, 3444–
Genetic heterogeneity among parapoxviruses isolated from sheep, cattle     3453.
and Japanese serows (Capricornis crispus). J Gen Virol 82, 1215–1220.      Seet, B. T. & McFadden, G. (2002). Viral chemokine-binding
Isaacs, S. N., Kotwal, G. J. & Moss, B. (1992). Vaccinia virus             proteins. J Leukoc Biol 72, 24–34.
complement-control       protein    prevents    antibody-dependent         Seet, B. T., McCaughan, C. A., Handel, T. M., Mercer, A., Brunetti, C.,
complement-enhanced neutralization of infectivity and contributes          McFadden, G. & Fleming, S. B. (2003). Analysis of an orf virus
to virulence. Proc Natl Acad Sci U S A 89, 628–632.                        chemokine-binding protein: shifting ligand specificities among a
McConahey, P. J. & Dixon, F. J. (1980). Radioiodination of proteins by     family of poxvirus viroceptors. Proc Natl Acad Sci U S A 100, 15137–
the use of the chloramine-T method. Methods Enzymol 70, 210–213.           15142.
McInnes, C. J., Wood, A. R. & Mercer, A. A. (1998). Orf virus encodes a    Shisler, J. L. & Moss, B. (2001). Molluscum contagiosum virus
homolog of the vaccinia virus interferon-resistance gene E3L. Virus        inhibitors of apoptosis: the MC159 v-FLIP protein blocks Fas-
Genes 17, 107–115.                                                         induced activation of procaspases and degradation of the related
McInnes, C. J., Deane, D., Haig, D., Percival, A., Thomson, J. & Wood,     MC160 protein. Virology 282, 14–25.
A. R. (2005). Glycosylation, disulfide bond formation, and the             Thurau, M., Everett, H., Tapernoux, M., Tschopp, J. & Thome, M.
presence of a WSXWS-like motif in the orf virus GIF protein are            (2006). The TRAF3-binding site of human molluscipox virus FLIP
critical for maintaining the integrity of binding to ovine granulocyte–    molecule MC159 is critical for its capacity to inhibit Fas-induced
macrophage colony-stimulating factor and interleukin-2. J Virol 79,        apoptosis. Cell Death Differ 13, 1577–1585.
11205–11213.                                                               Ueda, N., Wise, L. M., Stacker, S. A., Fleming, S. B. & Mercer, A. A.
Mercer, A. A., Ueda, N., Friederichs, S. M., Hofmann, K., Fraser, K. M.,   (2003). Pseudocowpox virus encodes a homolog of vascular
Bateman, T. & Fleming, S. B. (2006). Comparative analysis of genome        endothelial growth factor. Virology 305, 298–309.

http://vir.sgmjournals.org                                                                                                                   977
You can also read